RESPA Servicing Rule Assessment Report

Executive Summary

Mortgage servicers play a vital role within the mortgage market by undertaking the day-to-day management of mortgage loans. The Dodd-Frank Wall Street Reform and Consumer Protection Act (Dodd-Frank Act) amended the Real Estate Settlement Procedures Act (RESPA) to place certain new obligations on servicers and authorized the Bureau of Consumer Financial Protection (Bureau) to issue regulations found by the Bureau "to be appropriate to carry out the consumer protection purposes" of RESPA. The Bureau's initial RESPA mortgage servicing rule and certain amendments, which this report refers to collectively as the 2013 RESPA Servicing Rule or Rule, came into effect in January 2014.

Section 1022(d) of the Dodd-Frank Act requires the Bureau to conduct an assessment of each of its significant rules and orders and to publish a report of each assessment within five years of the effective date of the rule or order. The Bureau developed plans for assessments in 2015 and began work on this assessment in 2016. Pursuant to decisions made at that time, although this assessment addresses matters relating to the costs and benefits of the Rule, this report does not include a benefit-cost analysis of the Rule or parts of the Rule. For section 1022(d) assessments that the Bureau undertakes going forward, the Bureau in its discretion is reconsidering whether to include benefit-cost analysis in its assessments and its published reports. The Bureau expects that this report will help inform the Bureau's future policy decisions concerning mortgage servicing, including whether to commence a rulemaking proceeding to make the 2013 RESPA Servicing Rule more effective in protecting consumers, less burdensome to industry, or both.

Chapter 1 describes the requirement to conduct an assessment, provides an overview of the goals of the Rule, and discusses the methodology and data used in the report. Chapter 2 discusses the scope and requirements of the 2013 RESPA Servicing Rule, including relevant definitions and exceptions to the Rule. As further discussed in Chapters 1 and 2, many provisions of the Rule are intended to facilitate review of borrowers for foreclosure avoidance options (i.e., "loss mitigation" options) by requiring servicers to make certain disclosures and by requiring certain procedural steps and timelines when borrowers are applying for and being evaluated for loss mitigation options.

Chapter 3 provides an overview of certain trends in the mortgage servicing market that are relevant to understanding the effectiveness of the Rule. Starting after 2010 borrowers who became delinquent were increasingly likely to recover from their delinquency and t0 avoid foreclosure. This trend began before the Rule took effect in 2014 and has continued. In assessing the Rule, the Bureau attempts to distinguish between the effects of the Rule and the effects of pre-existing trends.

Chapters 4 and 5 analyze the effects of the Rule as a whole on borrowers and servicers, and the remaining chapters focus on particular provisions of the Rule. Chapter 4 looks at changes in the rates of foreclosures and of borrower recovery from delinquency before and after the Rule's effective date. Key findings include:

  • After controlling for trends in certain observable factors, loans that became delinquent were less likely to proceed to a foreclosure sale during the months after the Rule's effective date compared to months prior to the effective date. The Bureau estimates that if the Rule had not gone into effect in 2014, at least 26,000 additional borrowers who became delinquent that year would have experienced foreclosure within three years of becoming delinquent. However, these estimates assume that the data available to the Bureau are representative of the market as a whole, and there may be unobservable factors which cannot be controlled for in the analysis.
  • After controlling for trends in certain observable factors, loans that became delinquent were more likely to recover from delinquency (that is, to return to current status, including through a modification of the loan terms) following the Rule's effective date. The Bureau estimates that if the Rule had not gone into effect in 2014, at least 127,000 fewer borrowers who became delinquent that year would have recovered from delinquency within three years of becoming delinquent. The Report notes that this estimate is also subject to the same caveats stated above.

Chapter 5 analyzes overall effects of the Rule on servicing costs. Key findings include:

  • Data on servicing costs, from a trade association survey focused on large mortgage servicers, suggest that the cost of servicing mortgage loans increased substantially between 2008 and 2013. The estimated average annual cost of servicing performing loans increased from about $60 per loan in 2008 to about $160 per loan in 2013 and remained between $160 and $180 from 2014 to 2017. The estimated average annual cost of servicing loans in default increased from about $480 per loan in 2008 to about $2,410 per loan in 2013 and remained between $2,000 and $2,400 from 2014 to 2017. The
  • main increase in costs of servicing occurred before the Rule's January 2014 effective date and so is not attributable to the Rule. However, between 2009 and 2012 many servicers became subject to new servicing requirements, due to legal settlements or investor policies, that were similar to the requirements that were later incorporated into the Rule. Some of the increase in costs before 2014 could reflect the cost of complying with these earlier requirements.
  • In interviews, servicers described large one-time costs of implementing the Rule (including technology and personnel costs), with some larger servicers estimating onetime costs ranging widely from approximately $1.00 to $14.00 per loan. For context, industry estimates of average annual servicing costs are approximately $250 per loan in 2014. The estimated one-time costs of implementing the Rule are based on interviews with a small number of servicers, and the servicers emphasized that the estimates were approximate. If this range were applied to the approximately 53 million mortgage loans outstanding as of 2014, it would imply that the total one-time cost for industry to comply with the Rule was in the range of $53 million to $743 million. In interviews, servicers cited several factors that contributed to the costs of implementing the Rule. These factors included the complexity of implementing the large number of Rule provisions together with other mortgage rules that became effective at the same time as well as the challenge of coordinating with vendors and clients to make changes in different but interdependent servicing systems. As noted in the previous bullet, some servicers had already become subject to servicing standards that were similar to many of those mandated by the Rule, which reduced the costs to these servicers of coming into compliance with the Rule.
  • Some servicers reported significant ongoing costs of complying with the Rule. Larger servicers estimated that the Rule had increased annual costs by amounts ranging from approximately $3.00 per loan to more than $11.00 per loan. For context, industry estimates of average annual servicing costs are approximately $250 to $300 per loan since 2014. The estimated costs attributable to the Rule are based on interviews with a small number of servicers, and servicers emphasized that the estimates were approximate. If this range were applied to the approximately 52 million mortgage loans outstanding on average from 2014 to 2018, it would imply that the total ongoing cost for industry to comply with the Rule is in the range of $156 million to over $572 million annually. Small and mid-size servicers generally said that they were unable to estimate cost impacts of the Rule but that the Rule's requirements were consistent with their practices prior to the Rule's effective date. Sources of additional ongoing costs that servicers identified included the need for more robust control functions and higher personnel costs to support increased communication with delinquent borrowers.

Chapters 6 through 11 examine the available evidence with respect to the effects of particular provisions of the Rule. Many of the findings in these chapters are based on de-identified, loanlevel data on servicing operations that the Bureau obtained from seven servicers and from interviews with these and other servicers. These data generally reflect the experiences of these servicers and their borrowers in 2012 (pre-Rule) and 2015 (post-Rule) and are a unique source of information about the activities that were directly affected by the Rule. The data represent a large number of loans serviced by a range of different types of servicers. However, the data from these seven servicers may not be representative of all mortgage loans. Appendix C, which describes these data, notes other limitations.

Chapter 6 discusses the Rule's early intervention provisions, which require servicers to attempt live contact with delinquent borrowers and to provide written disclosures to delinquent borrowers. Key findings include:

  • Servicers interviewed generally said that the Rule's early intervention requirements were consistent with their practice prior to the Rule and did not require substantial operational changes other than tracking and monitoring compliance with the Rule's requirements. The Bureau did not determine the specific cost to servicers of complying with the early intervention requirements.
  • Consistent with what servicers reported, the data show little change in the timing of written notification to newly delinquent borrowers between the pre-Rule and post-Rule periods. The data do suggest that, post-Rule, delinquent borrowers are somewhat more likely than they were pre-Rule to start applying for loss mitigation earlier in delinquency. The share of borrowers initiating a loss mitigation application within six months of becoming 60 days delinquent increased from 39 percent in 2012 to 43 percent in 2015.

Chapter 7 discusses the Rule's continuity of contact provision for borrowers seeking loan modifications or other forms of loss mitigation. Chapter 7 also discusses other aspects of the loss mitigation provisions that apply before borrowers have completed loss mitigation applications. Chapter 8 discusses aspects of the Rule's loss mitigation provisions that apply when servicers are evaluating loss mitigation applications. Key findings include:

  • The data suggest that it took borrowers longer to go from initiating a loss mitigation application to completing the application in 2015 (post-Rule) than in 2012 (pre-Rule).
  • This may be because the Rule required servicers to define a complete application to be a more comprehensive package than what some servicers considered a complete application pre-Rule. Despite the longer time to complete applications once they had been initiated, borrowers who submitted complete applications in 2015 did so at a similar stage of their delinquency as borrowers who completed applications in 2012.
  • With respect to the Rule's loss mitigation provisions, many servicers said the most significant and costly changes they made were to comply with the Rule's requirements to: (1) provide a five-day acknowledgment notice for loss mitigation applications;
  • (2) evaluate borrowers for all available loss mitigation options at the same time; and (3) provide a decision letter that describes the outcome of an evaluation for all available options. Servicers generally said the Rule's loss mitigation process requirements other than the three listed above were consistent with their prior practice and did not require substantial operational changes other than tracking and monitoring compliance with the requirements. The Bureau did not determine the specific cost to servicers of complying with the Rule's loss mitigation provisions.
  • The data suggest that the time from borrower initiation of a loss mitigation application to short-sale offer increased in 2015 compared to 2012. In interviews, some servicers said this was likely due to the additional time required to collect the documents necessary to evaluate borrowers for all available loss mitigation options at the same time. Other servicers suggested that the increase in short-sale timelines may reflect an increase in the length of short-sale marketing periods post-Rule.
  • The data show that a larger share of borrowers who completed loss mitigation applications appealed the servicer's determination in 2015 compared to 2012. The proportion of appeals that were successful was lower post-Rule. The net effect was that there was no increase in the likelihood that a borrower whose application was denied successfully appealed that denial.

Chapter 9 discusses the Rule's provision that prohibits servicers generally from initiating a foreclosure proceeding until a borrower is more than 120 days delinquent. Chapter 9 also discusses the Rule's provision that prohibits servicers from taking certain steps toward a foreclosure sale for specified time periods once a borrower has submitted a complete loss mitigation application. Key findings include:

  • Consistent with the new requirement, the data show that in 2015 (post-Rule) servicers initiated foreclosure within the first 120 days far less often compared to 2012 (pre-Rule). The decrease in foreclosures initiated during that period was not offset by an increase in foreclosures initiated within the next several months, even after controlling for other factors. These facts suggest that the Rule's general prohibition on initiating foreclosure within the first 120 days of delinquency prevented rather than delayed foreclosures. Housing counselors surveyed generally said the foreclosure restrictions were the most important of the Rule's requirements in terms of helping their clients.
  • Servicers interviewed generally said they had to make significant changes to their foreclosure processes to ensure compliance with foreclosure restrictions and that these restrictions were among the more costly provisions of the Rule to implement. The Bureau did not determine the specific cost to servicers of complying with the Rule's foreclosure restriction provisions. Consistent with the Bureau's finding above, the servicers interviewed acknowledged that borrowers generally had benefited from these foreclosure restrictions. However, servicers also suggested that these restrictions may have disadvantaged some borrowers by causing them not to engage in the loss mitigation process until after they had fallen further behind on their mortgage payments.
  • Data indicate that a larger share of borrowers who completed loss mitigation applications in 2015 were able to avoid foreclosure than borrowers who completed loss mitigation applications in 2012. Post-Rule, loans had been delinquent for longer when servicers initiated foreclosure, compared to the pre-Rule period. Data suggest that the foreclosure restrictions have not increased the time it takes for servicers to go from initiating foreclosure to a sale.

Chapter 10 discusses the Rule's error resolution requirement. Key findings include:

  • In interviews, some servicers said that the Rule's error resolution provisions required them to make significant changes to how they track and respond to error assertions, whereas others said that for them the provisions required few changes. Servicers also said the Rule had little effect on whether borrowers submitted written error assertions. The Bureau did not determine the specific cost to servicers of complying with the Rule's error resolution provisions.
  • Data suggest that the rate of written error assertions per account fell by about one-half after the Rule's effective date compared to the prior three years. The decrease largely reflects a decline in error assertions related to loss mitigation. There was, however, considerable variation among servicers, with some servicers showing a substantial increase in written error assertions and others experiencing a substantial decrease. There is evidence that borrowers submitted fewer follow-up or repeat error assertions post-Rule, consistent with servicers becoming more responsive. On the other hand, there is no evidence that error assertions under the Rule generally changed the likelihood that borrowers' loss mitigation applications were approved.

Chapter 11 discusses the Rule's force-placed insurance requirements, which include a prohibition on charging borrowers for force-placed insurance unless the servicer has provided certain disclosures and complied with other requirements. Key findings include:

  • Servicers interviewed said that the Rule's requirements with respect to force-placed insurance were generally consistent with the force-placed insurance policies and procedures that they had in place before the Rule, so that the Rule's effects on borrowers and servicers were small. The Bureau did not determine the specific cost to servicers of complying with the Rule's force-placed insurance requirements.
  • The data show a moderate decrease post-Rule in the share of borrowers receiving forceplaced insurance, a trend that is consistent with the Rule's force-placed insurance requirements being effective but is also consistent with other potential explanations, such as changes in the insurance market that made it easier or less expensive for borrowers receiving a notice to obtain insurance.

1. Introduction

The mortgage market is the largest market for consumer financial products and services in the United States, with approximately $10.7 trillion in consumer mortgage loans outstanding. As described more fully later in this report, mortgage servicers play a vital role within the broader market by undertaking the day-to-day management of mortgage loans.

The number of borrowers who became delinquent on their mortgage loans increased sharply during the housing crisis that began in 2007. When borrowers become delinquent, servicers are typically responsible for attempting to collect overdue payments and, if borrowers are unable to repay, for resolving the delinquency either through a foreclosure sale or through an alternative to foreclosure. Foreclosure is costly, and when a borrower is unable to make required payments it may be in the interest of both the borrower and the loan's owner if the servicer works with the borrower to modify the loan agreement or find another alternative to foreclosure.

When the housing crisis began, servicers were faced with historically high numbers of delinquent mortgages, loan modification requests, and in-process foreclosures in their portfolios. The Bureau of Consumer Financial Protection (Bureau) said in 2013 that many of these servicers lacked the infrastructure, trained staff, controls, and procedures needed to manage effectively the flood of delinquent mortgages they were obligated to handle. Inadequate staffing and procedures led to a range of reported problems with servicing of delinquent loans, including some servicers misleading borrowers, failing to communicate with borrowers, losing or mishandling borrower-provided documents supporting loan modification requests, and generally providing inadequate service to delinquent borrowers. These reports led to efforts to introduce new standards for servicing delinquent mortgage loans, including new guidelines issued by major investors and federal government programs and settlement agreements with federal and state governmental authorities.6

The Dodd-Frank Wall Street Reform and Consumer Protection Act (Dodd-Frank Act) was signed into law on July 21, 2010.7 In the Dodd-Frank Act, Congress significantly amended the statutory requirements governing mortgage practices, addressing lending terms or practices that Congress concluded had contributed to or exacerbated the housing crisis.

In January 2013, the Bureau issued the "Mortgage Servicing Rules Under the Real Estate Settlement Procedures Act (Regulation X)" (2013 RESPA Servicing Final Rule).8 The 2013 RESPA Servicing Final Rule implemented certain new provisions of RESPA that were included in the Dodd-Frank Act, section 19(a) of RESPA, which allows the Bureau to "prescribe such rules and regulations…as may be necessary to achieve [RESPA's] purposes," and section 1022 of the Dodd-Frank Act, which authorizes the Bureau to adopt regulations "appropriate to carry out the consumer protection purposes" of RESPA. The Bureau amended the 2013 RESPA Servicing Final Rule several times before it took effect on January 10, 2014.9 This report refers to the 2013 RESPA Servicing Final Rule as so amended as the "2013 RESPA Servicing Rule" or the "Rule."

The Dodd-Frank Act's amendments to RESPA imposed new mortgage servicing requirements and prohibitions on servicers of federally related mortgage loans with respect to force-placed insurance, borrower assertions of error, and borrower requests for information. The Rule implemented these requirements. The Rule also included new requirements regarding servicing policies and procedures, early intervention with delinquent borrowers, continuity of contact with delinquent borrowers, and loss mitigation procedures which the Bureau found appropriate or necessary to carry out the consumer protection purposes of RESPA.

Section 1022(d) of the Dodd-Frank Act requires the Bureau to conduct an assessment of each significant rule or order adopted by the Bureau under Federal consumer financial law. As discussed further below, the Bureau has determined that, for purposes of section 1022(d), the 2013 RESPA Servicing Rule is a significant rule. Section 1022(d) also requires that the Bureau publish a report of the assessment within five years of the effective date of the significant rule or order. This document is the report of the Bureau's assessment of the 2013 RESPA Servicing Rule in accordance with section 1022(d). However, the Bureau also views this assessment and its report as a broader opportunity to advance the public's knowledge of this market and the effects of the 2013 RESPA Servicing Rule.

In May 2017, the Bureau published a Request for Information (or RFI) requesting public comment on its plans for assessing the Rule as well as certain recommendations and information that may be useful in conducting the planned assessment. The Bureau received approximately 40 comments in response to the RFI. The Bureau considered data and other relevant information provided by commenters, as well as comments on the assessment plan, as it conducted the assessment and prepared this report.

This report does not generally consider the potential effectiveness of alternative requirements on mortgage loan servicing that might have been or might be adopted, nor does it include proposals by the Bureau to modify any rules. The Bureau expects that the assessment findings made in this report and the public comments received in response to the RFI will help inform the Bureau's future policy decisions concerning mortgage servicing, including whether to commence a rulemaking proceeding to make the 2013 RESPA Servicing Rule more effective in protecting consumers, less burdensome to industry, or both. In future policy development, the Bureau expects to consider other public comments, including comments received in 2018 in response to a series of requests for information about Bureau activities.14 Those comments are not summarized in this report.

Finally, the Bureau's assessments pursuant to section 1022(d) of the Dodd-Frank Act are not part of any formal or informal rulemaking proceedings under the Administrative Procedure Act. This report does not represent legal interpretation, guidance, or advice of the Bureau and does not itself establish any binding obligations. Only the rules and their official interpretations (commentary) establish the definitive requirements.

1.1 Purpose and scope

1.1.1 Statutory requirement for assessments

Section 1022(d) of the Dodd-Frank Act requires the Bureau to conduct an assessment of each significant rule or order adopted by the Bureau under Federal consumer financial law. The assessment must address, among other relevant factors, the rule's effectiveness in meeting the purposes and objectives of title X of the Dodd-Frank Act and the specific goals stated by the Bureau. The assessment must reflect available evidence and any data that the Bureau reasonably may collect. Before publishing a report of its assessment, the Bureau must invite public comment on recommendations for modifying, expanding, or eliminating the significant rule or order.

The purposes and objectives of title X of the Dodd-Frank Act are set out in section 1021 of the Dodd-Frank Act. Pursuant to section 1021(a) of the Dodd-Frank Act, the purpose of the Bureau is to implement and, where applicable, enforce Federal consumer financial law consistently for the purpose of ensuring that all consumers have access to markets for consumer financial products and services and that markets for consumer financial products and services are fair, transparent, and competitive.18 The objectives of the Bureau are listed in section 1021(b) of the Dodd-Frank Act. Specifically, section 1021(b) provides that the Bureau is authorized to exercise its authorities under Federal consumer financial law for the purposes of ensuring that, with respect to consumer financial products and services:

  1. Consumers are provided with timely and understandable information to make responsible decisions about financial transactions;
  2. Consumers are protected from unfair, deceptive, or abusive acts and practices and from discrimination;
  3. Outdated, unnecessary, or unduly burdensome regulations are regularly identified and addressed in order to reduce unwarranted regulatory burdens;
  4. Federal consumer financial law is enforced consistently, without regard to the status of a person as a depository institution, in order to promote fair competition; and
  5. Markets for consumer financial products and services operate transparently and efficiently to facilitate access and innovation.

1.1.2 Goals of the Rule

The 2013 RESPA Servicing Rule in part implements section 1463 of the Dodd-Frank Act, which amended RESPA. Section 1463(a) imposed new mortgage servicing requirements and prohibitions under RESPA on servicers of federally related mortgage loans and also provided the Bureau authority to establish obligations on such servicers that are appropriate to carry out the consumer protection purposes of RESPA.

The servicing provisions included in the Dodd-Frank Act imposed new requirements with respect to force-placed insurance and established new timelines for responding to borrower assertions of error and borrower requests for information. The Rule implemented these requirements and included additional provisions that the Bureau found appropriate to carry out the consumer protection purposes of RESPA. These additional provisions largely relate to how servicers interact with delinquent borrowers and processes for considering borrowers for foreclosure alternatives. Specific provisions of the Rule address servicing policies and procedures, early intervention with delinquent borrowers, continuity of contact with delinquent borrowers, and loss mitigation procedures, as well as certain exemptions from coverage, all of which the Bureau found to be appropriate to carry out or necessary to achieve the purposes of RESPA and title X of the Dodd-Frank Act and prevent evasion of those laws.

The Bureau stated that the provisions of the Rule were intended to achieve the consumer protection purposes of RESPA. The Bureau said that, considered as a whole, RESPA, as amended by the Dodd-Frank Act, reflects at least two significant consumer protection purposes: (1) to establish requirements that ensure that servicers have a reasonable basis for undertaking actions that may harm borrowers; and (2) to establish servicers' duties to borrowers with respect to the servicing of federally related mortgage loans. The Bureau further stated that, specifically with respect to mortgage servicing, the consumer protection purposes of RESPA include:

  1. responding to borrower requests and complaints in a timely manner;
  2. maintaining and providing accurate information;
  3. helping borrowers avoid unwarranted or unnecessary costs and fees; and
  4. facilitating review for foreclosure avoidance options.

The Bureau further stated that each of the provisions adopted in the 2013 RESPA Servicing Rule was intended to achieve some or all of these purposes. The specific purposes of particular provisions of the Rule are detailed below in the chapters that discuss each provision. As discussed further below, many provisions of the Rule are aimed at facilitating review for foreclosure avoidance options by requiring servicers to make disclosures to borrowers about options for avoiding foreclosure, to help borrowers with the process of applying for those options, and to follow certain timelines and procedures in evaluating loss mitigation applications. The Bureau emphasized that its goal was not to achieve any particular target with respect to the number or speed of foreclosures, but rather to ensure that borrowers are protected from harm in the process of being evaluated for a loss mitigation option and in proceeding to foreclosure.

When the 2013 RESPA Servicing Final Rule was issued and in conjunction with later events, the Bureau released public statements that generally reiterated or elaborated on the purposes described above. The Bureau stated that the Rule provisions were designed to work together to help prevent avoidable foreclosure. The Bureau stated that borrowers generally should be made aware of their loss mitigation options, not face a foreclosure start until they have had time to explore these options, and not face a foreclosure sale before their applications have been evaluated.26 In addition, the Bureau stated that borrowers should not be subject to "runarounds" or surprised by payments, rates, or fees. Generally, the Bureau stated that practices prior to the Rule had too often led to unnecessary foreclosure.

In later remarks, the Bureau reiterated that problems in the mortgage servicing industry prior to the Rule may have caused too many borrowers to lose their homes. The Bureau has also stated that the Rule was issued to avoid the immediate and self-executing directives that Congress had enacted with respect to RESPA in title XIV of the Dodd-Frank Act, therefore reducing the burden on industry.

1.1.3 Determination that the 2013 RESPA Servicing Rule is a significant rule

The Bureau has determined that the 2013 RESPA Servicing Rule is a significant rule for purposes of section 1022(d) of the Dodd-Frank Act. As discussed in the RFI, the Bureau determined that the 2013 RESPA Servicing Rule is a significant rule partly on the basis of the estimated aggregate ongoing cost to industry of complying with the Rule.31 The Rule mandated several changes in mortgage servicing, including new disclosures for force-placed insurance, an expanded error resolution regime, and new servicing procedures and requirements for servicing delinquent loans, including mandated timelines and procedural rights in loss mitigation. These changes in turn required multiple changes in business operations, including adjustments in technology, training, and compliance. The Bureau noted in the preamble to the 2013 RESPA Servicing Final Rule that these changes would require servicers to modify systems and procedures and that the new requirements could require servicers to increase staff time devoted to certain activities and to hire more staff.32 Taking all of these factors into consideration, the Bureau formally determined in March 2017 that the 2013 RESPA Servicing Rule was "significant" for purposes of section 1022(d).33

In January 2013, the Bureau also issued separate "Mortgage Servicing Rules Under the Truth in Lending Act (Regulation Z)" (2013 TILA Servicing Final Rule).34 The 2013 TILA Servicing Final Rule became effective at the same time as the 2013 RESPA Servicing Rule. The Bureau has determined that the 2013 TILA Servicing Final Rule is not a significant rule (either individually or collectively with any amendments to the 2013 TILA Servicing Final Rule that took effect on January 10, 2014) for purposes of Dodd-Frank Act section 1022(d). That rule implemented the periodic statement requirement created by Dodd-Frank Act section 1420 and exempted small servicers from it. The 2013 TILA Servicing Final Rule also required a new initial adjustable-rate mortgage notice and revised certain existing disclosures and other servicing provisions under TILA. The estimated ongoing cost to servicers of complying with the 2013 TILA Servicing Final

Rule is small, as set forth in the Bureau's analysis of benefits and costs that accompanied the Rule.35 In particular, the 2013 TILA Servicing Final Rule generally modified important disclosures which consumers were already receiving.36 The 2013 TILA Servicing Final Rule's new disclosure requirements were intended to help certain groups of consumers make betterinformed decisions and were not expected to affect competition, innovation, or pricing in the mortgage market. These factors led the Bureau to conclude that the 2013 TILA Servicing Final Rule is not "significant" for purposes of section 1022(d).37

1.2 Methodology and plan for assessing effectiveness

In general, the Bureau methodology for the assessment consisted of three steps:

  • First, the Bureau considered the potential relevant effects of the Rule at a high level. These effects are the intended and unintended consequences of the Rule that would potentially be useful in evaluating whether the Rule, or a specific Rule requirement, furthers the goals of the Rule that were stated at the time of the rulemaking and, as relevant, the purposes and objectives of the Bureau or other relevant factors. The Bureau also considered the broader market context that could influence the effect of the Rule.
  • Second, the Bureau developed measures of the potential relevant effects and market conditions. The Bureau then collected available evidence and data that would allow the Bureau to compute these measures.
  • Third, the Bureau analyzed these measures and considered whether the Rule or specific Rule requirement furthered the goals of the Rule that were stated at the time of the rulemaking and, as relevant, the purposes and objectives of the Bureau or other relevant factors. In doing so, where possible, the Bureau compared the observed measures to what those measures would be under a counterfactual or "baseline."

Specifying a baseline against which to evaluate a rule's effects is necessary for both forecasting the future effects of proposed regulations and evaluating the historical effects of adopted regulations. When a regulation has already taken effect, however, it is often not possible to find firms or a part of the market that is neither subject to the rule nor indirectly affected by the rule—but is nevertheless subject to the same other determinants of prices, quantities, and other market outcomes—such that data about those firms or that part of the market provide a baseline for evaluating the effects of the rule. In general, retrospective analysis requires making a formal or informal forecast of the market absent a rule, or absent a specific provision of a rule, to serve as the baseline. Data limitations often make this difficult to do in practice.

For purposes of this assessment, the Bureau has generally used as a baseline the market absent the Rule as a whole or the specific provision being evaluated. When it is not possible to reliably estimate what a measure would have been under the baseline, the Bureau compares the relevant measure to its level before the Rule's effective date, thus capturing changes since the Rule took effect. Such changes are an imperfect measure of the effects of the Rule to the extent market changes that would have taken place even absent the Rule affect relevant measures.

In assessing the 2013 RESPA Servicing Rule, the Bureau collected information on a variety of measures that are relevant to the purposes and objectives of title X of the Dodd-Frank Act and goals of the Rule as described above. This report first describes overall trends in the market and in observable activities of servicers and borrowers that are associated with the Rule requirements. After providing this broader context of the market and the Rule, the report analyzes the effects of the Rule and its specific requirements.

Comments on the proposed plan received in response to the RFI informed the Bureau's plan for the assessment. These included comments on the assessment's focus, methodology, and sources of information. These comments are summarized in Appendix B. The Bureau revised its plan based on public comments and other information it received, although it was not possible to conduct all analyses proposed by commenters due to data and resource limitations.

The measures the Bureau analyzed include a number of servicing-related activities and outcomes, including:

  • Servicer activities undertaken to comply with the 2013 RESPA Servicing Rule, such as responding to loss mitigation applications and to borrower notices of error, including the timing of these actions;
  • Borrower activities, including: (a) use of the rights provided by the 2013 RESPA Servicing Rule, such as assertion of errors and use of appeals; and (b) borrower actions that may be prompted or enabled by the 2013 RESPA Servicing Rule, such as submission of loss mitigation applications or borrower verification of hazard insurance in response to a force-placed insurance notice sent by the servicer; and
  • Borrower outcomes that the 2013 RESPA Servicing Rule sought to affect, including, for example, fees and charges assessed and paid, incidence and severity of delinquency, how delinquency is resolved, and time to resolution of delinquency.

In addition to borrower outcomes that the Rule sought to affect, which are largely related to borrower delinquency, the assessment considers effects of the Rule on the costs of servicing. These costs affect servicers and may alter the cost and availability of mortgage credit for borrowers.

The Bureau's approach to analyzing these measures varies considerably depending on the specific question or Rule provision being addressed and the relevant data available. Much of the analysis in this report relies on comparing measures before and after the Rule took effect and focuses in particular on comparing measures as of 2012 and 2015. These periods were chosen to cover times shortly before and after the January 2014 effective date of the Rule but avoid the year immediately before and after the effective date, during which transition effects would confound the analysis.

As further discussed in Chapter 3, many aspects of the mortgage servicing market were changing rapidly in the years immediately before and after the Rule's effective date. The analysis in this report attempts, where possible, to isolate changes that resulted from the Rule by controlling for changes in observable factors. However, the analysis does not permit the Bureau to say definitively whether changes were caused by the Rule or by other factors in the market. As a result: (1) the Bureau is cautious about attributing observed changes to the Rule; (2) the report attempts to identify possible alternative explanations for changes between the pre-Rule and post-Rule period; and (3) the report attempts to present evidence in such a way that readers can gauge the strength of the evidence of particular effects.

1.3 Sources of data and information

This section briefly describes the major sources of information and data that the Bureau utilized and their limitations. Appendix C provides additional detail on some of these sources.

In conducting the assessment, the Bureau reviewed available public sources of data, including both publicly available loan-level data and published studies and reports pertaining to mortgage servicing. The Bureau's researchers also reviewed information the Bureau received in the normal course of its work and in response to the RFI. Based on its review of data from these sources, the Bureau concluded that the assessment required additional data and that these data were reasonable to collect. The Bureau collected additional data as described below.

In determining what additional data to obtain, the Bureau considered the strength of the evidence the source could provide about whether or not the Rule was effective as well as other factors, including cost to the Bureau and industry burden. The Bureau focused significant resources on collecting certain data and other information directly from seven servicers because other data sources that the Bureau identified did not include data on many of the specific activities affected by the Rule. The Bureau also considered collecting other types of data that it ultimately decided not to pursue. For example, the Bureau considered the possibility of conducting a large-scale consumer survey focused on the effects of the Rule, but decided that the significant cost and administrative challenges of conducting the study outweighed the likely benefits to the assessment of such a study.

1.3.1 Loan performance data from Black Knight and the GSEs

For some analyses, the Bureau used two sources of monthly, loan-level performance data. The first is the commercially available "McDash" data set from Black Knight (McDash Data), which as of February 2018 includes loan-level information on over 175 million mortgages and home equity loans. Loans in the McDash Data represented approximately 60 percent of loans in the market at the end of 2017. The second is publicly available data from the Federal National Mortgage Association (Fannie Mae) and the Federal Home Loan Mortgage Corporation (Freddie Mac), together known as the government-sponsored enterprises (GSEs). These data, referred to in this report as the "GSE Data," include information on fixed-rate, fully-amortizing mortgage loans that were acquired by the GSEs from 2000 to 2017, totaling approximately 48 million loans serviced during the period. Both of these sources provide data on loan characteristics and loan performance, including repayment status, foreclosure, and some information about loan modifications and other loss mitigation options. The McDash Data include more detailed information about loan modifications and other loss mitigation options obtained by borrowers, but this additional information is available for only approximately 60 percent of the loans covered and only starting in 2008.

An advantage of these sources is that they include information on a large number of loans from a broad set of servicers. However, they do not provide any data on some of the key activities that are governed by the Rule, such as initiating applications for loss mitigation, appealing loss mitigation decisions, or forced placement of insurance. The McDash Data and the GSE Data are discussed in more detail in Appendix C.

1.3.2 American Survey of Mortgage Borrowers

The American Survey of Mortgage Borrowers (ASMB) is part of the National Mortgage Database (NMDB®) program, which is jointly funded by the Federal Housing Finance Agency (FHFA) and the Bureau. The ASMB was first conducted in 2016, and subsequent waves of the survey took place in 2017 and 2018. The NMDB is a de-identified loan-level database of closed-end first-lien residential mortgages. It:

  1. is representative of the market as a whole;
  2. contains detailed, loan-level information on the terms and performance of mortgages, as well as certain characteristics of the associated borrowers and properties;
  3. is continually updated;
  4. has an historical component dating back before the financial crisis of 2008; and
  5. provides the set of borrowers from which ASMB respondents are chosen.

The information on borrowers and loans available is de-identified and does not include any directly identifying information such as borrower name, address, or Social Security number.

ASMB is designed to provide information, particularly related to delinquent mortgage borrowers, that is not available in the NMDB. Participation in the survey is voluntary, and its target universe is current and delinquent borrowers of closed-end first-lien residential mortgages from one year before the survey. To achieve this objective, ASMB draws its sample from mortgages that are part of NMDB. The ASMB is discussed in more detail in Appendix C.

1.3.3 Loan-level data on servicing operations from seven servicers

The Bureau collected de-identified loan-level data from seven mortgage servicers using its authority under section 1022(c)(4) of the Dodd-Frank Act (the Servicing Operations Data). The servicers were selected from among the largest 100 servicers and represent a range of servicer "types," including larger and smaller servicers within this range, depositories and nondepository institutions, and subservicers. The data request covered loans serviced at each entity as of January 2012 and January 2015. In total, the Servicing Operations Data cover more than three million loans, of which about a quarter became more than 30 days delinquent or had some loss mitigation activity during 2012 or 2015. The data include information about each loan's characteristics and repayment history as well as information about certain specific servicer and borrower activities, such as sending disclosures or completing loss mitigation applications.

The Servicing Operations Data are a unique source of information about the activities that were directly affected by the Rule. On the other hand, because they are drawn from the records of only seven servicers, the Servicing Operations Data may not be representative of data from all servicers. In addition, some of these seven servicers did not track consistently over time some of the data fields that the Bureau requested, which limits some of the analysis that can be done with the Servicing Operations Data. The Servicing Operations Data are discussed in more detail in Appendix C.

1.3.4 Survey of housing counselors and legal aid attorneys

The Bureau conducted an online voluntary survey of housing counselors and legal aid attorneys who help borrowers facing foreclosure (the Counselor Survey). The survey sought to understand counselors' and attorneys' perspectives on how the Rule affected borrowers they work with. Survey questions focused on whether, in the counselors' experience, particular provisions of the Rule are effective and reasons that those provisions are or are not effective. The Bureau sent the survey to approximately 800 organizations (comprising the population of HUD-certified housing counselors listed on HUD's website with an email address) and to a list of legal aid attorneys maintained by the Bureau. The survey received a total of 239 partial or complete responses, 140 from housing counselors and 99 from legal aid attorneys. As with any voluntary survey, respondents were self-selected, meaning that they may not be representative of the total population of HUD-certified housing counselors and legal aid attorneys. The survey is discussed in more detail in Appendix E.

1.3.5 Servicer interviews

The Bureau interviewed mortgage servicers and vendors about the effects of the Rule. These interviews included in-depth interviews with eight servicers, including all seven of those that provided Servicing Operations Data, and shorter interviews with 17 additional servicers and three vendors to the servicing industry. The interview questions focused on operational changes that servicers made as a result of the Rule, the perceived effects of the Rule on borrowers, the costs of complying with the Rule, and other effects of the Rule on servicers. In selecting servicers for interviews the Bureau sought a broad cross-section of servicers in terms of size and other characteristics. The servicers interviewed included 13 that reported servicing more than 50,000 mortgage loans (referred to in this report as "large servicers"), seven that reported servicing between 5,000 and 50,000 mortgage loans (mid-size servicers), and five that reported servicing 5,000 or fewer loans (small servicers). The information collected in interviews reflects a relatively small sample of servicers and in some cases reflects the individual experience of servicer personnel interviewed. However, it provides information on the effects of the Rule, including the costs to servicers of complying with the Rule, which cannot be obtained from administrative data. The interviews also provide context for understanding trends evident in the other data sources.

1.3.6 Consumer complaints submitted to the Bureau

The Bureau has handled more than 300,000 complaints from borrowers related to mortgage issues (including both servicing and originations) since it began accepting mortgage complaints on December 1, 2011. For the assessment, the Bureau analyzed mortgage servicing complaints taken by telephone or through the Bureau's website. Some analyses report findings from a randomly selected sample of complaints for which the Bureau manually reviewed and coded the complaint narratives and company responses.

In reviewing these data, the Bureau recognizes that it does not verify all the facts alleged in these complaints, but takes steps to confirm a commercial relationship between the borrower and the company. Additionally, only a small fraction of all borrowers submit complaints to the Bureau, and those who do are not necessarily representative of the larger borrower population as a whole. Further, the level of detail provided in complaint narratives and responses varies among both borrowers and servicers. That said, complaints are a rich source of information on the borrower's description of servicing practices. The complaint data used for this report are discussed in more detail in Appendix C.

1.3.7 MBA cost data

The Bureau obtained data on servicers' costs provided by the Mortgage Bankers Association (MBA). This includes data from 2009 through 2017 from the Annual Mortgage Bankers Performance Report (MBA Annual Performance Report Data) and summary information from the MBA's Servicing Operations Study and Forum (MBA Forum Data). The MBA Annual Performance Report Data are derived from data that independent mortgage companies and subsidiaries of banks, thrifts, and non-depository institutions provide to Fannie Mae, Freddie Mac, and the Government National Mortgage Association (Ginnie Mae) if they service loans owned or guaranteed by those entities. Although the sample of reporting companies changes over time, on average it includes approximately 140 servicers servicing approximately 9.5 million loans. The MBA Forum Data come from a survey of over 30 larger servicers that participate in MBA's Servicing Operations Study and Forum for Prime and Specialty Servicers. The MBA Forum Data are based on a smaller number of servicers but that number includes a large share of the largest servicers and represents both depository and non-depository servicers.

These sources are discussed further in section 5.1.1.

1.3.8 Evidence from comments received in response to the Request for Information

The Bureau also received approximately 40 public comments in response to the May 2017 RFI, many of which provided information about the effects of the 2013 RESPA Servicing Rule. Commenters reported on their own experiences, and provided information from surveys and other types of research, regarding the overall effects of the Rule and the effects of particular Rule requirements that are within the scope of the assessment. This information is summarized in Appendix B and incorporated into other parts of the report as appropriate. Overall:

  • Approximately 12 percent of the comments came from individual mortgage servicers that reported on their experiences with the 2013 RESPA Servicing Rule.
  • Approximately 30 percent of the comments came from trade associations that represent mortgage servicers or service providers to the servicing industry. These comments generally reported the experiences of the members of the organizations and in some cases reported the results of surveys of mortgage servicers.
  • Approximately 20 percent of comments came from legal aid organizations, publicinterest law firms, and consumer advocacy groups, one of which reported on a survey it conducted of housing counselors and others that work with delinquent mortgage borrowers.

The Bureau also received a number of comments that addressed other subjects on which the Bureau requested comment. These comments are also summarized in Appendix B.

1.4 Report overview

The next chapter provides an overview of the 2013 RESPA Servicing Rule. Chapter 3 discusses trends in the mortgage servicing market that are relevant to understanding data on the effectiveness of the Rule. Chapter 4 analyzes the effectiveness of the Rule as a whole, looking specifically at changes in the rates of foreclosures and of borrowers' recovery from delinquencies before and after the Rule's effective date based on the McDash and GSE Data. Chapter 5 analyzes overall effects of the Rule on servicing costs and the servicing market. Chapters 6 through 11 analyze the effectiveness of particular provisions of the Rule.

2. The 2013 RESPA Servicing Rule

Prior to the adoption of the Dodd-Frank Act, the mortgage servicing industry was subject to limited Federal consumer financial protection regulation. RESPA set forth basic protections with respect to mortgage servicing that were implemented by the U.S. Department of Housing and Urban Development (HUD). These included required disclosures at application concerning whether the lender intended to service the mortgage loan, disclosures upon an actual transfer of servicing rights, and disclosures and other substantive requirements for escrow account management. RESPA further required servicers to respond to ''qualified written requests''— certain written error resolution or information requests relating to the servicing of the borrower's mortgage loan.

Although prior to the Dodd-Frank Act RESPA did not impose many requirements on servicers, servicers were and still are required to navigate overlapping requirements governing their servicing responsibilities. In addition to federal law, servicers must consider the effects of state and even local regulation on mortgage servicing. Servicers also must comply with investor requirements to the extent they service loans owned or guaranteed by various types of entities. These include:

  1. servicing guidelines required by Fannie Mae, Freddie Mac, and Ginnie Mae;
  2. (2) government insured program guidelines issued by the Federal Housing Administration (FHA), Department of Veterans Affairs (VA), and the U.S. Department of Agriculture Rural Housing Service;
  3. (3) contractual agreements with investors (such as pooling and servicing agreements and subservicing contracts); and
  4. (4) bank or institution policies.

As discussed more fully in the next chapter, the number of borrowers who became delinquent on their mortgage loans and were seeking options to avoid foreclosure increased sharply during the housing crisis that started in 2007. Such options, generally referred to as "loss mitigation" options, can include modifying the loan by changing repayment terms to reduce monthly payments as well as other options to cure the delinquency without foreclosure. The housing crisis placed a spotlight on servicers' role in working with borrowers who were seeking loss mitigation options, because problems in the loss mitigation application process can have profound consequences for borrowers struggling to keep their homes.

The Dodd-Frank Act introduced new requirements for mortgage loan servicing and directed the Bureau to engage in rulemaking to implement those requirements. The Act also authorized the Bureau to adopt regulations "appropriate to carry out the consumer protection purposes" of RESPA. In addition, servicers, investors, guarantors, and state and federal regulators took a number of steps—both before and after the Dodd-Frank Act was enacted—to adjust servicer loss mitigation and foreclosure practices to help borrowers struggling with their mortgage payments and to address problems relating to evaluation of loss mitigation options. For example:

  • The U.S. Department of Treasury and HUD sponsored the Making Home Affordable program, which introduced several programs to encourage loss mitigation, most prominently the Home Affordable Modification Program (HAMP). HAMP provided servicers with financial incentives to modify loans and established servicing guidelines for federal government sponsored loss mitigation programs;
  • The FHFA directed Fannie Mae and Freddie Mac to align their guidelines for servicing delinquent mortgages they own or guarantee to improve servicing practices (the Servicing Alignment Initiative);
  • Prudential regulators, including the Board of Governors of the Federal Reserve System (Board) and the Office of the Comptroller of the Currency (OCC), undertook enforcement actions against major servicers, resulting in consent orders that imposed requirements on servicing practices;
  • In February 2012, 49 state attorneys general, joined by numerous federal agencies including the Bureau, entered into a National Mortgage Settlement (NMS) with the nation's five largest servicers at the time.52 The NMS imposed a number of obligations on servicers, many of which relate to loss mitigation evaluations and associated practices.53 Several other large servicers later became subject to similar settlements; and
  • States adopted regulations relating to mortgage servicing and foreclosure processing, including requiring servicers to evaluate borrowers for loss mitigation options under certain circumstances.54

Many of these requirements and guidelines coalesced around a common set of best practices for servicing delinquent loans—for example, practices for reaching out to delinquent borrowers regarding loss mitigation options, making personnel available to borrowers to assist with loss mitigation applications, and certain protections from foreclosure for borrowers seeking loss mitigation. In finalizing the 2013 RESPA Servicing Final Rule, the Bureau incorporated many of these concepts, as it sought consistency with servicing standards embodied in the NMS, FHFA's Servicing Alignment Initiative, HAMP guidelines, federal regulatory agency consent orders, and state law mortgage servicing statutory requirements.

The 2013 RESPA Servicing Rule addressed six major topics, which are summarized below. Most of these provisions have parallels in the NMS and other servicing standards noted above that pre-dated the Rule. In general, this report analyzes the effectiveness of the Rule's provisions summarized below in later chapters, and those chapters include more detailed descriptions of the provisions themselves and related servicing standards.55

Early intervention with delinquent borrowers. The Rule generally requires servicers to establish or make good faith efforts to establish live contact with borrowers by the 36th day of their delinquency (for each billing cycle for which a payment sufficient to cover principal, interest, and, if applicable, escrow is due and unpaid) and to promptly inform such borrowers, where appropriate, that loss mitigation options may be available. In addition, servicers must generally provide borrowers a written notice with information about available loss mitigation options by the 45th day of their delinquency.

Continuity of contact with delinquent borrowers.57 The Rule requires servicers to maintain reasonable policies and procedures to provide delinquent borrowers with access to personnel to assist them with loss mitigation options where applicable.

Loss mitigation procedures. The Rule requires servicers to follow specified loss mitigationrelated procedures for a mortgage loan secured by a borrower's principal residence. Servicers generally must provide a written notice acknowledging receipt of a borrower's loss mitigation application within five days and exercise reasonable diligence in obtaining documents and information to complete the application. For a complete loss mitigation application received more than 37 days before a foreclosure sale, the Rule generally requires the servicer to evaluate the borrower, within 30 days of receiving the application, for all loss mitigation options available to the borrower in accordance with the investor's eligibility rules. Servicers must permit borrowers to appeal denials for loan modifications under some circumstances. The Rule also prohibits a servicer from making the first notice or filing required by applicable law for any judicial or non-judicial foreclosure process until a borrower's mortgage loan obligation is more than 120 days delinquent and places certain restrictions on "dual tracking," where a servicer is processing a consumer's loss mitigation application at the same time that it advances the foreclosure process.

Error resolution and information requests. The Rule requires servicers to comply with certain error resolution procedures for written notices of error from borrowers relating to the servicing of a mortgage loan. Servicers generally are required to acknowledge the notice of error within five days of receipt and to investigate and respond in writing within 30 days of receipt, either correcting the error or notifying the borrower that no error occurred. Similar procedures and deadlines generally apply to servicer acknowledgment of and response to borrower written requests for information.

Force-placed insurance. The Rule prohibits servicers from charging a borrower for forceplaced insurance coverage unless the servicer has a reasonable basis to believe the borrower has failed to maintain hazard insurance as required by their loan agreement. If the borrower has an escrow account for the payment of hazard insurance premiums, the servicer is prohibited from obtaining force-placed insurance if the servicer can continue the borrower's homeowner insurance, even if the servicer needs to advance funds to the borrower's escrow account to do so. The Rule also requires servicers to send two notices before charging the borrower for forceplaced insurance coverage.

General servicing policies, procedures, and requirements. The Rule requires servicers to establish policies and procedures reasonably designed to achieve objectives specified in the Rule.

Many of the Rule's requirements do not apply to small servicers, which are generally defined as servicers that service 5,000 mortgage loans or fewer and only service mortgage loans the servicer or an affiliate owns or originated. Small servicers are exempt from: early intervention and continuity of contact requirements; most loss mitigation procedure requirements; certain requirements relating to obtaining force-placed insurance; and the provisions relating to general servicing policies, procedures, and requirements.

3. Market overview

This chapter provides context for the Bureau's assessment of the 2013 RESPA Servicing Rule by documenting the significant changes in the mortgage markets immediately before and after the Rule became effective in January 2014. The first section of this chapter describes the role of mortgage servicers and provides background on how servicers responded to the increase in delinquent mortgage loans during the housing crisis that started in 2007. The second section describes, primarily based on GSE and McDash Data, trends from 2004 to 2017 in how mortgage delinquencies were resolved.

3.1 Background on mortgage servicing

Servicers perform tasks such as billing borrowers for amounts due, collecting payments, disbursing funds to owners or investors, maintaining and disbursing funds from escrow accounts, credit bureau reporting, and maintaining vacant and abandoned properties. If borrowers become delinquent, servicers are typically responsible for pursuing collection and, if borrowers are unable to repay, servicers are generally responsible for managing the foreclosure process and for loss mitigation activities.

Mortgage servicing is performed by banks, thrifts, credit unions, and nonbank servicers under a variety of business models. In some cases, a loan's servicer is the creditor that originated the mortgage loan. Such servicers may continue to hold the loan in their portfolio or may sell the ownership of the loan but continue servicing the loan, either by retaining the mortgage servicing rights (MSRs) or through a subservicing agreement with the new owner of the MSRs. In other cases, servicers have no role at all in origination or loan ownership, but rather purchase MSRs for securitized loans or are hired as a subservicer to service loans on behalf of the owner of the MSRs. A subservicer does not own MSRs for loans it subservices but performs servicing functions in exchange for a fee; a subservicer may service such loans in its own name or in the name of the owner of the MSRs. Because MSRs can be transferred, a borrower's loan may be serviced by multiple servicers during the life of the loan. All servicers, including subservicers, are subject to the Rule, although small servicers are exempt from certain provisions.

Much of the Rule addresses how servicers interact with borrowers who become delinquent. This section provides background on the increase in delinquent loans that accompanied the housing crisis, the role of servicers when borrowers become delinquent, and economic research that provides evidence on the efficiency of the market for mortgage servicing during the period following the crisis.

3.1.1 Trends in delinquencies and foreclosures

When borrowers are unable to continue making their mortgage payments, they generally rely on their servicers to help them determine whether there are options for resolving the delinquency through a loan modification or other loss mitigation option. This aspect of servicing took on particular prominence when mortgage loan delinquencies increased sharply starting in 2007.

Figure 1 below shows the increase in both the rate at which borrowers became delinquent and in the number of loans that were seriously delinquent following the start of the housing crisis in 2007. The number of loans that transitioned from current to 90 or more days past due (seriously delinquent) increased from less than 0.5 percent of all mortgage loans at the end of 2006 to a peak of 2 percent at the end of 2008. The figure also shows that by the fourth quarter of 2009 about 4 percent of loans were in the foreclosure process and about 5 percent of loans were not in foreclosure but were at least 90 days past due. The rate of new serious delinquencies peaked in the first quarter of 2009 and returned to 2006 levels by the start of 2014, but the total number of loans in these seriously delinquent categories peaked a bit later and remained well above pre-crisis levels into 2017 as loans that had become delinquent in earlier periods worked their way toward resolution.

FIGURE 1: NEWLY DELINQUENT LOANS AND LOANS SERIOUSLY DELINQUENT OR IN FORECLOSURE PROCESS (NATIONAL DELINQUENCY SURVEY, CONSUMER CREDIT PANEL)

Borrowers can become delinquent on their mortgages for a number of reasons, often related to changing life circumstances. Job loss, health problems, divorce, or other life events can all reduce household income and increase expenses, making it difficult or impossible for borrowers to make monthly mortgage payments. In the 2007 to 2009 period, a number of factors in the broader economy contributed to the sharp increase in delinquency rates.

  • The economic downturn meant many consumers lost jobs and income, so that payments that had been manageable became unaffordable. From January 2007 to January 2010, U.S. unemployment rates increased from 4.6 percent to 9.8 percent (see Figure 2).
  • A widespread decline in property values reduced borrowers' equity in their homes, in some cases meaning borrowers owed more than the amount for which their homes could be sold (see Figure 3). This limited their ability to refinance the loan or to pay it off by selling the home and may have reduced the incentives of some borrowers to continue making payments on their mortgage loans.
  • In the years leading up to the crisis, there was a substantial rise in the number of higher credit risk borrowers who were able to obtain mortgage loans. Some loans obtained by these borrowers required mortgage payments that proved to be unsustainable given borrowers' financial circumstances.

These factors not only contributed to borrowers becoming delinquent but also limited servicers' and borrowers' options for resolving a delinquency. As discussed in the next section, before the crisis most borrowers who became at least 60 days delinquent cured the delinquency on their own, either by returning to current status or by prepaying the loan (which could follow either a sale of the home or refinancing the loan). During the crisis, the factors noted above meant it was harder for borrowers either to resume making payments or to sell the home.

FIGURE 2: CIVILIAN UNEMPLOYMENT RATE, JANUARY 2000 TO OCTOBER 2018 (US BUREAU OF LABOR STATISTICS)

FIGURE 3: NATIONAL HOME PRICE INDEX, JANUARY 2000 TO JULY 2018 (BLACK KNIGHT, CASESHILLER)

3.1.2 Loss mitigation options

Faced with a borrower who is delinquent and unable to make agreed payments, servicers may foreclose or may offer the borrower options for loss mitigation. Because foreclosure is a costly process, loss mitigation may in some cases be better for both the investor and the borrower. Loss mitigation options generally fall into the following categories:

  • Retention options are intended to return the borrower to a current status on the loan while retaining ownership of the home. One type of option is a modification of the loan terms. Loan modifications may involve extending the term of the loan, reducing the interest rate, forbearing or reducing the principal balance, or some combination of these changes. Another option is a repayment plan, sometimes combined with short-term forbearance, which is designed to cure the delinquency without modifying the loan terms.
  • Non-retention options permit the borrower to pay off the loan and transfer the property without a foreclosure sale. In a short sale, the lien holder permits the home to be sold for less than the borrower owes on the mortgage. In a deed-in-lieu of foreclosure transaction, the borrower voluntarily turns over ownership of his or her home to the servicer (acting on behalf of the investor) and avoids the foreclosure process.

A borrower seeking loss mitigation generally applies to be considered for loss mitigation options and is evaluated by the servicer. If servicers do not own the loans they service, contracts between the servicer and the mortgage loan owner generally set forth the steps the servicer can or should take when loans become delinquent, including evaluating offers of loss mitigation. In some cases investors must approve loss mitigation offers, and in other cases investors may have given the servicer discretion, subject to investor guidelines, to offer loss mitigation options on their behalf.

3.1.3 Relevant economic research on delinquent loan servicing

Several economic studies have examined topics related to loss mitigation and what determines outcomes for borrowers who become delinquent on their mortgage loans.

One strand of research has found that, during the pre-Rule period, delinquent borrowers who appear to have been in similar financial situations but had different servicers experienced very different outcomes. If servicers make decisions about loss mitigation that are in the best interest of the loans' investors, then one would expect similarly situated borrowers to obtain the same outcomes, whether that be foreclosure or a loss mitigation option, at least in cases where the investor is the same. This research has shown, however, that substantially similar borrowers whose loans were owned by the same investor received different loss mitigation decisions from different servicers. Research focusing on the HAMP program, analyzing loans over the period of 2008 to 2012, found that some servicers modified loans at half the rate of other servicers, even after taking into account borrower, loan, and investor characteristics. Another study, examining data from 2009 to 2012 for securitized, private label loans, found significant differences in delinquency cure rates after controlling for observable factors. This study also found that loans with nonbank servicers were more likely than loans with bank servicers to cure, were more likely to be modified with principal decreases, and were more likely to undergo second modifications.

Some of this research found that these differences in loan modification results depended on the characteristics of the servicing staff and the technology used by the servicer. In particular, the likelihood of approving modifications was positively correlated with the number of staff and the number of staff training hours. The likelihood of approving modifications was negatively correlated with staff workload volume and indicators of poor technology (such as the percentage of dropped calls and time callers spend on hold). This suggests that some servicers had invested more in the capacity to manage the loss mitigation process for delinquent consumers.

Some research has examined the incentives of servicers to "wait and see" whether a borrower will recover from a delinquency rather than either modifying the loan or pursuing foreclosure. Examples of factors that may affect servicers' incentives to wait include whether the servicer must make disclosures about their loss mitigation and foreclosure activity and whether the servicer owns a second lien on the same property. This research shows that a servicer's decision about whether and when to initiate foreclosure can be affected by a number of factors that may be unrelated to the 2013 RESPA Servicing Rule.

Research on the effectiveness of the HAMP program in helping borrowers recover from delinquency found that the effect of a HAMP modification increased following program changes in mid-2010. These program changes included some features similar to those included in the Rule, including requiring some servicers to provide borrowers with a single point of contact to help with loss mitigation applications.

3.2 Trends in outcomes of delinquency spells

This section describes overall trends in loss mitigation and foreclosure from the pre-crisis period to 2017.

Figure 4 shows the total number of GSE loans that were modified, foreclosed upon, or ended the year at least 12 months delinquent during each year from 2001 to 2017. The figure shows that all three categories increased dramatically starting in 2008 and subsequently declined.

Figure 4 also highlights how the relative frequency of the three categories shifted over this period. Prior to 2008, foreclosures were more common than modifications. Starting in 2009, the number of modifications was greater than or very close to the number of foreclosures completed. A major factor in this change was the federal government's introduction of several programs to help borrowers avoid foreclosure by encouraging loan modifications and other forms of loss mitigation. The most prominent of these was HAMP. HAMP required a uniform loan modification process for participating servicers and provided financial incentives for servicers and homeowners to modify loans under the program. GSE and government-insured loan programs also introduced changes to their servicing policies to encourage loss mitigation.

FIGURE 4: NUMBER OF GSE LOANS MODIFIED, FORECLOSED, AND AT LEAST 12 MONTHS DELINQUENT BY YEAR (GSE DATA)

Figure 4 also illustrates that between 2008 and 2011 there was a large increase in the number of loans that ended each year at least 12 months delinquent. Until 2008, this was relatively uncommon, meaning that most loans that began the year delinquent and did not cure on their own ended the year either foreclosed on or modified. This changed beginning in 2009, when much larger shares of loans remained at least 12 months delinquent at the end of each year.

One likely factor in the increase in year-end delinquency is a large increase in foreclosure timelines. As further discussed below in section 3.3, the large increase in delinquent loans was followed by increases in the time it took to move delinquent loans through the foreclosure process. After 2008, many loans that were more than 12 months delinquent at year end were likely at some stage of foreclosure proceedings. At the same time, HAMP and other changes to government policies and the fall in housing prices may have increased the benefit of modification relative to foreclosure.

By the time the Rule became effective at the start of 2014, these trends had started to moderate, but the numbers of foreclosures, modifications, and loans at least 12 months delinquent remained well above pre-crisis levels.

Figure 5 considers how delinquencies are resolved over time, depending on the year in which the loan first became 60 days or more delinquent (i.e., loans that have missed two payments as of the end of a month). For loans that became 60 days delinquent at some point during each year, the chart shows the share of loans in each status category 18 months after the loans became 60 days delinquent.

FIGURE 5: LOAN STATUS 18 MONTHS AFTER FIRST 60 DAY DELINQUENCY BY YEAR OF DELINQUENCY (GSE DATA)

Figure 5 shows that during the 2004 to 2006 period, most loans that became 60 days delinquent cured within 18 months either by becoming current or by prepayment, presumably through sale or refinancing. This began to change in 2007, and in 2008 and 2009 only about 20 percent of loans either became current (without modification) or were prepaid.

Starting in 2007 the share of delinquencies that ended in foreclosure within 18 months increased, but the share of loans that had been modified increased more significantly. During 2001 to 2006 less than 8 percent of loans that became 60 days delinquent were modified within 18 months, but this grew to 15 percent for loans becoming delinquent in 2008 and 29 percent for those becoming delinquent in 2009.

For loans that became delinquent in 2007 and 2008, when modifications were becoming more prevalent, a large fraction of loans that were modified were again delinquent 18 months after the original delinquency. This suggests that many modifications made in this period did not result in sustainable payments. However, this fraction changed somewhat for loans first becoming delinquent in 2009 onward, with roughly 80 percent of modified loans remaining current 18 months after the initial delinquency.

The figures above indicate that, following the crisis, many loans remained delinquent for a relatively long time before the delinquency was resolved. Many loans were modified when they had been delinquent for one year or more. Figure 6 uses the McDash Data to show the number of modifications completed each year by the number of missed payments when the loan was modified. The figure illustrates that starting in 2010 a substantial fraction of loans modified were behind by at least 12 months in payments, and from 2012 to 2016, more than 10 percent of the loans modified in each year had been delinquent for at least two years. Modification so far into a delinquency could mean that new modification options became available later in a delinquency, that servicers were reconsidering borrowers for options that had been previously denied, or that borrowers were seeking modifications later in delinquency (perhaps as foreclosure sale dates approached).

FIGURE 6: NUMBER OF MONTHS DELINQUENT WHEN LOANS ARE MODIFIED, BY MODIFICATION YEAR (MCDASH DATA)

Figure 7, below, uses the McDash Data to show the number of short sales and deeds-in-lieu of foreclosure completed by year. These types of non-retention foreclosure alternatives peaked in 2012, a few years after the peak in modifications. The frequency of these options fell from its peak even more sharply than that of modifications, with the 2014 level at about 27 percent of the 2012 level.

FIGURE 7: SHORT SALES AND DEEDS-IN-LIEU OF FORECLOSURE AS A SHARE OF TOTAL LOANS OUTSTANDING BY YEAR (MCDASH DATA)

3.3 Foreclosure timelines

The length of the foreclosure process has grown substantially since the crisis. For borrowers, longer timelines can provide more time to resolve a delinquency or find alternative housing, but can also delay the resolution of the situation and the opportunity to put the delinquency behind them.79 For servicers and investors, longer timelines can make the foreclosure process more costly, delaying the recovery of amounts owed by the borrower and potentially increasing costs from legal fees and property maintenance. Communities may also be harmed by longer timelines, particularly when properties remain vacant during the foreclosure process. Longer timelines affect the incentives of both parties prior to a foreclosure sale: borrowers may have reduced incentives to resolve a delinquency if foreclosure is further away, whereas servicers may have greater incentives to seek alternatives to foreclosure if a foreclosure sale takes longer or is more costly to accomplish.

Foreclosure processes, including timelines, are governed by state and local law and vary significantly among states. One important factor is whether an individual state follows a "judicial" foreclosure process, in which a court must review the case before a foreclosure sale can be completed, or a "non-judicial" foreclosure process, in which the lienholder can pursue a foreclosure sale without having to petition a court.

Figure 8 below shows trends in the time between a borrower's last paid installment and completion of a foreclosure sale for foreclosures completed each month, separately for judicial and non-judicial states. In judicial states, the median foreclosure took about 12 months in the earliest years shown. Median foreclosure times increased beginning in 2009, reaching about 30 months at the start of 2013 and remaining near that level before declining after mid-2015. Many foreclosures took much longer than the median: for most of the 2014 to 2017 period, more than 25 percent of foreclosures completed in judicial states took longer than four years. Timelines in non-judicial states are shorter but similarly show an increase in the median timeline and a larger increase in the 75th percentile over the time period.

FIGURE 8: MONTHS FROM LAST PAID INSTALLMENT TO FORECLOSURE SALE, BY MONTH OF FORECLOSURE SALE (BLACK KNIGHT MCDASH)

Even among states that follow either judicial or non-judicial foreclosure procedures, states differ in their foreclosure laws and procedures, and some of the variation in foreclosure timelines illustrated in Figure 8 is driven by differences across state systems. The next figure shows median time from last paid installment to foreclosure sale for the states with the highest number of foreclosures among judicial states. Median timelines that were fairly similar in the 2006 to 2008 period have diverged substantially since then. In New York and New Jersey, median timelines continued to increase for foreclosures completed through 2017, exceeding five years in many months, whereas in Florida, Illinois, and Ohio median timelines declined after reaching a peak for foreclosures that were completed in the 2012 to 2013 period.

FIGURE 9: MONTHS FROM LAST PAID INSTALLMENT TO FORECLOSURE SALE, BY MONTH AND BY STATE (BLACK KNIGHT MCDASH)

3.4 Other servicing market trends

In the years following the housing crisis there have been changes in the types of firms servicing mortgages, which some have attributed to increased regulation of servicing. Servicers that were not depository institutions ("non-depository servicers") have played a growing role in recent years: in 2010, only one of the 10 largest servicers was a non-depository, whereas at the end of 2013 five of 10 were non-depositories and at the end of 2017 six of the 10 largest servicers were non-depositories. Among the top 20 servicers in each year, the share of outstanding mortgage principal balances serviced by non-depositories grew from 12 percent in 2009 to 31 percent in 2014. Some large bank servicers have significantly reduced their mortgage servicing in recent years.

Figure 10 presents the market share of the largest four and the largest 15 servicers since 2004. Concentration increased in 2007 and 2008 as large servicers acquired servicers or servicing portfolios at the start of the housing crisis. Concentration subsequently fell gradually and by 2017 remained a bit higher than the 2006 level.

FIGURE 10: MARKET SHARE OF LARGEST MORTGAGE SERVICERS, 2004 TO 2017 (INSIDE MORTGAGE FINANCE)

In recent years, subservicing has become more prevalent in the mortgage servicing market. The share of total mortgage loans outstanding that is subserviced increased from 9 percent in the final quarter of 2013 to 22 percent in the third quarter of 2018.

3.5 Consumer complaints about mortgage servicing

The Bureau received approximately 300,000 mortgage complaints between December 1, 2011, and December 31, 2017. When consumers submit complaints, the Bureau's complaint form prompts them to select the consumer financial product or service as well as the type of problem they are having. These selections include issues that may arise during loan origination (i.e., applying for a mortgage or refinancing an existing mortgage, closing on a mortgage) or during loan servicing (i.e., struggling to pay mortgage, trouble during payment process). Figure 11 shows the volume of mortgage complaints, by issue, received between 2012 through 2017. In each year, a large majority of complaints regard servicing issues rather than origination and of these, a large fraction relate to struggles paying a mortgage, such as complaints regarding loan modification or foreclosure.

FIGURE 11: MORTGAGE COMPLAINTS BY ISSUE (BUREAU COMPLAINT DATA)

The Bureau searched keywords in consumers' narrative descriptions in mortgage servicing complaints to identify those related to loss mitigation. The Bureau received approximately 112,000 loss mitigation complaints from 2012 to 2017. Figure 12 reports these results. In 2012 and 2013 the majority of servicing complaints related to loss mitigation. Since then, loss mitigation complaints have been declining both in total and as a share of mortgage servicing complaints, whereas there is no particular trend in other mortgage servicing complaints. The different patterns in these two trends suggest that the decline in loss mitigation-related complaints is not driven by an overall decline in the likelihood that borrowers choose to complain to the Bureau about servicing issues.

The drop in complaints about loss mitigation from about 30,000 in 2013 to roughly 10,000 in 2016 is likely due in part to the drop in the number of delinquent loans over this period. As shown in Figure 1 above, approximately 5.4 percent of loans were seriously delinquent or in foreclosure in 2013, and this fell to about 3.1 percent in 2016. In the same timeframe the number of loss mitigation-related complaints fell by about 66 percent, more than the approximately 43 percent drop in the number of loans seriously delinquent or in foreclosure. While trends in serious delinquency may not be representative of trends in the number of borrowers seeking loss mitigation, the drop in the number of complaints related to loss mitigation nevertheless suggests that, in the years after the Rule took effect, borrowers who were in distress were less likely to contact the Bureau with complaints related to loss mitigation.

FIGURE 12: LOSS MITIGATION COMPLAINTS AND OTHER MORTGAGE SERVICING COMPLAINTS (BUREAU COMPLAINT DATA)

The Bureau accepts complaint submissions from individuals as wells as from agents, trustees, and others acting on behalf of an individual. The complaint form indicates whether the complaint came from an individual or an authorized third party on behalf of the borrower. Figure 13 shows that third-party complaints from attorneys and housing counselors increased when the Rule took effect in 2014. One possible explanation for the increase is that, with the Rule in place, attorneys had more reason to expect that servicers would address consumer complaints submitted to the Bureau, particularly those that raise the possibility of Rule violations. It is also possible that borrowers were more likely to work with an attorney or housing counselor after the Rule took effect.

FIGURE 13: COMPLAINTS SUBMITTED ON BEHALF OF SOMEONE ELSE, BY RELATIONSHIP TO BORROWER (BUREAU COMPLAINT DATA)

3.6 Compliance with the Rule

The Bureau began examining mortgage servicers for compliance with the Rule shortly after the January 2014 effective date. Supervisory examinations of mortgage servicers since 2014 have generally focused on reviewing for compliance with the Rule and for unfair, deceptive, and abusive acts or practices. The Bureau discusses in its Supervisory Highlights patterns and trends in exams that have taken place, and the Bureau specifically focused on mortgage servicing in a June 2016 special edition.

The Bureau's examinations have uncovered mixed levels of compliance with the Rule across the industry. While the servicing market has made investments in compliance, the magnitude and persistence of compliance challenges since 2014, particularly in loss mitigation communications, show that those investments have not always been sufficient to prevent Rule violations across the marketplace.

With respect to the following Rule provisions, the Bureau's examinations identified violations by one or more servicers, including:

Completing Loss Mitigation Applications
  • Failing to send loss mitigation acknowledgement within five days after receiving a borrower's application for loss mitigation, as required by the Rule, or failing to send any loss mitigation acknowledgement.
  • Requesting documents as part of the application that were inapplicable to borrower circumstances and unnecessary for evaluating borrowers for loss mitigation.
  • Failing to state additional documents and information required for borrowers to complete an application, such as income and tax forms that the servicer's internal records showed were necessary at that time.
  • Failing to include any reasonable date by which borrowers must return additional documents and information to complete an application.
  • Giving borrowers 30 days to submit additional documents, but then denying borrowers' applications before 30 days.
  • Failing to send loss mitigation acknowledgement notices to borrowers requesting shortterm payment relief.
  • Failing to identify and process material submitted by borrowers to complete an application.
Loss Mitigation Evaluations & Appeals
  • Failing to maintain policies and procedures that were designed to identify with specificity all loss mitigation options for which a borrower may be eligible. One or more servicers sent letters to borrowers soliciting loss mitigation applications when internal records showed that the borrowers were not eligible for any loss mitigation option.
  • Failing to maintain policies and procedures to properly evaluate a loss mitigation application for all options that the borrower may be eligible for based on the loan owner's requirements. Instead, the borrower was evaluated only for options that the servicer(s) preselected.
  • Sending loss mitigation denial notices that failed to communicate a borrower's specific right to appeal.
  • Treating borrower self-employed gross income as net income when evaluating loss mitigation applications, which may have led to less affordable modifications.
  • Failing to state the correct reasons(s) for denying a loan modification.
  • Failing to maintain policies and procedures reasonably designed to properly evaluate a borrower who submits a loss mitigation application for all loss mitigation options for which the borrower may be eligible.
Foreclosure Restrictions
  • Failing to classify loss mitigation applications as facially complete after receiving the documents and information requested in the loss mitigation acknowledgement notice, and failing to afford these eligible consumers with foreclosure protections.
  • Failing to maintain policies and procedures to facilitate sharing accurate and current information about the status of any evaluation of a borrower's loss mitigation application and foreclosure proceedings. For example, one or more servicer(s)' foreclosure attorneys sent a foreclosure referral letter to a borrower after the borrower had entered into a loss mitigation agreement.
Other Rule Provisions
  • Failing to include a statement that borrowers should consider contacting servicers of any other mortgage loans secured by the same property to discuss available loss mitigation.
  • Failing to maintain policies and procedures related to the oversight of service providers, or the facilitation of periodic reviews of service providers.
  • Failing to maintain policies and procedures reasonably designed to provide a borrower with accurate and timely information and documents in response to the borrower's request for information with respect to the borrower's mortgage loan. One or more servicers failed to provide information and loss mitigation application forms to a substantial number of borrowers who called in to request such information.
  • Failing to maintain policies and procedures reasonably designed to facilitate the sharing of loss mitigation information when necessary.

To date, the Bureau has announced five enforcement actions against mortgage servicers that include violations pertaining to the Rule. As with the Bureau's examinations, these violations were generally related to the Rule's provisions on completing loss mitigation applications, loss mitigation evaluation and appeals, and foreclosure restrictions.

Certain servicers have made efforts to properly implement effective compliance management programs. Some servicers have made significant improvements in the last several years, in part by enhancing and monitoring their servicing platforms, staff training, coding accuracy, auditing, and allowing for greater flexibility in operations.

In October 2015, the National Housing Resource Center conducted a national survey of housing counselors about compliance with the Bureau's mortgage servicing regulations. The survey asked respondents to rate how well each of 11 servicers complied with the Bureau's servicing regulations, including provisions established by the Rule. On average, respondents reported servicers were most likely to comply with the Rule's foreclosure restrictions (the restriction on initiating foreclosure within 120 days of delinquency, on initiating foreclosure while a complete loss mitigation application is pending, and on scheduling or conducting a foreclosure sale while a loss mitigation application is pending). Respondents reported lower levels of compliance with the Rule's requirements for continuity of contact and acknowledging within five days the receipt of loss mitigation applications.

In responses to open-ended questions, respondents to the Counselor Survey also highlighted compliance challenges. Many reported experiences with clients who did not receive a letter within five days acknowledging the receipt of loss mitigation applications, and several raised concerns about compliance with the provisions related to foreclosure restrictions and continuity of contact.

4. Overall effects of the Rule on foreclosures and modifications

The preceding chapter describes trends in mortgage servicing from the housing crisis to the present, spanning the effective date of the 2013 RESPA Servicing Rule. This chapter focuses on measuring the overall effect of the Rule on two key outcomes: the incidence of foreclosure and the incidence of recovering from delinquency. As discussed in Chapter 1, many provisions of the Rule are aimed at facilitating review for foreclosure avoidance options by requiring servicers to make disclosures to borrowers about options for avoiding foreclosure, by requiring servicers to help borrowers with the process of applying for those options, and by requiring certain timelines and procedures in evaluating loss mitigation applications and in pursuing foreclosures. The Bureau emphasized in the Rule that its goal was not to achieve any particular target with respect to the number or speed of foreclosures, but rather to ensure that borrowers are protected from harm in connection with the process of evaluating a borrower for a loss mitigation option and proceeding to foreclosure.

Later chapters of the report use the Servicing Operations Data and other sources of information to assess the effects of individual components of the Rule. In contrast, this chapter uses the McDash Data and GSE Data to study the overall effects of the Rule on the incidence of foreclosure and the incidence of recovery from delinquency, but does not provide information as to which parts of the Rule affected these outcomes.

The first section of this chapter describes the construction of the data used for the analysis of the overall effects of the Rule, with a high-level description of the methodology (technical details on the methodology are described in Appendix D). The second section presents results on the effect of the Rule on foreclosures, and the third section presents results on recovering from delinquency.

4.1 Data and methodology

For this chapter, the Bureau constructed four related datasets, two using the McDash Data and two using the GSE Data from Fannie Mae (Fannie Mae Data). Specifically, for each source of data, the Bureau constructed one dataset designed to study the incidence of foreclosure and one dataset designed to study the incidence of recovering from delinquency, including through a loan modification. For studying the incidence of foreclosure, the datasets include monthly observations for loans that become 90 days delinquent, starting from the month when the loan is first observed being 90 days delinquent, and continuing until the loan experienced a foreclosure sale or otherwise exited the dataset.118 To study recovery from delinquencies, the Bureau constructed a similar dataset which followed loans from the time they became 30 days delinquent until they either became current (through a modification or otherwise) or exited the data some other way. For each dataset, the Bureau excluded loans that experienced the outcome of interest prior to becoming the requisite number of days delinquent as well as loans that are never observed to become the requisite number of days delinquent. To make the data as comparable as possible between the McDash and Fannie Mae sources, the Bureau excluded a small number of loans from the McDash Data that were originated before 1999, and a small number of loans from the Fannie Mae Data that were located in the Virgin Islands or Guam—in each case such loans were not present in the other data source.121 For computational reasons, the McDash Data used in this chapter consist of a random sample of 10 percent of all loans in the data that meet the delinquency criteria stated above.

Data of this nature present two related challenges, which require specialized methods to address. First, the data do not always show whether or not an outcome eventually occurs. For instance, if a loan in the Fannie Mae Data becomes 90 days delinquent in January 2016 and has not experienced foreclosure by December 2017, the last month observed in the data, the data cannot speak to whether that loan eventually experienced foreclosure. This in turn complicates calculations, for instance, of the rate of foreclosures 36 months following a 90-day delinquency. This issue is known as "censoring," and can, under some assumptions, be readily addressed using the tools of duration analysis, also known as survival or hazard analysis.

Second, and closely related to the first issue, data on mortgage servicing presents an environment with competing risks. That is, a loan may exit the data not only because of the end of the data, but also because of other final outcomes. For example, if foreclosure is the outcome of interest, a loan may exit the data because the borrower pays off the loan early, such as through a refinance. This in turn will lead to biased estimates if the competing outcomes are correlated with the risk of the outcome of interest. In the case of mortgage servicing data, the competing outcomes are very likely to be correlated—borrowers who pre-pay likely have a very different underlying risk of foreclosure from those who do not, and as a result a simple analysis of borrowers who do not pre-pay will not be representative of the average risk of foreclosure. This chapter uses specialized duration analysis tools to analyze data with competing risks. For more technical detail on the methodology used in this chapter, see Appendix D.

4.2 Overall effects of the Rule on the incidence of foreclosure

As noted above and discussed in detail in Chapter 2, several provisions of the Rule might be expected to reduce the prevalence or incidence of foreclosure.124 This section examines the overall effect of the Rule on foreclosure based on datasets from the Fannie Mae Data and the McDash Data which follow loans until foreclosure or other exit (see section 4.1). In both datasets, the outcome of interest is a completed foreclosure, which might be a sale, auction, or transition to real estate owned (REO) status by the bank that owns the mortgage, depending on the foreclosure process in the state where the property is located and the choices of the servicer or the investor. This section first describes the patterns in the incidence of foreclosure and then presents formal statistical estimates of the Rule's effect.

4.2.1 Patterns in the incidence of foreclosure

The following analyses describe changes in the incidence of foreclosure since the effective date of the Rule. Any changes, of course, may not necessarily be caused by the Rule but may instead result from a variety of trends in the mortgage market, including the general improvement in the economy and in the housing market in many states. This section describes the changes that the Rule could potentially account for and identifies factors that the formal statistical analysis in the next section needs to control for to attempt to isolate the effect of the Rule.

Figures 1 and 2, below, plot the incidence of foreclosures over time since becoming 90 days delinquent, by the year a loan first became 90 days delinquent. The horizontal axis of each plot shows the number of months since becoming 90 days delinquent, and the vertical axis shows the cumulative incidence of foreclosure for such loans—in essence, the proportion that experienced a completed foreclosure by that time. Incidence rates are shown for 2005, 2008, 2011 and 2014, for three years after each loan first becomes 90 days delinquent. These years were chosen to present snapshots of the incidence rate before, during, and after the housing crisis as well as after the effective date of the Rule. Figure 1 uses the Fannie Mae Data, while Figure 2 uses the McDash Data.

FIGURE 1: CUMULATIVE INCIDENCE OF FORECLOSURE FOR LOANS THAT HAVE BECOME 90 DAYS DELINQUENT, BY THE YEAR THE LOAN BECAME 90 DAYS DELINQUENT (FANNIE MAE DATA)

FIGURE 2: CUMULATIVE INCIDENCE OF FORECLOSURE FOR LOANS THAT HAVE BECOME 90 DAYS DELINQUENT, BY THE YEAR THE LOAN BECAME 90 DAYS DELINQUENT (MCDASH DATA)

The incidence of foreclosure increases steadily over the first two years after a loan becomes 90 days delinquent, but increases less rapidly after two years. The rate of increase has changed substantially over time, however. Consistent with the trends presented in the preceding chapter, both figures show that the incidence of foreclosure increased dramatically for loans that first became 90 days delinquent in 2008 compared with loans that first became 90 days delinquent in 2005, before the housing crisis. After 2008, the incidence of foreclosure over time declined in both datasets, although this decline was more pronounced for the set of loans in the McDash Data. For loans in either dataset that first became 90 days delinquent in 2008, about 35 percent had experienced a completed foreclosure 36 months later. By comparison, for loans that became 90 days delinquent in 2014, about 23 percent of Fannie Mae loans and about 18 percent of loans in the McDash Data had experienced a completed foreclosure within 36 months. It is not clear whether the differences across the two data sources are due to the different samples of loans used in the two data sources or disparities in how the two datasets record foreclosures. Nonetheless, although the levels may differ, the broad trends over time are similar between the Fannie Mae Data and McDash Data.

The remaining descriptive analyses in this section focus on changes in the incidence of foreclosure between loans that first became 90 days delinquent in 2011, before the effective date of the Rule, and loans that first became 90 days delinquent in 2014, after the effective date. These analyses are intended to illustrate the changes in the incidence of foreclosure between the pre- and post-Rule periods, but should not be interpreted as the causal effect of the Rule—a number of other changes were occurring between these two years, including the National Mortgage Settlement (NMS) and the introduction of the GSE Streamlined Modification program. The next subsection uses regression analysis to attempt to account for some of these other factors.

The next two analyses focus on the Fannie Mae Data—results using the McDash Data are similar, differing in levels but not in patterns.

Loan modifications may be an important factor in helping delinquent borrowers avoid foreclosure. Many of the Rule's requirements address informing borrowers about loss mitigation options such as loan modifications, and the application and timely evaluation of borrowers for such options. Figure 3, below, breaks out the incidence of foreclosure in the Fannie Mae Data by year of delinquency and whether the borrower had a modification in effect as of their last observation in the data. For comparison, about 36 percent of loans that became 90 days delinquent in 2011 and 39 percent of loans that became 90 days delinquent in 2014 received a modification within 36 months.

FIGURE 3: CUMULATIVE INCIDENCE OF FORECLOSURE FOR LOANS THAT BECOME 90 DAYS DELINQUENT, BY YEAR AND WHETHER OR NOT THE LOAN WAS MODIFIED BY THE LAST MONTH IT WAS OBSERVED (FANNIE MAE DATA)

The incidence of foreclosure was much lower for borrowers who received a modification—fewer than 5 percent of such borrowers had foreclosure completed within three years of becoming 90 days delinquent in 2011. In contrast, 45 percent of borrowers with unmodified loans who became 90 days delinquent in 2011 experienced foreclosure within three years. The decline in the incidence of foreclosure for loans that became 90 days delinquent in 2014 occurred entirely for unmodified loans—if anything the incidence of foreclosure increased slightly for modified loans. Although not the focus of this analysis, the difference in the incidence of foreclosure between modified and unmodified loans suggests that loan modifications are, in general, effective at preventing foreclosure rather than simply delaying it, at least over a horizon of three years.

As discussed in the previous chapter, foreclosure rates differ substantially by state, particularly depending on whether the state has a judicial foreclosure regime. Figure 4 shows the cumulative incidence rates of foreclosure by year and type of state, again focusing on the Fannie Mae Data because the results in the McDash Data are very similar. Comparing the two panels shows that, for loans that became 90 days delinquent in 2011, significantly fewer experienced foreclosure within three years in states with judicial foreclosure. At the same time, the incidence rate declined only in states without judicial foreclosure. Other elements of state law besides a judicial foreclosure regime may influence the incidence of foreclosure as well. This comparison serves to illustrate the importance of controlling for baseline differences across states.

FIGURE 4: CUMULATIVE INCIDENCE OF FORECLOSURE FOR LOANS THAT BECOME 90 DAYS DELINQUENT, BY YEAR AND TYPE OF STATE FORECLOSURE REGIME (FANNIE MAE DATA)

The incidence of foreclosure before and after the Rule also varied across segments of the market. Figure 5, below, uses the McDash Data to calculate the cumulative incidence of foreclosure by the type of investor that owned the loan, grouped into government-insured loans securitized by Ginnie Mae (e.g., FHA, VA), GSE loans (e.g., Fannie Mae), loans held in the portfolio of the loan's servicer, and loans securitized by private label investors. The incidence of foreclosure for seriously delinquent loans varies widely across these categories. For loans that are observed becoming 90 days delinquent in 2011, the incidence of foreclosure after three years ranged from almost 25 percent for government-insured loans to around 14 percent for originator-owned loans. The incidence of foreclosure was lower for GSE loans that became 90 days delinquent in 2014, compared to such loans that became 90 days delinquent in 2011.

FIGURE 5: CUMULATIVE INCIDENCE OF FORECLOSURE FOR LOANS THAT BECOME 90 DAYS DELINQUENT, BY YEAR AND TYPE OF INVESTOR (MCDASH DATA)

Servicer practices are also a significant determinant of the risk of foreclosure—indeed, as discussed in Chapter 1, this was a primary motivation for the 2013 RESPA Servicing Rule. Figure 6, below, illustrates this using the Fannie Mae Data. The Fannie Mae Data include a servicer identifier for any servicers that hold more than 1 percent of the loan volume in the Fannie Mae data, with smaller servicers pooled in an "Other" category. There are about 35 servicers identified in this manner, including five entities for whom the requirements of the NMS became binding in April 2012 and four others that joined the settlement later, with the requirements binding in 2013. Figure 6 breaks out the cumulative incidence of foreclosure for loans that became 90 days delinquent by year and whether the loan was serviced by an NMS servicer. Note that the national servicing standards required by the NMS became binding for most servicers subject to it between 2012 and 2013, and expired three and a half years later. Thus, any changes in foreclosure incidence between 2011 and 2014 reflect both the Rule and the NMS itself in addition to any other trends and shifts that may have been occurring in the mortgage market. The figure shows a substantial decline in the incidence of foreclosure for delinquent loans serviced by parties to the NMS. For loans that became delinquent in 2011, the incidence of foreclosure for loans serviced by NMS servicers was around double that of nonNMS servicers at any point in time after becoming 90 days delinquent. NMS servicers actually had a slightly lower incidence of foreclosures compared to non-NMS servicers for loans that became 90 days delinquent in 2014. Given that a large fraction of loans in the McDash Data are serviced by large banks, it is possible that the differences in the overall incidence of foreclosure between the two datasets are due to a higher proportion of loans serviced by NMS servicers in the McDash Data compared with the Fannie Mae Data.

FIGURE 6: CUMULATIVE INCIDENCE OF FORECLOSURE FOR LOANS THAT BECOME 90 DAYS DELINQUENT, BY YEAR AND WHETHER SERVICER WAS PARTY TO THE NATIONAL MORTAGE SETTLEMENT (FANNIE MAE DATA)

The results in this section indicate that the incidence of foreclosure has declined substantially for loans that became 90 days delinquent since the effective date of the Rule. As the results in this section demonstrate, the reduction in the incidence of foreclosure was limited to particular sections of the market. In some cases it would make sense for the Rule to only affect those segments, but even then care should be taken in attributing these changes to the Rule. Above all, the results in this section illustrate that it is important to account for differences in foreclosure rates across loan products, states, and servicers, among other factors.

4.2.2 Modeling the effect of the Rule on foreclosure

Building on the descriptive results above, this section next presents statistical estimates of the overall effect of the 2013 RESPA Servicing Rule on foreclosures. The goal is to establish, to the extent possible given the limitations of the data, the causal effect of the Rule on the incidence of foreclosures. The results in this section are based on a regression model that accounts for both censoring and competing risks as well as loan characteristics and factors other than the Rule that may affect the incidence of foreclosures. The regressions use the full data described in section 4.1, rather than simply comparing loans that became 90 days delinquent in 2011 to those that became 90 days delinquent in 2014. In essence, the effect of the Rule is estimated by comparing similar loans that have had the same amount of time pass since becoming 90 days delinquent, but with or without the Rule in effect—equivalent to comparing the vertical distance between two lines in the cumulative incidence figures above, except with other loan characteristics and policies held constant. See Appendix D for technical details on the regression methodology.

As noted above, the key variable in the regression model is an indicator for whether or not the Rule was effective at the beginning of the current month. In addition to the indicator for the Rule, all of the regressions control for a set of policy variables (e.g., the NMS and the FHFA Servicing Alignment Initiative) and loan characteristics, including state, origination year, and the year the delinquency began. The full results of the regression models, with coefficients, are presented in Appendix D.

The regression model is then used to predict how many additional foreclosures would have occurred without the Rule for loans that became 90 days delinquent in 2014. Table 1, below, shows the results of this prediction exercise. The first column shows results based on the model using the Fannie Mae Data and the second column shows results based on the McDash Data. The first row shows, for loans that became 90 days delinquent in 2014, the predicted change in the incidence of foreclosure 36 months later due to the Rule. To put these numbers in perspective, the remainder of the table extrapolates this change in incidence to the change in the level of foreclosures in the entire mortgage market. The second row of Table 1 reports the total number of loans that became 90 days delinquent in 2014 in each dataset, and the third row of the table reports the change in the number of foreclosures after 36 months one would expect to see in the dataset based on the change in incidence. Of course, the datasets used for the models contain a subset of all seriously delinquent loans in the mortgage market.129 Data from the Federal Reserve Bank of New York indicate that about $218 billion of mortgage balances became 90 days delinquent in 2014, while the Fannie Mae Data used in this section represented $4.9 billion of mortgage balances and the McDash Data represented $6.5 billion.130 The fourth row of Table 1 reports the multiplier needed to make each dataset equivalent to the market as a whole.131 The final row of Table 1 reports an estimate of the total number of foreclosures within three years that could have been prevented had the Rule been in effect in 2011.

TABLE 1: PREDICTED CHANGE IN FORECLOSURES CAUSED BY THE RULE, BASED ON ESTIMATES FROM A COMPETING RISKS HAZARD MODEL

Fannie Mae Data

McDash Data

Predicted Change in Cumulative Incidence of Foreclosure

-0.023

-0.02

# Loans In Data That Became 90 Days Delinquent in 2014

37,810

38,228

Predicted Change in Foreclosures In Data

-856

-773

Multiplier to Match Data to Market

44.222

33.82

Predicted Change In Foreclosures in Market

-37,854

-26,142

Taking the model using the Fannie Mae Data as an example, the model predicts that for loans that became 90 days delinquent in 2014, the Rule lowered the cumulative incidence rate of foreclosure after 36 months by 2.3 percentage points. For reference, Figure 1 shows that about 22 percent of all loans that became 90 days delinquent in 2014 experienced a completed foreclosure within three years. The data used for the model contain about 38,000 loans that first became 90 days delinquent in 2014, and so the reduction in the cumulative incidence rate implies that the Rule reduced the number of completed foreclosures within three years in this dataset by 856. The roughly 38,000 loans in the Fannie Mae Data that became 90 days delinquent in 2014 represent roughly 2 percent of all mortgage balances that became 90 days delinquent in 2014, and so multiplying by about 44 gives approximately the effect on the whole market—a reduction of almost 38,000 loans. In other words, the model predicts that if the Rule had not gone into effect in 2014, approximately 38,000 more borrowers who became delinquent that year would have experienced foreclosure within three years. The model using the McDash Data predicts a somewhat smaller effect of the Rule on foreclosures—a reduction of about 26,000 foreclosures.

The analysis above is premised on several assumptions, which may not hold in practice. First, the bottom line estimate of foreclosures prevented in the market as a whole assumes that the data are representative of the mortgage market as a whole—or, more precisely, of the population of loans that became 90 days delinquent in 2014. The Fannie Mae Data represent the universe of fully-amortizing, fixed-rate, single family mortgages guaranteed by Fannie Mae, and the 10 percent random sample of the McDash Data used for the regression analysis is representative of single family mortgages reported to Black Knight, which make up a majority of loans in the market. However, if the Rule had substantially different effects on loans that are in neither the Fannie Mae nor the McDash Data, this analysis cannot account for those differences. The fact that the Fannie Mae and McDash Data ultimately give similar results, despite having very different sets of loans, may provide some indication that these results are representative of the market as a whole.

Second, although the regression models contain a variety of controls designed to account for factors unrelated to the Rule that may have affected the incidence of foreclosure, it is not possible to account for all factors that may affect foreclosure. In particular, if there was a downward trend in the incidence of foreclosure around the effective date of the Rule, a comparison of similar borrowers before and after the Rule could show a decline in incidence even if nothing changed. The many controls in the regression model are intended to adjust for any such trend, but they may not do so perfectly. There is a way to test for the presence of an unrelated trend around the effective date of the Rule: rather than using a single indicator for whether the current month is before or after the effective date of the Rule, one can estimate the model with a set of indicators denoting 12 or more months before the Rule, 9 to 12 months before the Rule, 6 to 9 months before the Rule, 3 to 6 months before the Rule, 0 to 3 months before the Rule, 0 to 3 months after the Rule, and so on with the final indicator being for 24 or more months after the Rule. The pre-Rule indicators will then capture any general trend in foreclosures around the time of the Rule after adjusting for other factors, and any difference between that trend and the effect of the post-Rule indicators will capture the effect of the Rule on the incidence of foreclosure. That is, if the Rule truly had an effect on the incidence of foreclosure, there should be a break in the trend around the time of the Rule. Figure 7, below, shows the results of this test using the Fannie Mae Data, plotting the model coefficients for the pre- and post-Rule indicators described above and a dashed line to show the slope of the preRule trend.

FIGURE 7: EFFECT OF THE RULE ON THE INCIDENCE OF FORECLOSURE: EVENT STUDY RESULTS FROM A COMPETING RISKS HAZARD MODEL (FANNIE MAE DATA)

The figure shows that even after controlling for other factors in the model, there was still a downward trend in the incidence of foreclosures before the effective date of the Rule. In other words, the controls used in the model do not fully explain why foreclosures were decreasing during the period leading up to the Rule. However, following the effective date of the Rule, the incidence of foreclosures fell further relative to the trend as illustrated by the sharp drop in the estimate six months after the effective date and continued decline through 20 months post-Rule. This result indicates at a minimum that the timing of the effect shown in Table 1 lines up with the effective date of the Rule, although the model cannot rule out the possibility that some other policy change or other factor that took place in January 2014 contributed to the decline in the incidence of foreclosures. Still, given the substantial effects of specific provisions of the Rule on foreclosure outcomes described later in this report, it seems plausible that the Rule is an important driver of this effect.

4.3 Overall effects of the Rule on the incidence of recovery from delinquency

As noted above and discussed in detail in Chapter 2, while the Rule does not require owners or investors to offer loan modifications or loss mitigation options, several provisions of the Rule were intended to inform borrowers about loss mitigation options and to establish a more fair application and evaluation process for such options. Although the Bureau was not seeking to achieve a particular level of loan modifications through the Rule, the expected effect of many provisions was to assist borrowers in retaining their property if they desired and to mitigate losses for investors. This section examines the overall effects of the Rule on loan modifications and self-cures (becoming current on the loan without assistance) jointly. There are two reasons for pooling these two outcomes, one substantive and one practical. Substantively, having borrowers retain their homes and resume making regular payments benefits both borrowers and investors regardless of how this occurs, and the Rule could have affected the incidence of both loan modifications and self-cures. On a practical level, loan modifications are not always recorded reliably in the data, particularly in the McDash Data. However, in both the Fannie Mae and McDash Data, loans almost always are marked current when they begin a permanent loan modification. Thus, becoming current following a delinquency is a good proxy for both self-cures and loan modifications, but does not distinguish between the two outcomes. This section refers to these joint outcomes as "recovery," and this section studies the effect of the Rule on the incidence of recovery.

The analysis in this section examines loans in the Fannie Mae Data and the McDash Data from the month they are first observed being 30 days delinquent until they recover or otherwise exit the data (see section 4.1). In both datasets, the outcome of interest is becoming current or having a flag for a permanent loan modification. Recovery is considered to occur in the first month the delinquent loan becomes current or has a permanent modification flag, but is not counted as a true recovery unless the loan remains current for at least three months. This section first describes the patterns in the incidence of recovery and then presents formal statistical results on the effect of the Rule.

4.3.1 Patterns in the incidence of recovery

The following analyses describe changes in the incidence of recovery since the effective date of the Rule. Note that any such changes were not necessarily caused by the Rule. There are a variety of trends in the mortgage market that could account for these changes, including the general improvement in the economy and in the housing market in many states. Instead, the results in this section describe the changes that the Rule could potentially account for, and identify important factors that the formal statistical analysis in the next subsection will need to control for.

Figures 8 and 9, below, plot the probability of a recovery over time since becoming 30 days delinquent, by the year a loan first became 30 days delinquent. The horizontal axis of each plot shows the number of months since becoming 30 days delinquent, and the vertical axis shows the cumulative fraction of such loans that recovered by that time. Figure 8 uses the Fannie Mae Data, and Figure 9 uses the McDash Data.

FIGURE 8: CUMULATIVE INCIDENCE OF RECOVERY FOR LOANS THAT BECOME 30 DAYS DELINQUENT, BY YEAR THE LOAN BECAME 30 DAYS DELINQUENT (FANNIE MAE DATA)

FIGURE 9: CUMULATIVE INCIDENCE OF RECOVERY FOR LOANS THAT BECOME 30 DAYS DELINQUENT, BY YEAR THE LOAN BECAME 30 DAYS DELINQUENT (MCDASH DATA)

For loans in both data sets, there was a noticeable shift in the timing and frequency of recovery between loans that became 30 days delinquent in 2005, before the housing crisis, and loans that became 30 days delinquent in 2008, during the crisis. The differences were most pronounced roughly one year after becoming 30 days delinquent, with about 15 percentage points fewer borrowers who became 30 days delinquent in 2008 recovering in one year, compared to borrowers who became 30 days delinquent in 2005. In both datasets the incidence of recovery was higher for loans that became 30 days delinquent in 2011 than in 2008, and the recovery rate for loans that became 30 days delinquent in 2014 exceeded the 2005 level. About 60 to 85 percent of all loans in both datasets recover within three years of becoming 30 days delinquent, with some variation depending on the year the borrower became delinquent. Fewer loans in the McDash Data recovered within a few months—the incidence of recovery in the first month after becoming 30 days delinquent was around one-third in each year, compared to around one-half for loans in the Fannie Mae Data.

Although there is an important interaction between recovery and foreclosures, the type of state foreclosure regime appears to have less of an effect on recovery than on foreclosures. Figure 10 breaks out the cumulative incidence of recovery in the Fannie Mae Data by the year each loan became 30 days delinquent and whether the state had a judicial foreclosure process. Recovery rates and differences in recovery rates for loans that became delinquent in 2011 versus 2014 are similar in both types of states.

FIGURE 10: CUMULATIVE INCIDENCE OF RECOVERY FOR LOANS THAT BECOME 30 DAYS DELINQUENT, BY YEAR AND TYPE OF STATE FORECLOSURE REGIME (FANNIE MAE DATA)

The difference in the incidence of recovery between loans that were serviced by parties to the NMS and those that were not is also less pronounced, compared to the incidence of foreclosures. Figure 11, below, shows the cumulative incidence of recovery for loans that became 30 days delinquent before and after the effective date of the Rule, broken out by whether the loan was serviced by an NMS servicer. The figure uses the Fannie Mae Data, as the McDash Data does not have servicer identifiers. Although loans serviced by NMS servicers were slightly more likely to recover earlier in delinquency, this difference persisted for loans that became delinquent after the effective dates of the Rule and the NMS.

FIGURE 11: CUMULATIVE INCIDENCE OF RECOVERY FOR LOANS THAT BECOME 30 DAYS DELINQUENT, BY YEAR AND WHETHER SERVICER WAS PARTY TO THE NATIONAL MORTAGE SETTLEMENT (FANNIE MAE DATA)

There is some variation in the incidence of recovery across loans held by different types of investors. Figure 12 uses the McDash Data to plot the cumulative recovery rate for loans that become 30 days delinquent in 2011 or 2014, separately for different types of investors. Both before and after the effective date of the Rule, loans held by the GSEs and government-insured loans had the highest incidence of recovery. Given that the FHFA Servicing Alignment Initiative imposed requirements around loss mitigation applications and was already in effect for loans that became delinquent in 2011, this result should be expected to some extent. Other types of loans had lower rates of recovery, but the cumulative incidence of recovery increased the most for private-label loans that became 30 days delinquent in 2014, compared to 2011.

FIGURE 12: CUMULATIVE INCIDENCE OF RECOVERY FOR LOANS THAT BECOME 30 DAYS DELINQUENT, BY YEAR AND INVESTOR TYPE (MCDASH DATA)

4.3.2 Modeling the effect of the Rule on recovery

Building on the descriptive results above, this section next presents statistical estimates of the overall effect of the 2013 RESPA Servicing Rule on recovery. Here the goal is to establish, to the extent possible given the limitations of the data, the potential causal effect of the Rule on the incidence of recovery. The results in this section are based on a regression model that accounts for both censoring and competing risks as well as loan characteristics and other factors that may affect the incidence of recovery besides the Rule. The regressions use the full data described in section 4.1, rather than simply comparing loans that became 30 days delinquent in 2011 to those that became 30 days delinquent in 2014. In essence, the effect of the Rule is estimated by comparing similar loans that have had the same amount of time pass since first being observed becoming 30 days delinquent, but with or without the Rule in effect. This is conceptually equivalent to comparing the vertical distance between two lines in the figures above while holding other loan characteristics and policies constant. See Appendix D for technical details on the regression methodology. A difference from the statistical results in the previous section is that foreclosure sales are treated as a competing event.

As noted above, the key variable in the regressions models in this section is an indicator for whether or not the Rule was in effect at the beginning of the current month. In addition to the indicator for the Rule, all of the regressions control for a set of policy variables (e.g., the NMS and the FHFA Servicing Alignment Initiative) and loan characteristics, including state, origination year, and the year the delinquency began. The full coefficient results are presented in Appendix D.

The regression model in each dataset is then used to predict how many fewer recoveries would have occurred without the Rule. Table 2, below, presents the predicted effect of the Rule on recovery. The methodology for this prediction exercise is identical to that described in section 4.2.2, above; see that section and Appendix D for details.

TABLE 2: PREDICTED CHANGE IN RECOVERY CAUSED BY THE RULE, BASED ON ESTIMATES FROM A COMPETING RISKS HAZARD MODEL

Fannie Mae Data

McDash Data

Predicted Change in Cumulative Incidence of Recovery

0.045

0.15

# Loans In Data That Became 30 Days Delinquent in 2014

105,644

71,811

Predicted Change in Recovery In Data

4,737

10,556

Multiplier to Match Data to Market

26.788

34.2

Predicted Change In Recovery in Market

126,892

360,971

Taking the model using the Fannie Mae Data as an example, the model predicts that for loans that became 30 days delinquent in 2014, the cumulative incidence rate of recovery after three years increased by 4.5 percentage points because of the Rule. The data used for the model contain about 105,600 loans that first became 30 days delinquent in 2014, and so the change in the cumulative incidence rate implies that 4,737 more loans in the data would have recovered from delinquency within three years. The roughly 105,600 loans in the Fannie Mae Data that became 30 days delinquent in 2014 represent about 4 percent of all mortgage balances that became 30 days delinquent in 2014, and so multiplying by about 27 gives approximately the effect on the whole market—an increase of about 127,000 loans. In other words, the model predicts that because of the Rule, 127,000 more borrowers recovered within three years. The model using the McDash Data predicts a significantly larger change of about 361,000 recoveries.

As with the analysis on foreclosures described in section 4.2.2 above, the analysis above is premised on several assumptions, which may not hold in practice. First, the bottom line estimates of recoveries in the market as a whole assumes that the data are representative of the mortgage market as a whole—or, more precisely, the population of loans that became 30 days delinquent in 2014. The Fannie Mae Data do represent the universe of fixed-rate, single family mortgages guaranteed by Fannie Mae, and the 10 percent random sample of the McDash Data used for the analysis is representative of single family mortgages reported to Black Knight, which make up a majority of loans in the market. However, if the Rule had substantially different effects on loans in neither the Fannie Mae nor the McDash Data, this analysis cannot account for those differences. The fact that the Fannie Mae and McDash Data ultimately give very different results for this analysis particularly indicates a need for caution.

Second, although the regression models contain a variety of controls designed to account for factors unrelated to the Rule that may have affected the incidence of recovery, it is not possible to account for all factors that may affect recovery. In particular, if there was an upward trend in the incidence of recovery around the effective date of the Rule, a comparison of similar borrowers before and after the Rule could show a decline in incidence even if nothing changed. The many controls in the regression model are intended to adjust for any such trend, but they may not do so perfectly. Similar to the analysis for foreclosures, a test for an unrelated trend is, rather than using a single indicator for before or after the effective date, to estimate the model with a set of indicators denoting 12 or more months before the Rule, 9 to 12 months before the Rule, 6 to 9 months before the Rule, 3 to 6 months before the Rule, 0 to 3 months before the Rule, 0 to 3 months after the Rule, and so on with the final indicator being 24 or more months after the Rule. The pre-Rule indicators will then capture any general trend in recovery around the time of the Rule after adjusting for other factors, and any difference between that trend and effect of the post-Rule indicators will capture the effect of the Rule on the incidence of recovery. That is, if the Rule truly had an effect on the incidence of recovery, there should be a break in the trend around the time of the Rule. Figure 13, below, shows the results of this test using the McDash Data, plotting the model coefficients for the pre- and post-Rule indicators described above and a dashed line to show the slope of the pre-Rule trend.

FIGURE 13: EFFECT OF THE RULE ON THE INCIDENCE OF RECOVERY: EVENT STUDY RESULTS FROM A COMPETING RISKS HAZARD MODEL (MCDASH DATA)

The figure shows that even after controlling for other factors in the model, there was a slight downward trend in the incidence of recovery around the effective date of the Rule. Following the effective date of the Rule, the incidence of recovery increased substantially, then began to trend upward somewhat. This result at a minimum indicates that the timing of the effect shown in Table 2 lines up with the effective date of the Rule, although the model cannot rule out the possibility that some other policy or other factor that took place in January 2014 contributed to the increase in the incidence of recovery.

4.4 Summary and conclusion

Many of the provisions of the 2013 RESPA Servicing Rule are intended to establish a fair process for review of loss mitigation applications to help borrowers avoid foreclosure when possible. Administrative data from Fannie Mae and McDash show that borrowers who became delinquent were less likely to experience foreclosure after the effective date of the Rule compared to before the effective date of the Rule. A statistical model of the incidence of foreclosure shows that foreclosures decreased at the time the Rule became effective, even after controlling for other policies, trends, and loan characteristics. The model suggests that the Rule prevented at least 26,000 foreclosures for loans that became seriously delinquent in 2014. The Rule also appears to have had an effect on borrowers' ability to recover from delinquency, whether by self-curing or by obtaining a loan modification. This effect appears both in the raw data and after controlling for other factors. A model of recovery from delinquency suggests that the Rule caused at least 127,000 additional borrowers to recover from delinquency.

The statistical evidence indicates that the changes in foreclosures and recoveries coincide with the effective date of the Rule, rather than being the continuation of a pre-existing trend. However, an important limitation of these analyses is that they cannot distinguish between the effect of the Rule and the effect of any other events that occurred on or around January 2014. In particular, the GSE Streamlined Modification program was rolled out widely to servicers in January 2014, and the Bureau's 2013 Ability-to-Repay/Qualified Mortgage Rule became effective on the same day as the 2013 RESPA Servicing Rule. Moreover, this analysis of the overall effects of the Rule cannot distinguish between the effects of individual provisions of the Rule. Based on this analysis alone, it is possible for individual provisions to have contributed to all, some, or none of the overall change in foreclosures and recoveries, and it is even possible that some provisions had a detrimental effect on foreclosures and recoveries that was offset by positive effects from other provisions. Subsequent chapters of this report examine the effects of individual provisions of the Rule.

5. Overall effects of the Rule on servicing costs

This chapter discusses what the Bureau learned about how the 2013 RESPA Servicing Rule affected the overall cost of servicing mortgage loans. The cost to servicers of complying with specific provisions of the Rule is discussed in later chapters that address those provisions.

The first two sections of this chapter discuss servicing costs generally, overall trends in those costs, and trends in labor productivity at mortgage servicers. The third section of the chapter discusses what the Bureau learned from servicer interviews and public comments about the effect of the Rule on servicing costs, and the fourth section discusses effects on costs that are specific to small and mid-size servicers.

5.1 Trends in servicing costs

The data described in this section indicate a significant increase in the cost of servicing mortgage loans since the beginning of the housing crisis, with an especially large increase in the cost of servicing delinquent loans. Some industry participants have attributed these cost increases in part to increased regulation, including the need to dedicate resources towards improving processes, quality assurance, and customer-facing practices to comply with those regulations. Industry participants have noted that many servicers remain on legacy software platforms that make system changes, including those needed to implement new regulations, time-consuming and costly.147 Some observers have said that increased regulatory oversight of servicing at both state and federal levels has contributed to increased ongoing costs, including staffing compliance departments and investing in regulatory tracking and monitoring systems. Some industry observers have suggested that recent cost increases have reduced consumer access to credit and have factored into the decision of some servicers, especially depository institutions, to leave the servicing business.

5.1.1 Available data on servicing costs

The discussion of servicing costs in this subsection is based on two data sources from the Mortgage Bankers Association (MBA). The first comes from a survey of around 30 larger servicers that participate in MBA's Servicing Operations Study and Forum for Prime and Specialty Servicers (MBA Forum Data). The second is from MBA's Annual Performance Reports (MBA Annual Performance Report Data), which include data from non-depository servicers that service loans owned or guaranteed by Fannie Mae, Freddie Mac, or Ginnie Mae and elect to make their data available to MBA for aggregate industry statistics. Each source has advantages and disadvantages. The MBA Forum Data are based on a smaller number of servicers but that number includes a large share of the largest servicers. The MBA Annual Performance Report Data include a larger number of servicers, but because this source covers only non-depository servicers that service for the GSEs or Ginnie Mae, it represents a relatively narrow segment of the servicing market.

There are limitations to these data sources, which present averages of costs and other measures across the limited sample of servicers who chose to participate in the respective survey in each period. Limitations include the following: (1) The servicers included in the sample change over time in ways that the Bureau cannot determine. Because of this, the Bureau cannot observe whether trends in the average costs reflect changes in the composition of the sample or changes in the underlying cost of servicing or other measures reported; (2) Participants in these surveys may not be representative of the servicing market as a whole; and (3) Because the data sources generally report only averages over all loans in their respective populations, changes in the numbers reflect changes both in the cost of servicing particular loans and in the types of loans being serviced. As discussed below, delinquent loans are more costly to service than nondelinquent loans. This makes it particularly difficult to interpret the MBA Annual Performance Report Data, which average the servicing cost for all loans, regardless of delinquency status.

The share of delinquent loans changes over time and is different for different servicers. Therefore, changes in the average cost of servicing a loan are likely to reflect changes in the delinquency rates (percentages of delinquent loans) of servicers participating in the survey as well as changes in the underlying cost of servicing a delinquent or non-delinquent loan.

Despite these limitations, these two data sets are the only sources of data on servicing costs the Bureau is aware of that reflect a broad range of market participants. The Bureau believes that the data are likely to reflect trends in underlying servicing costs even if the reliability or interpretation of any specific number is uncertain.

5.1.2 Average servicing costs 2008 to 2017

This section presents overall trends in servicing costs. To provide context for these trends, however, this section first discusses the components of servicing costs, and in particular the costs of servicing loans in default. Table 1 provides a breakdown of servicing costs by function, based on 2017 MBA Forum Data. Costs are divided into the "direct" costs of servicing loans and "indirect" costs, which include an allocation of corporate administrative expense and unreimbursed costs related to foreclosure and REO. Direct costs are further categorized according to costs associated with servicing all loans regardless of delinquency status and costs that are specific to loans in default.

The table shows that annual costs specific to loans in default made up $100 of $233 on a per loan basis, or approximately 43 percent, of servicing costs in 2017 for the servicers contributing to the MBA Forum Data. In other words, over two-fifths of all servicing costs are specific to the relatively small share of loans that are delinquent, which means that the costs of servicing a loan that is delinquent are much higher than the costs of servicing a loan that is current. One implication of this is that servicers' total costs are highly dependent on the number of delinquent loans in their portfolios. An increase in delinquency rates will increase servicing costs even without any change in underlying operational costs of servicing a loan. TABLE 1: DIRECT SERVICING COSTS BY FUNCTION (2017) (MBA FORUM DATA)

Function

Cost per Loan

Share of Total

Total non-default specific direct servicing costs

$107

 

46%

 

Customer Service

 

$26

 

11%

Setups, Transfers, & Payoffs

 

$9

 

4%

Escrow

 

$11

 

5%

Cashiering & Investor Accounting

 

$12

 

5%

Servicing Systems

 

$24

 

10%

Records

 

$5

 

2%

Quality Assurance

 

$5

 

2%

Other

 

$16

 

7%

Total default-specific direct servicing costs

$50

 

21%

 

Collections

 

$11

 

5%

Loss Mitigation

 

$15

 

6%

Foreclosure

 

$12

 

5%

Bankruptcy, REO, & Other Default

 

$11

 

5%

Total indirect servicing costs

$76

 

33%

 

Unreimbursed FC and REO expenses

 

$50

 

21%

Corporate administration expense

 

$26

 

11%

Total servicing costs

$233

 

100%

 

Figure 1, also based on the MBA Forum Data, looks separately at servicing costs per performing loan and servicing costs per loan in default (non-performing loans) from 2008 to 2017. The figure shows a very large increase in the per loan cost of servicing both performing and nonperforming loans during this period. The increase is especially large for non-performing loans, with an increase of nearly 400 percent in per loan costs over the period. The main increase took place between 2008 and 2013, with no clear trend between 2013 and 2017.

The Rule became effective at the start of 2014, so the increase in costs described in Figure 1 is not attributable the Rule itself. However, as described elsewhere in this report, many of the Rule's requirements are similar to requirements of the National Mortgage Settlement (NMS) and other requirements imposed by regulators and investors during the 2009 to 2012 period.

Thus, the increase in costs through 2013 may reflect in part the cost of coming into conformity with those earlier requirements, some of which became part of the Rule in 2014.

The cost increases are likely attributable to other factors as well. For example, as described in Chapter 3, there was a large increase in the average amount of time between initiating a foreclosure and completing a foreclosure sale, which may explain part of the increase in foreclosure-related costs.

FIGURE 1: AVERAGE SERVICING COSTS PER ACCOUNT, 2008-2017 (MBA FORUM DATA)

Figures 2 and 3 compare average servicing costs from the MBA Forum Data and the MBA Annual Performance Report Data since 2008. Both figures show an increase in average servicing costs from 2008 to 2017, although the trends are quite different. The MBA Forum Data show a leveling off and subsequent decrease in average costs starting in 2012, whereas the MBA Annual Performance Report Data show an increase in 2010, with little change thereafter until 2016. The overall increase appears greater in the MBA Forum Data, although this is largely driven by the low average costs for the MBA Forum Data in 2008; in dollar terms, the increase from 2009 to 2017 is similar in both data sources. Another notable difference is the much higher share of indirect costs shown in the MBA Forum Data.

The different trends in the two data sources likely represent differences in the populations of servicers covered by each source. Servicers represented in the MBA Annual Performance Report Data are non-depositories, relatively small on average, and, as discussed later in this chapter, have lower delinquency rates on average than larger servicers. Larger servicers may have been more affected by high delinquency rates prior to 2013 and may have had to do more to change practices for managing delinquent loans. Higher delinquency rates could also explain higher indirect costs, particularly unreimbursed foreclosure and REO expenses.

FIGURE 2: AVERAGE ANNUAL SERVICING COSTS PER ACCOUNT (MBA FORUM DATA)

FIGURE 3: AVERAGE ANNUAL SERVICING COSTS PER ACCOUNT (NON-DEPOSITORIES) (MBA ANNUAL PERFORMANCE REPORT DATA)

5.2 Labor costs and productivity

The cost of mortgage servicing is driven largely by labor costs. Based on the MBA Forum Data, 59 percent of direct cost in 2017 was personnel expenses. This suggests that an important part of the increase in servicing costs since 2008 is likely to be an increase in personnel expenses, either in terms of needing more employees to service the same number of loans or in terms of needing employees who are more highly-paid on average.

Figure 4 shows that, in the MBA Forum Data, the number of loans serviced per full-time equivalent (FTE) decreased from 2008 to 2013, flattening out from 2013 onward. This trend is consistent with the trends in costs shown in Figure 1, as the need for more employees to service the same number of loans translates into higher personnel cost per loan. There was an even greater drop in the average number of defaulted loans serviced per default FTE—overall average loans serviced per servicing FTE fell to about 46 percent of its 2008 level by 2017, whereas default loans serviced by default FTE fell to about 31 percent of its 2008 level. Changes in productivity are likely driven at least in part by the changes in the share of non-performing loans, with each non-performing loan requiring the work of personnel in departments that are not involved in servicing a performing loan. However, these data indicate that productivity did not increase again as the number of delinquencies fell after 2010, suggesting that productivity has remained low after 2012 for other reasons.

FIGURE 4: AVERAGE NUMBER OF LOANS SERVICED PER SERVICING FTE (MBA FORUM DATA)

Figure 5 below, using MBA Forum Data, shows that compensation for servicing personnel appears to have increased and the ratio of management to non-management FTEs has decreased. From 2008 to 2017, the total compensation (salaries, bonuses, incentives, and benefits) per FTE increased 40 percent. By way of comparison, during the same time, the average annual wage of a loan officer increased 23 percent, according to the U.S. Bureau of Labor Statistics (BLS).

FIGURE 5: SERVICER COMPENSATION AND RATIO OF NON-MANAMGENT TO MANAGEMENT FTES (MBA FORUM DATA)

5.3 Effects of the Rule on servicing costs

This section summarizes what servicers told the Bureau in interviews about the overall effect of the Rule on costs. In general, servicers had difficulties estimating how the Rule's provisions affected their costs. Servicers attributed this to a number of factors, including ongoing changes in delinquency rates, evolving investor requirements, state law changes, and changes in the size and composition of their servicing portfolios, all of which have effects on costs. Servicers' characterizations of the costs of the Rule varied considerably. Servicers said they incurred significant expense to come into compliance with the Rule initially. Some servicers said that the Rule had resulted in large ongoing costs and others said that the Rule had relatively little ongoing effect on costs. This section first discusses the one-time costs of implementing the Rule and then turns to ongoing costs of complying with the Rule.

5.3.1 One-time implementation costs

In general, servicers said that implementing the Rule required significant effort and expense over the one-year period from publication of the RESPA Servicing Final Rule to the effective date.

Overall, servicers reported that the changes required to comply with the loss mitigation evaluation, response, and appeals process provisions were the most burdensome to implement, largely due to the requirement to evaluate borrowers for all available options and provide a comprehensive decision letter. Servicers also said they incurred significant burdens to implement the error resolution and foreclosure provisions. Servicers further reported that the implementation of the loss mitigation application process provisions and the early intervention provisions were less burdensome for them than the error resolution and foreclosure provisions. Many servicers explained that they were already in compliance with at least some of the early intervention and loss mitigation application provisions because they were complying with similar GSE guidelines and other requirements that pre-dated the Rule. Servicers that were not already in compliance with these provisions said the required changes to their loss mitigation application and early intervention processes were relatively minor, such as changes to deadlines and letter content. Servicers generally said the force-placed insurance provisions were the least costly provisions to implement because they were largely compliant with these provisions before the Rule.

Overall costs of implementation

Many servicers said that they were not able to estimate the overall cost of implementing the Rule, and those that did provide them noted that the estimates were tentative. Some servicers said that the changes required by the Rule were undertaken at the same time they were implementing other changes, such as systems upgrades, that were unrelated to the Rule, and that they were not able to attribute implementation costs to the Rule specifically. Others simply said that they did not track costs related to implementation, in part because implementing the Rule provisions was mandatory.

For servicers that did provide estimates or estimated ranges of the Rule's one-time implementation costs, estimates ranged from about $1.00 per loan to about $14.00 per loan. For context, industry estimates of average annual servicing costs are approximately $250 per loan in 2014. If this range were applied to the approximately 53 million mortgage loans outstanding in 2014, it would imply that the total one-time cost for industry to comply with the Rule was in the range of $53 million to $742 million. However, servicers that did not provide estimates generally characterized the implementation as less burdensome than those that did provide estimates. Many servicers, particularly among small and mid-size servicers, said that one-time implementation costs were not that extensive and mostly reflected revisions to policies and procedures.

Some servicers, large servicers in particular, said that the largest costs they incurred arose from information technology changes. Some servicers also emphasized the need to train staff as a major source of one-time costs. Others, in particular those that said implementation costs were not that great, said that the main costs were the time of legal and compliance staff in developing policies and procedures to comply with the Rule.

Servicers generally said that a significant amount of leadership time was devoted to managing the implementation process during the year prior to the effective date. Some large servicers described weekly meetings of 10 to 25 senior staff that took place over the course of one year to oversee and coordinate implementation.

Some servicers emphasized that a large part of the cost of implementing the Rule was the opportunity cost of not making other changes to their servicing process that they believed could have helped borrowers. These servicers said that the resources they use to implement new regulatory requirements are the same resources they would use to make other improvements to their servicing. Some servicers said that the Rule had therefore meant that they delayed initiatives they thought would improve the borrower experience, such as enhancements to online or mobile communication with borrowers.

Timelines for implementation process

Although specifics varied across servicers, the implementation process generally involved the following steps:

  • A "gap analysis" or "impact analysis" to determine what changes from existing practices were required to comply with the Rule. Servicers generally said this was prepared by the compliance department in coordination with affected business units. Even if servicers determined that their existing practices complied with the Rule's provisions, servicers generally revised policies and procedures to reflect the Rule's provisions.
  • Having identified any gaps between existing practice and the Rule requirements, servicers developed and implemented a plan for closing those gaps. Servicers generally said they managed the implementation of different provisions as separate projects, although some servicers said that one person led the overall implementation process. Some servicers said they managed individual projects within affected business units, while others said they managed these projects centrally. Servicers emphasized the need for coordination across business units given that changes made by one business unit often affect others.
  • Testing systems and user testing.
  • Rollout, including technology updates and training.

These phases overlapped for different aspects of the Rule and also as servicers' understanding of the Rule requirements changed. Some servicers said that implementation continued after the effective date, in part because they revised their understanding of some Rule requirements during the course of implementation.

Particular implementation challenges

Many servicers cited complexity as a challenge in the implementation of the Rule. One source of complexity was that several changes to mortgage rules took place at the same time, including not only the various provisions of the 2013 RESPA Servicing Rule but also the 2013 TILA Servicing Final Rule and the Ability-to-Repay/Qualified Mortgage Rule. This meant that the same resources, particularly in information technology and compliance, were required to coordinate and conduct many efforts at once. It also increased complexity because of the interrelationships between different servicing systems, which means that changes made to one system had effects in other systems.

The implementation process was affected by servicers' reliance on vendors, in particular providers of software platforms for mortgage servicing. Servicers generally said they were in contact with technology vendors throughout the implementation period to understand how vendors' products would adjust to reflect the new Rule requirements, and some described participating in user forums to discuss changes to software platforms. Some servicers described their reliance on technology vendors as a challenge in implementing changes, noting that they were unable to test their revised processes until software updates were provided by vendors. In some cases servicers were dependent on vendors to help with changes to the servicer's specific implementation of the vendor's software. Some smaller servicers also said that it was sometimes a challenge to get the attention of technology vendors that were facing similar demands from many clients at once.

Some servicers said there was additional complexity associated with reconciling Rule requirements with different client and investor requirements and the need to work with clients and investors on how changes required by the Rule would be implemented.

Some servicers emphasized that the complexity of implementing the Rule was exacerbated by the number of amendments to the 2013 RESPA Servicing Final Rule during the implementation period,159 differences in how servicers and vendors interpreted some Rule requirements, and an industry interpretation of some Rule requirements that evolved during the year before the Rule was effective, partly in response to Bureau guidance provided during that period.

5.3.2 Ongoing costs

Servicers said there has been an increase in their ongoing costs as a result of the Rule, particularly the cost of servicing delinquent loans. As with one-time implementation costs, not all servicers were able to estimate the amount of this cost increase. Some servicers provided an estimated range for increased ongoing costs, and all servicers providing estimates or ranges noted that these were tentative. Larger servicers estimated that the Rule had increased annual costs by amounts ranging from approximately $3.00 to more than $11.00 per loan. For context, industry estimates of average annual servicing costs are approximately $250 to $300 per loan since 2014. If this range were applied to the approximately 52 million mortgage loans outstanding on average during 2014 to 2018, it would imply that the total ongoing cost for industry to comply with the Rule is in the range of $156 million to over $572 million annually.

The most commonly cited source of increased ongoing costs was an increase in compliance staff to monitor compliance with the Rule provisions. Some servicers also said they had to increase the number of staff in the customer relations or the loss mitigation departments as a result of the Rule. In addition to hiring additional staff, some servicers also said that because of changes they had put in place to comply with the Rule they needed to hire more highly qualified staff for certain roles, including compliance and customer service. This is consistent with the trend toward increased average compensation in Figure 5 above.

As noted in Chapter 3, the amount of time it takes servicers to complete a foreclosure has increased substantially since the crisis. This longer timeline adds costs to the foreclosure process, increasing servicer outlays and delaying the time until they are reimbursed through a foreclosure sale. Some research has estimated that, between 2005 and 2007, costs from foreclosure timelines amounted to 12 percent of the unpaid loan balance for loans defaulting, but had risen to 21 percent by 2013. Furthermore, other research has found that a borrower anticipating a longer foreclosure timeline may be more likely to default, as a result of the greater expected number of months before being required to vacate the property.

Relatedly, other research has found that the average loss-given default (loss severity) for certain conventional mortgages increased from 10 percent pre-crisis to 30 percent during the crisis, and has remained at 40 percent post-crisis. The Rule and other government interventions, as well as changes in business practices, may have contributed to this increase. Specifically, the researchers found a causal effect of both the Rule and the NMS on longer foreclosure timelines and, therefore, loss severity. The authors also found that the Rule increased loss severity unrelated to foreclosure delay by seven percentage points. Chapter 9 considers the effect of the Rule's foreclosure restriction provisions on the time from initiating a foreclosure to property sale.

5.4 The small servicer exemption and effects on small and mid-size servicers

Small servicers are exempt from certain parts of the 2013 RESPA Servicing Rule. A servicer is considered to be a small servicer if one of the following criteria applies: 1) the servicer, together with any affiliates, services 5,000 or fewer mortgage loans, and the servicer owns or originated all of them; or 2) the servicer is a Housing Finance Agency. As of the fourth quarter of 2015, the Bureau estimated that 95 percent of servicers that are depository institutions and 83 percent of servicers that are non-depository institutions serviced 5,000 or fewer loans.

Small servicers are exempt from early intervention and continuity of contact requirements; most loss mitigation procedure requirements; certain requirements relating to obtaining force-placed insurance; and the provisions relating to general servicing policies, procedures, and requirements. Representatives at community banks and credit unions interviewed by the Government Accountability Office (GAO) said that the small servicer exemption was "helpful in reducing…compliance requirements" and "has been helpful to their businesses and customers."169

Some small servicers the Bureau interviewed said that the exemption for small servicers was valuable, as smaller institutions are less likely to have the resources to track compliance with all of the Rule's requirements. However, in interviews several small servicers noted they voluntarily complied with at least some of the provisions from which they were exempt. Prior to 2014, some of these servicers had practices that were similar to, if not more stringent than, those required by the Rule. As such, these servicers maintained similar practices after the Rule became effective, but formalized their policies and procedures in areas such as early intervention, continuity of contact, and loss mitigation. One small servicer said that it voluntarily complied with the Rule in anticipation that, with fluctuations to its portfolio or future growth, it will no longer be eligible for the exemption. Others simply said that they choose to voluntarily comply with the Rule's provisions if practical.

5.4.1 Trends in costs and delinquency rates by servicer size

Figures 6 and 7 use the MBA Annual Performance Report Data to break out direct expenses per loan and serious delinquency rates, respectively, by servicer size for the non-depository servicers represented in that survey. These figures start in 2012, the first year that MBA broke out the data in these servicer size categories in the MBA Annual Performance Report Data. Note that because these data are collected from only non-depositories, their cost and delinquency rates may not be representative of those measures for depository institutions with similarly sized servicing portfolios.

Figure 6 shows that costs per loan serviced in these data were highest for the smallest servicers, and this group also had the largest increase in expenses during the period 2012 to 2017. Servicers of this size were not subject to most provisions of the Rule so long as all loans they serviced were loans they owned or originated. Servicers in the next size category—those servicing between 2,500 and 10,000 loans—also experienced increased costs per loan from 2012 to 2017. Some of those servicers (those servicing 5,000 loans or fewer) could qualify for the small servicer exemption.

Figure 7 shows that, while the rate of serious delinquency fell in all size categories from 2012 to 2017, the decline was much less pronounced among non-depository servicers that serviced fewer than 50,000 loans.

FIGURE 6: ANNUAL DIRECT EXPENSES PER LOAN, BY SERVICING VOLUME (NON-DEPOSITORIES) (MBA ANNUAL PERFORMANCE REPORT DATA)

FIGURE 7: SEVERE DELINQUENCY RATE, BY SERVICING VOLUME (NON-DEPOSITORIES) (MBA ANNUAL PERFORMANCE REPORT DATA)

5.4.2 Qualitative information: Interviews with small and midsize servicers

Broadly, the small and mid-size servicers that the Bureau interviewed said they agreed with the spirit of the Rule and believed the Rule increased consumer protection in the mortgage market. However, some of the small and mid-size servicers noted that the business model of credit unions and small banks allows them to provide more high-touch customer service and suggested that the Rule's requirements were not necessary for mid-size servicers with low delinquency rates.

Small and mid-size servicers generally had relatively low rates of serious delinquency and fewer foreclosures than larger servicers. As of the end of 2017, the small and mid-size servicers the Bureau spoke with each reported 90-day delinquency rates below 1 percent, compared to the national rate of 2.91 percent for all servicers. Even for mid-size servicers the number of foreclosure actions was relatively small. Of 12 small and mid-size servicers interviewed, none completed more than 70 foreclosures in 2017 and only one completed more than 25. The low delinquency rates reported by small and mid-size servicers are consistent with Figure 7 above. Perhaps because of the relatively small number of delinquent loans they service, small and midsize servicers said that prior to the Rule their policies and procedures related to delinquency and loss mitigation were less formal and they tracked compliance with those policies and procedures less systematically. Nevertheless, many small and mid-size servicers were already in compliance with some provisions of the Rule before its effective date, in part because they service GSE loans and the GSE guidelines include requirements similar to some Rule provisions.

Many of the small and mid-size servicers had practices that were already compliant with the early intervention and continuity of contact provisions, while others said they were following the general spirit of the requirements but had to adjust policies and procedures to fit the timelines and documentation specified by the Rule. For loss mitigation requirements, servicers noted the main changes to their policies and practices were the requirements to send a written determination letter and the need to institute a formalized appeals process, even though appeals were rare. Some servicers said that prior to the Rule their collections and loss mitigation functions were completely separate and that the Rule led them to combine these two functions or to educate collectors about loss mitigation in order to better assist borrowers.

While some small servicers were already in compliance with the force-placed insurance requirements prior to the Rule, others shifted to using a vendor or increased their oversight of their vendor to ensure compliance. One small servicer said that it shifted to using a vendor to track force-placed insurance because it would be more efficient for them to manage and oversee their vendor's work rather than tracking insurance in-house. Some small and mid-size servicers said the error resolution provisions required them to formalize policies and implement tracking systems to be in compliance, but they emphasized that their relatively small customer base allows them to be very responsive to borrower concerns and thus the Rule had little effect on their borrowers.

Small and mid-size servicers said implementing the Rule was costly but were not able to estimate the cost in dollar terms. Some servicers said the provisions of the Rule that affect technology and technology providers, such as tracking various activities and implementing specific notices with prescribed language, were more burdensome and meant costly changes to systems. For some servicers, staff training was one of the primary implementation costs. Even where Rule requirements were consistent with prior procedures, ongoing costs for these servicers have increased because of the need to add quality control and internal audit processes to ensure compliance. Other on-going costs have resulted from additional mailings and borrower outreach as required by the Rule. The Rule prescribes specific timelines and language for notices. For the small and mid-size servicers this has increased costs without necessarily affecting the servicers' approach or the borrowers' outcomes.

The findings discussed above are generally consistent with what the GAO found in interviews with community banks and credit unions that service mortgage loans. Some of these servicers told GAO that as a result of the Bureau's servicing rules (which could include both the 2013 RESPA Servicing Rule and the 2013 TILA Servicing Final Rule), they had increased staff, updated data systems, or hired third parties to assist with compliance activities. The GAO found that the share of mortgage loans serviced by community banks and credit unions has increased since 2008.175

6. Early intervention

This chapter considers the Rule's early intervention provisions, which require servicers to provide borrowers with certain information about delinquency and loss mitigation options, if appropriate or applicable, early in a borrower's delinquency. These provisions were designed to work in tandem with other provisions in the Rule to help prevent avoidable foreclosure. Main findings include the following:

  • Servicers interviewed generally said that the Rule's early intervention requirements were consistent with their prior practices and did not require substantial operational changes other than those related to tracking and monitoring compliance with the requirements. The Bureau did not determine the specific cost to servicers of complying with the early intervention requirements.
  • Survey data from 2016 indicate that, in the post-Rule period, a large share of delinquent borrowers speak to their servicers about loss mitigation options at some point.
  • Consistent with what servicers reported, the data obtained from seven servicers show little change in the timing of written notifications to delinquent borrowers between the pre-Rule and post-Rule periods. The data do show that, post-Rule, delinquent borrowers were somewhat more likely than they were pre-Rule to start applying for loss mitigation earlier in delinquency, with the share of borrowers initiating a loss mitigation application within six months of becoming 60 days delinquent increasing from 39 percent to 43 percent.

6.1 Background

Prior to the 2013 RESPA Servicing Rule, there were no federal statutory or regulatory provisions that addressed early intervention for delinquent mortgages generally. However, a number of early intervention standards existed for loans owned by the GSEs as well as for loans with insurance guarantees from government agencies such as the FHA or the VA. Freddie Mac recommended that servicers contact borrowers on the third day of delinquency and Fannie Mae had similar recommendations for borrowers at high risk of default. For FHA and VA insured loans, servicers were required to reach out within 20 days of delinquency occurring.

In addition, beginning in December of 2012, servicers participating in HAMP were expected to comply with early intervention standards that included multiple telephone calls and written notices prior to foreclosure. For servicers under the National Mortgage Settlement (NMS), servicer contact with borrowers was required to follow either the HAMP or GSE guidelines, whichever of the two guidelines called for a more expedited timeline.

Through the 2013 RESPA Servicing Rule, the Bureau incorporated many of the standards described above and intended to create a uniform minimum national standard for early intervention. The Rule includes early intervention requirements designed primarily to encourage servicers to communicate with borrowers about options for avoiding foreclosure. These communications include both written notices and attempts at live contact. Specifically, servicers must establish or make good faith efforts to establish live contact with a delinquent borrower by the 36th day of the borrower's delinquency. Live contact includes conducting an in-person meeting with the borrower or a telephone conversation, but not leaving a recorded message. In addition, promptly after establishing live contact, servicers must inform the borrower about the availability of loss mitigation options, if appropriate. Servicers must also provide a written notice by the 45th day of delinquency that includes information about the foreclosure process, housing counselors and the borrower's state housing finance authority, and, if applicable, information about loss mitigation options that may be available. The Rule includes model clauses for servicers to communicate as required by the Rule, but does not require servicers to use any particular language. The Rule also requires servicers to establish policies and procedures reasonably designed to provide accurate information to a borrower concerning loss mitigation options available to a borrower from the owner or assignee of the borrower's mortgage loan. The Rule exempts "small servicers," which includes any servicer that services 5,000 loans or fewer, all of which the servicer owns or originated, from these requirements.

Through the early intervention provisions of the 2013 RESPA Servicing Rule, the Bureau aimed to further the borrower protection purposes of RESPA, including to help borrowers avoid unwarranted or unnecessary costs and fees and to facilitate review of borrowers for foreclosure avoidance options. The Bureau noted that delinquent borrowers may not make contact with servicers to discuss their options because they may be unaware that they have options or that their servicer is able to assist them. At the same time, the Bureau was aware of evidence that the longer a borrower remained delinquent, the more difficult it could be to avoid foreclosure. Thus, the Bureau believed it was important for servicers to reach out early in a delinquency to determine whether loss mitigation options may be appropriate for a borrower and ensure that borrowers understand that they may have options to avoid foreclosure.192

This assessment evaluates how the early intervention provisions have affected communication between servicers and borrowers about their delinquency status, how such communications affect borrower outcomes, and what burdens these provisions have placed on servicers. The next section of this chapter describes what the Bureau learned from servicer interviews, comment letters on the RFI, and the Counselor Survey about how the early intervention provisions affected borrowers and servicers. The following section describes quantitative evidence on the effects of the early intervention provisions, using data from the ASMB and the Servicing Operations Data to analyze (a) how often servicers communicate with delinquent borrowers about loss mitigation options and at what point in their delinquency and (b) whether servicers' communication with borrowers are associated with different outcomes for borrowers.

6.2 Qualitative information: Servicer interviews and Counselor Survey

6.2.1 Effects on servicer practices

Servicers generally told the Bureau that the early intervention provisions did not significantly alter their practices for reaching out to delinquent borrowers. Servicers have incentives to contact borrowers when they do not make required payments and, as described in the prior section, other sources of servicing standards already required them to make similar disclosures for all or portions of their serviced portfolios.

With respect to live contact, servicers said that it was common practice before the Rule to call delinquent borrowers within the first few weeks of a missed payment, and that this practice continued after the Rule's effective date. Such calls, generally referred to as "collections" calls, focus on when payment can be expected and may lead to a discussion of loss mitigation options if the borrower indicates an ongoing difficulty in making payments. Servicers told the Bureau that, before the Rule's effective date, in most cases they were already placing collections calls well within 36 days of a missed payment, with most servicers saying they generally attempted to call borrowers starting around the 16th day after a missed payment. Most servicers said that their schedule for making outbound calls depends on the borrower. For example, borrowers with a history of delinquency or a higher estimated default risk might start receiving calls only a few days after they miss a payment. Although most servicers said either that the Rule had not affected their calling schedules or that it had led to only minor changes, some said that they had to change their calling practices to ensure that they attempted contact with all borrowers by day 36 of their delinquency. These servicers said that before the Rule they would have postponed calling some borrowers if the borrower's characteristics suggested that the borrower was likely to pay without outreach.

Some servicers said that, although the timing of their early telephone outreach did not change much, the Rule led them to revise call scripts and conduct staff training to ensure that, during their collection calls, staff inform borrowers about the availability of loss mitigation options, if appropriate. Servicers also described formalizing their procedures and implementing additional monitoring of calls and other measures to track compliance with the Rule's early intervention requirements.

With respect to written notices, most servicers reported that they were sending some form of early intervention letter before the Rule became effective and that the Rule required changes only to the content and timing of the notices. A few servicers interviewed were not sending any form of written early intervention letter before the Rule or were not sending them to delinquent second lien borrowers.

Some servicers said that the Rule led them to attempt live contact more frequently than before the Rule with borrowers who had longstanding delinquencies. Some servicers said that, absent the Rule, they did not attempt to contact some borrowers who had been delinquent for many months or when a foreclosure case was active, particularly when loss mitigation options appeared to have been exhausted.

6.2.2 Effects on borrowers

Counselor Survey: About 39 percent of respondents to the Counselor Survey said that the early intervention live contact requirements were effective, with an additional 43 percent saying they were "somewhat effective."195 Slightly more respondents said that the early intervention written notices were effective, with 42 percent describing them as "effective" and 50 percent describing them as "somewhat effective." When asked about reasons these provisions were effective or not effective, a majority of respondents said that when the provisions were effective it was because the provisions led their clients to receive information about their options in a timely manner and motivated clients to initiate a loss mitigation application. In cases where counselors believed the requirements were not effective, over 40 percent of respondents said that it was often or always because the information is not understandable to their clients, and several respondents said that notices were often incomplete, incorrect, or vague. On the other hand, with respect to written notices, several respondents said that, even if borrowers did not understand the content of the notices, counselors found them helpful for understanding their client's situation. With respect to written notices, 48 percent of respondents said that if the notice was not effective it was because the borrower attempted to contact the servicer about the notice, but the servicer did not respond to the attempt.

Servicer interviews: Servicers generally said the practice of making early intervention contacts helps borrowers to consider loss mitigation options early in their delinquency. However, they had mixed views on whether the early intervention provisions of the Rule itself had any effect on whether delinquent borrowers sought loss mitigation. Some said that the Rule had made borrowers more likely to engage with the servicer about loss mitigation early in their delinquency. Others said that there was little effect on borrowers because the early intervention provisions had changed little about their servicing practices. In addition, some servicers said that borrowers have become more aware of loss mitigation options in general since the onset of the housing crisis (perhaps due to media coverage, word of mouth, housing counseling, or other financial education sources) and are more likely to make contact with servicers to obtain assistance even absent outbound calling or letters. Also, some servicers said that many borrowers do not engage with them about loss mitigation options until foreclosure is imminent, which might limit the effectiveness of early intervention contacts for such borrowers.

Servicers expressed mixed opinions about the effectiveness of written early intervention notices. Many said that the information was helpful to borrowers. Others expressed doubt that many borrowers read the disclosures and suggested that making live contact, such as through telephone calls, is a more effective way of engaging delinquent borrowers in the loss mitigation process. Some servicers said they believed that borrowers have benefitted from greater standardization of content and timelines for written early intervention disclosures across investors and clients.

Several servicers said that early intervention communications (both live contact and written notices) generally were not helpful for borrowers with longstanding delinquencies, particularly if borrowers had already been denied for loss mitigation options. These servicers said that such borrowers are likely to have already been exposed to the disclosures and that the disclosures

they provide to comply with the Rule in some cases might even convey a false hope that options are available when in fact all options have been exhausted.

Servicers gave examples of other circumstances in which the Rule led them to make early intervention communications that they did not believe helped borrowers. Some servicers said that the practices they have adopted to comply with the Rule can lead them to make loss mitigation disclosures to borrowers who are not interested in them (e.g., cases in which a customer just wants to pay by telephone and has heard the disclosures before, or is calling to ask about a specific option and is not interested in hearing about other options). Some servicers also said the written early intervention disclosures they send have frustrated or confused some borrowers who were affected by a disaster and were given a forbearance, because the message of the notice, in particular the focus on foreclosure and delinquency, is inconsistent with the message the servicers provide such borrowers about the forbearance (i.e., that the borrower does not need to worry about making payments for a certain period of time).

6.2.3 Costs of the provisions and other effects on servicers

Servicers generally said that the early intervention provisions were among the least costly of the Rule provisions, given that prior to the Rule they were already making collections calls and, in most cases, sending written notices to delinquent borrowers. Servicers incurred one-time costs associated with setting up compliance systems and training staff on procedures to comply with the new requirements. Some servicers said they incurred ongoing costs from additional tracking and monitoring of these communications to ensure compliance with the specific Rule requirements.

Some servicers said the Rule caused them to make additional calls to borrowers, including calls to borrowers who they otherwise would not have called relatively early in the delinquency and borrowers with longstanding delinquencies. Some servicers also said that the early intervention provisions had increased average call times because staff ask borrowers more questions to determine whether loss mitigation is appropriate and spend more time discussing loss mitigation options with borrowers. Servicers also said that, if the Rule leads them to make early intervention disclosures that do not seem appropriate to the situation, borrowers may become confused and frustrated, which can mean the servicer must spend additional resources resolving questions and addressing complaints. Collectively, these factors imply that servicers need more staff to handle early intervention telephone calls, increasing costs.

Servicers that were not already sending early intervention letters pre-Rule incurred the additional expense of sending them post-Rule. Some servicers said that both they and borrowers would benefit from electronic delivery of these disclosures.

6.3 Quantitative analysis: ASMB and Servicing Operations Data

This section uses the ASMB and Servicing Operations Data to quantify: (a) how often and when servicers communicate with delinquent borrowers about delinquency and loss mitigation; (b) trends in whether delinquent borrowers initiate loss mitigation applications; and (c) the relationship between servicer communications and how borrowers' delinquencies are resolved.

The ASMB includes questions about whether a borrower had difficulty making mortgage payments in the recent past and what information borrowers obtained from servicers when they were having difficulty making mortgage payments. The questions discussed in this section were asked only of respondents who said they had had difficulty making their mortgage payments during the recent past. Because the survey began in 2016 it generally does not permit comparison of the pre-Rule and post-Rule periods. However, it provides a source of nationally representative data on what communications servicers had with borrowers about loss mitigation options after the Rule became effective. Among those respondents, results are reported separately for borrowers whose credit records indicated they were current on their mortgage one year before the survey was conducted (current borrowers) and borrowers whose credit records indicated they were delinquent one year before the survey was conducted (delinquent borrowers). Note that all respondents discussed below reported having difficulty making their mortgage payments during the recent past, regardless of whether credit records indicated they were current the prior year. The current borrowers may have resolved a past delinquency, may have encountered difficulties after the date as of which they were identified as current, or may have been able to avoid delinquency despite their payment difficulties.

The Servicing Operations Data include data on communications with borrowers related to delinquency or loss mitigation, including the timing of inbound and outbound communications and in some cases codes reflecting the content of the communication. The Bureau requested data on letters, telephone calls, and other forms of communication; however, most servicers were unable to provide complete data on telephone contacts. This limits what can be learned from the data about the early intervention live contact requirements.

6.3.1 Borrower awareness of early intervention communications

The next three tables present information from the ASMB about whether respondents who had experienced difficulties making their mortgage payment recalled certain communications from their servicer. The Rule requires servicers to make disclosures to certain delinquent borrowers about the availability of loss mitigation options if applicable as well as information about how to contact a housing counselor and certain other information. Thus, under the Rule, most delinquent respondents would be expected to have heard from their servicer on these topics, although consumers might not recall such communication at the time of the survey.

Table 1 summarizes responses to ASMB questions in 2016 about whether the borrower had gotten information from or talked with the servicer about various loss mitigation or other options that might help borrowers struggling with their mortgage payments. Among delinquent borrowers (for whom it seems most likely that the early intervention provisions would have been triggered at some point), 79 percent of respondents reported receiving information about a loan modification from their servicer, and 65 percent of delinquent borrowers reported receiving information about a way to get caught up on missed payments.

Only about one-third recalled receiving information about financial counseling.

TABLE 1: SHARE OF BORROWERS WHO REPORTED PAYMENT DIFFICULTY WHO RECEIVED INFORMATION FROM SERVICERS ABOUT VARIOUS OPTIONS, BY DELINQUENCY STATUS (PERCENT) (2016 ASMB)

Received information about:

Current

Delinquent

Refinancing the mortgage

44

47

A loan modification

60

79

Financial counseling

29

34

A way to get caught up on missed payments

37

65

Selling or giving up the property

17

42

These results indicate that, at least in the years after the January 10, 2014, effective date of the

Rule, a large majority of delinquent borrowers were aware of their servicer's efforts to provide information about modification options. This survey question did not distinguish between written and live disclosures or identify how soon in a delinquency the disclosures were received. They do indicate that servicers are generally successful at making delinquent borrowers aware of loan modification options. However, the data do not indicate whether consumers were aware of these options because of the Rule's requirements.

Although the ASMB began after the effective date of the Rule, borrowers were asked about when they "start[ed] having difficulty making mortgage payments," shedding some light on how servicer practices changed over time. Table 2 breaks out the results for delinquent borrowers according to when respondents first began having difficulties making their mortgage payments. This table enables comparison of borrowers who first experienced difficulties in 2013 or earlier, before the Rule was effective, and those who first had difficulties in 2014 or later, after the Rule was effective.

TABLE 2: SHARE OF DELINQUENT BORROWERS WHO REPORTED PAYMENT DIFFICULTY WHO RECEIVED INFORMATION FROM SERVICERS ABOUT VARIOUS OPTIONS, BY THE TIME WHEN THE BORROWER BEGAN HAVING DIFFICULTIES PAYING THEIR MORTGAGE (PERCENT) (2016 ASMB)

Received information about:

2013 or earlier

First half of 2014

Second half of 2014

2015 or later

Refinancing the mortgage

49

44

53

31

A loan modification

81

83

80

65

Financial counseling

39

22

32

29

A way to get caught up on missed payments

65

63

75

58

Selling or giving up the property

46

38

47

23

ASMB respondents who reported that they began having difficulties in 2013 or earlier do not appear to be less likely to have received information about these types of options. The responses do not indicate when the disclosures took place, and borrowers whose difficulties began before the Rule became effective may have received information from the servicer about options after the Rule's effective date. Nonetheless, these results are consistent with servicers' statements in interviews that early intervention communications took place before the Rule was effective.

In 2017, the ASMB included some more specific questions regarding contact with servicers about payment difficulties. Borrowers were asked about whether the servicer tried to contact the borrower, whether the borrower tried to contact the servicer, and whether the borrower in fact talked with the servicer. Table 3 reports results for these questions.

TABLE 3: COMMUNICATIONS WITH BORROWERS WHO REPORTED PAYMENT DIFFICULTIES, BY DELINQUENCY STATUS (PERCENT RESPONDING "YES") (2017 ASMB)

Question

Current borrowers

Delinquent borrowers

When you were having difficulties making the mortgage payment, did the servicer ever try to contact you by phone, mail, or other means?

52

92

Did you ever try to contact your servicer to discuss your difficulties?

56

87

Did you ever talk with the loan servicer or their representative?

53

90

These results indicate that, in the period after the Rule was effective, 92 percent of delinquent respondents were aware of contact attempts by the servicer, and 90 percent had in fact spoken with the servicer. For borrowers who were current but experienced difficulty paying their mortgage, about one-half were aware of contact attempts by the servicer and a slightly higher share had attempted contact themselves. The lower percentages for current borrowers are consistent with the possibility that many borrowers in this group experienced difficulties without in fact missing any payments and, therefore, did not trigger the Rule's early intervention requirements.

6.3.2 Timing of early intervention written notices

The next figure presents information about the mailing of written early intervention notices using the Servicing Operations Data. Although RESPA did not require written notices in the pre-Rule period, the Servicing Operations Data include the dates on which servicers sent notices that they considered to be equivalent during the pre-Rule period.

Figure 1 shows the distribution of the number of days from the beginning of a borrower's delinquency to the mailing of a notice, separately by servicer, using a box-and-whiskers plot in which the box represents the range from the 25th to 75th percentile and the heavy line represents the median. This analysis is limited to loans that went from current to at least 60 days delinquent during the relevant period. The analysis excludes two servicers that were unable to provide communications data for some accounts for inclusion in the Servicing Operations Data. The dotted horizontal line at 45 days corresponds to the requirement of the Rule.

Consistent with what servicers said in interviews with the Bureau, the data show little change in the timing of written notices between the pre-Rule and post-Rule periods. The median times before a written notice was sent are nearly the same for the pre-Rule and post-Rule periods at all five of these servicers, and for four out of five the median is well under 30 days.

FIGURE 1: DAYS FROM THE START OF A BORROWER'S DELINQUENCY TO THE MAILING OF FIRST WRITTEN NOTICE (SERVICING OPERATIONS DATA)

6.3.3 Changes in the rate and timing of loss mitigation applications

This section looks at the change from 2012 to 2015 in whether and when delinquent borrowers initiate loss mitigation applications. The analysis focuses on borrowers who became delinquent during 2012 or 2015—specifically, borrowers who went from being current to 60 days delinquent at some point during the first six months of the relevant year. Because the Servicing Operations Data include only applications initiated during either 2012 or 2015, the analysis restricts attention to borrowers who became delinquent during the first six months of the year to ensure that the data include all applications that are initiated within six months following the start of delinquency.

Figure 2 shows, for all borrowers who become 60 days delinquent during the applicable year, the share that six months later (a) had initiated a loss mitigation application, (b) had become current without initiating a loss mitigation application, or (c) remained delinquent and had not initiated a loss mitigation application. This figure excludes data from two servicers for which data on initiation of loss mitigation applications was unavailable or appeared unreliable for 2012 or 2015 or both. The figure shows that, for the five servicers for which the Bureau has these data, borrowers in 2015 were somewhat more likely to initiate a loss mitigation application within six months. In 2015, borrowers were also somewhat more likely to become current despite never initiating a loss mitigation application, and were less likely to remain delinquent for six months without either curing the delinquency or initiating a loss mitigation application.

FIGURE 2: OUTCOME 6 MONTHS AFTER BECOMING 60 DAYS DELINQUENT IN FIRST HALF OF YEAR (SERVICING OPERATIONS DATA)

Figure 3 shows the distribution of how long it took borrowers to initiate loss mitigation applications relative to when they became delinquent. The figure generally shows an increase in the share of borrowers who initiated loss mitigation applications within the first 60 days of delinquency. Together, Figures 2 and 3 indicate that borrowers in 2015 were more likely to initiate applications and to initiate them early in delinquency. These changes could reflect the effect of the early intervention provisions but could result from other factors.

FIGURE 3: DISTRIBUTION OF THE NUMBER OF DAYS BETWEEN DELINQUENCY AND INITIATION OF LOSS MITIGATION APPLICATION, BY YEAR (SERVICING OPERATIONS DATA)

6.3.4 Relationship between servicer contact and borrower outcomes

This section analyzes the relationship between servicer contact and borrower outcomes. As discussed, the data generally do not permit the Bureau to identify whether early intervention communications caused borrowers to seek loss mitigation options or otherwise address their delinquency. In particular, one might expect that whether a borrower engages with the servicer early in a delinquency is related to other important factors, such as whether the borrower expects to be able to cure the delinquency or whether the borrower expects foreclosure is imminent. Nonetheless, evidence that borrowers who speak with their servicer are more likely to cure their delinquency than borrowers who do not would be consistent with the early intervention provisions being effective in helping borrowers.

Figure 4 uses 2017 ASMB data to look at whether respondents said certain actions were taken to address their payment difficulties, for borrowers who did and did not report talking to their servicer when they began having payment difficulties. For delinquent borrowers, those who reported having spoken with the servicer were more likely to report that their loan was modified. They were also more likely to have lost the home through foreclosure. This suggests that in some cases contact with the borrower was driven by the foreclosure process, or perhaps that borrowers who knew their circumstances were more serious were both more likely to speak with the servicer and also more likely to lose their home to foreclosure. Borrowers who did not speak with the servicer were much more likely to report that none of the listed actions were taken to address their payment difficulties, which could mean that they were able to cure on their own or that they remain delinquent.

FIGURE 4: ACTION TAKEN TO ADDRESS PAYMENT DIFFICULTIES, PERCENT OF DELINQUENT BORROWERS (2017 ASMB DATA)

7. Completing loss mitigation applications

This chapter considers the Rule's provisions related to completing loss mitigation applications, which generally require servicers to provide certain delinquent borrowers with access to knowledgeable personnel to respond to the borrowers' inquiries, acknowledge loss mitigation applications in writing, and exercise reasonable diligence to make incomplete applications complete. Main findings include the following:

  • The data suggest that it took borrowers longer to go from initiating a loss mitigation application to completing the application in 2015 compared to 2012, before the Rule was effective. This may be because the Rule required servicers to define a complete application to be a more comprehensive package than what some servicers considered a complete application pre-Rule.
  • Despite the longer time to complete applications once they had been initiated, borrowers who submitted complete applications in 2015 did so at a similar stage of their delinquency as borrowers who completed applications in 2012.
  • Among the Rule's provisions related to completing loss mitigation applications, many servicers said the most significant and costly changes they made were to comply with the Rule's requirement to provide a five-day acknowledgment notice for loss mitigation applications. Servicers generally said that the Rule's other requirements discussed in this chapter were consistent with their prior practice and did not require substantial operational changes other than those related to tracking and monitoring compliance with the requirements.

7.1 Background

The 2013 RESPA Servicing Rule included a number of provisions that established standards for the loss mitigation application process. These provisions, among other things, established continuity of contact requirements, timelines for acknowledging receipt of an application, and the requirement that a servicer, upon receipt of an incomplete loss mitigation application, must exercise reasonable diligence in obtaining documents and information to complete a loss mitigation application for all loss mitigation options available to the borrower.

Prior to the Rule, an array of standards and guidelines existed for servicers surrounding the loss mitigation process. HAMP imposed a 10-business-day requirement for acknowledging receipt of a homeowner's loan modification package and a 30-calendar-day requirement to review documentation for completeness. Servicers covered by the National Mortgage Settlement (NMS) were required to provide written acknowledgement of receipt of documentation submitted in connection with an application within three business days and to notify the borrower of any known deficiency in the initial submission no later than five business days after receipt. Fannie Mae and Freddie Mac had guidelines for receipt of documentation and permitted the borrower 30 days from the date of the notification of any missing information or documentation to supplement the application prior to the servicer making a determination.203

Similar to the Rule's provisions, many servicer standards and investor guidelines required servicers to establish a single point of contact (SPOC) for certain borrowers. The largest 20 HAMP participants were required to provide a SPOC for all homeowners with non-GSE loans who were evaluated for HAMP. The NMS, which generally took effect in mid-2012 and expired in late 2015, required covered servicers to establish a SPOC and undertake outreach efforts to communicate loss mitigation options for first lien mortgage loans to all potentially eligible delinquent borrowers, following the timelines established by HAMP.205

The 2013 RESPA Servicing Rule includes provisions related to completing applications that are similar to the other sources of servicing standards. Particular features of the loss mitigation provisions related to completing applications include the following:

  • The Rule's "continuity of contact" provision requires servicers to maintain policies and procedures that are reasonably designed to provide certain delinquent borrowers with access to knowledgeable personnel who can respond to the borrowers' inquiries and, as applicable, to assist them with available loss mitigation options. These policies and procedures must include assigning personnel to a borrower no later than the 45th day of the borrower's delinquency.207
  • In general, the Rule requires servicers to acknowledge loss mitigation applications in writing. If a servicer receives a loss mitigation application 45 days or more before a foreclosure sale, the Rule requires the servicer to, promptly upon receipt, review the application to determine if the application is complete and notify the borrower in writing within five days after receiving the application that the servicer has determined that the application is complete or incomplete and, in the case of an incomplete application, state in the notice the additional documents and information the borrower must submit to complete the application and the applicable due date for the materials.
  • The Rule defines a complete loss mitigation application as an application in connection with which a servicer has received all the information that the servicer requires from a borrower in evaluating applications for loss mitigation options available to the borrower.
  • The Rule generally requires servicers to exercise reasonable diligence in obtaining documents and information to complete a loss mitigation application. For example, if a servicer needs additional information from the applicant, such as an address or a telephone number to verify employment, the servicer exercises reasonable diligence by contacting an applicant promptly to obtain the missing information. There are some exceptions to the reasonable diligence requirements, e.g., a servicer may offer a shortterm payment forbearance program to a borrower based upon an evaluation of an incomplete loss mitigation application.

The Rule exempts "small servicers," which are servicers that service 5,000 loans or fewer, all of which they own or originated, from all of the requirements discussed above.

As with the other provisions of the Rule related to loss mitigation, the Bureau established the Rule's provisions related to completing applications as part of its efforts to establish a fair process for review of loss mitigation applications to help borrowers avoid foreclosure when possible. The Bureau believed that these restrictions were necessary and appropriate to achieve the consumer protection purposes of RESPA, including facilitating the review of borrowers for foreclosure avoidance options. In addition, the Bureau adopted these provisions to implement a servicer's statutory obligation to take timely action to correct errors relating to avoiding foreclosure by establishing servicer duties and procedures that must be followed to avoid errors with respect to foreclosure. The Bureau also intended these provisions to further the goals of the Dodd-Frank Act to ensure a fair, transparent, and competitive market for mortgage servicing. Particular goals of the provisions covered in this chapter include:

Opportunity for loss mitigation evaluation. The loss mitigation provisions were motivated in part by concerns that some servicers were doing an inadequate job of communicating with borrowers regarding loss mitigation options, and that some servicers were unwilling to work with borrowers to reach agreement on loss mitigation options.221 The Bureau intended these provisions to help ensure that borrowers have a full and fair opportunity to be evaluated for loss mitigation options.

Access to knowledgeable personnel. With respect to the continuity of contact provisions in particular, the Bureau stated that the ability of borrowers to speak with knowledgeable personnel about available loss mitigation options and the actions necessary to obtain such options would make it easier for borrowers to effectively pursue loss mitigation.

Servicer responsiveness to borrowers. The Bureau intended these provisions to improve servicer responsiveness to borrowers seeking assistance with loss mitigation. The Bureau said it adopted the continuity of contact requirement in particular in response to widespread reports of communication breakdowns between servicers and delinquent borrowers, which the Bureau understood to be an impediment to the success of foreclosure mitigation programs.

The Bureau's assessment of the continuity of contact, written acknowledgement, and reasonable diligence provisions sought to understand how these requirements have affected the likelihood that borrowers complete applications and the timing for completing applications. The next section of this chapter describes what the Bureau learned from servicer interviews, comment letters on the RFI, the Counselor Survey, and consumer complaints about these provisions' effects on borrowers and servicers. The following section uses the Servicing Operations Data to analyze (a) the share of initiated applications that are completed, (b) the timing from application initiation to completion, (c) borrowers' delinquency status when they complete applications, and (d) outcomes for borrowers who initiate and complete applications.

7.2 Qualitative information: Servicer interviews, Counselor Survey, and consumer complaints

7.2.1 Effects on servicer practices

About one-half of the 13 large servicers the Bureau interviewed said that prior to the Rule they provided borrowers with a written notice acknowledging receipt of loss mitigation documents. Large servicers who serviced loans subject to HAMP and the NMS said they provided a written notice on the timeline required by those programs. Most of the six mid-size servicers interviewed said they did not send written acknowledgement notices prior to the Rule. One mid-size servicer said it started sending written acknowledgement notices during the housing crisis because it found that the written notice reduced the number of telephone calls from borrowers seeking to confirm that the servicer received their application. Some of the servicers that did not provide borrowers with a written acknowledgement notice prior to the Rule said that they instead acknowledged the receipt of application materials by telephone.

Even for servicers that acknowledged application materials in writing before the Rule, the Rule required updates to the content of the notice. Most servicers that were already acknowledging applications by telephone continued the practice post-Rule; thus, borrowers who submit loss mitigation documents to these servicers now receive a written acknowledgement notice and a telephone call. Some large servicers said that they send a written acknowledgement notice after the receipt of each document related to a loss mitigation application.

About one-half of the large servicers and some of the mid-size servicers the Bureau interviewed used some form of continuity of contact prior to the Rule, for instance by assigning delinquent borrowers either an individual or a team to facilitate completion of their loss mitigation applications. Most servicers that had a single point of contact model in place prior to the Rule had implemented it as a result of HAMP or NMS.

Interviews revealed little concrete information about how servicers' reasonable diligence efforts have changed as a result of the Rule. Many servicers said they made an effort pre- and postRule to make incomplete applications complete, but servicers did not provide detail about how the frequency or quality of those efforts have changed. As discussed in Chapter 8, some servicers said the Rule led them to collect more documents than they would have otherwise because of the requirement to evaluate borrowers for all available loss mitigation options simultaneously.

7.2.2 Effects on borrowers

Counselor Survey. About 56 percent of respondents to the Counselor Survey said the written acknowledgement notices required by the Rule were effective in helping their clients, with an additional 36 percent saying they were "somewhat effective." When asked about reasons this provision was effective, almost two-thirds of the respondents stated that if the written acknowledgement notice is effective, it is often or always because it prompts clients to take action to complete their loss mitigation application. When asked about reasons this provision was not effective, more than one-half of the respondents said that it is often or always because the notice is not specific enough. In responses to open-ended questions, several respondents stated their clients do not understand the written notice, but counselors and attorneys can interpret the notice for them which helps their clients understand how to complete their applications. Additionally, several respondents stated that, in spite of the acknowledgment notice, servicers still exercise a degree of subjective control over determining when an application is complete, making it challenging for borrowers, counselors, and attorneys to determine whether a servicer violated the foreclosure restriction provisions described in Chapter 9.

Ninety percent of housing counselors but only 68 percent of legal aid attorneys said the continuity of contact requirements were at least somewhat effective in helping their clients. More than one-half of counselors stated that if the continuity of contact requirements are effective it is often or always because the requirements help the borrower obtain accurate information about their loss mitigation application. In cases where the requirements are not effective, the most common reason selected by both housing counselors and legal aid attorneys was that the respondents or their clients have difficulty reaching servicer personnel or obtaining a timely response. All legal aid attorneys surveyed said that was at least sometimes the case. Additionally, all legal aid attorneys surveyed said that if the continuity of contact requirements are not effective, it is at least sometimes because servicer personnel are not able to provide accurate information about the client's loss mitigation application. This indicates that communication issues between borrowers and servicers still exist post-Rule.

Servicer interviews. The majority of servicers that did not send a written acknowledgement notice prior to the Rule stated that the notices are not helpful to borrowers because servicers were proactively contacting borrowers about applications by telephone before the Rule took effect. Other servicers said they believe the written acknowledgement notice is beneficial to borrowers. One servicer noted that the written acknowledgement notice helped their staff track missing information in incomplete loss mitigation applications. This servicer reported that its personnel refer to the notice when speaking with borrowers on the telephone about their application.

A number of servicers said that the SPOC model is beneficial for borrowers. One servicer that uses an individual SPOC model said the model allows borrowers to build trust and a relationship with one servicer staff person. Additionally, a number of servicers said they believe telephone conversations are more effective than written communications at helping borrowers complete their application. One servicer said borrowers express confusion about how to fill out application documents and speaking with a knowledgeable customer service representative by telephone is helpful.

Complaint data. Figure 1, below, shows results of an analysis of a random sample of consumer complaints from 2012 and 2015 related to loss mitigation. The Bureau manually reviewed complaint records of 400 individuals who submitted complaints about loss mitigation to the Bureau, including the narrative complaint by the borrower, the servicer's response, and any supporting documents submitted by the borrower or the servicer. Figure 1 shows the share of these borrowers that raised certain concerns about the loss mitigation application process. In 2012 about 35 percent of consumers in the sample described having to submit documents multiple times or complained of lost documentation; this share fell to 25 percent in 2015. Similarly, in 2012, 31.5 percent of consumers in the sample described difficulty communicating with someone knowledgeable about their application (e.g., a SPOC who did not return calls or did not respond to written messages). In 2015, this share fell to 27.5 percent. The figure indicates that there may have been some improvement in servicer-borrower communication from 2012 to 2015, but communication problems related to loss mitigation still exist. This is consistent with the housing counselor and legal aid attorney survey results described above.

FIGURE 1: SHARE OF CONSUMERS WHO DESCRIBED HAVING TO SUBMIT DOCUMENTS MULTIPLE TIMES AND WHO RAISED CONCERNS ABOUT THE LACK OF RESPONSIVENESS BY THEIR SERVICER, 2012 AND 2015 (BUREAU COMPLAINT DATA)

7.2.3 Costs of the provision and other effects on servicers

A few large servicers reported challenges in meeting the five-day deadline for sending the written acknowledgement notice. These servicers said that at times they send out notices that are inaccurate because they are rushing to meet the deadline. A few servicers who did not previously send a written notice said that adding the additional letter increased their mailing costs.

Some servicers said they had to hire additional staff or train existing staff to serve as SPOCs, as a result of the Rule, which may have led to an increase in personnel costs. Some servicers said they added a role of a "processor," partially as a result of the Rule, who evaluates applications for completeness before the application is sent to the underwriter. This person is typically responsible for ensuring the acknowledgement notice is sent within five days. At some servicers, the SPOC reviews applications for completeness and at others the SPOC and processor roles are performed by different personnel.

7.3 Quantitative analysis: Servicing Operations Data

To quantitatively assess the effects of the Rule provisions related to completing loss mitigation applications, the Bureau analyzed the information on borrowers' loss mitigation experience in the Servicing Operations Data. For loans included in the Servicing Operations Data, the data include information on all loss mitigation applications that borrowers started in 2012 and in 2015. The loss mitigation data include, among other information, the date the borrower first reached out by telephone to the servicer and requested help, the date the borrower first submitted written documents related to an application, and the date the borrower submitted a complete application.

As noted in Chapter 6, most servicers were unable to provide complete data on telephone contacts. Thus, the Bureau was not able to learn how the frequency and timing of telephone communications were affected by the Rule or whether certain telephone communication patterns affect the likelihood that a borrower completes an application. Similarly, only a few servicers provided complete data on the written acknowledgement notices. Among those servicers, all sent a written acknowledgement notice in 2012 and there was little variation in practices for the timing of the notice in 2015. Therefore, the Bureau was not able to observe how receipt of the written acknowledgement notice affects borrower behavior.

7.3.1 Rate of applications completed

The Rule's loss mitigation procedures, such as the continuity of contact and reasonable diligence requirements, were intended to facilitate the review of borrowers for loss mitigation, and one way they might do so is by increasing the share of initiated applications that borrowers ultimately complete. As shown in Chapter 6, the share of delinquent borrowers who initiated a loss mitigation application increased somewhat in 2015 compared to 2012. This section looks at the next step in the loss mitigation application process by analyzing the share of initiated applications that are completed. Figure 2, below, shows the percentage point change in the share of initiated applications that were ultimately completed in 2012 and in 2015, separately by servicer.

Figure 2 shows that for two of six servicers there was a decrease in the share of initiated applications that borrowers completed and for one servicer there was almost no change. However, there are two factors unrelated to the provisions themselves that may have reduced the number of completed applications observed in the 2015 data. First, servicers' definition of a complete application is likely more strict in the post-Rule period. Given that the Rule requires servicers to collect documents needed to evaluate borrowers for all available options, servicers may need to collect more documents before an application is considered complete; therefore, some applications may have been considered complete in 2012 that would not have been considered complete in 2015.235 Second, the introduction of the GSE streamlined modification program in July 2013, which was available to borrowers with loans owned or guaranteed by Fannie Mae and Freddie Mac, affected how some borrowers obtained modifications. Beginning

July 1, 2013, servicers of GSE loans were required to offer eligible borrowers who were at least 90 days delinquent a way to modify their mortgage without requiring financial or hardship documentation. Servicers were required to solicit borrowers for the streamlined modification program even if the borrower had started a loss mitigation application. The GSE streamlined modification was prevalent in 2015, and some borrowers who initiated but never completed an application in 2015 likely never completed it because they received a streamlined modification. Therefore, although Figure 2 shows that the share of applications that were completed increased for some servicers from 2012 to 2015, it is not surprising that the share decreased for others.

FIGURE 2: PERCENTAGE POINT CHANGE FROM 2012 TO 2015 IN SHARE OF LOSS MITIGATION APPLICATIONS THAT ARE ULTIMATELY COMPLETED, BY SERVICER (SERVICING OPERATIONS DATA)

7.3.2 Time from application start to completion

The previous section considers whether or not borrowers who initiated applications ultimately completed them. Figure 3, below, combines data from five servicers to show the distribution of the number of days between the date the borrower first submitted written documents in connection with an application and the date the borrower completed the application.

Pooling the five servicers together, the median number of days increased from 36 in 2012 to 63 in 2015. As noted in section 7.3.1, timelines from initiation to completion may be longer in 2015 because it may have taken borrowers longer to submit the documents required for an application to be considered complete in 2015 relative to 2012. Some servicers also said in interviews that the Rule led them to encourage borrowers who express any interest in loss mitigation to fill out application documents earlier than they would have before the Rule. These servicers suggested that this encouragement means application initiation happens earlier postRule than pre-Rule, but that in some cases accelerating the initiation of loss mitigation did not shorten the time it took borrowers to provide a complete application.

FIGURE 3: DISTRIBUTION OF THE NUMBER OF DAYS BETWEEN FIRST WRITTEN DOCUMENTS SUBMITTED AND APPLICATION COMPLETION DATE, BY YEAR. AGGREGATE DATA FOR FIVE SERVICERS. (SERVICING OPERATIONS DATA)

7.3.3 Delinquency status at application completion

As discussed above, a goal of the loss mitigation provisions was to increase the likelihood that delinquent borrowers are considered for a loss mitigation plan that could help them avoid foreclosure. Data from HAMP indicate that loans modified earlier in the borrower's delinquency are less likely to re-default. Figure 4, below, shows the percentage of applications completed in 2012 and 2015 by the borrower's delinquency status on the application completion date. The sample for this analysis is limited to loans that went from being current to 60 days delinquent during 2012 or 2015 to focus on borrowers who were early in delinquency and applying for the first time. Among these borrowers, about one-third completed their application when they were current or less than one month past due in both years. The share of borrowers who completed applications when they were less than 60 days past due was about the same in 2012 and 2015. However, the share of borrowers who completed an application when they were 60 to 89 days past due was slightly higher in the post-Rule period, with a corresponding decrease in completions when borrowers were 90 or more days delinquent. Overall, 71 percent of applications from these borrowers were completed when the borrower was less than 120 days past due in 2012 compared to 73 percent in 2015. While Figure 3 above shows the time from initiation to completion has increased post-Rule, Figure 4 shows that despite the increased timeline, borrowers generally completed applications at the same time in their delinquency post-Rule compared to pre-Rule.

Note that the Servicing Operations Data only include information about loss mitigation applications initiated in 2012 and 2015. Therefore, Figure 4 is necessarily limited to borrowers who initiated an application before the end of the year in which they became 60 days delinquent. There are likely additional borrowers who initiate and complete applications later in their delinquency. Of all applications initiated in 2012 (not just those where the borrower became delinquent during that year), 24 percent were initiated when the borrower was more than one-year past due. Of applications initiated in 2015, 22 percent were initiated when the borrower was more than one-year past due.

FIGURE 4: DAYS PAST DUE AT LOSS MITIGATION APPLICATION COMPLETION DATE FOR BORROWERS WHO TRANSITIONED FROM CURRENT TO DELINQUENT IN 2012 OR 2015 (SERVICING OPERATIONS DATA)

7.3.4 Outcomes for delinquent borrowers who initiate and complete applications

Previous sections of this chapter examine whether and when borrowers who initiate applications complete them. This section first considers the status of delinquent borrowers who initiate a loss mitigation application one year after they initiate the application. Next, for those borrowers who completed an application, it considers their status six months after the completion of the application.

Figure 5, below, includes borrowers who transitioned from current to 60 days past due in the first half of 2012 or 2015 and initiated a loss mitigation application in that year. Figure 5 shows the share of such borrowers in each year who, 12 months later: (1) had not completed their application and were still delinquent (or foreclosed upon); (2) cured without completing their loss mitigation application; or (3) completed their application. Borrowers in each year are categorized into one of the three groups such that the sum of the three columns in a given year is 100 percent. The figure shows that borrowers who became delinquent and initiated an application in 2015 were less likely than those in 2012 to complete the application, but this coincided with an increase in the likelihood of curing the delinquency without completing an application. This is consistent with improved economic conditions in 2015 allowing a larger share of delinquent borrowers to recover without loss mitigation compared to 2012. It is also possible that some of the borrowers in 2015 who cured without completing an application received streamlined modifications, which cannot be consistently identified in the Servicing Operations Data. The share of borrowers who had neither completed an application nor recovered from delinquency 12 months later was almost constant between the two years.

FIGURE 5: OUTCOMES 12 MONTHS LATER FOR DELINQUENT BORROWERS WHO INITIATE LOSS MITIGATION APPLICATIONS (SERVICING OPERATIONS DATA)

The Servicing Operation Data allow the Bureau to track delinquent borrowers who completed an application (those in the column on the right in Figure 5) and observe their status six months after they completed the application. Figure 6, below, shows that borrowers who became delinquent in 2015 and completed an application were somewhat more likely than delinquent borrowers who completed an application in 2012 to be approved for loss mitigation and somewhat less likely to still be delinquent or have experienced foreclosure six months after completing the application.

The Rule did not establish requirements for loss mitigation evaluation criteria nor did it require servicers or investors to offer a specific type of loss mitigation.248 The increase in the share of borrowers approved for loss mitigation could be due to a change in the types of borrowers that completed applications in 2015 compared to 2012. Improved economic conditions in 2015 may mean that delinquent borrowers in 2015 were less distressed on average than delinquent borrowers in 2012 and thus were better candidates for modifications. Additionally, the Rule created a definition of a complete application that was more comprehensive than the set of documents that may have been considered a complete application in 2012. This could lead to a selection effect: borrowers who were able to assemble and submit a complete application in 2015 may have been better candidates for modifications than applicants in 2012 who could have been evaluated based on a more limited application. Finally, the increase in approval rates could be related to the introduction of streamlined modifications. Borrowers in 2015 may have completed an application for a non-streamlined modification and been approved for a streamlined modification. Approval for GSE streamlined modification was not based on an evaluation of financial or hardship documentation, so borrowers who may not have qualified for other types of loss mitigation may have been approved for streamlined modifications. Overall, Figure 5 and Figure 6 show that delinquent borrowers in 2015 were somewhat less likely to complete an application, but those that did were somewhat more likely to be approved for a loss mitigation option compared to those in 2012.

FIGURE 6: OUTCOMES SIX MONTHS LATER FOR DELINQUENT BORRROWERS WHO COMPLETE APPLICATIONS (SERVICING OPERATIONS DATA)

8. Evaluations and appeals

This chapter considers the Rule's loss mitigation application evaluation and appeal provisions, which generally require servicers to evaluate applications within 30 days of receipt, evaluate borrowers simultaneously for all loss mitigation options available to them, and review borrower appeals of loan modification denials. Main findings include the following:

  • A larger share of borrowers received a decision on their loss mitigation application within 30 days in 2015 compared to 2012, before the Rule was effective.
  • The data suggest that the time from borrower initiation of a loss mitigation application to short-sale offer increased in 2015 compared to 2012, before the Rule was effective. In interviews, some servicers said this was likely due to the additional time required to collect the documents necessary to evaluate borrowers for all available loss mitigation options at the same time. Other servicers suggested that the increase in short-sale timelines may reflect an increase in the length of short-sale marketing periods post-Rule.
  • Among the provisions discussed in this chapter, many servicers said the most costly provisions were the requirements to evaluate borrowers for all available loss mitigation options at the same time and to provide a decision letter that describes the outcome of an evaluation for all available options. Servicers generally said that the 30-day evaluation timeline and appeal requirements were consistent with their prior practice and did not require significant operational changes other than tracking and monitoring compliance with the requirements.
  • The data show that a larger share of borrowers who completed loss mitigation applications appealed the servicer's determination in 2015 compared to 2012, before the Rule was effective. The proportion of appeals that were successful was lower post-Rule. The net effect was that there was no increase in the likelihood that a borrower whose application was denied successfully appealed that denial.

8.1 Background

Prior to the adoption of the 2013 RESPA Servicing Rule, federal regulations did not address servicers' procedures to evaluate loss mitigation applications. However, other national mortgage servicing standards, such as those adopted by FHFA, the National Mortgage Settlement (NMS), and HAMP, existed at that time for certain servicers. The loss mitigation evaluation and appeal standards established by the 2013 RESPA Servicing Rule generally tracked these other national mortgage servicing standards. The FHFA Servicing Alignment Initiative, the NMS, and HAMP all required servicers to review loss mitigation applications within 30 days of receiving a complete application. The FHFA Servicing Alignment Initiative also required servicers of GSE loans to consider borrowers for all loss mitigation options simultaneously. The NMS likewise required servicers to evaluate borrowers who had timely submitted a complete loan modification application "for all available loan modification options for which they are eligible" before referring a borrower for foreclosure, but did not require simultaneous review for home retention options, like loan modifications, and non-home retention options, like short sales. HAMP, the FHFA Servicing Alignment Initiative, and the NMS all provided for borrower appeals of loan modification denials.

The 2013 RESPA Servicing Rule sets certain deadlines and procedures concerning review of loss mitigation applications, but does not mandate specific loss mitigation programs or outcomes.255 In general, if a servicer receives a complete loss mitigation application more than 37 days before a foreclosure sale, the servicer must evaluate it within 30 days for all loss mitigation options available to the borrower and provide the borrower a written notice stating which loss mitigation options, if any, it will offer to the borrower. Servicers generally may not make a loss mitigation decision based upon evaluation of any information provided by a borrower in connection with an incomplete loss mitigation application. If the servicer denies the borrower for any trial or permanent loan modification option, the servicer must include in the denial notice, among other information, the specific reasons for denial for each trial or permanent loan modification option denied. The servicer must allow borrowers to appeal decisions regarding denials for trial or permanent loan modifications if a complete loss mitigation application is received 90 days or more before a foreclosure sale; in such cases the servicer must permit appeals made within 14 days after the servicer provides the determination to the borrower. The Rule also requires servicers to establish certain policies and procedures reasonably designed to properly evaluate a borrower who submits an application for a loss mitigation option for all loss mitigation options for which the borrower may be eligible pursuant to any requirements established by the owner or assignee of the borrower's mortgage loan. The Rule exempts "small servicers," which includes any servicer that services 5,000 loans or fewer, all of which the servicer owns or originated, from these requirements.

Through the loss mitigation evaluation and appeal provisions of the Rule, the Bureau sought to establish a fair process for review of loss mitigation applications to help borrowers avoid unnecessary foreclosure. The Bureau sought to further the consumer protection purposes of RESPA, including providing borrowers with timely access to accurate and necessary information regarding an evaluation for a foreclosure avoidance option and facilitating the evaluation of borrowers for foreclosure avoidance options, as well as the goals of the Dodd-Frank Act to ensure a fair, transparent, and competitive market for mortgage servicing. In addition, the Bureau adopted these provisions to implement a servicer's statutory obligation to take timely action to correct errors relating to avoiding foreclosure by establishing servicer duties and procedures that must be followed to avoid errors with respect to foreclosure. Particular goals of the evaluation and appeals provisions include:

  • Evaluation timelines. In the 2013 RESPA Servicing Final Rule, the Bureau stated that establishing the timelines described above was necessary to protect borrowers and achieve the consumer protection purposes of RESPA, including facilitating borrowers' review for loss mitigation options and avoiding errors with respect to foreclosure.267
  • Simultaneous review for all available options. The Bureau intended that the requirement that a servicer evaluate a borrower for all loss mitigation options available to the borrower, in combination with the notice requirements, would enable borrowers to understand (1) the loss mitigation options for which the servicer determined the borrower is eligible, (2) the results of the servicer's evaluation for any loan modification option, and (3) the reasons for denial for any loan modification option. The Bureau also believed that this would mitigate information disparities between servicers and borrowers. Previously, servicers often required borrowers to select a loss mitigation option for which the borrower wanted to be considered, but the Bureau believed that servicers were in a better position than borrowers to determine the loss mitigation programs for which borrowers may qualify. The Bureau believed that without the approach adopted in the Rule, borrowers may not have considered or pursued some options. The Bureau believed that requiring servicers to evaluate borrowers for all loss mitigation options available might avoid different outcomes for similarly situated borrowers due to the options for which individual borrowers applied. Likewise, the Bureau hoped to eliminate the need for borrowers to submit multiple applications for each possible alternative for which the borrower was potentially eligible, which was prevalent prior to the housing crisis and led to longer timelines for evaluating loss mitigation applications.
  • Appeals. Regarding the appeals provisions, the Bureau stated borrowers would benefit from the opportunity to have an independent review at a servicer if the borrower believes a mistake was made in the evaluation of a loan modification option resulting in an incorrect decision.

The Bureau stated that its goal was not to achieve any particular target with respect to the number or speed of foreclosures, but to ensure that borrowers are protected from harm in the process of servicers evaluating them for loss mitigation options and proceeding to foreclosure.

The Bureau's assessment of the evaluation and appeal provisions sought to understand how these Rule provisions have affected evaluation timelines for borrowers, adequacy of borrower information about the options available to them, borrower appeals of loss mitigation decisions, and the burdens these provisions have placed on servicers and borrowers. The next section of this chapter describes what the Bureau learned from servicer interviews, comment letters on the RFI, and the Counselor Survey about the evaluation and appeal provisions' effects on borrowers and servicers. The following section uses the Servicing Operations Data and GSE Data to analyze (a) evaluation timelines before and after the Rule was effective, (b) the effects of the requirement to evaluate borrowers simultaneously for all available options, and (c) borrowers' use of the appeals process and outcomes of appeals before and after the Rule's effective date.

8.2 Qualitative information: Servicer interviews and Counselor Survey

8.2.1 Effects on servicer practices

Servicers generally said in interviews that they were already evaluating loss mitigation applications within 30 days of receipt prior to the Rule and did not report challenges with meeting the 30-day deadline. Most small and mid-size servicers interviewed said that they generally review applications in one to four days, unless there is a short-term increase in application volume, which can occur in the event of natural disasters. Other servicers stated that evaluations typically take two to four weeks to complete. All large servicers the Bureau interviewed said they were sending written loss mitigation decision letters prior to the Rule.

Prior to the Rule, servicers said that they would review borrowers only for the option the borrower requested on their application or servicers would review borrowers for options on a rolling basis as they received the required documents. Of the small and mid-size servicers the Bureau interviewed, some stated they considered borrowers for all loss mitigation options simultaneously prior to the Rule. None of the large servicers the Bureau interviewed said they reviewed borrowers for all available loss mitigation options simultaneously prior to the Rule. A number of servicers cited the requirement to collect documents necessary to evaluate borrowers for all loss mitigation options available to the borrower as the most significant change they made as a result of the Rule.

Most large servicers reported they had an appeals process in place prior to the Rule, but they did not follow a standard set of appeal policies and borrowers may not have been aware of their appeal rights. Some form of an appeal process was required by the FHFA Servicing Alignment Initiative, HAMP, and the NMS. Many servicers reported that they initially designed their appeal process around the requirements for loans covered by those programs and expanded the process to encompass their entire portfolio as a result of the Rule. Some large servicers noted that they made changes to their processes to comply with the Rule's prohibition on making the first notice or filing for foreclosure until after the borrower has exhausted the appeal process in cases where the borrower submits a loss mitigation application and an appeal before the servicer initiates the foreclosure proceeding. Small servicers generally stated they did not have an appeal process prior to the Rule. These servicers consistently reported receiving few appeals in recent years; most small servicers said they had received fewer than five appeals since 2014. Several small servicers suggested appeal volumes are low because servicers exhaust loss mitigation options before denying an application to avoid foreclosure-related costs.

8.2.2 Effects on borrowers

Counselor Survey. Fifty-nine percent of housing counselors surveyed said that the 30-day evaluation timeline was effective in helping their clients, with another 34 percent saying the timeline requirement was "somewhat effective." Respondents said that when this requirement is effective it often or always helps their clients obtain a timely decision on their loss mitigation application. When asked about cases in which the provision is not effective, most respondents said it is often or always because the servicer takes more than 30 days to evaluate the application.

Less than one-half (47 percent) of housing counselors and legal aid attorneys surveyed said the appeal requirements are effective in helping their clients, with another 39 percent saying they were "somewhat effective." Given that servicers report receiving few appeals, it may not be surprising that a relatively large share of counselors report the provisions are only "somewhat effective." In responses to open-ended questions several respondents stated that servicers rarely or never reverse loss mitigation decisions based on an appeal, and a few said that notices of errors were a more effective method of appealing a decision.

Servicer interviews. Some servicers said the Rule created a consistent, relatively quick evaluation timeline, which is beneficial to borrowers. Some servicers said that borrowers sometimes express frustration to servicers about the process of collecting documents needed to complete an application that can be used to evaluate borrowers for all available options. For example, these servicers said borrowers who express interest in only a non-retention option (e.g., a short sale) express frustration when servicers request certain information, such as income documentation, which is required to evaluate the borrower for a modification but is not relevant for the non-retention option evaluation. One servicer said some borrowers may benefit from learning about options other than the one they were originally seeking, but the servicer has not observed this to be true in practice. On the other hand, some small and mid-size servicers who were already evaluating borrowers for all options simultaneously prior to the Rule said this practice is beneficial to borrowers and improves borrowers' ability to resolve their delinquencies.

Servicers the Bureau interviewed had varied opinions about the effects of listing the denial reason for each option the borrower was not approved for in the determination letter. Some servicers said that in their experience this causes borrower confusion. For example, these servicers reported that borrowers who are seeking non-retention options express confusion when they are denied for a loan modification and vice versa. However, other servicers believed that the added specificity in the determination notice is beneficial for borrowers. These servicers said increased transparency into the evaluation process is useful for borrowers; in their experience, borrowers today demonstrate a clearer understanding of what they were evaluated for and why they may have been denied than borrowers prior to the effective date of the Rule. One servicer said that in its experience, providing the borrower the reasons for denial makes it less likely that a borrower will appeal because the borrower better understands the reason for denial.

Most of the large servicers the Bureau interviewed said it is uncommon for a loan modification denial to be reversed as a result of an appeal. In cases where denials are reversed, it is typically because the borrower provided new income documentation as part of the appeal.

8.2.3 Costs of the provision and other effects on servicers

As mentioned in Chapter 5, servicers reported that the loss mitigation provisions were the most burdensome to implement, largely due to the requirement to evaluate borrowers for all available options and provide a comprehensive decision letter. Some of the large servicers the Bureau interviewed emphasized that developing the technological process to insert the appropriate denial reason language into loss mitigation determination letters was a major challenge and required significant labor hours to implement. One large servicer that uses a proprietary data system to track loss mitigation activity said that updating the content of the determination letters involved modifying their data system such that it captures the results of each evaluation, compares each result to the others and outputs the final decision in a format that is transferrable to the determination letter. Developing this process required repeated quality control testing. One servicer reported hiring an external contractor to assist with this aspect of implementation. Another servicer said that after each determination letter is sent to a borrower, the single-point of contact calls the borrower to explain the information in the determination letter, because the letter is so detailed. Additional time on the telephone with borrowers will generally increase personnel costs.

As mentioned in Chapter 5, servicers cited the Rule's complexity as a challenge in the implementation of the Rule. The Rule provisions related to loss mitigation determination letters are an area in particular where servicers described challenges in interpreting the Rule's requirements. Finally, borrower frustration about the collection of documents needed to complete an application, as described above, likely requires servicers to spend additional time on the telephone with borrowers addressing their concerns, which increases costs.

8.3 Quantitative analysis: Servicing Operations Data and GSE Data

To quantitatively assess the effects of the Rule's provisions related to evaluations and appeals, the Bureau primarily analyzed the information on borrower loss mitigation experience available in the Servicing Operations Data. For certain analyses, the Bureau also drew upon the GSE Data. The Servicing Operations Data include information on all loss mitigation applications that borrowers in the sample started in 2012 and all applications started in 2015. The same borrower could appear in both years if he or she started at least two separate loss mitigation applications, one in 2012 and one in 2015. The data include the date the borrower submitted a complete application (for applications that were completed); the appeal date (if the borrower appealed); an indicator for whether the appeal was successful; indicators for whether the borrower was considered for a loan modification, short term repayment plan, short sale, or deed-in-lieu of foreclosure; and the offer or rejection date for those loss mitigation options. For applications initiated in 2012 or 2015, the data include information from the time the borrower started the application through June 2017. As a result, the data include information about offers, rejections, or appeals even if those events occur in a later year than the year of the application start date.

8.3.1 Evaluation timelines

The Rule imposed a 30-day deadline for evaluating loss mitigation applications, although many servicers the Bureau interviewed reported completing evaluations within 30 days before the Rule took effect. To test whether this provision affected the speed of evaluations, the Bureau analyzed the number of days between the date an application was complete and the earliest date the borrower received an offer or denial for a loss mitigation option. Figure 1, below, shows the share of borrowers who had received a loss mitigation decision by a given number of days after their application was complete. For applications initiated in 2012, 64 percent of borrowers received a decision within 30 days compared to 88 percent in 2015. This is consistent with what the Bureau learned in servicer interviews; servicers were evaluating most, but not all, applications within 30 days pre-Rule. Figure 1 shows that the share of applications with decisions within the first week changed very little pre- and post-Rule. If anything, a larger percentage of applications had a decision within the first week in 2015 compared to 2012.

FIGURE 1: NUMBER OF DAYS FROM COMPLETE APPLICATION TO DECISION DATE (SERVICING OPERATIONS DATA)

8.3.2 Simultaneous review for all available options

As discussed above, the Bureau expected the requirement to review borrowers for all available loss mitigation options would mean that borrowers had better information about the options they qualify for, were less likely to need to submit multiple applications, and were more likely to receive fair evaluations for all options available to them. The Servicing Operations Data do not have any information regarding whether borrowers seeking loss mitigation are better informed about their options. The data show only what options borrowers were offered and cases where borrowers are offered more than one option are rare in the data. Thus, it cannot be determined from the data whether or not the provision led to consideration of borrowers for more options and perhaps consideration for options that they would not have applied for previously.286

Some servicers reported that an unintended consequence of the requirement to evaluate borrowers for all available options simultaneously is that it imposes unnecessary burden on servicers and borrowers when the borrower is interested in one particular option. In particular, servicers reported that borrowers seeking non-retention options (e.g., a short sale) have additional burdens of time, trouble, and aggravation due to the need to provide documentation that is not relevant to the short-sale application but is relevant for an evaluation for other options, such as a loan modification. One way to test this is to measure the number of days from when a borrower first reached out to their servicer about loss mitigation to the date he or she was offered a short sale. Figure 2 plots this below based on the Servicing Operations Data, using the earlier of the date the borrower first reached out for help orally or the date the borrower first submitted written documents to the servicer in connection with an application as the date of first outreach.287 For borrowers who initiated a loss mitigation application in 2012 and were ultimately offered a short sale, the median time from first outreach to short-sale offer was 63 days whereas the median was 115 days for borrowers who initiated an application in 2015. There was also an increase in the median number of days from date of first outreach to trial modification offer, not shown in the figure, but the change is not as significant as it is for shortsale offers (63 days compared to 78 days). The increase in time from first outreach to short-sale offer is consistent with additional time needed for borrowers seeking short sales to complete an application, but could also have other explanations. In particular, the time required before a short sale is offered may include time needed to market the property and negotiate a purchase contract with the buyer. Therefore, the increase in time from borrower's first outreach about loss mitigation to the contract approval date may reflect changes in the housing market or investor policies between 2012 and 2015, in addition to the Rule taking effect. For example, in a market where house prices are rising, such as in 2015, investors may be more likely to delay approval of a short-sale contract because they might expect to get a better offer by waiting.

Other factors may have affected short-sale timelines. The Rule's requirement to evaluate borrowers for all options available to them may have made it more likely that borrowers who were denied for modifications in 2015 were offered a short sale compared to those denied for a modification in 2012. This could mean that a larger share of the borrowers offered short sales in 2015 were originally seeking a home retention option compared to borrowers offered short sales in 2012. Thus, if the loss mitigation process is longer for borrowers who initially seek a modification, but are denied and offered a short sale than for those who initially seek a short sale and are approved, the Bureau would expect to see an increase in short-sale timelines from 2012 to 2015.

FIGURE 2: NUMBER OF DAYS BETWEEN BORROWER'S FIRST OUTREACH ABOUT LOSS MITIGATION AND SHORT-SALE OFFER (SERVICING OPERATIONS DATA)

Table 1, below, uses GSE Data from Fannie Mae to conduct a related analysis. The table shows the median number of months from last paid installment date to short-sale date by year for loans that were at least 60 days delinquent at the time of short sale. The Fannie Mae Data show that the median number of months increased from 10 to 12 months between 2011 and 2013 to 15 to 18 months between 2014 and 2016. This indicates that borrowers who ultimately sell their home through a short sale are taking longer to do so in the years since the Rule was effective. The short-sale date in the Fannie Mae Data represents the date the sale occurred, which is after the contract approval date, so the concept that investors may be more likely to delay approval of a short sale in a market with rising home prices is relevant for this analysis as well. These findings could be consistent with the hypothesis that the requirement to complete a loss mitigation application with enough information to evaluate the borrower for all options is causing a delay in short-sale timelines, but could also be explained by changing market conditions and change in the set of borrowers offered short sales.

TABLE 1: MEDIAN NUMBER OF MONTHS FROM LAST PAID INSTALLMENT TO SHORT-SALE DATE (FANNIE MAE DATA)

Year

Number of months

2000

6

2001

7

2002

8

2003

8

2004

9

2005

9

2006

10

2007

9

2008

9

2009

9

2010

11

2011

10

2012

10

2013

12

2014

15

2015

18

2016

17

8.3.3 Appeals

The Bureau analyzed how frequently borrowers appealed decisions on complete loss mitigation applications. The appeals analysis is limited to five servicers that were able to provide data on appeals in both the pre-Rule and post-Rule periods.

Figure 3, below, shows all borrower appeals and successful appeals as a share of complete loss mitigation applications in 2012 and 2015. The figure includes borrowers who appealed their denial for all loss mitigation options and those who appealed after the servicer approved them for a trial or permanent loan modification offer. In cases where a borrower appealed after receiving a loan modification offer, the Servicing Operations Data do not indicate whether the borrower is appealing the denial of another modification option (in which case the Rule's appeal provisions apply) or is appealing the terms of the offered modification (in which case the Rule's appeal provisions do not apply). The Bureau included borrowers who appealed after receiving a loan modification approval in the analysis below because it is possible these borrowers were denied for another loan modification and their appeal was an appeal of that denial.

Overall, only a small fraction of borrowers who complete applications appeal the servicer's decision. Borrowers included in the Servicing Operations Data utilized appeal rights more frequently in 2015 compared to 2012, with 4 percent of complete applications involving an appeal in 2015 compared to 1 percent in 2012. Although a larger share of borrowers appealed the decision on their complete application in 2015, the share of complete applications with a successful appeal was similar in 2012 and 2015 (0.2 versus 0.1 percent), implying that a much smaller share of appeals was successful in 2015. In interviews, some servicers stated that, while they had a process for appeals in place prior to the Rule, borrowers were not necessarily aware of their appeal rights prior to the Rule. These servicers said borrowers are more aware of their appeal rights now, in part because the Rule requires that the written loss mitigation determination letter include language about the appeal process.

FIGURE 3: SHARE OF COMPLETE APPLICATION WITH BORROWER APPEAL AND SUCCESSFUL APPEAL (SERVICING OPERATIONS DATA)

The Bureau compared appeal rates for borrowers who were denied for all offers and those who were approved for a trial or permanent modification. Figure 4, below, shows that a larger share of borrowers who received no offers appealed the decision in 2012 and 2015 compared to those who received a modification offer. The appeal rate was almost 4 times larger for both groups in 2015 compared to 2012, with the appeal rate for those who received a modification offer increasing from 0.5 to 1.9 percent and the rate for those denied for all options increasing from 1.2 to 6.1 percent. Successful appeals as a share of complete applications for both groups was close to constant from 2012 to 2015.

FIGURE 4: APPEAL RATES BY WHETHER OR NOT BORROWER RECEIVED A LOAN MODIFICATION OFFER (SERVICING OPERATIONS DATA)

9. Foreclosure restrictions

This chapter considers the Rule's foreclosure restriction provisions, which place limits on servicers' ability to initiate and complete the foreclosure process, particularly while a servicer is evaluating a complete loss mitigation application. The main findings include the following:

  • Borrowers who were between 90 and 120 days delinquent were much less likely to have foreclosure initiated in 2015 compared to 2012, before the Rule was effective. This decrease in foreclosures initiated was not offset by an increase in foreclosures initiated within one year of the first 90-day delinquency, even after controlling for other factors. These facts suggest that the Rule's general prohibition on initiating foreclosure within the first 120 days of delinquency prevented rather than delayed foreclosures. Housing counselors surveyed generally said that the Rule's foreclosure restrictions were the most important of the Rule's requirements in helping their clients.
  • Servicers interviewed generally said they had to make significant changes to their foreclosure processes to ensure compliance with the Rule's foreclosure restrictions and that these restrictions were among the more costly provisions of the Rule to implement. The Bureau did not determine the specific cost to servicers of complying with the Rule's foreclosure restriction provisions. Consistent with the Bureau's finding above, the servicers acknowledged that borrowers generally had benefited from these foreclosure restrictions. However, the servicers also suggested that these restrictions may have disadvantaged some borrowers by causing them not to engage in the loss mitigation process until after they had fallen further behind on their mortgage payments.
  • Data indicate that a larger share of borrowers who completed loss mitigation applications in 2015 were able to avoid foreclosure, compared to similar borrowers in 2012, before the Rule was effective. There also was a shift from pre-Rule to post-Rule toward loans being delinquent for longer before servicers initiated foreclosure. The data also suggest that the foreclosure restriction provisions have not increased the time it takes for servicers to go from initiating foreclosure to a sale.

9.1 Background

Prior to the effective date of the 2013 RESPA Servicing Rule, no federal statutory or regulatory provisions prescribed general rules for servicers with respect to starting or completing foreclosures. However, there were and continue to be state law requirements that govern foreclosure timelines.

In the wake of the housing crisis, state and federal regulators imposed requirements on servicers with respect to their foreclosure practices. The FHFA's Servicing Alignment Initiative required covered servicers that received an application for a loan modification option from a borrower within 120 days of delinquency to evaluate the borrower for a loan modification option before referring the borrower's mortgage to foreclosure. Prudential regulators, including the Federal Reserve Board and the OCC, undertook enforcement actions against major servicers, resulting in consent orders that imposed requirements on servicing practices, including those related to foreclosure. The National Mortgage Settlement (NMS) prohibited servicers from referring a mortgage to foreclosure if the servicer had received a substantially complete loan modification application before day 120 of delinquency. Some states adopted or proposed regulations relating to foreclosure processing, which included prescribing timelines for initiating foreclosure.

The 2013 RESPA Servicing Rule includes three primary foreclosure restrictions. First, the Rule generally prohibits a servicer from making the first notice or filing required for a foreclosure process until a mortgage loan account is more than 120 days delinquent if the mortgage is secured by the borrower's principal residence. Second, even if a borrower is more than 120 days delinquent, if a borrower submits a complete loss mitigation application, a servicer may not make the first notice or filing required for a foreclosure process unless: 1) the servicer informs the borrower that the borrower is not eligible for any loss mitigation option and any appeal has been exhausted; 2) a borrower rejects all loss mitigation offers; or 3) a borrower fails to comply with the terms of a loss mitigation option. Third, if a borrower submits a complete application for a loss mitigation option after the foreclosure process has commenced but more than 37 days before a foreclosure sale, a servicer may not move for a foreclosure judgment or order of sale, or conduct a foreclosure sale, until one of the same three conditions above has been satisfied. For the purposes of this chapter, the assessment refers to these three foreclosure restrictions, respectively, as: the "minimum delinquency requirement;" the "prefiling dual tracking restriction;" and the "post-filing dual tracking restriction."

As with other provisions of the Rule related to loss mitigation, the Bureau established the Rule's restrictions on foreclosure as part of its efforts to establish a fair process for review of loss mitigation applications to help borrowers avoid foreclosure when possible. The Bureau believed that these restrictions were necessary and appropriate to achieve the consumer protection purposes of RESPA, including facilitating the review of borrowers for foreclosure avoidance options. In addition, the Bureau adopted these provisions to implement a servicer's obligation to take timely action to correct errors related to avoiding foreclosure by establishing servicer duties and procedures that must be followed to avoid errors with respect to foreclosure. The Bureau also intended these provisions to further the goals of the DoddFrank Act to ensure a fair, transparent, and competitive market for mortgage servicing. Particular goals of the foreclosure restriction provisions include:

Uniform loss mitigation standards. The foreclosure restriction provisions intended to help delinquent borrowers avoid unnecessary costs and fees associated with foreclosure. While the Bureau had no specific target with respect to the number or speed of foreclosures, the Bureau intended to ensure that borrowers were protected from unnecessary harm in connection with foreclosure proceedings. The Bureau believed that establishing uniform mortgage servicing standards was necessary and appropriate to protect borrowers in connection with foreclosure proceedings and achieve the consumer protection purposes of RESPA. The Bureau believed that such standards would establish appropriate expectations for loss mitigation processes and ensure that borrowers have a full and fair opportunity to receive an evaluation for loss mitigation before suffering the unnecessary harms associated with foreclosure.

Reasonable opportunity for borrowers to submit loss mitigation applications. The minimum delinquency requirement was intended to provide borrowers with a reasonable opportunity to submit loss mitigation applications prior to servicers commencing foreclosure proceedings. The Bureau believed that establishing a loss mitigation plan would benefit borrowers by reducing borrower costs and limiting the possible negative effects of foreclosures on borrowers' credit reports. The Bureau believed that it was necessary and appropriate to implement the consumer protection purposes of RESPA by ensuring that borrowers, servicers, and courts have a specified early period during which a servicer could not begin the foreclosure process.

Restrict dual tracking. The Bureau adopted the foreclosure restriction provisions to address the borrower harm that it believed may result when a servicer pursues a foreclosure during the loss mitigation process. By prohibiting a servicer from pursuing a foreclosure while a complete loss mitigation application is pending, the 2013 RESPA Servicing Rule was designed to eliminate the clearest harms to borrowers resulting from servicers pursuing loss mitigation and foreclosure proceedings concurrently.

Prevent strategic delays of foreclosure. The 2013 RESPA Servicing Rule's loss mitigation provisions were designed to set timelines for loss mitigation evaluation that could be completed without requiring a suspension of the foreclosure sale date to avoid strategic use of these procedures to extend foreclosure. The Bureau did not intend that borrowers should use the foreclosure restriction provisions to strategically delay or derail the foreclosure process.

The Bureau's assessment of the foreclosure restriction provisions sought to understand how these provisions have affected foreclosure rates both for delinquent borrowers who have applied for loss mitigation and for borrowers who have not applied. The next section of this chapter describes what the Bureau learned from servicer interviews and the Counselor Survey about the foreclosure restriction provisions' effects on borrowers and servicers. The following section uses the Servicing Operations Data to analyze (a) general foreclosure rates and timing, (b) how many borrowers with complete loss mitigation applications also experience the initiation or completion of foreclosures, (c) the timing of loss mitigation applications relative to foreclosure, and (d) the direct effects of each of the three main provisions on foreclosure.

9.2 Qualitative information: Servicer interviews and Counselor Survey

9.2.1 Effects on servicer practices

Most servicers that the Bureau interviewed said the Rule's foreclosure restrictions had affected how they manage the foreclosure process. Some small- and mid-size servicers, however, said that, because they initiate so few foreclosures each year, the Rule's foreclosure provisions had little practical effect.

Most servicers said that the minimum delinquency requirement had led them to delay initiation of foreclosure. Servicers generally said that before the Rule they would refer some loans for foreclosure when they were 90 days past due and that the Rule had caused them to not refer loans to foreclosure counsel until after the 120th day. In principle, the Rule permits referring a loan to foreclosure counsel earlier, so long as the servicer does not make the first notice or filing316 in a foreclosure action within the first 120 days of delinquency. However, servicers generally said that in practice they apply the restriction to the date they refer cases to an attorney.

Most servicers also said that the restrictions on initiating foreclosure while evaluating a complete loss mitigation application had led to at least some changes in their practices. Some servicers said that, prior to the Rule, foreclosure and loss mitigation processes proceeded on completely separate tracks and that the foreclosure process would not be affected by a loss mitigation application unless and until a trial modification was in place. Others said that even prior to the Rule they would take no further steps in the foreclosure process once a complete application was received, but that the Rule had led them to introduce additional controls to ensure this policy is carried out consistently.

Most servicers said that their practice is to halt all work on an ongoing foreclosure case once a complete application is received, even though the Rule generally permits servicers to continue steps toward a foreclosure sale that fall short of scheduling or conducting a sale. Some servicers said they did so because they thought that made the most sense for borrowers or was what borrowers expect, while others emphasized that they thought there was too much compliance risk associated with relying on foreclosure counsel not to take prohibited steps while a complete loss mitigation application was in process.

Many servicers said they would halt a foreclosure sale even if a loss mitigation application was received less than 37 days before a scheduled foreclosure sale.

9.2.2 Effects on borrowers

Counselor Survey. About two-thirds of housing counselors surveyed said that the foreclosure restriction provisions were effective in helping their clients, with another 30 percent saying that the provisions were "somewhat" effective. About two-thirds of respondents said that when the provisions were effective, it was often or always because the provisions helped clients initiate a loss mitigation application. Several respondents stated that the provisions are especially helpful for borrowers who do not contact a counselor or understand their loss mitigation options until late in foreclosure. In instances in which the provisions were reportedly not effective, respondents most commonly said the servicer made it difficult for the borrower to complete a loss mitigation application. Several respondents stated that the leeway servicers have in defining when an application is complete makes it challenging for borrowers to determine whether a servicer violated the foreclosure restriction provisions.

Servicer Interviews. Servicers generally said borrowers had benefited from the Rule's foreclosure restrictions. Some said that the minimum delinquency requirement gave some borrowers additional opportunity to apply for loss mitigation and to avoid fees related to foreclosure. Some servicers also said that the provisions have reduced borrower confusion because loss mitigation and foreclosure activities are no longer occurring simultaneously. Some servicers also said there was a psychological benefit to borrowers to have foreclosure activity stop once the borrowers started the loss mitigation process.

On the other hand, some servicers said that they believe the minimum delinquency requirement can do a disservice to certain borrowers by delaying the first steps of the foreclosure process. These servicers said that for some borrowers the start of the foreclosure process provides the motivation to start engaging with the loss mitigation process or to otherwise resolve their situation. In such cases, these servicers said, the delay in a foreclosure caused by the minimum delinquency requirement can mean that the borrower may miss an additional payment by the time they first engage with the servicer about their delinquency, and the additional amount owed can make it more difficult for the borrower to find a workable solution.

9.2.3 Costs of the provision and other effects on servicers

Servicers said that the foreclosure restrictions were among the more costly provisions of the Rule to implement. Most servicers said they incurred both one-time and ongoing costs to ensure that the Rule's foreclosure restrictions were not violated. Many servicers described steps they took to better integrate loss mitigation and foreclosure functions to ensure that the foreclosure process stops when a loss mitigation application is completed. Servicers said they needed to develop and implement a system for communicating the status of a loss mitigation application to the foreclosure department and foreclosure counsel. Servicers generally implemented this in the first instance through an automated process in which the loss mitigation department places a "hold" on an account that is automatically communicated via the servicing platform to the foreclosure department and foreclosure counsel. However, servicers said that this process also requires some manual checks to prevent violations of the dual tracking rules.

Some servicers emphasized that much of the cost associated with the dual tracking restrictions arises from the need to avoid taking certain actions to advance a foreclosure even when they receive a complete application just before the action was scheduled to take place. Servicers noted that it can take time to determine whether an application is complete, both when an application is received for the first time and in instances when the borrower provides a new document that could complete an existing incomplete application. Some servicers said that significant resources are spent shortly before making the first notice or filing in a foreclosure action and before scheduling a foreclosure sale to ensure that an application has not been completed at the last minute.

Specific to the minimum delinquency requirement, some servicers said that waiting to refer an account for foreclosure until the account is more than 120 days delinquent sometimes makes it more difficult to comply with investor requirements to commence foreclosure within certain timelines. Also, many servicers said that the minimum delinquency requirement is a problem if a property has been abandoned, because the Rule makes it harder for them to take earlier steps to preserve the property's condition. While the foreclosure restrictions in the Rule do not apply where the property is not the borrower's "principal residence," and therefore would not apply if the property is truly abandoned, servicers said that determining whether a property is in fact abandoned is complex and expensive, and that they therefore observe the 120-day restriction in all or almost all cases even if they believe the property is abandoned.

Servicers generally said that the foreclosure restrictions had lengthened the foreclosure process on average. Some noted that this meant additional costs to conduct property inspections and in some cases to maintain vacant properties over a longer period than pre-Rule.

9.3 Quantitative analysis of the foreclosure restriction provisions

The analyses in this chapter draw information from several components of the Servicing Operations Data. The primary data are monthly loan performance records, which show current delinquency and payment status for each month from January 2012 to June 2017, for loans serviced in 2012, and from January 2015 to June 2017, for loans serviced in 2015. These data are limited to loans that were more than 30 days delinquent or completed a loss mitigation application as of the end of a month during 2012 or 2015. These data are further linked to loanlevel information on, among other things, loss mitigation applications and the dates, if any, that the servicer made the first notice or filing of foreclosure (initiated foreclosure, henceforth), and of the foreclosure sale. Note that there is a difference in timing between the monthly performance data and the data on loss mitigation applications. The Bureau's data requests specified that information on loss mitigation applications only be provided for applications initiated during 2012 or 2015; as noted above, the monthly performance data covers a broader period.

9.3.1 Foreclosure rates and timelines before and after the Rule

This section describes the frequency of foreclosure in the Servicing Operations Data. Figure 1 shows the cumulative proportion of loans for which the servicer initiated foreclosure and the proportion of loans for which a foreclosure sale was completed. The top panel shows the rates for loans that became more than 30 days delinquent or began a loss mitigation application in 2012, while the bottom panel shows the same statistics for loans that became more than 30 days delinquent or initiated a loss mitigation application in 2015. Consistent with the results in Chapter 3, foreclosure rates for loans that were delinquent or had applied for loss mitigation declined substantially between 2012 and 2015. By January 2013, servicers had initiated foreclosures for almost 20 percent of loans that were more than 30 days delinquent or had applied for loss mitigation in 2012, and servicers completed a foreclosure sale for more than 10 percent of these loans. This is roughly double the rate in January 2016 for loans that became delinquent or applied for loss mitigation in 2015.

FIGURE 1: RATE OF INITIATING FORECLOSURE AND FORECLOSURE SALES BY YEAR FOR LOANS (SERVICING OPERATIONS DATA)

Figure 2 shows histograms of the number of days borrowers were delinquent at the date when the servicers initiated foreclosure. The top panel shows the distribution of days delinquent when foreclosure was initiated in 2012, and the bottom panel shows the same distribution for loans where foreclosure was initiated in 2015. The number of days shown is truncated at two years (730 days) delinquent to highlight the distribution for most of the data, but a substantial percentage of foreclosures were on loans delinquent by more than two years. A vertical line in each plot denotes 120 days, the minimum delinquency requirement in the Rule.

As Figure 2 shows, there was a shift from 2012 to 2015 toward loans being delinquent for longer before servicers initiated foreclosure. This goes somewhat beyond shifting foreclosures initiated at less than 120 days delinquency to being initiated at more than 120 days delinquency. The share of foreclosures servicers initiated at between 120 and 135 days delinquent increased substantially between 2012 and 2015, by about 15 percentage points. By comparison, the share of foreclosures initiated before 120 days delinquent went from about 28 percent to zero. The share of foreclosures servicers initiated between five to eight months of delinquency declined as well. At the same time, the share of foreclosures servicers initiated between eight months and two years delinquency increased from about 20 percent in 2012 to about 30 percent in 2015.

FIGURE 2: DISTRIBUTION OF THE NUMBER OF DAYS BORROWERS WERE DELINQUENT AT THE TIME WHEN SERVICER INITIATED FORECLOSURE (SERVICING OPERATIONS DATA)

9.3.2 Foreclosure activity and loss mitigation applications

The Bureau intended the dual tracking restrictions of the Rule to prevent harm to borrowers from the foreclosure process while a loss mitigation application is pending. Before assessing the effect of the Rule on this goal, a threshold question is how often servicers took formal steps toward foreclosure against borrowers who also applied for loss mitigation at some point. The mere fact that the borrower both completed a loss mitigation application and his or her servicer ultimately initiated foreclosure or a foreclosure sale does not necessarily mean that the servicer was simultaneously processing the application and the foreclosure. However, there cannot have been this kind of "dual tracking" unless some of the borrowers who experienced foreclosure actions also applied for loss mitigation.

One way to measure this overlap is to focus on borrowers who completed loss mitigation applications initiated in 2012 or 2015. Figure 3 shows the proportion of these borrowers who ever had a foreclosure initiated, and the proportion who ever have a foreclosure sale completed, by the year of the loss mitigation application. Note that this is not a perfect measure, as loans that have a loss mitigation application late in the relevant year are observed for a shorter period of time. The data show that a large proportion of borrowers with completed loss mitigation applications in 2012 also had either an initial foreclosure filing or foreclosure sale at some point. About 40 percent of borrowers with a completed application had foreclosure initiated within the time frame of the Servicing Operations Data, and about 7 percent had a complete foreclosure sale at some point. The rate of servicers initiating foreclosure among borrowers who completed loss mitigation applications declined in 2015, although the decrease is small relative to the overall decrease in foreclosure activity reflected in Figure 1 above. The proportion of borrowers who completed loss mitigation applications and also experienced a foreclosure sale did not change between 2012 and 2015.

FIGURE 3: PROPORTION OF BORROWERS WITH LOSS MITIGATION APPLICATIONS WHO ALSO HAD FORECLOSURE INITIATED OR COMPLETED (SERVICING OPERATIONS DATA)

Another way to measure the extent of overlap between borrowers with foreclosure activity and those with completed loss mitigation applications is by focusing on whether an initiated foreclosure resulted in a foreclosure sale. That is, limit the sample to borrowers who had foreclosure initiated, and measure the proportion for whom a foreclosure sale eventually occurred, and how this differs if a loss mitigation application was completed after the servicer initiated foreclosure. Note that a servicer need not have been engaging in dual tracking for this to occur. For example, if the loss mitigation application is denied it is possible to observe a completed application during the foreclosure process even if the servicer completely halted the foreclosure process while considering the application.

Figure 4 shows the share of loans for which foreclosure was initiated that eventually saw a foreclosure sale, broken out by year and whether or not a loss mitigation application was completed after foreclosure was initiated. For this figure, the sample excludes borrowers who had foreclosure initiated in 2012 and were still in the foreclosure process in July 2014—loans that had foreclosure initiated in 2015 and have a sale after July 2017 are excluded because the data end in June 2017.

Both before and after the effective date of the Rule, a foreclosure sale occurred much less often for loans with a loss mitigation application that was completed after the servicer initiated foreclosure. Still, foreclosure sales were completed for a substantial fraction of loans for which a loss mitigation application was completed after foreclosure was initiated, about one-sixth of such loans in both 2012 and 2015. The share of loans for which a foreclosure had been initiated that ended in a sale during the period of analysis remained roughly the same in the pre- and post-Rule periods. The data indicate that the share of loans with a foreclosure sale increased slightly for loans where a loss mitigation application was completed after foreclosure was initiated. This may indicate that the post-filing dual tracking restriction had little effect on borrowers, or it may be that the pool of borrowers applying for loss mitigation after foreclosure was initiated has changed. For example, it may be that more meritorious loss mitigation applications tend to be completed earlier in delinquency, both before and after the Rule, but after the Rule foreclosure tended to be initiated later in delinquency, as shown in Figure 2.

FIGURE 4: SHARE OF LOANS IN FORECLOSURE THAT ULTIMATELY RESULTED IN A FORECLOSURE SALE, BY YEAR AND WHETHER A LOSS MITIGATION APPLICATION WAS COMPLETED BEFORE FORECLOSURE WAS INITIATED (SERVICING OPERATIONS DATA)

Together, Figure 3 and Figure 4 show that many borrowers who apply for loss mitigation also have foreclosure initiated and foreclosure sales completed at some point. Again, this does not mean that the servicers represented in the Servicing Operations Data were engaging in dual tracking, but it does indicate at the very least that there could have been such behavior for the Rule to potentially affect.

9.3.3 Timing of loss mitigation applications relative to foreclosure

One preliminary way to test whether the pre-filing dual tracking restriction—the restriction on initiating a foreclosure proceeding once a completed application was filed and pending review— was binding for practical purposes on servicers is to focus on loans where a loss mitigation application was completed and the servicer ultimately initiated foreclosure. When dual tracking occurs, one would expect to see servicers sometimes initiating foreclosures shortly after borrowers complete loss mitigation applications, since completing an application does not necessarily affect the servicer's foreclosure timeline. Conversely, if servicers are not dual tracking, it should be rare to see foreclosure initiated shortly after an application is completed, because one would expect some gap in time while the servicer evaluates the application. As a result, under the pre-filing dual tracking restriction in the Rule, foreclosures initiated shortly after a loss mitigation application is completed should be less common than prior to the Rule, because a servicer cannot initiate foreclosure until it has processed a complete loss mitigation application. A complete loss mitigation application would only appear in the data shortly before the servicer initiates foreclosure if the servicer quickly denies the application, and even then the 14-day waiting period for appeals would apply. This analysis also tests whether the minimum delinquency requirement allowed more borrowers to complete loss mitigation applications before servicers initiated foreclosures. If the minimum delinquency provision was helpful to borrowers in this regard, post-Rule there should be relatively more loss mitigation applications completed before the servicer initiates foreclosure than there were pre-Rule.

Figure 5 plots the smoothed distribution of when loss mitigation applications were completed relative to the servicer initiating foreclosure, with separate distributions for loans where the loss mitigation application was started in 2012 or 2015. For loans with both a loss mitigation application and foreclosure initiated, the concentration is greatest just to the right of zero for both years; in other words, the loss mitigation application is most often completed shortly after the initiation of foreclosure. There are notable shifts in the distribution from 2012 to 2015. The share of applications completed immediately after foreclosure was initiated and the share of applications completed more than 90 days before both increased. There was also a sharp decline from 2012 to 2015 in the proportion of loss mitigation applications that borrowers completed right before the servicer initiated foreclosure, which likely reflects the pre-filing dual tracking restriction. Together, these shifts suggest that after the Rule borrowers were more able to complete loss mitigation applications prior to foreclosure than they were before the Rule.

FIGURE 5: DISTRIBUTION OF THE TIMING OF COMPLETED LOSS MITIGATION APPLICATIONS RELATIVE TO INITIATING FORECLOSURE (SERVICING OPERATIONS DATA)

One potential problem with the analysis in Figure 5, however, is that the distribution of loss mitigation application timing may differ across different types of loans, and the changes between 2012 and 2015 may reflect unrelated changes in the frequency of particular types in the sample. To explore this, Figure 6, below, plots the smoothed distribution of loss mitigation application timing relative to initiating foreclosure, for different subsets of loans. Panel (a) limits the sample to loans that were less than one-year delinquent at the time of initiating foreclosure, while panel (b) limits the sample to loans that were more than one-year delinquent at that time. For loans that were less delinquent at the time of foreclosure, there was less of an increase between 2012 and 2015 in the proportion of loans that had a loss mitigation application completed more than 90 days before, compared to the sample of all loans. Loans that were more severely delinquent at the time of initiating foreclosure were more likely to have completed their loss mitigation applications more than 90 days prior. Regardless of the delinquency length, there is still a sharp decline in the proportion of loans that have a loss mitigation application immediately prior to initiating foreclosure. Panel (c) limits the sample to borrowers in states without judicial foreclosure proceedings, and panel (d) is limited to borrowers in states with judicial foreclosure. Whether a state has judicial foreclosure proceedings make relatively little difference for the distribution of loss mitigation application timing relative to foreclosure.

FIGURE 6: DISTRIBUTION OF THE TIMING OF COMPLETED LOSS MITIGATION APPLICATIONS, RELATIVE TO INITIATING FORECLOSURE, BY DELINQUENCY AND JUDICIAL FORECLOSURE STATUS (SERVICING OPERATIONS DATA)

<365 Days Delinquent at Filing (c) Non-Judicial States

>365 Days Delinquent at Filing (d) Judicial States

9.3.4 Effects of the minimum delinquency requirement

As discussed above, the purpose of the minimum delinquency requirement was to provide borrowers more time to avoid going into foreclosure, whether through loss mitigation or other means. For instance, requiring that borrowers generally be at least 120 days past due before the servicer initiates foreclosure was intended to allow more time for borrowers to apply for loss mitigation and go through the application process. The additional time may also enable borrowers to refinance or to gather enough money to cure the delinquency.

The minimum delinquency requirement may also provide borrowers more time to cure their delinquency or otherwise avoid foreclosure without loss mitigation. This can be tested by examining foreclosure outcomes for borrowers who are observed becoming 90 to 120 days delinquent. Prior to the Rule, servicers generally did not initiate foreclosure before a borrower became 90 days delinquent, and the minimum delinquency requirement has no direct effect on borrowers beyond 120 days delinquent. Thus, if the minimum delinquency requirement enabled borrowers to avoid foreclosure, there should be a reduction in the probability of foreclosure while 90 to 120 days delinquent after the Rule, without a corresponding increase a few months later. While other provisions of the Rule may also have reduced the probability of initiating foreclosure, including the pre-filing dual tracking restrictions and the provisions around early intervention and completing applications, there is no reason that these provisions would specifically affect the probability of initiating foreclosure between 90 and 120 days.

To conduct this test, the Bureau estimated a regression model using the first observation for each loan (if any) for which the borrower was between 90 and 120 days delinquent. The outcome of interest in the model is a categorical variable denoting how many months later foreclosure occurred, with categories for 0 to 2 months, 3 to 4 months, 5 to 6 months, 7 to 12 months, and 13 months or more, where the last category includes loans that never have foreclosure initiated within the range of the Servicing Data. The key independent variable is an indicator for whether the loan became 90 days delinquent before or after the Rule. If the minimum delinquency requirement prevented foreclosures, the expected result is a negative effect for the bottommost category, 0 to 2 months, and a matching increase in the 13 months or never category. Conversely, if the minimum delinquency requirement merely delayed foreclosures, one would expect to see a negative effect in the bottommost category and positive effects for the intermediate categories of 3 to 4 months, 5 to 6 months or 7 to 12 months. Note that if the minimum delinquency requirement merely shifted foreclosures a month later in delinquency, this would still be captured in the bottommost category, producing an effect of zero. The regression controls for whether the borrower applied for loss mitigation, as well as state-level average house prices in the current month, servicer, origination year, state, loan balance, monthly payment, interest rate, and product type (i.e., GSE, government, or privatelyheld loan).

Figure 7 presents the results of this test. The figure plots the average marginal effect of becoming 90 to 120 days delinquent after the Rule instead of before the Rule on the probability of initiating foreclosure in each of the five categories discussed above. The figure shows that the probability of initiating foreclosure within 0 to 2 months of being observed at 90 to 120 days delinquent fell by about 6 percentage points. For comparison, about a quarter of loans that became 90 days delinquent in 2012 or 2013 had foreclosure initiated in 0 to 2 months. At the same time, the estimates show no increase in the probability of initiating foreclosure between 3 and 12 months later. Instead, the increase comes entirely in foreclosures happening more than one year later or never. It is likely that these effects were driven by the minimum delinquency requirement, rather than other provisions of the Rule or general economic trends. While other factors may well have led to a decline in the probability of foreclosures, it is unlikely that any other factor would have a disproportionate effect on loans in the two months immediately after becoming 90 to 120 days delinquent, particularly given that the regression holds constant whether or not the borrower had applied for loss mitigation.

FIGURE 7: AVERAGE MARGINAL EFFECT OF THE RULE ON THE PROBABILITY OF FORECLOSURE WITHIN A PERIOD OF MONTHS AFTER BEING OBSERVED BECOMING 90 DAYS DELINQUENT FOR THE FIRST TIME (SERVICING OPERATIONS DATA)

To further explore this issue, Figure 8 breaks out loans that first became 90 days delinquent before and after the Rule's effective date based on the loan's status six months after being observed becoming 90 days delinquent. Consistent with Figure 7, after the Rule, loans were less likely to be in foreclosure or have a foreclosure sale six months after becoming 90 days delinquent. Although the proportion of loans that were more than 120 days delinquent but not in foreclosure increased following the effective date of the Rule, there was also an increase in the proportion of loans that had cured and become current, including loans that were modified and current on the modified terms. There was also a substantial increase in loans that were still delinquent, but by less than 120 days. This consists of both borrowers who cured but fell behind again and borrowers who made payments but did not fully catch up on their loans and become current.

FIGURE 8: LOAN STATUS SIX MONTHS AFTER FIRST BECOMING 90 DAYS DELINQUENT, BY YEAR OF INITIAL DELINQUENCY (SERVICING OPERATIONS DATA)

9.3.5 Effects of the pre-filing dual tracking restriction

The dual tracking restrictions were intended to help borrowers who were applying for loss mitigation to avoid foreclosure when possible. This section focuses specifically on the pre-filing dual tracking restriction. Prior to the Rule, it was possible for a borrower to apply for loss mitigation and have the servicer still initiate foreclosure, even if the loss mitigation application was ultimately successful. As discussed above, the foreclosure process is costly for both borrowers and investors, and avoiding initiating foreclosure in cases that can be resolved with loss mitigation has benefits to all parties.

The challenge in assessing the effectiveness of this provision of the Rule is that not all loss mitigation applications are successful. If the application is denied, the borrower refuses all options offered, or the borrower fails to perform on a modification, initiating foreclosure will simply be delayed, rather than prevented. Such delays are costly to servicers and investors and can also impose costs on borrowers, with limited, if any, benefit.

If the pre-filing dual tracking restriction helped borrowers avoid initial foreclosure filings, there should be an increase in accepted loss mitigation offers occurring before the servicer initiates foreclosure or in instances when the servicer never initiates foreclosure. To test this, the analysis focuses again on a subsample of the data consisting of loans that have become at least 90 days delinquent for the first time. This analysis further restricts attention to borrowers who completed a loss mitigation application on or after first becoming 90 days delinquent, but prior to the servicer initiating foreclosure (loans that never have foreclosure initiated are included).

Figure 9, below, shows the proportions of these borrowers that ultimately accepted a loss mitigation offer prior to the servicer initiating foreclosure and the proportion that had foreclosure initiated within one year of first becoming 90 or more days delinquent. From 2012 to 2015 there was an increase in the proportion of borrowers with complete loss mitigation applications who accept an offer after becoming 90 days delinquent. There was also a decline in the proportion of such borrowers who had foreclosure initiated within one year of becoming 90 days delinquent.

FIGURE 9: PROPORTIONS OF BORROWERS WHO HAVE BECOME 90 OR MORE DAYS DELINQUENT FOR THE FIRST TIME AND COMPLETED A LOSS MITIGATION APPLICATION WHO ACCEPT A LOSS MITIGATION BEFORE AN INITIAL FORECLOSURE FILING, AND THE PROPORTION WHO HAVE FORECLOSURE INITIATED WITHIN 1 YEAR (SERVICING OPERATIONS DATA)

Quantifying how the pre-filing dual tracking restriction affected the rate of foreclosures presents some analytical challenges. First, a number of other provisions of the Rule affect the probability of foreclosure. Second, there is not a good control group that captures how a completed loss mitigation application would affect the probability of foreclosure absent the Rule. Finally, there may be other factors unrelated to the Rule that have changed for the loans in the Servicing Operations Data, or the average characteristics of the pool of loans in the data may have changed. The challenge of distinguishing the effect of the pre-filing dual tracking restrictions from other provisions of the Rule can be overcome by focusing on borrowers with a loss mitigation application completed on or after becoming 90 days delinquent and before the servicer initiates foreclosure. While several other provisions affect loss mitigation applications and foreclosure, the pre-filing dual tracking provision is most relevant in this circumstance, as these other provisions generally govern servicer conduct either prior to receipt of a completed application, prior to the borrower becoming 90 days delinquent, or both. In the absence of a good baseline, an estimate of the effect of the provision can be obtained in two stages, equivalent to the presentation in Figure 9 above—first estimating the effect of the Rule on a complete application resulting in an accepted offer, and then the effect of an accepted offer on the probability of foreclosure. Finally, a regression model can account for other unrelated factors that may be coincidentally correlated with the Rule.

Specifically, the Bureau estimated two probit regressions. The first regression models the probability that a borrower with a complete loss mitigation application ultimately accepts a loss mitigation offer prior to the servicer initiating foreclosure, as a function of the year in which the borrower first became at least 90 days delinquent. The second regression models the probability of the servicer initiating foreclosure within one year as a function of whether or not the borrower ultimately accepted a loss mitigation offer. Both regressions control for differences in house prices, state, age of loan, servicer, and loan characteristics.

The results of these regressions are presented in Table 1, below. Controlling for other factors, becoming 90 days delinquent in 2015, under the Rule, rather than in 2012, before the Rule, is associated with increasing the probability of a borrower who completes a loss mitigation application ultimately accepting an offer by 20.4 percentage points. At the same time, on average a borrower accepting a loss mitigation offer reduces the probability that a servicer initiates foreclosure within one year by 22.9 percentage points, regardless of which year the application was completed. This would suggest that the Rule was associated with an increase in borrowers obtaining loss mitigation, but no change in the likelihood that loss mitigation agreements prevent foreclosure. The final column of Table 1 shows the total change in the probability of initiating foreclosure within one year, obtained by multiplying together the estimates from the first two columns. The pre-filing dual tracking restriction is associated with reducing the probability of foreclosure for consumers who complete applications within one year by 4.7 percentage points, from an average of about 19 percentage points.

Note that although these regression estimates control for a variety of observable loan characteristics, the change in the foreclosure rate for borrowers with completed loss mitigation applications may not have been caused by the Rule. In particular, between 2012 and 2015, numerous parties including the Bureau, the GSEs, and housing counselors invested resources in public education to encourage delinquent borrowers to apply for loss mitigation, and this could have led to an increase in accepted offers even without the Rule. Other trends in the housing market or the broader economy could account for this result as well, including the increased use of streamlined modifications.

TABLE 1: EFFECT OF THE PRE-FILING DUAL TRACKING RESTRICTION ON FORECLOSURE WITHIN 1 YEAR OF BECOMING 90 DAYS DELINQUENT, FOR BORROWERS WHO COMPLETED A LOSS MITIGATION APPLICATION

 

Effect of Rule on Probability of Accepted LM Offer

Effect of an Accepted Offer on Probability of Foreclosure Within 1 Year

Net Effect on Foreclosure

Estimate

0.204***

-0.229***

-0.0467

 

(0.0113)

(0.0051)

 

***P < .001

Estimates are average marginal effects from a probit regression. The sample consists of loans that first become 90 days delinquent in either 2012 or 2015 and complete a loss mitigation application between the month the borrower becomes 90 days delinquent and when the servicer initiates foreclosure. Standard errors reported in parentheses.

9.3.6 Effects of the post-filing dual tracking restriction

The post-filing dual tracking restriction prevents servicers from proceeding with a foreclosure sale while a complete loss mitigation application is pending if the application is completed more than 37 days before a scheduled foreclosure sale date. Although industry stakeholders, including servicers the Bureau interviewed for the assessment, asserted that this provision has increased costs by extending foreclosure timelines, both the evidence previously presented in this report and other feedback from industry stakeholders suggest that this provision had little aggregate effect on either borrowers or industry. Figures 5 and 6, above, show that loss mitigation applications are most often completed shortly after the servicer initiates foreclosure. This is consistent with servicer feedback to the Bureau's interviews. Together these facts suggest that the post-filing dual tracking restriction would not be binding in most cases: Both the Rule and many pre-Rule servicing standards required servicers to render a decision on a complete application within 30 days. At the same time, the timeline to complete a foreclosure is generally much longer than 30 days, even in states without judicial foreclosure. Nonetheless, this section explores further whether the post-filing dual tracking restriction meaningfully affected loans in foreclosure.

While analysis earlier in this chapter shows that servicers are initiating foreclosure later in delinquency after the Rule took effect, this chapter has not yet discussed the timeline after the first notice or filing. Figure 10, below, shows the average number of days between initiating foreclosure and the eventual sale, broken out by the year foreclosure was initiated and whether a loss mitigation application was completed after initiating foreclosure. Note that the data cannot capture a lag of more than two and a half years between initiating foreclosure and sale for loans where foreclosure is initiated in 2015 because the data end in June 2017. To ensure that the data are comparable across years, the figure excludes loans where foreclosure is initiated in 2012 and the ultimate foreclosure sale was after June 2014. The figure shows that, in fact, at least for the servicers in the Servicing Operations Data, there has been almost no change in the average timeline between initiating foreclosure and eventual sale. If anything, the average number of days has decreased slightly. In theory the post-filing dual tracking restriction may lead to longer foreclosure timelines for borrowers who apply for loss mitigation. However, in practice, there was little change in the timeline for borrowers with loss mitigation applications, although both before and after the Rule loans with loss mitigation applications completed after the servicer initiates foreclosure take longer to result in a sale. Note that this result is consistent with the results in Chapter 3 using the McDash Data, which show that the median number of months between foreclosure referral and sale stayed constant in states without judicial foreclosure, and declined somewhat in states with judicial foreclosure.

FIGURE 10: NUMBER OF DAYS BETWEEN INITIATING FORECLOSURE AND EVENTUAL SALE, BY YEAR AND WHETHER OR NOT THE BORROWER COMPLETED A LOSS MITIGATION APPLICATION (SERVICING OPERATIONS DATA)

Servicers also noted instances of borrowers "gaming" the post-filing dual tracking restriction to stave off a foreclosure sale. However, as noted above, both before and after the effective date of the Rule, most borrowers who apply for loss mitigation and have foreclosure initiated complete their applications either before or shortly after foreclosure is initiated. More than two-thirds of such borrowers complete their applications on or before 90 days following the servicer initiating foreclosure. The data show that most foreclosures in states with judicial foreclosure take significantly longer than 90 days from initiation to sale, and even in non-judicial foreclosure states a majority of foreclosures take longer than 90 days from initiation to sale. As a result, a minority of applications were completed late enough in foreclosure to potentially interfere with a planned sale.

One way to test for the prevalence of strategic loss mitigation behavior is to look at the frequency of loss mitigation applications very close to a foreclosure sale. If borrowers are using the provision strategically to delay foreclosure, one would expect to see an increase in applications completed just over 37 days from the scheduled date of the foreclosure sale. The loss mitigation application information in the Servicing Operations data sometimes include the date of the scheduled foreclosure sale as of the date of the loss mitigation application. These data are limited, however, as servicers were often unable to provide a date, likely because a sale was not scheduled at the time of the loss mitigation application. As an alternative, it is possible to test whether borrowers frequently complete loss mitigation applications around a foreclosure sale using the distribution of the timing of completed loss mitigation applications relative to the actual sale date. This will not capture strategic timing on the part of borrowers as reliably, because the actual sale date may have been rescheduled following a strategically timed loss mitigation application. However, the data are more complete for actual foreclosure sale dates.

Figure 11 shows the distribution of completed loss mitigation applications in 2012 and 2015 relative to the scheduled foreclosure date at the time the application is completed, focusing on applications completed at most 90 days prior to a scheduled sale. A vertical gray line denotes 37 days before the scheduled sale. There is cyclical pattern within weeks, likely because servicers predominately record loss mitigation applications as being complete on weekdays, while the data show that foreclosure sales are disproportionately scheduled for Tuesdays. Despite this cyclical pattern, there is no evidence of borrowers "bunching" their loss mitigation applications around 37 days before a scheduled foreclosure sale. There is a relatively large share of applications completed less than 30 days before scheduled sale dates, but these would not be affected by the post-filing dual tracking restriction of the Rule as they are past the 37 day cut-off in the provision. Moreover there was no change in the share of applications completed less than 37 days before a scheduled foreclosure sale between 2012 and 2015. If anything, there may have been an increase in applications received just after the 37 day cut-off in 2015 compared to 2012.

FIGURE 11: DAYS REMAINING BEFORE SCHEDULED FORECLOSURE SALE AT THE TIME A LOSS MITIGATION APPLICATION IS COMPLETED, FOR LOANS WHERE AN APPLICATION WAS COMPLETED BETWEEN 0 AND 90 DAYS BEFORE SCHEDULED SALE (SERVICING OPERATIONS DATA)

Figure 12 shows the distribution of loss mitigation applications relative to the actual sale date for completed sales, again for loans in 2012 and 2015, denoting 37 days with a vertical gray line. Again there is no evidence of bunching. The distribution is somewhat noisy, but there is no evidence of a change on either side of 37 days prior to the eventual sale. There is a relative increase in applications completed close to the ultimate sale date.

FIGURE 12: DAYS REMAINING BEFORE ACTUAL FORECLOSURE SALE DATE WHEN LOSS MITIGATION APPLICATION COMPLETED, FOR LOANS WHERE AN APPLICATION WAS COMPLETED BETWEEN 0 AND 90 DAYS BEFORE ACTUAL SALE (SERVICING OPERATIONS DATA)

To summarize the results on the post-filing dual tracking restriction, Figure 10 above suggests that the post-filing dual tracking restriction had little effect on foreclosure timelines, and Figures 11 and 12 indicate that the provision does not appear to have been used to delay foreclosure timelines, strategically or otherwise. Thus, it does not appear that the provision can have had much effect on whether or not borrowers ultimately experienced a foreclosure sale. Conversely it does not appear that the provision created unintended costs by unnecessarily delaying foreclosures, either by extending timelines in general or by encouraging strategic behavior among borrowers. It is possible that strategic behavior by borrowers does occur upon occasion, and it is very likely that dealing with such cases is disproportionately frustrating, time consuming, and costly for servicers, compared to handling the cases of other borrowers who do not attempt strategies to delay foreclosure. However, the data indicate that such behavior is not widespread.

10. Error resolution

This chapter considers the Rule's error resolution provisions, which require servicers to follow certain procedures if a borrower claims that the servicer made an error, including providing a response within a certain amount of time. The Bureau's main findings about these procedures include the following:

  • In interviews, some servicers said the Rule's error resolution provisions required them to make significant changes to how they track and respond to error assertions, whereas others said that for them the provision required few changes. Servicers also said the Rule had little effect on whether borrowers submitted written error assertions. The Bureau did not determine the specific cost to servicers of complying with the Rule's error resolution provisions.
  • Data suggest that the rate of written error assertions per account fell by about half after the Rule's effective date compared to the prior three years. The decrease largely reflects a decline in error assertions related to loss mitigation. There was, however, considerable variation among servicers, with some servicers showing a substantial increase in formal error assertions and others experiencing a substantial decrease.
  • Servicers may have become more responsive to error assertions by borrowers, with some servicers providing written acknowledgments of an assertion of error faster and more frequently. There is also evidence that borrowers submitted fewer follow-up or repeat error assertions after the Rule, consistent with servicers becoming more responsive. On the other hand, there is no evidence that error assertions under the Rule generally changed the likelihood that borrowers' loss mitigation applications were approved.

10.1 Background

Since 1990, RESPA has required servicers to respond to ''qualified written requests'' (QWRs)— written requests from a borrower that seek to resolve an error or seek information relating to the ''servicing'' of the borrower's mortgage loan. Servicers were required to: acknowledge receipt of a QWR within 20 business days; conduct an investigation, make appropriate corrections to the borrower's account; and respond within 60 business days, providing the borrower with a written explanation of the outcome and a contact number of an individual who could assist the borrower. Section 1024.21(e) of Regulation X implemented these statutory requirements prior to the 2013 RESPA Servicing Rule.

In addition, since 2012, servicers that were covered by the National Mortgage Settlement were required to acknowledge certain complaints within 10 business days and "provide a substantive written response within 30 days," but these requirements applied only to complaints from borrowers who were in default or had applied for loan modifications that were forwarded via designated federal agencies or state attorneys general.

Servicers that serviced loans for Freddie Mac or Fannie Mae also had general obligations to respond to borrower complaints. Freddie Mac's 2012 Servicing Guide directed servicers to

"have procedures in place to respond to complaints," while Fannie Mae's 2012 Servicing Guide gave similar instructions for "borrower inquiries," requiring servicers to respond promptly and in accordance with RESPA requirements. In the case of "escalated complaints," Freddie Mac imposed a timeline and specific business capability requirements.

The Dodd-Frank Act amended RESPA to include additional servicer obligations to borrowers who submit QWRs. Specifically, the Dodd-Frank Act amendments established shorter timeframes for servicers to acknowledge, investigate, and resolve borrowers' error assertions. The amendments also described specific categories of asserted errors to which servicers must respond in a timely manner, including "borrower's requests to correct errors relating to allocation of payments, final balances for purposes of paying off the loan, or avoiding foreclosure, or other standard servicer's duties."

The 2013 RESPA Servicing Rule implemented these Dodd-Frank Act amendments. Under the Rule, in general, if a borrower sends a written request asking a servicer to resolve an error (a notice of error or "NOE"), the servicer must: (1) within five business days, acknowledge the notice of error; and (2) within 30 to 45 days, correct the error and provide the borrower written notification of the correction, or conduct an investigation and provide the borrower written notification that no error occurred.347 Similarly, when a borrower sends a written request that the servicer send them information about their account (information request), the servicer must: (1) within five business days, acknowledge the request; and (2) within 30 to 45 business days, provide the information or conduct a reasonable search for the requested information and provide the borrower with a written notification explaining why the information is not available.348 The Rule specifies that a QWR that asserts an error relating to the servicing of a mortgage loan is a notice of error for purposes of the error resolution provision.349 The Rule does not require servicers to respond to duplicative or overly broad NOEs and information requests.350 Servicers are permitted to set up a dedicated address to receive NOEs and information requests, as long as they notify borrowers of this address.351 The Rule does not exempt "small servicers," which includes any servicer that services 5,000 loans or fewer, all of which the servicer owns or originated,352 from these requirements.353 The Rule also requires a servicer to establish policies and procedures reasonably designed to investigate, respond to, and, as appropriate, make corrections in response to complaints asserted by a borrower. A servicer also must provide a borrower with accurate and timely information and documents in response to a borrower's request for information with respect to the borrower's mortgage loan.354

In amending Regulation X, the Bureau sought to address pervasive consumer protection problems across major segments of the mortgage servicing industry, caused, in part, by servicers' failure to maintain and provide accurate information to borrowers about their mortgage loan accounts. The Bureau noted that a servicer's obligation to maintain accurate and timely information regarding a mortgage loan account to a borrower is one of its most basic servicing duties. The Bureau stated that it aimed to further the consumer protection purposes of RESPA, including ensuring responsiveness to borrower requests and complaints and the provision and maintenance of accurate and relevant information.

In addition, the Bureau intended for the error resolution and information request provisions of the Rule to provide borrowers with better tools and information when dealing with their mortgage servicers. The Bureau expected the greater scope and clarity of the new error resolution process would allow borrowers who would otherwise not assert errors at all or would assert them informally to obtain the benefits of the new investigation and response requirements of the error resolution process.358

The Bureau's assessment focused on the error resolution provisions and generally did not analyze the information request provisions. The analysis of the error resolution provisions sought to understand how these provisions have affected servicers' responses to borrowers' formal assertions of error. The next section of this chapter describes what the Bureau learned from servicer interviews, comment letters on the RFI, and the Counselor Survey about the error resolution provisions' effects on borrowers and servicers. The following section uses the Servicer Operations Data to analyze (a) trends in the number of error assertions received by servicers, (b) changes in the timelines for servicers to respond to error assertions, (c) evidence on whether borrowers had less need to make repeated or follow-up complaints due to the Rule, and (d) evidence on whether the error resolution provisions assisted borrowers with pending loss mitigation applications.

10.2 Qualitative information: Servicer interviews and Counselor Survey

10.2.1 Effects on servicer practices

Servicers offered very different views on the effect of the Rule's error resolution provisions. Prior to the Rule, all servicers interviewed had procedures in place for responding to error assertions and were subject to QWR requirements. Most small servicers and some large servicers said that this approach to borrower assertions of error was already in place and complied with the Rule's requirements, and that the only changes they made were updating policies and procedures to reflect the specific Rule requirements and adding monitoring and quality control processes to ensure and demonstrate compliance. Other servicers, however, said they made significant changes to their procedures for responding not only to notices of error as defined by the Rule but also to other borrower complaints.

Servicers that made significant changes to their procedures said that specific provisions of the Rule that drove these changes were the timing requirements for acknowledging and responding to errors and the new classification of specific categories of complaints. Some also said they reworked their error or complaint response processes as a result of not only the error resolution provision itself but also the greater regulatory focus on servicing policies and procedures. Partly because of this, some said that changes to their processes encompassed both written notices of error that are covered by the Rule and other asserted errors that are not covered by the Rule, such as those made by telephone or email. Some said they expanded the types of error assertions that were subject to the procedures specified in the Rule beyond a relatively narrow set of written error assertions that they would have previously considered subject to the QWR requirements. To comply with shorter timelines, some servicers said they changed their staffing practices, with some saying that they increased the level of specialization among staff responsible for responding to error assertions to improve their ability to quickly and consistently respond to errors. Some servicers created a designated address for submission of notices of error.

10.2.2 Effects on borrowers

Counselor Survey. Almost one-half of the housing counselors surveyed said they rarely or never discuss the error resolution provisions with their clients. Overall, housing counselors provided fewer responses about the error resolution provisions than any of the other provisions discussed in the survey. Of those who discussed the provisions, 38 percent of housing counselors surveyed said that the error resolution provisions were effective in helping their clients, and an additional 49 percent said that the provisions were "somewhat effective." When asked about the reasons these provisions were effective or not effective, a majority of respondents said that when the provisions are effective it is because they help their clients obtain information about their mortgage loans. In cases where counselors believe the requirements are not effective, more than one-half said it was because the servicer does not follow the Rule. In responses to open-ended questions, several housing counselors said that the error resolution provisions were very helpful in getting servicers to respond to their clients. Other open-ended responses indicated that in some cases servicers respond to formal error assertions with form responses that do not address the issue raised in the complaint. Related to the survey results, in a comment letter, a consumer group cited other survey results indicating that most housing counselors believed that as a result of the Rule it was easier to correct errors or obtain requested information.

Servicer Interviews. Most servicers said that borrowers with concerns about their loans are most likely to contact their servicers by telephone before sending a written notice of error. Servicers generally did not believe that the Rule's error resolution provision had had any effect on how likely borrowers are to assert an error or whether they did so in writing.

Some servicers said the Rule had led to shorter timelines for responding to error assertions and a more consistent treatment of error assertions, which they identified as consumer benefits. Most servicers said they believed acknowledgment notices were helpful for borrowers in reassuring them that servicers had received their error assertion and were considering them. However, some servicers said that standard-form acknowledgment notices they provide to comply with the Rule could be confusing to borrowers in circumstances where they seem inconsistent with other communications about the alleged error—for example, if the borrower has been discussing the alleged error by telephone or if the resolution notice arrives at about the same time as the acknowledgment notice.

Some servicers said that by redesigning their error resolution practices they are now better able to track and analyze complaints and error assertions, and that internal review of improved complaint data led to process improvements that can help borrowers, particularly in how servicers communicate with them about topics that are common sources of complaints.

10.2.3 Costs of the provision and other effects on servicers

For servicers that fundamentally changed their system for responding to complaints and error assertions, this provision was among the most costly to implement, with changes to staffing practices and tracking systems. Some servicers reported staffing increases to respond to errors and to monitor compliance with the Rule. Servicers that did not make significant changes to their error resolution practices naturally did not report significant costs of complying with the error resolution provision. However, such servicers generally did need to update policies and procedures and add monitoring to ensure that their practices complied with requirements of the

Rule.

Some servicers emphasized that the requirement to resolve errors within 30 days had caused them to add staff to help with error processing. As with certain other timing requirements, some servicers said that while it is not a challenge to meet the deadline in most cases, there is a relatively small subset of cases for which it is a challenge to do so. These servicers said that a disproportionate part of the cost of complying with the provision is trying to avoid exceeding the Rule's deadline for a response in this relatively small number of cases. For example, one commenter said that the 30-day response timeline is difficult or impossible to meet under some circumstances, such as where action by a court or another third party was necessary to correct the error.

Some servicers said they receive many overly broad or duplicative error notices and that these can be costly to deal with. Under the Rule, servicers are not required to respond to requests that are duplicative or overly broad. However, some servicers said that they hesitate to decline to respond on those grounds because of the risk of non-compliance if the error were later determined not to be overly broad or duplicative. As a result, these servicers said that they attempt to respond to even extremely broad error notices or subject such notices to additional scrutiny, including by legal counsel, before declining to respond to them.

10.3 Quantitative analysis: Servicing Operations Data

For purposes of this chapter the Bureau assembled a dataset from several of the tables included in the Servicing Operations Data. The primary data consist of a record for each communication in which the borrower asserted an error (error assertions) received by each servicer for each loan serviced as of January 1, 2012, and for each loan serviced as of January 1, 2015. The Bureau's data requests specified for servicers to provide complaints, QWRs in which an error was asserted, and NOEs. However, four servicers also provided some communications which they classified as information requests, perhaps because these requests included a complaint. For servicers that provided a random sample of loans, the records related to error assertions for sampled loans only. The data cover communications received between January 1, 2011 and June 30, 2017, although for purposes of this chapter, the Bureau utilized only error assertions received before January 1, 2014, for loans that were serviced by a given servicer in 2012, and only error assertions on or after January 1, 2014, for loans serviced in 2015. Cutting the data in this manner prevents duplication for loans that were serviced by the same servicer in both 2012 and 2015, and it enabled the Bureau to more cleanly associate error assertions with other data provided by servicers.

The error resolution data have particular limitations that are distinct from other types of data included in the Servicing Operations Data. First, to a greater extent than other types of data, different servicers tracked these data differently. For instance, some servicers were able to provide information on oral complaints, while others were not. Second, as discussed above, in interviews some servicers reported that they improved their tracking of assertions of error in response to the error resolution provisions of the Rule. For example, some servicers began tracking additional information about complaints in defined fields rather than in free-form note fields. As a result, changes between the pre-Rule and post-Rule periods may reflect changes in what is systematically tracked rather than changes in actual error assertions by borrowers.

Another limitation is that the Servicing Operations Data do not indicate whether or not servicers resolved the error assertions in borrowers' favor. Although this information was requested from servicers, only two servicers were able to determine the resolution of an error assertion from their databases without manually reviewing the records of each error assertion. Of these two servicers, one had very few error assertions to begin with, and the other was only able to report a resolution for a small subset of error assertions. Thus, the Servicing Operations Data does not address whether the error resolution provisions have helped borrowers to obtain their desired resolution more frequently than occurred before the Rule.

For some analyses discussed below, the Bureau linked the error assertion data to data on any loss mitigation applications that were completed around the same time as the error assertions were received by the servicers.

10.3.1 Trends and averages of error assertions

This subsection provides a basic description of error assertions included in the Servicing Operations Data. Figure 1 shows the frequency of the various types of error assertions reported by the servicers in the data. To account for the differences in portfolio size across servicers and changes in portfolio size over time, these frequencies are presented as a fraction of the total number of accounts serviced by each servicer. In other words, for each servicer the frequencies are the total number of error assertions of each type received by that servicer in each period (even if one borrower submitted multiple error assertions of a given type), divided by the total number of accounts serviced in that period.

FIGURE 1: ERROR ASSERTIONS PER ACCOUNT, FOR LOANS SERVICED IN 2012 AND FOR LOANS SERVICED IN 2015 (SERVICING OPERATIONS DATA)

The figure shows that formal written error assertions per account fell by about two-thirds since the effective date of the Rule. More precisely, the servicers in the Servicing Operations Data reported receiving about one-third as many NOEs per account between 2014 and 2017 as they received QWRs in which the borrower asserted an error between 2011 and 2013. There was also an increase in reported complaints per account. Although the Rule did not remove the QWR definition from RESPA, borrower contacts that servicers categorized as QWRs essentially disappeared starting in 2014 and were replaced by NOEs. Across different servicers, the proportion of error assertions between 2014 and 2017 that were deemed to be formal NOEs varied between 5 percent and 32 percent.

There were many more error assertions per account from 2014 to 2017 that servicers categorized as complaints or similar classifications, compared to 2011 to 2013. However, this change likely does not represent a true increase in complaints received by servicers. As noted above, several of the servicers the Bureau interviewed said that, since implementation of the Rule, they upgraded their systems to allow for better tracking of complaints that assert errors, beyond formal written requests as required by the Rule. That is, the servicers may have received a similar number of complaints per account between 2011 and 2013, but may not have stored data on those complaints in such a way that they could provide them to the Bureau. It is also possible that some error assertions that servicers would have identified as QWRs in 2011 to 2013 were categorized as complaints or information requests in the 2014 to 2017 period. In addition, the Bureau's own complaint mechanism may have played a role in the apparent reduction in formal error assertions. Consumers who might otherwise have sent their servicer a QWR in 2012 to address an error may instead have filed a complaint with the Bureau in 2015. Although one servicer in the Servicing Operations Data included complaints referred by the Bureau in its error assertion data, this was not generally the case, and moreover the servicer in question classified Bureau complaints as complaints rather than NOEs.

Both before and after the Rule became effective, error assertions were made disproportionately by borrowers who were having trouble paying their mortgages. Figure 2, below, breaks out the number of formal error assertions (QWRs or NOEs) per account by whether or not the borrower was more than 30 days delinquent or began a loss mitigation application at any point in the same calendar year that the error assertion was made. Note that being in the delinquent group does not necessarily mean the borrower was delinquent or applying for loss mitigation at the time of the error assertion, although these events had to occur in the same calendar year. There were roughly 10 times more formal error assertions per account submitted by borrowers who were more than 30 days delinquent or applied for loss mitigation during the same year, compared to borrowers who were never delinquent and never applied for loss mitigation. Consistent with Figure 1, above, the number of NOEs in 2015 was about one-third the number of QWRs in 2012 for the delinquent group, while for non-delinquent borrowers, the number increased slightly, from a very low base.

FIGURE 2: FORMAL ERROR ASSERTIONS PER ACCOUNT, BY YEAR AND WHETHER THE ACCOUNT WAS DELINQUENT OR APPLIED FOR LOSS MITIGATION IN THAT YEAR (SERVICING OPERATIONS DATA)

Many borrowers submit multiple error assertions in the data, so the share of borrowers that have asserted an error is smaller than the ratio of error assertions to accounts. Overall, 1.8 percent of borrowers submitted a QWR during the 2012 period and 0.9 percent of borrowers submitted a NOE during the 2015 period.

As noted above, the changes in error assertions between 2012 and 2015 may largely reflect changes in how servicers tracked and categorized error assertions, which varied by servicer. Figure 3 plots for each servicer the percent change in QWRs and NOEs per account serviced between 2011 and 2013 and between 2014 and 2017. To maintain servicer anonymity the Bureau randomly assigned each servicer a letter to represent it. Note that, as with servicerspecific results in previous chapters, the Bureau randomly re-assigned letters for subsequent figures, such that servicer "A" in Figure 3 is not necessarily the same as servicer "A" in any other figure. One servicer reported receiving almost no QWRs from accounts it serviced in 2012, and so it is omitted to avoid erroneously skewing the scale of the figure.

FIGURE 3: PERCENT CHANGE BETWEEN 2011-2013 AND 2014-2017 IN FORMAL NOTICES OF ERROR OR QUALIFIED WRITTEN REQUESTS PER ACCOUNT, BY SERVICER (SERVICING OPERATIONS DATA)

The overall decrease in NOEs compared to QWRs shown in Figure 1 is the combination of substantial increases from some servicers, in one case an increase of over 200 percent, and moderate to large decreases from others.

The proportion of borrowers who submitted at least one formal error assertion also varied widely across servicers. Figure 4 shows the percentage of overall accounts at each servicer that submitted at least one formal error assertion, with separate results for before and after the Rule. Note that this is a different calculation than the previous three figures—each borrower is counted once, no matter how many error assertions they submitted. With the exception of one servicer that received formal error assertions from more than 13 percent of loans in the data before the Rule, formal error assertions came from fewer than 5 percent of borrowers at all servicers before and after the Rule.

FIGURE 4: PERCENTAGE OF ALL BORROWERS WHO SUBMIT A FORMAL ERROR ASSERTION, BY TIME PERIOD (SERVICING OPERATIONS DATA)

Error assertions were made for a variety of reasons, reflected in the wide range of codes used by servicers in the data provided to the Bureau. The Bureau grouped each servicer's error codes into eight broad categories that are consistent across servicers. Figure 5 plots the total number of formal error assertions (QWR's and NOE's) per account serviced that were received before and after the effective date of the Rule, broken down by the category of the error reason. Note that this is more precisely the total number of formal error assertions with a reported error reason; no reason was reported for about 43 percent of formal error assertions in 2012. Less than 1 percent of formal error assertions in 2015 had a missing reason.

Almost one-half of formal written error assertions received by servicers from accounts they serviced in 2012 were about loss mitigation, and about 6.5 percent related to foreclosure. Other formal error assertions were split about evenly between other types of issues such as payments, a category which includes payment application, billing statement, and loan payoff issues. In the post-Rule period, covering loans serviced in 2015 and error assertions received between 2014 and 2017, formal written error assertions about loss mitigation represented a smaller share of the total formal written error assertions and a much smaller number of such assertions overall. However, the rate of initiated loss mitigation applications also declined over this period, and the decline in loss mitigation applications is roughly proportional to the decline in formal error assertions related to loss mitigations. The share of foreclosure-related error assertions also declined somewhat. The remaining error assertions continued to be roughly evenly split between the remaining categories although the levels generally increased. However, once again it is important to note that servicers were more able to provide error reasons in the post-Rule period, and so the change in levels may reflect improved reporting rather than an actual increase in errors of these types.

FIGURE 5: REASONS FOR ERROR ASSERTIONS, BEFORE AND AFTER THE RULE (SERVICING OPERATIONS DATA)

10.3.2 Timelines of responses to error assertions

As noted above, one of the specific requirements of the Rule was to set shorter deadlines for acknowledging and responding to error assertions. The Servicing Operations Data include the date the servicer received each error assertion, the date of any acknowledgement, and the date of any response. With these data, the Bureau calculated the distribution of acknowledgement and response times. This analysis is limited to written formal error assertions for this analysis, but the results look substantially similar with all error assertions included.

Some servicers frequently do not record a date when an acknowledgment was sent, even for QWRs and NOEs. While this may include cases where an acknowledgement was not sent, it also may reflect the servicers improving their tracking of written acknowledgements for error assertions. In particular, before the Rule servicers had 20 days to send an acknowledgment letter, and may have omitted that letter when they had already resolved the error assertion prior to the deadline. Figure 6, below, shows the proportion of formal error assertions with a nonmissing acknowledgement date for each servicer both before and after the Rule. While three of seven servicers almost always recorded an acknowledgement date both before and after the Rule, the remaining servicers had an acknowledgement date for a minority of formal error assertions before the Rule, increasing substantially after the Rule. In contrast, all servicers except for one were able to provide a date for the letter reporting the outcome of the error assertion for all or almost all formal error assertions.

FIGURE 6: PROPORTION OF FORMAL ERROR ASSERTIONS WITH NON-MISSING ACKNOWLEDGEMENT DATE, BY YEAR RECEIVED (SERVICING OPERATIONS DATA)

Figure 7 shows the distribution of the number of days from receipt to acknowledgment in a boxand-whisker plot, broken out by servicer and time period. Note that because the size of the box in a box-and-whisker plot is based on the difference between the 25th and 75th percentiles, one servicer's plot collapses at zero because the servicer reported providing an acknowledgement for more than 75 percent of all error assertions on the day the error assertion was received. The dotted horizontal line at five days corresponds to the requirement of the Rule.

FIGURE 7: DISTRIBUTION OF DAYS FROM RECEIPT OF AN ERROR ASSERTION TO SENDING AN ACKNOWLEDGMENT, BY SERVICER AND TIME PERIOD WHEN THE ASSERTION WAS RECEIVED (SERVICING OPERATIONS DATA)

The timeline for acknowledgements largely accelerated in the post-Rule period, particularly at two servicers that previously had relatively long timelines for providing acknowledgments. At the same time, prior to the Rule four servicers in the data were already providing acknowledgments in less than five days in most cases, and the Rule does not appear to have affected their behavior.

Figure 8 shows the distribution of the number of days between date of receipt of an error assertion and the date the servicer sent notice of the outcome of its resulting investigation. Again the distributions are broken out by servicer and time period. As noted above, the Rule generally requires that servicers respond to assertions of error within 30 days, with the possibility of a 15-day extension in some cases. A horizontal dotted line in the figure denotes 30 days.

FIGURE 8: DISTRIBUTION OF DAYS FROM RECEIPT OF AN ERROR ASSERTION TO SENDING NOTICE OF A CORRECTION OR THAT NO ERROR OCCURRED, BY SERVICER AND TIME PERIOD WHEN THE ASSERTION WAS RECEIVED (SERVICING OPERATIONS DATA)

It appears that servicers in the Servicing Operations Data already provided most responses to error assertions in less than 30 days in the pre-Rule period. Indeed, if anything it appears that for six out of the seven servicers in the data, the median time between the receipt of an error assertion and the servicer's official response was longer post-Rule. It is not certain how this increase in response times affected borrowers. It may be that servicers are using the additional time afforded by the Rule to consider and investigate error assertions more thoroughly than previously. It is also possible that this is simply a selection effect—complaints that the servicers began tracking only after the effective date of the Rule may take longer than those that they tracked before the Rule became effective. However, given that the results are similar including the broader set of informal complaints, this seems less likely as an explanation.

10.3.3 Repeated or follow up complaints

One potential benefit of the error resolution provisions is that making servicers more responsive to error assertions (either due to the required timelines or more standardized policies and procedures) may reduce the need for borrowers to file multiple complaints or follow-up with time-consuming telephone calls. The Bureau cited this as a potential benefit in the Rule itself, and multiple servicers interviewed by the Bureau cited reduced follow-up complaints as a benefit of the Rule.

The next analysis examines cases where a borrower submits a formal error assertion (QWR or NOE) soon after the submission of another assertion. Table 1 shows how often this occurs. For QWRs received by servicers between 2011 and 2013, almost 17 percent were submitted within 30 days after another error assertion by the same borrower. Extending the time period to 90 days, this figure rises to almost 28 percent. Clearly, it was relatively common for borrowers to submit multiple error assertions to their servicer even within a relatively short timeframe. This phenomenon decreased somewhat after the effective date of the Rule. Just over 12 percent of all error assertions received between January 2014 and June 2017 were received within 30 days of another error assertion by the same borrower. There were larger decreases using a longer time period. It is also possible to restrict the analysis to potentially identical error assertions—that is, those formal error assertions that are made using the same method (mail vs. electronic) and have the same reported error reason. Potentially identical follow-up error assertions make up more than half of all follow-up error assertions.

TABLE 1: RATES OF FOLLOW-UP ERROR ASSERTIONS BY THE SAME BORROWER (SERVICING OPERATIONS DATA)

Time Period Error Assertion Received:

2011-2013

2014-2017

Proportion with Additional Error Assertion Within:

30 Days

0.168

0.123

60 Days

0.237

0.181

90 Days

0.275

0.214

Proportion potentially Identical Error Assertion Within:

30 Days

0.0880

0.0679

60 Days

0.130

0.104

90 Days

0.153

0.124

10.3.4 Error resolution and loss mitigation applications

As discussed in Chapters 7, 8, and 9, the Bureau sought to establish procedures for loss mitigation that would help borrowers avoid foreclosure when possible. While servicers' procedures for handling error assertions do not directly relate to loss mitigation, Figure 5 shows that a large fraction of error assertions relate to loss mitigation. Thus, to the extent that the Rule affected the way that servicers resolved error assertions regarding loss mitigation applications, it may affect the ultimate outcome of those applications.

The Bureau analyzed the effect of an error assertion on the success of a loss mitigation application, specifically whether the Rule's error resolution provisions help borrowers resolve problems with their loss mitigation applications. A challenge with this analysis is that error assertions are not made randomly. The fact that the borrower felt the need to submit an error assertion to his or her servicer indicates that the borrower felt the servicer had made an error, presumably one that the borrower believed would hurt his or her prospects for a successful loss mitigation application. All else equal, then, loss mitigation applications for which the borrower submitted an error assertion may be less likely to succeed, regardless of how the servicer handles the error assertion. However, it is plausible that this difference between loss mitigation applications with and without an error assertion is constant over time and would not change following the effective date of the Rule. Thus, it is possible to use the difference between loss mitigation applications with and without an error assertion in the pre-Rule period as a baseline or control group.

Specifically, the Bureau estimated a difference-in-differences regression model, comparing the change in the success rate of loss mitigation applications before and after the Rule, and with and without an error assertion. For purposes of this analysis, an observation is a loan with at least one loss mitigation application completed in 2012 or 2015, and a successful loss mitigation application is defined as one for which the borrower was offered at least one loss mitigation option, regardless of whether the borrower accepted the offer. A borrower is flagged as having submitted an error assertion if there was at least one error assertion received by the servicer after the borrower's first observed loss mitigation application was complete, and no more than 90 days after the borrower's last observed loss mitigation application. The regression allows for separate effects for each servicer, accounting for the very different trends that each servicer had in both error assertions and loss mitigation applications, and controls for differences in success rates across states and product types (e.g., GSE, FHA, portfolio).

Figure 9 shows the results of the difference-in-differences analysis, plotting the size of the effect of an error assertion under the Rule at each servicer. The results do not show evidence that error assertions under the Rule generally assist borrowers with resolving problems with their loss mitigation applications. For two of the five servicers in the regression, the estimated effect is essentially zero, while for two others, the probability that a successful loss mitigation application resulted in an offer actually decreased when an error assertion was received by the servicer in 2015, compared to when an error assertion was received by the servicer in 2012. For one servicer the effect of the Rule was positive, with the probability of a loss mitigation offer increasing with an error assertion received in 2015, although this estimate is sufficiently noisy that a 95 percent confidence interval includes zero. The results in Figure 9 use all error assertions to define the treatment variable in order to maximize the sample size, but the results are substantively similar if the sample is restricted to QWRs and NOEs.

Note that the key assumption of this analysis is that the differences, on average, between borrowers with loss mitigation applications who submitted error assertions and those who did not remained the same from 2012 to 2015 except for the direct effect of the error resolution provisions of the Rule. This may not in fact be the case. Other factors including other provisions of the Rule may have increased this difference. That is, the population of borrowers who submitted error assertions in 2015 may have been those who were inherently less likely to receive loss mitigation options compared to the population in 2012. For instance, the continuity of contact requirements discussed in Chapter 7 may have allowed more errors to be resolved informally, such that the remaining error assertions were more serious cases. Thus, it may not be the case that for two of five servicers, asserting an error under the Rule's error resolution provisions reduced the likelihood of receiving a loss mitigation offer; regardless, there is no evidence that the error resolution provisions of the Rule generally assisted borrowers in addressing problems with their loss mitigation applications.

FIGURE 9: EFFECT OF THE ERROR RESOLUTION PROVISIONS ON THE PROBABILITY THAT A LOSS MITIGATION APPLICATION PRODUCES AT LEAST ONE LOSS MITIGATION OFFER. (SERVICING OPERATIONS DATA)

11. Force-placed insurance

This chapter considers the Rule's force-placed insurance provisions. These provisions require servicers to send certain notices before charging a borrower for force-placed insurance and also include other restrictions on force-placed insurance. Key findings include:

  • Servicers interviewed said that the Rule's requirements with respect to force-placed insurance were generally consistent with the force-placed insurance policies and procedures that they had in place before the Rule was effective, so the Rule's effects on borrowers and servicers were small. The Bureau did not determine the specific cost to servicers of complying with the Rule's force-placed insurance requirements.
  • The data show a moderate decrease post-Rule in the share of borrowers receiving forceplaced insurance, a trend that is consistent with the Rule's force-placed insurance requirements being effective but is also consistent with other explanations, such as changes in the insurance market that made it easier or less expensive for borrowers to obtain insurance.
  • Borrowers were more likely to have insurance force-placed during a given year if they experienced delinquency or applied for loss mitigation during that year. Such borrowers also appear to respond less frequently than others do to force-placed insurance notices.

11.1 Background

Virtually all mortgage loan contracts require borrowers to maintain hazard insurance during the term of their loans. These contracts also permit lenders to charge borrowers for any hazard insurance lenders obtain to cover the mortgaged property if borrowers fail to maintain hazard insurance coverage. Force-placed insurance is hazard insurance that servicers obtain on behalf of the owner or assignee of a mortgage loan when the servicer is unable to obtain evidence that the borrower has complied with the obligation to maintain hazard insurance. A September 2015 GAO report found that force-placed insurance affects approximately 1 to 2 percent of all mortgaged properties and that premium rates for force-placed insurance are generally higher than rates for borrower-purchased insurance.

Prior to the effective date of the 2013 RESPA Servicing Rule, no federal regulations addressed communications with borrowers pertaining to force-placed insurance. However, investor guidelines and the National Mortgage Settlement (NMS) contained requirements regarding force-placed insurance. Fannie Mae discouraged servicers from using force-placed insurance. It required servicers to make multiple attempts to contact borrowers to obtain evidence of insurance and provide borrowers with at least 60 days from contact to provide proof of insurance before charging for force-placed insurance.377 Fannie Mae also required servicers to make certain disclosures concerning force-placed insurance cost and coverage and required servicers to refund overlapping charges if borrowers subsequently provided proof that a hazard policy was in place. Similarly, the NMS required servicers to provide notice to borrowers before obtaining force-placed insurance and to refund charges for overlapping coverage. The NMS also included other force-placed insurance provisions which were generally similar to those of the 2013 RESPA Servicing Rule, discussed below.

The Dodd-Frank Act amended RESPA and established certain limits on the use of force-placed insurance.380 The 2013 RESPA Servicing Rule's force-placed insurance requirements generally implemented specific force-placed insurance provisions that were included in the Dodd-Frank Act's amendments to RESPA.

The Rule's provisions include that servicers may not assess a premium or charge for forceplaced insurance unless the servicer: (1) has a reasonable basis to believe that the borrower has failed to comply with the mortgage contract's requirement to maintain hazard insurance; (2) has sent two notices to the borrower within specified timeframes with certain information before assessing the borrower any premium charge or fee related to the force-placed insurance; and (3) has not received in response to these notices evidence that the borrower has had required hazard insurance in place continuously. In addition, servicers must cancel forceplaced insurance within 15 days of receiving evidence that borrowers have required hazard insurance in place on their mortgaged property and refund any fees or charges for periods of overlapping coverage. Force-placed insurance charges imposed by servicers on borrowers, beyond those subject to state regulation as insurance charges, must be bona fide and reasonable.384 Servicers may not purchase force-placed insurance for certain borrowers with escrow accounts if servicers are able to disburse funds from the borrowers' escrow accounts to pay hazard insurance premium charges, regardless of whether funds in the escrow account are sufficient to cover the premium charges. "Small servicers" are generally not exempt from the requirements discussed in this paragraph.

The Bureau adopted these requirements to achieve Congress' intent to protect borrowers from the unwarranted and avoidable force-placement of hazard insurance, which the Bureau believed generally provides substantially less coverage for borrowers' property at a substantially higher cost than borrower-obtained hazard insurance policies. The Bureau adopted these provisions to achieve the consumer protection purposes of RESPA, which include the avoidance of unnecessary and unwarranted charges and fees and the provision to borrowers of accurate and relevant information.388

The Bureau's assessment of the force-placed insurance provisions sought to understand how these provisions have affected whether borrowers are charged for force-placed insurance, the circumstances under which force-placed insurance is obtained, and the types of borrowers that are ultimately charged for force-placed insurance. The next section of this chapter describes what the Bureau learned about the force-placed insurance provisions' effects on borrowers and servicers from servicer interviews, consumer complaints, and the Counselor Survey. The following section uses the Servicing Operations Data to analyze (a) how many borrowers received force-placed insurance notices and how many of those borrowers are charged for forceplaced insurance, (b) the timing of force-placed insurance notices relative to when borrowers are charged for force-placed insurance and relative to when existing hazard policies expire, and (c) the geographic distribution of force-placed insurance coverage.

11.2 Qualitative information: Servicer interviews, Counselor Survey, and consumer complaints

11.2.1 Effects on servicer practices

Servicers generally said that the Rule's force-placed insurance provisions had little effect on their practices. Some said this was because prior to the Rule they had already implemented the NMS requirements related to force-placed insurance. However, even those servicers that were not subject to the NMS said that the Rule's requirements were consistent with their practices pre-Rule.

All servicers the Bureau spoke with used a vendor for some or all activities related to forceplaced insurance, including tracking hazard insurance coverage, mailing notices, placing policies, and canceling policies when borrowers provide proof of hazard insurance. Some servicers said that after the Rule they began using a vendor, or using a vendor for a broader range of activities, although none of the servicers that made such changes said the changes were motivated solely by the Rule.

Based on interviews with servicers and vendors, the Bureau understands that before the Rule became effective servicers generally mailed at least two force-placed insurance notices prior to charging for force-placed insurance and that the timelines for these notices were consistent with the Rule requirements. In addition to mailing notices, servicers said they or their vendor attempted to obtain proof that hazard insurance was renewed directly from the previous hazard insurer and some servicers said that they attempted to telephone borrowers if they do not respond to written notices. Some servicers said they sent one additional written notice beyond those that are required, and one servicer said that before the Rule it sent three notices, but that it reduced that to two to conform their policies more closely to the Rule's requirements. Servicers had to make changes to the notices to include the specific content required by the Rule. A few servicers said that when mailing notices they include a separate page in addition to the required notice in order to provide additional information that is not in the required notice.

As with the notice requirements, servicers generally said that the Rule's other force-placed insurance provisions also did not require them to change their practices. Servicers said that where insurance is paid through an escrow account it was already their practice to advance funds to prevent an existing hazard policy from expiring (regardless of whether the escrow account has a positive balance), and some said that this was a general industry practice prior to the Rule. Servicers generally said that they did not add any charges for force-placed insurance beyond state-regulated premiums charged by the insurer, so that the requirement that any additional charges be "bona fide and reasonable" did not change their practice. Similarly, servicers said that before the Rule they already provided refunds of force-placed insurance premiums if the borrower provided evidence of a coverage overlap, so the Rule did not change their practices.

11.2.2 Effects on borrowers

Counselor Survey. About 42 percent of respondents to the Counselor Survey said the forceplaced insurance provisions were effective in helping their clients, with another 35 percent saying the provisions were somewhat effective. When asked about ways in which the provisions were effective, respondents were mostly likely to say that it was because it prompted their client to provide proof of insurance, with 42 percent of respondents saying this was "often" or "always" the reason. When asked about cases when the provision was not effective, respondents were most likely to say that it was not effective because the disclosures were not understandable to borrowers. Thirty-eight percent of respondents said that among cases in which the notices were not effective it was "often" or "always" because the notices were not understandable even for clients that understood English well. In responses to open-ended questions, several respondents stated that the notices provided by servicers fail to inform borrowers when force-placed insurance is more expensive or provides less coverage than borrower-obtained hazard insurance. Some stated that, as a result, borrowers may not seek to obtain insurance until after a closer review of their mortgage statements.

Servicer interviews. Several servicers told the Bureau that sending notices to borrowers was effective in encouraging them to provide proof of hazard insurance, but servicers generally said the Rule did not make a difference in this regard because they were already sending similar notices before the Rule. Similarly, the servicers said they were already following other forceplaced insurance provisions of the Rule.

Some servicers said that they had needed to add the estimated cost of a force-placed policy to their notices as a result of the Rule, and some of those that made this change said that they believed this helped motivate some borrowers to provide proof of hazard insurance. On the other hand, some servicers said that an unintended consequence of disclosing the cost of a force-placed policy was that for some borrowers this cost was less than the premium for a hazard policy they could obtain themselves, and that borrowers focused on the lower costs of force-placed insurance even though force-placed insurance provides much less comprehensive coverage.

Some servicers said that some borrowers complained to them about the language of the forceplaced insurance notices, which they found harsh or accusatory.

Complaint data. To identify complaints related to force-placed insurance, the Bureau used a keyword search in the consumer narrative field of mortgage servicing complaints. The Bureau received approximately 2,400 borrower complaints related to force-placed insurance between December 1, 2011 and December 31, 2017. Figure 1 below shows results of an analysis of a random sample of consumer complaints from 2012 and 2015 related to force-placed insurance. The Bureau manually reviewed complaint files, including the narrative complaint by the borrower, the servicer's response, and any supporting documents submitted by the borrower or the servicer. The analysis divided the resolution of consumer complaints into four categories based on how the servicer categorized the resolution of the complaint: (1) cases where a policy was correctly placed and was not removed; (2) cases where a policy was correctly placed and was removed (presumably after the borrower provided proof of an existing hazard policy); (3) cases where a policy was incorrectly placed; and (4) cases where a policy was not placed. The results show a drop in the share of cases in which the force-placed policy is removed, and an increase in the share in which it is not removed. That is, an increasing share of complaints about force-placed insurance arose from cases in which the servicer concluded that there was an ongoing lack of coverage. This is at least consistent with a reduction in the frequency with which borrowers are charged for force-placed insurance when they in fact have a hazard policy in place or are willing and able to obtain one.

There was also an increase in the share of complaints in which there was no placement of forceplaced insurance, although the numbers of complaints in this category were small in both 2012 and 2015. These may be cases in which consumers reacted negatively to notices they received before their servicers obtained force-placed insurance. The increase could mean that more servicers are providing notices or are providing them earlier, or could reflect some consumers reacting negatively to a change in the content of the notices.

FIGURE 1: RESOLUTION OF COMPLAINTS RELATED TO FORCE-PLACED INSURANCE, 2012 AND 2015 (BUREAU COMPLAINT DATA)

11.2.3 Costs of the provision and other effects on servicers and borrowers

In light of the small effect on practices, servicers generally said they incurred few additional costs as a result of the force-placed insurance provisions. Servicers made one-time changes to the content of the notices they sent, and some servicers said that they increased monitoring of their vendor to ensure compliance with the Rule provisions.

Many servicers said that after implementing the required language in force-placed insurance notices they received complaints from some borrowers who found the language harsh or accusatory. However, servicers interviewed generally said the overall effect of such complaints on their business was small.

11.3 Quantitative analysis: Servicing Operations Data

To assess quantitatively the effects of the force-placed insurance provisions of the Rule, the Bureau analyzed the Servicing Operations Data. The Servicing Operations Data include information about borrowers who received force-placed insurance notices during 2012 and 2015, including the dates notices were sent, whether and when insurance was force-placed, any refunds provided, and certain other details. For certain analyses, the Bureau combined these data on force-placed insurance with other variables in the Servicing Operations Data, such as delinquency status or the use of an escrow account.

Servicers had difficulty providing some of the information on force-placed insurance that the Bureau requested for the Servicing Operations Data. This difficulty stemmed in part from the fact that all servicers used a vendor for force-placed insurance, and as such some of these data were less readily accessible in the servicers' systems of record. The force-placed insurance data provided to the Bureau generally had more gaps and inconsistencies than other types of information discussed in this report. One servicer was unable to provide any data on force placed insurance for loans serviced in 2012. Despite these difficulties, the Bureau judged that the data used for the analysis in this chapter were sufficiently reliable to use for this analysis.

11.3.1 Share of borrowers receiving notices and being charged for force-placed insurance

Figure 2 focuses on borrowers who did not have force-placed insurance in place on January 1 of the relevant year, but who did receive a force-placed insurance notice during that year. The two panels of Figure 2 each show the share of borrowers who receive a first and second force-placed insurance notice, the share who were charged for force-placed insurance, and the share for whom force-placed insurance had not been cancelled within six months of placement. The analysis splits borrowers into two groups, based on whether borrowers showed evidence of having trouble paying their mortgage during the year the force-placed insurance notice was sent. Specifically, the analysis is split based on whether borrowers were more than 30 days delinquent at any point during the year or applied for loss mitigation during the year.

Among the borrowers who did not experience delinquency or apply for loss mitigation (the left panel), about 7 percent of borrowers were sent the first force-placed insurance notice, but only about one-third of those were sent the second notice (presumably because the servicer obtained proof of insurance before the second notice was sent). Of those receiving the second notice, about one-half actually had insurance force-placed. The drop in the share of borrowers who were sent the second notice relative to the first was more pronounced in 2015 than in 2012, which is consistent with the notices used in 2015 being more effective but could also reflect other changes in the market.

For borrowers who evidenced some difficulty making payments during the year (the right panel), the drop off following each notice is less pronounced—about half of borrowers who were sent the first notice also were sent the second notice, and of those the large majority had insurance force-placed. Although fewer first notices were sent in 2015 than in 2012, the drop off between each stage appears similar between the two years for these borrowers.

FIGURE 2: SHARE OF BORROWERS REACHING CERTAIN STAGES OF THE FORCE-PLACED INSURANCE PROCESS, ACCORDING TO DELINQUENCY STATUS DURING THE YEAR (SERVICING OPERATIONS DATA)

To better compare what happened after borrowers received the first notice in 2012 and 2015, Figure 3 repeats the analysis of Figure 2 as a share of all borrowers who received the first notice. The figure shows that, regardless of delinquency status, in 2015 a smaller fraction of borrowers who received a first notice also received the second notice and a smaller fraction of such borrowers had insurance force-placed. This is consistent with the notices being more effective in 2015, although it is also possible that other factors explain the change. For example, it may be that in 2015 changes in the property insurance market made it easier for borrowers to obtain a hazard policy.

FIGURE 3: SHARE OF BORROWERS REACHING CERTAIN STAGES OF THE FORCE-PLACED INSURANCE PROCESS, ACCORDING TO DELINQUENCY STATUS DURING THE YEAR (SERVICING OPERATIONS DATA)

Figure 4 splits borrowers by whether the borrower's insurance is paid through an escrow account and again shows the share of borrowers who received force-placed insurance notices and the share the had insurance force-placed. For borrowers whose insurance is paid through escrow, hazard insurance should not be cancelled by the insurer for non-payment, because the Rule requires the servicer to forward funds to the borrower's escrow account if necessary to make premium payments. Figure 4 shows that notices were sent to a higher fraction of borrowers for whom insurance was not paid through escrow; however, the fraction of borrowers who ultimately have insurance force-placed is similar for both groups.393

FIGURE 4: SHARE OF BORROWERS REACHING CERTAIN STAGES OF THE FORCE-PLACED INSURANCE PROCESS, ACCORDING TO WHETHER HAZARD INSURANCE PREMIA ARE PAID THROUGH ESCROW ACCOUNT (SERVICING OPERATIONS DATA)

Figure 5 repeats the analysis of Figure 4 showing shares relative to the number of borrowers who were sent the first notice. Relative to 2012, borrowers in 2015 who received the first notice were less likely to receive the second notice and less likely to have insurance force-placed, whether or not their insurance was paid through an escrow account.

FIGURE 5: SHARE OF BORROWERS REACHING CERTAIN STAGES OF THE FORCE-PLACED INSURANCE PROCESS, ACCORDING TO WHETHER HAZARD INSURANCE PREMIA ARE PAID THROUGH ESCROW ACCOUNT (SERVICING OPERATIONS DATA)

11.3.2 Timing of notices

The Rule requires servicers to mail borrowers a force-placed insurance notice at least 45 days before charging them for that insurance, to give borrowers an opportunity to avoid being charged for force-placed insurance. To look at the timing of force-placed insurance notices, the Bureau used data from four servicers that were able to provide data on both the dates notices were sent and the dates borrowers were charged for force-placed insurance in 2012 and 2015. Figure 6 plots the distribution of the time from the first notice date to the date borrowers are charged for force-placed insurance by servicer and year using a box-and-whiskers plot in which the box represents the range from the 25th to 75th percentile and the heavy line represents the median. The dashed line represents 45 days from the date the first notice was sent, which is the earliest point at which the Rule permits servicers to begin charging for force-placed insurance. The typical timeline varies widely across the four servicers. For each servicer, the median number of days between the first notice and the charge date was greater than 45 days both in 2012 and 2015. There is no clear pattern of changes between 2012 and 2015. Servicers B and C, which had the widest variation in dates in 2012, both reduced this variation by 2015. Similarly, servicers B and D, the only servicers that charged borrowers within 45 days at least 25 percent of the time in 2012, both reduced the share of cases in which the notice was provided fewer than 45 days before the borrower was charged. The other two servicers provided the notice at least 45 days in advance in more than 75 percent of cases both before and after the Rule.

FIGURE 6: DAYS BETWEEN FIRST FORCE-PLACED INSURANCE NOTICE AND DAY BORROWER IS CHARGED FOR FORCE-PLACED INSURANCE (SERVICING OPERATIONS DATA)

Force-placed insurance is generally effective retroactively, going into effect when the borrower's previous insurance expired. Figure 7 shows box plots by servicer for the time between the effective date of force-placed insurance (i.e., the expiration date of the prior hazard insurance policy) and the date of the first notice, for five servicers that were able to provide data on both dates. Figure 7 shows relatively little change in the timing of servicers sending the first forceplaced insurance notices relative to the hazard insurance expiration date; if anything, Figure 7 shows that some servicers provided these notices a bit later in 2015. Four of five servicers generally began sending the first force-placed insurance notices shortly after the date the hazard policy expired, with the remaining one servicer sending the first notice well before policy expiration.

FIGURE 7: DAYS BETWEEN EFFECTIVE DATE OF FORCE-PLACED INSURANCE AND FIRST FORCEPLACED INSURANCE NOTICE (SERVICING OPERATIONS DATA)

11.3.3 Geographic distribution of force-placed insurance

As shown in Figure 2 above, around 1 percent of borrowers who did not experience delinquency or seek loss mitigation were newly charged for force-placed insurance each year, and most had not canceled the force-placed insurance six months later. Because these borrowers were making their mortgage payments, a failure to provide proof of hazard insurance may not be driven by an inability to pay premiums. To better understand the effectiveness of the force-placed insurance provisions in providing borrowers with relevant information, it would be helpful to know whether these borrowers made an informed choice not to obtain their own hazard insurance.

The Bureau understands from discussions with servicers that in some cases borrowers do not obtain their own hazard insurance because they are unable to obtain hazard insurance or because they find force-placed insurance to be a less expensive option. This could reflect individual risk characteristics or claim history, or it could reflect the risk characteristics associated with the location of the property. This section looks at geographic variation in forceplaced insurance rates.

Figures 8 and 9 show the rate of force-placed insurance by state on January 1, 2012 and 2015, respectively, using the Servicing Operations Data. To focus on borrowers who did not experience payment difficulties, these figures are restricted to borrowers who were not 60 days delinquent and did not apply for loss mitigation in the applicable year. The figures generally show greater rates of force-placed insurance in the southeastern states, Texas, and Puerto Rico, areas with greater than average exposure to natural disasters. This suggests that some degree of force placement could be explained by the greater difficulty some borrowers face in obtaining affordable hazard insurance.

FIGURE 8: 2012 SHARE OF LOANS WITH NO DELINQUENCY OR LOSS MITIGATION APPLICATION DURING THE YEAR THAT HAVE FORCE PLACED INSURANCE ON JANUARY 1ST (SERVICING OPERATIONS DATA)

FIGURE 9: 2015 SHARE OF LOANS WITH NO DELINQUENCY OR LOSS MITIGATION APPLICATION DURING THE YEAR THAT HAVE FORCE PLACED INSURANCE ON JANUARY 1ST (SERVICING OPERATIONS DATA)

APPENDIX A: The 2013 Respa Servicing Rule and Bureau Purposes and Objectives

Introduction

As discussed in Section 1 of this report, section 1022(d) of the Dodd-Frank Act requires the Bureau to conduct an assessment of each significant rule or order adopted by the Bureau under Federal consumer financial law. Section 1022(d) requires that the assessment address, among other relevant factors, the rule's effectiveness in meeting the specific goals stated by the Bureau, as well as the Bureau's purposes and objectives specified in section 1021 of title X of the DoddFrank Act. Whereas the body of the report addresses the specific goals stated by the Bureau, this appendix highlights certain core findings in the body of the report with respect to the DoddFrank Act's purposes and objectives.

Purposes and objectives

Purposes

Under section 1021(a) of the Dodd-Frank Act, "[t]he Bureau shall seek to implement and, where applicable, enforce federal consumer financial law consistently for the purpose of ensuring that all consumers have access to markets for consumer financial products and services and that markets for consumer financial products and services are fair, transparent, and competitive."

All consumers have access to markets for consumer financial products and services.

Consumers generally do not choose their servicers. The servicing of loans is arranged by the loan's owner, and the Bureau would not expect that the Rule would affect directly consumers' "access" to mortgage servicing. However, lenders must factor the cost of servicing into the expected profitability of a mortgage loan when deciding whether to extend credit. An increase in the cost of servicing loans reduces the expected profitability of each loan, which may cause lenders to tighten lending standards or increase borrowing costs, either of which can reduce access to mortgage credit. The evidence suggests that the Rule increased the cost of servicing mortgage loans, but that this increase was not especially large relative to the overall cost of servicing. The estimated annual increase in servicing costs was less than $10 per loan for most servicers interviewed, whereas survey evidence suggests that average annual servicing costs have ranged from $250 to $300 per loan since 2012. Given that the cost of servicing is itself a relatively small part of the overall cost of mortgage lending, the effect of the Rule on access to mortgage credit is likely small. The overall effects of the Rule on servicing costs are discussed in sections 5.3 and 5.4, and costs of complying with particular Rule provisions are discussed in sections of Chapters 6 through 11.

A stated goal of the Rule was to facilitate servicer review for loss mitigation options, which could have the effect of increasing consumer access to loan modifications and other types of loss mitigation. Overall, the data indicate an increase in the fraction of delinquent borrowers who initiate applications for loss mitigation, from 39 percent of delinquent borrowers to 43 percent of delinquent borrowers, although the Bureau cannot determine whether this increase is attributable to the Rule. The data do not indicate that borrowers are more likely to complete applications that they initiate, although again this may be affected by other market factors. The direct and indirect effects of particular Rule provisions on the review of consumers for loss mitigation options is discussed in Chapters 6 through 9 and section 10.3.4.

Markets for consumer financial products and services are fair, transparent, and competitive.

Fairness

Many provisions of the Rule may be viewed as improving fairness, including procedural requirements for evaluating loss mitigation applications and error resolution provisions. The Rule's loss mitigation procedures and restrictions on foreclosure are designed to ensure a fair process for borrowers to apply and be evaluated for the loss mitigation options made available by the owners of their mortgage loan. Evidence from servicer interviews suggests that, as a result of the Rule, borrowers are treated more consistently post-Rule when applying for loss mitigation options. The Rule led some servicers to make changes to how they evaluate borrowers for loss mitigation, notably in evaluating borrowers for all available loss mitigation options (rather than only those that the borrower or servicer believed would be most appropriate) and increasing the share of borrowers who received a decision on their loss mitigation application within 30 days. Both of these changes suggest that borrowers are more likely to be treated consistently during the evaluation process. The evidence indicates that more borrowers are taking advantage of the opportunity to appeal loss mitigation denials than before the Rule was in effect, which may reflect greater procedural fairness in the evaluation process. Chapter 7 examines evidence of the effectiveness of the provisions related to completing loss mitigation applications, and Chapter 8 examines evidence regarding the evaluation of loss mitigation applications and appeals of denials for loss mitigation options.

The Rule's error resolution provisions provide timelines under which servicers must acknowledge, investigate, and resolve certain borrower assertions of error and requests for information. The evidence does not suggest that consumers are more likely to assert errors following the Rule's effective date, but there is some evidence that borrowers are less likely to submit repeated error assertions, suggesting that the Rule may have made servicers more responsive to borrower assertions of errors. Chapter 10 examines evidence on consumer utilization of error resolution provisions and effects on servicer responsiveness.

Transparency

With respect to transparency, the Rule requires servicers to provide several disclosures related to delinquency and the loss mitigation process as well as disclosures related to error resolution and force-placed insurance.

The Rule's early intervention disclosures establish requirements for servicers to contact delinquent borrowers early in the delinquency to inform them about loss mitigation options and provide other delinquency-related information. In interviews, servicers generally said that the early intervention requirements were similar to their prior practices. Consistent with this, the data show little change in the timing of written notification to delinquent borrowers between the pre-Rule and post-Rule periods. Chapter 6 describes the Bureau's analysis of these provisions.

For borrowers applying for loss mitigation, the Rule's continuity of contact provisions are intended to help ensure that certain delinquent borrowers have access to servicer personnel who can help them obtain information they need to complete their applications and understand their applications' status. The Rule also addresses written notices that servicers must send to borrowers at certain points during the loss mitigation application process that detail information about the application requirements, the status of the application, the outcome of the evaluation of the application, and the appeals process. The housing counselors and legal aid attorneys the Bureau surveyed said that these provisions helped borrowers obtain information about loss mitigation applications, suggesting that these provisions may have made the loss mitigation process more transparent for borrowers. On the other hand, some servicers interviewed said that they did not believe these provisions led them to provide more information to loss mitigation applicants than they did before the Rule's effective date. The Bureau's findings regarding these provisions are discussed in Chapter 7.

The Rule's error resolution disclosures relate to transparency about how borrower error assertions are being handled and the results of the servicer's investigation, which can increase transparency with respect to a borrower's mortgage. The evidence suggests that borrowers submitted fewer follow-up or repeat error assertions after the Rule, which is consistent with borrowers receiving clearer information from servicers about their investigations into asserted errors. The error resolution provision of the Rule is discussed in Chapter 10.

The Rule requires servicers to send written notices sent to borrowers prior to charging a borrower for force-placed insurance. Servicers interviewed said that these notices were generally consistent with notices provided before the Rule became effective. The data show a moderate decrease in the share of borrowers receiving force-placed insurance, which may be consistent with the disclosures required by the Rule making borrowers more aware of the need to provide proof of hazard insurance coverage so that they can avoid force-placed insurance.

The Bureau's findings regarding these notices are included in Chapter 11.

Competitiveness

With respect to competitiveness, consumers generally do not choose their servicers and therefore mortgage servicing is not a market in which servicers compete for consumers. Rather, mortgage servicers compete with each other to offer and provide their services to the owners of mortgage loans. If the Rule increased costs for some servicers more than for others, then this could have the effect of making these servicers less competitive relative to other servicers in offering and providing their services to the owners of mortgage loans. Small servicers are exempt from some provisions of the Rule; however the evidence does not suggest that after the Rule's effective date the cost of servicing has increased more for servicers that are not small servicers than it has for small servicers. The effects of the Rule on costs are discussed in Chapter 5.

Objectives

The objectives of the Bureau are listed in section 1021(b) of the Dodd-Frank Act.

Consumers are provided with timely and understandable information to make responsible decisions about financial transactions.

As discussed above with respect to transparency, the Rule requires certain disclosures related to the servicing of the mortgage account. The data suggest that timelines for providing many mortgage servicing disclosures required by the Rule generally have not decreased relative to timelines for providing similar disclosures in the pre-Rule period, although there is evidence that some servicers provided written acknowledgments of error assertions more quickly than before the Rule's effective date. The Bureau's analysis of how the Rule affected the timeliness of these disclosures is included in sections 6.3.2, 7.3.2, 10.3.2, and 11.3.2.

Prior to issuing the Rule, the Bureau conducted consumer testing to assess whether certain draft model disclosure forms were comprehensible to consumers and found that the forms were comprehensible. The Bureau did not conduct additional consumer testing for this assessment, but did receive information from its survey of housing counselors and legal aid attorneys and from servicer interviews about whether disclosures were understandable; these findings are discussed in sections 6.2.2, 7.2.2, and 11.2.2 and in Appendix E.

Consumers are protected from unfair, deceptive, or abusive acts and practices and from discrimination.

The specific goals of the Rule, which are noted in Chapter 1, do not explicitly include protecting consumers from unfair, deceptive, or abusive acts and practices or from discrimination. Although many of the protections in the Rule might prevent or deter such acts or practices or discrimination, the information and data the Bureau obtained and generated in conducting this assessment do not provide a basis for the Bureau to offer views as to any meaningful effect the Rule may have had on this general Bureau objective.

Outdated, unnecessary, or unduly burdensome regulations are regularly identified and addressed in order to reduce unwarranted regulatory burdens.

The Rule implemented Dodd-Frank Act amendments to RESPA and established a number of new mortgage servicing requirements. The specific goals of the rulemaking, as outlined in Chapter 1, did not seek to identify or address any pre-existing regulations that were outdated, unnecessary, or unduly burdensome.

However, in developing the Rule, the Bureau took steps to mitigate some of the burden it imposed in the Rule. More specifically, as described in the summary of the provisions and the evolution of the Rule in Chapters 1 and 2, the Bureau sought to mitigate regulatory burden in developing the Rule by, among other things, (i) exempting small servicers from certain of the Rule's requirements, as described in Chapter 2, (ii) adopting requirements for servicing policies and procedures that permit servicers flexibility in how to implement the requirements, (iii) limiting the scope of the Rule's error resolution provision to apply to only written, rather than oral, assertions of error and specifying that a QWR that asserts an error relating to the servicing of a mortgage loan is a notice of error for purposes of the error resolution provision, avoiding overlapping requirements, (iv) amending the Rule to clarify certain aspects of the Rule and to provide more flexibility for servicers with certain requirements related to loss mitigation applications, and (v) providing compliance guidance for implementing certain provisions in the Rule, including those related to successors-in-interest, early intervention, and concerning servicers' obligations to provide certain notices/communications to borrowers who have exercised their right under the Fair Debt Collection Practices Act barring debt collectors from communicating with them. Overall, although the Bureau undertook the measures above to mitigate regulatory burden, the Bureau did not obtain or generate data in this assessment

that would allow it to estimate the decreased burden associated with those measures individually or collectively.

Note also that the Bureau examined certain elements of the Rule about which stakeholders raised questions about the burden imposed on industry relative to the benefit to consumers. Section 5.4 discusses the effects of the Rule on mid-size servicers, which do not qualify for the small servicer exemption. Sections 7.3.2 and 8.3.2 analyze the effect of the Rule on how long it takes borrowers to complete loss mitigation applications and to receive short-sale offers, respectively. Section 9.3 considers the effect of the Rule on the time it takes servicers to complete a foreclosure. In addition, Sections 6.2.3, 8.2.3, 9.2.3, 10.2.3, and 11.2.3 describe findings from servicer interviews about the burdens of the Rule that apply in particular circumstances, such as when borrowers have already exhausted all loss mitigation options or when properties are vacant or abandoned.

Federal consumer financial law is enforced consistently, without regard to the status of a person as a depository institution, in order to promote fair competition.

The specific goals of the Rule, which are noted in Section 1, do not explicitly include whether Federal consumer financial law is enforced consistently without regard to status as a depository or non-depository institution.

The Bureau has enforcement authority with respect to non-depository mortgage servicers and depositories with assets over $10 billion and the prudential regulators have enforcement authority with respect to smaller depositories. As section 3.6 reports, to date, the Bureau has brought five enforcement actions against mortgage servicers for violating the Rule.

The Bureau has supervisory authority with respect to depositories with assets over $10 billion and non-depository mortgage servicers. As discussed in section 3.3.2, the Bureau has conducted examinations among large depositories and non-depository mortgage servicers.

Markets for consumer financial products and services operate transparently and efficiently to facilitate access and innovation.

Potential effects of the Rule on transparency and access are discussed above. However, the Bureau does not have sufficient evidence to conclude whether the mortgage servicing market is operating more or less transparently and efficiently because of the Rule. With respect to innovation, Chapter 3 discusses changes in the types of loss mitigation options that servicers offer, including the increasing use of streamlined modifications. These innovations in loss mitigation have taken place while the Rule has been in effect, suggesting that the Rule has not prevented such innovation, although the Bureau cannot rule out the possibility that greater innovation of this type would have taken place absent the Rule. In addition, as discussed in section 5.3.1, some servicers said that the time and effort needed to implement the Rule meant that fewer resources were available to make innovations in how they service loans, so that in this sense the Rule would have slowed innovation.

APPENDIX B: Comment Summaries

On May 11, 2017, the Bureau published a request for information (RFI) on the 2013 RESPA Servicing Rule assessment and invited the public to submit comments and information on a variety of topics. The public comment period closed on July 10, 2017. The Bureau received approximately 35 comments in response to the RFI. The Bureau summarizes the comments and information received on certain topics below and the full comments are available on www.regulations.gov.

Generally, commenters reported on their own experiences, and provided information from surveys and other types of research, regarding the overall effects of the Rule and the effects of particular requirements that are within the scope of the assessment report. This information is summarized here and incorporated into other parts of the report as appropriate. See Chapter 1, "Sources of information and data," for a summary of the data and information used in the assessment. This appendix also contains a summary of recommendations for modifying, expanding, or eliminating the Rule.406 Finally, section IV of the RFI described the assessment

plan, and the Bureau also invited comments on the plan. These comments are summarized below. The Bureau continued to develop the assessment plan after publishing the RFI, taking into account the comments received.

While most comments addressed the major topics the Bureau identified for public comment, some comments related to topics beyond those set forth in the notice. For example, some commenters addressed mortgage servicing rule amendments that the Bureau issued after the Rule's effective date. The Bureau has not considered such non-responsive comments in this assessment, but may consider such comments in the context of future policy development.

Evidence about 2013 RESPA Servicing Rule effects

Regarding the overall effect of the Rule, a consumer advocacy group reported results of a survey it conducted in June 2017 of consumer advocates, which found that 85 percent of respondents believed the Rule had benefited homeowners and 86 percent believed it had helped more homeowners avoid foreclosure. A trade association said that, based on a survey of customer satisfaction with mortgage servicers, industry's increased investment in default servicing "had not been commensurate with better service for consumers." A number of trade associations and one mortgage servicer cited an industry survey indicating large increases in the cost of mortgage servicing since 2008. According to this survey, the per-loan cost of servicing performing loans increased from $59 per year in 2008 to $163 per year in 2016, and the per-loan cost of servicing delinquent loans increased from $482 per year to more than $2,113 per year over the same period.

Regarding specific implementation challenges, one servicer said that when the Rule became effective many third-party servicing system providers were not able to accommodate the tracking and reporting requirements of the Rule, requiring servicers to create new user fields and monitor manual reports to maintain compliance.

A number of commenters provided evidence of several market changes since 2008 related to an increase in regulation of mortgages or mortgage servicing. A trade association cited a survey of credit unions that found that 44 percent of respondents that offered mortgages sometime during the previous five years have either eliminated certain mortgage products and services (33 percent) or stopped offering them (11 percent), primarily due to burden from the Bureau's regulations. That survey also found that credit unions with assets of less than $100 million are the group most apt to have dropped their mortgage program altogether. Another trade association cited findings from Inside Mortgage Finance that as of 2016 five of the top ten servicers were nonbanks, whereas in 2010 only one of the top ten was a nonbank. A servicer said that the share of mortgage loans serviced by banks had fallen from 89 percent in 2008 to 68 percent in the third quarter of 2016.

Some commenters provided evidence about changes to consumer access to mortgage credit. One trade association said that servicers "were compelled to adopt conservative underwriting and servicing practices to offset the rising costs and manage the demands associated with the unpredictable nature of servicing" and that "in response to a more stringent regulatory and enforcement regime, lenders have introduced credit overlays that restrict access to credit." One servicer cited an Urban Institute report indicating that the mean FICO score at origination was 27 points higher in 2017 than it was in 2007.

Some commenters said that certain disclosure requirements in the Rule caused unintended consequences for borrowers. Regarding content, some trade associations that represent mortgage servicers said that their members receive complaints about the force-placed insurance notices required by the Rule, describing consumers as being "surprised and angered by the tone" or finding the notices "confusing" or "off-putting." Regarding frequency, one servicer said that borrowers complain that attempts at early intervention live contact required by the Rule are excessive. Regarding timing requirements, one servicer and one trade association said that borrowers sometimes submit overly broad notices of error or information requests as a tactic to attempt to delay foreclosure. However, another commenter said that the Rule's provision that servicers need not respond to such notices within seven days of a foreclosure sale date "has proven to be effective in preventing frivolous requests that are made just prior to a scheduled foreclosure sale."

Some commenters noted challenges associated with the Rule's loss mitigation provisions. Some commenters noted challenges with mandated timelines. Two commenters said that servicers are unable to adequately assess an application's status in time to provide the acknowledgment notice within five days and that this necessitates additional follow up between the borrower and servicer to complete the application. One commenter described state or local requirements related to loss mitigation that the commenter described as conflicting with some timelines in Bureau rules, which the commenter said can cause borrower confusion. A number of commenters noted challenges with the dual-tracking provisions. One trade association representing law firms said that servicers sometimes place a "CFPB hold" on foreclosure proceedings if there is any loss mitigation activity, regardless of whether there is a complete application. Another trade association said that it takes time for law firms representing servicers to stop the foreclosure process when required to comply with the Rule. A servicer and a trade association said that courts may be reluctant to postpone foreclosure sales postjudgment and that some jurisdictions require by statute or court rule that foreclosure sales occur within a certain number of days of the foreclosure judgment.

Recommendations to modify, expand, or eliminate the 2013 RESPA Servicing Rule

This section first summarizes general recommendations for modifying, expanding, or eliminating the Rule, and then discusses recommendations regarding specific Rule provisions.

General recommendations about the Rule

Several industry commenters suggested modifications to the Rule to reduce the cost of complying with the Rule. A number of industry commenters stated that certain Rule provisions duplicate or conflict with state law requirements and urged the Bureau to amend the Rule to preempt more restrictive state laws and regulations. An industry commenter asserted the application of the Rule was unclear in certain circumstances. This commenter suggested that the Bureau issue additional written guidance to reduce uncertainty. According to another industry commenter, the Rule's requirements are overly prescriptive and this has the effect of limiting benefits, diverting resources away from consumer-friendly innovation, and restricting the amount of credit available to borrowers. This commenter urged the Bureau to amend the Rule to provide servicers more discretion when complying with the Rule.

Some industry commenters stated that some servicers, particularly mid-size servicers, have exited the mortgage servicing market as a result of the Rule. To reduce the effect of the Rule on mid-size servicers and small servicers seeking to grow, several industry commenters and a government agency suggested that the Bureau expand the small servicer exemption to cover more servicers, including those operating in local or regional markets.

Several industry commenters provided general feedback with respect to required disclosures, both in general and with respect to specific disclosures. Some industry commenters stated that the formatting and content requirements for required disclosures are too prescriptive and undermine borrower communication. Accordingly, these commenters urged the Bureau to allow servicers more flexibility to tailor required disclosures. Several industry commenters suggested that the Bureau expand the safe harbor provided by the use of model forms to allow servicers to make non-substantive changes. An industry commenter suggested that the Bureau engage in consumer testing for all model disclosures to help ensure consumer understanding and usability. This commenter suggested that disclosure requirements should take into account modern communication technologies.

Early Intervention Requirements for Certain Borrowers (§ 1024.39)

Several industry commenters recommended modifications to the Rule's early intervention requirements. A number of industry commenters stated that early intervention outreach may confuse or annoy borrowers who are serially delinquent (i.e., routinely fail to remit timely payments, but consistently cure the delinquencies) or who have loans with late-stage delinquencies and no loss mitigation options. Accordingly, some industry commenters suggested that the Rule should require servicers to provide written early intervention notices once per delinquency cycle. A number of industry commenters also suggested that the early intervention requirements should not apply to late-stage delinquency loans or loans subject to foreclosure proceedings. A number of industry commenters stated that the live contact requirements may conflict with the Telephone Consumer Protection Act and requested additional guidance.

One consumer advocacy group suggested expanding the scope of loans subject to early intervention requirements, including loans serviced by reverse mortgage servicers. A consumer advocacy group also stated that servicers should be required to provide new early intervention notices when they resume servicing a charged-off loan where account activity had been suspended.

Continuity of Contact (§ 1024.40)

One industry commenter supported the Rule's reasonable policies and procedures requirement, as opposed to more prescriptive requirements, noting that the approach allows servicers flexibility to comply with the Rule in a manner consistent with individual company practices. However, the industry commenter urged the Bureau to exempt servicers from continuity of contact requirements for loans where borrowers have filed for bankruptcy. Noting the complexities of bankruptcy law and the legal risk associated with automatic stay violations, the commenter stated that reassigning borrowers in bankruptcy to personnel with specialized bankruptcy expertise would be beneficial.

Loss Mitigation Procedures (§ 1024.41)

Consumer advocacy groups generally supported the Rule's loss mitigation provisions, which they stated increase transparency, streamline loss mitigation procedures, promote timely decisions by servicers with respect to loss mitigation applications, and prevent avoidable foreclosures. One consumer advocacy group cited, among other things, a national survey of consumer advocates conducted by the National Consumer Law Center suggesting that the Rule helped borrowers to access loss mitigation options and prevent avoidable foreclosure. A consumer advocacy group stated that requiring a servicer to provide specific reasons for denying a borrower loan modification options helps the borrower to understand the basis of the determination and appeal denials when appropriate. The consumer advocacy group further stated that borrowers benefit from the requirement that a servicer assess a borrower for all available loss mitigation options, as well as from the 14-day timeframe for a borrower to accept, reject, or appeal a loan modification option or determination.

Several consumer advocacy groups suggested modifications to the loss mitigation requirements. Among other things, consumer advocacy groups recommended that the Rule: prohibit servicers from requesting duplicative information from borrowers in response to loss mitigation applications; extend foreclosure protections to borrowers with incomplete loss mitigation applications or pending notices of error relating to foreclosure; apply the 120-day foreclosure waiting period to vacant properties; and require servicers to inform borrowers when a loss mitigation application is deemed duplicative. A number of consumer advocacy groups suggested that the Bureau require servicers to accept duplicative requests, either in the event of a borrower's changed circumstances, or if a sufficient amount of time has passed since the prior request. Consumer advocacy groups also urged the Bureau to address additional communication concerns between borrowers and servicers about loss mitigation applications in certain circumstances. Several consumer advocacy groups stated that servicers should be required to offer affordable loan modification options to borrowers facing hardship.

Industry commenters commented on the standard for loss mitigation application review. An industry commenter stated that the Rule gives servicers no choice but to delay a foreclosure even in response to an incomplete loss mitigation application and recommended that borrowers should be required to provide material information, such as pay stubs or tax returns, before a servicer becomes subject to dual tracking provisions. Another industry commenter suggested that the Bureau create standardized loss mitigation definitions and procedures to provide clearer rules for servicers when evaluating loss mitigation applications. The industry commenter further requested that the Bureau create a uniform loss mitigation application. Another industry commenter recommended that the Bureau permit servicers to provide a streamlined loss mitigation application process for borrowers who express a preference for certain loss mitigation solutions. Noting that in its experience most borrowers at risk of foreclosure want to retain their property, an industry commenter recommended that the Bureau reconsider whether servicers should be required to offer home retention options at the same time as liquidation options.

Several industry commenters suggested that the Bureau extend the timeframes for servicers to provide certain loss mitigation notices, including the acknowledgement notice. According to industry commenters, the five-day timeframe for the acknowledgement notice does not provide sufficient time for servicers to determine whether the loss mitigation application is complete, particularly when a servicer must obtain additional information from a borrower or third party.

A number of industry commenters commented on loss mitigation application review timeframes and dual tracking restrictions when loss mitigation applications are pending. One industry commenter stated that the loss mitigation rules should be amended to prevent "abuse of the privilege," particularly with respect to multiple requests for loss mitigation by the same borrower. Another industry commenter suggested that dual-tracking restrictions should apply only when borrowers express a genuine interest in loss mitigation early in the process. Industry commenters stated that timeframes for reviewing loss mitigation applications when foreclosure sales are pending are too short. An industry commenter stated that the timeframe to review loss mitigation applications should be extended when an application is received less than 60 days prior to a foreclosure sale. Some industry commenters stated that a servicer in receipt of a complete loss mitigation application may need time to determine whether the application is complete and to coordinate with a third-party service provider, such as a foreclosure attorney, to suspend a foreclosure. These commenters suggested that the Rule should provide a grace period between the receipt of a complete application and the point at which the Rule's dual-tracking restrictions apply, with one commenter recommending a five-day period.

Several industry commenters addressed specific foreclosure-related provisions. For example, the Rule generally prohibits a servicer from initiating foreclosure if a loan is not more than 120 days delinquent; however, this provision does not apply if the property is not a borrower's principal residence. An industry commenter and a government agency stated that it is costly for servicers to prove that a vacant or abandoned property is not a borrower's principal residence and suggested that the Bureau streamline the foreclosure process for such properties. An industry commenter said that the prohibition on initiating foreclosure if a loan is not more than 120 days delinquent is unnecessary when a consumer, such as a successor in interest, intends to surrender the property; another industry commenter urged the Bureau to allow servicers to foreclose on a loan when a borrower remains continuously 90 or more days delinquent and does not qualify for, or rejects, loss mitigation options. An industry commenter asserted that restrictions on foreclosure referrals and sales pursuant to 12 CFR 1024.41(f) and (g) should not apply to vacant or abandoned properties because these provisions delay resolution of a borrower's delinquency and depress neighboring home values.

Servicing transfers (General Servicing Policies, Procedures, and Requirements, § 1024.38)

The Bureau received several comments related the general servicing policies, procedures, and requirements provisions related to transfers of servicing. One consumer advocacy group stated that the requirements related to the transfer of information during servicing transfers are ineffective for loans with pending loss mitigation applications. The commenter suggested that the Rule should impose specific requirements on servicers in lieu of the current general policies and procedures requirements. Consumer advocacy groups urged the Bureau to require transferor servicers to transfer all pending loss mitigation requests and documents when transferring loans and to notify borrowers about the status of pending loss mitigation applications and disputes. These commenters also suggested that the Rule should require transferee servicers to convert trial modifications and honor permanent modifications. Another industry commenter recommended that servicers to which loans are transferred should not be strictly liable for servicing errors made by the transferor servicer prior to the transfer, arguing that strict liability makes servicers unwilling to accept servicing of some loans and may have the effect of "trapping" consumers with a failing servicer.

Error Resolution Procedures and Requests for Information Requirements (§§ 1024.35, 1024.36)

Industry commenters provided feedback with respect to the error resolution and information request provisions. Industry commenters stated that servicers should not be required to respond to overbroad or unduly burdensome notices of error or information requests, such as non-specific form letters. Several industry commenters commented on the timing requirements for the error resolution and information request provisions. Some industry commenters suggested extending the timeframes for servicers to acknowledge and respond to notice of errors and information requests, and one suggested that the Bureau extend the period during which a servicer can avoid sending an acknowledgment notice by providing the information requested by the borrower. An industry commenter stated that some communications are ambiguous and recommended that the Rule require servicers to make only reasonable efforts toward identifying whether a communication constitutes a notice of error or request for information, and that the timeline for acknowledging and responding should be triggered only by receipt of a communication which is reasonably understood to be such a notice or request. Another industry commenter stated that responding to a notice of error or information request sometimes requires the involvement of a third party, such as a court or county recorder's office, which the servicer cannot control. Accordingly, the commenter stated that the Rule's timing requirements should toll while a servicer's response is delayed by a third party. An industry commenter requested additional guidance to clarify when a notice of error or information request provided to a third party, such as a vendor or service provider, triggers a servicer's obligation to acknowledge and respond to that communication.

Consumer advocacy groups generally supported the error resolution and information request provisions. For example, a consumer advocacy group stated that the provisions allow borrowers to resolve disputes without resorting to litigation, which benefits servicers and consumers. Some consumer advocacy groups suggested modifying the error resolution and information request provisions. For example, several consumer advocacy groups recommended that the Rule require servicers to stay foreclosure proceedings when a servicer receives a notice of error relating to the foreclosure. These consumer advocacy groups proposed expanding the types of borrower communications that trigger the error resolution and information request provisions. For example, these consumer advocacy groups suggested that notices of error and information requests provided by a borrower's agent, as well as such communications provided to an address other than a designated address pursuant to 12 CFR 1024.35(c) and 1024.36(b), should trigger a servicer's obligation under the provisions.

Force-placed Insurance (§ 1024.37)

Consumer advocacy groups generally supported the Rule's force-placed insurance requirements, stating that they benefit the mortgage servicing market by increasing transparency and reducing consumer harm. For example, one consumer advocacy group cited a national survey of consumer advocates conducted by the National Consumer Law Center suggesting that fewer borrowers experienced force-placed insurances abuses as a result of the Rule.

While generally supportive of the Rule, several consumer advocacy groups proposed modifying the force-placed insurance requirements. Some consumer advocacy groups stated that servicers should be required to advance hazard insurance payments for all borrowers, including borrowers without escrow accounts. A consumer advocacy group urged the Bureau to prohibit servicers from accepting payments, including reinsurance deals or discounted administrative services, from force-placed insurance providers. This consumer advocacy group suggested that the Bureau should limit the amount of force-placed insurance that servicers may purchase on behalf of a borrower and prohibit servicers from retroactively charging borrowers for more than 60 days of force-placed insurance coverage. Another consumer advocacy group urged the Bureau to cover hazard insurance required by the Flood Disaster Protection Act of 1973 under the Rule's force-placed insurance requirements.

Industry commenters generally stated that the Bureau should modify the force-placed insurance requirements to provide more flexibility to servicers. An organization that provides insurance tracking services urged the Bureau to create a safe harbor for servicers to purchase force-placed insurance for borrowers with escrow accounts whose hazard insurance has been cancelled or not renewed for nonpayment of premium charges where a servicer made a good faith effort to restore a borrower's lapsed hazard insurance policy. Several industry commenters commented on force-placed insurance disclosures. An industry commenter stated that the required disclosures are too prescriptive and prevent servicers from including additional information, such as the availability of lender-placed insurance, which may be more affordable for consumers. Industry commenters stated that the language on force-placed insurance model disclosures is too harsh and should be modified to make them more "consumer-friendly." With respect to reminder notices required by 12 CFR 1024.37(d), an industry commenter stated that the Rule should not require a servicer to estimate the cost of the force-placed insurance, if the servicer does not know the cost. According to the commenter, estimating the cost of forceplaced insurance imposes a significant burden on servicers and estimated costs may confuse borrowers, if the actual cost is different.

The assessment plan

Comments on the scope of the assessment

Several industry commenters expressed concern that Bureau's plan as described in the RFI would focus too narrowly on delinquent borrowers and urged the Bureau to also focus attention on the costs to industry of the Rule and related market effects, emphasizing that the Dodd-Frank Act objectives require the Bureau to assess the regulatory burden of the Rule, whether it promotes fair competition, and whether under the Rule markets operate transparently and efficiently to facilitate access and innovation. In this context, some industry representatives said the assessment should include analysis of how the Rule has affected access to credit for consumers. A trade association representing credit unions said the Bureau should focus in particular on costs and benefits of the Rule for credit unions and on small servicers. Consumer advocacy groups and legal aid organizations supported a focus on delinquent borrowers, but also said the assessment should address whether and how the Rule had increased costs for servicers.

Some trade associations said that the assessment should also encompass the 2013 TILA Servicing Final Rule, stating that implementation of that rule has been costly and that servicers treated the 2013 TILA Servicing Final Rule and the 2013 RESPA Servicing Rule as a combined set of new regulations. One servicer said it did not object to the determination that the 2013 TILA Servicing Final Rule was not significant, but said it might be appropriate for the Bureau to evaluate a rule's significance based on actual costs rather than original estimates of the costs. Some industry trade associations said that the Bureau should include the 2016 amendments to the 2013 RESPA Servicing Rule within the scope of the assessment and one government agency said that the Bureau should delay its assessment until those amendments are in effect. A trade association said that the assessment should also encompass the "trickle-down" effect of ATR/QM underwriting requirements on servicing.

Some commenters identified particular aspects of the Rule that they said the assessment should address. One trade association said the Bureau should assess whether the burdens of the error resolution and request for information provisions are outweighed by benefits to borrowers and whether the early intervention provisions had reduced consumer confusion and improved engagement with loss mitigation. Another trade association said that the assessment should address whether the Rule improved the engagement of delinquent borrowers and how the Rule affected borrowers' ability to pursue their preferred loss mitigation options.

A trade association said that the Bureau should evaluate lessons learned from the adoption and implementation of the Rules, including why multiple amendments were needed, how oral guidance affected compliance, and how the Bureau could have encouraged third-party providers to timely deliver products necessary to comply. A law professor commented that the assessment should evaluate whether servicers are complying with the Rule and whether servicers that fail to comply are implementing remediation plans. Finally, one trade association said that the assessment should address the sufficiency of the Bureau's initial burden analysis in order to inform future rulemakings.

Comments on sources of information

Some commenters, including several legal aid organizations, one large servicer, and a consumer advocacy organization, noted the limitations of existing data sources on loss mitigation and servicing, and said that collecting loan-level data would benefit the assessment. Legal aid and consumer advocacy organizations supported the Bureau's plan to survey housing counselors, and some recommended that the Bureau survey or interview borrowers as part of the assessment. A trade association said that the Bureau should conduct qualitative interviews with servicers to understand how the Rule affected business decisions around servicing, and a government agency also said the Bureau should conduct outreach to industry.

Some servicers and industry trade associations emphasized that any new data collected from servicers should be targeted at specific measures of the Rule's effectiveness. These commenters also said that data collected from servicers should be collected and restricted in certain ways, treated as confidential, and used only for purposes of the assessment.

An industry trade association said that the Bureau should seek information from court records to assess foreclosures that were put on hold because a borrower expressed an interest in loss mitigation. A law professor said the Bureau should review court filings alleging servicer abuses.

One servicer recommended that for future rulemakings the Bureau establish prospectively how it intends to measure rule effectiveness so that industry can capture relevant data in the course of business.

A consumer advocacy group expressed concern about the Bureau relying on certain publicly available information on servicing costs to assess the effect of the Rule on servicing costs, noting that such data is less informative if it does not distinguish among costs for small, medium, and large servicers or break down costs into components, and also noting that factors other than the Rule may increase costs.

Other comments on methodology

Several commenters, including industry and consumer advocate representatives, said that the Bureau should take into account that servicing outcomes before and after the Rule's effective date were affected not only by the Rule but also by other developments in the market, including economic factors, investor guidelines, and other state or federal rules.

Some industry commenters proposed metrics that the Bureau should use to measure the effectiveness of the Rule. One trade association said that the Bureau should focus on access to delinquency resolution rather than on outcomes in assessing the Rule because under the Rule loss mitigation outcomes are left to investors. One servicer proposed a list of metrics that the Bureau should use in assessing specific provisions of the Rule, such as changes in the timing of loss mitigation applications initiated and changes in the number of foreclosure referrals prior to day 120 of a delinquency. Another proposed certain metrics specific to the Rule's error resolution and request for information provisions, such as the number of letters from borrowers asserting an error and the number of valid errors identified. One trade association and one servicer suggested that the Bureau engage an outside expert to determine the theoretical baseline that would exist if the Rule were not in effect.

APPENDIX C: Data Sources

As discussed in Chapter 1, the Bureau considered a number of sources of information in conducting the assessment. Below, some of these data sources are described in more detail, including the source of information, limitations of the data, and summary statistics.

A. Data from government-sponsored enterprises

Both of the government-sponsored enterprises (GSEs), Fannie Mae and Freddie Mac, publish loan-level performance data on a portion of the mortgages that they purchase or guarantee.

Each GSE publishes and regularly updates two data files: one on origination loan characteristics, and one with monthly performance records for each loan, dating back to the loan's origination.

The GSEs publish similar information in both file types.

FANNIE MAE

Fannie Mae's data include a subset of Fannie Mae's fully amortizing, full documentation, singlefamily fixed-rate mortgages with a term of 30 years or less, including all such loans that Fannie Mae acquired between January 1, 2000 and December 31, 2017.

According to Fannie Mae, the following types of mortgages are excluded from the data:

  • Adjustable-rate mortgage loans, balloon mortgage loans, interest-only mortgage loans, mortgage loans with prepayment penalties, government-insured mortgage loans, Home Affordable Refinance Program (HARP) mortgage loans, Refi Plus™ mortgage loans, or non-standard mortgage loans.
  • Certain types of mortgage loans, such as mortgage loans with original LTVs greater than 97 percent, Alt-A, other mortgage loans with reduced documentation and/or streamlined processing, and programs or variances that are ineligible today.410
  • Mortgage loans originated prior to 1999, mortgage loans subject to long-term standby commitments, sold with lender recourse or subject to other third-party risk-sharing arrangements, or that were acquired by Fannie Mae on a negotiated bulk basis.

FREDDIE MAC

Freddie Mac's data include fully amortizing 15-, 20-, and 30-year full documentation fixed-rate mortgages that were originated from January 1, 1999 forward.

According to Freddie Mac, the following types of mortgages are excluded from the data:

  • Adjustable Rate Mortgages (ARMs), Initial Interest, balloons, and any mortgages with step rates;
  • Government-insured mortgages, including Federal Housing Administration/Veterans
  • Affairs (FHA/VA), Guaranteed Rural Housing (GRH), and HUD-Guaranteed section 184 Native American Mortgages;
  • Home Possible®/Home Possible Neighborhood Solution® Mortgages and other affordable mortgages (including lender-branded affordable loan products);
  • Mortgages delivered to Freddie Mac under alternate agreements;
  • Mortgages for which the documentation is not verified or not waived;
  • Mortgages associated with Mortgage Revenue Bonds purchased by Freddie Mac; and
  • Mortgages delivered to Freddie Mac with credit enhancements other than primary mortgage insurance, with the exception of certain lender-negotiated credit enhancements.

Differences and limitations

While the data each GSE publishes are comparable, there are a few differences between the datasets:

  • The Fannie Mae Data includes servicer name with the monthly performance records, allowing for comparisons between servicers. The Bureau used this as a way to compare loans between servicers that were subject to the NMS and those that were not.
  • In the Fannie Mae Data, the modification flag is populated with 'Y' in the month the modification occurs and continues to be populated as such in all subsequent months. Because of this, it is only possible to observe the first modification a loan receives and not whether a loan is modified more than once. In the Freddie Mac data, the modification flag is populated in the month that each modification occurs. As such, it is possible to observe if a loan is modified more than once.

Because the GSEs' data only include conventional fixed-rate mortgages, loan performance may not fully represent the conditions and activity present in the mortgage market as a whole. As of the end of 2017, mortgages sold to Fannie Mae and Freddie Mac comprised about 45 percent of the total mortgage debt outstanding on single family residences.411

Table 1 shows summary statistics for the subsamples of the GSE data that are most often used in the report, which are limited to mortgages for primary residences that have ever been at least 30 days delinquent.

TABLE 1: SUMMARY STATISTICS ON GSE DATA – LOANS ORIGINATED ON SINGLE-FAMILY PROPERTIES FOR PRIMARY RESIDENCES THAT HAVE EVER BEEN 30 DAYS DELINQUENT, WITH MONTHLY PERFORMANCE RECORDS THROUGH DECEMBER 2017

Metric

Fannie Mae

Freddie Mac

Number of total loans

3,430,050

2,347,619

Average credit score at origination

696

709

Share that ever become 60+ days delinquent

83%

45%

Share that complete foreclosure

11%

10%

Share that had a permanent modification

11%

12%

As of Jan 1, 2012

Number of active loans

1,641,838

1,151,896

Average outstanding balance

$139,928

$152,905

Average interest rate

5.47%

4.98%

Share of loans originated 2008 or later

26%

26%

Share of loans 90+ days delinquent

11%

13%

Share of loans 365+ days delinquent

6%

6%

As of Jan 1, 2015

Number of active loans

1,305,638

841,519

Average outstanding balance

$138,412

$147,872

Average interest rate

4.93%

4.34%

Share of loans originated 2008 or later

43%

38%

Share of loans 90+ days delinquent

7%

9%

Share of loans 365+ days delinquent

4%

4%

B. Data from Black Knight – McDash Analytics

Source of data

To support its market monitoring and research functions, the Bureau purchases a loan-level performance database from McDash, a subsidiary of Black Knight, Inc. As of February 2018, McDash had loan-level information on over 175 million mortgages and home equity loans. The proportions of GSE, FHA, and portfolio loans in the McDash Data are fairly similar to the comparable proportions in the aggregate market. Additionally, the Bureau subscribes to McDash's Loss Mitigation Module, which provides additional fields on loss mitigation activity, including permanent and trial modifications, repayment plans, short sales, and deeds-in-lieu. The Loss Mitigation Module is comprised of all loans serviced by a subset of the servicers that report to the broader McDash performance data. The earliest monthly records in the Loss Mitigation Module are for January 2008, even if a given loan is in the broader McDash database prior to January 2008.

Limitations

Only loans serviced by servicers that use the Black Knight MSP Servicing Platform are reflected in McDash Data. As such, there are loans for which the monthly performance records in McDash do not show the full history of the loan. In some cases, the first record in McDash is after the origination date or the last record reflected has a payment status that is not a terminal event;415 in other cases, the loan performance records skip months or the loan exits permanently due to a servicing transfer, presumably to a non-reporting servicer.

The Loss Mitigation Module covers about 60 percent of all loans in the broader McDash performance data and is primarily reported by larger depository institutions. In conducting analysis of certain loss mitigation trends for this report, the data are limited to loans in the loss mitigation module to accurately identify loss mitigation outcomes; these are less representative of the market as a whole. Short sales and deeds-in-lieu are identified only in the loss mitigation modules. The performance records in the broader McDash Data set may over count the number of foreclosures, as the payment status field uses the same code for short sales, deeds-in-lieu, and foreclosures.

Table 3 shows summary statistics for the subsample of the McDash data that is most often used in this report, which is limited to a 10 percent random sample of mortgages on primary residences that have ever been observed at least 30 days delinquent.

TABLE 2: SUMMARY STATISTICS ON MCDASH DATA - LOANS ORIGINATED ON SINGLE-FAMILY PROPERTIES SINCE 1999, WITH MONTHLY PERFORMANCE RECORDS THROUGH DECEMBER 2017 THAT HAVE EVER BEEN 30 DAYS DELINQUENT

Metric

McDash

Number of total loans

1,745,428

Average credit score at origination

664

Share by investor

Fannie

15%

Freddie

7%

Ginnie

10%

Private

26%

Portfolio

41%

Share that ever become 60+ days delinquent

82%

Share that ever have foreclosure initiated

33%

Share that complete foreclosure

15%

Share that had a permanent modification

20%

As of Jan 1, 2012

Number of active loans

780,948

Average outstanding balance

$179,589

Average interest rate

5.58%

Share of loans originated 2008 or later

26%

Share of loans 90+ days delinquent

23%

Share of loans 365+ days delinquent

13%

Share of borrowers in bankruptcy

4%

As of Jan 1, 2015

Number of active loans

553,287

Average outstanding balance

$162,844

Average interest rate

4.92%

Share of loans originated 2008 or later

39%

Share of loans 90+ days delinquent

16%

Share of loans 365+ days delinquent

9%

Share of borrowers in bankruptcy

4%

C. Servicing Operations Data

Source of data

The Bureau collected de-identified loan-level data from seven mortgage servicers using its authority under section 1022(c)(4) of the Dodd-Frank Act. The servicers were selected from among the largest 100 servicers to represent a range of servicer "types" based on servicing portfolio size, depository or non-depository status, subservicing activity, delinquency rates, and other characteristics. The Servicing Operations Data collection followed a few steps. First, the Bureau conducted interviews with the seven servicers to understand how operational activity is tracked in their data system. The Bureau developed a standardized data request based on these conversations that included data fields the Bureau expected the servicers could reasonably provide. After sharing the draft request with the seven servicers, the Bureau modified the request to each servicer so that the servicers were not required to provide data they reported would be overly burdensome to produce. All seven servicers provided data for most of the fields requested, but no servicer was able to provide data on all the requested fields. The servicers produced data over the course of several months and in some cases reproduced data if the Bureau's data validation process revealed an error.

Scope of data

The servicers provided data on all, or a sample of, mortgages they serviced (including loans the institution subserviced or had subserviced on their behalf) as of January 2012 and January 2015. The loan-level data include information on loan characteristics; loss mitigation applications initiated during 2012 and 2015; borrower communications relating to delinquency and loss mitigation; force-placed insurance; and communications regarding complaints, notices of error, and qualified written requests. Servicers that serviced more than 300,000 loans had the option of providing data on a random sample of 300,000 loans, oversampling loans that were more than 30 days delinquent or for which borrowers had applied for loss mitigation.

Additionally, each servicer provided monthly performance records for all, or a sample of, loans that were more than 30 days delinquent or made an initial application for loss mitigation at some point in calendar year 2012 or 2015. The Bureau asked the servicers who reported sampled data to oversample loans that met these criteria.

Limitations

The Servicing Operations Data are a unique source of information about the activities that were directly affected by the Rule; however, because they are drawn from the records of only seven servicers, the data may not be representative of the experience of borrowers at all servicers. Additionally, during the period for which the Bureau requested monthly records, some loans were transferred from one of the seven servicers to a servicer that did not report data to the Bureau, leading to incomplete monthly records.

Some of the information the Bureau requested for the Servicing Operations Data was not tracked consistently across the seven servicers or over time. In some instances, the implementation of the Rule led the servicers to update their systems to allow for better tracking of certain events. For example, several servicers that contributed to the Servicing Operations Data said that, since implementation of the Rule, they updated their systems to better track certain types of complaints that assert errors. These servicers, therefore, were likely able to provide information about events in 2015 that was unavailable or incomplete for events in 2012. In other instances, the Rule established new definitions for certain practices which may have affected how data was tracked and reported. The Rule, for instance, established a definition for a complete loss mitigation application which likely affected how the data were reported in 2015 as opposed to 2012.

Most servicers also had difficulty providing some of the information the Bureau requested. For example, all seven servicers used a vendor for force-placed insurance, and as such, some of these data were less readily accessible in the servicers' systems of record. In another instance, although the Bureau requested information on telephone calls related to delinquency or loss mitigation, most of the seven servicers were unable to provide complete data.

Summary of data

Information on all loans in the Servicing Operations Data, both delinquent and current. Table 3 provides an overview of all mortgages in the Servicing Operations Data for which the seven servicers provided loan-level data. As noted above, some of the servicers provided a sample of loans they serviced. The Bureau asked these servicers to oversample loans that were more than 30 days delinquent or made an initial application for loss mitigation in 2012 or 2015.

As such, the numbers provided below should not be interpreted as representative of the seven servicers that provided the data, but only representative of the data itself.

TABLE 3: SUMMARY STATISTICS ON ALL LOANS IN THE SERVICING OPERATIONS DATA

Metric

2012

2015

Number of loans

1,493,703

2,472,811

Share of loans that are more than 30 days delinquent or initiate a loss mitigation app in the year

22.4%

24.9%

Share of loans originated 2008 or later

53.9%

67.7%

Average credit score at origination

702

709

Share by investor

Fannie

39.2%

37.9%

Freddie

9.5%

10.6%

Ginnie

13.3%

17.9%

Private

21.2%

21.0%

Portfolio

16.8%

12.6%

Share with a first notice or filing for foreclosure January 2009 – June 2017

13.8%

9.2%

Share with a completed foreclosure sale January 2009 – June 2017

5.8%

3.3%

Share for which force-placed insurance was in place on Jan 1*

2.3%

1.5%

Share for which force-placed insurance was not in place on Jan 1, but the servicer sent an initial notice related to force-placed insurance during the year*

8.1%

7.5%

Share of accounts with error assertions between Jan. 2011 and June 2017

Complaints

1.8%

4.2%

Notices of error

0.0%

0.9%

Requests for information

1.8%

0.4%

Qualified written requests

1.8%

0.0%

* Includes five of seven servicers, as two servicers we unable to provide FPI data for both years

Information on all loans more than 30 days delinquent or with an initial application for loss mitigation in 2012 or 2015 in the Servicer Operations Data, both delinquent and current. For certain loans, the Bureau requested monthly performance records and other information pertaining to loss mitigation. Table 4 provides an overview of these loans. In addition to overall statistics we report the lowest and highest values among the seven servicers.

TABLE 4: SUMMARY STATISTICS ON ALL LOANS IN THE SERVICING OPERATIONS DATA THAT ARE MORE THAN 30 DAYS DELINQUENT OR SUBMIT AN INITIAL APPLICATION FOR LOSS MITIGATION IN 2012 OR 2015

Metric

2012 Overall

2012

Lowest servicer value

2012

Highest servicer value

2015 Overall

2015

Lowest servicer value

2015

Highest servicer value

As of January 1 of year, average outstanding balance

$187,755

$147,405

$260,751

$174,680

$136,631

$233,810

As of January 1 of year, average interest rate

6.2%

4.9%

7.2%

5.5%

4.5%

6.1%

As of January 1 of year, share 60 or more days delinquent

53.6%

15.6%

69.6%

50.9%

12.4%

71.7%

As of January 1 of year, share 90 or more days delinquent

45.6%

10.4%

62.9%

43.4%

7.1%

63.3%

As of January 1 of year, share 365 or more days delinquent

24.8%

2.2%

49.0%

27.3%

0.9%

42.4%

Share that initiate a loss mitigation app within the year*

34.6%

11.9%

42.9%

32.6%

13.1%

40.0%

Share of those that initiate that accept a trial mod within 18 months of year end*

17.7%

11.8%

36.7%

19.4%

6.2%

31.4%

Share of those that initiate that accept a permanent mod within 18 months of year end**

16.9%

8.9%

30.0%

18.6%

6.1%

26.6%

Share of those that initiate that accept a short-term payment forbearance program within 18 months of year end*

1.2%

0.2%

4.4%

2.8%

0.0%

9.5%

Communication attempts per account related to loss mitigation and delinquency from servicer to consumer

Written / Postal

26.3

2.1

60.1

21.1

2.5

36.3

Phone

78.1

0.5

335.8

63.7

0.3

180.7

* Includes six of seven servicers, as one servicer was unable to provide reliable loss mitigation application data.

** Includes five of seven servicers, as two servicers were unable to provide reliable permanent modification data.

Information on loans that were current then became 60 or more days delinquent in 2012 or 2015 in the Servicer Operations Data. The Servicing Operations Data includes information about loss mitigation applications initiated only during 2012 and 2015 and monthly performance records for these loans from January 1st of the application year onward. Therefore, to focus on borrowers who are early in delinquency and applying for loss mitigation for the first time, certain analyses related to loss mitigation are limited to borrowers who were current before transitioning to 60 or more days delinquent in 2012 or 2015. For certain analyses on the initiation of loss mitigation applications after the start of delinquency, the population is further limited to borrowers who were current before becoming 60 or more days delinquent in the first six months of each year. This provides a population of borrowers for which any loss mitigation application initiated within six months of becoming 60 or more days delinquent is included in the Servicing Operations Data. There are additional borrowers who initiate applications later in their delinquency who are excluded from analyses that use these two populations: of all applications initiated in 2012 (not just those where the borrower transitioned from current to delinquent during the year), 39 percent and 24 percent were initiated when the borrower was more than six months and one year past due, respectively. Of all applications initiated in 2015, 31 percent and 22 percent were initiated when the borrowers were more than six months and one year past due, respectively.

D. American Survey of Mortgage Borrowers

The American Survey of Mortgage Borrowers (ASMB) is part of the National Mortgage Database (NMDB®) program, which is jointly funded by the Federal Housing Finance Agency (FHFA) and the Bureau. The ASMB was first conducted in 2016, and subsequent waves of the survey took place in 2017 and 2018. The NMDB is a de-identified loan-level database of closed-end first-lien residential mortgages. It: (1) is representative of the market as a whole; (2) contains detailed, loan-level information on the terms and performance of mortgages, as well as certain characteristics of the associated borrowers and properties; (3) is continually updated; (4) has an historical component dating back before the financial crisis of 2008; and (5) provides the set of borrowers from which ASMB respondents are chosen.

The core data in NMDB are drawn from a random 1-in-20 sample of all closed-end first-lien mortgage files outstanding at any time between January 1998 and the present. A random 1-in-20 sample of newly reported mortgages is added each quarter. Mortgages are followed in the NMDB database until they terminate through prepayment (including refinancing), foreclosure, or maturity. The information on borrowers and loans available is de-identified and does not include any directly identifying information such as borrower name, address, or Social Security number.

ASMB is designed to provide information, particularly related to delinquent mortgage borrowers, that is not available in the NMDB. Participation in the survey is voluntary, and its target universe is current and delinquent borrowers of closed-end first-lien residential mortgages. To achieve this objective, ASMB draws its sample from mortgages that are part of NMDB.

Beginning with 2016, a random sample of 10,000 loans per year was drawn from loans in the NMDB. In 2016, the survey sample was chosen such that according to credit records, 70 percent of survey recipients were delinquent at the beginning of 2015, one year prior to the survey, and 30 percent were current at that time. In 2017, the sample was chosen such that 65 percent of those chosen were delinquent at the beginning of 2016, 25 percent were current, and 10 percent were delinquent Hispanic borrowers. In 2018, the same proportions were used as in 2017. The sample of Hispanic borrowers that were delinquent received both English and Spanish versions of the survey. All sampled borrowers were mailed a copy of the survey, but the survey provided the option to respond on-line. In 2016, 28 percent of responses were provided using the on-line system, and in 2017, this number rose to 41 percent after the printed survey highlighted the existence of the web site.

The ASMB survey response rates have been falling in recent years. In 2016 the survey received a response rate of 23.4 percent and in 2017 the survey received a response rate of 19.3 percent. Response rates continued to fall for 2018, with a rate of 17.5 percent. Response rates were higher for borrowers that were current on their mortgage than for those that showed a delinquency, but the count of responses was higher for delinquent borrowers because of their higher sampling rate. The survey results exclude responses that were unusable because respondents indicated they did not have a mortgage at the end of the preceding year, because responses related to a loan other than the loan in the NMDB, or because respondents left too many survey questions incomplete.

E. Consumer complaints

The Bureau began accepting mortgage complaints on December 1, 2011. As of December 31, 2017, the Bureau had handled over 300,000 mortgage complaints and another 9,000 debt collection complaints where the underlying type of debt was a mortgage. Mortgage is the third most complained about financial product or service, behind only debt collection and credit or consumer reporting.

Intake channels. The Bureau receives complaints through its website, by referral from the White House, congressional offices, and other federal and state agencies, and by telephone, mail, email, and fax. Results in this report are based on complaints received from the Bureau's web and telephone channels, which comprise 64 percent of all mortgage complaints, to ensure a consistent data collection.

Participating companies. The Bureau routes complaints about financial products and services directly to financial companies and works with them to get consumers a timely response. If a consumer submits a complaint against a company that is not currently participating in the complaint program, the Bureau contacts the company, works to get the company boarded to the company complaint portal, which companies use to provide responses to complaints, and sends the complaint for a response.

In 2012, 361 companies responded to mortgage complaints. In 2015, 621 companies responded to mortgage complaints. Despite this increase, only a small fraction of responses to mortgage complaints in 2015 came from companies that did not respond to complaints in 2012. Of the complaints submitted in 2015 that received a response, approximately 97.5 percent received a response from a company that was responding to complaints in 2012.

The Bureau's complaint form prompts consumers to select the consumer financial product or service with which they have a problem, as well as the issue that best describes their complaint. Consumers also respond to product-specific questions (e.g., are you concerned about losing your home to foreclosure?). Additionally, consumers provide narrative text describing their situation and a fair resolution to their issue.

The complaint form allows consumers to make only one issue selection. These issues correspond to the part of the mortgage process with which they are having a problem. Reflecting the complexity and interrelated nature of mortgages and mortgage issues, consumers are not asked to provide further specificity by selecting a sub-issue. While a single selection captures what consumers perceive to be the primary issue or problem they are having, it does not capture other topics that a consumer may raise in the narrative fields (e.g., a consumer who selects struggling to pay mortgage, but describes difficulties with a loan modification and problems with his or her credit report). To compensate for this limitation, the Bureau developed a list of keywords to identify complaints related to loss mitigation and force-placed insurance. The result of these keyword searches are used to identify "loss mitigation complaints" and "force-placed insurance complaints" in this report. The Bureau randomly sampled 400 loss mitigation complaints and 300 force-placed insurance complaints to be manually coded for certain analyses in the report.

APPENDIX D: Econometric Methodology for Estimating the Overall Effects of The Rule

Chapter 4 of this report presents estimates of the effect of the Rule on the incidence of foreclosure and the incidence of recovering from delinquency. This appendix details the methodology used for those estimates and presents additional supporting results.

Data construction

For the cumulative incidence of foreclosure, the outcome of interest is a completed foreclosure, while prepayment is treated as a competing risk. Loans that have a short sale, that are transferred out of the data or reach the last calendar month available in the data without an outcome are considered censored. For the cumulative incidence of recovery, the outcome of interest is becoming current for at least three consecutive months, whether through a self-cure or a loan modification. For the analysis of recovery, a completed foreclosure sale or prepayment are treated as competing risks, with other exits considered censoring.

Cumulative incidence functions

Chapter 4 presents a variety of figures describing the incidence of the outcomes of interest, foreclosure and recovery, respectively. These figures all show a non-parametric estimate of the cumulative incidence function of these outcomes. The cumulative incidence function for outcome j at time t is formally defined as the unconditional probability of experiencing outcome j at or before time t. The non-parametric estimate presented in the graphs, Ij(t) is computed as follows:

where S(t) is the Kaplan-Meier survival function for all outcomes, including competing outcomes, dij is the number of individuals that experience an outcome of type j at time t, and nt is the number still at risk (have not been censored or experienced any outcome) at time t. Using cumulative incidence instead of the survival function or the cause-specific hazard function makes it possible to handle competing risks that are correlated with the outcome of interest. In particular, the Kaplan-Meier estimator for the survival function tends to over-estimate incidence in the face of competing risks.

Competing risks regression models

Following the descriptive analysis of each outcome, Chapter 4 presents estimates of the total change in each outcome caused by the Rule. These estimates are based on a competing risks hazard regression model. At a high level such a model can be thought of as follows, using foreclosure as an example: At every point in "event time" (here, the number of months since becoming either 90 days delinquent) where a foreclosure occurs, compare the characteristics of the loan that experienced a foreclosure to all other loans that are still at risk at the same time (i.e., neither foreclosed upon or censored). The coefficients are then selected such that at every event time, the loan actually foreclosed upon is predicted to be the most likely loan to be foreclosed upon. A positive coefficient indicates that the loan characteristic is associated with an increase in the cumulative incidence of foreclosure, and a negative coefficient indicates the opposite.

Estimating the effect of the Rule, accounting for other important factors in the housing market, requires a model with at least three features: First, the model must be able to account for competing risks. Standard hazard regression models are based on the assumption that any censoring is non-informative, or at random. Prepayment is very likely correlated with the underlying risk of foreclosure, and both prepayment and foreclosure are likely correlated with the underlying risk of recovery. Second, the model must be able to handle time-varying covariates. A number of approaches to hazard modeling require all characteristics of each individual to be fixed in time. However, a number of characteristics of the loans in the Fannie Mae and McDash Data change over time. Most importantly, when time is specified as months since the beginning of delinquency, the Rule itself is a time-varying covariate. Finally, given the size of the Fannie Mae and McDash datasets, the model must be relatively efficient computationally. The Fannie Mae Data, for instance, contain more than 35 million loan-month observations. There are ways to estimate a competing risks hazard model with time varying covariates that are simply too slow to compute in a reasonable amount of time. The regression analysis in Chapter 4 employs the sub-distribution hazard model of Fine and Gray. The FineGray model handles competing risks and time varying covariates, at the cost of some difficulty in directly interpreting the coefficients.

The Fine-Gray model is relatively straightforward, computationally. The model is an extension of the commonly used Cox proportional hazards model. The Cox model assumes proportional hazards. That is, if Tj is the time when the outcome occurs for individual i, then,

where 0(t) is the baseline hazard, and Xit is a vector of potentially time-varying independent variables. In the Cox model, a partial likelihood function is constructed with one term for each time period in which an outcome occurs. If individual i is the individual that experiences the outcome at time t, then the likelihood term is the probability that i experiences that outcome, given that an outcome occurs at time t, which by Bayes Rule is:

where Rt is the set of individuals at risk at time t, and the baseline hazard 0(t) cancels from numerator and denominator. A tie-breaking procedure is used for cases where more than one individual experiences an outcome at a particular period, and the coefficients b are estimated via a maximum likelihood procedure. In other words, coefficients are chosen such that the individuals that actually experience the outcome in each period are predicted to be the most likely individuals to do so.

The difference of the Fine-Gray model from the standard Cox model is that individuals who experience a competing risk remain in the risk set Rt in future periods, even after they have exited the data. This construction may seem counter-intuitive, as by definition individuals who have experienced a competing risk cannot have the outcome of interest occur. However, as Fine and Gray show, mathematically this treatment of the risk set leads to a proper partial likelihood function for the cumulative incidence of the outcome interest.

To account for the possibility that individuals who experienced the competing risk would have been right censored by another means (e.g., the end of the time period covered by the data), the additional observations are assigned a weight less than one, equal to an estimate of the probability that the individual would have remained uncensored absent the competing risk. The weights are computed as the Kaplan-Meier survival function where the outcome is censoring. An advantage of this approach for computational efficiency is that the weights are not dependent on the vector of covariates and thus can be calculated once before running the regressions, which in turn are simply a standard Cox regression program.

For the regressions in Chapter 4, the X variables for the loans with a competing risk are generally carried forward from the last period they are observed in the data. The exception is for variables driven by policy and the value of the monthly house price index, where it is clear what the value would be if the loan had not exited from the competing risk.

To make clear how a loan that experiences a competing risk is treated in the regression analysis, consider a loan in the foreclosure analysis that becomes 90 days delinquent in July 2013, but is pre-paid in November 2013. For purposes of this loan, July 2013 corresponds to t=1. For terms in the likelihood corresponding to t≤ 5, this loan is included in the denominator as part of the risk set, with the actual values of its covariates in those months, and a weight of one. Starting with period t=6, the loan is still included in the risk set in the denominator of the partial likelihood function, but it is assigned a weight of less than one, reflecting the probability of being censored after period 5 in the data as a whole. For variables such as unpaid balance and the previous month's delinquency status, the values from November 2013 are carried forward for this loan, but the Black Knight house price index is updated to the December 2013 value for the borrower's state. However, starting with period t=7, which corresponds to January 2014 for this loan, the indicator for the Rule being in effect switches from zero to one, as any loan present in January 2014 or after would have had the Rule in effect. Ideally, other time-varying covariates such as unpaid balance would adjust as well, but it would require many strong assumptions to determine how, for instance, unpaid balance would evolve had the loan not been prepaid.

A major drawback of the Fine and Gray model is that the coefficients produced by the model are not readily interpretable by themselves. In the Cox model, the exponentiated coefficients can be interpreted as hazard ratios, such that a one unit increase in X multiplies the hazard rate at every point in time by exp(b). The addition of individuals who have experienced competing risks in the risk set breaks this interpretation. Instead, a change in X multiplies what Fine and Gray refer to as the sub-distribution hazard, which does not have any natural interpretation. The coefficients can be indirectly interpreted as coefficients for a complementary log-log generalized linear model for the cumulative incidence function, which means that a positive coefficient indicates that the associated covariate increases the cumulative incidence and a negative coefficient decreases it. However, interpreting the magnitudes of the coefficients is still problematic, as the complementary log-log link function is also difficult to interpret with respect to magnitudes. For relatively low levels of incidence, the complementary log-log link function and the logistic link function are approximately the same, and the coefficients from the latter can be interpreted as log-odds ratios. However, in practice the cumulative incidence of foreclosure, and particularly the cumulative incidence of recovery, are too large for this approximation to hold. Despite this drawback, it is possible to use the coefficients of the FineGray model to predict the level of the cumulative incidence function for specific values of the covariates.

Foreclosure regression coefficients

As noted above, it is difficult to interpret the magnitude of the coefficients in a Fine-Gray regression model. However, the signs of the coefficients are meaningful, indicating the direction of the effect of covariates on the cumulative incidence function, and inference is also informative. The tables below present the coefficients from models with foreclosure as the outcome of interest. As noted above and in Chapter 4, for this analysis loans are followed from the first month they are observed being 90 or more days delinquent until they experience a completed foreclosure, a prepayment, or are censored by the end of the data, a transfer, or similar phenomena. Foreclosure is the outcome of interest, while prepayment is a competing risk.

The key variable in the regressions is an indicator for whether or not the Rule was effective at the beginning of the current month. Because of the structure of the Fine and Gray model, in essence the coefficient on this indicator is determined by comparing similar consumers for whom the same amount of time has passed since becoming 90 days delinquent, but with or without the Rule in effect. In addition to the indicator for the Rule, all of the regressions control for a set of policy variables and loan characteristics. In both datasets, all regressions include year-of-initial-delinquency fixed effects, the month-over-month change in the (log) Black Knight house price index by state, the loan-to-value ratio at origination, an indicator for whether the FHFA Servicing Alignment Initiative was in effect in the current month, indicators for whether a loan was currently experiencing the foreclosure moratoria enacted by the GSEs in 2008 or by the NMS parties between September 2010 and December 2013, an indicator for whether a loan modification was in effect in the current month, the number of months the loan was delinquent at the start of the month,42 the current interest rate, the current unpaid balance and an indicator for whether the loan was originated for the purpose of refinancing. Regressions using the Fannie Mae Data further control for the debt-to-income ratio and borrower credit score, both calculated at origination, and an indicator for whether the NMS applied to a loan (i.e., for loans serviced by NMS servicers after the NMS effective date). Regressions using the McDash Data also control for the current monthly payment (principal and interest), the original loan term, and whether the loan has a fixed rate.

Table 1 presents the competing risks hazard regression coefficients for the outcome of foreclosure using the Fannie Mae Data. Cluster-robust standard errors clustered by loan are reported in parentheses. Column (1) shows results of a simple version of the model, using only the control variables described above. Column (2) reports results including year-of-origination fixed effects and column (3) further adds state and servicer fixed effects. The coefficients for the effect of the Rule are negative, indicating that the Rule reduced the incidence of foreclosure once it was in effect. Adding more controls has almost no effect on the coefficient, and the effect of the Rule is statistically significant in all specifications. This consistency suggests that the estimate of the effect of the Rule is truly due to the Rule rather than other factors that are not included in the regression. If, in contrast, the estimate were highly sensitive to the other controls included, that might suggest that indicator for the Rule was picking up other factors changing at the same time as the Rule. The other coefficients in Table 1 largely match expectations—a modification in effect reduces the incidence of foreclosure, while falling further behind on a mortgage increases it. Foreclosures are more common for loans with higher interest rates and less common for loans with higher balances.

TABLE 1: COMPETING RISKS HAZARD MODEL RESULTS FOR THE HAZARD OF FORECLOSURE: FANNIE MAE DATA

(1)

(2)

(3)

After Rule Effective Date

-0.319***

-0.312***

-0.163***

(0.00907)

(0.00907)

(0.00927)

Modification in Effect

-0.128***

-0.162***

-0.213***

(0.00706)

(0.00743)

(0.00734)

Months Delinquent

0.0943***

0.0941***

0.107***

(0.000188)

(0.000189)

(0.00018)

Current Interest Rate

-0.00514**

-0.0205***

-0.00405

(0.00228)

(0.00259)

(0.00262)

Current Balance

-0.418***

-0.456***

-0.33***

(0.00257)

(0.0027)

(0.00301)

Loan Characteristics at Origination

Yes

Yes

Yes

Indicators for Pre-Rule Policies

Yes

Yes

Yes

Origination Year FE

No

Yes

Yes

State FE

No

No

Yes

Servicer FE

No

No

Yes

Year of Initial Delinquency FE

Yes

Yes

Yes

Number of Loans

961180

956276

960069

*** P < 0.01; ** P < 0.05; * P < 0.1

Notes: Statistics are coefficients from the Fine and Gray (1999) subdistribution hazard regression model using fixed-rate, 30 year mortgages from Fannie Mae. A completed foreclosure sale is the outcome of interest and prepayment is a competing risk. Loans are considered to be at risk once they first become 90 days delinquent. Standard errors clustered by loan are reported in parentheses. Loan characteristics at origination include borrower credit score, DTI and an indicator for refinance.

Table 2 shows results using the McDash Data. The layout of the table is the same as Table 1, save that instead of controls for servicer, column (3) contains investor type fixed-effects (e.g., Fannie Mae, Freddie Mac, Private/Portfolio), as the McDash Data do not contain servicer identifiers. These results also indicate that having the Rule effective reduces the incidence of foreclosure, with little change as more covariates are added.

TABLE 2: COMPETING RISKS HAZARD MODEL RESULTS FOR THE HAZARD OF FORECLOSURE: MCDASH DATA

(1)

(2)

(3)

After Rule Effective Date

-0.159***

-0.21***

-0.181***

(0.0155)

(0.0155)

(0.0156)

Modification in Effect

-0.671***

-0.626***

-0.628***

(0.00845)

(0.0085)

(0.00855)

Months Delinquent

0.0448***

0.0471***

0.0536***

(0.000112)

(0.000122)

(0.000124)

Current Interest Rate

0.0205***

0.0416***

0.0466***

(0.00129)

(0.00137)

(0.00148)

Current Balance

0.105***

0.112***

0.111***

(0.00118)

(0.00124)

(0.0011)

Loan Characteristics at Origination

Yes

Yes

Yes

Indicators for Pre-Rule Policies

Yes

Yes

Yes

Origination Year FE

No

Yes

Yes

State FE

No

No

Yes

Investor FE

No

No

Yes

Year of Initial Delinquency FE

Yes

Yes

Yes

Number of Loans

797557

796932

794086

*** P < 0.01; ** P < 0.05; * P < 0.1

Notes: Statistics are coefficients from the Fine and Gray (1999) subdistribution hazard regression model, using 30 year mortgages from the Black Knight McDash data. A completed foreclosure sale is the outcome of interest and prepayment is a competing risk. Loans are considered to be at risk once they first become 90 days delinquent. Standard errors clustered by state are reported in parentheses. Loan characteristics at origination include indicators for refinance and fixed rate.

Recovery regression coefficients:

The tables below present the coefficients from models with recovery as the outcome of interest. As noted above and in Chapter 4, for this analysis loans are followed from the first month they are observed being 30 or more days delinquent until they become current, pre-pay, have a completed foreclosure, or are censored by the end of the data, a transfer, or similar phenomena. Loans must become current for at least three consecutive months to be coded as having recovered, but the timing of a recovery is coded as the first month a loan becomes current following a 30-day delinquency. Recovery is the outcome of interest, while prepayment and foreclosure are competing risks. The control variables are the same as for the foreclosure regressions described above, except that the indicator for having a loan modification is excluded, as loans that are modified are counted as recovering, making it collinear with the outcome.

Table 3 presents the competing risks hazard regression results for the outcome of recovery using the Fannie Mae Data. Column (1) shows results of a simple version of the model, using only the control variables described above. Column (2) reports results including origination year fixed effects, and column (3) further adds state and servicer fixed effects. Regardless of the specification, the model indicates that the Rule increased the incidence of recovery. The coefficients on other variables in the model largely align with expectations. The more delinquent a loan is at the beginning of the month, the less likely it is to recover during that month, although the effect is small. Consumers with higher interest rate loans (and presumably higher monthly payments) are less likely to recover, as are consumers with higher balances. As with foreclosure, the likely amount of equity the borrower has is important—increases in house prices are associated with a higher incidence of recovery, while loans with high loan-to-value ratios have a lower incidence of recovery.

TABLE 3: COMPETING RISKS HAZARD MODEL RESULTS FOR THE HAZARD OF RECOVERY: FANNIE MAE DATA

(1)

(2)

(3)

Rule Effective

0.146***

0.147***

0.15***

(0.00647)

(0.00648)

(0.0065)

Months Delinquent

0.00812***

0.00791***

0.00678***

(0.000274)

(0.000276)

(0.000274)

Current Interest Rate

-0.455***

-0.604***

-0.608***

(0.00105)

(0.00131)

(0.00131)

Current Balance (00000s)

-0.0364***

-0.0347***

-0.0294***

(0.000816)

(0.000849)

(0.000936)

Monthly Percent Change in House Prices

9.98***

9.55***

9.59***

(0.0952)

(0.0954)

(0.0982)

Loan to Value at Origination

-0.00498***

-0.00486***

-0.0052***

(0.0000443)

(0.0000445)

(0.0000468)

Loan Characteristics at Origination

Yes

Yes

Yes

Indicators for Pre-Rule Policies

Yes

Yes

Yes

Year of Initial Delinquency FE

Yes

Yes

Yes

Origination Year FE

No

Yes

Yes

State FE

No

No

Yes

Servicer FE

No

No

Yes

Number of Loans

2891394

2891394

2891394

*** P < 0.01; ** P < 0.05; * P < 0.1

Notes: Statistics are coefficients from the Fine and Gray (1999) subdistribution hazard regression model using fixed-rate, 30 year mortgages from Fannie Mae. Recovery, defined as beginning a spell of current payment status of at least 3 months or receiving a permanent loan modification, is the outcome of interest and prepayment or foreclosure are competing risks. Loans are considered to be at risk once they first become 30 days delinquent. Standard errors clustered by loan are reported in parentheses. Loan characteristics at origination include borrower credit score, DTI and an indicator for refinance.

Table 4 shows matching results using the McDash Data. The layout of the table is the same as Table 3, save that instead of controls for servicer, columns (2) and (3) contain controls for investor type (e.g., Fannie Mae, Freddie Mac, Private), as the McDash Data do not contain servicer identifiers. The coefficient on the Rule shows the same pattern as in the Fannie Mae Data, with little change regardless of the specification. The coefficients in the McDash Data are larger in absolute value, although it is important to remember that these are not comparable across datasets. Unexpectedly, in the McDash Data loans with higher LTV had a higher incidence of recovery.

TABLE 4: COMPETING RISKS HAZARD MODEL RESULTS FOR THE HAZARD OF RECOVERY: MCDASH DATA

(1)

(2)

(3)

Rule Effective

0.468***

0.468***

0.478***

(0.00879)

(0.0088)

(0.00879)

Forwarded to Foreclosure Attorneys

-0.415***

-0.41***

-0.409***

(0.0056)

(0.0056)

(0.00562)

Months Delinquent

-0.037***

-0.037***

-0.0368***

(0.000281)

(0.000282)

(0.000281)

Current Interest Rate

-0.288***

-0.297***

-0.298***

(0.000769)

(0.000796)

(0.000808)

Current Balance (00000s)

0.0749***

0.074***

0.0736***

(0.000307)

(0.00031)

(0.000324)

Monthly Percent Change in House Prices

14.1***

13.8***

13.3***

(0.134)

(0.134)

(0.137)

Loan to Value at Origination

0.00000138*

0.00000138*

0.00000141*

(0.000000745)

(0.000000744)

(0.000000736)

Loan Characteristics at Origination

Yes

Yes

Yes

Indicators for Pre-Rule Policies

Yes

Yes

Yes

Year of Initial Delinquency FE

Yes

Yes

Yes

Origination Year FE

No

Yes

Yes

State FE

No

No

Yes

Investor FE

No

No

Yes

Number of Loans

1561881

1561881

1557683

*** P < 0.01; ** P < 0.05; * P < 0.1

Notes: Statistics are coefficients from the Fine and Gray (1999) subdistribution hazard regression model, using 30 year mortgages from the McDash data. Recovery, defined as beginning a spell of current payment status of at least 3 months or receiving a permanent loan modification, is the outcome of interest and prepayment or foreclosure are competing risks. Loans are considered to be at risk once they first become 30 days delinquent. Standard errors clustered by loan are reported in parentheses. Loan characteristics at origination include indicators for refinance and fixed rate.

Event study analysis

The regression results presented above suggest that the Rule may have reduced the incidence of foreclosure and increased the incidence of recovery. In general, more evidence is needed to show whether this apparent effect is truly causal or the result of other trends in the mortgage market. However, note that the results of the model are not simply a before and after comparison, which could easily be biased by other trends—the structure of the model is such that the comparison is between borrowers at the same stage of delinquency, before or after the Rule, and holds a number of other factors constant. Still it is possible for a bias from unrelated trends to persist. Chapter 4 plots the results of a distributed-lags type event study. In this context, an event study is a more flexible version of the models presented above—instead of a single indicator for whether the current month is before or after the effective date of the Rule, the model includes a set leads and lags around the Rule in three- month bins (e.g., 9 to 12 months before the Rule, 3 to 6 months after etc.). If there were no unrelated trends after holding constant the other factors in the model, the coefficients on the pre-Rule indicators would make a horizontal line. The difference between the coefficients of the pre-Rule indicators and those of the post-Rule indicators would indicate the effect of the Rule. Note that the effect of the Rule need not be constant over time, but could grow or shrink, and the post-Rule coefficients will capture any such changes.

In practice, the pre-Rule indicators shown in Chapter 4 do not form a horizontal line, indicating the presence of some trend in the incidence of foreclosures and recoveries even after controlling for other factors. However, these trends appear roughly linear, and so it is possible to recover the effect of the Rule net of the pre-trends using a parametric event study.429 This is implemented by replacing the pre-Rule indicators in the event study model with a linear trend in calendar time. Since the post-Rule indicators non-parametrically account for any changes post-Rule, this time trend will capture only the pre-Rule trend, and the post-Rule coefficients can then be interpreted as changes compared to the pre-Rule trend. These coefficients could be plotted directly, but the figures in Chapter 4 instead plot the estimated linear pre-trend on top of the non-parametric event study coefficients, to allow the reader to visually compare the effects of the Rule to the pre-trend. The slope of the dashed pre-trend lines in the figures are equal to the coefficient on the linear calendar time trend.

Predicting the effect of the Rule:

Tables 1 and 2 in Chapter 4 present estimates of the change in foreclosures and recoveries caused by the Rule. While the magnitude of the coefficients for the effect of the Rule are difficult to interpret by themselves, it is valid to use the model to predict the cumulative incidence of foreclosure. In other words, one can calculate, based on the results of the model, what fraction of loans that became 90 days delinquent in 2014 would have experienced foreclosure if the Rule had not gone into effect in January 2014, or what fraction of loans that became 30 days delinquent would have recovered absent the Rule. The procedure used for the first row of the tables is straightforward: first, use the model to predict the cumulative incidence of foreclosure for loans that become 90 days delinquent in 2014, using the actual characteristics of those loans over the next 36 months—including the status of the Rule, which was in effect for all such loans over that period. Then calculate the same prediction but set the indicator for the Rule to zero— assuming the Rule had not actually gone into effect in January 2014. The tables report the difference in the predicted cumulative incidence functions at t = 36. The text in Chapter 4 explains the computation of the remaining rows of the tables, used to scale up the predicted change in cumulative incidence to the full U.S. mortgage market.

APPENDIX E: Survey of Housing Counselors and Legal Aid Organizations

The Bureau surveyed housing counselors and legal aid attorneys who counsel clients regarding mortgage delinquency and default resolution to learn about their experiences and perspectives.

In total, the Bureau collected 239 partial or complete responses, 140 from housing counselors and 99 from legal aid attorneys.

TABLE 1: FREQUENCY OF DISCUSSING THE 2014 RESPA SERVICING REQUIREMENTS WITH CLIENTS

Question

Rarely or Never

Sometimes

Often or always

n

How often do you discuss the servicer's practices for making live contact?

18%

21%

61%

173

How often do you discuss servicer's early intervention written notices?

11%

24%

65%

169

How often do you discuss the servicer's written acknowledgement notices?

4%

23%

72%

166

How often do you discuss the servicer's continuity of contact policies?

12%

32%

57%

161

How often do you discuss the servicer's reasonable diligence policies?

12%

26%

63%

160

How often do you discuss the servicer's timeline for evaluating your client's application?

3%

20%

77%

160

How often do you discuss the requirement to evaluate only applications that are complete?

6%

16%

78%

154

How often do you discuss the servicer's appeal policies?

6%

16%

78%

154

How often do you discuss the restrictions on beginning a foreclosure action?

7%

24%

69%

193

How often do you discuss the restrictions on foreclosure sale?

6%

16%

78%

180

How often do you discuss servicer's error resolution practices?

33%

37%

30%

201

How often do you discuss the servicer's force-placed insurance practices?

19%

43%

38%

154

TABLE 2: GENERAL EFFECTIVENESS OF THE 2014 RESPA SERVICING REQUIREMENTS

Question

No

Somewhat

Yes

n

In general, are the live contact requirements effective in helping your clients?

19%

43%

39%

140

In general, are the early intervention written notice requirements effective in helping your clients?

8%

50%

42%

149

In general, are the continuity of contact requirements effective in helping your clients?

18%

49%

33%

141

In general, are the reasonable diligence requirements effective in helping your clients?

8%

44%

48%

141

In general, is the 30-day evaluation timeline effective in helping your clients?

7%

34%

59%

152

In general, is the requirement to evaluate only applications that are complete effective in helping your clients?

14%

39%

47%

144

In general, are the appeals rules effective in helping your clients?

14%

39%

47%

144

In general, are these restrictions on beginning the foreclosure process effective in helping your clients?

6%

30%

64%

171

In general, are these restrictions on foreclosure sale effective in helping your clients?

4%

30%

65%

164

In general, are the error resolution rules effective in helping your clients?

13%

49%

38%

127

In general, are the force-placed insurance requirements effective in helping your clients?

22%

35%

42%

125

Part A: Early Intervention

TABLE 3: REASONS WHY THE LIVE CONTACT REQUIREMENTS ARE EFFECTIVE OR NOT EFFECTIVE

Question

Rarely or Never

Sometimes

Often or always

Don't Know or N/A

n

When the live contact requirements are effective in helping your clients, how often is it due to the following?

They help your clients understand their options.

17%

41%

41%

1%

140

Your clients receive information about the delinquency or their options in a timely manner.

16%

36%

49%

0%

140

They motivate your clients to begin a loss mitigation application.

12%

34%

54%

1%

140

They help your clients complete a loss mitigation application.

27%

35%

36%

1%

140

They help your clients resolve the delinquency without a foreclosure sale.

19%

46%

31%

4%

140

When the live contact requirements are NOT effective in helping your clients, how often is it due to the following?

The servicer does not establish contact with your client in a timely manner.

30%

39%

29%

3%

140

The servicer's communications are not understandable to clients who have limited English proficiency.

17%

27%

39%

16%

140

The servicer's communications are not understandable to clients who understand English well.

16%

37%

41%

6%

140

The servicer does not provide useful information.

19%

38%

40%

3%

140

The client does not respond even though the servicer's communications provide useful information.

22%

52%

22%

4%

140

Other problems with early intervention or loss mitigation mean the live contact requirements are not beneficial to your clients.

33%

39%

13%

15%

140

TABLE 4: REASONS WHY THE WRITTEN NOTICE REQUIREMENTS ARE EFFECTIVE OR NOT EFFECTIVE

Question

Rarely or Never

Sometimes

Often or always

Don't Know or N/A

n

When the early intervention written notice requirements are effective in helping your clients, how often is it due to the following?

They help your clients understand their options.

7%

48%

44%

1%

149

Your clients receive information about the delinquency or their options in a timely manner.

3%

41%

54%

1%

149

They prompt your clients to initiate a loss mitigation application.

6%

40%

53%

1%

149

They help your clients resolve the delinquency without a foreclosure sale.

12%

45%

40%

3%

149

When the written notice requirements are NOT effective in helping your clients how often is it due to the following?

The servicer does not comply with the written notice requirements.

42%

28%

25%

5%

140

The servicer's communications are not understandable to clients who have limited English proficiency.

15%

25%

44%

15%

140

The servicer's communications are not understandable to clients who understand English well.

14%

38%

43%

5%

140

The servicer does not provide useful information.

24%

44%

28%

3%

140

The client does not respond even though the servicer's communications provide useful information.

25%

46%

28%

1%

140

The client attempts to contact the servicer about the notice, but the servicer is unresponsive.

15%

36%

48%

1%

140

Part B: Loss Mitigation

TABLE 5: REASONS WHY THE WRITTEN ACKNOWLEDGEMENT REQUIREMENTS ARE EFFECTIVE OR NOT EFFECTIVE

Question

Rarely or Never

Sometimes

Often or always

Don't Know or N/A

n

When the written acknowledgement requirements are effective in helping your clients, how often is it due to the following?

They help your clients understand what additional documents and information the servicer requires from your clients for a complete application.

4%

33%

62%

1%

156

They prompt your clients to take action to complete their loss mitigation application.

4%

31%

63%

1%

156

When the written acknowledgement requirements are NOT effective in helping your clients, how often is it due to the following?

The servicer does not promptly provide the written acknowledgment notice.

18%

37%

44%

1%

156

The content of the written acknowledgment notice is not understandable to clients who have limited English proficiency.

12%

20%

47%

21%

156

The content of the written acknowledgment notice is not understandable to clients who understand English well.

12%

40%

44%

5%

156

The written acknowledgement notice is not specific enough.

15%

29%

54%

2%

156

The client does not respond even though the servicer's communications provide useful information.

37%

47%

12%

4%

156

Other problems with loss mitigation mean the written acknowledgment requirements do not make a difference.

26%

38%

15%

20%

156

TABLE 6: REASONS WHY THE CONTINUITY OF CONTACT REQUIREMENTS A EFFECTIVE RE EFFECTIVE OR NOT

Question

Rarely or Never

Sometimes

Often or always

Don't Know or N/A

n

When the continuity of contact requirements are effective in helping your clients, how often is it due to the following?

They help your clients obtain accurate information about loss mitigation applications.

11%

44%

44%

1%

141

They help your clients complete a loss mitigation application.

21%

42%

37%

1%

141

They reduce the amount of time it takes for your clients to complete a loss mitigation application.

29%

40%

29%

1%

141

They help your clients resolve the delinquency without a foreclosure sale.

20%

46%

31%

3%

141

When the continuity of contact requirements are NOT effective in helping your clients, how often is it due to the following?

You or your client have difficulty reaching servicer personnel or obtaining a timely response to questions.

6%

20%

73%

1%

141

Servicer personnel are not able to provide accurate information about the client's loss mitigation application.

7%

34%

59%

1%

140

The servicer does not adequately communicate with consumers who have limited English proficiency.

18%

24%

35%

23%

141

TABLE 7: REASONS WHY THE REASONABLE DILIGENCE REQUIREMENTS ARE EFFECTIVE OR NOT EFFECTIVE

Question

Rarely or Never

Sometimes

Often or always

Don't Know or N/A

n

When the reasonable diligence requirements are effective in helping your clients, how often is it due to the following?

They help your clients understand what documents and information the servicer requires for a complete application.

8%

38%

53%

1%

141

They prompt your clients to submit additional documents and information for the loss mitigation application.

6%

33%

60%

1%

141

They help your clients complete a loss mitigation application.

13%

40%

46%

1%

141

They help your clients resolve the delinquency without a foreclosure sale.

13%

43%

39%

5%

141

When the reasonable diligence requirements are NOT effective in helping your clients, how often is it due to the following?

The servicer is not reasonably diligent in attempting to reach the client.

15%

36%

47%

2%

141

The servicer does not communicate what documents and information the client must provide to complete the application.

15%

35%

48%

1%

141

The servicer does not adequately communicate with consumers who have limited English proficiency.

18%

20%

37%

26%

141

The client is unresponsive to the servicer's reasonable diligence efforts.

37%

48%

13%

1%

141

Other problems with loss mitigation mean reasonable diligence rules do not make a difference.

23%

37%

13%

27%

141

TABLE 8: REASONS WHY THE 30-DAY EVALUATION TIMELINE IS EFFECTIVE OR NOT EFFECTIVE

Question

Rarely or Never

Sometimes

Often or always

Don't Know or N/A

n

When the 30-day evaluation timeline is effective in helping your clients, how often is it due to the following?

         

It helps your clients timely obtain a decision on their complete loss mitigation application.

4%

36%

59%

2%

152

It helps your clients obtain loss mitigation offers.

5%

43%

49%

3%

152

It helps your clients resolve the delinquency before the servicer starts the foreclosure process.

13%

37%

45%

5%

152

When the 30-day evaluation timeline is NOT effective in helping your clients, how often is it due to the following?

It takes significantly more than 30 days for the servicer to evaluate the application and notify the client because there is a delay in receiving information from a third party.

25%

34%

37%

4%

152

It takes significantly more than 30 days for the servicer to evaluate the application and notify the client for a reason other than a delay in receiving information from a third party.

14%

34%

51%

1%

152

The servicer timely notifies the client of its decision, but the client does not understand the determination letter.

28%

46%

24%

2%

152

The servicer timely notifies the client of its decision, but does not offer any options that are helpful to the client.

14%

51%

35%

1%

152

Other problems with loss mitigation mean the 30-day evaluation timeline does not make a difference.

25%

36%

16%

24%

152

TABLE 9: REASONS WHY THE REQUIREMENT TO EVALUATE ONLY APPLICATIONS THAT ARE COMPLETE IS EFFECTIVE OR NOT EFFECTIVE

Question

Rarely or Never

Sometimes

Often or always

Don't Know or N/A

n

When this requirement to evaluate only applications that are complete is effective in helping your clients, how often is it due to the following?

         

It helps your clients understand their options.

22%

35%

40%

2%

144

It helps your clients obtain loss mitigation.

15%

47%

37%

1%

144

It helps shorten the loss mitigation application timeline.

28%

33%

37%

1%

144

It helps your clients resolve the delinquency without a foreclosure sale.

19%

42%

35%

4%

144

When this requirement to evaluate only applications that are complete is NOT effective in helping your clients, how often is it due to the following?

The servicer does not follow the rule – it evaluates the application before it is complete.

56%

23%

17%

4%

144

The application is unnecessarily delayed because it takes a long time for the servicer to collect documents and information relating to all available loss mitigation options.

7%

39%

52%

2%

144

The servicer loses documents or makes duplicative requests for documents, which makes the process take a long time

3%

22%

73%

1%

144

Other problems with loss mitigation mean this requirement does not make a difference.

24%

35%

17%

24%

144

TABLE 10: REASONS WHY THE APPEALS RULES ARE EFFECTIVE OR NOT EFFECTIVE

Question

Rarely or Never

Sometimes

Often or always

Don't Know or N/A

n

When appeals rules effective in helping your clients, how often is it due to the following?

They help the servicer correct an error in evaluating the complete loss mitigation application.

20%

42%

34%

4%

137

They help your clients obtain a loss mitigation option the servicer would have otherwise denied.

23%

45%

28%

4%

137

When appeals rules are NOT effective in helping your clients, how often is it due to the following?

The servicer does not permit or timely permit an appeal when required under the Mortgage Servicing Rule.

39%

35%

22%

4%

137

The servicer permits an appeal, but the client does not make an appeal.

31%

58%

9%

2%

137

The client makes an appeal, but the servicer wrongly denies the appeal.

16%

42%

38%

4%

137

The client makes an appeal, but the servicer appropriately denies the appeal.

26%

53%

15%

5%

137

The servicer takes too long to review the appeal.

33%

39%

23%

5%

137

Other problems with loss mitigation mean the appeals rules do not make a difference.

30%

33%

8%

29%

137

Part C: Foreclosure

TABLE 11: REASONS WHY THE RESTRICTIONS ON BEGIN EFFECTIVE OR NOT EFFECTIVENING A FORECLOSRE ACTION ARE

Question

Rarely or Never

Sometimes

Often or always

Don't Know or N/A

n

When these restrictions on beginning the foreclosure process are effective in helping your clients, how often is it due to the following?

They help your clients initiate a loss mitigation application.

7%

26%

66%

2%

171

They help your clients complete a loss mitigation application.

11%

30%

58%

2%

171

They help your clients obtain loss mitigation.

10%

42%

47%

2%

171

They help your clients resolve the delinquency without a foreclosure sale.

9%

37%

51%

2%

171

When these restrictions on beginning the foreclosure process are NOT effective in helping your clients, how often is it due to the following?

The servicer does not follow these rules.

30%

30%

37%

3%

171

The client does not act promptly enough to complete a loss mitigation application before the 120-day period expires and the servicer begins foreclosure proceedings.

16%

43%

39%

2%

171

The servicer makes it difficult for the client to complete an application.

14%

30%

54%

2%

171

Other problems with loss mitigation mean these restrictions do not make a difference.

29%

33%

18%

20%

171

TABLE 12: REASONS WHY THE RESTRICTIONS ON FORECLOSURE SALE ARE EFFECTIVE OR NOT EFFECTIVE

Question

Rarely or Never

Sometimes

Often or always

Don't Know or N/A

n

When these restrictions on foreclosure sale are effective in helping your clients, how often is it due to the following?

It helps your clients initiate a loss mitigation application.

4%

29%

67%

0%

165

It helps your clients complete a loss mitigation application.

4%

29%

66%

1%

165

It helps your clients obtain loss mitigation.

6%

40%

54%

0%

164

It helps your clients resolve the delinquency without a foreclosure sale.

9%

38%

53%

1%

165

When these restrictions foreclosure sale are NOT effective in helping your clients, how often is it due to the following?

The servicer does not follow these rules.

25%

35%

37%

3%

161

The client does not act promptly enough to complete a loss mitigation application more than 37 days before a foreclosure sale.

16%

48%

33%

3%

165

The servicer makes it difficult for the client to complete an application.

15%

33%

49%

4%

165

Other problems with loss mitigation mean this restriction does not make a difference.

30%

35%

16%

19%

165

Part D: Error Resolution

TABLE 13: REASONS WHY THE ERROR RESOLUTION PROVISIONS ARE EFFECTIVE OR NOT EFFECTIVE

Question

Rarely Often or

Sometimes

or Never always

Don't Know or N/A

n

When the error resolution rules are effective in helping your clients, how often is it due to the following?

They help your client resolve miscommunications with the servicer.

15%

45%

38%

2%

127

They help your client obtain information about the mortgage loan.

7%

31%

61%

2%

127

They help your client save time or money.

28%

43%

26%

2%

127

They help your client fix a problem with the mortgage loan.

20%

44%

35%

2%

127

When the error resolution rules are NOT effective in helping your clients, how often is it due to the following?

The servicer does not follow this rule.

15%

24%

56%

5%

127

The client submits the written notice to an incorrect address.

66%

19%

9%

6%

127

The client incorrectly states an error occurred.

61%

29%

5%

5%

127

The servicer conducts an investigation but does not fix any error.

20%

28%

45%

6%

127

The servicer agrees on the facts stated by the client, but does not agree that an error occurred.

24%

41%

29%

6%

127

Other problems with the error resolution requirements mean the rules do not make a difference.

38%

24%

20%

18%

127

Part E: Force-placed Insurance

TABLE 14: REASONS WHY THE FORCE-PLACED INSURANCE REQUIREMENTS ARE EFFECTIVE OR NOT EFFECTIVE

Question

Rarely or Never

Sometimes

Often or always

Don't Know or N/A

n

When the force-placed insurance requirements are effective in helping your clients, how often is it due to the following?

They cause the borrower to provide the servicer with proof of hazard insurance.

14%

37%

42%

6%

125

They cause the servicer to maintain the borrower's existing hazard insurance coverage rather than obtaining force-placed insurance.

20%

31%

42%

7%

125

They prevent the servicer from renewing an existing force-placed insurance policy because the borrower provides proof of hazard insurance before an annual forceplaced insurance policy is renewed.

18%

41%

34%

7%

125

When force-placed insurance requirements are NOT effective in helping your clients, how often is it due to the following?

The servicer does not send the force-placed insurance notice.

31%

34%

20%

15%

125

The servicer obtains force-placed insurance even though it could have maintained existing hazard insurance coverage.

18%

36%

32%

14%

125

The force-placed insurance notice is not understandable to clients who have limited English proficiency.

8%

26%

40%

26%

125

The force-placed insurance notice is not understandable to clients who understand English well.

16%

34%

38%

12%

125

The borrower is unable to obtain a hazard insurance policy or finds that the forceplaced insurance policy is the most costeffective option.

47%

33%

12%

8%

125

Other problems with force-placed insurance mean the rules do not make a difference.

26%

30%

7%

37%

125