Investors often have strong opinions, either positive or negative, about the use of long duration Treasury bonds within a portfolio. While many defined benefit plans make use of long Treasury bonds for hedging purposes, we also believe that total return investors can benefit from a strategic allocation to the asset class. More broadly, we believe this asset class can and should play a role in a long-term portfolio that has significant exposure to risk assets (e.g., equities and credit). When implementing a long Treasury allocation, we would not suggest moving to that allocation in its entirety immediately, but rather over time. The balance of this paper will review the properties of long duration Treasury bonds, examine the role of fixed income in a portfolio, discuss the implications of rising interest rates and evaluate the merits of long Treasury bonds in a portfolio.
There are a number of key properties of long duration Treasury bonds that are important to keep in mind. We believe it is reasonable to think of them as (credit) risk-free, given that they are backed by the U.S. government. Said differently, we do not expect there to be any real risk of default or impairment of any U.S. Treasury issues. The cash flow that will be realized by an individual bond is known at purchase and the return one can expect is well-defined (though like all fixed income investments which mature, still subject to reinvestment risk). The long duration nature of these securities provides notable market price exposure to changes in interest rates. We recognize that this mark-to-market volatility should not be dismissed out of hand, but we believe it is much more important to consider volatility in the context of a portfolio rather than to focus on the stand-alone volatility of each asset class within a portfolio. We discuss diversification in greater detail later in this paper. At this point, we will simply note that duration risk and reinvestment risk typically offset each other over time for an ongoing investor (for instance, short-term drawdowns come with greater reinvestment yields).
In thinking about long duration Treasuries, we would first consider the various roles of high-quality fixed income in a conventional investment portfolio:
At a minimum, long duration Treasuries tend to serve each of these roles as well as or better than a broader, lower-duration fixed income portfolio (i.e., one built around the Barclays U.S. Aggregate index).
If held to maturity, long duration Treasuries do quite an effective job of preserving wealth. Obviously, over shorter time periods, this wealth preservation attribute can be somewhat compromised by the price volatility of long Treasuries. Long duration Treasuries are also one of the most liquid asset classes an investor can hold. As for diversification, we believe long duration Treasuries can fill this role in an amplified (and capital efficient) way. That is, in comparison to shorter-duration fixed income, the greater line item risk associated with the market return on long Treasuries is often desirable in the context of a portfolio for which the remaining assets are volatile on a mark-to-market basis and generally driven by economic growth, assuming that Treasuries tend to move in the opposite direction of these assets. This property may also allow for an investor to achieve the same degree of diversification that would be expected from fixed income but importantly, in a more capital efficient manner. Further, while equity-like assets typically rely on economic growth (or market expectations for growth), fixed income securities, particularly Treasuries, tend to do well in low-growth and deflationary/disinflationary environments as well as broad "risk-off" environments. While we may have certain expectations for a growth/inflation environment, a well-diversified portfolio should be constructed to recognize, to some degree, the possibility that expectations are not realized.
Interest rates in the U.S. have experienced secular declines for decades and are near historical lows. However, just as we would not extrapolate this trend going forward, we submit that we should not rely on the expectation that rates will mean revert to historically normal levels. In fact, evidenced by a significant percentage of developed country global bond markets trading at yields well below those in the U.S., it is entirely possible that U.S. rates go lower from here. U.S. interest rates' significant yield premium relative to government bonds in other developed markets, could—all else equal—increase global demand for U.S. Treasury bonds. Also consider the simpler fact that other developed countries, such as Japan, have experienced lower interest rates for many years, giving us precedent to believe the same is possible in the U.S.
All this being said, we are not suggesting that we have an expectation that rates will stay where they are or that they will fall further. We are simply suggesting that the possibility be considered. In fact, in our latest capital markets model, dated March 31, 2016, Treasury yields on average are projected to exceed implied forward rates by a healthy margin (see Figure 1). Long duration bonds do not necessarily underperform shorter duration bonds when rates rise. In order for this to happen, rates would need to rise more than those implied by the forward curve, which can be thought of as a breakeven curve for duration. Despite these lower expected returns, we still believe long Treasuries have a place in a portfolio that is largely exposed to equity-like assets because of the powerful diversification benefit investors derive from a long Treasury allocation.
Figure 1: Historical 10-Year Treasury Yield & Rocaton Forecasted Yields. Rates are shown as of March 31 of each year.
Investors who rely substantially on equity markets or other growth-related assets to generate returns in portfolios typically will allocate small amounts to investment grade fixed income. The rationale for this allocation would typically be to provide liquidity and diversification. Diversifying allocations should at a minimum protect capital when equity markets fall, but even better would be if these allocations rise in value when equity markets fall. While we often use historical correlations and returns to make a judgment about the co-movements we might expect in the future, history readily demonstrates that observed correlations can differ wildly depending on the data used.
Figure 2: Core Fixed Income and Long Treasuries rolling 3-year correlations to global equities
First, we should consider when we care most about diversification from our fixed income assets. Presumably, we are looking for the "most diversification" when the rest of our portfolio (we will continue to proxy this with global equities) is doing the most poorly. In times of modestly negative or positive equity performance, the value of diversification is arguably less meaningful. In an ideal scenario, there would be no diversification when equity performance is excessively good.
Instead, we believe investors should focus on the diversification benefits of long Treasuries in the worst global equity market scenarios. We can see from Figure 3 that in moderately poor to very good equity markets, fixed income markets in general experience a wide range of returns. And indeed, returns on long Treasuries are far more dispersed than those for broad market fixed income. However, in the worst equity markets, long Treasuries tend to do the most to offset the damage. It is also worth noting that when equity markets experienced a decline of 10% or more over rolling one-year periods back to 1988, long Treasury bonds have always delivered a positive return, often above that of the Barclays Aggregate.
Figure 3: Rolling one-year returns for fixed income versus global equities
Ultimately, we believe the diversification we tend to see in the worst equity market environments is a function of low growth/low inflation or, more broadly, "risk-off" markets where investors flee "risk assets" to what they consider to be the safest assets (e.g., Treasury bonds). Investor behavior in these cases often leads to a re-pricing of risk premiums, pushing prices down and risk premiums up for riskier assets and lowering interest rates. This is something which we have observed empirically and which we generally expect to hold going forward. Naturally, we cannot guarantee this, and we have seen market environments where, because of central bank intervention, for instance, bad news becomes good news and vice versa. However, again, in risk-off markets driven by fear, for example, both history and logic lead us to expect investors to seek out Treasury assets.
Despite the current low interest rate environment, we are supportive of a long-term strategic allocation to long duration Treasuries. We believe long Treasuries function at least on par with or often better than core fixed income in upholding what we believe to be the three key purposes of holding fixed income: wealth preservation, liquidity, and diversification. While there are times that are less attractive to make such a shift—for instance, we likely would not invest in long Treasuries when the yield curve is flat or inverted—we believe that, over time, some position in these assets should be established. We also believe that this is not a position that should be fully entered into immediately or in a very short time period. Instead, we believe now is a reasonable time to set a plan for reaching a strategic level and to adopt a plan to transition to that level over time.