Head of U.S. Fixed Income Greg Wilensky and Portfolio Manager Michael Keough discuss the challenges in forecasting interest rates, particularly in the wake of major market shocks like COVID-19.

Key Takeaways

  • Forecasting the direction, magnitude and volatility of interest rate changes is difficult.
  • Maintaining humility in those forecasts, coupled with a diversity of risks, can potentially improve returns.
  • We believe thinking in terms of “excess” returns and “empirical” durations can also help.

Computers in the future may weigh no more than 1.5 tons."                      - Popular Mechanics magazine, 1949

Popular Mechanics was absolutely right. They correctly – perhaps brilliantly, considering the silicon chip wasn’t invented for another decade – predicted that the size and weight of computers would fall in the future. But the extent to which they were wrong about the magnitude of that change is breathtaking.

No matter how right you may be on the “what,” there is so much room to be wrong on the “how.” To put it in terms that many investors are thinking about today, no matter how confident you may be that interest rates will rise as the U.S. economy recovers, there is still much you can be wrong about.

Old classic computer sitting on an art deco retro desk. The wall is covered in a wallpaper with a striped wallpaper.Had an investor correctly predicted a global pandemic would erupt in the first quarter of 2020, we can imagine they would have sold stocks and bought government bonds before it started. But one could argue that without a detailed map of the markets’ exact paths, the knowledge that a pandemic was coming would have been less valuable than you might think. When would the investor have decided to reverse their more cautious position? What is the probability they covered their position a little or a lot, too early or too late and thus lost as much money in late March as they made in early March?

Correctly predicting the general direction of financial markets far into the future is certainly helpful but gauging the timing and magnitude of the moves is just as important and often more difficult. Despite the shock of COVID-19, many experienced portfolio managers were able to not only navigate the volatility, but also find opportunities to add value for their clients, whether through lower volatility relative to their benchmark, higher returns, or both. Indeed, the task of portfolio management is, in our view, to build portfolios that are designed to perform well given both our predictions and the potential uncertainty that surrounds them and the world we live in.

The Value of Humility

As good as any portfolio manager might be, surprises will happen. That is not to say that some managers aren’t better than others at predicting the direction of certain sectors, industries, or even companies. But humility is important.

The fact that a major shock like COVID-19 recently happened to the market may make it seem less likely that something of that magnitude will happen again soon. But the probability of a shock in any given year is largely independent of whether a shock has recently occurred. The probability that markets will experience a shock is best assumed to be a constant, if small, possibility.

Given that these low-probability events (sometimes called “black swans” given their rarity) can have an outsized impact on returns, a humble investor who recognizes the challenge of forecasting the next black swan should, in our view, prudently build portfolios that help keep the downside risk from one of these events within acceptable limits.

The Value of Diversity

To paraphrase the old saying, we don’t believe you should put all your eggs into one forecast. A portfolio should be filled not just with a diversity of sectors and securities, but a diversity of risks. Even if the probability of being right in any one forecast is high, the probability of a portfolio performing well over time will be higher if there are more forecasts that are as independent of each other as possible.

To give you a sense of the math, if you allocated your portfolio’s risk across 10 different market forecasts that were reasonably uncorrelated to each other, each having a good chance of being right, your portfolio would likely result in a better risk-adjusted return than if you had all your risk allocated to one forecast that had a higher chance of being right.

Managing interest rate risk is one valuable, and important, tool in portfolio positioning, but we think it’s also critical to incorporate humility when budgeting the amount of risk allocated to positioning based on one’s interest rate expectations, so as not to place too much confidence in the accuracy of just one forecast.

The Shortcomings of Standard Duration Measures

While there is a general understanding that non-Treasury securities exhibit less interest rate sensitivity than Treasury securities with the same duration, many investors do not systematically account for this in their investment decisions. The standard calculation for duration provides the expected return for a change in the investment’s yield, but this mathematical calculation can be unhelpfully simplistic.

Consider the figure below, which shows the historical relationship between changes in the 5-year U.S. Treasury bond yield and the return of the investment-grade corporate bond index. When Treasury yields rise, corporate bond returns have tended to fall. However, the sensitivity of the relationship does not match the mathematical duration of the corporate bond index. Instead, it has empirically behaved like it has a significantly lower duration. The average mathematical duration of the corporate bond index over the period shown was near 7.4 years. However, the slope of the line graphed shows the sensitivity to be closer to 4.7 years – a significant difference, with potentially significant effects on returns.1

U.S. Treasury Yield Changes vs. Corporate Bond Returns

Source: Bloomberg, Janus Henderson, as of 31 March 2021. Corporate Bond returns are represented by the Bloomberg Barclays U.S. Corporate Bond Index, which measures the U.S. dollar-denominated, investment-grade, fixed-rate corporate bond market.

 

Even if an investor properly forecasts where 10-year Treasury yields will go, knowing how a portfolio will behave empirically is crucial to delivering the intended outcomes. We believe that taking a systematic approach to estimating and using “empirical durations” will lead to better risk-adjusted performance. Failure to do so could lead to underperformance even if your interest rate forecast is accurate.

Thinking About Excess Returns

Rate forecasts aside, we believe investors should always be looking to investments with the potential to outperform similar duration U.S. Treasuries. The technical term for this is “excess return” insofar as these investments generate a return in excess of the return generated by Treasuries.

Corporate bonds, for example, pay a yield higher than Treasuries. As long as the spread (the difference in those yields) remains constant, owning corporate bonds will result in an excess return over Treasuries because of this spread. Of course the spreads can widen, which hurts the excess return, or tighten, which improves the excess return. Whether the spread expands or contracts can be correlated to the direction of interest rates but it can also be influenced by other factors, such as the conditions and outlook for the economy as a whole, a particular industry or a specific company.

Ultimately, both corporate bonds and the wide range of securitized assets available in today’s bond market typically provide higher yields than U.S. Treasuries while providing some diversification. Treasury Inflation-Protected Securities (TIPS) and floating-rate investments like many commercial mortgage-backed securities (CMBS) and collateralized loan obligations (CLOs), can provide even more diverse profiles of interest rate risk and income.

The Logical Conclusion

Volatility in markets is not a bad thing. The volatility in 2020, coupled with the interventions by the U.S. Federal Reserve and historic fiscal stimulus, created opportunities for active managers to express strong convictions. There is, in our view, nothing wrong with feeling confident about a forecast. The trick is to make sure the resulting positions are sized appropriately within an overall risk budget.

Just as diversifying risks and holdings is key within core fixed income portfolios, we believe investors should be mindful of that same diversity across their overall portfolio of investments. The ultimate mix is, ideally, a result of individual goals, risk tolerance and investment views. Given a strong conviction that rates will rise sharply, investors may consider either reducing overall fixed income exposure or making adjustments within fixed income allocations to increase exposure to strategies with lower interest rate sensitivity, including shorter-duration or higher-income strategies.

It is also important to remember that, in markets with a wide range of outcomes, allocating too much of your risk budget to a single forecast, such as the direction of interest rates, can result in increased risks. A little humility and diversification in portfolio construction can go a long way.

Fixed Income Perspectives

Connecting you to the latest thinking from our fixed income teams.

Explore Now

 

1Source: Janus Henderson Investors, as of 31 March 2021. In the regression equation y=mx+b, m is the slope of the line, in this case, showing the relationship between changes in Treasury yields and corporate bonds returns. The equation for the regression graphed is y=-4.7x+0.36. Thus, the corporate bonds returns have exhibited a duration of 4.7 years.

Collateralized Loan Obligations (CLOs) are debt securities issued in different tranches, with varying degrees of risk, and backed by an underlying portfolio consisting primarily of below investment grade corporate loans. The return of principal is not guaranteed, and prices may decline if payments are not made timely or credit strength weakens. CLOs are subject to liquidity risk, interest rate risk, credit risk, call risk and the risk of default of the underlying assets.