How Important is the Shape of the Yield Curve in Managing Operating Assets?

by David Hertan, CFA, Managing Director, Commonfund

• Investing in longer maturities in virtually any yield curve environment should allow an investor to capture excess returns over shorter maturities.
  
• There is no value in trying to “beat the market” by extending out the curve when it is steep and being more defensive when it is flat or inverted.
  
• Active management of treasury portfolios based solely on the shape of the yield curve is unlikely to add value over the long term

Inevitably, when speaking with organizations about ways to more effectively invest operating assets, the discussion turns to the shape of the yield curve. Many investors in short-duration assets believe the curve is a critical determinant of fixed income returns. And, such investors can mistakenly view active management narrowly – solely driven by the shape of the curve. In fact, a “flat” yield curve is often cited as a reason not to invest in longer term fixed income securities. Undoubtedly it is easier for an investor to lengthen maturity in a fixed income portfolio when the curve is steep – increasing yield is an attractive proposition. However, our analysis suggests that investors should not limit themselves to extending maturity only in a steep yield curve environment. Investing in longer maturities in virtually any environment should allow an investor to capture excess returns over shorter maturities.

Greenspan’s Conundrum

The dynamics of the yield curve have received considerable attention subsequent to the pronounced flattening of the curve after the Fed began to raise short-term interest rates in June 2004. This unanticipated market behavior was famously labeled a “conundrum” by former Federal Reserve Chairman Alan Greenspan in his testimony to Congress in early 2005. The most recent thinking regarding interest rate and yield curve theory is that bond yields may be separated into two components: 1) an expected rate component that reflects the short-term interest rates that are expected to prevail during the maturity of the bond; and 2) a term premium that reflects the uncertainty of knowing what those future short-term rates will actually be. Both components fluctuate. While rate expectations have moved both up and down, studies have suggested that the term premium has been in a declining trend for the past two decades. This narrowing of the term premium may be due to less volatile economic and inflationary conditions as well as supply and demand factors in the fixed income market. It is difficult to precisely explain behavior of the yield curve because neither component of bond yields nor their determinants are readily and directly observable.

Commonfund Analysis

A critical issue for fixed income investors is whether the shape of the curve provides any information that may help generate superior returns. Commonfund recently did a statistical analysis of the relationship between the steepness of the yield curve and monthly returns in the fixed income market. Two common measurements were used to describe the shape of the yield curve:

• 3-month Treasury Bills versus 10-year Treasury Notes
• 2-year Treasury Notes versus 10-year Treasury Notes

To capture absolute and relative returns across a range of maturities, we evaluated five different measures of fixed income performance:

• Return of 3-month T-bills
• Return of the Merrill Lynch 1-3 year Treasury Index
• Return of the Lehman Aggregate Bond Index
• Excess return of the Merrill Lynch 1-3 year index over 3-month T-bills
• Excess return of the Lehman Aggregate index over 3-month T-bills

Our analysis looked at the period that included all available data and the period beginning in January 1988. We are inclined to focus on the later period since it generally reflects Fed policy and economic volatility similar to what we might expect in the future. A summary of our findings is shown below:

2-year/10-year Yield Curve
	ML 3-month T-Bill Auction Average	ML 1-3 Year Treasury Index	Lehman Aggregate Bond Index	Excess Returns of ML 1-3 Year	Excess Returns of Lehman Aggregate
1/78 to 12/06 (all available data)
R^2	51.60%	1.07%	0.05%	1.53%	1.86%
Beta	-0.22	-0.10	0.04	0.12	0.27
t-stat	-19.21	-1.94	0.41	2.32	2.56
1/88 to 12/06
R^2	57.13%	2.21%	0.16%	1.12%	0.52%
Beta	-0.15	-0.09	-0.05	0.06	0.10
t-stat	-17.35	-2.26	-0.60	1.60	1.08
3-month/10-year Yield Curve
	ML 3-month T-Bill Auction Average	ML 1-3 Year Treasury Index	Lehman Aggregate Bond Index	Excess Returns of ML 1-3 Year	Excess Returns of Lehman Aggregate
1/82 to 12/06 (all available data)
R^2	8.39%	0.01%	0.06%	0.92%	0.51%
Beta	-0.05	0.00	0.03	0.05	0.08
t-stat	-5.22	-0.15	0.43	1.66	1.24
1/88 to 12/06
R^2	42.90%	3.29%	0.54%	0.13%	0.05%
Beta	-0.09	-0.08	-0.07	0.01	0.02
t-stat	-13.03	-2.77	-1.10	0.55	0.35

Key Findings

The steepness of the yield curve has a negative correlation with the return on 3-month T-bills as shown by its negative beta. Its beta is statistically significant in each of the four scenarios; and, when we look at the 2-year/10-year curve, this negative correlation explains over 50 percent of the variation in return (R-squared). The relationship makes intuitive sense since an inverted yield curve would imply high short-term rates and suggest a high return on T-bills. For the same reason, it is not surprising that the absolute return on the Merrill Lynch 1-3 year Treasury index is also negatively correlated with steepness of the curve but, although significant, the explanatory power of this relationship is limited.

Looking at excess returns, or returns above those available on T-bills, provides some perspective on the opportunity to benefit from changing investment strategy based on the steepness of the yield curve. There is very limited evidence of a relationship between the steepness of the curve and the excess returns that have historically been available in fixed income indices. The relationship is only significant in one of the four data sets and explanatory power is minimal in all four. We also did regression analysis using various moving averages and lagged data with similar results. The chart below gives a graphical illustration of the relationship between the 2-year/10-year curve and excess returns of the Lehman Aggregate.

Looking at Different Interest Rate Environments

Although the steepness of the yield curve has not been very useful in predicting excess returns, we were curious as to whether extending out the curve in any particular environment produced excess returns. Accordingly, we divided the data into quartiles based on the steepness of the curve and looked at excess returns earned during periods when the yield curve was “Very Flat”, “Flat”, “Steep” or “Very Steep”. Mean excess returns, as shown in the table below, were all positive and, although they exhibited some differences, there was no clear pattern.