Riding the Bill Curve

Abstract

W ith a positively sloped U.S. Treasury bill curve, an investor receives additional yield for holding a bill with extended maturity. This additional yield is compensation for the additional risk of the longer security or the market's implicit forecast of a rise in interest rates. Investors who seek to profit from the tendency for yields to fall relative to this forecast as bills age are said to be pursuing a strategy of "riding the yield curve." We examine the effectiveness of this strategy using a comprehensive sample of U.S. Treasury bills.The strategy is as simple as it is well-known. Suppose an investor has a three-month holding period and considers two potential strategies. First, buy a three-month bill and mature it. Second, buy a sixmonth bill (or longer), and sell it after three months. If the yield curve is upward-sloping and does not change over the next three months, the six-month bill will earn a higher return because of the increase in price due to the decrease in yield relative to the forecast at which it is priced. McEnally [1981] refers to this effect as the price change due to maturity shift. Accordingly, investors will collect an additional return.For example, suppose that a 91-day bill and a 182day bill are yielding 5% and 5.25% on a discount basis, respectively (5.06% and 5.39% as money market yields). Buying and maturing the 91-day bill will generate a 91day return of 1.28%. Buying the 182-day bill and selling after 91 days will generate a 1.43% return over the same 91-day period if the yield curve remains unchanged. ROBIN GRIEVES is with HSBCSecurities in New York (NY 10005).

Do Traders Benefit from Riding the T-Bill Yield Curve?

“…Accordingly, this uncertainty motivated our reexamination of the T-bill trading strategies of Grieves et al [1999] and Bieri and Chincarini [2005] to determine if higher returns are available using transaction data.…”

“…9 For these same days, we also collected daily price/yield information from the Federal Reserve Bank of St. Louis FRED database for 91-day and 182-day T-bills (see the following discussion). Grieves et al [1999] used daily closing prices from HSBC to represent the prevailing market price at which participants could have transacted. 10 To provide comparability with Grieves et al we used daily 91-day and 182-day T-bill yields from FRED.…”

“…11 Our sample period is January 4, 2001, through September 27, 2007, which corresponds to the dates for which GovPX data is available. Because Grieves et al [1999] analyzed data from January 2, 1984, through April 20, 1997, we were able to determine if the higher returns observed in previous studies persist into a more recent time period. 12 The sample period is interesting because of the economic shocks and alternative interest rate environments it spans.…”

“…Instead, our study is an ex post analysis of the returns available from riding the yield curve. Grieves et al [1999] calculated returns to 91-day rides on a Thursday-to-Thursday basis because newly issued T-bills become available on Thursdays. We followed their approach and conducted our primary analysis using Thursday price/yield data, followed by robustness checks using non-Thursday data.…”

“…2 Using three-and six-month constant-maturity rates from the Federal Reserve, Bieri and Chincarini [2005] found strong and consistent excess returns from riding the U.S. T-bill yield curve. Like other studies, Grieves et al [1999] used end-of-day T-bill prices from dealer quotes. From 1987 through 1994, they found that every riding strategy they examined provided higher returns on average.…”