by Ruben C. Trevino, Ph.D., and Barbara M. Yates, Ph.D.
Executive Summary
- The objective of this paper is to examine the historical performance of Treasury bills relative to bonds and stocks under different inflationary conditions.
- Over the period studied, Treasury bill yields were positively correlated with inflation and on average provided nominal returns before taxes above the inflation rate. However, on an estimated after-tax basis, the real returns became negative and thus did not preserve an investor’s capital.
- Furthermore, the long-term average real return on Treasury bills was lower than the average real return on bonds and stocks, especially on an estimated after-tax basis. Thus, stocks and bonds provided better long-run protection against inflation.
- However, Treasury bills performed better than bonds and stocks in years when inflation was very high. Treasury bill returns were highly correlated with inflation levels on a year-to-year basis, while bond and stock returns were not.
- Similarly, Treasury bills performed better than bonds and stocks in years when inflation increases were large. Stocks and bonds were adversely affected by increases in inflation, while Treasury bill returns were not.
- Historical experience suggests that over the long run Treasury bills are not a good hedge against inflation. However, because stocks and bonds are more adversely affected by high inflation levels and/or large increases in inflation, Treasury bills could temporarily serve as a refuge.
Ruben C. Trevino, Ph.D., is an associate professor of finance at Seattle University, teaching in the area of investments and financial planning. He is a former portfolio manager for Colonial Management Associates in Boston.
Barbara M. Yates, Ph.D., is a professor of economics and chair of the department of economics at Seattle University. She served as director of the master of science in finance program.
Recent economic conditions have drawn many to seek out less risky and more liquid assets, such as U.S. Treasury bills. At the same time, large projected government deficits and massive increases in the monetary base have caused many to express concerns about future inflation. Thus, a look at how Treasury bills have fared historically relative to long-term bonds and stocks under different inflationary conditions could be relevant and pertinent.
Treasury bills are considered poor investments over the long run in periods of inflation as they have provided returns barely above the inflation rate. However, over short periods, stocks and bonds could be more adversely affected than Treasury bills when inflation is high or increasing. If this is the case, Treasury bills might be an attractive temporary alternative.
The primary objective of this paper is to examine the performance of Treasury bills, bonds, and stocks under different inflationary conditions. We address three specific questions: (1) Do Treasury bills protect us against inflation over the long run? (2) Which asset class, Treasury bills, bonds, or stocks, performs best during individual years of high inflation? (3) Which asset class, Treasury bills, bonds, or stocks, performs best during individual years with inflation increases?
Our analysis of the data for the 1954–2007 period shows that Treasury bill returns are highly correlated with inflation levels, but stock and bond returns are not. Over the long run, of the three main financial asset classes, Treasury bills provided the lowest, though still positive, average real return before taxes. However, Treasury bills generally performed better than stocks and bonds in years when inflation was very high or when increases in inflation were large. We begin with a brief review of the existing literature on inflation and then present the empirical analysis of the performance of Treasury bills, bonds, and stocks as it relates to inflation.
Review of Literature
According to the theory of interest rates attributed to Irving Fisher (1930), nominal risk-free interest rates should be equal to the expected inflation rate plus a real rate of return. If this is the case and real returns are somewhat constant, Treasury bill rates will move closely with inflation rates and constitute, to a certain degree, an inflation hedge. The Fisher hypothesis has been extensively studied in the economics and finance literature. The studies look at both cross-country data and time series data for individual countries. While the results vary across countries, time, and the measure of inflation expectations used, there is in general, strong support for an empirical relationship between nominal interest rates and expected inflation. Most studies report a stronger correlation over long periods than over short periods, suggesting a lag exists in adjustments of nominal interest rates to inflation expectations. For a sample of this literature see Berument and Jelassi (2002), Cooray (2003), Crowder and Hoffman (1996), and Fahmy and Kandil (2002).
Inflation expectations should have a similar effect on long-term bond nominal interest rates. As inflation expectations increase, bond yields to maturity must increase to compensate investors. However, given their constant coupon payments and par values, higher bond yields can only be achieved by a downward adjustment of current bond prices. This drop in the price results in lower realized rates of return over the short run. Empirical studies do support this negative relationship between increases in inflation and lower bond returns. See, for example, Smirlock (1986). Thus, long-term bonds could be poor performers over short periods when inflation increases.
The relationship between inflation and stock returns is more complicated. Some economists argue that companies, for the most part, will be able to pass cost increases to the consumer and ensure that profits are preserved during inflationary times. If this is the case, stock prices would not suffer with increases in inflation. However, as is true in the case of bonds, when inflation increases, the expected nominal returns on stocks should increase, and for this to happen, current stock prices need to adjust down. Most studies document a negative correlation between inflation increases and stock returns over short periods. Thus, at least for the short run, stocks could be poor performers during inflationary periods. For a sample of this literature, see Cambell and Vuolteenaho (2004), DeFina (1991), Gultekin (1983), and Sharpe (2002).
The current study extends this literature by examining the performance under different inflationary conditions of these three main asset classes simultaneously. The emphasis is on the relative performance. Also, we concentrate on one-year returns and emphasize the short term. Our focus is on the practical implications for the investor, rather than testing particular theories, such as the Fisher hypothesis.
Data
Our study uses data for the period between 1954 and 2007, a period that includes significant variability in interest rates and inflation. For interest rates on Treasury bills we used the yields to maturity on 90-day bills at the end of each year. This data came from the Federal Reserve. Treasury bill returns, long-term Treasury bond returns, large company stock returns, and the inflation rate were obtained from the 2008 Ibbotson SBBI Classic Yearbook. Treasury bill returns are computed by rolling over four consecutive 90-day Treasury bills. The Treasury bond returns series is based on a one-bond portfolio with a maturity of approximately 20 years and a coupon close to the one on a recently issued par bond. The large company stock returns series is computed using the Standard and Poor’s Composite 500 Index. This index includes 500 of the largest stocks in terms of market value in the United States. Dividends are assumed to be reinvested.
The annual percentage change in the CPI (Consumer Price Index for All Urban Consumers, not seasonally adjusted) was used as the measure of inflation. Annual real returns were computed using the formula: real returns = [(1 + nominal returns) / (1 + inflation rate)] – 1. This is the formula used by Ibbotson to convert nominal returns into real returns. Because we were interested in how the investor actually fared, we used actual inflation rather than expected inflation.
Empirical Analysis
For Treasury bills to constitute an effective inflation hedge, interest rates on Treasury bills must increase when the rate of inflation increases. In addition, the yields must be above the rate of inflation and provide a positive real return. Figure 1 depicts the history of annual Treasury bill yields and inflation rates over the past 54 years. The strong co-movement of these two series is evident. The correlation coefficient is a high 0.79.
Treasury bills and inflation over the long run. Most of the time, the Treasury bill yield series was above the inflation series, resulting in positive average real returns. In fact, the average nominal return on Treasury bills was 5.17 percent and the average inflation rate was 3.92 percent, resulting in a compounded average real return of 1.23 percent. However, these real returns compare poorly with the real returns on bonds and stocks of 3.02 percent and 8.76 percent, respectively. So, these results confirm what is generally known that even though Treasury bills provide a positive real return before taxes, their returns are lower than the real returns of bonds and stocks. (See Table 1.)
Incorporating the effect of taxes into the long-term average returns will make the performance of Treasury bills even worse in absolute and in relative terms. The interest on Treasury bills is considered current income for tax purposes and is therefore taxed at the ordinary income tax rate. Similarly, most of the return from bonds comes from coupon interest, which is also taxed as ordinary income. Stock returns, on the other hand, receive a more favorable tax treatment. Their returns are composed of capital gains and dividends. Unrealized capital gains are untaxed, while realized long-term capital gains have traditionally been taxed at much lower tax rates. Dividends were taxed as ordinary income for many years, but have received more favorable tax treatment recently.
Tax rules and rates have changed substantially over time, and marginal tax rates on income depend on the individual investor’s income tax bracket. Thus, the effect of taxes can only be approximated. For simplicity, we will illustrate the possible effect of taxes on the absolute and relative returns by assuming that the returns of Treasury bills and bonds are taxed every year at the tax rate of 28 percent. In contrast, we assume that the tax rate on stock returns will be only 15 percent. The after-tax results are shown in Table 1. We can see that the after-tax real average return on Treasury bills is small, but negative, at –0.19 percent. Treasury bills over the long run no longer preserve an investor’s capital. In contrast, Treasury bonds still have a positive, although small, after-tax real average return of 1.01 percent. Stocks are the best long-term performers with an after-tax real average return of 6.74 percent.
The empirical evidence on the relationship between Treasury bill yields and inflation suggests that they might provide protection against inflation over the long run. However, once taxes are considered, their real returns are likely to turn negative. Furthermore, the returns on average are lower than those of bonds and stocks. In the next two sections, we examine in greater depth the short-term inflation protection properties of Treasury bills in comparison to those of bonds and stocks. The question here is whether an investor would be better off staying in Treasury bills or temporarily switching to Treasury bills from bonds or stocks during individual years when inflation is unusually high and/or rising.
The performance of Treasury bills, bonds, and stocks during years of different inflation levels. For descriptive purposes, we grouped the years by the inflation level and computed average annual returns, both real and nominal, for Treasury bills, long-term bonds, and stocks. The results are shown in Table 2, Part A. The inflation rate categories were chosen to provide even ranges around the long-term average for the 54-year period of 3.92 percent, while also ensuring a reasonable number of observations in each category. For low levels of inflation between 0 and 2 percent, stock returns were higher than the returns on Treasury bonds and bills by a large margin. Stocks continued to do better than bonds and bills for inflation levels between 2 and 6 percent, but by a smaller margin. All three assets yielded positive average real returns for years when inflation remained below 6 percent.
However, for years when inflation was above 6 percent, the average nominal return for Treasury bills was 8.38 percent, while the average nominal return for bonds was 0.84 percent and for stocks was 2.97 percent. The corresponding averages of real returns were –0.61 percent for bills, –7.48 percent for bonds, and –5.60 percent for stocks. Therefore, even though Treasury bills didn’t earn positive real returns during high inflation years, Treasury bonds and stocks actually did worse. Given the size of the average returns of Treasury bills compared to bonds and stocks, the outperformance will still hold even after taxes are considered.
In Part B of Table 2 the relationship between nominal returns and inflation is analyzed more formally using regression techniques. The regression results indicate that there was a strong positive correlation between Treasury bill annual nominal returns and the inflation rate. The slope coefficient was a positive 0.68 and the intercept was 2.50. The slope coefficient was different from zero at a statistical significance of 1 percent. The R-squared value of 51 percent indicates a reasonably good fit.
For Treasury bonds and stocks the relationships between nominal returns and inflation was much weaker. In the case of bonds, the slope coefficient was insignificantly different from zero and the R-squared was only 3 percent. For stocks, the slope coefficient was negative and significant at the 5 percent level. The R-squared was only 8 percent. These results support the hypothesis that Treasury bill returns correlate positively with inflation, and thus they provide a certain degree of short-term inflation protection. Such was not the case for bonds and stocks. Their nominal returns do not seem to be positively correlated with inflation levels, and for years with very high inflation levels, they provided lower short-term returns than Treasury bills.
The performance of Treasury bills, bonds, and stocks during years of different inflation changes. Investors are interested not only in how different assets perform at different levels of inflation but they would also like to know how they perform during years of different inflation changes. Therefore, we estimate the sensitivity of Treasury bill, bond, and stock returns to changes in inflation. In Table 3, Part A, annual nominal and real returns for the three assets are grouped by changes in inflation. As shown in the table, for inflation increases of more than 1 percent, average real returns were negative for all of the three asset classes. However, the average real returns were worse for bonds (–5.42 percent) and stocks (–4.66 percent) than for bills (–0.45 percent). Over the short run, bond and stock returns were more negatively affected by large inflation increases than Treasury bill returns. Given the size of returns on Treasury bills compared to those on bonds and stocks, the outperformance would still exist even after taxes are considered.
Table 3, Part B shows the results of the more formal regressions of nominal returns on changes in inflation. The findings substantiate the inflation risk associated with bonds and stocks. In both cases, the slope coefficients were negative and significantly different from zero at the 1 percent level. The slope coefficients estimate the sensitivity of Treasury bill, bond, and stock returns to changes in inflation. The results show that if inflation rises by 1 percent over a year, bond returns will tend to decrease on average by 2.34 percent and stock returns will tend to decrease on average by 3.88 percent in that same year. Stock returns are slightly more sensitive to changes in inflation than bond returns, but in both cases, returns are negatively related to increases in inflation. The R-squared values were 19 percent and 20 percent, respectively.
There was no statistically significant relationship between Treasury bill returns and inflation changes. The R-squared was only 1 percent. Treasury bill nominal returns are more highly correlated with the level of inflation and less so with whether last year’s inflation rate was higher or lower than this year’s. For example, if the inflation rate moves from 12 to 13 percent, the nominal returns will tend to be close to 13 percent, but if the inflation rate moves from 2 to 3 percent, the Treasury bill nominal returns will tend to be around 3 percent. Thus, Treasury bill nominal returns reflect inflation levels more than changes in inflation.
Practical Implications for Financial Planning
As part of the financial planning process, investors must decide on their most appropriate asset allocation, typically among stock, bonds, and Treasury bills. Their appropriate asset allocation depends on factors such as the overall goal, the investment horizon, risk tolerance, etc. A young professional, for example, might choose to invest her 401(k) portfolio as follows: 70 percent stocks, 20 percent bonds, and 10 percent Treasury bills. The 10 percent allocation to Treasury bills could satisfy the need for liquidity and will reduce the overall risk of the portfolio. However, any amount invested in Treasury bills will earn lower returns over the long run and will be subject to reinvestment risk. Our findings indicate that the returns on Treasury bills will barely compensate for inflation on a pre-tax basis.
However, if a period of rising inflation is anticipated, what should an investor do? The empirical evidence presented above indicates that stocks and bonds don’t fare well under these conditions. The investor has two options: stay the course (70 percent, 20 percent, and 10 percent) or move part of her investments in stocks and bonds temporarily into Treasury bills. Moving into Treasury bills increases reinvestment risk, but reduces the overall portfolio risk. Once rising inflation is no longer a concern, the investor can return to the target asset allocation. Treasury bills might be a relatively safe place to be during the inflationary storm.
Conclusions
Over the long run Treasury bill yields have generally moved with inflation. Most of the time, Treasury bill yields have been higher than inflation, providing investors with a positive average real return. For the 1954–2007 period used in this study, the average real return was a positive 1.20 percent. However, on an estimated after-tax basis, the average real returns became negative and thus did not preserve an investor’s capital.
Furthermore, the average real return on Treasury bills was lower than the average real return on bonds and stocks. This was especially true on an estimated after-tax basis. Over the full 1954–2007 period, bonds and stocks clearly provided better inflation protection, corroborating conventional wisdom.
However, there were years when Treasury bills did relatively better. Specifically, in years when inflation levels were high, the average real returns on bonds and stocks tended to be lower than those on Treasury bills. This is a consequence of the fact that Treasury bill nominal returns move more closely with inflation levels.
Similarly, we found that in years of large inflation increases, Treasury bills performed better than stocks and bonds. The real returns on stocks and bonds were more negatively affected by inflation increases.
In sum, history shows that in the long run stocks and bonds do better than Treasury bills in protecting investors against the effects of inflation. Nonetheless, long-term investors can take some comfort in knowing that any portfolio allocation to Treasury bills for safety or liquidity considerations will tend to keep up with inflation and provide a positive, if small, pre-tax average real return. However, on an after-tax basis, the long-run average real return is likely to be slightly negative.
However, there are instances when Treasury bills can temporarily do better than stocks and bonds. Our study shows that in years with large inflation increases or high inflation levels, stocks and bonds do worse than Treasury bills. In this relative sense, Treasury bills can offer a temporary refuge from the damaging effects of inflation.
References
Berument, Hakan, and Mohamed Mehdi Jelassi. 2002. “The Fisher Hypothesis: A Multi-Country Analysis.” Applied Economics 34: 1645–1635.
Campbell, John Y., and Tuomo Vuolteenaho. 2004. “Inflation Illusion and Stock Prices.” The American Economic Review 94, 2 (May): 19–23.
Cooray, Arusha. 2003. “The Fisher Effect: A Survey.” The Singapore Economic Review 48, 2: 135–150.
Crowder, William J., and Dennis L. Hoffman. 1996. “The Long-Run Relationship Between Nominal Interest Rates and Inflation: The Fisher Equation Revisited.” Journal of Money, Credit, and Banking 28, 1 (February): 102–118.
DeFina, R. 1991. “Does Inflation Depress the Stock Market?” Business Review (Federal Reserve Bank of Philadelphia) (November): 1–10.
Fahmy, Yasser A.F., and Magda Kandil. 2003. “The Fisher Effect: New Evidence and Implications.” International Review of Economics and Finance 12, 4: 451–465.
Fisher, Irving. 1930. The Theory of Interest. New York: Macmillan.
Gultekin, N. Bulent. 1983. “Stock Market Returns and Inflation: Evidence from Other Countries.” The Journal of Finance 38, 1 (March): 49–65.
Ibbotson SBBI. 2008. Classic Yearbook. Morningstar Inc.
Sharpe, Steven A. 2002. “Reexamining Stock Valuation and Inflation: The Implications of Analysts’ Earnings Forecasts.” The Review of Economics and Statistics 84, 4 (November): 632–648.
Smirlock, Michael. 1986. “Inflation Announcements and Financial Market Reaction: Evidence from the Long-Term Bond Market.” The Review of Economics and Statistics 68, 2 (May): 329–333.
If you have any questions or concerns regarding this article, please send comments to JFPfeedback@fpanet.org
