Chapter

Annex II Information Content of the Yield Curve

Author(s):
International Monetary Fund. Research Dept.
Published Date:
February 1995
Share
  • ShareShare
Show Summary Details

The slope of the yield curve, summarized by the difference between current long-term and short-term interest rates, is one of many indicators used to assess economic conditions.1 It is used as a predictor of growth, inflation, and future interest rates, and it is often taken as an indicator of the stance of monetary policy. A positively sloped yield curve has been associated with an increase in output in the period ahead, and with an increase in future inflation and short-term interest rates. A negatively sloped, or “inverted,” yield curve has been taken as an indicator of future declines in these variables.

This interpretation is consistent with developments in the yield curve and in growth over the past two decades in several countries.2 In the United States, Germany, and Canada, changes in the slope of the yield curve have been followed fairly consistently by corresponding changes in the growth of output roughly one year later (Chart 25). In France, a similar pattern is evident in the 1970s and again since 1988; the relative constancy of the yield curve during most of the 1980s highlights the fact that not all changes in growth rates are preceded by changes in the yield curve. In the United Kingdom, the relationship appeared to break down in the mid-1980s, but it was quite consistent before and after the mid-1980s. In Japan and Italy, the relationship at this lag is not as apparent.3 With a longer and more variable lag—on the order of two to four years—the slope of the yield curve also appears to predict inflation in several countries.4

Chart 25.Major Industrial Countries: Yield Curve Slope and Real Growth

1 Defined as long-term interest rate minim short-term interest rate. Long-term interest rates refer to the following instruments: for the United States, yield on ten-year treasury bonds; for Japan, over-the-counter sates yield on ten-year government bonds with longest residual maturity: for Germany, yield on government bonds with maturities of nine to ten years; for France, long-term (seven- to ten-year) government bond yield (Emprunts d’Etat à long terme TME); for Italy, before June 1991, secondary market yield on fixed-coupon (BTP) government bonds with two- to four-year residual maturity, thereafter ten-year government bond rate: for the United Kingdom, yield on medium-dated (ten-year) government stock; and for Canada, average yield on government bonds with residual maturity of over ten years. Short-term interest rates refer to the following instruments: for the United States, three-month certificates of deposit (CDs) in secondary markets; for Japan, before July 1984, three-month Gensaki rate, thereafter three-month CDs; for Germany, France, and the United Kingdom, three-month interbank deposits; for Italy, three-month treasury bills; and for Canada, three-month prime corporate paper.

2 GDP data refer to west Germany through 1990 and to unified Germany thereafter.

A common explanation for the observed leading relationship between the slope of the yield curve and both output and inflation in several countries is that the slope of the yield curve is an indicator of the stance of monetary policy. The yield curve might also be expected to contain information about future economic developments because market assessments of a variety of factors in addition to monetary policy—such as fiscal policy developments, shocks to money demand or commodity markets, expected inflation or exchange rate movements—may be reflected in long-term interest rates and thus may affect the slope of the yield curve directly.

Underlying Relationships

Several fundamental relationships determine market interest rates and hence the slope of the yield curve. The first is summarized in the Fisher equation, which specifies that the nominal interest rate, at any maturity, is the sum of the real rate of interest and the rate of inflation expected over the relevant period. Underlying this equation is the observation that investors with funds to lend have the option of investing those funds in physical capital with its corresponding real rate of return. For a nominally specified loan contract to be attractive, it must offer a competitive real return plus compensation for the expected change—and the associated uncertainty—in the purchasing power of the money with which the loan will be repaid.

Market participants have alternatives besides real investment, and this provides a second, complementary way to characterize interest rate determination. This relationship is summarized in the “expectations hypothesis,” which holds that the long-term interest rate equals the geometric average of the short-term interest rates expected to prevail over the lifetime of the long-term contract. In addition, the long-term interest rate may contain a risk premium, to compensate for uncertainty regarding the future course of interest rates and policy actions, for example, as well as a term premium, to compensate investors for the liquidity lost by locking into a long-term agreement.5 The expectations hypothesis is based on arbitrage reasoning as well: an investor can enter into a long-term loan contract or can plan a sequence of consecutive short-term loans, rolling the funds over and earning the market-given short-term interest rate each time. If the returns on one of these two strategies were expected to be higher (taking account of transaction costs and risks), investors would shift funds to that strategy, and interest rates would adjust.

Country-specific factors affect the extent to which long-term interest rates reflect market expectations of future economic variables. Administrative controls on interest rates, credit controls, limitations on market access, or collusive pricing of loan rates, for example, tend to interfere with the information content of market interest rates. Differences in the importance of such obstacles across countries, or over time within a country, would tend to alter the relationships. In addition, the size of the risk and liquidity premiums in long-term interest rates may be affected by policy developments and by institutional features of financial systems. Finally, as international financial markets have become more integrated, and particularly for countries whose exchange rates are linked, economic and policy developments in other countries have significant effects on domestic interest rates.

In principle, the market-driven relationships hold for interest rates at any maturity, but in practice very short-term interest rates are overwhelmingly influenced by monetary policy actions.6 This is the case whether the central bank is directly using an interest rate as a policy target—as the U.S. Federal Reserve Board did before 1979—or is targeting a quantity variable such as the growth of a monetary aggregate through actions in the interbank reserve market, which in turn affect interest rates. Similarly, central banks achieve exchange rate objectives primarily through their control of short-term interest rates. Because the monetary authorities’ actions are the principal determinant of movements in very short-term interest rates, there is a shift in the mix of processes determining interest rates as one looks along the maturity spectrum; at the very short end, it is reasonable to view interest rate movements as being largely determined by current monetary policy; at the long end, interest rates are anchored by market participants’ expectations of inflation and exchange rates, and by the long-run real rate of return in the economy.

Interpreting Yield Curve Movements

In general, changes in the slope of the yield curve have mainly reflected policy-related changes in short-term interest rates (Chart 26). In most cases, a steepening of the yield curve has occurred when short-term interest rates have fallen, and long-term interest rates have followed suit but have decreased by less. Similarly, declines in the slope of the yield curve have typically been the result of increases in short-term interest rates that were accompanied by smaller increases in long-term interest rates.

Chart 26.Major Industrial Countries: Yield Curve Slope and Interest Rates1

(In percent)

1 See note to Chart 25.

The change in long-term interest rates that accompanies a change in short-term interest rates depends on several factors. Consider, for example, the possible effects on the yield curve of an easing of monetary policy that results in a decline of 100 basis points in short-term interest rates. If these lower short-term rates are expected to be maintained for some time, then long-term interest rates would also tend to decline, although by somewhat less. The size of the change in the long-term rate would depend on how long the change in short-term interest rates is expected to be maintained, on market perceptions of the degree of slack in the economy, and on what impact the policy change is expected to have on inflation. The resulting change in the slope of the yield curve would be different depending on market conditions and expectations and thus would provide information not contained in the policy-induced change in short-term interest rates alone.

The yield curve may also adjust in the absence of current monetary policy actions if, for example, expectations of future monetary policy were revised, or if market perceptions of fiscal pressures or other factors affecting future real interest rates changed. An example of such an episode is the steepening of the yield curve that occurred as the fiscal implications of German unification became apparent. The anticipated increase in real activity and pressure on future real interest rates raised long-term interest rates relative to short-term interest rates, causing a steepening of the yield curve without a shift in official short-term interest rates.

It is possible for movements in long-term interest rates to more than offset a policy-induced change in short-term interest rates. Thus, although an increase in official short-term interest rates typically results in a moderate flattening of the yield curve, if markets considered the tightening to be insufficient, then long-term interest rates would tend to rise by much more, resulting in a steepening of the yield curve. Alternatively, a reduction in policy-controlled short-term interest rates typically leads to a steepening of the yield curve, but if markets anticipated much larger future reductions in short-term interest rates—because inflationary pressures were considered to be very low for example, and economic slack was thought to be considerable—then long-term interest rates would fall much further than short-term interest rates, and the slope of the yield curve would decrease.

An example in which a change in the yield curve’s slope was dominated by a change in long-term interest rates is the United States in 1983-84.7 When long-term interest rates rose roughly 1¼ percentage points between May and August 1983, the Federal Reserve responded by raising the federal funds rate by 1 percentage point, leaving the yield curve slightly steeper. Long-term interest rates began to rise again in early 1984, and by midyear they were another 1½ percentage points higher. Short-term interest rates also moved higher in 1984, but the increase in long-term rates dominated these movements until midyear, and the yield curve steepened further. Taken in isolation, the steepening of the yield curve might have been interpreted as an indicator of higher growth in the following year. Growth actually declined throughout 1984 and into early 1985, however, and the increase in long-term rates probably reflected an increase in inflationary expectations. A similar episode occurred in the United States in 1987. A third example, with changes in the opposite direction, is Germany in 1976-77, when long-term interest rates declined by roughly 2 percentage points between the fourth quarter of 1976 and the first quarter of 1978, but short-term interest rates declined by just 1 percentage point. The yield curve’s slope declined, but, with roughly a one-year delay, growth began to increase.

Implications and Recent Developments

One must, of course, be cautious in drawing strong conclusions, but the analysis above suggests some interesting observations. The first is that the widely cited evidence of the predictive power of the yield curve does appear to be closely related lo the interaction between the stance of monetary policy and the yield curve. In most cases, changes in the slope of the yield curve have been driven principally by changes in short-term interest rates. In cases in which the opposite held true—when changes in long-term rates, and thus presumably changes in market expectations rather than current monetary policy actions, dominated shifts in the yield curve—the usual “predictions” of the yield curve slope for future growth have not been fulfilled. This does not imply, of course, that these yield curve movements did not contain relevant information for other economic variables. Moreover, the monetary authorities’ subsequent response may have reversed developments that would otherwise have “validated” the market expectations revealed in the yield curve’s movement. In all cases, it is clear that a range of economic indicators must be considered when interpreting changes in the yield curve.

The analysis can be usefully applied to recent movements in the yield curve in several major industrial countries where long-term interest rates have increased since early 1994. In the United States, the yield curve steepened with an increase in long-term interest rates that exceeded the rise in short-term rates. The resulting movement in the yield curve may have reflected market assessments of a strengthening recovery and building inflationary pressures, and hence expectations that the Federal Reserve would raise short-term interest rates further in the future.

Long-term interest rates have also recently increased in Europe and Japan, resulting in a sleeper yield curve. In contrast to the United States, however, ii seems unlikely that these yield curve developments reflected fears of higher inflation, in view of the large margins of slack in these economies. Instead, part of the increase is probably due to spillover effects of the increase in U.S. long-term interest rates, which made dollar bonds relatively more attractive. In Europe, there has also been a perception that the long period of falling bond yields has come to an end, and that the pace of monetary easing is likely to be slower than had earlier been expected. This shift in sentiment appears to have triggered an unwinding of positions in bond futures markets. In Japan, the 1 percentage point rise in long-term interest rates since the beginning of the year appears to have reflected the perception that increased fiscal deficits may put pressure on asset markets, as well as the liquidation of bond holdings by banks attempting to restructure their balance sheets. More recently, signs that the economic down-turn may have bottomed out, and possible spillover effects from U.S. interest rate movements, appear to have been the predominant factors.

This annex was prepared by Bas B. Bakker and Monica Hargraves.
1The yield curve refers to the plot, for a given date or time period, of interest rates ordered by the time to maturity of the underlying securities (from short to long maturities). The difference between long-term and short-term interest rates provides a simple characterization of the yield curve slope. More detailed measures of the slope at different points along the yield curve can also be informative.
2For an econometric analysis of the relationship between the slope of the yield curve and subsequent real output growth, see James H. Stock and Mark W. Watson. “New Indexes of Coincident and Leading Economic Indicators,” in Olivier Jean Blanchard and Stanley Fischer, NBER Macroeconomics Annual, 1989 (Cambridge, Massachusetts and London: MIT Press, 1989), pp. 351-94; Arturo Estrella and Gikas A. Hardouvelis, “The Term Structure as a Predictor of Real Economic Activity.” Journal of Finance, Vol. 46 (June 1991), pp. 555-76; and Zuliu Hu, “The Yield Curve and Real Activity,” Staff Papers (IMF), Vol. 40 (December 1993). pp. 781-806.
3The evidence for Italy is not fully comparable with that for the other countries because, for the sample period in question, the available long-term interest rate is for two- to four-year maturities, which is considerably shorter than the ten-year interest rates available for other countries.
4See, for instance, Frederic S. Mishkin, “The Information in the Longer-Maturity Term Structure About Future Inflation,” Quarterly Journal of Economics, Vol. 105 (1990), pp. 815-28; and Phillippe Jorion and Frederic Mishkin, “A Multicountry Comparison of Term-Structure Forecasts at Long Horizons,” Journal of Financial Economics, Vol. 29 (1991), pp. 59-80.
5For example, the average differential between long- and short-term interest rates in the United States over the past twenty-five years has been 1½ percentage points, and the steepness of the yield curve can be judged relative to this “norm.”
6This is not to suggest that market forces play no role, particularly for interest rates on private sector securities, hut rather that monetary policy actions have a large influence on all such interest rates.
7This episode has been characterized as an “inflation scare” that required a large policy response from the Federal Reserve; see Marvin Goodfriend, “Interest Rate Policy and the Inflation Scare Problem: 1979-1992,” Federal Reserve Bank of Richmond Economic Quarterly, Vol. 79 (Winter 1993), pp. 1-24.

    Other Resources Citing This Publication