The expectations hypothesis (EH) of the term structure plays an important role in the analysis of monetary policy, where shorter-term rates are assumed to be determined by the market’s expectation for the overnight federal funds rate. With two exceptions, tests using the effective federal funds rate as the short-term rate easily reject the EH. These exceptions are when the EH is tested over the nonborrowed reserve targeting period and when the test is performed only using data for settlement Wednesdays – the last day of bank reserve maintenance period. This paper argues that these exceptions are anomalous: in the former case, the failure to reject the EH occurs when economic analysis suggests that the market should be less able to forecast the federal funds rate. In the latter case, it occurs when there are sharp spikes in the funds rate that cannot improve materially the market’s ability to forecast the funds rate. Additional analysis shows that these anomalous results are a consequence of the procedure used to test the EH.

It is widely believed that the Fed controls the federal funds rate by altering the degree of pressure in the reserve market through open market operations when it changes its target for the funds rate. Recently, however, several analysts have suggested that the Fed need not conduct open market operations to change the funds rate. Rather, they argue it is sufficient that the Fed indicate its desire for the funds rate. This paper notes that there is yet a third alternative, the interest-rate-smoothing hypothesis, that suggests that the Fed does not move rates per se but, rather, smooths the transition of rates to the new equilibrium required by economic shocks. This paper tests the open market and open mouth alternatives using a methodology first used by Cook and Hahn [Journal of Monetary Economics (1989a) 331]. Finding no evidence that either open market operations or open mouth operations can account for the close relationship between the funds rate and the funds rate target, a variety of evidence consistent with the interest-rate-smoothing hypothesis is considered. The results suggest that many changes in the Fed’s funds rate target are an endogenous response to economic events and suggest that an alternative way to identify exogenous changes in policy is to identify exogenous changes in the Fed’s funds rate target.

Recently, a number of researchers (Christiano and Eichenbaum, 1992; Christiano et al., 1996a,b, 1997; Evans and Marshall, 1998; Strongin, 1995; Pagan and Robertson, 1995; Brunner, 1994) claim to have found evidence of a statistically significant liquidity effect in a recursive structural VAR using nonborrowed reserves (NBR). It is claimed that innovations to NBR reflect the exogenous policy actions of the Fed. This paper argues that the opposite is true. Specifically, I show that the Fed has an incentive to offset bank-initiated discount window borrowing when it implements the Federal Open Market Committee’s policy directive, and that it has done so since the late 1950s. This practice has created a negative contemporaneous covariance between NBR and the funds rate that has been incorrectly attributed to the liquidity effect. By showing that these models capture the endogenous response of the Fed to bank borrowing on NBR, rather than the effect of exogenous policy actions on the funds rate, this paper also resolves the puzzle of the vanishing liquidity effect noted by Pagan and Robertson (1995) and Christiano (1995).

A considerable volume of research shows that asset prices respond to changes in the Federal Reserve's discount rate. While several competing hypotheses have been advanced to explain the market's response to discount rate announcements, comparatively little effort has been made to differentiate among alternative hypotheses. The result is an abundance of evidence establishing that asset prices respond to discount rate announcements, but little if any agreement about why markets respond. This article attempts to fill a void in the literature by pointing out how competing hypotheses differ and by constructing tests explicitly designed to differentiate among competing explanations. The evidence suggests that the market's reaction to discount rate changes is purely an announcement effect, i.e., a reaction to new information contained in the announcement, that the direct effect of discount rate changes on market rates is nil, that the announcement effect is invariant to the Federal Reserve's operating procedure and that, generally speaking, changes in the discount rate do not signal a change in monetary policy. The announcement effect appears to vary with both the nature and extent of the information that the announcement of a discount rate change is believed to contain.

Daniel L. Thornton, Maximum Likelihood Estimates of a Partial Adjustment-Adaptive Expectations Model of the Demand for Money, The Review of Economics and Statistics, Vol. 64, No. 2 (May, 1982), pp. 325-329

[This article investigates the out-of-sample predictability of bond excess returns. We assess the economic value of the forecasting ability of empirical models based on longterm forward interest rates in a dynamic asset allocation strategy. The results show that the information content of forward rates does not generate systematic economic value to investors. Indeed, these models do not outperform the no-predictability benchmark. Furthermore, their relative performance deteriorates over time.]

One of the most influential tests of the expectations hypothesis is Mankiw and Miron [Q. J. Econ. 101 (1986) 211], who found that the spread between the long-term and short-term rates provided predictive power for the short-term rate before the Fed's founding but not after. They suggested that the failure of the expectations hypothesis after the Fed's founding was due to the Fed's practice of smoothing short-term interest rates. We show that their finding that the expectations hypothesis fares better prior to the Fed's founding is due to the fact that the test they employ tends to generate results that are more favorable to the expectations hypothesis during periods when there is extreme volatility in the short-term rate.

This article investigates the out-of-sample predictability of bond excess returns. We assess the economic value of the forecasting ability of empirical models based on long-term forward interest rates in a dynamic asset allocation strategy. The results show that the information content of forward rates does not generate systematic economic value to investors. Indeed, these models do not outperform the no-predictability benchmark. Furthermore, their relative performance deteriorates over time.

This article examines the dynamic relationship between two key US money market interest rates––the federal funds rate (FF) and the 3-month Treasury bill rate. Using daily data over the period from 1974 to 1999, we find a long-run relationship between these two rates that is remarkably stable across monetary policy regimes of interest rate and monetary aggregate targeting. Employing a nonlinear asymmetric vector equilibrium correction model, which is novel in this context, we find that most of the adjustment toward the long-run equilibrium occurs through the FF. In turn, there is strong evidence for the existence of significant asymmetries and nonlinearities in interest rate dynamics that have implications for the conventional view of interest rate behavior.