As the US economy recovers in fits and starts, market and policymaker attention is turning to the exit strategy. How will the Fed exit from its loose monetary policy? In particular, how will it winding down its second bout of quantitative easing, known universally as QE2?
The speed of the exit strategy is likely to hinge in part on the amount of inertia inherent in US monetary policymaking process, i.e. the speed at which policy adjusts to incoming information. Some anecdotal evidence points towards significant inertia, such as the following quote from the Federal Open Market Committee (FOMC) minutes of 15 October 2010:
In their discussion of the relative merits of smaller and more frequent adjustments versus larger and less frequent adjustments …, [FOMC] participants generally agreed that large adjustments had been appropriate when economic activity was declining sharply in response to the financial crisis. In current circumstances, however, most saw advantages to a more incremental approach that would involve smaller changes … calibrated to incoming data. (FOMC 2010)
Because the Federal Reserve has used non-traditional monetary policy tools in response to the financial crisis, there are few directly comparable historical precedents to assess the likely speed of the exit strategy. Nevertheless, we can still use the available historical evidence on interest rate decisions to determine how much policy inertia is typically present in the Fed’s decision-making process.
The source of persistence in Fed decision-making has critical implications for how we might forecast future monetary policy. The sources, however, are subject to significant controversy in the research literature.
In new research (Coibion and Gorodnichenko 2011), we shed light on this using a variety of methods – all of which point to very significant levels of policy inertia. In short, we should not expect rapid policy changes in the near future – barring clear signs of economic distress.
The analytical framework
Since Taylor (1993) macroeconomists have relied on simple interest rate reaction functions to characterise the endogenous response of monetary policymakers to economic fluctuations. Our own baseline formula for predicting monetary policymakers’ desired interest rate is an extension of the classic “Taylor rule”; it looks at the central bank’s forecast of inflation, the growth rate of output, and the output gap. Our rule departs from the classic Taylor specification in that it allows for responses to both the output gap and the growth rate of output and also in that it allows for the central bank to respond to the forecast of future macroeconomic variables consistent with the notion that monetary policy changes take time to affect the economy so policymakers should be forward-looking in their policy decisions.
The left panel of Figure 1 plots the predicted interest rate from this Taylor rule estimated using the forecasts made by the staff of the Federal Reserve prior to each FOMC meeting (the Greenbooks) relative to the actual interest rate set by the FOMC over most of the Greenspan era.
As emphasised by Taylor (1993), a simple specification such as this can account for much of the policy changes over this time period. However, the predictions of the Taylor rule are noticeably more volatile than actual interest rates. The average size of the predicted change in interest rates (in absolute value) is approximately 60% larger than actual quarterly changes in interest rates (57 basis points to 35 basis points). Actual interest rates are also significantly more persistent than predicted interest rates (autocorrelations are 0.98 versus 0.93) and the residuals are positively serially correlated.
To account for this difference between the behaviour of actual interest rates and those predicted from baseline Taylor rules, two explanations have been suggested.
- The first and most common interpretation is policy inertia.
That is, policymakers do not set interest rates equal to the desired rate each period but rather move interest rates in a sequence of steps towards the desired interest rate (Clarida et al. 2000).
This is commonly modelled as “interest rate smoothing”. Applying such formulas to our baseline formula, we find a very high levels of interest smoothing (see paper for details). The assumption that interest rate persistence primarily reflects policy inertia has dominated both empirical and theoretical macroeconomics research.
The second interpretation is that the observed serial correlation in policy rates reflects persistent monetary policy shocks (Rudebusch 2002, 2006). This can be modelled by assuming the errors in our baseline formula are serially correlated.
The right panel of Figure 1 shows the fitted values of the Taylor rule under the two interpretations of the Fed’s behaviour – a policy inertia and persistent shocks – are essentially indistinguishable to the naked eye (see paper for more precise comparisons and technical details).
Importantly, the two interpretations have very different implications for understanding the determination of monetary policy.
Differentiating between policy inertia and persistent shocks
In Coibion and Gorodnichenko (2011), we provide robust evidence that policy inertia is a more likely source of the persistence in interest rates than the persistent shocks hypothesis. The key pieces of evidence are:
- Nested specifications
We show that once we allow for sufficiently general specifications of both interest smoothing and persistent shocks (i.e. more than first-order processes), specifications of the Taylor rule which include both policy inertia and persistent shocks strongly favour the inertial policy interpretation.
- Conditional identification
A key difference between the two explanations is that, under policy inertia, the gradual adjustment of interest rates should occur irrespective of the underlying source of economic fluctuations, whereas the alternative points to additional persistence only after monetary policy shocks. Using exogenous shocks to identify innovations to the Fed’s forecasts of future macroeconomic conditions that are not driven by monetary policy shocks, we continue to find high estimated levels of interest smoothing, consistent with the policy inertia interpretation.
- Interest Rate Predictability
Rudebusch (2002) argues that if policy inertia was important, then future interest rate changes should be quite predictable, yet he documents that financial futures markets fail to predict future interest rate changes beyond the one-quarter horizon. However, the inability of financial markets to predict future interest rate changes could also reflect uncertainty about the policy rule or more limited information about the economy than what is available to the Fed. Consistent with this, we find that the assumptions about future interest rates made by the staff of the Fed for each set of Greenbooks do a significantly better job of predicting future interest rate changes than private sector forecasts.
Was the Fed responding to other factors?
Rudebusch (2002) also suggests that serially correlated shocks in the Taylor rule should really be interpreted as the result of the Fed responding to time-specific concerns not controlled for in the Taylor rule. While we can never completely rule out omitted variables, we consider estimates of the Taylor rule augmented with a variety of measures of some of the more likely candidates for omitted variables.
- Financial market factors
The most likely sources of policy actions not directly tied to output and inflation are credit conditions and financial considerations. However, when we include measures of credit spreads, stock prices, and uncertainty in the Taylor rule, these are generally insignificant and do not qualitatively affect the estimated degree of interest smoothing.
- Real-time forecast revisions
Another omission from the baseline Taylor rule which could potentially and misleadingly lead to the appearance of policy inertia is the importance of data lags and revisions of Fed forecasts about the current state. When we augment the Taylor rule to control for changes in the Fed’s forecast of macroeconomic conditions, we find no evidence of significant Fed responses to these measures and the estimates of policy inertia continue to be high.
- Private sector forecasts
We also consider the possibility that the Fed responds not just to its forecasts of future macroeconomic conditions but also those of the private sector. This could occur if the central bank is unsure about the quality of its forecasts when they differ from those of other agents or if the central bank is concerned about the effect of its policy decisions on the expectations of other agents. We find statistically significant evidence that the Federal Reserve does respond to deviations of its forecasts from those of the private sector, but controlling for these differences eliminates the evidence of persistent shocks while leaving the estimate of policy inertia unchanged.
The historical decisions of the Federal Reserve with respect to interest rates, at least during the Greenspan period, consistently point to very significant inertia in the policymaking process. To the extent that this translates to the exit strategy and the non-interest rate tools used by the Fed during the current crisis, our results suggest that the policy reversal is likely to be gradual, in the absence of additional significant economic shocks.
Clarida, Richard, Jordi Galí, and Mark Gertler (2000), “Monetary Policy Rules and Macroeconomic Stability: Evidence and Some Theory”, Quarterly Journal of Economics 115(1), 147-180.
Coibion, Olivier, and Yuriy Gorodnichenko (2011), “Why are target interest rate changes so persistent?”, SSRN Working Paper 1738512.
FOMC (2010), “Minutes of the Federal Open Market Committee”, October 15 and November 2-3.
Rudebusch, Glenn D (2002), “Term structure evidence on interest rate smoothing and monetary policy inertia”, Journal of Monetary Economics, 49(6):1161-1187.
Rudebusch, Glenn D (2006), “Monetary Policy Inertia: Fact or Fiction?”, International Journal of Central Banking, 2(4):85-135.
Taylor, John B (1993), “Discretion versus Policy Rules in Practice”, Carnegie Rochester Conference Series on Public Policy, 39:195-214.