A Model of Monetary Policy Shocks for Financial Crises and Normal Conditions

Abstract

In their classic 1999 paper, Monetary policy shocks: What have we learned and to what end?, Christiano, Eichenbaum, and Evans (CEE) investigate one of the most widely used methods for identifying monetary policy shocks of its time. Unfortunately, their approach is no longer viable, at least not in its original form. A major problem stems from the recent behavior of two key variables in their model, the Fed Funds rate and non-borrowed reserves. We develop a new identification scheme that remedies these difficulties but maintains the basic CEE framework. Our empirical specification is motivated by a standard New Keynesian DSGE model augmented by a simple financial structure. The model provides theoretical support for variables we use in place of certain variables that were used in the classic VAR approach outlined in CEE. One significant innovation is our use of Divisia M4, the broadest monetary aggregate currently available for the United States, as the policy indicator variable. We obtain four major empirical results that support the use of a properly measured broad monetary aggregate as the policy variable. First, policy shocks have significant effects on output and on the price level, even when an interest rate is included in our model-contradicting the New-Keynesian argument that monetary aggregates are redundant. Second, we develop a model that is not subject to the output, price or liquidity puzzles common to this literaturecontradicting the view that using the interest rate as the policy indicator generally yields more reasonable responses than a monetary aggregate. Third, during normal conditions policy shocks from our Divisia-based model have similar effects on variables to those found in the Fed Funds model of monetary policy, and where there are differences our model with Divisia M4 obtains results that are more consistent with standard economic theory. Fourth, our preferred specification produces plausible responses to a monetary policy shock in samples that include or exclude the recent financial crisis.

A Classical View of the Business Cycle

“…Bernanke (2006) discusses the historical and intellectual trends that caused the Fed to gradually deemphasize the monetary aggregates in its policymaking process, beginning in the early 1980s. The results presented here and in Keating et al (2019), however, reveal that valuable information is contained in Divisia monetary aggregates, both about monetary policy and its effects on the economy, even in data samples including the 1980s and running through the present. Perhaps, after a long hiatus, research focusing the classical channels of transmission discussed by Fisher and Working can contribute importantly to the design of effective monetary policy strategies.…”

“…can be interpreted as an expanded Taylor (1993) Rule that includes the contemporaneous rate of nominal money growth, as well as inflation and the output gap, on the short list of variables to which the Federal Reserve is assumed to respond in setting its target for the federal funds rate. 17 Alternatively, this monetary policy rule can be seen as taking the same general form as those used by Sims (1986), Leeper and Roush (2003), Leeper and Zha (2003), Sims and Zha (2006), Belongia and Ireland (2015b, 2016a, and Keating et al (2019) to identify, within structural vector autoregressions, monetary policy shocks based on the contemporaneous effects those policy disturbances are allowed to have on both the federal funds rate and the rate of money growth. According to this interpretation, (10) associates a contractionary monetary 17.…”

Money and Output: Friedman and Schwartz Revisited

“…Three other recent papers take different, but complementary, approaches to modeling, statistically, the links between Divisia money, output, and prices. Keating et al (2014) experiment with various triangular identification schemes in VARs that include Divisia money.…”