Deconstructing Monetary Policy Surprises—The Role of Information Shocks

This 2020 American Economic Journal: Macroeconomics paper by Marek Jarociński and Peter Karadi argues that conventional high-frequency-identified monetary policy surprises conflate two economically distinct shocks — a genuine monetary policy shock and a “central bank information shock” — and shows that separating them substantially changes conclusions about how powerfully monetary policy affects the economy. The key identifying insight is that a pure monetary policy tightening should raise interest rates while lowering stock prices (the standard asset-pricing prediction), whereas a central bank information shock — in which the central bank’s own announcement conveys good news about the economic outlook that partly offsets a simultaneous tightening — should raise both; using a Bayesian structural VAR combining high-frequency surprises (three-month fed funds futures and S&P 500 changes around 240 FOMC announcements, 1990-2016) with sign restrictions to disentangle the two, the authors find that around one-third of FOMC announcements historically show this “wrong-signed” positive interest-rate/stock-price co-movement. The purified monetary policy shock produces a more persistent decline in output and prices and a rise in the excess bond premium, while the information shock raises both output and prices and lowers the excess bond premium — and because these two shocks move macro variables in opposite directions, the paper shows that the conventional (unpurified) high-frequency-identified shock, which implicitly attributes all surprises to monetary policy, systematically understates the true effectiveness of monetary policy and generates spuriously large and persistent interest-rate responses. Structurally estimating a New Keynesian model with financial frictions to match the two sets of impulse responses, the authors find the conventional (contaminated) shock requires implausibly extreme price stickiness and negligible financial frictions to fit the data — essentially reproducing Nakamura and Steinsson’s (2018) puzzle — whereas the purified monetary policy shock is consistent with more moderate, empirically plausible price stickiness and substantially larger financial frictions, leading the authors to conclude that failing to control for central bank information shocks can seriously distort inferences about the transmission mechanism, including the perceived importance of financial frictions.

Monetary Policy Surprises, Credit Costs, and Economic Activity

This 2015 American Economic Journal: Macroeconomics paper by Mark Gertler and Peter Karadi uses a “proxy SVAR” — a structural VAR identified with external instruments rather than timing restrictions — to show that modest, high-frequency-identified surprises in monetary policy produce disproportionately large movements in private credit costs, driven mainly by term premia and credit spreads rather than by revisions in the expected path of short-term rates. Using surprises in federal funds and Eurodollar futures around FOMC announcements (1991-2012) as instruments for a monthly VAR (1979-2012) in which the one-year government bond rate serves as the policy indicator, the authors find that a monetary tightening that raises the one-year rate by about 20 basis points produces a roughly 15 basis-point increase in corporate bond rates and a roughly 7 basis-point increase in mortgage rates, an 8 basis-point rise in the Gilchrist-Zakrajšek excess bond premium persisting for about eight months, and a significant, fairly rapid decline in industrial production reaching its trough after about 18 months — while the response of the CPI is not statistically significant and virtually all the effect on real activity operates through real rather than nominal rates. Decomposing the response of longer-term rates, the paper finds that for two-, five-, and ten-year maturities “virtually all” of the rate increase is due to the term premium rather than the expected path of short rates, and that credit spreads and term premia together (the “excess premium”) account for essentially the entire increase in private borrowing costs — a pattern the authors argue is inconsistent with the standard frictionless monetary transmission mechanism and instead consistent with a credit channel operating alongside limited participation or limits to arbitrage in longer-term bond markets. A key auxiliary finding is that forward guidance matters: substituting the federal funds rate for the one-year rate as the policy indicator, normalized to the same funds-rate movement, produces an output contraction “more than 50 percent smaller,” indicating that surprises which revise expectations about the future path of policy have substantially stronger real and financial effects than surprises confined to the current policy rate.

Peter Karadi | Macro Paper Warehouse

Deconstructing Monetary Policy Surprises—The Role of Information Shocks

Monetary Policy Surprises, Credit Costs, and Economic Activity