Research

Asset-price surprises around FOMC announcements reflect four distinct shocks: monetary policy, central bank information, the Fed's systematic response to public news, and the risk premium. Existing high-frequency methods identify at most three at a time. I identify all four jointly from the responses of four asset prices in a narrow window around each announcement, using sign and magnitude restrictions derived from a simple asset-pricing model. Over 1991–2023, the risk premium shock dominates the long end of the yield curve: it accounts for about 54 percent of the variance of the ten-year yield surprise and is the largest contributor in every specification I consider. Omitting it, or the Fed's response to news, contaminates conventional measures — roughly a quarter of the news-response shock is misattributed to monetary policy, and about a tenth of the risk premium shock to central bank information. What moves long-term interest rates on announcement days is therefore mostly not news about policy, but a change in the compensation investors require for bearing risk. 

The dots in the Federal Reserve's Summary of Economic Projections (SEP) reveal that FOMC members disagree about the appropriate policy rate, but not what they disagree about. Using members' individual projections from 2012 to 2019, I estimate each member's policy reaction function and decompose the cross-member dispersion in the dot plot into differences in economic outlook and differences in reaction functions. The composition of this disagreement shifts systematically with the forecast horizon: at the nearest horizons it reflects mainly differences in how members would respond to a given outlook, at intermediate horizons differences in the outlook itself, and at the longest horizons differences in members' perceived long-run neutral rate. Reaction-function differences therefore dominate at both ends of the horizon spectrum, including the near-term dots most relevant for the coming decision. These differences are large enough to move the implied rate path by 25 to 75 basis points, are systematically related to observable characteristics such as experienced inflation and educational background, and help predict dissenting votes. The individual dots thus carry behaviorally meaningful information about how members would set policy that the median projection discards.

I investigate whether the Fed pays attention to the forecast’s discrepancies with the market. Using the quarterly data sample 1987-2007, I find that the Fed reacts to the GDP growth discrepancy with the Survey of Professional Forecasters (SPF). Specifically, the Fed is more inclined to follow the market if disagreement in market forecasts is smaller. However, it seems not to be the case with forecast discrepancies of the GDP deflator. I show that positive tail risks, approximated by the skewness of SPF individual forecasts, are important factors for the Fed to follow the market. One percentage point increase in the quarterly GDP deflator nowcast difference between the Fed and the market is associated with a 0.4 p.p. lower Fed Fund Rate if the market distribution is symmetric, whereas the effect is halved if the distribution is substantially skewed to the right (skewness = +2). However, the rapid changes in economic conditions during the GFC presumably contributed to the more mixed results obtained from the bigger sample (1987-2017).

This paper examines the impact of monetary policy on financial and macroeconomic variables in Russia. We distinguish between two types of monetary policy shocks: (1) those driven by changes in current policy rates, and (2) those driven by other factors such as forward guidance, communication, or central bank information. We find that these two types of shocks have distinct effects on financial variables. The first type primarily explains movements in interest rates across the yield curve, while the second type better accounts for fluctuations in the exchange rate and stock market indices. Moreover, we show that monetary policy transmission from interest rates to inflation operates with a lag of about one year, and that this effect is only temporary. 

Russian monetary policy may spill over to the member countries of the Eurasian Economic Union (EAEU) through several channels, but evidence on the significance of these effects remains scarce. This paper estimates the influence of Russian monetary policy shocks, proxied by shocks to the Moscow Interbank Actual Credit Rate, on the EAEU economies. Monetary policy shocks are first identified from a FAVAR model of the Russian economy, estimated on monthly data for more than 50 indicators. These shocks are then used in separate VAR models for each member country, and both impulse response functions and forecast error variance decompositions are analyzed. Transmission appears to operate through financial variables rather than the real economy: the shocks significantly affect money supply, nominal exchange rates, and money market rates in some member countries, while their effects on industrial production and inflation are not statistically significant. The estimated effects are nonetheless mostly small and heterogeneous.