Research

Abstract: 

Different series of high-frequency monetary shocks can have a correlation coefficient as low as 0.5 and the same sign in only two-thirds of observations. Both data and methods drive these differences, which are starkest when the federal funds rate is at its effective lower bound. Methods that exploit the differential responsiveness of short- and long-term asset prices can incorporate additional information. After documenting differences in monetary shocks, we explore their consequence for inference. We find that empirical estimates of monetary policy transmission from local projections and VARs are less affected by shock choice than forecast revision specifications. 

• FEDS working paper

Abstract: 

Eurodollar futures were the bedrock for constructing high-frequency series of monetary policy surprises, so their discontinuation poses a challenge for the continued empirical study of monetary policy. We propose an approach for updating the series of Gürkaynak, Sack, and Swanson (2005) and Nakamura and Steinsson (2018) with SOFR futures in place of Eurodollar futures that is conceptually and materially consistent. We recommend using SOFR futures from January 2022 onward based on regulatory developments and trading volumes. The updated series suggest that surprises over the recent tightening cycle are larger in magnitude than those seen over the decade prior and restrictive on average.

• Article

• Data

Abstract: 

We present an indicator of job loss derived from Twitter data, based on a fine-tuned neural network with transfer learning to classify if a tweet is job-loss related or not. We show that our Twitter-based measure of job loss is well-correlated with and predictive of other measures of unemployment available in the official statistics and with the added benefits of real-time availability and daily frequency. These findings are especially strong for the period of the Pandemic Recession, when our Twitter indicator continues to track job loss well but where other real-time measures like unemployment insurance claims provided an imperfect signal of job loss. Additionally, we find that our Twitter job loss indicator provides incremental information in predicting official unemployment flows in a given month beyond what weekly unemployment insurance claims offer.

• FEDS working paper

Abstract: 

The appropriate calibration of monetary policy requires an evaluation of the size and timing of the effects of policy changes on the economy.  Despite decades of research, significant uncertainty about those effects remains.  Using a large collection of structural macroeconomic models and three different monetary policy rules, we simulate the effects of monetary policy shocks on the size and timing of fluctuations in output and inflation. We then relate those effects to a set of model and non-model attributes, including the monetary policy rule employed, specific model features such as nominal and real rigidities, and other non-model features, such as whether the model is estimated or calibrated, when the model was formulated, and the backgrounds of the authors.  We find that the range of the resulting impulse responses is wide, with some models showing very rapid and large effects, while others show much more gradual or more modest effects.  Most of the persistence of the effects of monetary policy shocks stems from the inertia embedded in the policy rules rather than from propagation from elsewhere in the models.  In addition, the effects of policy are more drawn out when the model is estimated rather than calibrated and if the authors include central bank staff.  Even after accounting for these and other attributes, there remains considerable variation in the effects of policy across models.  We conclude that policymakers need to be humble about their knowledge of the effects of monetary policy and approach their policy decisions with a risk management framework.

• Current draft