Nowcasting the Output Gap

Title: Nowcasting the Output Gap

Authors: Tino Berger, James Morley, and Benjamin Wong

Register to attend

Series format:

60 minutes Zoom presentation

30 minutes of Q&A

Co-authors can answer clarification questions posed in the Q&A throughout the presentation. A moderator will collect the more substantive questions to be answered in the Q&A session. During this time, participants can also raise their hand and pose live questions. Everybody will be muted, and all interactions with the speaker are going through the chat with the moderator.

Abstract

We propose a method to directly nowcast the output gap using the Beveridge-Nelson decomposition based on a Bayesian mixed-frequency VAR. The mixed-frequency VAR produces very similar estimates of the U.S. output gap in comparison to a quarterly VAR, but is able to provide timely updates to estimates within a quarter as higher-frequency data become available. We find that monthly indicators for consumer sentiment, initial claims, employment growth, and the unemployment rate all contain useful information about the output gap, with initial claims particularly relevant for tracking large changes in the output gap before current quarter real GDP is released. An out-of-sample analysis of the COVID-19 crisis suggests a massive drop in the nowcast for the U.S. output gap in 2020Q1 from 3.2% to -6.7% with the release of March data and our scenario forecasts suggest the output gap could fall further to between -14.6% and -15.9% in 2020Q2.