A Novel Work-around to Data Availability Lags

Published on by ajfieldhouse

Data availability lags pose a significant problem for policymakers in the midst of a recession or financial crisis. The U.S. economy is surely careening into a deep recession, but we have no clue how deep the hole will be. For what it’s worth, JP Morgan recently revised their forecast to a -10% annualized contraction in the first quarter and a -40% annualized contraction in the second quarter, for a cumulative decline of ~15% for the first half of the year (take such ad hoc forecasts with a huge grain of salt). Preliminary data for U.S. real GDP in the second quarter of the year won’t be released until July 30.

Notwithstanding the fact that the last three reports showed a staggering 16.8 million unemployment claims since mid-March, these data availability lags are one reason that weekly initial unemployment insurance (UI) claims—a leading economic indicator—get so much attention during recessions.

Economists Daniel Lewis, Karel Mertens, and James Stock have an interesting new data project and related working paper, “U.S. Economic Activity During the Early Weeks of the SARS-COV-2 Outbreak,” that tries to work around data availability lags in our headline economic indicators. The authors also published a shorter, more accessible overview of their work on the New York Fed’s blog:

The authors have constructed a new Weekly Economic Index (WEI) to track real economic activity closer to real time; the index is being updated every Tuesday and Thursday. The WEI is constructed using data available at a weekly frequency or higher, including UI claims and measures of federal tax withholdings, retail sales, railroad traffic, steel production, and electricity usage, among others. The index is extracted from these high-frequency data series using a method called principal component analysis.

Scaled to U.S. real GDP growth, the WEI suggests that we experienced a collapse in U.S. economic activity by the week ending March 28 equivalent to a 6.2% annualized contraction in real GDP—if that lower level of activity persisted for a full quarter:

WEI_Fig_4

As you can see, that would be a more abrupt and far steeper rate of decline than we ever experienced during the Great Recession. Stay tuned for more updates…

Leave a Reply