Abstract:
Accurate identification of economic recessions in a timely fashion is a major macroeconomic challenge. The most successful early detector of recessions, the Sahm rule, relies on changes in unemployment rates, and is thus subject to measurement errors in the U.S. labor force statuses based on survey data. We propose a novel misclassification-error-adjusted Sahm recession indicator and provide empirically-based optimal threshold values. Using historical data, we show that the adjusted Sahm rule offers earlier identification of economic recessions. Based on the newly released U.S. unemployment rate in March 2020, our adjusted Sahm rule diagnoses the U.S. economy is already in recession, while the original Sahm rule does not.
