Did COVID-19 induce a reallocation wave?

We critically examine the hypothesis that COVID-19 has ushered in a large reallocation shock in the USA, beyond typical business cycle patterns. We take a broad perspective, and first consider data from the CPS and JOLTS; there is no noticeable uptick in occupation or sector switches, either at the aggregate level or in the cross-section. The dispersion of sectoral growth rates over the three years before the pandemic was similar to the previous period. The recovery from the initial shock was characterized by very high quits and low layoffs, patterns indicative of a strong labour market, not excessively high reallocation relative to previous business cycles. High growth of small employers in the recovery, and larger ones once the labour market tightened, is also a common cyclical pattern. We then examine whether mismatch unemployment rose as a result of the pandemic; using an off-the-shelf multisector search and matching model, there is little evidence for an important role for mismatch in driving the unemployment rate during the pandemic. Finally, we employ a novel Bayesian Structural Vector Autoregression framework with sign restrictions to identify a reallocation shock; we find that it has played a relatively minor role in explaining labour market patterns in the pandemic, at least relative to earlier episodes.