The brevity and violence of contractions and expansions

Early studies of business cycles argued that contractions in economic activity were briefer (shorter) and more violent (rapid) than expansions. This paper systematically investigates this claim and in the process discovers a robust new business cycle fact: contractions in employment are briefer and more violent than expansions but we cannot reject the null of equal brevity and violence for expansions and contractions in output. The difference arises because employment typically lags output around peaks but they coincide in their troughs. We discuss the performance of existing business cycle models in accounting for this fact, and conclude that none can fully account for it. We then show that a business cycle model with asymmetric adjustment costs on employment and a choice of when to scrap old technologies can account for the business cycle fact both qualitatively and quantitatively.