On Tests of Reallocation

Scott Schuh and Robert Triest have written a useful and interesting paper, collecting existing evidence on job flows and providing us with a few brand new facts to chew on. I wish to touch on some of the many good things in the paper, but it comes with the discussant’s job description that I must focus on the parts of the paper that I view as more problematic. Few economists would deny that an ongoing process of factor reallocation is essential to the economic growth and prosperity of a market economy. The field is more divided on whether or not the churn (that is, the ongoing processes of creation and destruction) has a significant effect on the economy at business cycle frequencies. But even if we accept its relevance to the business cycle, it is a large leap to claim that reallocation shocks are a substantial source of business cycles, at least in the United States. Perhaps constrained by the shocks theme of the conference, Schuh and Triest chose to emphasize this last, most debatable claim in many passages of their paper, including the introduction. I suspect this is a mistake at this stage, when we still have plenty to learn about the two less controversial claims, and the intermediate one in particular. Indeed, many of the recent developments in the reallocation literature relate to the intermediate claim, that is, that the job reallocation process is important for business cycle considerations. Moreover, most of the survey and “testing” part of the paper is not about models that are based on reallocation shocks, but about the response of the churn to changes in business cycle conditions. I share their eagerness to go beyond this stage