In order to measure the flexibility of the labor market, evaluate the job-worker matching process, and model business-cycle dynamics, economists have studied the flows of workers across the labor market states of employment, unemployment, and not in the labor force. One important flow that has been poorly measured is the movement of workers from one employer to another without any significant intervening period of nonemployment. This paper exploits the "dependent interviewing" techniques used in the Current Population Survey since 1994 to estimate such flows. We find that they are large, and their omission significantly understates the degree of mobility in the labor market. In 1999, for example, on average more than 4,000,000 workers changed employers from one month to the next, about the same number as left the labor force from employment and more than twice the number that moved from employment to unemployment. Close to half of the new jobs started in 1999 represented employer changes, as did close to half of the separations. Consistent with previous studies of younger workers, teenagers exhibit the highest rates of employer-switching, and the rate declines through about age 40. However, even among prime-aged workers, about 2 percent change employers each month. Contrary to the implications of many business cycle models, we find no evidence that employer-to-employer flows are procyclical, at least not as the labor market tightened between 1994 and 2000. This finding raises questions about the ways in which stylized facts about labor market flows have been used.
