Just how tight is the labor market? It depends on which measurement you pick!
Some economists look to the unemployment rate, or to its difference from some estimated, but not measured, “natural level” to try to assess whether the labor market and economy are in equilibrium (we are not fans of this approach). Others use the vacancy-to-unemployment ratio as a summary of the demand vs. supply balance of the labor market (we are also not fans of this approach).
Still others, us included, have relied on the quit rate from the Job Openings and Labor Turnover Survey (JOLTS), citing its high predictive power in explaining wage growth. The intuition for this metric is that a tight labor market is one where good jobs are easily found, and already-employed workers either job-hop or feel comfortable quitting with the confidence that they’ll be able to find a new job.
The JOLTS measures quit rates by asking establishments how many employees they had and how many employees quit in a given month. We can measure this same concept in a different way using the Current Population Survey (CPS) to see how many employed people left their job in the previous month to another job, unemployment, or left the labor force. The response rate of JOLTS has fallen greatly since 2020, so it is good practice to get another perspective on labor market churn.
CPS, unlike JOLTS, allows us to distinguish between quits to other employers or quits to unemployment or out of the labor force. As Mike Konczal pointed out, an increase in employer-to-employer (EE) transitions is usually a positive for the economy. More EE transitions generally mean that the labor market is more dynamic as more workers move to better, higher-paying matches.
In the graph below, I plot the JOLTS quit rate against the employer-to-employer transition rates (as reported by Fujita, Moscarini and Postel-Vinay (2023), who do the heroic work of adjusting the CPS data for changes in survey methodology) and a CPS analog of the JOLTS quit rate. I construct the CPS quit rate analog following Moscarini and Postel-Vinay (2022) by summing the employer-to-employer transition rate, the transition rate of the employed into unemployed job leavers, and the transition rate of the employed out of the labor force (except for into retirement and disability, which are supposed to be excluded from JOLTS quits).
A few things stand out. First, the increase in the JOLTS quit rate during the COVID recovery was far more dramatic than the increase in either the EE transition rate or the CPS quit rate. Second, the EE transition rate is essentially back to pre-COVID levels, while total quit rates (using either CPS or JOLTS) show a somewhat tighter labor market.
Moscarini and Postel-Vinay (2022) present a labor market slack measure which they call the “acceptance ratio”, which is the ratio of the EE transition rate and the unemployment to employment (UE) ratio, which measures labor market misallocation. The intuition behind this measure is that labor demand raises both the UE and EE transition rate, but misallocation leads to higher job market churn as workers find higher productivity jobs faster. Thus, the higher the acceptance ratio, the more misallocation there is in the labor market.
Through this lens, the labor market was characterized by an increase in misallocation (and a corresponding increase in labor market churn) in 2021. Crucially, Moscarini and Postel-Vinay (2022) argue that job changes due to this type of churn actually moderate inflationary labor cost pressures, since this churn leads to workers finding better matches. This bump in labor market churn has since faded, and has returned to pre-COVID levels.
In sum, the CPS data on EE transitions have three implications. First, by this metric, the post-COVID labor market was not as tight as other metrics, such as the traditional JOLTS quit rate, unemployment, or the vacancy-to-unemployment ratio suggested. Second, part of the labor market recovery in 2021 was due to workers finding better jobs. To the extent that this represents workers moving to higher productivity jobs, this was not nor should have been interpreted as inflationary. Third, measures of labor market churn, (the acceptance ratio and the EE transition rate) are back to normal pre-COVID levels.