Labor market in March deteriorated much more quickly than the unemployment rate suggests

Today’s employment report, a snapshot taken in the second week of March before mass stay-at-home orders, showed that businesses reacted surprisingly quickly to cut employment in response to the novel coronavirus.  In March, employment dropped by 700,000, and the unemployment rate shot up to 4.4 percent from 3.5 percent.  Less widely noted, the labor force participation rate declined by 0.7 percentage point to 62.7 percent—already close to its nadir in the Great Recession.  This sharp decline in the labor force participation rate last month suggests that the unemployment rate is significantly understating the deterioration in the economy.

The labor force participation rate typically moves in a predictable relationship with the unemployment rate. For instance, research I have done with co-authors suggests that a sustained 1 percentage point increase in the unemployment rate would be expected to decrease the participation rate by ¼ – ½ percentage point. The two send a similar message about slack in the labor market.  In a sense the unemployment rate is what we call a “sufficient statistic” for slack.  However, if the labor force participation rate moves in an unusual way, then that movement contains additional information.  In March, the labor force declined much more than would be expected given the rise in the unemployment rate, suggesting that there is considerably more slack in the labor market than would be suggested by the unemployment rate alone.

Every month workers leave employment, either because they are fired or voluntarily.  Some become unemployed and some drop out of the labor market.  (The former tell government survey takers that they are actively looking for work; the latter say they haven’t looked for work in the past couple of weeks. ) In March, a much larger than usual share of those who left employment dropped out of the labor market.  In fact, if we look at individual labor market transitions from February to March we see that 5.8 million individuals left employment and moved out of the labor force entirely [1]. This represents an increase of 1.2 million people dropping out of the labor force relative to the average in prior months.  If these individuals had followed more typical patterns, more of them would have transitioned into unemployment, and the unemployment rate last month would stand at 5.2 percent.

Obviously, we know from the unemployment insurance claims data that the unemployment rate is likely already above 10 percent, so the point here isn’t about measuring the unemployment rate in February.  Instead, these data are important because they suggest the potential magnitude by which the unemployment rate is understating slack.  Consistent with the view that slack has increased beyond that suggested by the headline unemployment rate, the number of people out of the labor force (and thus not counted in the headline unemployment rate) who want a job rose by 500,000.  And even among those who kept their jobs, the number of individuals working part-time for economic reasons (those who had their hours cut due to business conditions) shot up.

It makes sense that individuals would choose to drop out of the labor force rather than remain unemployed at the current time.  To be counted as unemployed in the household survey you have to say you were actively looking for work—and right now governments are calling on individuals to stay home in order to avoid spreading the virus.  In recognition of this reality, many states have changed the rules governing their unemployment insurance programs so that individuals are not required to be looking for work in order to collect benefits.  Thus, this policy change reduces the incentive that individuals eligible for those benefits have to look for work. Instead, these individuals might just drop out of the labor force.  The result again, is that the rise in the unemployment rate is not completely capturing the decline in work and economic activity more broadly.

Over the longer term, the rise in the share of individuals staying out of the labor force could have implications for the recovery.  The unemployed, on average, are the most likely to find a job.  Individuals who are out of the labor force but want a job are also quite likely to become employed, although they do so at lower rates than the unemployed–they tend to behave similarly to the long-term unemployed.  And, even many of those who say they do not want a job (and are not disabled or retired) do transition back into the labor force, although they do so at much lower rates.  Individuals who are out of the labor force make decisions about their time and finances that could change their propensity to work, at least in the near-term. For instance people could opt to take care of family members or reduce their spending on housing.  This is likely one reason why the pickup in the labor force participation rate typically lags that of the unemployment rate.  Indeed, this dynamic could be seen during the recovery from the Great Recession, particularly around 2014, when the rise in the labor force participation rate lagged what would have been expected given the improvement in the economy. All in all, the decrease in the labor force participation rate suggests that social distancing increases detachment from the labor force. This detachment could slow re-entry to employment when the economy does begin to open up.

[1] The labor force status numbers presented in the Employment Situation represent stocks, which usually show small changes from month to month.  However, the small changes in stocks elide a huge amount of churning in the labor market each month as millions of individuals find or change jobs and move in and out of the labor force.  Each month in conjunction with the Employment Situation, the BLS also releases data on individual transitions between employment, unemployment, and nonemployment.