The pandemic has been a continuing nightmare for parents.This has been particularly true for mothers. Even before the pandemic, child care duties fell disproportionately on women, and this disparity has only grown.But figuring out how many mothers have been pushed out of the labor force specifically because of school closings can be tricky.Back in March and April, most government surveys didn’t ask households directly about school closings. The closings happened suddenly and broadly across the country, at the same time the labor market was collapsing from the pandemic. And the downturn has weighed more heavily on female-dominated sectors like leisure and hospitality anyway, regardless of whether workers were parents.So how many parents have exited work because their children’s schools have closed? Here’s our best guess and how we arrived at it.(For the purposes of this article, “parent” means anyone with school-age children, from 5 to 17.)There are 1.2 million parents out of the labor force since February.A simple measure is to look at the change in parental labor force participation since February, just before the pandemic. (This is a measure of anyone with a job, actively looking for a job, or expecting to be called back from furlough.)Between February and September, participation for mothers declined by 3.3 percentage points, adjusting for normal seasonal variations. For fathers, it declined by 1.3 percentage points. That’s the equivalent of 900,000 fewer mothers and 300,000 fewer fathers in the work force over the last seven months.But there might be many reasons that a parent might leave the labor force unrelated to school closures. So an alternative is to look at parents who say they aren’t in the labor force because of home or family concerns. Parents who passed on job searches because of school closings might list that as a reason in labor force surveys. Because this is just a subset of all people out of the labor force, the numbers are smaller: Home/family nonparticipation rose by 2.5 percentage points (700,000 people) for mothers of school-age children, and 0.6 percentage points (150,000) for fathers, between February and September.But if our goal is to pin down the labor-force effects of school closings specifically, these numbers are not entirely satisfying. They may be too high if they include broader pandemic effects on the parental labor force. But they may be too low if the parental labor force would have, absent the coronavirus, grown since February (as it had been doing in the years before the pandemic), or if affected parents gave a reason other than “home/family care” for exiting the labor force.More schools are open than back in the spring.One key measure that helps us is the variation in how many schools are closed. Back in April and May, they were closed across the country, so it was impossible to distinguish labor market outcomes for parents by the prevalence of closed schools.By September, sharp differences had emerged. The Census Bureau’s Household Pulse Survey, a new high-frequency survey of how households are coping with the pandemic, found that in early May almost 100 percent of South Dakota parents with children in school reported that classes had either moved to remote learning or been canceled entirely. In early September, that share was just 36 percent. By contrast, the numbers in Hawaii barely changed between May and September.School closing rates alone won’t solve our mystery, however — they’re probably strongly correlated with other coronavirus-related factors, like public health restrictions and anxieties, that would affect nonparents as well.So the big question is how much more sensitive parents are to school closings than, say, nonparents who are identical in other regards. A cursory look at state-level averages suggests that the differences are meaningful here. In states with higher school closure rates, mothers of school-age children tended to see slower job growth than women without children over the summer of 2020 versus the same span in 2019.But to get an even more precise read, I used the Current Population Survey — a monthly survey of households that is the source for many official labor market measures, including the unemployment rate — to examine individuals rather than states. I looked at the same people over time: those who were in the survey over eight different months, June through September in both 2019 and 2020. This allowed a more precise comparison, because I could adjust for the unique characteristics of each individual, that person’s past labor market behavior, as well as some state-level differences, including demographics and industry mix.Analyzing the data this way suggests that a 10 percent rise in the school closing rate in September was associated with summer labor force growth that was 1.5 percentage points lower for mothers than for childless women. Nationally, that’s the equivalent of 1.6 million fewer mothers in the labor force in September. All of this difference was because of home/family care. By contrast, fathers’ labor force growth wasn’t any more sensitive to school closures than it was for childless men.A crucial point is that the pandemic is not over yet, so these estimates can’t tell us what the impact on the labor market would be if schools that are currently closed reopened immediately. Reopening schools unsafely might have muted or even negative impacts on the parental labor force, to say nothing of risks to the children themselves, since doing so might lead to higher caseloads, raise health anxieties further, and prompt even tighter restrictions.School disruptions have set us back the equivalent of a month.The pandemic caused the deepest recession since the Great Depression, as well as disproportionately driving mothers out of the labor force and wiping out gains that women had made in recent years.Over the last three months, the economy has gradually been healing, with payroll employment growing at an average pace of 1.3 million jobs per month.But the declines in the parental labor force suggest that school closings have shaved the equivalent of around a month off the recovery. These declines are one of the most visible demonstrations of what we have learned over the last seven months: The virus dictates the pace of the recovery, and until it’s under control, economic life will not return to normal.Ernie Tedeschi is an economist and head of fiscal analysis at Evercore ISI. He worked previously at the U.S. Treasury Department. The analysis here is solely his own. The full methodology and results for this analysis are available here. You can follow him on Twitter at @ernietedeschi.