Thread by @jasonfurman

Expanded version of our childcare paper on @nberpubs today, with @kearney_melissa and Willie Powell.

Finds that the differential impact on women with younger children explains ~1% of the labor market worsening from Jan/Feb-2020 to Jan/Feb-2021.

nber.org/papers/w28934?…

Our Q: how much of the 3.6pp decline in EPOP was due to childcare issues. Intuition for small result:

1. Women with younger children ~10% of workforce

2. Need to look at their *differential* reduction in work, not entire reduction

3. If you include father's goes the other way.

The @nberpubs version goes through many, many, many sensitivity tests with different control groups, different concepts, but keeps getting roughly the same result. Because even if #2 on the previous were 2X that would take the result from 1% to 2%. So robustness not surprising.

I addressed many of these sensitivity issues in a previous thread (the numbers don't exactly match the @nberpubs version but they're close).

https://twitter.com/jasonfurman/status/1396923831889977346?s=20

A reminder: we are NOT asking whether childcare has been a challenge (it has). We are also not asking whether school closures have been bad for children (it has).

We are also not trying to precisely estimate the impact of childcare challenge's on women's work. To the degree we find that mother's of young children have reduced their work somewhat compared to other control groups.

Instead we're asking a macroeconomic question: how much of the total job loss is due to this factor. Every other attempt to get at this question I have seen also finds a small result.

This paper is an attempt to better understand what was going on in the world. @nberpubs does not allow policy conclusions & wouldn't have wanted one regardless.

But here is my own policy view: our childcare system was problematic pre-pandemic, strongly support the Families Plan.