Childcare & Employment Outcomes

  • Recent research likely understates the impact of children and lack of childcare on the economic recovery.
  • Many people with children may be underemployed and hours worked should be a key economic variable in a study of the impact of children on employment outcomes.
  • The impact of children and childcare on employment outcomes is concentrated on the part of the workforce that is of child-rearing age.  An analysis including people who are older and retiring will understate the impact of childcare on employment outcomes.
  • Additional research should consider multivariate models based on repeated cross sections and on longitudinal datasets.  The research should consider multiple existence of children definitions and educational attainment groups.

In a recent article, Jason Furman and coauthors argue that lack of childcare had only a small impact on the U.S. jobs recovery at the end of the COVID pandemic.    

Furman’s focus is on the change in the employed to population ratio for people with children under 13 and for people with no children over 13.  He finds a small decrease in labor force participation for women without a bachelor’s degree but no impact for any other group.

The research paper is narrowly focused on labor force participation and on a single definition of households with children.   Several additional tests are needed.

  • Many people without childcare are employed part time and would like to work additional hours.  It would be useful to examine how the pandemic impacted hours worked and underemployment for people with and without small children. My hypothesis is the existence of children of certain ages has a larger impact on underemployment than on labor force participation or unemployment.
  • The article does not document a pre-pandemic baseline on the impact of children and childcare on employment patterns.  It is likely that in normal times many people with children were already out of or marginally attached to the workforce.
  • Furman and coauthors consider only one classification of people with children — at least one child under 13.  Other existence of children variables including a child not yet in school, a school-age child, and multiple children in certain age groups should be considered.     
  • Furman and coauthors consider the impact of two education groups – no bachelor’s degree and bachelors or above on the impact of childcare on labor market outcomes.  It would be useful to consider three groups – less than bachelor’s degree, bachelor’s degree, and above bachelor’s degree.   I suspect this partition will show a significant impact of children on labor force outcomes for the large cohort of people ending their educational career with a bachelor’s degree.  Additional education beyond a four-year degree may reduce the impact of children on labor force outcomes.
  • The bivariate results presented by Furman and coauthors can be misleading because of omitted variables.   The bivariate framework does not allow for consideration of multiple child-existence variables or multiple education groups.  (See two comments above.). The bivariate framework also does not account for the impact of age on workforce outcomes and the interaction of an age and children on work outcomes.  
  • An analysis of the population in peak working years age 26 to 50 is likely to show a larger impact than an analysis of the entire population.  Older workers who are retiring may be out of the workforce regardless of whether they have children.  Furman’s finding that the impact of children on employment outcomes is small may be the result of the existence of a large number of older workers who are leaving the workforce.  The pertinent population for this question is workers of child-rearing age only.
  • The speed of the recovery and the impact of childcare on the labor market is likely to differ across industries.  Labor shortages have been most pronounced in restaurant and hospitality industry.  It would be useful to determine if the impact of childcare on employment and the labor market recovery is larger in some industries than other industries, especially since certain industries have a more pronounced labor shortage and are more highly dependent on female workers with young children.
  • The CPS can be used to track individuals over a period of time.  It would be highly useful to use longitudinal CPS data to compare month to month changes in labor market outcomes – fully employed, unemployed, underemployed, non-participant, from month to month.

Concluding Thoughts:  The analysis by Furman and coauthors likely substantially understates the impact of childcare on the economic recovery.  Much more analysis on this topic is needed.

Jason Furman, Kearney, M., and Powell, W. How much have childcare challenges slowed the US jobs market recovery.

Julia A. Rivera Drew, Flood, S, and Warren J.R. Making Full Use of the Longitudinal Design of the Current Population Survey:  Methods for Linking Records Across 16 Months


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.