Shift-Share Estimation of Child Care Labor Supply

• In Event: Empirical Evidence on Child Care Supply and the Child Care Workforce

Abstract

How much would it cost to expand the child care workforce? Several existing federal programs aim to improve child care access and quality, and recent years have seen substantial discussions of child care expansions that would bring US investments in early childhood closer to those made by OECD peers. Improvements to either quantity (access) or quality would require new ECE teachers, which can be expected to require wage improvements. But how large should we expect these wage increases to be?

Surprisingly few credible estimates are available to inform policy simulations. Blau (1993) uses data from the Current Population Survey (CPS) to estimate a labor supply elasticity of 1.2-1.9. However it is open to question to what extent this study, based on data from 1977-87, remains applicable given substantial changes to the child care market. Cunha and Lee (2023) use worker turnover data from Texas to estimate a structural model that implies a labor supply elasticity of 2. Bassok et al (2021) and Urban Institute (2025) find that child care wage interventions have large effects reducing teacher turnover, which would tend to suggest a high value for labor supply elasticity. However, there remains a need for more evidence about this critical parameter.

We try to close this gap with new estimates that describe how child care wages and employment respond to local demand shocks. Our approach blends two classic techniques. We construct Bartik shift-share instruments (Bartik 1991) to obtain a measure of plausibly exogenous shocks to employment in industries that employ likely child care consumers. These measures then serve as a demand-shock instrument (Wright 1928) to estimate the supply curve in the market for early childhood education workers.

Our approach relies on identifying industries that employ child care consumers but not potential child care workers. In order to accomplish this, we conduct an analysis of industry workforce “relatedness” to child care. We use two approaches: First, we examine CPS data on transitions of workers between industries. We argue that employment shocks in industries that have comparatively few transitions to or from child care industry are unlikely to affect the child care worker supply curve. Second, we use ACS PUMS data estimate a predictive model of whether an individual works in child care as a function of individual and household demographics. This model is then used to score industries based on how similar their workers are to child care workers.

Initial results suggest that the two approaches to industry “relatedness” yield similar results, and identify several large industries that employ likely consumers of child care but are not highly connected to the childcare workforce. Further analysis using county-level Quarterly Workforce indicators data will use shift-share shocks constructed based on these demand industries to estimate how child care employment and wages respond to local demand shocks and obtain new estimates of labor supply elasticity.

Author

• Jonathan Borowsky, University of Minnesota - Twin Cities
  Presenting Author