The effects of childcare costs on women’s labor force participation rates

By Marianna Anagnostou (’24), Zach Everett (’25), Jonathan Farbman (’26), and Seona Maskara (’26)

Note: This is one of a series of guest posts from students in the Tufts Political Science Research Methods course.

It’s been three years since the Biden Administration introduced the “American Families Plan,” which proposed spending $225 billion on childcare. The plan sought to significantly increase the size of the childcare sector, raise the minimum wage of childcare workers to 15 dollars an hour, and ensure that low and middle-income families are paying at most 7% of their income on childcare. To provide some perspective, most families with children spend over 20% of their income on childcare. Policy on childcare has been creeping forward over the last 10 years, but still, the American Families Plan — or any major federal legislation like it — never passed.

While childcare costs affect families as a whole, more of the burden has historically been placed on women, who are traditionally viewed as the primary caregivers. This can affect women’s financial independence, deepen the pay gap, and perpetuate gender inequality. Although infant and toddler care centers can provide some assistance, it is an expensive choice that not everyone can afford to make. Thus, we wanted to conduct research on childcare through the lens of gender to answer the question: How do infant & toddler childcare costs affect women’s labor force participation in the United States?

How we did our research

In order to estimate the effect of childcare costs on women’s labor force participation rates, we ran two separate OLS regression models — one for toddler center-based childcare costs, and one for infant center-based childcare costs, which both come from the U.S. Department of Labor’s National Database of Childcare Prices. In both models, the cost of center-based childcare as a percent of median household income was the primary independent variable, and the women’s labor force participation rate was the dependent variable. Given that the figures are similar, we’ve only included the scatterplot and OLS regression model for toddler center-based childcare costs, as they cover a wider age range. We ran our analysis on county-level data from all US states, with the exception of AK, CO, HI, IN, MO, and NM, since these states had no — or almost no — data in our childcare cost dataset. In order to isolate the effect of the cost of childcare, we controlled for the following additional variables in our model: women’s average commute time to work and the percentage of women who work from home, both from American Community Survey 5-year estimates; the percentage of women with a bachelor’s degree, also taken from the American Community Survey; the average annual cost of living, from the Economic Policy Institute’s Family Budget Calculator; racial composition, from the original DOL dataset; and population density, from census data. Finally, we controlled for state, in order to account for any of a number of potential state-specific effects, such as family leave policies, child tax credits, etc. All data was from 2018, when available, except for the cost of living data which was only available for 2024.

What we found

This scatterplot visualizes the relationship between women’s labor force participation rate toddler center-based costs as a percentage of median income. The size of the point is proportional to the population of the county.

The next figure visualizes the multiple regression analysis of women’s labor force participation rate and toddler center-based costs as a percentage of median income, with controls. The analysis also controls for state, although those effects are not visualized in the figure. Each point represents the coefficient estimate for that variable, and the horizontal bars represent 95% confidence intervals.

Controlling for population density, percentage of population that is white, average women’s commute time, percent of women working from home, percent of women with a bachelor’s degree, annual cost of living, and state, we found that higher center-based childcare costs in a given county are associated with lower women’s labor force participation rates (i.e., the variables have an inverse relationship). The finding is significant at a 95% confidence level, and therefore, we can safely reject the null hypothesis. For every percentage point increase in infant center-based childcare cost as a percentage of median household income, the women’s labor force participation rate decreased by .5962 percentage points. For every percentage point increase in toddler center-based childcare costs as a percentage of median household income, the women’s labor force participation rate decreased by .749 percentage points. The magnitudes of both these relationships are quite sizable. We speculate that the greater effect that toddler center costs have on women’s labor force participation than infant center costs do may be due to the impact of long-term childcare costs. This is a notable finding, as it demonstrates that childcare costs at different stages of a child’s growth may incentivize or discourage women’s participation in the workforce at different rates.

Conclusions

Our research suggests that the price of infant/toddler care centers predicts the women’s labor force participation rate with an inverse relationship. There is a possibility of reverse causality with our findings, where women’s labor force participation rate may actually be driving the prices of infant/toddler center-based childcare more than the reverse; however, we see a more plausible real-world explanation for our hypothesized relationship of childcare costs driving women’s labor force participation.

The connection between gender inequity and childcare is clear as higher childcare costs are associated with fewer women working. This hurts women’s financial autonomy and only deepens preexisting inequity, demonstrating a potential need for policy that supports lower childcare costs.

We also want to acknowledge that our research is most representative of the dynamics of a heterosexual partnership which inherently excludes a number of queer identities and non-traditional family structures. Additionally, we recognize that looking at women’s participation rate may exclude other marginalized gender identities that also bear the burden of high childcare costs. Further research should be conducted with data that accounts for these identities, and which is based on family-level analysis, which could present a more nuanced and granular picture of how family dynamics mesh with childcare costs to determine which caregiver(s) in a family work, in what jobs, and for how long.