How the provision of childcare affects attitudes towards maternal employment

Anja Roth•View original
HighlightsTechnicaladela voice
A government builds childcare facilities expecting public gratitude. But does exposure to those facilities actually change how people think about mothers working? That's the question Anja Roth sets out to answer, and the reason it's hard is the endogeneity problem. Attitudes shape which institutions get built, and institutions shape attitudes. Separating those two directions of causation requires a clever natural experiment. Roth finds one in Switzerland. In 2010, the cantons of Bern and Zurich introduced a regulation requiring municipalities to assess annual demand for lunchtime and after-school care, and to provide it if ten or more children signed up for any slot. Neighboring cantons had no such requirement. That geographic discontinuity, combined with Switzerland's direct democracy, where citizens vote on specific policy bills roughly four times a year, gives Roth a differences-in-differences design with unusually precise measurement of attitudes. Ballot outcomes reveal policy-specific preferences, not the blurry aggregates of party affiliation. The design exploits two treatment margins. First, cantonal regulation: did your canton impose the new requirement? Second, municipal provision: did your municipality actually open the facilities? Municipality fixed effects and ballot fixed effects absorb time-invariant local characteristics and nationwide swings in opinion. The key identifying assumption is parallel pre-trends, which Roth supports with pre-treatment comparisons, a subsample restricted to municipalities that opposed the 2004 maternity-leave bill, and placebo exercises using healthcare ballots — ballots where the reform should produce no effect, and doesn't. The headline result is that the provision of lunchtime and after-school care raises voter support for policies promoting maternal employment by roughly 1.8 to 2.9 percentage points across specifications, with a baseline estimate of 2.91 percentage points. That's within roughly three years of introduction. The effect isn't driven only by parents. Only about sixteen percent of voters have school-aged children, yet approval shifts by nearly 3 percentage points — so the norm change propagates through the broader population. Post-vote surveys confirm this: treated cantons became measurably more favorable toward promoting gender equality overall, with a regulation-times-post coefficient of 0.166. Then the cost finding complicates the story. In Bern, total childcare costs in 2013 ran to about fifty point seven million Swiss francs, with the public covering thirty-eight point eight million — roughly 0.27 percent of public expenditures. As per-capita public costs rise, the provision benefit erodes. The interaction coefficient is around negative 0.12, meaning each additional franc of per-capita cost reduces approval by about that amount. The regulation margin itself is associated with a decline of 8.7 to 9.7 percentage points, suggesting that when voters see the bill, enthusiasm cools sharply. Roth's contribution is rare: credible causal evidence that local institutions move gender norms, not just reflect them. The practical implication is that how policymakers sequence and communicate childcare costs matters as much as the facilities themselves. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.

A government builds childcare facilities expecting public gratitude. But does exposure to those facilities actually change how people think about mothers working? That's the question Anja Roth sets out to answer, and the reason it's hard is the endogeneity problem. Attitudes shape which institutions get built, and institutions shape attitudes. Separating those two directions of causation requires a clever natural experiment. Roth finds one in Switzerland. In 2010, the cantons of Bern and Zurich introduced a regulation requiring municipalities to assess annual demand for lunchtime and after-school care, and to provide it if ten or more children signed up for any slot. Neighboring cantons had no such requirement. That geographic discontinuity, combined with Switzerland's direct democracy, where citizens vote on specific policy bills roughly four times a year, gives Roth a differences-in-differences design with unusually precise measurement of attitudes. Ballot outcomes reveal policy-specific preferences, not the blurry aggregates of party affiliation. The design exploits two treatment margins. First, cantonal regulation: did your canton impose the new requirement? Second, municipal provision: did your municipality actually open the facilities?

Municipality fixed effects and ballot fixed effects absorb time-invariant local characteristics and nationwide swings in opinion. The key identifying assumption is parallel pre-trends, which Roth supports with pre-treatment comparisons, a subsample restricted to municipalities that opposed the 2004 maternity-leave bill, and placebo exercises using healthcare ballots — ballots where the reform should produce no effect, and doesn't. The headline result is that the provision of lunchtime and after-school care raises voter support for policies promoting maternal employment by roughly 1.8 to 2.9 percentage points across specifications, with a baseline estimate of 2.91 percentage points. That's within roughly three years of introduction. The effect isn't driven only by parents. Only about sixteen percent of voters have school-aged children, yet approval shifts by nearly 3 percentage points — so the norm change propagates through the broader population. Post-vote surveys confirm this: treated cantons became measurably more favorable toward promoting gender equality overall, with a regulation-times-post coefficient of 0.166. Then the cost finding complicates the story. In Bern, total childcare costs in 2013 ran to about fifty point seven million Swiss francs, with the public covering thirty-eight point eight million — roughly 0.27 percent of public expenditures. As per-capita public costs rise, the provision benefit erodes.

The interaction coefficient is around negative 0.12, meaning each additional franc of per-capita cost reduces approval by about that amount. The regulation margin itself is associated with a decline of 8.7 to 9.7 percentage points, suggesting that when voters see the bill, enthusiasm cools sharply. Roth's contribution is rare: credible causal evidence that local institutions move gender norms, not just reflect them. The practical implication is that how policymakers sequence and communicate childcare costs matters as much as the facilities themselves. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.

More in Social Sciences →

Related lectures