The Long-Run Effects of Cesarean Sections

↵Notes: This table shows the estimates of the marginal effect of an unplanned C-section on different neonatal health indicators by OLS, 2SLS, bivariate probit, and differences-in-differences estimation (see Equations 1, 2, 3, and 6). Specifications as detailed in Section III.B, with the full set of fixed effects and controls. The sample for bivariate probit and 2SLS includes only singleton first births; for differences-in-differences, the sample includes both first and second singleton births from families where the first child is born by vaginal delivery (and the second is born by vaginal delivery or unplanned C-section). The OLS estimation includes both first and higher order singleton births. Robust standard errors (in parentheses) in first three panels, and standard errors clustered at the family level in the differences-in-differences panel. First-stage F-statistic from 2SLS and bivariate probit specifications. Significance: *p < 0.10, **p < 0.05, ***p < 0.01, standard p-values. Romano–Wolf multiple hypothesis adjusted p-values in square brackets.

↵Notes: This table shows the estimates of the effect of an unplanned C-section on the probability of the child having each diagnosis by age by OLS, 2SLS, bivariate probit (marginal effects), and differences-in-differences estimation (see Equations 1, 2, 3, and 6). Specifications as detailed in Section III.B, with the full set of fixed effects and controls. The sample for bivariate probit and 2SLS includes only singleton first births; for differences-in-differences, both first and second singleton births from families where the first child is born by vaginal delivery (and the second is born by vaginal delivery or unplanned C-section). The OLS estimation includes both first and higher order singleton births. Robust standard errors (in parentheses) in the first three panels, and standard errors clustered at the family level in the differences-in-differences panel. First-stage F-statistic from 2SLS specifications. Significance: *p < 0.10, **p < 0.05, ***p < 0.01, standard p-values. Romano–Wolf multiple hypothesis adjusted p-values in square brackets.

↵Notes: This table shows the results of regressing different proxy measures of quality of care on the instrument (Normal shift * Pre-leisure), controlling for normal shift, pre-leisure day, hospital, year, and month-of-birth fixed effects and the full set of controls as described in Equation 2. The outcome in Column 1 is a dummy equal to one if the birth was planned (planned C-section or induction) and zero otherwise. In Column 2, the dependent variable is a dummy for shoulder dystocia during childbirth, and in Column 3, a dummy for brachial plexus injury. In Column 4, the dependent variable is mother’s length of stay (in days), and the sample is restricted to mothers who delivered by C-section. Finally, in Column 5, the outcome is a dummy equal to one if the mother is diagnosed with any complication related to the puerperium within one week from the delivery (ICD-10 Codes: O85–O92). Here the sample is restricted to single births, unscheduled C-sections, and vaginal births that take place on working days. Robust standard errors in parentheses. Significance: *p < 0.10, **p < 0.05, ***p < 0.01.

↵Notes: Column 1 shows a placebo regression where the outcome is birth weight. Columns 2 and 3 gives the results from the bivariate probit (top) and the differences-in-differences (bottom) estimation of the impact of unplanned C-section on the probability of asthma diagnosis by age five restricting the sample to births taking place on Thursdays or Fridays (Column 2) or to noninduced births (Column 3). Specifications are as detailed in Sections III.B.1 and III.B.2, with the full set of fixed effects and controls. Robust standard errors (in parentheses) for bivariate probit results, and standard errors clustered at the family level in the differences-in-differences panel. Significance: *p < 0.10, **p < 0.05, ***p < 0.01.