Measuring Coverage, Explaining Constraints: A Mixed-Methods Study of Childcare and Women's Employment in the Philippines

This dataset supports the study “When Childcare Systems Bind: A Mixed-Methods Exploration of Childcare and Women’s Employment in the Philippines.” It contains the analytical dataset used to examine whether childcare coverage is associated with employment rates in the Philippines, with particular attention to gender differences in employment outcomes. The dataset was constructed from two sources. The first is the Philippine Statistics Authority’s merged 2023 Family Income and Expenditure Survey and Labor Force Survey (FIES-LFS) microdata, which contains 438,199 individual-level observations. The second is barangay-level childcare data from the Child Development Council, covering 42,010 barangays and indicating whether each barangay has a childcare center. Because the FIES-LFS public-use file does not contain barangay identifiers, the FIES-LFS and childcare datasets could not be merged at the barangay level. Both datasets were therefore aggregated to the province or province-equivalent level using the province identifier. Barangay-level childcare records were converted into province-level childcare coverage rates, while individual-level FIES-LFS records were aggregated into province-sex-level employment rates and covariate means or shares. The final analytical dataset contains 170 observations, representing 85 province-equivalent areas separately observed for men and women. The outcome variable is the province-sex-level employment rate. Individual employment status is coded as one if the respondent is employed and zero otherwise, and then computed as the weighted mean within each province-sex cell. The key explanatory variable is childcare coverage, measured as the ratio of barangays with childcare centers to total barangays in each province-equivalent area. The corrected coverage measure ensures that the rate does not exceed one. For the National Capital Region, the four NCR districts are retained as province-equivalent areas following the FIES-LFS structure, but the same corrected NCR childcare coverage value is assigned to all four districts because childcare coverage is available only at the NCR level. The dataset also includes province-sex-level control variables, including the natural logarithm of household income per capita, household headship, illness or disability, education composition, marital-status composition, urban residence, certification, age, age squared, and household size. These variables are measured as weighted means or shares after aggregation. The dataset is intended for replication, verification, and further analysis of the quantitative results reported in the manuscript. It supports pooled baseline, sex-differentiated, and controlled sex-differentiated models estimating the association between childcare coverage and employment rates. The estimates derived from this dataset should be interpreted as associations, not causal effects, because the analysis is cross-sectional and cannot fully address unobserved provincial characteristics, reverse causality, or simultaneity between local employment conditions and childcare availability. This dataset contains no personally identifiable information. The raw qualitative data from the focus group discussions and key informant interviews are not included because confidentiality and anonymity of participants must be protected.