An important contributor to cross-country differences in income and productivity is the sectoral composition of the economy. Poor countries employ a large fraction of their workforce in low-productivity agriculture. Previous work has studied these productivity and employment outcomes from the perspective of the individual. However, agriculture in poor countries typically involves the employment of many household members on a household farm. Additionally, since agriculture is concentrated in rural areas, leaving agriculture for other sectors requires a physical move of the household. Whether a move takes place depends on opportunities for all household members, not lone individuals. Any factor making female non-agricultural employment less attractive acts as a cost to mobility out of agriculture for the entire household and will slow down structural transformation. This project measures individual choices of work in their household context, investigates their determinants, and studies to what extent barriers to urban employment, in particular for women, affect structural change.
The research team studies these questions using two unique large-scale micro datasets they assembled. The first dataset harmonises over 1,700 household and labour force survey datasets from 105 countries. The second dataset harmonises over 130 time-use survey datasets from 42 countries. The team use these datasets to measure labour supply, home hours and wages by gender, marital status, and sector/location. Informed by these data, they write down a two-sector model where time-use and location choice are made by single and married households. The quantitative analysis yields insights into the determinants of female labour force participation, notably on the roles of gender wage gaps, social norms, income effects and home production. The team then use their model to provide evidence on the implications of female choices and their determinants for household choices and, in turn, for aggregate agricultural employment and productivity.
Comparable information on labour market outcomes for detailed demographic subgroups currently does not exist. This research project is a stepping-stone towards such a comprehensive, globally comparable dataset. In addition to the data collection effort, the quantitative analysis speaks directly to the determinants of female labour force participation and sectoral choice, and can be used to better our understanding of barriers to female employment, and the aggregate implications of gender-based policies.
The project’s initial empirical analysis makes three contributions. First, it shows that the drop in the agricultural employment rate across country income levels is indeed largely driven by changes in the share of households working in agriculture, with a much smaller role for within household changes. Second, the project documents that in low-income countries, in urban areas, married females supply much less market hours than their rural counterparts. Third, it provides the first measures of home hours both for domestic services and care work across country income groups both at the aggregate and group level. These facts feed into the quantitative model, for which results are still pending.