Abstract

We document the positive correlation between part-time work and hourly earnings in Latin America. Using Chile as an example, we analyze the causal effect of working part-time on hourly earnings. And we provide explanations behind the change in sign (from positive correlation to negative causality). The positive correlation between part-time work and hourly earnings has already been documented for Honduras and Chile. We provide evidence that it is also present in several Latin American countries. Using a novel econometric technique that does not rely on the presence of an exclusion restriction for identification, we have causally estimated the effects on hourly earnings of working part-time, which are negative and significant for a subsample of female workers (those in a dependency relation or in formal settings). The estimated penalty is economically important in magnitude, as it is of about 30 per cent. We suggest that the reason for the sign change is the presence of heterogeneity, which summarizes in a point estimate (the OLS) many different situations, encompassing multiple effects (measurement errors, precarious features). Once those situations are considered and the heterogeneity is taken into account, there is a negative causal effect from working part-time on hourly earnings.