Job Market Paper
Workhour Normality and the Gender Gap: Evidence from Brazilian Exporters
Workhours are ‘normal’ when they match the rest of the economy and are regular. Exporting firms’ foreign trading partners often operate in different time zones and can induce abnormal working hours. Using a comprehensive employer-employee matched dataset from Brazil, I show that the further away a trading partner is temporally, the wider are the gaps in employment and wages in exporting firms. Two extra hours of difference of an exporter from their trading partners can lead to a decrease in the proportion of female employees by 3.5 percentage points, and a fall in women’s wages by 2.8 percent. This can account for more than 10 percent of the gender wage gap seen in exporting firms. Men’s wages are unaffected by the temporal distance to the trading partners. Using a panel event study design, I also find that new mothers are significantly more likely to leave exporting firms with far-away trading partners than they are to leave domestic firms a year after giving birth. This suggests that women, especially mothers, find it more difficult than men to work abnormal hours, and this explains a significant portion of the gender gap in employment and earnings.
[draft available soon]
Publications
Redistributing Teachers using Local Transfers
with Athisii Kayina, Abhiroop Mukhopadhyay, and Anugula N. Reddy
World Development (2018), 110: 333-344. PDF
In this paper we show that local redistribution of educational resources via teacher transfers between neighboring public schools can improve equity in access to teachers. Transfers from teacher surplus schools to deficit schools within a 10 km radius in Haryana, a state of India for which we have geo-coded location of schools in 2013, enables 19 percent of deficit schools to meet the minimum requirement. We use the mandated norms in the Right to Education Act in India, to define deficit and surplus schools. In the process we also provide a characterization of schools that are in deficit and those in surplus. We find that connectedness, the social composition of the enrolled students, the income of the neighborhood are important determinants of a school being in deficit. Surplus schools mirror the results on deficit, but not always so: they are far more heterogenous, leading to possibilities that they may in fact be no different than some low shortage deficit schools. Keeping in the background this heterogeneity in surplus schools, we design local transfers between schools and evaluate them on how well they match characteristics of the donor and recipient schools. The chosen algorithm is compared to another transfer rule that reduces the variance of shortages across schools and is found to be better in matching characteristics, that is, the donor and recipient schools are, on an average, matched in characteristics: in terms of the development of the region, its rural/urban location, connectivity and school characteristics. A comparison of transfers that follow our redistribution rule to transfers resulting from an actual transfer policy shows that while our rule removes deficits in rural areas, the actual transfers favored more developed regions.
Dispatches 