Utilizing comprehensive administrative data from Brazil, we investigate the impact of peer effects on wages, considering both within-gender and cross-gender dynamics. Since the average productivity of both individuals and their peers is unobservable, we estimate these values using worker fixed effects while accounting for occupational and firm sorting. Our findings reveal that within-gender peer effects have approximately twice the influence of cross-gender peer effects on wages for both males and females. Furthermore, we observe a reduction in the disparity between these two types of peer effects in settings characterized by greater gender equality.

This paper partially identifies population treatment effects in observational data under both non-random treatment assignment and sample selection. Bounds are provided for both average and quantile population treatment effects, combining assumptions for the selected and the non-selected subsamples. We show how different assumptions help narrow identification regions, and we illustrate our methods by partially identifying the effect of maternal education on the 2015 PISA math test scores in Brazil. We find that while sample selection increases considerably the uncertainty around the effect of maternal education, it is still possible to calculate informative identification regions.

We present new empirical evidence on the implications of minimum wages on employment in the uncovered sector in developing countries, analyzing a unique dataset assembled from a set of micro surveys collected in 49 low- and middle-income countries. Our identification strategy exploits relative bindingness in minimum wages across socio-demographic labor market groups, and allows for heterogeneous effects depending on the degree of compliance with minimum wage legislation. We find that a higher minimum wage is associated with a larger self-employment share, the effect being statistically significant only in contexts where the rule of law is more respected. The estimated impact of the minimum wage on the uncovered sector, once non-compliance with the law has been considered in the analysis, is economically significant.

This paper examines the adoption and diffusion of a computer-assisted learning (CAL) platform in 188 low-performing public schools in Lima, Peru. Using a large-scale randomized control trial with over 1,600 teachers and 50,000 students in 4th-6th grade we evaluated three treatments to facilitate adoption: (i) workshops for all teachers, (ii) workshops for selected teachers, and (iii) workshops plus personalized coaching for selected teachers. We find that treated teachers and their students had significantly greater engagement with the platform compared to non-treated teachers; teachers who received both workshops and coaching had significantly more students connecting regularly and completing exercises than those who only received workshops. Treated teachers reported having more information, more knowledge, and more favorable attitudes towards technology. Treatment effects were heterogeneous by gender, age, digital skills, and attitudes towards technology. We also found positive spillovers among non-treated teachers. Platform use declined in subsequent years, but supplementary workshops were significantly more effective at boosting utilization for teachers who had already received our interventions.

We characterize the presence of product and labor market power for a wide range of countries with different levels of development. To do so, we rely on data representative of small, medium, and large formal manufacturing firms in countries around the globe from the World Bank Enterprise Survey (WBES) and use the production function approach to estimate markups and labor wedges (gaps between wages and marginal revenue products of labor). We propose a measure of combined market power that jointly measures markups and labor wedges and discuss its conceptual relevance. We conclude that: 1) Although the average firm holds significant market power in countries of all income levels, both markups and markdowns are higher in less developed economies. 2) There is significant cross-firm dispersion of markdowns and labor wedges around that mean, however, more so in lower-income economies. 3) In all regions larger firms exhibit higher labor market power, but not higher product market power. Implications for the functional distribution of income across levels of development are discussed.

This study explores how perceptions of corruption influence the perception of inequality and the attitudes towards redistributive policies. We rely on two survey experiments administered in eight Latin American countries. We find that perceptions of widespread corruption and tax evasion by wealthy individuals and corporations increase awareness of unequal opportunities and unfairness in income distribution. Consequently, there is an increase in the support for taxing the rich. However, these perceptions do not significantly alter support for other taxation forms aimed at expanding social policies, nor do they affect attitudes towards key anti-poverty initiatives such as conditional cash transfers and non-contributory pensions. We argue that, since corruption and tax evasion also erode trust in the government, individuals do not do not translate their inequality concerns into a higher demand for redistributive policies.