![Sample our Global Development journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5929/TOC-Subject-Sample-2021-Global-Development.jpg"})
Abstract
Climate shocks have well-documented impacts on the short-term welfare of rural households in developing countries. We investigate the impact of droughts on Brazilian labour markets. We find long-term rural wage losses beyond the immediate impact of the drought, with affected workers taking five years to catch up with their peers. Findings are robust to model specification. The severity of the losses varies with the dependence on agricultural income, supporting the notion of diversifying portfolio strategies in rural areas to reduce climate-related income risk.
Acknowledgements
We thank Cristian Pop-Eleches, Glenn Sheriff, Eric Verhoogen, and two anonymous referees for comments on earlier versions of this paper. Valerie Mueller acknowledges financial support from the Rosenstiel School of Marine and Atmospheric Sciences at the University of Miami, and the Postdoctoral Fellowship Program at the Earth Institute at Columbia University.
Notes
1. We include state fixed effects because we are mechanically unable to identify the effect of droughts on wages when including municipality fixed effects. There is not enough temporal variation in the dataset to be able to identify effects when including municipality fixed effects. For example, 99 per cent and 96 per cent of the municipalities have smaller within-municipality variance than across municipality variance for the treatment variables. Thus, with municipality fixed effects we have near multicollinearity and larger standard errors. Regression results including municipality fixed effects were insignificant or had counter-intuitive signs. We do offer state fixed effects to account for spatial heterogeneity, however, this many not take into account bias from across municipalities, for example, if agricultural investment occurs less in areas prone to droughts.
2. Due to the cross-sectional nature of the data, there is potentially an identification problem if agricultural investment is correlated with unpredictable weather. We include the banks variable to control for this effect.