Identification of earnings dynamics using rotating samples over short periods: the case of Chile

Abstract

Due to high informality and sparse longitudinal data, empirical studies often ignore labor income dynamics in developing economies, where earnings inequality is highest and social insurance is weakest. I propose a dynamic earnings process with two distinct shocks: unemployment spells and the wages of workers who stay employed. This income process can be estimated from employment surveys with a rotating sample design, which are available for several countries. Applying this procedure to Chilean data, I show that wage volatility and unemployment rates are highly heterogeneous across workers. Unemployment spells are the most important source of earnings risk for workers.