Job Market Paper

Abstract

In this paper I develop and estimate a quantile-based panel data model to explore the nature of income dynamics over the business cycle. The logarithm of household income net of observable characteristics follows a general first-order Markov process, which depends on observable characteristics and a latent factor that captures the business cycle. The model is estimated using administrative panel data covering the entire Norwegian population between 1979 and 2013. I find that the income process over the business cycle differs systematically by age, income, skill level and their interaction. My findings also suggest that the redistributive nature of the Norwegian tax and transfer system plays a key role in attenuating the increased left-skewness risk of income shocks during recessions, especially among the low skilled. The estimated income process is then used to study the implications for the welfare costs of business cycles and its distribution among the population in a dynamic life-cycle model of consumption and savings.