Changes in the Distribution of Income Volatility

Research documenting an increase in income risk in recent decades has treated its proxy -- income volatility, the expected magnitude of income changes -- as if it were the same for everyone. In reality, some people face more risk than others. We develop a Markovian hierarchical Dirichlet process (MHDP) prior to model the evolution of the distribution of income volatilities, not just the mean. This augments the recently-developed hierarchical Dirichlet process (HDP) prior to accommodate the serial dependence of panel data. We find that increases over time in mean volatility can be attributed solely to increases in volatility at the right tail; the risky are getting riskier while the rest of the distribution remains unchanged. This has strong welfare implications, as those facing increasing volatility may also be the most risk tolerant, as evidenced by their willingness to take on income risk.