Labor Income Indices Designed for Use in Contracts Promoting Income Risk Management

Labor income indices are created for groupings of individuals, using data from the Panel Study of Income Dynamics. People are grouped by a clustering algorithm based on an estimated transition matrix between jobs, by education level, and by skill category. The groups are defined so that relatively few people move between them. For each of the groupings, we generate a labor income index using a hedonic repeated-measures regression methodology. Similarities between pairs of indices and between indices and individual labor incomes are described. It is argued that indices like those presented here might someday be used in settlement formulae in contracts promoting income risk management.