Sources of measurement errors in earnings data : new estimates of intergenerational elasticities

Abstract

Using Norwegian intergenerational data with a substantial part of the life-cycle earnings of children and almost the entire life-cycle earnings for their fathers, we present new estimates of intergenerational mobility. Extending the length of the fathers’ earnings windows from 5 to 30 years increases the estimated elasticities. Varying the age of father at observation has the opposite effect. Our findings indicate that intergenerational earnings mobility may have been strongly overstated in many earlier studies with shorter earnings histories. Biases in the estimated elasticities appear to be related to age and/or life-cycle measurement errors more than persistency in the transitory innovations.