The implications of selective attrition for estimates of intergenerational elasticity of family income.

Numerous studies have estimated a high intergenerational correlation in economic status. Such studies do not typically attend to potential biases that may arise due to survey attrition. Using the Panel Study of Income Dynamics - the data source most commonly used in prior studies - we demonstrate that attrition is particularly high for low-income adult children with low-income parents and particularly low for high-income adult children with high-income parents. Because of this pattern of attrition, intergenerational upward mobility has been overstated for low-income families and downward mobility has been understated for high-income families. The bias among low-income families is greater than the bias among high-income families implying that intergenerational elasticity in family income is higher than previous estimates with the Panel Study of Income Dynamics would suggest.