Measurement Error in Surveyed Data: Revisiting the Study of Income and Consumption Dynamics

Nayoung Lee, University of Southern California

This paper uses data from the Korean Labor and Income Panel Study to examine whether measurement error in survey data has the potential to generate biases for studies on income and consumption dynamics. A first-differenced dynamic panel model is estimated with lagged incomes and consumptions as internal instruments, and two external instruments - individuals’ satisfaction regarding their household income and per capita household eating-out consumption - are used. This study suggests that time-varying measurement error plays an important role especially in the study of consumption dynamics. The combined effect of unobserved heterogeneity and time-invariant measurement error leads to an upward bias for the studies of income and consumption dynamics. However, the effect of time-varying measurement error offsets the combined effect of unobserved heterogeneity and time-invariant measurement error in the study of consumption dynamics.