Estimating Vector Autoregressions with Panel Data

TLDR

The study develops and tests vector autoregression techniques for panel data, applied to wage and hours dynamics among American males. The model incorporates nonstationary individual effects and is estimated via instrumental variables on quasi‑differenced autoregressive equations. Results indicate lagged hours do not predict wages, yet lagged hours are significant predictors of hours worked. © 1988 The Econometric Society.

Abstract

This paper considers estimation and testing of vector autoregressio n coefficients in panel data, and applies the techniques to analyze the dynamic relationships between wages an d hours worked in two samples of American males. The model allows for nonstationary individual effects and is estimated by applying instrumental variables to the quasi-differenced autoregressive equations. The empirical results suggest the absence of lagged hours in the wage forecasting equation. The results also show that lagged hours is important in the hours equation. Copyright 1988 by The Econometric Society.