Abstract

This paper considers the estimation of common break dates in panel data models when the break structure is misspecified. The common change point estimator has been shown to be consistent, assuming that a change has taken place in each series at an unknown common point. In practice, however, the common break assumption is restrictive, and the change points may vary significantly across individuals. Recent clustering literature suggested modeling the panel data using a grouped pattern. The regression coefficients containing break dates are heterogeneous across groups but homogeneous within a group. In this paper, we consider a case where the break points are common in one group but can vary across groups. The asymptotic analyses and simulation results show that if the single-break model is estimated, the estimated common break point is consistent for one of the true break points.