Testing for Error Correction in Panel Data

This paper proposes four new tests for the null hypothesis of no cointegration in panel data that are based on the error correction parameter in a conditional error correction model. The limit distribution of the test statistics are derived and critical values are provided. Our Monte Carlo results suggest that the tests have reasonable size properties and good power relative to other popular residual-based cointegration tests. These differences arises because latter imposes a possibly invalid common factor restriction. In our empirical application, we present evidence suggesting that international health care expenditures and GDP are cointegrated once the possibility of an invalid common factor restriction has been accounted for.