Missing Values in Panel Data Unit Root Tests

Missing data or missing values are a common phenomenon in applied panel research and of great interest for unit root testing. The standard approach the literature is to balance by removing units and/or trimming time period all units. However, this can be costly terms lost information. Instead, existing tests could extended case unbalanced panels, but often difficult because observations affect bias correction which usually involved. This paper contributes two ways; it extends popular allow values, secondly, employs asymptotic local power functions analytically study impact various missing-value methods on power. We find that zeroing-out method results greater test power, result holds deterministic component specifications, such as intercepts, trends structural breaks.