SMALL SAMPLE PROPERTIES OF PANEL COINTEGRATION TESTS IN THE PRESENCE OF STRUCTURAL CHANGE

Panel tests for non-stationarity are increasingly popular in recent years, also for macroeconomic data. Given that panels used in practice are rather small, there is a need of exploring the small sample properties of the tests in various cases. For annual data the N dimension of the panels is limited to no more than 25, and spatial dimension is also limited, because of the nature of studied entities. So, the main concern of researchers remains the relatively small panels – theoretical critical values should be applied with caution, given that they are taken in limits. An additional feature of macroeconomic panels are cycles – with business cycles one can expect even more than one structural break in the series, because up to 3 major cycles can fit in a series with T=25. In the paper small sample properties for the three “group” statistics of Pedroni (1999) under presence of structural breaks are explored. A set of Monte Carlo experiments is applied to processes with a structural break for three possible break dates at 0.3T, 0.5T and 0.7T. Tested for power against the general alternative.