Further investigation of the uncertain unit root in GNP

A more powerful version of the augmented Dickey-Fuller test and a test that has trend stationarity as the null are applied to U.S. gross national product. Simulated critical values generated from plausible trend- and difference-stationary models are used to minimize possible finite-sample biases. The discriminatory power of the two tests is evaluated using alternative-specific rejection frequencies. For postwar quarterly data, these two tests do not provide a definite conclusion. When analyzing annual data over the 1869-1986 period, however, the unit-root null is rejected, but the trend-stationary null is not.