Hypothesis Testing with Real GDP: Somewhat Mixed Results

Many of the hypothesis test results appear in this chapter of the book and the online-only appendix. The tests have attempted to correct for possible violations of standard regression assumptions. This chapter and its online appendix present Chow (1960) breakpoint test results and breakpoint unit root test results. This chapter also has Wald coefficient restriction test results and t-statistic test results seeking a change in the intercept over subsample periods. This chapter also offers Chow forecast test results. Although less clear than in Chapter 3 , many of the results support the view that at least some negative aggregate demand shocks have resulted in downward breaks in U.S. real GDP in log level and ‘change in the log of’ form (with a unit root interpreted as a break in each period). Does a break in the intercept with data in ‘change in the log of’ form suggest a break in the time trend with log level data? At least one downward break in U.S. per capita output likely also occurred. This suggests that changes in population (or its growth rate) alone cannot explain all downward breaks in U.S. real output that seem to have been associated with decreases in aggregate demand.