WenShwo Fang, Stephen M. Miller
This study examines the relationship between U.S. output growth and its volatility over the period 1876:I to 2012:II. We adjust the data for outliers and structural breaks. We employ generalized autoregressive conditional heteroskedasticity (GARCH) and exponential GARCH (EGARCH) specifications. Normality and homoskedasticity appear only in the GARCH or EGARCH model that corrects for the outliers. When including the break in the mean equation, high volatility persistence remains. After also accommodating the breaks in the variance equation, the integrated GARCH effect proves spurious, either for the symmetric or the asymmetric model. Finally, our empirical results suggest that the finding of higher output growth volatility stimulating output growth and higher output growth reducing its volatility obtained from the symmetric GARCH-in-mean (GARCH-M) model also proves spurious as a result of the emergence of an asymmetric effect. Our more appropriately specified asymmetric EGARCH-M model suggests positive volatility-in-mean and level effects in the long-period real gross national product series.