(d*) The Breusch-Godfrey test statistic is statistically significant for 1 lags; the p value is 0.01. For
        two lags and beyond, the p vaules were higher. In the interest of parsimony, one should use
        one lags. One chooses the p order through experimentation.