The Breusch-Godfrey test is a statistical tool used to detect autocorrelation in the residuals of regression models. It’s an extension of the Durbin-Watson test, allowing for higher order autoregressive processes.
In application, we first estimate our model and obtain the residuals. Then, we regress these residuals on the original independent variables and lagged values of the residuals. The null hypothesis assumes no autocorrelation among residuals. If the calculated LM statistic exceeds the critical chi-square value, we reject the null, indicating presence of autocorrelation.
