Say what serial correlation is and what it does to statistical inference

Serial correlation means errors are correlated across observations, common in time series. Positive serial correlation usually makes standard errors too small, so t-statistics are inflated and Type I errors rise. The coefficients stay consistent unless a lagged value of the dependent variable is one of the regressors, in which case they become inconsistent too. The Durbin–Watson statistic tests first-order correlation, while the Breusch–Godfrey test regresses residuals on the regressors and p lagged residuals and handles several lags. The correction is serial-correlation consistent (Newey–West) standard errors, which also allow for heteroskedasticity. A common slip is to forget that a Durbin–Watson test is invalid when a lagged dependent variable is present.