Managing Autocorrelation

There are a number of ways to manage autocorrelation. These include:

• Adding predictors (explanatory variables) to a model.

• Cochrane-Orcutt method.

• Using lagged versions of the explanatory and/or response variables.

• Using a regression model that allows adjustment of the model covariance structure. (autocorrelated or time series residuals)

While the methods listed can be effective, some prefer methods based on ARIMA, GARCH, or similar stochastic models.