Question

Explain the problem of Autocorrelation. What are the consequences of Autocorrelation? What are the corrections to Autocorrelation?

Answer 1

Autocorrelation is a problem in time-series regression. It refers to a problem in data collected repeatedly over a period of time. The data iscorrelated with itself. As a consequence of autocorrelation, the error term observations follow a pattern. From such patterns, we can infer that something is wrong.

Corrections to Autocorrelation:

Method 1: The model specification is left as it is, but the confidence intervals are expanded around the regression coefficients to account for the violation of the model assumption of non-autocorrelated errors.

Method 2: The model specification is changed to obtain non-autocorrelated errors.