|
When the errors in a regression model have AR(1) serial correlation, why do the OLS standard errors tend to underestimate the sampling variation in the j? Is it always true that the OLS standard errors are too small? |
| New search. (Also 1294 free access solutions) |