A report on some Monte Carlo experiments which show that collinearity is not so important a problem as might be inferred from the frequency with which it is cited to explain poor results. One of the reasons given for concern about collinearity is that it may impair prediction. If one is willing to assume that the structure (the relationships among the explanatory variables) will remain unchanged, it is well known that collinearity does not impair prediction. The experiments described show that only in the most extreme case is predictive ability impaired when the underlying pattern of collinearity is altered.
Newhouse, Joseph P., Is Collinearity a Problem?. Santa Monica, CA: RAND Corporation, 1971. https://www.rand.org/pubs/papers/P4588.html. Also available in print form.
Newhouse, Joseph P., Is Collinearity a Problem?, Santa Monica, Calif.: RAND Corporation, P-4588, 1971. As of October 06, 2021: https://www.rand.org/pubs/papers/P4588.html