Are regression approach futures hedge ratios stationary?

In contrast to some recent research, this article finds that regression approach futures hedge ratios are stationary. It shows that a previous study's failure to reject the random walk null hypothesis was due to its small sample size and the overlapping hedge ratio calculation approach's bias toward accepting the random walk hypothesis. The impact of overlap on the Dickey‐Fuller full model intercept and slope estimates is demonstrated analytically and numerically. Finally, the article shows that out‐of‐sample hedging performance is not significantly improved by updating the hedge ratios. © 1998 John Wiley & Sons, Inc. Jrl Fut Mark 18:851–866, 1998