Working Paper: CEPR ID: DP10191
Authors: Jan C. van Ours; Martin A. van Tuijl
Abstract: This paper studies the causes and consequences of in-season changes of the head-coach of association football teams. We exploit data from the highest level of Dutch professional football during 14 successive seasons. An in-season change of the head-coach depends on recent match results and the difference between actual results and expectations as measured using bookmaker data. We find that, after the head-coach has been replaced, teams perform better than before. However, the performance is also better than before for a control group of coach replacements that did not occur. From this we conclude that replacement of head-coaches does not improve team performance.
Keywords: Association Football; Coaches; Performance
JEL Codes: J44; L83
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| in-season head coach dismissals (J63) | team performance improvement (L10) |
| control group of coach replacements that did not occur (Z22) | team performance improvement (L10) |
| recent match results (C72) | probability of coach dismissal (Z22) |
| cumulative surprise indicator (C43) | probability of coach dismissal (Z22) |
| team performance improvement (L10) | regression to the mean (C29) |