
A 2021 paper in the Journal of Business Economics finds that the introduction of VAR eliminated the stoppage-time bias of Italian referees (or to be more precise, referees of Italian Serie A games). The statistical analysis shows that, before the introduction of VAR, Italian referees reduced the amount of stoppage-time when the home team had a slight lead. After the introduction of VAR there was no evidence of such bias. By contrast, those refereeing in German games did not show a stoppage-time bias (in either the pre-VAR or post-VAR periods).
How might we explain the statistical results in the last paragraph? As economists, possibly ones influenced by the presentation of material in books like Freakonomics, we might look to the incentives. Let us look at the financial incentives. The first footnote in the same Journal of Business Economics paper says German and Italian referees are paid different amounts. Both are paid a fix annual fee of €80,000. However, those doing German games get €5,000 per game whereas those doing Italian games get €3,800 per game. Now we can tell a story or two. Better pay means better performance. Or Italian referees are borderline corrupt due to "low" pay and the introduction of monitoring technology like VAR is elimination some of these questionable practices.
The above is my spin on some of the results presented in a paper by Ulrike Holder, Thomas Ehrmann and Arne Konig. It is my manipulation of their result. The authors are much more measured in the presentation of their results. They explicitly say that their paper shows that referees are NOT the channel through which potential home bias might operate. Unlike my presentation above, they highlight the small size of the bias in Italian games. It is less than 9 seconds. When are games likely to be decided? In those 9 seconds or in the other 5,400+ seconds? Or in any one of the 5,409+ seconds? If one wants to rely on the statistics and probabilities then one should do it consistently.
Last week I posted a piece about how the Holder, Ehrmann and Konig paper raises important issues in the identification of home bias and the channels through which home bias might operate (here). Hidden in the text and reference list is another paper worth some consideration. It gets away from the preoccupation with psychological explanation of home bias to examine the impact of physiology. That will be the subject of my next post.