When Should a Football Manager be Sacked?
Football is renowned for its kneejerk analysis and a debate on yesterday evening’s Monday Night Club on Radio Five Live plumbed new depths in speculating over the relationship between Steve Bould and Arsene Wenger at Arsenal, with opinions scattered around like so much confetti and little evidence to back the arguments up.
It’s good, therefore, when a serious minded, analytical study emerges to throw light on an important issue and today, we are pleased to highlight a newly published paper that brings sophisticated statistical methods to bear on a key issue. The journal Economics & Finance Research is available open access and the article in question is available for download at this link. It certainly provides food for thought and here is Danny Lovatt from the publisher Routledge with a short introduction to an important contribution.
Was Sven unlucky to be dismissed at Manchester City? Did Mick McCarthy survive on borrowed time at Sunderland? Researchers from Henley Business School’s ICMA Centre have developed a model that could give directors of football clubs vital data when deciding on their manager’s future.
The model crucially evaluates the managers’ performance on their genuine ability to get the best out of the team. It shows managers’ points per game performance, when factors beyond their control such as transfer spending, wages and the absence of players are taken out of the equation. The research team comprised the ICMA Centre’s Professor Chris Brooks, Professor Adrian Bell and Tom Markham.
The researchers used a statistical technique, called ‘bootstrapping’ which has been used in the world of finance to evaluate the ability of mutual fund managers to outperform the markets, but in this case allows researchers to assess the performance of football managers on their ability alone. The technique compares the actual performance of managers over a period of time compared to a model that predicts a range of outcomes of what a team’s performance should have been. The real manager’s performance can then be compared to see if his team has performed significantly better or worse in real life as a result of his intervention.
The researchers employed match results and data from seasons 2004-05 to 2008-09 and the results echo the views of football experts. David Moyes was widely regarded to be over-performing in his role as Everton manager given the resources at his disposal. The research demonstrates that during the sample period, Moyes scored on average 0.54 points per game more than the 1.02 points per game expected which, over the course of a Premier League season, suggests Moyes’ influence made the difference between Everton pushing for a place in the Champions League and fighting a relegation battle.
The model also argues that several managers were, based on match performances, sacked unjustifiably during the period. Sven-Gà¶ran Eriksson’s contract was terminated by Manchester City at the end of the 2007-08 season even though evidence shows he outperformed 95% of results forecast in the researchers’ model that ignored managerial influence, and thus was performing significantly above expectations.
The model also identifies several managers who were performing significantly worse than expected before they were eventually dismissed – for instance, Mick McCarthy during his spell at Sunderland and Aidy Boothroyd while at Watford. The researchers conclude that, using their model, it could be possible to show that managers who are performing significantly below expectations after just 10 games are unlikely to improve their performance, and it might be best for the club to remove them from post sooner rather than later. On the other hand, managers who are only somewhat disappointing, rather than significantly underperforming, should be given more time to develop their teams.
Tom said: ‘In 2009, only six of the English Premier League’s clubs managed to make a pre-tax profit despite the Premier League clubs’ combined revenue of £2 billion. This is in large part due to the failure of clubs to control costs. The appointment and subsequent dismissal of the wrong manager can be extremely costly as managers are entitled to compensation if their contracts are terminated early.
‘It is widely believed that rich teams can buy success. However, our method is a big step forward as for the first time we are able to identify whether the number of points per game secured by the manager is due to the characteristics of the team or managerial skill.’
Read the full text of ‘The Performance of Football Club Managers: Skill or Luck?’ here.