Throughout the pandemic, the policy tagline has been ‘flatten the curve’. Usually in relation to exponential growth in cases and hospitalisations, but one side effect of the pandemic has been that the exponential growth in managerial turnover has been flattened too.
Some things remain the same though – tried and tested, and not necessarily new, or dynamic, managers still retain their positions at the top of the game. In December, struggling Premier League team West Brom sacked their manager and appointed Sam Allardyce, who at that point had 1,057 matches of experience as a head coach.
Economists Thomas Peeters, Stefan Szymanski and Marco Terviö have conducted research into this “inefficient advantage of experience in the market for football managers”. They find that “in around one quarter of all cases, where a firm hires an experienced worker, this experienced worker has an estimated ability below the average ability of recent labor market entrants”. Allardyce has won just 2 of his 17 matches in charge of West Brom and the team remains heavy favourites to be relegated out of the Premier League.
Thinking more broadly about managerial turnover, if we take the data that Soccerbase provides, and plot the number of managerial changes per year, it looks a lot like the plots of Covid-19 cases – exponential growth before finally the growth abates. The black line below is all managerial changes per season, whereas the red line is all managerial changes for English clubs only – the website is dominated by English football.
The difference between the black and red lines shows that the exponential growth is mainly related to the expansion in data coverage on the website since around 2000, adding in a significant number of non-English leagues. That said, the red line, English clubs only, is still increasing up until recently.
It looks like Covid-19 has helped flatten the curve though, since the observations for the current and last season, both of which heavily influenced by the virus, show sharp falls. This doesn’t seen controversial. Soccer club owners are perhaps less likely to bite the bullet and sack a manager if fans are not in the stadium shouting abuse at the manager (and the owners too).
We can take a look at the average length of a managerial tenure to potentially shed a little more light on this hypothesis. In the plot below, in black is the mean tenure (number of matches) for a managerial spell ended in a particular season, and in red is the median tenure.
The last two observations show that the mean tenure of all managerial spells ended in 2020/2021 season has been 58 matches, down from 85 in 2019/2020 season, 79 in 2018/2019, and 91 in 2017/2018. The median tenure has been essentially unchanged in that time. So managerial spells don’t seem to be getting any longer, even if the number of severances has fallen.
It is also the case that if we divide by the number of soccer clubs in the dataset, to get an idea of the number of managerial changes per club, we get a different picture too.
We see from this picture that the number of changes per club was increasing linearly throughout soccer’s existence since the 1880s, until around the year 2000. Since then it’s been quite flat, with a bulge in the latter half of the 2010s, which has dissipated since then. Currently there’s about 0.4 managerial changes per soccer club per season – hence if a manager lasts two seasons, the one to come is more than likely his last one.
That drop back from the late 2010s bulge does coincide with Covid-19, however. Fewer managers are being chopped, but they haven’t really lasted longer since fewer games have taken place thanks to Covid-19 in the last couple of years.