Why doesn't Sean Dyche sign more black players?
This year is the 20th anniversary of a landmark research paper in the economics of football. Stefan Szymanski, then at Imperial College and now at the University of Michigan, wanted to measure the degree of prejudice against black footballers in England from the late 1970s through to the early 1990s. Reasoning that performance in football should reflect a team's spending on players, he compared wage bills versus success in the table to see whether teams fielding more black players seemed to finish higher than expected. If they did, then black players might have been underpriced in the market, suggesting racial discrimination. The paper did find evidence for this notion and was published in the Journal of Political Economy, arguably the most prestigious journal in academic economics.
As we reconsider the treatment of black footballers and coaches today, we thought it would be interesting to study the recruiting patterns of coaches at the highest level. So we studied the histories of all 20 coaches in the Premier League, gathering all their signings in the top four tiers of English (and Welsh) football all the way back to 2001*. Without inferring anything about performance, we wanted to see how uniform the recruitment of black players was among the coaches in this internationally diverse group.
The purpose of this article is not to accuse Sean Dyche, or any other coach, of prejudice against black players. Rather, the purpose is to start a discussion about factors that may affect the recruitment of black players in football, and Dyche's signings display an unusual tendency that is statistically striking but difficult to explain with data.
Before we start, a disclaimer: We classified a player as black if his biography indicated ethnicity that would generally be called black or Afro-Caribbean in United Kingdom. If we have misclassified anyone, we are truly sorry and will make any necessary changes as soon as possible. Also, the diversity in a squad clearly does not come just from black players; we are not measuring recruitment of players with East Asian, South Asian, North African, Arab, Latino or other ethnicities here.
The first thing we noticed was the concentration of coaches around a share of roughly one third black players among their signings. Of course, there were also outliers on both sides:
Because the number of signings for each coach can vary so much, a better way to view the data might be on a chart like this one, where we've added a line that crosses through all points with a third of signings estimated to be black players:
We'd like to believe that the majority of coaches did not discriminate or have any preference related to black players, so we might assume that roughly a third of the player pool for the Premier League consisted of black players. Under that assumption, some coaches signed significantly more or fewer black players than expected.
We say "significantly", because we can compare these coaches' signings to the results of a random recruitment process to draw some statistical conclusions. Let's start with Steve Bruce's signings.
Bruce signed 77 black players out of a total of 213. Our hypothetical recruitment policy would pick 213 players at random out of an infinite pool. Under our assumption, we'd expect to pick about 71 black players, or one in three. So how unusual was it to pick 77? Well, if we went through our random recruitment process 10,000 times – and we actually had a computer do this – then we'd pick 77 or more black players around 20% of the time. That's not too frequent, but in Bruce's case there's still a decent chance that it was a statistical fluke.
On the other side of the dotted line in the chart above is Chris Wilder. He signed 48 black players out of 210, where a random recruitment process might have chosen 70. In fact, there was less than a 0.05% chance – that's one twentieth of one percent – of picking 48 or fewer black players randomly. The same went for Dyche, who has had fewer signings since 2001 but an even smaller share of black players.
There are bigger outliers than Bruce above the line, too. With our assumptions, there was only about a 15% chance of randomly picking as many black players as Brendan Rodgers chose to sign. In Slaven Bilic's case, it was around 4%. But neither of these was as much of a statistical improbability as Wilder or Dyche's dearth of black signings. In fact, Bilic's recruitment record was about 100 times more likely to occur randomly than Wilder's or Dyche's. We'd only expect to see a record like Dyche's once among 2,000 coaches, if he truly chose black and non-black players at random. In this study, we're looking at 20 coaches.
So why didn't Dyche and Wilder sign more black players? The answer doesn't have to be prejudice. After all, Nuno Espirito Santo, the only black coach in the Premier League, has also signed rather fewer black players than most of his colleagues – likely a consequence of Wolves' Portuguese-heavy recruitment strategy.
One possible explanation is that there were fewer black players in the lower tiers of the EFL, where both Wilder and Dyche operated during the sample period. There are plausible reasons for this to be true. First, lower-tier clubs may rely more on their local catchment areas and may also be located in towns where fewer black players grow up. If that's the case, then we shouldn't be comparing EFL recruitment records using the 1-in-3 standard assumed above.
But that's not the only potential reason for seeing fewer black players in the EFL. For one thing, discrimination could come under less scrutiny in areas with smaller black communities. And clubs with smaller budgets may also find it easier to discriminate in the way Szymanski documented, since their revenue generally depends less on performance than it would in the Premier League. In the top tier, every place in the table comes with a different payoff, with the added extremes of relegation and qualification for Europe. That's not the case further down the pyramid, where moves in midtable offer no financial reward.
In any case, since Dyche has come to the Premier League, he has signed only 3 black players out of 29. For Wilder, that figure is 10 out of 22. So if the EFL effect is real, then Wilder seems to have left it behind upon joining the top tier. Dyche apparently has not.
So it would be great to hear Dyche's own take on the paucity of black players in his squads. Even Burnley's U23 squad, which is not under Dyche's direct control, has a higher share of black players than the first team. Is it just a statistical quirk, or can he offer other reasons that might shed some light on the challenges black players continue to face?
* We considered a player to be a new signing if he had not already been at the club contiguously under the coach in question. So a loan player who became a permanent signing under the same coach would only count once. However, a loan player who was at the club when the coach arrived and then became a permanent signing under the coach would still be counted. A player who came to the club, left for at least a season, and then returned would also be counted upon his return.
[Photo: Brian Minkoff/London Pixels]
