By Pete Dawson
Since the coronavirus pandemic many sports have been, and continue to be, played behind “closed doors” (i.e. in empty stadiums). In England, this has included: Premier League and Championship football; Test cricket and limited overs contests; rugby league and rugby union; and two Formula 1 races at Silverstone, amongst others.
The absence of crowds clearly has financial implications for the revenue of teams. The fact that these events took place at all indicates the importance of broadcast revenue, especially at the elite level of sport. In the case of professional football, for example, matchday revenue accounts for less than 20% of total revenue of teams in the top tier of the so-called big five leagues as reported by Deloitte in their most recent review of football finance (see Figure 1 below). In contrast, broadcast revenue accounts for well over 50% in most of these leagues. Further down the football pyramid the reliance on matchday income increases although there is significant variation between clubs.1 If the requirement of empty stadiums continues well into the 2020-21 season, then many smaller teams could be faced with the very real prospect of going out of business.
As well as the obvious loss of matchday revenue, another side effect of playing in empty stadiums is the effect on home advantage. In a previous blog earlier in the year (see here), I suggested that playing in a empty stadium would lead to a reduction in home advantage because the (social) pressure exerted by a largely partisan home crowd that would benefit the home team – either directly because of improved performance of the home team relative to the away team, or indirectly via the match official who would tend to give decisions that favour the home team – would now be absent.
Between mid-July and early August, I employed three interns – Lucy Day, Akshil Shah and Martha Skillern – as part of the School of Economics Summer Internship Scheme to help investigate this. We went about collecting and analysing team performance, referee performance and match outcomes across the top leagues in England, Germany, Italy and Spain.
Our principle aim was to see if referee decision-making changed when matches were played in empty stadiums. We provide here our initial findings based on preliminary data analysis.
Figures 2 and 3 show the mean number of yellow and red cards with and without crowds across the four tier 1 leagues (Premier League, Bundesliga 1, Serie A and La Liga) for the 2019-2020 season. Both figures suggest that the mean number of cards fell for the away team when crowds were absent. For yellow cards, offences fell by close to 21% on average, with red cards falling by 23.5%.
In the case of penalties awarded (Figure 4) we see that the away team received a penalty once every 6.5 games with a crowd but once every 5 games without a crowd. Although an increase was also observed for the home team too, the magnitude of the increase was smaller (and not significant). Overall, therefore, the away team incurs fewer yellow and red cards and more penalties in the absence of a crowd, which is consistent with the view that in the absence of social pressure decision-makers – in this case match officials – appear to make fairer (less biased) decisions.
But what about the performances of the players / teams? Does the absence of a crowd affect them too? Could this provide an explanation for observed differences in referee behaviour? To consider this we looked at possession and pass success and found very little difference with and without crowds. We also found that in terms of penalties awarded, the likelihood of scoring (or missing) remained almost the same with and without crowds for both the home team (85.39% with versus 84.34% without) and away team (77.71% with versus 78.75% without). What we did find was the home team made fewer shots and had fewer corners. In the case of the former, the home team, on average, made 1.5 fewer shots per game but otherwise not much of significance.
So, it appears the observed differences are largely due to differences in the behaviour of referees. Ultimately, did it affect match outcomes? Well, rather confusingly yes, and no! On the one hand the percentage of home wins across matches did fall, from 44.08% to 41.38%, however this difference was not significant in a statistical sense, so overall rather inconclusive.
Perhaps it affects some teams more than others? To consider this we looked at differences in match outcomes between “big” teams and “smaller” teams. Here we define a “big” club objectively using valuation data provided by respected German-based website Transfermarkt.com and the 2020 Football Money League produced by Deloitte. From this we identified 14 teams across the four leagues who had valuations significantly greater than the other teams (and if you are a football fan, you can probably guess quite a few of these). Perhaps unsurprisingly, when these teams played in their home stadiums home advantage was higher when crowds were present (69%). But interestingly, the home advantage of the larger teams increased further (to 77%) when crowds were absent, although the difference was not significant at conventional levels. In contrast, for the “smaller” teams, home advantage drops from 39% with crowds to 33% without, and the reduction is statistically significant.
What is the explanation for this? Clearly both big teams and small teams have been affected by the absence of crowds, but the bigger teams with larger squads and greater depth in terms of playing talent were better able to take advantage of the (temporary) increase in number of substitutions allowed (in the English Premier League, for example, the increase was from three to five) as well as the increased demands associated with playing the remaining games of the season in such a short period of time.
One final test we consider here is to change the reference point. In the analysis provided thus far we have used the first part of the 2019-2020 season as the reference point. An alternative reference point would be to look at the corresponding period, or more precisely the corresponding games, from the 2018-19 season. This would make sense especially if one suspects that home advantage might be affected more by matches that take place towards the end of the season where more contests have “high-stakes”, and the pressure from the (home) crowd becomes more vociferous. Our analysis confirms this. But it now identifies a larger reduction in home advantage (on average, by 6 percentage points) which is now significant. As before, the effects are more pronounced – that is the reduction is larger – for the smaller teams.
The 2020-21 season has now started and this last weekend (19th / 20th September) saw a small number of fixtures from the English Football League welcoming spectators, albeit very limited in number, to their venues. It remains to seen what impact, if any, this has on home advantage if this pilot continues and, ultimately (and hopefully), extended. However, it will still be some considerable time before stadiums will be operating anywhere near full capacity, which will clearly continue to affect the financial position of teams, especially those that rely on match day revenue. But as we have seen here it is also likely to continue to affect home advantage and potentially widen the gulf between the bigger, wealthier teams and the smaller, poorer teams.
Footnote
[1] This variation not only occurs between clubs but also the same club over time, especially for those who have received parachute payments, which are given to clubs for a limited number of seasons following relegation from the Premier League. Take Norwich City as an example. In the four seasons between 2014-15 to and 2017-18, the proportion of matchday revenue accounted for between 11.77% and 20.54%. During these seasons Norwich were either in the Premier League or in receipt of parachute payments. In 2018-19, without Premier League status or parachute payments, the figure jumped to 28.78%.
Reference
Deloitte (2020), Annual Review of Football Finance.
Banner Image by Vienna Reyes on Unsplash