[This article was first published on R Bloggers on , and kindly contributed to R-bloggers]. (You can report issue about the content on this page here)
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

In this blog post, I use the market value estimations from the Transfermrkt website to evaluate the influence of money in the top-tier of European football. I trace the evolution of market value for the top 5 European leagues, compute a measure of the market concentration as a proxy for competitiveness within each league, and then estimate the relationship between market share and performance by way of points accumulated by each club per season. I find that the top 5 European football leagues have relatively low levels of market concentration compared to industrial benchmarks, though levels of market concentration do vary from across leagues and over time. Regression results indicate that an increase of a club’s share of the total market by 1% is associated with an increase of 2.68 points earned per season.

## Background

Over the holidays, I had a bit of free time–and quite honestly, was in need of a little distraction–so I decided to have a look at the relationship between money and football using the wealth of data available on the excellent Transfermrkt website.

To the attentive football fan, the influx of money pouring into the top leagues in England, Spain, Italy, Germany, and France over the past decade or so is difficult to ignore: the splashy entry of foreign owners and corporate management of top flight clubs, increasingly frequent record-breaking transfers of players in all positions, new stadiums, new facilities, new global partners–the list goes on and on. It’s easy to conclude that this increased flow of money is having a negative effect on competition. It certainly seems that the biggest, i.e. wealthiest, clubs in each league are the only ones truly competing for domestic and European titles. In this analysis, I hope to ground this discussion in some data to provide a clearer picture on the size of this flow of money into top European football, to see how much of this money is being channeled towards a handful of top clubs in each league, and finally, to quantify what effect, if any, this has had on competition.

## Data

The main impetus behind this post is the availability of a large database of consistent market valuations for players in all of the top European football leagues available on the German website Transfermrkt. For the top five leagues in England, Spain, Italy, Germany and France, Transfermarkt provides a continuous series of data with the market valuations of all teams from the 2005-2006 season to the present. While these valuations are estimates, they are widely considered to be some of the best estimates available and the site is regularly cited by media and used for academic studies1.

A typical entry for a single season in each league contains the following information:

From these tables, we are able to calculate a total market value for all players for each league in each season, and likewise, compute each team’s share of the total market based on the valuation of their squad.
## Total Market Value

The first thing we will look at is the evolution of the total market value of the top 20 most valuable leagues in Europe. The list was taken from Transfermarkt’s current ranking, as of January 2020:

The site has complete market value data going back to the 2005-2006 season for the top five leagues (England, Spain, Italy, Germany, and France), and full data for most of the other teams in the top 20 going back to the 2006-2007 season. While I am primarily interested in examining the top five, I include the other countries in some of the plots below to provide some additional context.

### Total Market Value, 2018-2019 Season

When we examine the most recent season where we have a complete set of data, we can immediately see that there is a big difference between the top five leagues and the rest of Europe. The figure below shows the total market value for all first division players in the top 20 leagues in Europe for the 2018-2019 season.

Looking at the plot, we can see that Ligue 1 in France, for example, has a total market value of over 3 billion euros, which is three times greater than the next biggest league outside the top 5 (i.e. Portugal). In fact, the Portuguese league is the only other one outside the top five that has a total market value of over 1 billion, and we can see that over half of the top 20 European leagues have a total market value of less than 500 million euros.

### Total Market Value, 2005-2018

The evolution of the market value of the top 5 leagues between the 20005-2006 season to the 2018-2019 season further highlights the disparity between the top 5 and the rest in Europe.

The figure below plots the total market value of each league in millions of euros (€). While the gap between the total market value of the top European leagues and the rest of Europe was already present in 2005, over time, this gap has widened, and even seems to accelerate in more recent years.

### Total Market Value Growth Rates

To get a sense of the growth rates of the top leagues over time, we plot, below, the log transformation of the total market value of each league. In this figure, the movement up or down of each line corresponds to a percentage increase/decrease in total market value for the league in question.

What can be seen in this figure is that the growth rates for the top five European leagues seem to move mostly in parallel, and that the percentage growth is constant and positive over the entire time span. In contrast, looking at the movement of the rest of the top 20 European leagues (lines in grey), we see that the up and down movement of the lines is erratic, dramatic and often negative for long stretches over the same time period.

The difference between the constant positive growth rates of the top five leagues and the volatile rate of change in the rest of the European leagues can be seen even more clearly in the figure below, which plots the percentage change in total market value from season-to-season of all the leagues. As we inferred from the log-plot, above, the rate of change of the top 5 mostly moves in parallel, and is quite constant and positive throughout the time series. In fact, visualized this way, it is clear that, since 2016, the French and Italian leagues have seen a slightly sharper increase in the growth of their total market value compared to the others. Also, as before, the wide swings of the grey lines, i.e. the other European leagues, show that their total market values are not as stable from season-to-season, no doubt because their total market values are smaller and, thus, are more susceptible to the out-sized influence of big transfer deals into and out of those leagues.

Finally, we can look at the average growth rate of the leagues over the entire time period. Because our percentage growth swings between positive and negative values for many of the country leagues, I calculate a simple average annual growth rate and plot it below. We can see that the average annual growth in market value for top 5 leagues is around, or just below, 10% each year. While there are a few countries with higher annual growth, looking back at our first figure, we see that these countries are among the smallest in Europe’s top 20, and are unlikely to approach the size of top leagues anytime soon.

## Market Concentration

In this part of the analysis, I calculate a measure of market concentration for each of the first-division football leagues. Specifically, I use the Herfindahl-Hirschman Index (HHI), which is used by economists to describe how competitive a given industry is by measuring the relative market shares of firms within that market or industry.

The HHI is calculated by summing the squared market share of each firm in an industry, so that:
$HHI = \sum_{i=1}^n s_i^2$

where $$s_i$$ is the market share of each firm ($$i$$), and $$n$$ is the total number of firms in the given industry. An HHI score of 1, for example, would mean that the market shares are evenly distributed among all firms in a market (perfectly competitive market). The HHI increases as the proportion of small firms holding large market shares increases, up to a maximum of 10,000, which would mean there is one firm serving the entire market (a perfect monopoly). Squaring the shares of each firm in this calculation gives greater weight to the largest firms in the market.

The U.S. Department of Justice, for example, uses the HHI to quantify the potential impact of business mergers and acquisitions and decide whether they are in possible violation of anti-trust laws. According to the U.S. Federal Trade Commission (FTC) and the Dept. of Justice (DOJ), markets can be classified into three types based on the HHI score:

• Highly Concentrated Markets: HHI above 2500
• Moderately Concentrated Markets: HHI between 1500 and 2500
• Unconcentrated Markets: HHI below 1500

Here we use the HHI to provide a way for us to quantify the concentration of market power in each of our football leagues. Or to put it differently, the HHI provides us with a way of measuring how equally the shares of the total market value for players is distributed among all of the competing clubs in each league.
The table below shows the HHI scores for the top 5 as of the 2018-2019 season. By the criteria set out by the U.S. FTC/DOJ all of these leagues would be considered unconcentrated, competitive markets. While it might seem like seem like only a handful of rich clubs in each country are truly competing for titles–and in fact, this might still be true, as we will later–in terms of oligopoly-like concentration of market power among the few, by industry standards, low HHI scores here indicates that the leagues would be considered competitive. When one considers that top-division football is a very regulated market to begin with, then this result seems less surprising.

 country season hhi ESP 2018 1267.906 FRA 2018 1207.440 GER 2018 995.611 ITA 2018 862.154 ENG 2018 804.856

The figure below shows the HHI scores for the top 20 leagues by market value in Europe for the 2018-2019 season. While we can see there is some variation between countries, only a handful of leagues are considered moderately concentrated (Austria, Ukraine, Scotland, and Portugal), the majority are considered unconcentrated, competitive markets by our measure of market value concentration.

Looking over the entire time span covered by our dataset, the figure below plots the HHI scores for all leagues from 2005-2018, where the size of the point is given by the total market value of the league or country in question. In this figure, we can see that the English Premier League and the Italian Serie A have the lowest market concentration scores throughout the series, despite their growth in total market value over time. Conversely, we can see that the Spanish and French leagues appear to be increasing in market concentration over the years, with the French Ligue 1 really increasing its relative rate of market concentration noticeably beginning with the 2011-2012 season onward. The German Bundesliga seems to lie somewhere in the middle, though it’s overall trend over time is a slight increase in market concentration.

With regards to the HHI scores for the rest of the European leagues, corresponding to their volatile growth rates, we also see a correspondingly wide range of values, though with few exceptions, larger markets tend to have lower HHI scores, and very concentrated markets are mostly found in the smaller leagues.

## Impact of market share on points won

Finally, we can estimate the impact of market share on the performance of each club in their respective leagues. While clearly there is a relationship between the amount spent on players and the total points won by club per season, we can try to put an exact number on this relationship with a simple linear regression. In this part of the analysis, I will restrict the sample data to only the top five leagues of primary interest.

The figure below plots the relationship between the share of the total market value each club has in their respective league, per season, against the total points won that season. The data points in this plot are pooled from all clubs within each league over all seasons, and broken up by country. We can clearly see the strong, positive linear relationship between each club’s share of the total market and the points won. Similarly, we can see some evidence of the different levels of market concentration in the leagues we alluded to earlier–specifically, this is captured by the wider spread of points drifting to the right-hand side of the plots for Spain, France and Germany, signalling those leagues contain more individual clubs with disproportionately larger shares of the total market than the rest. Conversely, the tighter bunching of the points for England and Germany is evidence of the relatively less concentrated markets in those leagues.

When we look at this same plot broken down by season, as in the figure below, we can see how the market concentration levels have evolved over the years. Plotted this way, we can clearly see where, beginning in the 2009-2010 season, a few clubs in Spain and Germany really begin to pull away from the rest of the teams in their respective leagues. By the 2012-2013 season, the influence of the PSG takeover is also clearly evident in the French league. While there are clearly a few exceptions in every league in every season where the team that won the most points (i.e. the highest point on the y-axis) is not the one with the greatest share of the market (i.e. the furtherst point to the right on the x-axis), the exceptions are rare, and the positive relationship between the two variables is quite clear over all teams and leagues.

### Regression results

As a last step, we can estimate a simple linear model in order to quantify the effect of club market share on performance in the form:
$Y_{l,s} = \theta_l + \theta_s + \beta \, \text{ClubMarketShare}_{l,s} + \varepsilon_{l,s}$

where $$Y$$ is the number of points won by a team in each league ($$l$$) per season($$s$$). The two theta terms control for the fixed-effects of each league ($$\theta_l$$) and season ($$\theta_s$$), and $$\beta$$ captures the impact of each club’s market share on their point total. The regression results reported below, state that, a 1% increase in each club’s share of the league’s total market value is associated with a 2.68 increase in total points won. The effect of the club market share on points won is statistically significant.

 Estimate Std. Error t value Pr(>|t|) CI Lower CI Upper DF 2.68 0.087 30.8 1.86e-158 2.51 2.85 1353

## Some final remarks

Over the past decade we’ve seen enormous sums of money flowing into top flight European football, and the data presented here gives a clearer picture on how much the leagues have changed from the 2005-2006 season to 2018-2019.

 country mv_pct.change ENG 252 ESP 243 FRA 209 GER 239 ITA 176

The top divisions in England, Spain, France, and Germany have all seen increases in the total market value of players in excess of 200%, and the Italian league has also seen an increase of 175% over this time period.

On the other hand, while it may seem that all of this money is flowing into just a handful of top clubs, the data shows that, the concentration of wealth is more evenly distributed than it may seem–dependent on your league of reference. On my end, for example, as an avid watcher of La Liga, my impression that the money is being concentrated with the top 2, and increasingly 3 teams, is in fact, born out by the data. Since 2005, the Herfindahl Index for La Liga has increased by 25%. Similarly, the concentration of market value into the top clubs in Germany and France has become even worse.

On the other hand, the biggest league by total market value and the one that has also grown the most–the English Premier League (EPL)–also has the most even distribution of market shares among the top five leagues. Furthermore, the EPL has remained remarkably consistent in its very low levels of market concentration over the entire time period covered in this analysis.

 country hhi_change hhi_pct.change ENG 57.9 7.75 ESP 257 25.4 FRA 506 72.2 GER 234 30.8 ITA -155 -15.2

There are many other questions we can ask with this data, and I was not able to show everything in a single post here, but as always, constructive comments, questions and suggestions are always most welcome!

## References

1. Herm, Steffen & Callsen-Bracker, Hans-Markus & Kreis, Henning, 2014. “When the crowd evaluates soccer players’ market values: Accuracy and evaluation attributes of an online community,” Sport Management Review, Elsevier, vol. 17(4), pages 484-492.

To leave a comment for the author, please follow the link and comment on their blog: R Bloggers on .

R-bloggers.com offers daily e-mail updates about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

# Never miss an update! Subscribe to R-bloggers to receive e-mails with the latest R posts.(You will not see this message again.)

Click here to close (This popup will not appear again)