Tuesday 2 April 2013

The Relationship between Spending and Performance – Correlation and Causality



Introduction


The rhetoric from UEFA regarding Financial Fair Play seems to be that it is essential for the long-term health and sustainability of the game that clubs are encouraged to be responsible and ‘live within their own means’. A joint statement released last year from the European Commission and UEFA outlines that the aims of FFP are to:

        Improve the economic and financial capability of clubs;
        Increase transparency and credibility;
        Improve governance standards in football;
        Encourage clubs to operate on the basis of their own revenues;
        Introduce more discipline and rationality in club finances;
        Protect the integrity and smooth running of UEFA club competitions;
        Encourage responsible spending for the long-term benefit of football;
        Protect the long-term viability and sustainability of European club football.

Furthermore it is claimed that “These are objectives which UEFA, as governing body for football in Europe, will promote in a balanced and proportionate way, acting in accordance with all applicable legal rules and, in particular, within the framework of EU law” http://ec.europa.eu/competition/sectors/sports/joint_statement_en.pdf.

We have seen in the past that football is not immune to general European legislation, in particular we can look at the Bosman Ruling of 1995 as an indication that any infringement on general European Union Law, such as the free movement of workers, is likely to end up in the ECJ. European Union Competition Law is intended to prevent companies abusing their market power, indeed with regards to sport the European Commission states:

“Most sport cases have been handled under EU antitrust rules, which prohibit anti-competitive agreements and practices as well as abuse of a dominant position. These cases concerned revenue-generating activities connected with sport, such as media rights and ticket sales and regulatory/organisational aspects of sport” http://ec.europa.eu/competition/sectors/sports/overview_en.html.  

It is therefore perplexing that the EC appear to be fully supportive of FFP when there is strong evidence that clubs in dominant financial positions are going to benefit the most from the new regulations. Recently, Arsenal, Liverpool, Manchester United and Tottenham submitted a statement to the FA requesting tighter financial control for the Premier League in line with UEFA’s FFP regulations. It is no coincidence that these are traditionally wealthy clubs with high revenues, seeking to limit spending to a club’s self-made resources. Additional measures for the Premier League have now been passed by the narrowest of majorities meaning FFP is not just something that clubs wanting to compete in Europe should be concerned with. I among others have argued that these regulations are anything but supportive of a sustainable sporting environment as UEFA claim, primarily because they limit competition, which in the long run will be detrimental to the health of the game.

Data


The data (kindly supplied by Stefan Szymanski) comes from all four professional English leagues over a ten-year period from 2000/01 to 2009/10. The data is not complete with figures missing from a number of clubs, particularly from lower divisions, however the volume is sufficient for this analysis. The data covers Final League Positions, Revenue, Wages, Profit/Loss and I have added the additional variable Spending. I have calculated spending as a team’s revenue minus or plus their profit or loss respectively, i.e. a team with a revenue of £100 million and a profit of £10 million has only spent £90 million, whereas a team with the same revenue and a loss of £10 million has spend £110 million.

Assumptions


It is commonly thought that football clubs that spend more are likely to perform better in the long run. This is because of the logical assumption that performance is based on talent i.e. a team with more talented individuals is more likely to succeed in the long run than a team with less talented players, as well as the equally logical assumption that more talented players will cost more money to obtain (transfer fees), and maintain (wages). This is an important assumption for analysing the relationship between performance and spending.
 

Part 1 – Correlation


Correlation analysis is a statistical measure of how two variables relate to each other. I am interested in how performance compares to spending and will examine the data holistically, annually and annually by division.

Overall Relationship


The graph below shows all ninety-two teams over all ten years. I have applied a Spline Interpolation Line to the data.
 
 
The first thing you will undoubtedly notice is a sharp rise in spending from around 20th up to 1st demonstrating the obvious financial clout of the Premier League. For the remaining leagues there is a gradual slope as the line progresses further down, but the gradient becomes almost negligible in League 2 (Division 4). There is a clear difference in variance between Premier League team spending and the remaining three leagues, which appear much closer. For example, results from the Championship range between £2.6 million and £72 million, whilst the Premier League has a range from about £21 million to over £320 million. You can even go as far as to say that the Premier League has its own split with teams regularly finishing above 8th exhibiting much larger spending than others. It is also interesting to note that no team has won the Premier League spending less than £100 million during this period. I have also calculated the Pearson Correlation for the entire data set and the value of 0.629 suggests a strong linear relationship between spending and performance.

Annual Relationship


To explore the annual relationship I have selected seasons 2000/01, 2004/05 and 2009/10 to provide a spread across the entire period. I have not included all years for the sake of repetition, but rest assured they all look pretty similar to these three. I have used LOESS regression as this allows us to identify an approximation using a localised trend, i.e. the calculations give higher weighting to data closer to the point in question.
 
 
 
 
 
 
 
As you can see the pattern is very similar to that already established. You will also note that lower league spending has hardly changed over the ten years whilst Premier League spending has more than quadrupled from just over £50 million to just under £220 million, using the LOESS curve. Pearson Correlation gives us 0.736, 0.606 and 0.668 for 2001, 2005 and 2010 respectively, suggesting the linear relationship has remained strong over the entire period.
 

Annual Divisional Relationship

 
Using the same seasons as above I have analysed the data for each division separately. I have left out the fourth tier due to the lack of data.

The Premier League





The graphs indicate that teams in the lower half of the league tend to be reasonably close in terms of their annual spend, whilst the top positions tend to go to teams that spend the most. Teams that appear to over perform in terms of their relative spend do so because of bumper profits i.e. Arsenal in 2001, Everton in 2005; or turn out to be one season wonders i.e. Ipswich 2001. It is clear that in the Premier League money is not sufficient to guarantee success, but it is certainly necessary for sustained competitiveness at the highest level.

Championship


The relationship in the Championship is a lot less clear. The correlation is a lot less defined than in the Premier League, however, the winners of the league in each season tends to be one of the top spenders. In the entire period the winners of the English second tier only spent below £20 million on two occasions, and on only one occasion, West Brom in 2007/08, did any winning team avoid making a loss. This raises the question of whether these successful teams would have achieved this under stricter financial restraints; a question that we are likely to see answered over the next decade of football under such conditions. The average correlation for the period was 0.451. Clearly the relationship is nowhere near that of the Premier League, however, there is still empirical evidence to suggest that money plays a pivotal role in determining success.

League 1


The English third tier, League 1, is similar to the Championship. The correlation is a lot less significant than the Premier League, yet the data shows a slight upward inflection towards the top of the league. Average linear correlation is 0.499, so again there is a case for the importance of spending for success.

Part 2 – Causality


Correlation ≠ Causation

Correlation and Causation is not the same thing. If one variable causes another then they are almost certainly correlated, but just because two variables are correlated does not mean that one necessarily causes the other.

It is difficult to claim solely from correlation analysis that spending is the key driver of performance, and unlike correlation, causality is extremely difficult to establish using statistics. However, it would take a brave individual to argue that money is not a critical factor in determining long-term success.

Spending is the most plausible determinant of performance if we accept the critical assumption that talent drives performance and that more talent costs more money. If we accept that there is a link between talent and performance, and that better players cost more money to buy (transfer fee) and to maintain (wages), it seems irrefutable that the amount of money a club spends is going to significantly shape performance.

The graph below shows the relationship between Wages and Performance.
 
 
 
The outcome is, unsurprisingly, almost identical to that between Spending and Performance. This is because wages make up the bulk of every clubs annual outgoings. It is increasingly rare for a club to spend less than 50% of their revenue on wages. Despite the lack of robust scientific control studies it seems appropriate to call the correlation between Spending and Performance a causal relationship.

Conclusion


In previous articles I have argued that competition is critical to the long-term health of the game. This is fundamentally due to the fact that football, like any other professional sport, requires fans that spend money on ‘the product’ i.e. watching games (at the ground, on T.V.) and buying merchandise. In turn, media coverage grows with fan interest, whilst worldwide interest grows with the increased media attention. The amount of money television companies are willing to pay for the rights to show games also increases with a rise in interested spectators e.g. there is a reason the Premier League has a much higher total T.V. deal than La Liga. Competition is key!

Under Financial Fair Play spending is likely to become synonymous with revenue for the vast majority, meaning performance will be somewhat restrained and predictable. The idea that FFP will secure a more viable future for the game is delusional. In the short-term the clubs making losses will inevitably spend less, and their performance is likely to suffer for it. In the long-term teams in dominant financial positions will be able to consolidate their sporting positions at the top with an ever-increasing gap that will be impossible to traverse (see my previous article discussing Sass http://journaloffootball.blogspot.co.uk/2012/10/normal-0-part3-why-ffp-is-end-of.html). Predictability will ultimately destroy consumer interest in football; perhaps not for every supporter, but the global nature of the modern game means that, financially at least, the majority of clubs will be worse off under FFP.

If it really is the European Commission’s mandate to prevent the abuse of a dominant position, then to allow, and even support UEFA’s FFP regulations is nonsensical.