In the marble halls

Competitive inequality in Spanish football

This post analyses La Liga, as part of ITMH’s competitive inequality project, tracking the impact of financial inequality upon football. The introductory post explains the measures that are used in the analysis. See the other posts in this series for commentary on top flight football in Germany, Italy and England, or the overall discussion of trends within Europe’s top leagues.

La Liga, the Spanish Primera Division, was founded in 1929, but the analysis below tracks performance across the period since its resumption after the end of the Spanish Civil War. The common perception of the league is that it is pretty much a duopoly between Real Madrid and Barcelona, who have recorded 35 and 27 titles respectively. Other prominent Spanish sides include Atletico Madrid, Valencia and Athletic Bilbao.

Spain also possesses the most success in the European Cup/Champions League, mainly due to Real Madrid’s 14 victories, with Barcelona also having lifted the trophy five times. These two clubs, as well as Atletico Madrid, were participants in the proposed European Super League, with Real’s president, Florentino Perez, slated to be the chairman of the new league. As of 6th June this year (when Juventus finally abandoned the project), Real and Barcelona remain the only clubs still publicly committed to the breakaway proposals.

Fig 1: correlation

(Bars here represent the individual correlation value between each season’s finishing positions and those in the preceding season. The red line is a five-year rolling average to pick up on trends in the medium term. The dotted line is an overall trendline representing the overall direction of travel for this measure across the results. See the introductory post to this series for an explanation of the choice of measurements.)

The chart here displays a general upward trend. The initial period after the Civil War was one of real flux, including one season where finishing positions were negatively correlated: i.e. a team with a high finishing position the previous season was actually (slightly) more likely to subsequently finish in the lower parts of the league and vice versa. In general, the results vary widely from year to year, with some years of greater stability and others of greater variation. The past ten years mark something of a departure from this, with consistent figures of around 0.6 or higher. As a result, the rolling average has hit higher levels than any previous period. This points to a greater level of predictability of league finishing positions over this period.

Fig 2: turnover

(Lines here represent five-year rolling averages for the turnover measure, assessing levels of variation among clubs finishing in the top two, top four, top six and top ten (in seasons with a 20+ team league) of the league. NB – since the calculation of this measurement relies upon future finishing positions, the current season’s figure can only be known in three year’s time. The latest figure, therefore, is for the 2019-20 season. See the introductory post to this series for an explanation of the choice of measurements.)

As noted above, the perception of La Liga is that it is almost always simply about Real Madrid and Barcelona, so I’ve included a top two turnover calculation. In fact, what this shows is that is has rarely been the case that these two finish first and second season after season – it has only really been the case that this has occurred at all in the late 1950s/early 1960s and from the mid-2000s onwards. So, while these two clubs are the undoubted powerhouses of Spanish football, they have always faced a challenge from other clubs for absolute supremacy, with Atletico providing the most recent disrupting force.

There may, however, be the threat of a triopoly taking control at the top. The most notable recent trend here is a sharp decline over the past fifteen years in top four turnover. In fact, the last time any of the three Spanish clubs involved in the European Super League finished outside of the top three was in 2011-12, when Atletico finished fifth. This does represent a real worry for the state of competition in Spanish football.

Fig 3: points per game

(Bars here represent the average points won per game by the team that finished champions that season. The red line is a five-year rolling average to pick up on trends in the medium term. The dotted line is an overall trendline representing the overall direction of travel for this measure across the results. See the introductory post to this series for an explanation of the choice of measurements.)

The rolling average here displays three clear and prominent waves, corresponding to the eras of three great teams, with performance falling away in between. The first era is that of Real Madrid dominance in the 1950s and 60s, stretching from the Di Stefano/Puskas team that swept all before it in European football through to the Yé-yé team who claimed a sixth European Cup in 1966. The second, slightly lower wave corresponds to the era of Real’s excellent Quinta del Buitre team of the late 1980s who, while ultimately falling short in the European Cup, won five successive La Liga titles.

Then, the most recent – and highest – of the waves is that initiated by Pep Guardiola taking up the coaches position at Barcelona. Subsequently, that Barcelona team won eight of the next eleven titles. This era has also marked a predominance of Spain’s top two clubs in European football, winning seven out of ten Champions League trophies from 2009 to 2018. While the graph shows Spanish football is clearly on the down-slope from that peak of brilliance, present levels remain consistently higher than all but the highest performances from the past.

Fig 4: goal difference per game

(Bars here represent the average goal difference per game recorded by each season’s title winners. The red line is a five-year rolling average to pick up on trends in the medium term. The dotted line is an overall trendline representing the overall direction of travel for this measure across the results. See the introductory post to this series for an explanation of the choice of measurements.)

The same three wave formation is visible when looking at goal difference trends, yet with the first and third waves being even more prominent, attesting to the greatness of sides in those eras. It does become clear here how much the period that has just ended in Spain, of Messi at Barcelona and Ronaldo at Real Madrid, will come to be seen as a time of otherworldly prowess. Just take a second to reflect upon this: over the period Ronaldo was at Real (2009-2018), he and Messi made 601 La Liga appearances between them, scoring a total of 640 goals. That’s an average of 1.06 goals per game each. For a period of nine seasons. (For perspective Erling Haaland’s scoring exploits in this season’s Premier League actually fall short of that figure, with him averaging only (!) 1.03 goals per game.)

By the way, that dip right in the centre of the latest wave – that’s the season Atletico battled their way to the title in 2014. While they were far from playing a spoiling style, they were unable to match the extraterrestrial performance of their rivals.

The past five seasons have seen this measure return to those more typical of overall averages.

Summary

Although its history has witnessed several spectacularly brilliant teams, La Liga has probably been more competitive than many would imagine. Indeed the excessive prominence of its two biggest sides is something that has only really become a phenomenon in the 21st century. Even now, however, the worry for levels of domestic competition may be more about whether the top three sides are attaining levels that the rest of the league is increasing unable to match.

In respect to the status of Spanish teams in Europe, the recent era saw a period when Real Madrid and Barcelona were regarded as the ultimate club sides in the world. For any player with aspirations of being considered the world’s best, playing for one of these sides seemed to be non-negotiable. As financial might has shifted towards the Premier League in recent seasons, this currently no longer seems the case. It will be interesting to see whether this is simply a short-term lull in their fortunes, or whether they are in danger of being permanently eclipsed at the top of the club game. The ongoing financial woes of Barcelona and the persistence with which they and Real have clung to the wreck of the European Super League suggests the latter is a very real possibility.

Competitive inequality in Italian football

This post analyses Serie A, as part of ITMH’s competitive inequality project, tracking the impact of financial inequality upon football. The introductory post explains the measures that are used in the analysis. See the other posts in this series for commentary on top flight football in Germany, Spain and England, or the overall discussion of trends within Europe’s top leagues.

Serie A has formed the single top national league in Italy since the 1929-30 season. The analysis below tracks performance across the period since its resumption after WWII. Over this period, three teams have consistently claimed most success: Juventus (36 championships in total, 29 of which are post-war); Internazionale (19 championships – 14 post-war); Milan (19 championships – 16 post-war). Other major contending teams include the two Rome clubs (Roma and Lazio), Napoli and Fiorentina.

The three dominant Italian clubs have also won 12 European Cup/Champions League trophies between them (Milan 7, Inter 3, Juventus 2). All three participated in the European Super League proposal, with the Juventus chairman, Andrea Agnelli, being thought to be one of the prime protagonists behind the plans. Indeed, Juventus only signalled their intention to pull out of the enterprise in recent days (6th June 2023).

Methodological mithering: in keeping with the national talent for Machiavellian subterfuge, Italian football has frequently descended into scandal, resulting in much greater use of points deductions than in other major leagues. This leaves a choice to be made when assessing league finishing positions between what has been determined by performance on the field and that subsequent to the application of penalties. E.g. in 2005-6, Juventus finished at the top of the Serie A table, but were subsequently penalised for their involvement in the Calciopoli scandal by being officially placed bottom of the table and relegated to Serie B. For the sake of this exercise, should their finishing position be recorded as 1 or 20; should it be their record, or that of the team subsequently awarded the title (Inter) that counts to points per game and GD per game calculations? My choice has been to, throughout the Italian results, use a raw league table based upon finishing positions as if there were no deductions, as this feels truer to the purpose of the assessment by reducing any artificial volatility.

Fig 1: correlation

Of the four leagues I’m looking at as part of this exercise, Serie A generally exhibits the highest average correlation between finishing positions from year to year. While the individual bars show great variation, the rolling average is typically operating at a slightly higher level than other leagues. This suggests that Italian football may possibly have developed an established order of clubs at an earlier stage, and have maintained this more thoroughly than other countries. It certainly is pronounced that Italian football is dominated by teams from its largest (and in the case of its three dominating clubs, its most affluent) cities, with fewer provincial upstarts than elsewhere. This lends credence to the idea that finance has mattered more, for longer, in the Italian league than with other major European leagues. Further support for this idea would come from the fact that, while the overall trend for this measure is upwards, the increase is less pronounced than that seen in some other leagues, suggesting that the effects of money are already baked into Italian football to a greater degree.

Also of note is the fact that recent years (since 2011) have been characterised by sustained higher levels of correlation, which have driven the averages up steadily.

Fig 2: turnover

(Lines here represent five-year rolling averages for the turnover measure, assessing levels of variation among clubs finishing in the top four, top six and top ten (in seasons with a 20+ team league) of the league. NB – since the calculation of this measurement relies upon future finishing positions, the current season’s figure can only be known in three year’s time. The latest figure, therefore, is for the 2019-20 season. See the introductory post to this series for an explanation of the choice of measurements.)

There is little in the way of overall trend for the turnover measurements for both top four and top six finishing positions in Serie A. With each, there is fluctuation between periods of greater or lesser turnover, yet it tends to consistently regress towards a mean of around 70% for top four turnover and 55% for top six turnover. The fact that top six turnover is in this range in a league with three dominant clubs suggests that, in most years, it is highly likely that these three finish in the top six, although they have not been able to colonise the top four, as evidenced by higher turnover at this level.

The measure for top ten turnover is only calculated following the expansion of the league to 20 teams in 2004-5 – it does not seem particularly worthwhile to calculate earlier than this for a group that would comprise more than half of the league. Yet, the short period in which this has been tracked sees a rapid decline in turnover levels, suggesting some degree of consolidation among more successful clubs that will be worth watching in future.

Fig 3: points per game

The rolling average here traces a loose upward sloping u-shape, being higher in the immediate post-war, lower through the 1970s, 80s and 90s (with odd exceptional higher performance such as from Juventus in 1976-77 and Inter in 1988-89), then generally higher levels of performance in the 21st century, capped by Juventus’s romp to 102 points in 2013-14.

Fig 4: goal difference per game

Italian football is all about tight, stifling defensive tactics, right? Well, yes, but it wasn’t always like that. You can clearly see the period in which Catenaccio fully emerged and became the defining Italian football mindset from the chart’s results – Helenio Herrara’s Inter team, who deployed it effectively the dominant force in both Italian and European football in the mid-1960s, became a model for much of what was to follow.

Prior to that, however, the immediate post-war period featured some incredible free-scoring performances from Italian champions. In 1950-51, Juventus, Inter and the eventual champions – a Milan side led by their Gre-No-Li Swedish attacking trio – each scored more than 100 goals in a 38-game season. But, the absolute pinnacle of post-war Serie A were the exploits of the dominant but ill-fated Grande Torino team who won five titles over the period of the war and its immediate aftermath. Their finest performance came in 1947-48, a season in which they scored 125 goals in 40 games, recording five or more strikes in half of their home matches, which included a 10-0 victory over Alessandria.

Unfortunately, tragedy would strike that team. In May 1949, in poor visibility from adverse weather conditions, the plane carrying them home from a friendly with Benfica in Lisbon crashed into one of the walls of the Basilica of Superga, which stands on a hill over Turin. None of the 31 on board, including team players and coaches survived the crash. Unlike Manchester United, who were able to return to prominence after a similarly tragic crash in the following decade, Torino have never again been a force in Italian football – they managed a solitary title in 1975-76, but have also had spells in Serie B and currently are a resolutely mid-table Serie A club.

In recent years, Italian title winners have been recording higher goal difference levels than the 1960s-1990s period. They have not, however, been able to get close to the incredible performance of that legendary Torino side.

Summary

This project is primarily interested in the impact of recent flows of money into football, and how this might be making the game more unequal and less competitive. The Serie A results show some evidence for this, with slight upward trends in correlation, reduced turnover in higher league positions and stronger performance from champions. However, it is also clear that Italian football has always displayed signs of stratification of clubs, making these trends less pronounced than they are elsewhere.

Competitive inequality in German football

This post analyses the Bundesliga, as part of ITMH’s competitive inequality project, tracking the impact of financial inequality upon football. The introductory post explains the measures that are used in the analysis. See the other posts in this series for commentary on top flight football in Italy, Spain and England, or the overall discussion of trends within Europe’s top leagues.

The Bundesliga was founded in the early 1960s, much later than other major national leagues. Since its inception, one club – Bayern Munich – has been consistently preeminent (winning 32 of a possible 60 titles), with the likes of Borussia Mönchengladbach (1970s), Hamburg (late 1970s/early 1980s) and Borussia Dortmund (1990s-present) providing the primary challenge.

German clubs have claimed eight European Cup/Champions League trophies, with Bayern the most successful, with six titles. No German clubs were involved in 2021’s abortive European Super League plans, although there are good grounds for thinking that, had the proposal been successful, two of the 15 proposed permanent members spots would have been offered to Bayern and Dortmund.

Fig 1: correlation

Despite the single team dominance of the league, results of the correlation analysis show consistent variation throughout the league’s existence. There are periods of greater stability (1980s, noughties), but also seasons characterised by major flux (1978-79, 1995-96, 2010-11, 2011-12). The overall trend across the results shows a gradual increase in correlation levels. In the shorter term, since the early 2010s there has been a steep upswing in correlation, meaning that we are currently in a period where finishing positions are more predictable and stable from year to year.

Fig 2: turnover

(Lines here represent five-year rolling averages for the turnover measure, assessing levels of variation among clubs finishing in the top four and top six of the league. NB – since the calculation of this measurement relies upon future finishing positions, the current season’s figure can only be known in three year’s time. The latest figure, therefore, is for the 2019-20 season. See the introductory post to this series for an explanation of the choice of measurements.)

The turnover measurements for both top four and top six finishing positions in the Bundesliga track each other pretty closely. For both, there are periods of relative consolidation in the early 1980s and across the 21st century. The graph shows the current period has been marked by a rapid short term decline in turnover, point to greater consolidation of top positions among a small number of clubs. Indeed, four teams (Bayern, Dortmund, RB Leipzig & Bayer Leverkusen) have all finished in the top six in every season since 2017-18. The chart shows that this level of concentration is something new within German football.

Fig 3: points per game

In the early years of the Bundesliga, the title-winners frequently ended up with a points total with an average of less than two points per game, with Bayern’s all-conquering early-1970s team providing an early exception by recording more than 2.3 points per game in both 1971-2 and 1972-3 seasons.

Since 2010, the chart displays a significant jump upwards, with subsequent champions (well, almost entirely Bayern – who have won the last 11 consecutive titles) frequently recording records that match or surpass that of the Beckenbauer/Gerd Muller era team. While the overall trendline shows a steady rise in performance by title winners, the recent trend records a decline over the past few seasons with a marked drop-off in 2022-23, as Bayern’s current team have declined (despite winning the title this season, it was clear to all that this year’s Die Roten side should no longer be considered a major force among Europe’s best). It will be interesting to watch whether this measure drops back to consistent historical levels (c. 1.9 to 2.2 points per game), marking Bayern of the 2010s as something rather special, or whether this season’s relative low is the aberration on a overall shift to a general higher level for title winners.

Fig 4: goal difference per game

This measure confirms the analysis of points per game outlined above: Bayern’s team of the early 1970s were something well above the general level of performance seen by German champions; however, their record has been consistently matched and sometimes surpassed by Bayern teams of recent vintage. Again, while the overall trend is upwards, the past few seasons have seen a slight decline. Future results will be worth watching, to assess whether or not recent levels of dominant performance by Bayern represent a new norm for German football.

Summary

In the Bundesliga trends across the data show greater correlation of finishing positions from season to season, declining turnover of teams at the top of the league and a shift towards more dominant performances by league champions. All of these measures are consistent with a decline in competitiveness and greater predictability of league outcomes.

ITMH’s competitive inequality project – introduction

Earlier this week, under angry, thundery skies, Manchester City paraded the spoils of a dominant season through the streets of Manchester. While their first Champions League victory allows them to finally claim a much longed-for place at the head of European football, their domestic dominance was already assured, with this season seeing them claim a fifth Premier League in the past six years, a period in which they have also managed six domestic cup victories.

Throughout football’s history, there have always been dominant teams. In the earliest years of the English Football League, Preston North End, Sunderland and Aston Villa vied for dominance; more recently Liverpool and Manchester United have both held the title claimed by City today of being English football’s unarguably pre-eminent power. In Europe, clubs like Real Madrid, Juventus and Bayern Munich have records of domestic success beyond anything displayed by any English team. Is it justified, therefore, for City’s current success to be consistently accompanied by hand-wringing about inequality? Is the situation today any less competitive than when Liverpool finished in the top two for ten straight seasons from the 1980s to early 1990s?

The posts within this project utilise four simple measures to attempt to assess the health of competition within the top European leagues and examine whether the sense that many fans possess – that the level of resources available to the teams at the top is harming the game – is reflected on the pitch.

Firstly, in a healthy, open competitive environment, you would expect to see a measure of flux in terms of league finishing positions. Each team is attempting to outperform others, to find tactical innovations or smart transfers that allow them to gain an edge over rival teams, so we should expect to see turnover in standings, both across each country’s highest league and also among the very top positions, as teams go through periods of relative success and failure. On the contrary, were success on-the-field to largely be a function of the off-the-field resources available to a club, we might expect to see a greater level of predictability and stability in finishing positions, with less movement and the emergence of some degree of sorting of clubs into tiers of relative performance.

The first two proposed measurements (adapted from measures used within a similar effort to measure levels of competition in commercial markets in Thomas Philippon’s The Great Reversal) try to capture this:

Year-to-year correlation of finishing position. Simply by assessing how much a team’s finishing position in a league table correlates to their position the year before, we can track the extent to which a finishing place one year might be thought to influence that the following year: e.g. if a team finishes 6th, if correlation is higher, they are more likely to finish in a position close to 6th the following year; a lower figure would suggest a greater range of probable outcomes. This measure therefore allows the overall stability of league finishing positions to be tracked. If correlation is increasing, this would be consistent with the claim that money was undermining the unpredictability of league competitions.
Levels of turnover at the top of the league. While the level of flux across an entire league is important, it means little if there is massive fluidity of teams lower in the league, yet the top few places are dominated by a small elite. So we might also look to assess the probability that a team who finish in the top few places dropping away to a lower finishing position at any time over the subsequent three seasons. This can give us an indication of the extent to which positions towards the top of a league become easier for elite clubs to defend. A lower figure here indicates a lower probability of top teams subsequently failing to finish in higher league positions. In discussion of specific leagues I’ll mention calculations based on the percentage of the top two, top four, top six and top ten places, depending upon the size of the league and an assessment of the shape of competition at the top.

Yet, as well as assessing the levels of fluidity in leagues as a whole, another key consideration is the scale of dominance exercised by the clubs at the very top. Is the gap between league champions and the also-rans wider than it was in the past? Two further simple measurements can capture this:

Points-per-game of champions. If there is a greater gap between the elite and the rest, you would expect elite clubs would record more victories and fewer defeats than champions of the past, which would result in higher points totals. (To allow for comparison, points have been calculated based on league records on the basis of three points per win, applying this even to leagues that were played under a two points per win rule. In a future post, I will consider whether the shift to three points for a win led to a different mindset and approach, which could therefore impact upon the figures used here.)
Average goal difference per game. Similar to the above, increases in the gap at the top would be likely to register in greater margins of victory and avoidance of anything but narrow defeats.

An increase in either of these measures over time would be evidence of a greater gap between teams at the top and the remainder of their respective leagues.

Armed with these four measurements, the next few posts will track trends in competition across some of Europe’s top leagues.