By Robbie Butler

At the start of the year I had an entry about the ever-expanding World Cup Finals and attempted to put some logic behind FIFA's plans to increase the number of teams at the Finals in 2026.

​Prior to this weekend's world cup qualifying fixtures, I read a Sky Sports article focusing on the potential of a World Cup without Argentina. Since the weekend, the chances of this have reduced thanks to a Lionel Messi penalty in a 1-0 win over Chile. The current qualifications standings are below.

Between 4 and 5 teams can qualify from this confederation. In percentage terms, this is far better than any other confederation. For example, if this were the end table, Ecuador could still qualify despite losing five games. No UEFA country could expect to qualify with this loss rate. Even controlling for the number of games played, five defeats in CONMEBOL qualifying would compare to almost 3 defeats in UEFA qualifying. I doubt any European country that loses three games will make a play-off, let alone qualify for the tournament in Russia.

The South America countries have things much more straight-forward, and not just over the European. The table below shows the percentage of countries that have qualified from each confederation since the World Cup 

​expanded to 32 teams in 1998. European countries have the second highest number of teams qualifying as a percentage of the number of entrants. 

The decision to increase the number of teams for further tournaments has been met with much criticism, and raised issues of the quality of teams that could now reach the finals. If 16 additional countries are to be given places, surely Africa, Asia, Oceania and North and Central American countries should be first in the queue. The quality may not improve but the distribution of places will be much closer to a model consistent with a "World Cup". 

By Robbie Butler

FIFA recently announced that the World Cup Finals will be expanding from 32 teams to 48 for the 2026 World Cup. This decision has been largely criticised. Fears of diluting the quality and player burn-out are commonly cited as reasons to oppose the move.

From the organiser’s perspective, the World Cup is big business. Broadcasting revenue alone from the most recent tournament in Brazil was worth $2.4 billion. In total, FIFA made a profit of approximately $2.6 billion from the 2014 tournament. Expanding the tournament seems like a no-brainer.
​
There is another reason to expand the competition, which has been rather overlooked. There are many more national associations now than there were even twenty years ago.

​A historical analysis of previous World Cup qualifications can help understand the evolution toward greater numbers of national associations. Starting in 1934 (the inaugural tournament in 1930 was invitation only), 1 out of 2 countries entering the qualification stage reached the Finals in Italy. This ratio has never been topped. The graph below plots the number of entrants, number of team qualifying (left-hand axis) and the percentage of countries qualifying (right-hand axis) from 1934 to 2026 (assuming the number of associations remains at 210 for the 2022 and 2026 tournaments). 

The expansion to 48 countries means just over one-fifth of all entrants will reach the finals. The 1982 expansion to 24 teams, and the 1998 expansion to 32, were outpaced by the rapid increase in associations, particularly in UEFA following the collapse of the Soviet Union and break-up of Yugoslavia.
​
Consider the UEFA confederation alone is worthwhile.

From a European perspective, an increase in the number of UEFA places must be a priority. The second graph below shows just how hard it has become to qualify for the World Cup as a UEFA member, with just 13 places for 54 countries trying to get to Russia 2018. This compares very poorly to the 13 spots available for Italia ’90 when just 33 countries entered the qualification phase. 

Almost 40% of UEFA countries entering the qualification stage reached the finals in 1990. With an identical number of places, just 24% of associations will be at Russia next year. ​Commentators can criticise the move, but let us not ignore the explosion in the number of FIFA and UEFA members. 

by Declan Jordan

​One of the explanations for the persistence in relative age effects in sport is the physiological benefits that players born closer to the age cut-off for under-age competition (usually earlier in a calendar year) have over players born later in the period. At a young age this can mean up to almost a year in the development of players in the same age category. This physiological advantage may be compounded by the advantages of getting more game time and subsequently more coaching. This could mean some young players dropping out of the game, leading to a persistent age effect.

In football this physiological difference may be more beneficial in different positions. For example, goalkeepers tend to be taller than outfield players, and defenders also tend to be taller and bigger than midfielders and attackers (we can of course think of many midfielders and attackers that are tall and big).

The graphs below show the breakdown of quarter of birth for elite under-17 footballers - i.e. those representing their countries at FIFA U-17 World Cup Finals between 2007 and 2014. There are four tournaments each for male and female players over the period. The data comes from FIFA and player position is designated on the website for each tournament.

The figures show, from the flatter distribution of the female players, that relative age effects are far more pronounced for males than for females.   

However, within each cohort there is no difference in the relative age effect across different positions. This means that the proportion of players in each cohort born in each quarter of the calendar year is similar to proportions in each field position. This suggests that physiological differences at early stages of development may affect the likelihood of selection (and subsequent coaching attention) in every position. Unlike rugby, it seems there isn't a position for all body shapes in football. Perhaps an interesting question would be to look at relative age effects by position in a sport like rugby, where physical size matters for different positions.

by Declan Jordan

The news that broke recently that all but two of Nigeria's Under-17 first team failed age tests for the forthcoming African Cup of Nations was shocking, but not surprising. Nigeria have been remarkably successful in the last decade at U-17 level. Nigeria won the U-17 World Cup in the last two tournaments in 2013 and 2015. They were runners-up in 2009 (when they hosted the tournament) and won as well in 2007. They have won the competition 5 times in all.

However, there have been question marks about Nigerian players' true ages for many years now. Indeed there are strong suspicions that African football,and to a lesser extent other regions, have flouted the age rules for many years. In the 2013 tournament, 3 players from each of Morocco, Congo, and Nigeria were sent home from the U17 World Cup for failing the age test. In 2009 MRI bone scans were introduced to test players were the age they claimed to be.

It may not be surprising that countries would want to use older players, since there is a physical advantage for older players in underage tournaments. Indeed, there is a substantial literature on the 'relative age effect', a topic on which my colleagues on this blog have published research. FIFA rules dictate that a player is eligible for a competition if he or she was born in the designated calendar year, so for the U17 tournament in 2015 players must have been born on or after 1 January 1998 (17 years previously). This means players born earlier in the year (in January for example) can have up to 12 months age advantage (and associated physical advantage) - details on player eligibility is available on page 20 here. There is now substantial evidence that football squads are over-represented by players born in the early months of a calendar year relative to the general population.

What might be expected is that countries who are looking to enter players who are older than they actually are is that they would state an age earlier in the calendar year. For example, in this article it is reported that the Syrian squad for the U19 World Cup in 2012 included six players born January 1, 1993 and that the entire team had January birth dates. That approach seems to be a little bit obvious of course, so perhaps it would be better not to draw attention to your squad by spreading the month of birth out.

The graphs below show the profile of month of birth (by quarter) for four editions of Men's and Women's U-17 World Cup Finals from 2007 to 2014. There are 24 countries in the Men's World Cup and 16 in the Women's. It is noticeable that the age dispersion is greater for women, more players are born in the later quarters of the year (the difference is statistically significant). The graphs show quite a remarkable difference between the dispersion for Nigerian squads and other squads (excluding Nigeria). Nigeria appeared in all of the four Women's Finals and three of the four Men's Finals. The graphs show that Nigeria had a statistically different month of birth distribution than the average of the other countries.

Extending the analysis to look at regional differences in relative age effects, the tables below show the classification of players by the half of the year in which they were born (January to June or July to December) for each global region.

Africa stands out as the only continent that shows a greater proportion of players born in the second half of the calendar year (the only exception is Oceania for the Women's Finals). The proportion born in the second half of the year is higher for African countries in the Women's Finals. Again, the difference between regions for men and women is statistically significant. 

It is difficult to know the reasons for Africa's exceptional result. If age fraud is a substantial problem there then perhaps the ages being registered are designed to deflect attention by having much less January birthdates. It may be that the birthdate registered is less important where it is different than an actual birthdate. In the absence of age fraud, perhaps there is less physical difference between African athletes of less than a year's age difference, or perhaps African football has avoided the traps of other countries that tend to select more older players and thereby lose younger players to the game. These latter explanations though would likely result in a flat distribution (similar to the dispersion of months of birth in the general population) rather than having such a large proportion in the second half and final quarter of the calendar year.

by Declan Jordan

​In a recent post my colleague Robbie Butler raised the thorny issue of FIFA country rankings. He bemoaned the problem of countries achieving a higher world rank than should perhaps be warranted because some countries will play more competitive matches against tougher opposition because of their continental affiliation than others. He’s right of course. But the problem of games against tougher opposition is perhaps one that some countries might welcome.
 
The FIFA ranking (or to give them their full title the FIFA/Coca Cola World Ranking) is largely a marketing exercise and is of little consequence. Teams from different continents rarely play each other. In the World Cup finals, teams from various continents are generally kept apart in the group stage draw. This means the overall ranking has little impact, with a country’s position relative to other countries in the same confederation being the critical aspect. For this reason, comparing positions across confederations is of limited value.
 
It does however prompt comparison of how countries accumulate ranking points and inevitably calls into question the conduct of competitive qualification tournaments in each confederation. I have written before about my views on the unfairness of the African qualification process relative to Europe and South America. The African and European contexts are similar to the extent that both confederations have about the same number of affiliated members (54 and 55 respectively). The root of the unfairness lies in huge disparity between the number of qualifiers each continent has – Europe 13 has spaces (not including hosts Russia) and Africa has 5.
 
The African qualifying campaign for Russia 2018 has already begun, and for 13 countries has already ended. In the middle of this month it will have ended for another 20 countries – this is 30 months before the finals will be held. For me this was brought home when looking at the fixtures for my football-mad son, who is Ethiopian, and seeing that they managed to overcome Sao Tome e Principe after being surprisingly beaten 1-0 in the first leg. Their reward is a game against Congo (the eighth ranked African team) in two weeks' time. It seems he has a much better chance supporting Ireland.

Could football fans and national associations in Europe countenance a situation where the lowest 26 ranked teams would have a play-off to continue with World Cup qualifying and risk being eliminated from the World Cup in October 2015 for a 2018 World Cup Finals tournament? The bottom 26 nations in Europe currently includes Turkey, Serbia, Hungary, Bulgaria, Finland, Norway, and more importantly the Republic of Ireland.
 
The first play-offs in Africa were held last month and 13 countries have been eliminated from the 2018 World Cup – having played two matches, one at home. The second round puts the 13 winners from the first round with the remaining 27 African nations who play in a home and away qualifier – the 20 successful teams will progress to play in five groups of four teams with the winner of each group going to the finals in Russia. This means the majority of African countries (33 out of 53 in the current campaign) have less than four World Cup qualifying matches and two at home. 
 
(This contrasts hugely with the South American qualifying campaign where teams play 18 games in a league format with the top four  In fact, Ecuador and Uruguay both won less than half of their matches (7 from 16) in qualifying for the 2014 World Cup – there were 16 games in qualifying last time because Brazil qualified as hosts).
 
The argument of course is that the European teams are better and so deserve more places at the finals (one hopes it is not based on market size or Asia and Africa will soon be entitled to a lot more spaces on that criteria). The FIFA rankings can be used to justify the superiority of the European (and South American) nations. However, this is a circular argument where nations can achieve more points by playing more competitive games, with some of them against higher ranking teams). Playing these (and more of them) fixtures makes teams better – or at least gives them an opportunity to improve more rapidly. Is it possible that Wales’ dramatic improvement from 112th in the world in 2010 to 8th in 2015 could have been achieved in a qualifying environment like Africa’s? This shows the benefit of having games against the 'best' to develop players at international level.

By Robbie Butler

I recently spoke about the challenges facing the Qatar 2022 World Cup, the first World Cup in history not to be played during the traditional, Northern Hemisphere summer months. Both FIFA and the organisers in Qatar have gone to great lengths to explain the need for the switch, and argue that football simply cannot be played in conditions of 45 -50 degrees Celsius during the months of June and July in the tiny Gulf state.

And therein lies another problem. Qatar is tiny by global standards. 

Depending on what you classify as a country, Qatar comes in as the 164th biggest state on the planet, one place behind the Falkland Islands! Since the World Cup went to a sixty-four game tournament in 1998 thirty-two teams have competed in the Finals. World Cup Finals have been staged in France (1998), Korea & Japan (2002), Germany (2006), South Africa (2010) and Brazil (2014). The illustrations below show how Qatar (purple) compares in terms of its geographic size to each of these former hosts (pink).

Another problem is surely the location of the twelve proposed hosting venues, spread across seven locations six of which are located on the eastern side of the Qatar Peninsula.  Doha (4) and Al Rayyan (3), which lie just 55km away from one another, will together provide seven of the stadia for the event. The locations of Al Daayen, Al Khor, Ash-Shamal, Al Wakrah and Umm Salal will each provide one stadium.  

Finally, just for the Irish perspective, here’s how Qatar compares to Ireland. Ireland (70,273 km²) is 6.07 times as big as Qatar (11,571 km²). In fact, counties Cork and Kerry combined cover some 12,300km², roughly 760km² more than Qatar.

Maybe this Rugby World Cup bid isn’t a bad idea after all…;-)

by Declan Jordan

The debate on immigration that has seemed to dominate British politics recently also has its echoes in football discussions. I have even in Ireland heard how there are too many foreign players in the English game. When I point out that Irish players are also foreign in the English game this prompts an exception. Even at the top of the English game the argument for limiting the number of foreign players seems to be getting traction - despite the lack of evidence and faulty logic. 

The emphasis in English commentary is on the number of players immigrating into the Premier League. This is probably because so few English players emigrate to ply their trade in foreign leagues. So while the argument goes that it may be bad for the English national side to have so many foreign players blocking chances for local players to progress, there is a corollary there that the English national team may benefit from English players going to play in other leagues.

I was pleased to come across a paper in the area in Applied Economics Letters co-authored by Grant Allan and John Moffatt from Strathclyde and Durham respectively. Grant is a colleague of mine in the Regional Science Association so we have even more to discuss at our annual conferences.

The implication of the study for English football is to embrace the 'other' aspect of openness of international labour markets by encouraging English footballers to play abroad, broadening their experience and developing skills and awareness. The paper considers two phenomena - muscle drain (the loss of players to foreign leagues) and brain gain (the hiring of foreign managers to manage the national side). The paper considers the effect of these two issues on national team performance.

The paper finds, using data on national team ranking, league standing and numbers of foreign-based players and imported managers, that the 'player export' variable is robustly positive and significant, suggesting that having players playing abroad has a beneficial effect on the national team ranking. The 'manager import' variable however is negative, suggesting that there may be cultural differences between players and managers that may be difficult to surmount.

It's interesting to consider these findings in the context of performances at this year's World Cup in Brazil. In their squad England had one foreign based player - Fraser Forster the third choice goalkeeper who plays with Celtic in Scotland. Of the semi-finalists, Brazil had 19 players out of 23 playing outside Brazil, Germany had 7 playing abroad (in England, Italy and Spain), Argentina had 20 playing abroad and the Netherlands had 13 playing abroad. Of the other teams in their group, the number of domestic players in each squad were 20 for Italy, 10 for Costa Rica, and 1 for Uruguay. The latter two qualified for the quarter final. This is an interesting site for data on this. 

While England had an English manager at a World Cup for the first time since 1998 it wasn't enough to overcome the 'player export' variable. Maybe it's worth noting that that the semi-finalists all had domestic managers too.

by Declan Jordan

There is a sizeable literature on the economic impacts of major sporting events. Robbie Butler has shown in previous posts such as here and here that the evidence suggests that hosting major sports events, such as the World Cup and the Olympics, rarely pay for themselves financially.

There is an interesting contribution to the literature on the economic effects of hosting major events in the latest issue of Kyklos. An article by two Turkish economists, Veysel Avsar and Umut Unal, looks at the effects on trade between host and participating countries at FIFA World Cups. The article can be accessed here, though there is open access earlier version on SSRN here.

The paper finds strong evidence for a trading effect of the World Cup using bilateral trade data for 196 countries between 1950 and 2006. Pointing to the showcase effect of hosting the World Cup and the opportunities it provides for developing networks and trade channels, the paper finds that the World Cup significantly increases exports from participant countries to host countries and that trade is higher for host participant pairs compared to other country pairs. The authors find that the effects are decreasing over time. A personal view is that this may reflect the diminishing importance of showcase effects as the cost of information and communication diminishes in an increasingly digital society.

The paper is an interesting contribution to the literature and points to economic effects that may be under-appreciated or overlooked in the studies of big event impact analysis.

by Declan Jordan

For me one of the stars of the World Cup this year is sports journalist Ken Early who has been filing really enjoyable and insightful articles for the Irish Times and contributing to the Second Captains online radio show. One of his articles on the dramatic humiliation of Brazil at the hands of Germany referred to the 'export' of Brazilian football players at very young ages. He argues that has led to Brazilian football losing its sense of identity.

Of course the causes of the decline and fall of Brazilian football are varied and multifaceted - some being proximate causes (David Luiz's positioning inability) and some long-term (the pile 'em high approach referred to in the same article by Early). The exodus of Brazilian talent prompted me to consider the experiences of other countries to see if there is something unique about the Brazilian experience. Perhaps the idea struck a chord because in Ireland we have a similar problem of 'talent drain' and so the only thing we have in common with Brazil is not getting a hammering at home from Germany in this World Cup competition.

The blog has seen posts recently on the 'granny rule' and the effects of previous generations' migration on international football teams' performances. Perhaps a related issue is the immigration of young professional footballers and the effect that might have on a national team's identity and performance.

I looked at the starting 11 of the four semi-finalists (Brazil, Germany, Argentina and the Netherlands). The experience of the players in terms of their migration patterns are quite striking. Each country has a strong domestic league.

In Brazil's case, as Ken Early pointed out, most of the players left Brazil at a young age after relatively few appearances in the Brazilian league. The most extreme case is Hulk who played only once for Vitoria before leaving at 19, initially to Japan. Of the 11 starters only Fred plays in Brazil currently, though he also left Brazil when he was 22 after 69 league appearances. On average players left Brazil at 21 years old and having made 51.5 appearances in domestic football.

The German experience is remarkably different. Of their 11 starters, 8 have never played club football outside Germany (and Miroslav Klose is one of the three having joined Lazio at 33 years of age in 2011). The other 2, Khedira and Ozil, left Germany at 22 and 23 respectively but both had made over 100 Bundesliga appearances at that stage. 

The comparison between the South American and European participants in the other semi-final is similar. All of the Argentinian players play outside their native country. The average age when they left is 20 and the average number of appearances before they emigrated is 52.5 (very similar to the Brazilian statistics). Even excluding Lionel Messi (who moved to Barcelona at 13) the average age only rises to 20.5 and the appearances to 57. Enzo Perez is a bit of an exception among the team. The Benfica player played 103 matches in Argentina before his move to Portugal.

Six of the Dutch players play abroad, with the other 5 never having played outside the Netherlands. However, the average age of the 6 when they left the Netherlands is 23 and the average number of appearances before leaving is 132. For example, Dirk Kuyt made 261 appearances for Utrecht and Feyenoord before joining Liverpool in 2006 and Ron Vlaar made 142 appearances for AZ 67 and Feyenoord before joining Aston Villa. The table shows the comparison between each country.

It's clear that while Brazil suffer from an exodus of players from their domestic leagues (mostly to European leagues) and this may damage the identity of the team with its fans, it is not a problem unique to them. This is probably a function of the money in European football relative to South America (and other continents). 

There are important implications in this for Brazilian attempts to address its football decline. After the abysmal performance of Germany in Euro 2000 the football association and clubs revamped football structures and player development and the fruits of that are seen now (at both national and club levels). The experience of player exodus poses the difficulty for Brazil that while it can enhance player development structures this is likely to simply make those players even more attractive to European clubs. Brazil needs to address player development (avoiding the 'law of numbers' approach referred to by Ken Early) but also address the talent drain.

By David Butler

Over June and July RTE, Ireland's national television and radio broadcaster, will beam images from World Cup 2014 in Brazil to T.V sets throughout the Irish state. For many sports enthusiasts this will bring great satisfaction. For the non-football fan however, what appears to be wall to wall football coverage, can test one's mental endurance.

At the start of the Premier League last  year I calculated how much football could be watched early in the season. Below is the math and some RTE entertainment comparisons to put into context the extent of World Cup coverage on RTE relative to a years worth of other T.V entertainment.

A dye in the wool football fan can view 5040 minutes of football action on RTE over the month (5 group games across the eight groups, as the last two kick off together, and 16 post group ties). This figure is a minimum and doesn't take into account opening/closing ceremonies, added time, extra time, highlight shows, the musings of any pundit or the superb Après Match.

To put this number (5040) in perspective, the 2012 season of the Australian soap Home and Away aired by RTE lasted 4950 minutes, that's 225 episodes that ran for 22 minutes each. In terms of just the 90 minutes of football, if one was to watch the entire World Cup on RTE it would be similar to watching just over an entire season of Home and Away in one month.

Season 51 of the Late Late Show, Ireland's Premier National talk show, lasted 4440 minutes, that's 37 episodes lasting 120 minutes each. Again, spending this time with Ryan Tubridy is not far off the extent of World Cup footage on RTE. 

In terms of other popular drama and entertainment shows on RTE, watching the entire World Cup would equate to viewing about 229 episodes of The Big Bang Theory, about 8 seasons of Celebratory Masterchef and would be close to seeing Dermot and his team renovate nearly 100 houses in Room to Improve!

Of course, the World Cup only happens every 4 years, so perhaps we should only think of a quarter of this viewing time or alternatively we could appreciate that soap fans just smooth their consumption and shouldn't be too frustrated. For me however these stats (especially the Home and Away one!) put into perspective what those who have no preference for football have to endure for June and July of 2014.