What football teaches economists
Whereas some scientists are trying to understand how football crowds function, others are taking data off the pitch to see how the rest of us interact. Chris Smith spoke to Edoardo Gallo, an economist from Cambridge University with an academic interest in football.
Chris - So, what can football teach us about economics then?
Edoardo - Well, economists are interested in how individuals behave and how they respond to the incentives they're facing in everyday life. Football players may seem very different from us, but in reality, they are people and they're actually professional, performing their everyday job, who respond to the incentives in this job. So, this is exactly a type of setting that us, as economists are interested in. On top of this, there are at least three reasons why football may be a particular interesting settings for economists were to study human behaviour. The first reason is that incentives are extremely clear in football - you want to win. There are rules that you have to follow or you can break to try to win. The second one is that the stakes are extremely high. So, the incentives are very strong. And the third one is that we actually have a large amount of data now on what players are actually doing on a football pitch. We can use this data to study human behaviour.
Chris - When you say, "We've got a lot of data from players on the football pitch...", how is that data gathered? Do you sit in your office at the university with a massive great plasma, tune this in and you sit there with your pen?
Edoardo - Well, I watch quite a lot of football, but not to collect data.
Chris - At work?
Edoardo - I'd love to have a big plasma in my office to watch football games. No, jokes apart, actually, there are some companies that collect an extensive amount of data. For example, the data that we have is from one of these companies. It has time stamped events by the second of every event that happens on a football pitch including the X,Y spatial coordinates.
Chris - In other words, you're tracking people around the football field so you can see where the players are going and you know where they were on the field at any one moment in time for every game.
Edoardo - Exactly, for every game. So for every game, we have more than a thousand data points and we can know exactly what is actually happening in those games and this is for example for the English Primer League for two seasons.
Chris - So, what can you learn from that resolution of data from knowing how the players interacted like that?
Edoardo - One of the projects that we did and was recently published was to look at the issue of discrimination. So, discrimination is something that social scientists and economists have been interested in for a long time, it's definitely something very important - both for its moral implication but also its economic implications. What we do is to look at this data and to show that actually, referees in the English Primer League, in the two seasons that we looked at, discriminate against a particularly group of players by awarding more yellow cards against them. And given that the data is so detailed, we can also see for example that these players does not behave differently in terms for example aggressive behaviour.
Chris - So, when you say they're giving more yellow cards and they're discriminating against certain players, who were they discriminating against?
Edoardo - Players that are non-white and they are foreign and they belong to the minority groups in the UK. So, it's not on simple racial lines. It's not a simple black or white.
Chris - What about the refs? Are the refs all white or are they a mixture of black and white refs? Is there any bias in that direction?
Edoardo - It would've been great to see if there was a significant group of referees that were black and a similar group that was white, whether there was any difference between the two. Unfortunately, the data that we look at, there was only one black referee that refereed very few games, so we cannot look at that.
Chris - So, are you saying that basically, referees on the basis of your data are racists or are you saying that there's something else going on which is making them discriminate against certain groups of players?
Edoardo - So, this is a great question and as economists are interested in the type of discrimination that is actually happening because for example, this is very important for policy reasons, now, there are different types of discrimination, one type is for example what economists call a preference based. So, you just don't like someone from a certain social group. So you therefore discriminate against them. Another discrimination for example is called statistical discrimination which means that there's a group of people. They behave in a certain way. There's a certain tendency to behave more in that way than another group and therefore, you discriminate against them. A typical example could be that if you believe that Spanish players dive more, then the referees may award them more yellow cards for diving even though maybe they're not actually diving more.
Chris - Is there any evidence for that in what you've looked at?
Edoardo - No, this was just an anecdote from watching a lot of football. The third type for example is what psychologists call implicit discrimination which is something unconscious that you do because you're associating a negative attribute to a certain social group and you do it without realising it. What we show in the data is that actually, there's evidence that discrimination that is happening is actually implicit discrimination because it's only happening when the referee is time pressured to make a decision to award the yellow card.
Edoardo - That's right. You use these mental processes that are associated with implicit discrimination, which are intuitive mental processes only when you're time pressured. We show that referees are time pressured by players to award the yellow card becuse if a foul happens, they only have a certain amount of time before they can award the yellow card. This varies across the football pitch. There is more time in the defence seen or the attacking part than there is in the midfield. We actually showed discrimination only happens in a midfield area. If the discrimination was for example preference based, there should be no difference across the different parts of the pitch. Well, if it's implicit, we'd expect that there's more discrimination in the middle third which is exactly what we found.
Chris - What can we do about that?
Edoardo - A general message from that research is that it's very difficult to change them. Of course, awareness may help a bit or we also know that awareness does not help in a lot of those domains. So of course, awareness is one potential solution but it doesn't get us all the way.