Friday 5 November 2010

Birth Lottery

Your date of birth is random. Of course, your parents’ wishes and desires had something to do with it but by and large the date of conception is determined by chance (luck or accident, whichever you wish to call it). The length of pregnancy is, similarly, difficult to predict with a great deal of certainty: some are born before the due date, others after.

Does it matter then when you happened to be born? As it turns out, it does matter quite a lot. Among the first to understand this were sport scientists: they noticed that a lot of top footballers and ice-hockey players were born during a particular period of the year. For example, an article by Baker and Logan (Developmental contexts and sporting success: Birth date and birthplace effects in national hockey league draftees 2000–2005, British Journal of Sports Medicine, 2007) found that, during the first half of the 2000s, 64% of NHL players were born in the first half of the year (National Hockey League covers both the US and Canada). Sherar et al. (Do physical maturity and birth date predict talent in male youth ice hockey players?, Journal of Sports Sciences, 2010) found an even more skewed pattern when looking at teenage ice hockey players from Saskatchewan. Among their initial sample, 55% were born in the first half of the year. After two rounds of rigorous selection, 78% of the final and much reduced sample were born in the first half of the year.

To make things even more complicated, being born in the first half of the year is not always the key to success in sports. Another article, by Delorme, Boiché and Raspaud (Relative age effect in elite sports: Methodological bias or real discrimination?, European Journal of Sport Science, 2010) found that significantly more (grown-up) French footballers were born during the months of August to October than in any other three-month interval.

If this all seems puzzling, the explanation is in fact rather simple and centers on the notion of ‘relative age’. Football and ice hockey clubs recruit new players from among children born within a particular age cohort and the span of that cohort is typically one year. This means that some children within a given cohort are relatively old while others are relatively young; the difference between the oldest and the youngest may be as much as one year less one day. North American ice hockey clubs recruit children born in a given calendar year: the Saskatchewan study covered youths born in 1989. French football clubs, in contrast, recruit players born between August of one year and July of the following year (or, more precisely, used to do so until 1982 when they switched to calendar years). The clubs in principle seek to select players based on talent. However, in sports where physical strength is important, older children have a physical advantage. Correspondingly, they may be perceived as being more talented. As a result, the relatively older players do well and succeed while those who are relatively younger do poorly – and are more likely to give up too.

Now how about economics? It turns out that when you were born also has economic implications – and I don’t mean that if you were born in December, you are less likely to earn the kind of salary that David Beckham gets (who, incidentally, was born in May). School enrolment is also organized by cohorts. In most European countries, a particular year includes children born from autumn of one year until the end of summer of the following year. Here, physical strength hopefully does not play much of a role but the relative maturity may be important: at the age of 5 or 6, a difference of almost one year is potentially large. The relatively older children, correspondingly, perform better in school. This effect is surprisingly long lasting: a Dutch study by Eric Plug (Season of birth, schooling and earnings, Journal of Economic Psychology, 2001) found that boys born in autumn (October to December, to be precise) are 12% more likely to attend university than the rest; autumn girls are 16% more likely to do so. A Japanese study by Daiji Kawaguchi (Actual Age at School Entry, Educational Outcomes, and Earnings, to be published in the Journal of Japanese and International Economies) similarly reports that those born in April to June (in Japan, the critical date for school entry is 1st April) have better test results and acquire more education than those born between January and March. Moreover, while Plug only found some tentative indications that relative age affects earnings, Kawaguchi’s results are more conclusive: Japanese men born from April to June earn 4% more than those born from January to March (the difference in earnings is not significant for women). 4% may not seem like a lot, but over one’s lifetime it would make a rather large sum.

Could these differences in education and earnings based on one’s month of birth be reduced by policy? The answer, in my opinion, is yes. Relative age effects would be smaller if school enrolment is staggered: it could take place in cohorts corresponding to three or six months, for example. The relative age differences within any given cohort would then be halved or quartered. Many schools are large enough to accommodate multiple cohorts per year. Universities could easily admit students in each term rather than only in Autumn. An added advantage would be that the job market for graduates would not be flooded by new entrants during the summer but the inflow would become staggered too.

A policy change that would deliver more egalitarian distribution of educational attainments and earnings at little or no extra cost might well be the proverbial 100$ bill lying on the sidewalk that is just waiting to be picked up.