It’s fairly well known that people with higher levels of education tend to be better off than those who left school with low-level or no qualifications: they’re more likely to be in work and if they have a job, are paid more. Yes, there’s the odd Lord Sugar-esq example that runs to the contrary, showing that having a raft of qualifications under your belt is not strictly necessary for success. But, the general rule still holds: the better you do in school, the better your chances in the jobs market are when you leave school.
Is that it? Are there no other benefits associated with education, other than jobs and pay? Well actually, there might be. Some economists have looked at what they term ‘the wider benefits of learning’. These can encompass anything other than the standard wage-employment outcomes that economists try to measure as the ‘benefits’ of education. Lots of things have been covered, from teenage pregnancy, to crime, to voting behaviour. The link between education and health outcomes is also one of the well-studied areas in the wider benefits literature. So are people with more education healthier, too? Let’s take a look.
In theory, education could affect health outcomes either directly or indirectly
Education might have a direct effect on health and healthy behaviours, because people may learn about the consequences of unhealthy behaviours at school, college or university – perhaps through their A Level biology classes, for example. Also, more educated people tend to be better able to grasp the basic consequences of unhealthy living, even if they don’t understand the complex biology behind it.
That’s the direct effect, but education may also have a number of indirect effects on health, too. If the better educated end up in secure, higher paying jobs, they’re more likely to be able to afford the things that help them to live a healthier lifestyle, like gym memberships for example. Reduced income volatility may also lead to less stress and more educated people usually work in safer environments. In addition, research has shown that they also put a higher value on their future life outcomes (they discount the future less heavily, in the jargon), making them less-likely to participate in unhealthy behaviour, like smoking. If more educated people tend to work and socialise with others who have a similar education level, then peer effects might play a part – if your neighbour or workmates lead a healthy lifestyle, you’re more likely to follow suit.
What does the evidence say about the extent of this link?
There is little evidence that looks at the relative importance of all these mechanisms put together, but a fair amount of UK evidence points, at the very least, to a correlation between education and health outcomes. Over a number of years, research from the Centre for Research on the Wider Benefits of Learning[i] has found education to be beneficial across a range of health areas:
- Depression: If 10 per cent of women in the UK who hold no qualifications were to obtain them at Level 1, the resulting reduction in the incidence of depression could lead to savings of up to £34 million per year. Taking women without qualifications to Level 2 could reduce their risk of depression at age 42 by 15 per cent. This could save the country up to £200 million per year (in 2002 money)[ii].
- Obesity: Analysis on data from the 1958 National Child Development Study shows that men and women with no qualifications were slightly more likely to be obese than those who held qualifications at Level 1[iii].
- Smoking: Individuals educated to Level 2 or below are 75 per cent more likely to be smoking age 30 compared to a similar individual educated to degree level or higher[iv].
- Take up of preventative care: Women with Level 2 or above qualifications have a higher probability of having 3 or more cervical screenings in 11 years than women with qualifications lower than Level 2[v].
However, the authors of these studies caution that these measured effects are not necessarily causal, as it’s very difficult to take into account all the things that might influence someone’s health outcomes. Therefore, the results from this type of work should be seen as an upper bound of the potential effect of education on health outcomes.
Has anyone attempted to identify a true causal effect?
Some researchers have, but the results are mixed. In a UK setting, two separate studies make use of the Raising of the School Leaving Age (RoSLA) in Britain in both 1947 and 1973 to try and identify if a causal relationship between education and health outcomes exists. They are able to do this because the changes in the school leaving age generated sharp differences in educational attainment among people born just months apart. Take for example the 1973 RoSLA. Those pupils whose 15th birthday fell before the 1st September 1972 were allowed leave school at age 15. Those whose birthday fell on or after the 1st September had to stay on to complete an additional year of education. In effect, this is akin to a natural experiment in science: there’s no reason to expect the health outcomes of the two cohorts of children to be different overall, not when they are born just months apart. So all else being equal, any improvements in health outcomes are likely to have a causal relationship with the extra education attained in that year. Both studies confirm through their analysis that the cohorts just affected by these changes completed significantly more education than previous cohorts subject to the old laws.
However, the two studies generate conflicting results. One, by Mary Silles in 2009,[vi] finds evidence of a causal relationship between education and health, with an extra year of education increasing the probability of ‘good’ health by between 4 and 6 per cent. It also finds that the probability of having no long-standing illness increases by between 5 and 7 per cent and that the probability of having no work-preventing illness increases by around 1 per cent.
The second study, by Damon Clark and Heather Royer,[vii] attempts to estimate the causal effect of education on a range of health outcomes, including mortality rates and a number of self-reported health problems and behaviours, such as smoking and alcohol intake. This study finds little or no evidence that the additional year of education improved health outcomes or changed health behaviours.
At first glance, these results seem rather intriguing. After all, they both use the same data and similar methods, but would seem to generate completely different results. So, why might this be? Is one ‘wrong’ and the other ‘right’? Or could they both be right, but in their own way?
Although the two studies use the same data and similar methods, they do look at different outcomes, which could have some part to play in explaining the conflicting results. However, there is also one key methodological difference: the second study defines its cohorts by month of birth and the first by year of birth. Since the compulsory schooling changes were introduced mid-year (April 1st 1947 and September 1st 1972), the month of birth approach ensures that people are more accurately assigned to the ‘before the change’ and ‘after the change’ groups. All other things equal, the month of birth approach should yield more accurate results.
Is Clark and Royer’s superior methodology enough for us to conclude that that there isn’t much in the way of a causal link between education and health outcomes? Probably not: there really isn’t much in it. If the authors of the first study repeated their analysis following Clark and Royer’s methodology, their results might still hold. And it may be that education can affect the outcomes studied in the second report, but they’re driven by participation in post-compulsory schooling instead.
Peer affects may also play a part. In other words, because whole cohorts of children were affected by the change at the same time, these children would have left school with the same types of peers as the unaffected students. Therefore, they may have been just as likely to engage in unhealthy activities as older cohorts.
It’s a shame that the Silles research uses year of birth and not month of birth, as it would have been interesting to be able to compare both sets of results more equally. As is nearly always the case in economics, more research is needed; and it would be worthwhile trying to replicate Silles’ results, using month of birth as a way of defining cohorts. Nevertheless though, it does enough to encourage me to believe that education is likely to have a causal relationship with at least some health outcomes.
But what do you think?
[i] Feinstein, L. Budge, D. Vorhaus, J and Duckworth, K. (2008): The social and personal benefits and of learning: A summary of key research findings. Centre for Research on the Wider Benefits of Learning
[ii] Feinstein, L. (2002): Quantitative estimates of the social benefits of learning, 2: Health (Depression and Obesity), Centre for Research on the Wider Benefits of Learning Report no. 6.
[iii] Feinstein (2002) op. cit.
[iv] Bynner, J., Dolton, P., Feinstein, L., Makepeace, G., Malmberg, L., Woods, L. (2003): Revisiting the Benefits of Higher Education. London: The Smith Institute.
[v] Sabates, R. and Feinstein L (2004): Education, training and the take-up of preventative healthcare, Centre for Research on the Wider Benefits of Learning Report no. 12.
[vi] Silles, M. (2009): The Causal Effect of Education on Health: Evidence from the United Kingdom, Economics of Education Review
[vii] Clark, D and Royer, H. The Effect of Education on Adult Health and Mortality: Evidence from Britain (forthcoming, American Economic Review),