University degrees still offer value in long run
Some assumptions held by many are simply not true, writes Ross Finnie.
Once upon a time, the bookstore was a place a writer would come to give a talk, meet readers, sign books. A bookstore was a forum, a stage, a theatre. No longer. —
Andrew Cohen
If you are a young person thinking about going on to post-secondary education (PSE), you probably have lots of questions. What should I study? How much money can I expect to earn right after graduation? How will my earnings grow after that?
If that is the case, there is good chance that you will be interested in what follows here — in fact, your parents may be just as interested, if not more. (It’s probably OK if they look.)
Last week, we at the Education Policy Research Initiative (EPRI), released three research briefs that answer these questions — at least in part — in a totally new and unique way. We did this by partnering with the University of Ottawa and Statistics Canada to link data on students to their tax records in order to follow the earnings of bachelors level graduates on a year-by-year basis following graduation.
We did this separately for all cohorts of graduates from 1998 through 2010, following everyone through to 2011, breaking the results down by area of study. In a context where the value of a university degree is often questioned, we find perhaps surprisingly strong outcomes.
Social sciences graduates tended to start with average earnings of around $40,000 immediately after graduation, but these grew substantially, almost doubling to just under $80,000 13 years later. These seem like reasonable outcomes — and seem to point to careers being successfully established.
Their counterparts with humanities degrees had similar starting earnings levels, and also experienced steady earnings growth, although they finished a little lower, just under the $70,000 mark.
These findings demonstrate the importance of having accurate data on student outcomes over a longer-run perspective than the usual six-month to two-year view which are the current norm — at least as most provincial “key performance indicators” attempt to tell the story. Quite simply, the short-run outcomes miss an important part of the story — and probably the most critical aspect: longer-run outcomes. (Those indicators also have other problems, including sample sizes and accuracy of the income information they rely on.)
This growth in earnings in the years following graduation should probably not be surprising, since these areas of study (social sciences and humanities) do not tend to be associated with particular skill sets which lead directly to specific kinds of jobs, but instead reflect the development of more general skills which open a range of employment opportunities and diverse career paths which often take some time to develop. This is normal.
Conversely, health graduates (including nurses, physio/occupational therapists, and so on) started with higher average earnings but experienced slower earn- ings growth over time. This pattern is consistent with the development of tighter sets of employment skills that lead directly into specific kinds of jobs — often in more structured (and unionized) work situations.
Thus, we have three ( broad) areas of study, and three distinct stories. As interesting, these patterns held relatively consistently across all cohorts of graduates, for those who left school from the late 1990s through the post-2008 crisis.
But the importance of being able to separate graduates’ earnings paths by year of graduation are demonstrated by our finding that math, computer science, engineering, and business graduates had generally higher earnings than others, but faced much more volatile outcomes over time. The boom and bust of ICT (information and communications technology) graduates’ earnings is particularly notable in this regard, with starting earnings being as high as $75,000 when the sector was at its strongest, only to crash to approximately $42,000 by 2004, after the bubble had burst.
Interestingly, the greatest number of ICT graduates joined the labour force precisely when first-year earnings were at their lowest, reflecting those who had started university at the peak of the boom only to graduate at the nadir of ICT fortunes. The numbers of ICT graduates then dropped sharply, while earnings have recovered only moderately.
We also split each set of graduates into quintiles — that is, the top 20 per cent of earners, the next 20 per cent, and so on. The spread between the highest and lowest earners was greatest for business and engineering and
Math, computer science, engineering, and business graduates had generally higher earnings than others, but faced much more volatile outcomes over time.
computer science graduates, thus reflecting another kind of variability of earnings that young people should be aware of when making schooling choices. That is, if you do well in these areas, you may do very well, but if you are more in the middle of the pack, your earnings will be much more … well, middling. In fact, we find that those in the third quintiles — the true middle earners — did about the same among business, social science, and humanities graduates. Who would have thought that would be the case? We didn’t, until the numbers showed us.
Also of interest is that among every group of graduates, the bottom quintile did not do very well at all. The question is: who are these people, and why were their earnings so low? One way to get at this would be to see how much those low earnings are related to low schooling performance (grades) both while the student was going through university and even earlier, at the end of high school (that information is available). Were those poor outcomes the result of bad things happening to good students, or related to struggling students continuing to struggle after they hit the labour market?
The gender earnings gap is also worth mentioning since it grows over time among graduates of all areas of study. Engineering graduates are of particular note in this regard, as they start with a gender earnings gap of around $15,000, as compared to the negligible gap in some other areas, which then widens to $30,000 after 13 years. These gaps may be related to different career and family choices men and women make over time, or to gender-related labour market discrimination that evolves in the years following graduation. Our current analysis cannot separate these sources, but future work may help answer these questions.
As interesting as these results are on their own, their greatest value may be to demonstrate the feasibility of this approach for tracking — and analyzing — graduates’ early labour market outcomes. We are now broadening the analysis to a set of 12 postsecondary institutions across the country. We also plan to deepen it by probing into how labour market outcomes are related to the broad range of student characteristics and schooling experiences that are captured in these data.