Accurate estimates of the returns to different higher education courses are vital. Information on the performance of the various options that students face is essential to enable them to make an informed decision about which subjects and at which university to study. High-quality evidence on returns is also crucial for policy makers considering the design of the higher education system and skills development in the economy. Recent work by Belfield et al. (2018) provided extensive evidence on the returns to different tertiary education options using newly linked administrative data. A common criticism of this work is that it examines the returns to gross earnings and does not take into account cost of living differences across the country. In this report, we investigate how cost-of-living adjustment affects the returns to different higher education options.
It is well known that housing costs vary widely across the country. For example, last year the average house price in Camden hit over £ 1.1million, compared to just £ 140,000 in County Durham. The prices of other goods and services show similar (albeit less extreme) trends and as a result there is huge variation in the cost of living across the UK. Figure 1 highlights this variation, showing how the overall cost of living may be more than 10% lower than the UK average in the north of the country, but more than 30% higher than the national average in London and the South-East. A salary that could allow individuals to lead a very comfortable life in County Durham could therefore leave them struggling to get by in London.
This can have important implications for how we think about going back to college. We know that graduates are much more geographically mobile than non-graduates and more likely to move to large cities (see Britton et al., 2021). If this means that graduates face a higher cost of living than non-graduates, the differences in gross income between these groups will overestimate the differences in their disposable income and standard of living. Moreover, not only are there differences in place of residence between graduates and non-graduates, but there are also very large differences in place of residence between graduates who attended different universities. Figure 2 compares residency at age 27 for graduates from the University of Bolton and the London School of Economics (LSE). While virtually all University of Bolton graduates remain in the North West, those of LSE remain concentrated in London and the South East. Thus, while it is true that income from LSE attendance is considerably higher than income from University of Bolton attendance, the differences in the cost of living in the areas where graduates of these respective institutions live. after leaving college suggest that the average University of Bolton student will need a lower income level than the average LSE graduate to achieve the same standard of living. A similar argument can be applied to different subjects – for example, medical graduates end up working in hospitals across the country, while graduates of some other subjects are much more concentrated in large cities.
Figure 1: Change in the cost of living in England
This leads us to ask how adjusting for differences in the cost of living based on the area of residence of individuals affects the estimates of returns to different universities and subjects. To answer this question, this report uses newly linked data on area of residence in early adulthood to explore where graduates live in early adulthood and estimate how returns to tertiary education (ES ) change once we adjust gross incomes for cost of living differences across the country. .
Figure 2: 27-year destination of University of Bolton (left) and LSE (right) graduates
We start by showing that there are indeed big differences in the place of residence of graduates of different universities after the end of their studies. Around 60% of those who attended a university in London still live there at 27, while less than 20% of graduates from institutions outside London live in London at 27. For universities outside London, there is a strong negative relationship between selectivity and the proportions of graduates remaining in the region and a strong positive relationship between selectivity and moving to London. For example, around half of Oxbridge graduates live in London at age 27, while this proportion is less than 10% for the less selective universities.
As London has the highest cost of living in the country, it also means London universities and the more selective institutions outside London are seeing the biggest drops in performance – up to 20%. percentage points – once we adjust income for differences. in the cost of living. On average, people who have attended university live in more expensive areas than those who have not, and adjusting income to the cost of living decreases the average yields of institutions by about 3 percentage points. percentage. However, a few universities show very small increases in yields. These are usually institutions located in low cost areas with a high proportion of graduates from and staying in the local region, as well as specialized agricultural colleges.
While the performance changes may be quite large for some institutions, and London institutions in particular, the cost of living adjustment does not significantly change the overall ranking of universities. The correlation between the rank of corrected and unadjusted returns is 0.97 for men and 0.96 for women. It’s important to note that no university goes from a significantly negative performance to a significantly positive performance or vice versa, once we have adjusted the cost of living.
When we look by topic, perhaps unsurprisingly, we only see very small changes in return estimates when we adjust income for cost of living. The three main counter-examples to this are economics (where returns are decreasing), medicine and veterinary science (where returns are increasing). However, we do not see the overall order of subject returns changing much, with economics and medicine still leading the way, and veterinary science moving from one of the least profitable subjects to a rank subject. intermediate. Otherwise, the subject’s estimates are broadly unchanged.