5 Differences in the growth of degree holders across the urban-rural hierarchy

View the most recent version.

Given that the incidence of degree holders differs greatly across urban and rural areas (see Figure 2), it is also true that this gap has been widening over time. From 1996 to 2001, the incidence of degree holders increased the most in large urban areas and the least in rural areas (Table 1). Large urban areas increased their share of degree holders by 2.9% over the period. The increase in share falls to 1.6% for medium urban areas or just a little over half of the increase for that oflarge urban areas. Small urban areas and rural areas increased their shares the least, 1.1% and 0.9%, respectively, or about one third of the increase in degree-holder share experienced by large urban areas.

Table 1
Degree holder percentage shares across urban and rural classes, 1996 to 2001

Recall that the question at hand concerns the relative roles of in situ and net migratory effects as drivers of human capital growth. We start with an aggregate view of the system by considering changes in human capital across the urban-rural hierarchy (Table 2). At the national scale, the number of degree holders increased by nearly 670,000 from 1996 to 2001, with nearly two thirds of the growth in human capital attributed to in situ growth, with the remaining one third attributable to immigration, given that there is no net migration effect at the national scale.

Large cities were by far the major beneficiaries of this growth in human capital. While accounting for only 51% of Canada's population aged 20 and above in 2001, large urban areas accounted for almost 75% of the growth in degree holders. The major contributors to this growth are in situ and net international migration, which accounted for 50% and 42%, respectively. Net internal migration accounted for only 8% of their growth in degree holders. Hence, the relative strength in the growth of degree holders in large urban areas can be attributed to strong internally generated growth ( in situ ) and international migration effects, but not to internal migration.

Moving down the urban-rural hierarchy, each remaining urban-rural class experienced an increase in degree holders. Net internal migration reduced the stock of degree holders in each of these regions, particularly medium centres. In total, these lost 23,830 of their degree holders through net internal migration, which amounts to a 35% reduction in the growth of their stock of degree holders. Small urban areas behaved similarly, with internal migration reducing the growth of degree holders by 22%, or 12,545. In both cases, in situ growth was the primary source of growth. For rural areas, growth in human capital came exclusively through in situ growth; while rural areas had a modest net inflow of immigrants, their contribution to human capital was washed out by out-migration.

Table 2
Components of degree holder growth, numbers and percent, by urban-rural class, 1996 to 2001

Much of the growth of degree holders occurs, naturally, at younger ages. Moreover, it is the accumulation of degree holders during the early part of their careers that has the strongest impact on the share of degree holders in a place in the longer run, because mobility tends to decrease with age. Hence, it is important to obtain not only an overall picture of degree-holder growth across urban and rural areas, but also of growth across different age classes. Beyond these reasons for looking at degree-holder growth at different ages, as will become apparent, there are also limitations to our census-based breakdown that influence our conclusions about the significance ofinternal migratory flows. The growth in degree holders by age group nationally is presented in Table 3.

Table 3
National components of degree holder growth, by age group, 1996 to 2001

Recalling that in situ growth accounted for nearly two thirds of the growth of degree holders nationally, it is quickly evident from the data that the young—the 20-to-24 and 25-to-29 age classes—contribute to over two thirds of this growth. Within these two age groups, 96% and 87%, respectively, of the increase in their share of degree holders comes from in situ growth. While the contribution of immigration to human capital growth is modest in these two age groups—4% and 13% for the 20-to-24 and 25-to-29 groups, respectively—its contribution quickly increases among those aged 30 and over. Above the age of 30, immigrants made the largest contributions in all but the oldest age group. In other words, immigration substantially raised the level of human capital throughout most labour force age groups.

Disaggregating by city size complements the results noted above and highlights the differential contributions by age group and across the urban-rural hierarchy (Table 4). Recalling that in situ growth contributed 50% of the growth in degree holders in large urban areas, this growth was again largely confined to young adults (aged from 20 to 29). Net domestic migration into large centres enables a modest growth in human capital through most labour force age groups, but net immigration is the most significant contributor, particularly among those aged from 30 to 54, those age groups that capture the majority of recent arrivals. Among older age groups, net migration and in situ growth replace immigration as the drivers of human capital growth.

At the other end of the urban-rural spectrum, changes in the level of human capital in rural areas are primarily associated with in situ growth. While there is a very modest increase in the level of human capital over that observed in 1996 (47,694), neither immigration nor net internal migration had a significant impact on the level of human capital in these areas, consistent with Looker (2001). However, underneath this general pattern there are some interesting patterns that emerge.

Net internal migration does drive human capital growth in rural areas for those at the start and the end of their careers. For those aged from 20 to 24 there are substantial losses of degree holders. If these newly minted degree holders had stayed, the degree-holder growth would have risen by 50%. However, for those aged from 30 to 39, net migration turns positive. It is unclear why this is the case, but it may be because degree holders are returning home. For those in the later part of their career—those aged from 50 to 64—we also see a positive effect of net migration, perhaps reflective of return migration to rural areas toward the end of the labour force career or in retirement (Newbold and Bell 2001). Among those aged 65 and over, the decline in human capital is partially attributable to death.

Medium and small urban areas share similar profiles. In both cases, migration and in situ growth were the drivers of human capital change. In situ dominates the growth of degree holders in medium and small urban areas for younger ages (from 20 to 29) as net international migration is weak and the effect of net internal migration is negative.

Perhaps the most distinctive characteristic of medium and small urban areas is the significant loss of degree holders because of net internal migration among degree holders in their 20s. In other words, had there been no migration out of these communities, the level of human capital would have been much higher. This result likely reflects out-migration of the young from medium urban centres and, to a lesser degree, from small urban centres to large urban centres following completion of their postsecondary education.

This result has to be interpreted with considerable care. On the one hand, the large losses of degree holders from small and medium urban areas suggest that these urban classes are able to generate a large number of degree holders, but that they are unable to retain them as large urban labour markets prove to be far more attractive. On the other hand, we also have to recognize that many of these small- and medium-sized cities may be university towns (for example, Guelph, Kingston or Lethbridge); that is, they specialize in the export of education services. Hence, a place like Kingston may not expect to keep the majority of the degree holders it is able to produce. A telltale sign of this is the larger number of degree holders in the 25-to-29 age group that leave medium urban areas. These degree holders are likely noting on their census forms the place where they were attending university five years before, rather than their home town. In effect, the census is, at least in part, identifying where degree holders obtained their degrees and not where they spent their formative years. In fact, many leaving medium urban areas may be returning home. In this sense, we are observing a 'pass through' effect rather than a loss of the young that grew up in these small urban and medium urban areas. The bottom line is that it is very difficult to interpret the flows of degree holders in the 25-to-29 age group and therefore we should treat it with caution.

A more reliable measure of the effect of migration on degree-holder growth is to focus on the 20- to-24 age group. Five years previous, these new degree holders ranged in age from 15 to 19. Therefore, they are more likely to indicate that they were residing in a place where they spent their formative years, rather than where they attended university.

Looking across the urban-rural hierarchy, the migration of these new graduates follows a distinct pattern. Rural and small urban areas lose a significant number of degree holders, reducing their number of degree holders by 49% and 26%, respectively. Medium urban areas effectively experience no net migratory effect, while large urban areas experience a net positive effect. There, the number of degree holders in the 20-to-24 age group increased by about 11% because of net internal migration. Based on this group of new university graduates, it is apparent that migration does matter, potentially driving down the incidence of degree holders in the smaller urban and rural parts of Canada. But, although these flows appear to be large for these areas, they amount to just a small fraction of the growth of new graduates in larger urban areas.

In total, two broad patterns emerge in the growth of degree holders across urban and rural areas. The very strong growth in the share of degree holders in large urban areas is driven by relatively high rates of in situ growth and immigration. Internal migration plays a much weaker role. In effect, large flows out of medium and small urban areas to large urban areas are swamped by the volume of degree holders generated within large urban areas. In this sense, we cannot account for the high rates of degree attainment in large urban areas by the sorting of degree holders across space, at least within Canada. In medium urban cities, in situ growth dominates, and immigration plays only a very minor role. For small urban and rural areas, net migration plays an important role, reducing the number of degree holders in these areas. This raises the question that if there were no migration, would the incidence of degree holders in smaller urban areas and rural areas match that of medium and large urban areas, or do their populations in general obtain degrees at a relatively low rate. This question is addressed in the next chapter.

Table 4
Components of human capital growth, numbers and percent, by urban-rural class and age group, 1996 to 2001