[Statistical] Thoughts on recruiting seniors as economic development
David Clark / February 28, 2016
What follows is not a prediction of what will happen, but rather a message of caution that we need to look deeper than what appears, on the surface, to make common sense: that recruiting seniors is a sound economic development strategy. Maybe it is not.
We know that Canada has a growing “senior” population both in absolute numbers and as a proportion of the population. In 1996, the senior population was 12.2%, 2001 it had grown to 13.0%, in 2006 to 13.7%, and 14.1% in 2011. Ontario, the study area for this article, has shown a similar pattern, growing from 12.9% in 1996 to 14.6% in 2011.
Ontario’s senior population has grown, in absolute numbers, from 1.3 million in 1996 to 1.9 million in 2011, an increase of 40.7%. In 1981 the senior population was 868,000, or 10.0% of the population.
Gerald Hodge, retired planning professor and author (The geography of aging, 2008) has stated that it is not the “total numbers of seniors or their rate of growth that signifies population aging” but rather their proportion of the total population that is the issue. With this as my focus, I looked at several characteristics of communities (at the census division level or county/region level) and how they are related to the senior population.
In a brief study I undertook in 2015 of community characteristics and median income (2006 census data) an interesting phenomenon was noted, specifically that a higher proportion of seniors in a community was strongly negatively correlated (statistically significant) with lower median incomes (r=-.68). This means that the higher the proportion of seniors, the lower the community’s median income. Each senior sub-cohort showed strong negative correlations for those aged up to 79, then moderate, negative correlations for those aged 80-years and older. (The correlations are 65 to 69 (r=-.676), 70 to 74 (r=-.728), 75 to 79 (r=-.668), 80 to 84 (r=-.553), and 84 plus (r=-425)). The correlations for other age groups to median income was only moderately positive; (birth – 19 years; r=.485) and ages 20 – 64, r=.545).
Correlations were similar for other census years, too: 2001 was -.647 and 2011 was -.597. (Note that in 2011 the government made the long-form census voluntary instead of mandatory so data may be biased based on who sent in returns. As such, I will wait for the 2016 census to see if an apparent downward trend, therefore increased proportions of seniors is correlated to higher median incomes, as shown in 2011 is real.)
To assess what the financial impact on a community might be, I ran a simple linear regression of the 2006 data. The results were that for each percentage point of increase in the senior population there is a drop of $623 in the median income.
So what is the pattern of distribution of median incomes across the province? Figure 1 provides a spatial view of the median incomes at county levels. An obvious pattern is that rural areas tend to have lower median incomes and higher median incomes are aligned with major highways and major cities, especially east-west.
I also looked at where seniors are concentrated in higher proportions, using location quotients (LQ). An LQ “shows the extent to which each of a set of areas departs from some norm” (David M Smith, Patterns in human geography, 1975); in this case how each census division differs from the province. LQs are, essentially, the same concept as proportions. LQs were calculated for the number of seniors for each census division in Ontario and compared to the provincial mean. The results (Figure 2) indicate that the relationship between median income and LQs for seniors (aged 65 years and older) is a strong, negative relationship (r=-.671).
Figure 2 shows higher proportions of seniors (purple and blue) in rural areas and the as expected and lower than expected LQs (red, orange, yellow) generally in urban areas aligned with major highways.
Recruitment of seniors is often supported suggesting that there will be more jobs for those in the health care and social service sectors. To test this assumption I looked at the percentage of workers in these sectors and compared them with median incomes and percentages of seniors (2006 census). If higher percentages of seniors do indeed bring in more (higher paid) health and social service sector workers we should see higher median incomes.
The results indicate there is a positive but week correlation of seniors to health care workers (r= .404) so there is a modest increase of workers with higher percentages of seniors). But the increase is only 1/10th of a worker for each percentage point increase in seniors. The correlation of health care workers to median income is moderate but negative (r= – 447), indicating that an increased number of health care workers means a lower median income. And combining social service and health care workers is similar, showing a week negative correlation (r= – .355).
What does this all mean? Well, correlation is not causation so we cannot state that more seniors in a community is the cause of a lower median income. We need to better understand seniors’ spending habits and consider this information when looking to recruit seniors as an economic development strategy.
What this does mean, in my opinion, is that there is something in these correlations and we need to do some deeper analyses to fully understand the relationship of seniors and community income. Blindly stating that senior recruitment will bring wealth, as I have heard stated in my own community from politicians and chambers, is not responsible planning. (And the numbers discussed here suggest this might not be the case.) Senior numbers will increase naturally within communities by aging in place. In addition, some seniors will move “back home” while still others will move communities where they have had a cottage, or move somewhere different for a change of scenery and new adventures. And, what will happen to the housing market, especially at the community level, when a very many baby-boomers, or their estates, put a lot of houses on the market?
Seniors are not a homogenous group. As such it will not be easy to assess the various scenarios that might be impacted by higher proportions of seniors, especially at the community level. But using a more strategic and statistical approach will be important to understanding potential impacts.