2.0 Housing affordability profile

View the most recent version.

2.1 Longitudinal estimates show dynamics of housing affordability

Cross-sectional estimates indicate that around 20% of Canadians (5.5 million) lived in households spending more than the affordability benchmark on shelter costs in 2002. Similar estimates are found for 2003 and 2004 (table 1).

SLID longitudinal data provide a different perspective by estimating the number of people exceeding the benchmark over a period of time. They indicate that 8.6% of Canadians (2.1 million) were living in households that persistently spent above the affordability benchmark between 2002 and 2004, while another 19% (almost 4.9 million) lived in households occasionally spending above the benchmark. In total then, about 28% (7.0 million) lived in households that ever exceeded the affordability benchmark during the study period. (table 2) This proportion is considerably higher than the 19% of people doing so in any single year.

2.2 Low-income households are more likely to spend above the housing affordability benchmark

Household income is a key determinant of STIRs. On average, income-constrained households have higher shelter cost burdens and are more likely than others to spend more than the housing affordability benchmark. In fact, SLID data confirm that, in 2002, over 80% of people living in households exceeding the affordability benchmark fall into the two bottom quintiles and hence are of low or modest income (tables 3a, 3b, 3c). In contrast, those with incomes in the top two quintiles account for only about 6% of all people exceeding the affordability benchmark. This group likely spends above the benchmark out of choice as opposed to necessity.

Nearly 57% of people in the lowest income quintile live in households spending more than the affordability benchmark. Their median STIR, at almost 50%, tends to be a consequence not only of their low income levels but also their relatively high shelter costs¹. A variety of reasons exist for having high shelter costs. For owners it may be high mortgage payments – once mortgages are paid off, STIRs naturally drop considerably. Renters may find that, unless they are in subsidized housing, there is a floor to rents below which accommodation cannot be obtained. In addition, families in this situation may live in cities with more expensive housing, need a bigger dwelling to accommodate a larger family, or lack the resources (social or financial) to search out less expensive accommodation.

While the most obvious reason for low household income is having a low-paying job, there are also other causes. These include: having only one earner, family breakup, job loss, and business or investment losses that are larger than income received (especially for the self-employed). Some households with low income may have other sources of revenue such as capital gains, savings, bursaries, loans, gifts and even charitable support.

In addition, some of these reasons for high shelter costs and low household incomes resulting in high STIRs are only transitory. Finding a job, getting married or moving are examples of events that could lower the STIR. SLID longitudinal data enable us to identify households that made these transitions and track their movements above and below the benchmark over the period of study.

Table 4 presents estimates similar to tables 3a, 3b and 3c but from a longitudinal perspective, not a cross-sectional one. This means that, instead of considering whether people are in households above or below the affordability benchmark at a given point in time, table 4 shows whether they are above or below the benchmark for one, two or three years out of the three-year study period.

Like cross-sectional estimates, longitudinal numbers show that as household income increases, a lower percentage of people live in households that ever spend above the affordability benchmark. However, longitudinal estimates of the percentage of people living in a household ever spending above the benchmark are higher than cross-sectional percentages, regardless of income quintile. As would be expected, over a longer period of time, more people live in households that spend above the affordability benchmark. The comparison also reveals that the higher the income quintile, the wider the percentage difference between the longitudinal and cross-sectional estimates. (table 5). As household income increases, there is a much larger turnover or change in those living in households spending 30% or more of household income on shelter. People in higher income quintiles who ever live in households spending above the benchmark simply do not tend to do so repeatedly or persistently. Instead, new people are entering as others are leaving the group from one year to the next, which leads to higher longitudinal than cross-sectional estimates.

Another way of looking at this is to examine the share of those persistently (all three years) exceeding the affordability benchmark during the three-year period compared to those ever exceeding it (at least one year). In the lowest quintile, almost half of those ever exceeding the benchmark did so for all three years. In contrast, only 7% of those in the highest quintile did so. (The percentages for the second, third and fourth quintiles are 21%, 16% and 8% respectively.) Thus, the higher the income quintile the larger the proportion of people moving back and forth across the affordability benchmark, indicating that the causes of exceeding the benchmark may often be temporary. But for the lower income quintiles, especially the lowest, there is a much higher proportion of people whose STIRs persistently exceed the benchmark, indicating that they are less able to adjust their incomes or shelter costs.

2.3 Who exceeds the affordability benchmark most often cross-sectionally and longitudinally?

As would be expected, results confirm the fact that a higher proportion of renters spend above the affordability benchmark. (table 6). In 2004, roughly one-third of renters (paying either market or subsidized rent) lived in households spending above the affordability benchmark, well over the percentages of 23% for owners with mortgages and 4% for owners without mortgages. Longitudinally, well over 40% of renters ever exceeded the affordability benchmark over the 2002-2004 study period, a much higher proportion than for owners. Meanwhile, those living in households that changed tenure during this period were much more likely ever to exceed the benchmark but less likely to exceed it persistently. This seems to indicate that changing tenure could be associated with temporary affordability difficulties. The study period is, however, too short to have a good understanding of all the dynamics and a few more years are required to gain a better understanding.

Results confirm findings from the Census, Survey of Household Spending and Survey of Financial Security that female lone-parent families and those living alone are the most likely to spend above the affordability benchmark at 44% and 42% respectively in 2004. The proportion of people exceeding the affordability benchmark in each of those groups was more than double the proportion in the population as a whole (20%). Those living alone must pay the entire cost of shelter themselves and rely on whatever income they alone can generate. Those supporting children alone face the double challenge of having only one income and needing to pay for larger accommodation.

Those whose family type changed over the study period are among the most likely ever to spend above the benchmark (39%). As with tenure-changers, this group may not remain long in this state since their three-year rate (7.1%) is below the national average. Other attributes such as number of years since immigration, visible minority status, and some geographical locations also seem to be associated with higher rates of ever or persistently exceeding the affordability benchmark. Recent immigrants in particular had notably higher percentages exceeding the benchmark, both cross-sectionally and longitudinally. These percentages declined as the length of time in Canada increased.

Geographically, Vancouverites were more likely to live in households exceeding the benchmark, with 33% doing so in 2004 and 44% ever exceeding the affordability benchmark over the three-year study period. Similar to those who changed tenure or family type over the study period, a relatively high percentage (41%) of those who changed place of residence exceeded the affordability benchmark at least once during the three-year period. But the higher STIRs associated with the change in residence seem to be temporary, since only 6.6% persistently exceeded the affordability benchmark, well below the national average of 8.6%.

More years of data will be required before strong conclusions can be reached. The rest of this report will start to examine the relationships between these socio-economic variables and housing affordability. Via two regression models, it isolates the individual influence of socio-economic characteristics to explain high probabilities of ever and persistently exceeding the affordability benchmark.

____________

1. Median Shelter-cost-to-income-ratio (STIRs) in this report include households with STIRs equal to or greater than 100%. Overall, roughly 3% of households have such STIRs. However, since a given income quintile (or other sub-population) may have a higher or lower percentage, the effects of this inclusion may vary. Normally, CMHC excludes these households from its affordability studies since it is difficult to interpret their financial circumstances. Possible reasons for STIRs being greater than 100% include: different reference periods for shelter and income; the collection of shelter costs that seem too high (perhaps because, if there is a business operated from home, there is difficulty separating shelter costs from business expenses); fluctuations in self-employment income; and the household having revenue other than standard income to put towards shelter.