4.0 Conclusion

View the most recent version.

In this report, the traditional cross-sectional analysis of housing affordability using shelter cost-to-income ratios (STIRs) has been extended by adding an examination of longitudinal data. Until this report, all studies on housing affordability in Canada were cross-sectional in nature, i.e., referring to a specific point in time, often to a single year.

The report begins by drawing on 2002 through 2004 SLID cross-sectional data to show that the proportion of Canadians living in households spending above the affordability benchmark of 30% at any one point in time — in any single year — tends to be stable at around one in five or 20%. It then capitalizes on SLID's longitudinal capabilities to study the dynamics of housing affordability of Canadians over a longer period of time by creating longitudinal estimates of housing affordability for the same three year period.

While cross-sectional data indicate a stable 20% of Canadians living in households spending above the affordability benchmark for shelter in any single year, longitudinal data give a broader perspective. When measured over a three year period, 28% of Canadians reported ever living in a household exceeding the affordability benchmark, indicating that this situation is more prevalent than shown by cross-sectional estimates alone.

Another way of looking at this is to divide those who ever exceeded the benchmark into those who did so for one year (12%), for two years (7%) and for all three years (9%). This shows that roughly one-third of those exceeding the benchmark at least once during the study period can be considered to be persistently doing so, while the other two-thirds are moving in and out of this state. However, a three-year study period is rather short. Some of the seemingly transient group may be ending or starting a prolonged period of exceeding the benchmark. When more years of data are available, we should be able to understand more about the dynamics of these households.

To identify the factors associated with spending above the affordability benchmark, two logistic regression models were run. They examined the correlates of a Canadian living in a household either "persistently" (all three years) or "ever" (at least one year) exceeding a STIR of 30%. Both models corroborated the findings from the crosssectional and longitudinal analyses presented in the descriptive section of this report. They found the attributes associated with the highest probabilities of living in a household spending above the affordability benchmark were: living alone, being a female lone parent, renting, being an immigrant, and living in Vancouver or Toronto.

In addition, the "ever" regression model contributed the important finding that those living in households experiencing some kind of transition during the study period have a higher probability of exceeding the benchmark at least once during the study period. Such transitions included renters with a change in rent subsidy status, those who changed housing tenure from owner to renter or vice versa, those who changed family type, for example, by marrying or divorcing, and those who moved between the cities examined in this study. Perhaps equally important was the finding that those experiencing these transitions did not do so persistently.

The models also show that renters in subsidized housing for all three years of the study period, while experiencing probabilities similar to market renters of exceeding the benchmark for at least one year, had lower probabilities of persistently doing so. This was in spite of having a median income approximately half that of market renters.

To conclude, the longitudinal capability of SLID enhances our understanding of housing affordability in two key ways:

• First, the proportion of Canadians ever living in households spending 30% or more of their before-tax household income on shelter is considerably larger than that indicated by traditional cross-sectional data. This study's three year panel data for 2002-2004 identifies almost 50% more Canadians experiencing these conditions over three years than did so in any single year according to traditional cross-sectional data; and
• Second, the majority (about two-thirds) of Canadians reporting ever living in households spending above the 30% STIR benchmark do not do so on a persistent basis year over year, but rather on an occasional basis, and that often the factors associated with their being in a household spending 30% plus are associated with some kind of transitional change.

This study, the first to explore the SLID longitudinal shelter cost data, was designed to advance our knowledge of patterns of affordability and build a base upon which further research can be carried out.