Concentrations of Poverty and Distressed Neighbourhoods in Canada - January 1997
| Previous | Contents | Next |
4. Results from the Canadian Data
4.1 Concentration of Low Income
Figure 1: Percentage of Poor Canadian Families in Very Poor Neighbourhoods 1980 and 1990
In 1980 tracts with at least double the family22 low income rate for Canada accounted for 11.8 percent of all low income families in Canada. By 1990, as the preceding chart indicates, such tracts accounted for 17.3 percent of all low-income families in Canada.23 Thus, over the decade, low-income families became more concentrated in census tracts with extremely high family low-income rates, just as in the USA.
As the following table shows, this resulted from an increase in the number of tracts with extremely high low-income rates between 1980 and 1990.
| Extremely Poor | Total | 1980 and 1990 | 1980 but not 1990 | 1990 but not 1980 |
| 1980 | 334 | 263 | 71 | NA |
| 1990 | 507 | 263 | NA | 244 |
There were 334 census tracts with family low-income rates over 26.0 percent in 1980 of which 263 also had low income rates over 26.4 percent in 1990. These 263 tracts accounted for 9.9 percent of all low-income families in 1980, but for only 9.2 percent of low-income families in 1990. It was the addition of 244 tracts which met the 26.4 percent threshold in 1990, but not the 26.0 percent threshold in 1980 which accounted for the increased share of low income families living in tracts with extremely high family low income rates.
This is similar to the pattern observed by Jargowsky in the USA. Increases in the concentration of the poor there occurred not because the population of high poverty tracts went up. Indeed tracts which were extremely poor in both 1980 and 1990 tended to suffer net population losses. The concentration of the poor increased in the USA because the number of tracts afflicted by extreme poverty rates went up as adjoining neighbourhoods also crossed the 40 percent threshold.24
Another interesting aspect of the growing concentration of low income families in tracts with abnormally high family low income rates in Canada is that it was focused in that country's nine largest census metropolitan areas.25 These centres accounted for over 100 percent of the entire increase in the number of low-income families in Canada between 1980 and 1990. Moreover, extreme poverty tracts within many of these metropolitan areas accounted for a much larger share of the total number of low income families in 1990 than in 1980.
| Census Metropolitan Area | Concentration of Poor 1980 | Concentration of Poor 1990 |
| Montreal | 30.1% | 40.1% |
| Winnipeg | 23.5% | 39.0% |
| Edmonton | 4.1% | 28.3% |
| Calgary | 6.4% | 20.3% |
| Quebec City | 20.8% | 26.6% |
| Vancouver | 7.2% | 15.5% |
| Toronto | 14.7% | 21.4% |
| Ottawa-Hull | 27.5% | 24.1% |
As can be seen from the preceding table, the "concentration of the poor", as defined by Jargowsky, rose from 30.1 percent to 40.1 percent in Montréal between 1980 and 1990, from 23.5 percent to 39.0 percent in Winnipeg, from 4.1 percent to 28.3 percent in Edmonton, from 6.4 percent to 20.3 percent in Calgary, from 20.8 percent to 26.6 percent in Québec City, from 7.2 percent to 15.5 percent in Vancouver and from 14.7 percent to 21.4 percent in Toronto. Only in Ottawa-Hull among the nine largest Census Metropolitan Areas (CMAs) was there a reduction in the concentration of the poor (from 27.5% in 1980 to 24.1% in 1990).
As Table 3 reveals, a similar trend was found with neighbourhood poverty rates.
| Census Metropolitan Area | Neighbourhood Poverty Rate 1980 | Neighbourhood Poverty Rate 1990 |
| Montreal | 14.0% | 20.4% |
| Winnipeg | 9.0% | 15.7% |
| Edmonton | 1.6% | 12.8% |
| Calgary | 2.2% | 8.9% |
| Quebec City | 9.0% | 11.2% |
| Vancouver | 2.5% | 6.1% |
| Toronto | 4.7% | 7.9% |
| Ottawa-Hull | 10.0% | 8.4% |
With the exception of Toronto, the large increases in the concentration of poor families and in neighbourhood poverty rates occurred in CMAs where real average census family incomes grew less than the national average between 1980 and 1990. The large jumps in Calgary, Edmonton Montréal and Vancouver coincided with either a decline (in the case of Edmonton) or an increase well below the national average in this indicator.26
In all cases where the concentration of the poor rose, it was an increase in the number of extreme poverty tracts within these CMAs between 1981 and 1991 which accounted for the increase. In Montréal and Toronto alone the number of such tracts rose from 167 to 273. In Ottawa-Hull, in contrast, the number dipped from 22 to 21.
While the overall family poverty rate outside the nine largest CMAs fell from 13.0 percent in 1981 to 11.7 percent in 1991, in the nine largest CMAs it rose from 13.0 percent to 16.5 percent.
Turning to Jargowsky's other indicator, the Neighborhood Poverty Rate, or the proportion of all families residing in tracted areas with extremely high family poverty rates, this stood at 7.4 percent for Canada in 1981 and at 9.9 percent in 1991. These are higher in level but similar in trend to the USA rates for tracted urban areas alone (3.3% in 1980 and 4.5% in 1990). However, as noted earlier, it is much easier to qualify as an extremely poor tract in Canada than it is to exceed the 40 percent poverty rate employed by Jargowsky as a threshold in the USA.
What is significant is how the increase in both the concentration of the poor and in the neighbourhood poverty rates between 1980 and 1990, mirror the trends in the USA for this decade.
As we have seen, the sources of these increases were also similar-a rise in the number of high poverty tracts in large urban centres; as tracts adjacent to existing high poverty tracts crossed the threshold into high poverty status themselves. Even the bulk of the increase occurred in a few similar large urban centres with Dallas, Houston, Detroit, Cleveland, Milwaukee, Minneapolis and Los Angeles exhibiting growth in the number of their high poverty tracts similar to that experienced in Calgary, Edmonton, Toronto, Québec City, Montréal, Winnipeg and Vancouver. Even the dip in the concentration of the poor in the national capital CMA of Ottawa-Hull was matched by a similar decline in Washington, DC.27
4.2 Distressed Neighbourhoods
Using the indicators and thresholds described in the section on data, 106 (2.7%) of the 3914 census tracts in Canada in 1991 with 1000 or more people met all five criteria of distress. This compares to 880 of 42,865 tracts in the USA (2.1%) which met all four of Ricketts' and Sawhill's underclass criteria in 1980.28 There were also 16 tracts which did not meet the low-income criteria, but which met all four of the other criteria of social and economic distress.29
Of the 106 tracts meeting all five criteria, 60 were found in just three Census Metropolitan Areas?Quebec City (10), Montréal (38) and Winnipeg (12). Over half (54) were in CMAs in the province of Québec.
The total population of these 106 tracts was 330,979 in 1991 or 1.21 percent of the national total. This compares to 2,484,000 people living in the "underclass" census tracts identified by Ricketts and Sawhill in the USA for 1980 which accounted for 1.10 percent of the USA population in that year.30
Over one-third (1364) of all census tracts in Canada with a population of 1000 or more met the criterion for at least one distressed neighbourhood indicator in 1991. The indicator for which the criterion was most commonly met was the proportion of families with children headed by a lone parent (954 tracts), followed by a low income rate more than one standard deviation above the national mean (690 tracts), a percentage of the 15-24 population attending school full-time one standard deviation lower than the national mean (595 tracts), a percentage of tract household income coming from government transfer payments one standard deviation above the national mean (550 tracts) and a percentage of males 15 and over who worked for pay full-time full year in 1990 one standard deviation lower than the national mean (466 tracts).
It should be noted at this point that some of these characteristics, in isolation, are flawed indicators of "distress". For example, a neighbourhood with a high concentration of retired people will, all else equal, have a low percentage of its male population 15 and over working for pay. Moreover, government transfer payments such as Old Age Security and Canada and Quebec Pension Plan benefits may make up a large proportion of neighbourhood income. However, since all but the most affluent seniors receive OAS, and C/QPP retirement benefits are not income tested, neither of these facts would necessarily indicate distress as commonly understood.
Similarly if a neighbourhood's 15-24 population were heavily weighted towards the upper end of the age range, the proportion attending school full-time could be low even if the age-specific school attendance rate was at or above the average.
Thus, it is only when we find tracts which meet the criteria for all or all but one of our indicators that we can be confident that we have found a neighbourhood with significant symptoms of distress.
That being said, clearly the best indicator of the five for identifying a "distressed" neighbourhood is an abnormally low percentage of males 15 or over working full-time, full-year in 1990. Of the 466 tracts meeting this criterion, 22.7 percent met all the other criteria as well. The next most reliable indicator was an abnormally high dependence on government transfer payments as a source of income with 19.3 percent of tracts exhibiting this characteristic meeting all four others. The least reliable indicator was a high proportion of families with children at home headed by lone parents with only 11.1 percent of the tracts with a high concentration of lone mothers having all the other characteristics of distressed neighbourhoods as well.
Interestingly, of the 219 census tracts meeting the criteria for four of the five indicators, the one most often not present was an abnormally low proportion of the 15-24 population attending school full-time. In 133 of these tracts (60.7%) it was this indicator which prevented designation of the neighbourhood as one "in distress." The second most common missing indicator for these "near-distressed" neighbourhoods was an abnormally low percentage of their 15+ male population working for pay full-time full year in 1990 with 49 of the 219 tracts satisfying all the criteria except this one.
As readers will recall, the high school dropout rate for 16-19 year olds was also the characteristic where Ricketts' and Sawhill's "underclass" tracts differed most from extreme poverty tracts (i.e., those with a poverty rate of 40.0% or more).
Of the 219 "near-distressed" census tracts, almost half were also in the Montréal (85), Québec City (13) and Winnipeg (8) Census Metropolitan Areas.
22 Families here refer to economic families of two or more persons. An economic family is a group of individuals related by blood, marriage, adoption or a common-law relationship sharing a common dwelling unit.
23 Of course the low-income rate in 1980 is calculated using the 1978 base Low Income Cutoffs while that in 1990 is calculated using the 1986 base LICOs. This means that some tracts in 1990,which would not have exceeded this threshold using the 1978 base cutoffs, did so because the 1986 base cutoffs were used. For all urban centres of 500,000 or more persons the family low income rate in 1990 using the 1978 base cutoffs was 12.2%, while, using the 1986 base cutoffs, it was 14.5%. For urban centres with populations of 500,000 or more the 1986 base cutoffs ranged from 8.5% (for a family of four) to 14.3% higher (for a family of two) than the 1978 base cutoffs. As mentioned earlier, it was unfortunately not possible to apply a consistent low-income threshold to census tract data for 1980 and 1990, since the differing threshold levels were hard coded into the data for these two years.
24 See Jargowsky, pp. 35-36.
25 These are, in order of 1991 population size, Toronto, Montréal, Vancouver, Ottawa-Hull, Edmonton, Calgary, Québec City, Winnipeg and Hamilton.
26 Real average census family income grew by 7.9% in Canada between 1980 and 1990. It fell by 2.6% in Edmonton and rose by 2.9%, 3.9% and 4.1% in Calgary, Vancouver and Montréal respectively.
27 See Jargowsky pp. 222-232 and p. 250. Neighbourhood Poverty Rates and the Concentration of the Poor for all tracted Canadian Census Metropolitan Areas and Census Agglomerations for the 1980 and 1990 income years appear in Appendices I and II.
28 See Ricketts and Sawhill, p. 323.
29 Of these eight were in Ontario, one was in Alberta and the other seven were in British Columbia.
30 See Ricketts and Sawhill, p. 322. Again, these numbers are not truly comparable because the selection criteria for the distressed neighbourhood tracts in Canada are different from those for the "underclass" tracts in the USA.
| Previous | Contents | Next |