Poverty in Canada (2nd Edition)

THE NECESSITIES APPROACH to the measurement of poverty is based on the premise that it is important for a society to know how many of its members cannot afford to provide themselves with all the basic necessities of life. In the previous two chapters, we have attempted to determine, with some rigour, the cost of basic necessities in Canada in the reference year, 1988. Essential food costs have been estimated by family size for 25 major metropolitan areas in each of the 10 provinces. Essential shelter costs by family size have been determined for metropolitan centres with a population of at least 50,000—53 cities in all. And finally, the cost of other necessities by family size has been calculated in such a way as to reflect an approximate Canadian average. Therefore, we have sufficient information to determine "poverty lines" for (at least) 25 major Canadian communities. The combined 1986 population in these 25 communities is 14,318,851 or about 57 percent of the overall population of Canada.

Using the necessities approach, the poverty line for a given family size is simply the sum of the costs of food, shelter and other necessities. This dollar value tells us the minimum income required to purchase all basic needs. The results of these calculations, by major city and by size of family are displayed in table 7-1.

Table 7-1: Poverty Lines ($)—1988, by Major City and Family Size
		Family size
City	Category	1	2	3	4	5	6
St. John's	Food	$1,111	$2,078	$3,091	$4,183	$5,171	$6,108
	Shelter	4,145	4,920	5,618	5,913	5,913	5,913
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	6,495	8,868	11,215	13,233	14,857	16,426
Charlottetown	Food	1,012	1,892	2,815	3,811	4,710	5,564
	Shelter	3,469	4,036	4,985	5,956	5,956	5,956
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	5,720	7,798	10,306	12,904	14,439	15,925
Sydney	Food	1,078	2,016	2,999	4,059	5,018	5,927
	Shelter	2,815	3,469	4,385	6,153	6,153	6,153
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	5,132	7,355	9,890	13,349	14,944	16,485
Halifax	Food	1,023	1,913	2,846	3,852	4,762	5,625
	Shelter	4,342	5,051	6,011	7,244	7,244	7,244
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	6,604	8,834	11,363	14,233	15,779	17,274
Moncton	Food	1,001	1,872	2,785	3,769	4,659	5,504
	Shelter	3,273	4,091	4,680	5,269	5,269	5,269
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	5,513	7,833	9,971	12,175	13,701	15,178
Saint John	Food	979	1,831	2,723	3,686	4,557	5,383
	Shelter	3,153	3,687	4,255	4,658	4,658	4,658
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	5,371	7,388	9,484	11,481	12,988	14,446
Chicoutimi	Food	$1,232	$2,304	$3,427	$4,639	$5,734	$6,774
	Shelter	3,044	3,633	4,265	4,593	4,593	4,593
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	5,515	7,807	10,198	12,369	14,100	15,772
Quebec	Food	1,155	2,160	3,213	4,349	5,376	6,350
	Shelter	3,665	4,647	5,193	5,989	5,989	5,989
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	6,059	8,677	10,912	13,475	15,138	16,744
Trois Rivieres	Food	1,067	1,995	2,968	4,081	4,966	5,867
	Shelter	3,404	3,749	4,036	4,451	4,451	4,451
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	5,710	7,614	9,510	11,606	13,190	14,723
Sherbrooke	Food	1,078	2,016	2,999	4,059	5,018	5,927
	Shelter	2,869	3,545	4,156	4,931	4,931	4,931
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	5,186	7,431	9661	12,127	13,722	15,263
Montreal	Food	1,100	2,057	3,060	4,142	5,120	6,048
	Shelter	3,862	4,800	5,564	6,545	6,545	6,545
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	6,201	8,727	11,130	13,824	15,438	16,998
Hull	Food	902	1,687	2,509	3,396	4,198	4,959
	Shelter	3,742	4,538	5,073	5,716	5,716	5,716
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	5,883	8,095	10,088	12,249	13,687	15,080
Ottawa	Food	$  935	$1,748	$2,601	$3,521	$4,352	$5,141
	Shelter	4,047	5,138	6,480	7,735	7,735	7,735
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	6,221	8,756	11,587	14,393	15,860	17,281
Toronto	Food	1,067	1,995	2,968	4,018	4,966	5,867
	Shelter	4,462	5,378	6,502	8,051	8,051	8,051
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	6,768	9,243	11,976	15,206	16,790	18,323
Hamilton	Food	1,045	1,954	2,907	3,935	4,864	5,746
	Shelter	3,404	4,244	5,149	6,480	6,480	6,480
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	5,688	8,068	10,562	13,552	15,117	16,631
London	Food	1,100	2,057	3,060	4,142	5,120	6,048
	Shelter	3,589	4,484	5,564	6,862	6,862	6,862
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	5,928	8,411	11,130	14,141	15,755	17,315
Sudbury	Food	990	1,851	2,754	3,728	4,608	5,443
	Shelter	3,633	4,364	5,062	5,422	5,422	5,422
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	5,862	8,085	10,322	12,287	13,803	15,270
Thunder Bay	Food	1,111	2,078	3,091	4,183	5,171	6,108
	Shelter	3,087	4,691	5,891	6,295	6,295	6,295
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	5,437	8,639	11,488	13,615	15,239	16,808
Winnipeg	Food	$  902	$1,687	$2,509	$3,396	$4,198	$4,959
	Shelter	3,458	4,549	5,651	6,785	6,785	6,785
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	5,599	8,106	10,666	13,318	14,756	16,149
Regina	Food	946	1,769	2,632	3,562	4,403	5,201
	Shelter	2,836	4,320	5,345	6,284	6,284	6,284
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	5,021	7,959	10,483	12,983	14,460	15,890
Saskatoon	Food	946	1,769	2,632	3,562	4,403	5,201
	Shelter	2,989	4,102	4,898	5,575	5,575	5,575
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	5,174	7,741	10,036	12,274	13,751	15,181
Edmonton	Food	1,001	1,872	2,785	3,769	4,659	5,504
	Shelter	3,415	4,320	5,258	5,782	5,782	5,782
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	5,655	8,062	10,549	12,688	14,214	15,691
Calgary	Food	1,067	1,995	2,968	4,018	4,966	5,867
	Shelter	3,404	4,440	5,705	6,360	6,360	6,360
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	5,710	8,305	11,179	13,515	15,099	16,632
Vancouver	Food	1,023	1,913	2,846	3,852	4,762	5,625
	Shelter	4,353	5,258	6,884	8,073	8,073	8,073
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	6,615	9,041	12,236	15,062	16,608	18,103
Victoria	Food	$1,045	$1,954	$2,907	$3,935	$4,864	$5,746
	Shelter	3,556	4,331	5,618	5,924	5,924	5,924
	Other	1,239	1,870	2,506	3,137	3,773	4,405
	PL	5,840	8,155	11,031	12,996	14,561	16,075

Regrettably, we cannot directly determine the number of poor persons in our major cities. There exists no published statistical information on income distribution by city and family size. The census provides city income distributions but with families of all sizes lumped in together. Statistics Canada annual household survey Income Distributions by Size in Canada presents income distributions by family size for provinces but not cities. Thus, the data do not permit the measurement of poverty in Canadian cities. However, this is really not a serious limitation. In fact, it is important to determine poverty lines and the extent of poverty in the whole of Canada and not just in the roughly 57 percent of the population represented in table 7-1. Therefore, it is appropriate to move on to the construction of essential living costs by province.

In determining provincial poverty lines, the major issue to be resolved is how to treat the numerous towns, villages and rural communities for which food or shelter data are not available. Casual observation of the data we do have suggests that food and shelter should be treated differently. For example, it appears, from table 6-4 that shelter costs are positively correlated with city population. If this is true, then the thousands of smaller communities not included would have lower shelter costs than those which are included. With food, there does not appear to be any particular relationship between costs and size of community. In fact, food costs seem to have something of a regional pattern—generally higher in the East and lower in the West—if anything at all.

Statistical analysis confirms these suspicions. [Based on regressions done by the author for this study.] City population is found to be strongly significant in explaining shelter costs in major cities across Canada. This result also holds for cities within our two largest provinces—Ontario and Quebec. With food on the other hand, the size of city does not help explain essential food costs and there is no significant correlation between city population and essential food costs. Given this evidence, it would not be inappropriate to calculate a weighted average of essential food costs for each province using just the cities in table 5-8. We have no reason to believe that food costs for those communities omitted would be significantly different. [At this point, some readers will argue that food costs in small towns and rural areas are higher because major grocery stores are not as accessible. While many residents of rural communities make frequent treks to cities for shopping and medical needs, it is true that some day-to-day items may cost more. However, these additional food expenses may be offset by garden plots and direct purchases from farmers. Indeed in 1986, average family food expenditure in both small urban and rural areas of Canada were below the overall average. (Famex (1986), pp. 40-1.)]

It would be quite inappropriate, however, to do the same for shelter costs. As was pointed out in chapter 6, apartment rents are lower in towns with populations between 10,000 and 50,000 than in cities with populations above 50,000. There is no data on shelter costs in smaller communities and rural areas. In the absence of complete information, it is assumed that, within each province essential shelter costs in the omitted areas are approximately equal, on average, to the least expensive shelter cost among included cities. In Ontario, for example, the city of Sudbury had the lowest overall shelter costs in 1988 among all listed cities. Our assumption means that shelter costs in the smaller communities in Ontario, comprising about 22 percent of the population, would be no higher, on average, than those in Sudbury.

In each of three of Canada's provinces—Newfoundland, Prince Edward Island and Manitoba, there is only one city with a population of 50,000 or more people. In these cases average essential shelter costs for the province will be clearly overestimated. However, in these areas it is preferable to let major city costs reflect overall provincial costs than to make some arbitrary adjustment.

Finally, the cost of "other" necessities (clothing, transportation, personal hygiene and household items) have been calculated for Canada as a whole (not by province or city) and vary only with family size.

These assumptions permit the construction of provincial poverty lines by family size. Weighted [Weighted by (1986) population of metropolitan area.] average food and shelter costs among the cities in each province are summed and "other costs" added in to get provincial totals. The results are presented in table 7-2. As might be expected, there is a somewhat smaller range of poverty line values among provinces than was the case with cities. Specifically, there is only a 16 percent difference between the least expensive (New Brunswick) and most expensive (Ontario) provinces compared with a range of about 25 percent between the least and most expensive cities. It would not be incorrect to say that there is broad similarity in provincial average essential living costs across the country.

Click here to view Table 7-2: Poverty lines in Canada by province and family size ($),1988

The estimation of poverty in Canada

The basic source of data used to determine the extent of poverty in Canada is Statistics Canada's Microdata File for Economic Families (1988). The information on the tape is compiled from the Survey of Consumer Finances which Statistics Canada has conducted annually since 1971. The sample survey, over 36,000 households in total, is representative of virtually all private households in Canada except for residents of Yukon and N.W.T., households located on Indian reserves and inmates of institutions. The current survey was conducted in April 1989 and respondents were asked by interviewers to supply information on their 1988 incomes. The following income sources were included: wages and salaries, net income from self employment, investment income (interest and dividends), net rents, estate income, government transfer payments from all levels of government including child and sales tax credits, pensions of all types, as well as income from scholarships, alimony and miscellaneous income. Specifically excluded as sources of income are: gambling gains and losses, money inherited during the year in a lump sum, capital gains or losses, receipts from the sale of property or personal belongings, loans received, lump sum settlements on insurance policies and rebates of property taxes and other taxes. All "in-kind" forms of income were similarly excluded. Finally, the income distribution publications exclude family units and individuals whose major source of income was military pay and allowances.

Estimation of the incidence of poverty using StatsCan's published income distribution tables (such as Income Distributions by Size) would require a process known as linear interpolation. This is because published ranges will not correspond to calculated poverty lines. Linear interpolation, a simple method of determining within range percentages, will result in fairly accurate aggregate estimates. However, the use of the microdata file, which includes every household in the sample, permits precise calculations and is therefore preferable. [The use of linear interpolation results in slightly (about 5 percent) exaggerated estimates of poverty. This is clearly because incomes are not evenly distributed within the selected lower income ranges but are rather "bunched" somewhat towards the higher end of the range.]

For those who are unfamiliar with the published income distribution tables, an example is reprinted below in table 7-3. This table shows the distribution of income in 1988 by family size for Ontario. Statistics Canada groups families with five or more persons together because of the small numbers of households involved. I will do the same both for convenience and to ensure close correspondence between my tables and published tables.

Click here to view Table 7-3: Percentage distribution of families by income groups, provinces, and size of family, 1988

Using the microdata file it was found that the average size of families with 5 or more persons was almost exactly 5.5 persons. Using this value, the poverty line for the 5 and over category is simply the arithmetic average of the poverty lines for families of sizes 5 and 6. This methodology is employed for simplicity for all provinces.

Applying the provincial average poverty lines to the microdata file, the extent of poverty by family size in each province in Canada was determined. The percentages as well as the absolute number of persons who reported incomes less than the given poverty lines are summarized below in table 7-4. At the bottom of the table, the aggregate values for Canada are given. In some cases, the household counts in the microdata file differ slightly from the published counts (i.e., in Statistics Canada's Income Distributions by Size in Canada, cat. 13-207). These differences are due to the masking of special family units to ensure confidentiality.

Click here to view Table 7-4: Estimation of poverty by province, 1988

This approach, based on the cost of necessities in the various parts of Canada, reveals that just under 1 million Canadians or about 4 percent of the population are poor. The rates of poverty are somewhat higher in Atlantic Canada but are fairly consistent across the country. Undoubtedly, there would be greater variations between cities. The poverty rate averages about 9 percent for unattached individuals and about 3 percent for persons in families. By comparison, Statistics Canada incidence of "low income" for 1988 (1986 base) was 37.7 percent for unattached individuals and 12.2 percent for families. To the extent that poverty is appropriately defined as insufficient income to afford all of the basic necessities, StatsCan's (1986 base) "low income" percentages represent a roughly four-fold exaggeration of the extent of poverty in Canada. Table 7-5 presents a full comparison of poverty rates using the various approaches for 1988.

Click here to view Table 7-5: Comparison of poverty rates and number of poor, 1988

One of the major advantages of the "necessities approach" is that it enables us to measure changes in the incidence of poverty over time in the absence of changes in the distribution of income. In other words, it allows us to track our success or failure in improving the lot of the least advantaged in our society. With that in mind, poverty lines were constructed for 1985 using the same methodology. While it would have been preferable to make comparisons to earlier periods, data on rents and food prices were not consistently available before 1985.

Income distributions for 1985 lump together the four "maritime" provinces and the three "prairie" provinces rather than treating each province separately, as has been done more recently. Therefore, a weighted average poverty line for each of these two regions was constructed. Using Income Distributions by Size in Canada (1985), [In this case the microdata file was not available so estimates were made using linear interpolation within published income ranges.] the extent of poverty was determined for the five areas of Canada—Maritimes, Quebec, Ontario, Prairies and B.C. The results, including the aggregate totals for Canada, appear in table 7-6.

Click here to view Table 7-6: Estimation of poverty by province, 1985

While some general patterns of poverty were similar, for example, somewhat higher rates in the Maritimes and for unattached individuals, the overall incidence of poverty in Canada was higher in 1985 than was the case in 1988. In fact, about 42 percent higher. [This figure takes into account the likely error due to interpolation.] In 1985 there were about 400,000 more people living in households where the reported income was insufficient to acquire all of the basic necessities, than was the case in 1988. This is not surprising considering the state of the economy in the mid 1980s and the steady improvements after that time. The national unemployment rate fell from about 10.5 percent in 1985 to 7.8 percent in 1988. [Statistics Canada, Canadian Economic Observer, cat 11-210, "Historical Statistical Supplement," 1988/89, table 2-5, p. 34.] The unemployment rate among young people less than 25 years old, the age group with by far the highest incidence of poverty, declined from 16.4 percent to 12 percent over the same period. [Ibid.] At the same time, there was no change in the underlying distributions of income, as represented by the quintile proportions in table A-1. This implies that the significant decline in the number of poor has been due primarily to real economic growth.

Between 1985 and 1988 our poverty lines have kept pace with the inflation rates in the various regions, increasing an average of just over 4 percent per year. Since 1988, the inflation rate has crept up to about 5 percent. Using this value, the poverty lines in table 7-2 can be "updated" for 1989 and 1990, and forecast for 1991 and 1992. The results are displayed in table 7-7. Inflation adjusted poverty lines should not be confused with revised poverty lines. The latter involves a complete recalculation of the cost of basic necessities using actual relevant costs and prices while the former uses a single average rate of inflation to increase all existing poverty lines uniformly. As long as poverty lines are revised every three or four years, [Because the necessities approach uses a simple, direct methodology and published data (largely), there is no reason why these poverty lines could not be revised annually. There would, of course, be a delay of approximately 1 to 2 years in obtaining relevant information from such sources as Statistics Canada and the CMHC.] inflation adjusted poverty lines are a reasonably good proxy in the interim.

Click here to view Table 7-7: Poverty lines in Canada by province and family size ($)