Parliamentary Research Branch

PRB 99-1E

COUNTING THE HOMELESS

What is a homeless person? How many are there? These are the two main questions driving the discussion on homelessness. The two questions, albeit distinct, are in fact closely linked; the answer to the first will determine the second. Thus, the estimated size of the group directly depends on the criteria used to define homelessness; the definition of homelessness will itself determine their number. The more inclusive the criteria, the larger the estimate, and vice versa.

We should not be surprised, therefore, to find some significant variations in the estimates. Two U.S. researchers have given estimates for the United States as a whole varying between 250,000 and 3,000,000 homeless.(1) This debate, which is primarily between activists and public servants, is particularly lively in the United States. Estimates world-wide, according to the United Nations, vary between one hundred million and more than one billion homeless. The differences in numbers are explained by the fact that the definition proposed by the UN includes those living in various degrees of unsatisfactory housing: those without a roof over their heads, those who sleep in temporary shelters or institutions, and those living in unsanitary or low-quality accommodation. Depending on which of these definitional components is used, the estimates of the homeless can vary from several million to more than one billion people.(2)

There is much at stake politically in counting the homeless. Very often, human and financial efforts to manage the problem are justified according to the head count. Generally speaking, the greater the estimate, the greater the number of services that will be directed to the homeless.

In comparison to the ongoing debate in the United States, the discussion on the numbers of the homeless is relatively recent in Canada. In fact, it was not until 1987, the International Year of Shelter for the Homeless, that Canadian researchers became interested in counting this population.

The first Canadian attempt to estimate the numbers of homeless people was carried out in 1987 by the Canadian Council on Social Development (CCSD). Its survey was aimed at shedding light on the causes of homelessness, profiling the homeless, determining the scope of the problem and developing some strategies to eliminate it.

All the agencies in Canada that provided temporary or emergency shelter, as well as those providing particular services to the homeless, were sent questionnaires to be filled out on 22 January 1987. Out of a total of 472 questionnaires distributed, only 283 were returned completed.

The survey disclosed that 10,762 people were staying in the shelters inventoried in the survey. These people were for the most part in Ontario (42%), Quebec (17.5%) and Alberta (14%). According to the agencies, during the year preceding the survey there were between 130,000 and 250,000 homeless people in Canada; that is, many thousands of men, women and children had no housing or were poorly housed.(3)

The validity of this estimate has been widely disputed, however. Major criticisms included the failure to include those who on 22 January 1987 were not staying in the documented shelters, the low participation rate of the agencies (283 out of 472), and the exclusive reliance on service providers as informants. With respect to the first criticism, it should be noted that the survey had omitted persons who, at the time of the survey, were doubled up with friends or family members, were sleeping in hotels or homes not included in the survey, were in prisons, hospitals, detoxification centres, or who were sleeping in entrances to apartment houses or abandoned buildings. Consequently, the results underestimated the scope of the problem in Canada. The strategy used by the CCSD has been characterized as an ineffective way of coming up with an overall picture of the homeless.

The second Canadian attempt to count the homeless, carried out by Statistics Canada during the 1991 Census, used a strategy analogous to that used by the CCSD and subsequently perceived as controversial and methodologically suspect.

The Statistics Canada survey, which took place in the course of one day in June 1991, covered about 90 soup kitchens located in 16 Canadian cities. Census-takers based in the agencies asked clients where they had spent the previous night. In contrast to the CCSD survey, however, no results were published by Statistics Canada; in 1995 the agency officially announced that results would not be published because of the mediocre quality of the data.

Not surprisingly, given the method used, the Statistics Canada survey, like the CCSD survey, proved controversial. Some criticisms were that the homeless had been polled at the beginning of the month, when poor people have less recourse to soup kitchens because they have just received their welfare cheques; and that the poll had relied on a single type of agency (soup kitchens) for information when even the officials of such agencies state that these are seldom frequented by some subgroups of the homeless, including young people.

To this day, Canada has no official data on homelessness, a situation that has come in for comment from the United Nations Committee on Economic, Social and Cultural Rights.

   A. The Committee’s Comments

In 1993, after considering Canada’s second report on economic, social and cultural rights, the U.N. Committee explicitly regretted the lack of Canadian data on homelessness:

The Committee notes the omission from the Government’s written report and oral presentation of any mention of the problems of homelessness. The Committee regretted that there were no figures available from the Government on the extent of homelessness, on the numbers of persons evicted annually throughout the country, on the lengths of waiting lists or the percentage of houses accessible to people with disabilities.(4)

Since Canada was unable to provide the Committee with such data, the Committee repeated its criticism in June 1998, when Canada filed its third report. This time, the Committee asked the government, by way of supplementary questions, to

Please provide any available data on the extent of homelessness in various cities in Canada. At what point would the government consider homelessness in Canada to constitute a national emergency?(5)

After the two attempts, in 1987 and 1991, to collect national data on homelessness, the Government of Canada told the UN Committee that the data obtained in these enumeration attempts were neither reliable nor representative. It added that, although some Canadian cities had tried to estimate the scope of the problem, the strategies used and the definitions of homelessness varied so much that it was impossible to compare the results. At the same time, the government spoke of the difficulties associated with gathering data on homelessness.

Enumerating the homeless is indeed an enormous task. Researchers face a number of obstacles, such as the lack of consensus on an operational definition of homelessness; double-counting of the population; the geographical and durational variations in homelessness; and the high costs of enumeration.(6)

The governments of the provinces and territories were also asked to reply to the supplementary questions of the UN Committee on Economic, Social and Cultural Rights arising out of its consideration of the Canadian report. All of them replied that there were no government data on the scope of the homelessness problem in various cities in Canada. Moreover, although there are some data originating from private sources, only the Quebec and Alberta governments presented the Committee with any of these statistics.

The Alberta government told the Committee, that on the basis of a survey conducted by private sources, it was estimated there were between 100 and 1,000 homeless in Calgary, out of a total population of 800,000 inhabitants.(7) The Quebec government reported that some studies had estimated the number of homeless in the province as a whole at 15,000 persons, with 10,000 in the City of Montreal alone. It was noted, however, that these data did not reflect the number of people without shelter every night, but rather the number who had had some experience of homelessness for some period during a year.(8) Moreover, the Quebec government also stated that estimates vary widely depending on the definition of homelessness used, and commented on the difficulty of counting homeless people.(9)

The federal government provided the Committee with the sole fact that just under 26,000 people had used the shelter network in Toronto in 1996.(10) This information was selected on the grounds that "the new City of Toronto has the largest and probably the most robust data series in the country with regard to homeless persons."(11)

The governments could, however, have reported other data to the Committee that would have given different impressions of the scope of homelessness in Canada. In Canada, as elsewhere, opinions differ widely as to the seriousness of the situation, and the political implications of the estimates of the numbers of the homeless are formidable: the size of the estimate directly affects the funding of relevant services, as well as the evaluation of the criteria for access to decent accommodation and for building low-cost housing. (For further information, see the section entitled "Definition of Homelessness.")

To correct the lack of reliable and representative data on homelessness in Canada, and to add to information about the problem as a whole, the federal government told the UN Committee that since 1994 the federal agency charged with implementing the National Housing Act, the Canada Mortgage and Housing Corporation (CMHC),(12) has made homelessness a research priority. For example, in the spring of 1996 the CMHC organized a three-day workshop on the problems of enumerating the homeless. Based on Canadian and U.S. research experience, the workshop helped to identify the most appropriate means of carrying out this task.(13) The federal government also told the UN Committee that the CMHC is developing a computerized research tool that will standardize the collection and management of admission data for services directed to the homeless. This should help to provide a uniform head count of those using such services in Canada and should soon be established in shelters for the homeless.

Research certainly continues to be the best way of advancing our knowledge of homelessness. It is still hard to produce data on the scope of the problem in Canada, whether nationally or provincially, but it is hoped that, thanks to the research undertaken by the Canada Mortgage and Housing Corporation, it will soon be possible to quantify the problem on a national scale. It must be stressed, however, that the major obstacle to counting the homeless continues to be the lack of a consensus on how to define them. (For further information, consult the section entitled "Definition of Homelessness.")

(1) A. B. Shlay and P. H. Rossi, "Social Science Research and Contemporary Studies of Homelessness," Annual Review of Sociology, Vol. 18, 1992, p. 129-160.

(2) Consult the UNICEF Web site at http://www.unicef.org.

(3) Canadian Council on Social Development, Homelessness in Canada, Ottawa, 1987. Note the comment: "A less than successful attempt by Statistics Canada to enumerate the homeless in the 1991 Census means that the CCSD figure represents the best and, to date, only estimate of the number of homeless people in Canada," in T. Peressini, L. McDonald and D. Hulchanski, Estimating Homelessness: Towards a Methodology for Counting the Homeless in Canada, CMHC, Ottawa, 1996, p. 2.

(4) United Nations, Committee on Economic, Social and Cultural Rights, Consideration of Reports Submitted by States Parties under Articles 16 and 17 of the Covenant, 3 June 1993.

(5) United Nations, Committee on Economic, Social and Cultural Rights, Points from the Examination of the Third Periodic Report of Canada on the Rights Covered by Articles 1-15 of the International Covenant on Economic, Social and Cultural Rights, (e/c.12/Q/CAN/1) 10 June 1998.

(6) For a detailed analysis of the various research studies conducted in this field, please see: Daniel Bentley, Measuring Homelessness: A Review of Recent Research, Institute of Urban Studies, University of Winnipeg, Winnipeg, 1995.

(7) The available estimates of the situation in Calgary vary. According to a recent study, close to 3,000 persons do not have access to stable housing in that city. The authors say, however, that a study by the Horizon Housing Society estimated the number of homeless in Calgary in 1989 at between 5,000 and 7,000. For further information, consult H. L. Holley and J. Abroeda-Florez, Calgary Homeless Study: Final Report December 1997, Calgary, 1997. This information was not presented to the UN Committee by the Alberta government.

(8) The estimates on the number of homeless also vary in Quebec. A recent survey by Santé Québec reveals that 28,000 people used shelters and soup kitchens in 1996 in the city of Montreal. The Gazette commented that "The new figure of 28,000 homeless represents a much more troubling reality than the figure of 15,000 that has been used during the last 10 years or so...." ("Homeless Problem Grows," 25 November 1998, p. A5).

(9) Canada, Canadian Heritage, Review of Canada’s Third Report on the International Covenant on Economic, Social and Cultural Rights; responses to the written questions from the Committee on Economic, Social and Cultural Rights (e/c.12/Q/CAN/1) on the occasion of the review of the third periodic report of Canada concerning the International Covenant on Economic, Social and Cultural Rights (E/1994/104/Add17), paragraph 41, provincial and territorial replies section, November 1998.

(10) Ibid., paragraph 41, Canada section.

(11) Ibid.

(12) The purpose of this Act is to improve the housing and living conditions of Canadians.

(13) The participants recommended, inter alia, that estimates of the number of homeless be based solely on services, in view of the high costs associated with a census of street people who use neither services nor shelters for the homeless and the fact that their exclusion would result in the total size of the homeless population being underestimated. For further information, see T. Peressini, L. McDonald and D. Hulchanski (1996).