PRB 00-02E




Prepared by:
Jean Dupuis
Economics Division
12 September 2000





   A. The Decline of Affordable Housing Stock

   B. Rent Control and Other Regulatory Constraints on Housing Markets

      1. Rent Controls

      2. Other Regulatory Constraints

   C. The Disappearance of the Casual Labour Market

   D. Deinstitutionalization

   E. The Crack Cocaine Epidemic






Although homelessness is a persistent manifestation of urban poverty, in recent years it has become the focus of increased attention by politicians, social scientists, urban planners and activists. What are the forces that promote homelessness and why this sudden interest in it?

This paper will attempt to identify the root causes of homelessness by comparing the U.S. and Canadian experiences, especially from the perspective of prominent Canadian and U.S. experts on homelessness and housing. It will suggest several of the influences that may have contributed to the rise of homelessness and brought it into current prominence. Differences between the characteristics of earlier and contemporary homeless populations will also be described.


Throughout U.S. history, the population has contained a sizeable segment of homeless persons. During colonial times, the homeless and destitute received very little sympathy and were sometimes driven out of towns and villages or shunted from community to community, lest they become a serious burden on the relief rolls.

At the end of the U.S. Civil War, released war veterans joined migrant workers to swell the homeless population. Migrant groups of hoboes drifted from town to town in search of shelter and employment. Like homeless people today, they were often treated with suspicion and distrust. Public policy, where any existed, often consisted of warning them to leave town or putting them temporarily in jail.

The consequences of such societal attitudes were severe, given the absence in those days of a "safety net" to protect families against unemployment. In the nineteenth century, the loss of a job – or the death or incapacitation of the household’s wage earner – introduced a real risk of destitution. Peter Rossi(1) notes that communities made a distinction between local homelessness, where families and individuals known to the community fell upon hard times, and transient homelessness, usually involving single persons believed to be homeless as a result of their irresponsible lifestyle. Unattached homeless individuals attracted scorn, while destitute families received more sympathetic treatment.

In the late nineteenth century, homelessness became institutionalized and concentrated in the poorer sections of industrial cities, known as Skid Row. These areas – with their collection of cheap hotels, religious missions and flop houses – were integral components of the urban economy, providing cheap, unskilled manual labour to factories, lumber mills and railroad yards.

The Skid Row population continued to grow until the early decades of the twentieth century, when mechanization of industrial processes – particularly in earth moving and materials handling – reduced the demand for unskilled labour, leaving the Skid Row population without a market for its services.

The Great Depression in the 1930s saw the collapse of economic activity and a substantial jump in unemployment, which lasted for more than a decade and at one point affected 25% of the workforce.(2) Many were compelled to leave their homes and communities in search of job opportunities elsewhere; as a result, the number of the homeless swelled. Attempts were made to enumerate its size during the Depression years: according to these estimates, the homeless population ranged between 200,000 to more than 1.5 million individuals. Rossi notes that this figure is very similar to contemporary estimates of today’s homeless (between 350,000 and several million).(3) The U.S. population was smaller then, however, so the homeless population represented a larger fraction of the total population than it does today.

Urban centres acted as magnets to people seeking employment. Many cities were swamped as a result of this migration, and poorer neighbourhoods saw their populations grow substantially. Because the homeless were often too numerous to be properly housed, authorities had to resort to various types of emergency housing, sometimes even jails. Some cities simply moved the homeless into camps in rural areas; many transients themselves built improvised shanty-towns on the outskirts of cities.

During the Second World War, as the homeless were absorbed into the armed forces and war industries, the problem virtually disappeared. In certain quarters there were fears that economic stagnation would return by the end of the war, and with it unemployment and homelessness. However, the accumulated private savings, the rapid conversion of military industries into civilian production, and a pent-up demand for private goods and services led to a post-war economic boom that kept the rates of unemployment and homelessness very low.

In the following decades, Skid Row areas – with their collection of establishments and businesses catering to a poor and aging population – slowly contracted, and their role in a modern urbanized economy became progressively marginalized. Many social scientists believed that these areas would eventually disappear completely from the urban landscape.

In the late 1950s, academic interest in Skid Row areas and their inhabitants resumed as the renewal of city centres compelled many urban planners and municipal authorities to confront the problems of older and poorer neighbourhoods. The focus was on finding out as much as possible about Skid Row areas and how their populations would fare once the areas were demolished to make way for urban renewal.

During the late 1950s and 1960s, many attempts were made to estimate the size of the homeless population in various urban centres. Rossi mentions a 1963 study by Donald Bogue(4) which used 1958 census data to count some 12,000 homeless persons in Chicago alone; he estimated that approximately 200,000 homeless, mostly elderly men, lived in the 41 largest U.S. cities. Another study done in 1960 by Bahr and Caplow(5) listed about 8,000 homeless men living in New York’s Bowery, and 30,000 additional homeless persons living elsewhere in the city. Another homeless count(6) listed 2,000 homeless persons in Philadelphia.(7)

Despite the economic expansion of the period, homelessness persisted. As Rossi stated:

The Skid Rows may well have been dying out; indeed, given the advanced age of the residual Skid Row population, the impending demise of Skid Row was widely and confidently predicted. But it was obvious that the final death throes would be neither merciful nor swift.(8)

In the 1950s, the notion of homelessness differed considerably from the one used today. At that time, homelessness implied more the lack of support, whereas today the term implies deprivation of shelter. In fact, most of the homeless men in Bogue’s study had stable shelter of one sort or another. Four-fifths rented windowless cubicles in flophouse hotels on a daily (or nightly) basis. These cubicles were small, partitioned spaces sometimes just large enough to accommodate a cot. The remaining fifth lived in private rooms in inexpensive single-room-occupancy (SRO) hotels or in mission dormitories. According to Bogue, only a small minority of homeless men actually lived on the street. Other studies made the same observation.

The studies were also remarkably similar in their other findings. They described the Skid Row population as very homogeneous, consisting mostly of white men whose median age was 50. Of these men, one-quarter were Social Security pensioners stretching their meagre benefits by renting the cheapest accommodation available. Another one-quarter suffered from chronic alcoholism. The remaining half of the Skid Row population was composed of the physically disabled (20%), chronically mentally ill persons (20%), and those suffering from what was then called "social maladjustment" (10%).

With the exception of the pensioners, most Skid Row residents earned their living through menial low-paid employment. Those without work could find shelter and food in municipal shelters and mission dormitories, considered to be the least desirable forms of housing.

Finally, Bogue’s study noted the social isolation of Skid Row inhabitants. Virtually all of the men were unmarried and most had never been married. Though many had families, the study noted that the ties of kinship were very tenuous. Some claimed to have friends but these relationships were very superficial.

Other studies confirmed that the Skid Row population was characterized by:

  1. extreme poverty, arising from unemployment or sporadic employment, chronically low earnings and low levels of government assistance;

  2. disability, arising from advanced age, alcoholism, and physical or mental illness; and

  3. social disaffiliation, and tenuous ties to family and kin, with few and no friends.(9)

Many social scientists noted the disappearance of the casual labour market and the end of the economic function of Skid Row along with the progressive mechanization of low-skilled jobs. However, Skid Row areas did not disappear altogether, as many academics had predicted. In most cities, smaller versions of them grew up near areas with SRO hotels and rooming houses.

In the late 1960s and early 1970s, many U.S. cities underwent considerable physical transformation and renewal in which much of the older and cheaper housing stock was being torn down or converted into more profitable structures/uses such as office buildings and parking lots. At the same time, the U.S. federal government passed legislation improving the amount and coverage of social security benefits for the elderly and the physically and mentally handicapped. These expanded benefits, together with the provision of subsidized public housing, enabled some of the Skid Row population to move into more adequate housing.


In the 1980s and 1990s, important modifications to the urban economy, housing markets, public policy, and demographic trends acted to transform the nature and characteristics of homelessness as described in the preceding section of this paper. Rossi has identified the differences and similarities in the old and new homeless populations.(10)

The most striking difference is the visibility and number of the homeless today. No longer confined to poor neighbourhoods, people living and sleeping on the street or in public places such as bus terminals and railroad stations have become a commonplace sight in most cities. In the past, the "homeless" managed in one way or another to find nightly shelter; the new homeless clearly suffer more severely from housing deprivation.

Another difference is that today’s homeless population is on average considerably younger and more varied in terms of gender and ethnic representation. In Bogue’s 1958 study of the homeless in Chicago, women accounted for about 3% of Skid Row residents. In a 1985-1986 study carried out in the same city, women made up almost one-quarter of the homeless population, a statistic confirmed by other recent surveys.

The racial composition of the homeless has also changed considerably over time, from being predominantly white – 82% on Chicago’s old Skid Row – to being more racially mixed. Among the new homeless, racial and ethnic minorities are heavily represented. In one study of Chicago, 54% of the homeless were black. In most cities, other ethnic minorities – principally Hispanics and Native Americans – are also represented disproportionately among the homeless, with the precise ethnic mix apparently determined by the ethnic composition of the local poor.

In the past, most Skid Row inhabitants (other than pensioners) could find enough sporadic employment to give them enough income to secure some kind of accommodation. However, the new homeless are largely unemployed (97%, according to one study), and much poorer as a result.

In spite of all these differences between the old and new homeless populations, the similarities are even more striking. Both groups demonstrate extreme poverty, intermittent and unpredictable income (if any), inability to afford stable and adequate housing, a high percentage of mental and/or physical disabilities, and a higher likelihood of alcoholism and drug abuse.

Finally, Rossi makes an interesting comparison between the homeless and those who are extremely poor but domiciled. Although some estimates of the total number of homeless people barely exceed one million individuals, the number of the poor stands at around 35 million; thus, the number of poor exceeds the number of homeless by a very wide margin.

In attempting to explain this finding, Rossi goes on to define what he calls the extremely poor: those whose household annual income is equal to or less than the median income of poverty households (52% of the official U.S. poverty level). In 1980, the U.S. federal government defined the poverty line as $8,414 per year; the extremely poor family of four had an annual income of $4,396, or $3 per person per day. According to this criterion, the number of the extremely poor was around 17 million in 1980 – still far more than the number of the homeless. The question, according to Rossi, is not why there were so many homeless people, but rather why there were so few, compared to the number of the extremely poor, whose financial situation is so similar.

Using the 1984 survey of General Assistance (GA) recipients as a database, Rossi compares the extremely poor with the homeless. The GA is a welfare program for poor people who are not eligible for such support programs as Social Security, Assistance for Families with Dependent Children (AFDC), or any other disability program. GA recipients must be able-bodied, single adults whose annual income does not exceed $1,800 and who possess no significant assets. He notes that the GA clientele and the homeless were socio-economically very similar except that the overwhelming majority of GA beneficiaries (92%) lived in conventional housing and in households, usually with their parents. Of the third who lived alone, more than half were receiving financial support from friends or family.

Thus, the principal difference between the homeless and the vast majority of the extremely poor GA recipients is that the latter either live with family or can rely on financial support from friends or relatives.

The above results make it apparent that the network of kith and kin is the last line of defence against homelessness; the homeless, in turn, are apparently those for whom this network has been destroyed in a process no doubt strongly tied to disabilities such as alcoholism or mental illness.(11)


Of the many possible reasons for the recent rise in homelessness, the most often cited are: the decline of affordable housing stock; rent controls and other regulatory constraints on housing markets; the decline in the casual labour market; deinstitutionalization; and the crack cocaine epidemic.

   A. The Decline of Affordable Housing Stock

As Rossi stated in Without Shelter,

one must remember that homelessness is a housing problem. Homelessness on the scale seen today is in large part an outcome of the shortage of inexpensive housing for the poor, a shortage that began in the 1970s, and has accelerated in the 1980s.(12)

Using the U.S. Census Bureau’s Annual Housing Surveys (AHS), Rossi observed sharp declines in the stock of affordable housing in many urban centres. For example, between 1977 and 1981, the survey reported declines in the supply of affordable multi-room housing units (i.e., housing that rents for 40% or less of poverty-level incomes) ranging from a low of 12% (Baltimore, Maryland) to as high as 58% (Anaheim, California). In 12 large cities surveyed between 1978 and 1983, the supply of inexpensive rental housing available to low-income families dropped by an average 30%.(13)

The stock of single-room-occupancy (SRO) units ordinarily rented by low-income single adults declined even more precipitously. Chicago’s Planning Department reported the demolition or conversion of about 18,000 single-person dwelling units between 1973 and 1984. Seattle, Boston, New York, Nashville, Philadelphia and many other cities reported the same trend; Los Angeles reported losing more than 50% of its single-occupancy rental units between 1970 and 1985.

Rossi partly attributes the decline in affordable housing stock to funding cuts in federal housing programs that had supported the construction of public housing or provided housing subsidies to poor households during the 1980s. At the same time, the number of poor urban households increased by 36%.(14) The combined effects of these two trends resulted in a severe shortage of affordable rental housing for low-income households.

The diminishing stock at the lower end of the urban housing market limits the options available to low-income households; they must either allocate a greater portion of their income for housing, leaving less for other necessities, or leave the housing market altogether.

U.S. sociologist Christopher Jencks takes a slightly different view.(15) First, he notes the difficulty of identifying and counting SRO units. Although the term is typically used for older buildings divided into single rooms that do not meet a city’s current standards for new construction, the definition varies from city to city and over time. According to Jencks, in the simplest approach, the U.S. Census data can be used to trace changes in the supply of single rooms. The best survey on single rooms – the American Housing Survey (AHS), initiated in 1973 – suffers from three major limitations:

  • it does not survey many one-room units in any given year;

  • it does not cover tenants in hotels patronized mainly by transients staying there for at least six months; and

  • in 1985, it changed the way it counted rooms. From 1973 to 1983, the AHS let the tenants determine the number of rooms they had; after 1985, the AHS defined this number, thereby re-classifying one-quarter of the single-room population into two-room units.

In response to the AHS report that the number of SRO units declined nationwide between 1973 and 1989 from 1.1 million one-room rental units to about 789,000, Jencks notes that the number was more or less stable between 1973 and 1983 and between 1985 and 1989. He concludes that the decline observed between 1983 and 1985 was a statistical phantom resulting from problems in counting SROs and the 1985 changes to the AHS survey design.

Although the number of one-room units remained more or less stable over the sample period, according to the Census the population living in them declined appreciably, from 314,000 in 1973 to 162,000 in 1989.

Most SROs were being demolished in the 1960s and 1970s, but there was still a sufficient supply of these units for rents to remain stable. Supply of and demand for SROs stayed mostly in balance during the 1970s. Moreover, both real wages and government benefits continued to rise until the early 1970s, permitting even irregularly employed individuals to gain access to better-quality housing and lowering demand for lower-quality SROs.

Although real wages and government benefits to the poor stopped rising and the demand for the lower end of the housing market stopped declining after 1973, homelessness became visible only during the 1980s. Jencks attributes this lag to the steadily declining purchasing power of the poor over the same period. He argues that the shortage of accommodation for the poor resulted not just from an insufficient supply of cheap rooms, but also from excess demand driven by the rise of long-term male joblessness and lagging government benefits for the unemployed. The imbalance between low housing supply and high demand caused SRO rents to rise faster than the general level of prices.

One would have expected the excess demand for cheap accommodation to be temporary and to have been absorbed as rising rents and profits provided incentives for entrepreneurs to increase the supply. Jencks suggests, however, that the prevailing municipal housing regulations and policies (e.g., rent controls) in many major cities prevented adjustment to the new market conditions, i.e., by building new units or converting conventional housing into cheaper one-room-occupancy units.

Once their incomes began stagnating, a growing proportion of the poor could no longer afford their existing accommodation, yet they were being deprived of alternative shelter. The extremely poor either resorted in greater numbers to emergency shelters or were compelled to live on the streets.

In summary, Jencks does not believe that the destruction of Skid Row areas provides an adequate explanation of the rise in homelessness. He concludes from the statistical evidence that homelessness was the result of increasing poverty and regulatory constraints imposed on the low-cost housing market in the late 1970s and early 1980s.

   B. Rent Control and Other Regulatory Constraints on Housing Markets

Certain social researchers attribute homelessness to public policy initiatives and urban housing regulations. This is an extension of the argument in the section above that suggests homelessness results from the erosion of the low-cost rental housing stock.

Housing markets are not really national in scale but are rather the sum of many regional and local markets; thus, the municipal regulatory environment can have a tremendous impact on the kinds and availability of low-cost rental housing. These regulations can take the form of price ceilings on rental units (rent controls), restraints on the type of construction permitted (zoning regulations), or growth control ordinances that limit or prevent certain activities or businesses from being conducted in order to impose more "orderly" urban development.

      1. Rent Controls

Usually enacted during periods of rapid urban expansion when rental housing is scarce, rent controls are essentially ceilings on the price level or price increases of rental units. Rents are set below what the market would allow in the absence of controls.

Rent control has been in force in a number of North American cities for many decades. For example, the City of New York retains rent controls imposed under temporary wartime price rationing during World War II.

Many U.S. and Canadian cities adopted rent controls in the 1970s, in response to high inflation. In 1971, the Nixon administration imposed a nationwide system of wage and price controls, which was later repealed; however, many cities retained the controls on rents. By the mid-1980s, more than 200 separate municipalities in the U.S. – encompassing about 20% of the population – were living under rent control.(16) In Canada, a majority of provinces enacted rent controls at the request of the federal government in October 1975, as part of the federal Anti-Inflation Programme.(17)

Standard supply and demand theory predicts that any price ceilings, including rent controls, will produce an excess of demand over supply – in other words, an economic "shortage."(18) Although the stated goal of rent control is to make rental housing more affordable, its principal result is an overall shortage of reasonably priced rental units.

Because the mandated rents are set below the market rate, a wedge is driven between the quantity demanded by consumers and the quantity supplied by producers. Consumers will demand more rental units at the artificially cheaper mandated price. Entrepreneurs, on the other hand, will be less willing to supply rental units because of the low return on their investment. In the absence of alternatives, a portion of the demand will remain unsatisfied.(19)

This is the textbook way of describing rent controls, or price ceilings. In real life, however, it is very onerous for governments to control the entire supply of a commodity; controls or price ceilings are usually applied to a portion of any market. Depending on the degree to which the authorities are willing to tolerate it, the uncontrolled portion serves to absorb or mop up the excess demand resulting from the controlled market, and the uncontrolled market may be considered legal, semi-legal or illegal.

Unlike the rent for controlled units, which are set at a mandated level, the rents in the uncontrolled segment of the market are determined by the interaction of supply and demand. The unsatisfied demand from the controlled segment will spill over to the unregulated segment, pushing rent levels even higher as more tenants compete for the limited unregulated supply and creating a widening price gap between the controlled and the uncontrolled portion of the market. The result is a redistribution of income that benefits the tenants living in controlled units at the expense of tenants whose rents are uncontrolled.

Once controls are imposed, the tenants in regulated units have a strong incentive to remain there for as long as possible, even for a lifetime, perhaps in some cases passing on the tenancy to descendants. In this way, rent controls reduce tenant turnover and, given the lack of affordable alternatives, further restrict the supply of housing.

Some rent control regulations permit landlords to charge higher rents when the original occupants move out. In order to avoid forced eviction, rent control legislation has often included strong anti-eviction provisions. The inclusion of this provision, however, makes it virtually impossible to evict tenants of rent-controlled units, even if they are delinquent or destructive.

Thus, unable to evict delinquent tenants and seeing their revenues fall, landlords of rent-controlled buildings may curtail expenditure on maintenance and let the controlled units deteriorate. Rent controls thus remove any incentive for entrepreneurs to expand supply through building new units or maintaining the existing stock (e.g., rental units).

Although the effects of rent controls on the quantity and quality of the rental housing supply are well known and documented, the causal link between homelessness and rent controls is not as apparent. As Jencks points out: "Even if rent control really is correlated with homelessness, it does not follow that one causes the other."(20) For example, rent controls have been in place for decades in many North American cities, yet homelessness became more visible only during the 1980s and 1990s.

What can be affirmed, however, is that rent control has exacerbated the problem of homelessness by restricting the supply of affordable housing. Those who truly benefit from rent controls are the tenants who occupied the units at the time rent ceilings were imposed. The rent for such a unit is very affordable but obtaining a unit is next to impossible, because existing tenants have the incentive to maintain occupancy at all costs. Moreover, under a rent control regime, entrepreneurs have little or no incentive either to provide new housing or to maintain existing housing. Thus, poor individuals or households are virtually excluded from the private rental market; they must take more expensive accommodation they can barely afford, seek other forms of public shelter, or be forced to live on the streets.

      2. Other Regulatory Constraints

In the past, homeowners could rent out basements, attics, or spare bedrooms and in this way extend the supply of low-cost accommodation for individuals. Now, through restrictive municipal regulations, many cities prevent such practices and permit only high-priced single-family homes, thereby depriving consumers of cheap alternatives.

William Tucker argues that a community’s regulatory regime may be the most critical reason for the rise of homelessness and has a tremendous impact on the housing supply. According to him, "Housing shortages are local problems created by local regulation, which is the work of local municipal governments."(21)

In attempts to overcome the effects of rent controls, homeowners and landlords often resorted to demolishing or renovating their rental property or converting it for higher-value uses such as condominiums or co-operatives. In response, to prevent a substantial portion of the housing stock from being excluded from rent controls, local governments often enacted new provisions that prohibited or restrained these demolitions, renovations or conversions.

As Lawrence B. Smith has said:

The economic consequences of these prohibitions on demolition, conversion, and major renovations are to reduce the economic value of the rental units and accelerate the deterioration of buildings by reducing their value, by constraining and by reducing incentives to maintain their quality. Controls thus generate behaviour that induces additional regulations that may exacerbate many of the deleterious consequences of the controls.(22)

Building codes designed to drive out "undesirable housing" in local jurisdictions have also prevented or inhibited the building of low-cost housing.

Through regulation, most cities and towns hold a tight rein on their housing markets. Tucker points out that suburbs are particularly exclusionary, zoning out everything but high-priced single-family homes and prohibiting the rental of rooms or apartments.

In addition, efforts in the name of "urban renewal," and municipal campaigns to "clean up downtown," often led to the tearing down of Skid Row areas with their vast supply of tenements, single-room-occupancy (SRO) hotels, sub-standard shelters, and other cheap accommodation.(23) Unfortunately, these urban renewal efforts were not often accompanied by efforts to provide cheap alternatives to offset the loss of the rental housing stock.

   C. The Disappearance of the Casual Labour Market

A market for casual labour is important for unskilled persons who, for various reasons, cannot hold down a regular full-time job. For example, alcoholics or schizophrenics may work effectively only on an intermittent basis. The presence of a casual labour market thus enables marginally skilled persons to have some income and to pay for some type of shelter.

In the past, poor neighbourhoods played an integral role in the urban economy by providing local industries with a ready supply of unskilled labour. According to Rossi, a major factor in the decline of the Skid Row areas was the disappearance of the casual labour market. He cites a 1980 study by Barrett Lee, which analyzed the Skid Row populations in 41 U.S. cities between the 1950s and the 1970s. This demonstrated that, as the proportion of each city’s labour force employed in the unskilled and service occupations declined, so did the Skid Row population.

In the earlier decade of the analysis, urban employers [who] needed muscle power to wrestle with cargo apparently put up with the low productivity of Skid Row men because they could be hired as needed and at low wages. The advent of forklift tractors and other highly efficient materials-handling technology meant that casual labourers were no longer cost-effective; the declining demand for casual labour put the homeless and Skid Row out of business. The continuing lack of demand for unskilled labour still contributes to today’s homelessness. But there is another element involved, one that also helps us to understand the declining average age of homeless persons. The past decade has seen a bulge in the proportion of persons between the ages of twenty and thirty-five, a direct outcome of the post-war baby boom. The consequence of this "excess" of young persons, especially males, was a depressed earning level for young adults, and an elevated unemployment rate.(24)

Between the mid-1960s and the mid-1980s, as the earning profiles and employment opportunities for U.S. workers aged under 35 deteriorated, the numbers of the homeless increased and their average age declined. These developments in demographic and labour market trends had serious consequences for the formation of households and families. The observed rise in recent decades of single-parent households, particularly those headed by females, partly results from the deterioration of the economic prospects of young men. In these circumstances, young men are less willing or able to form households and take on the economic role of husband and father.

Jencks notes that, although there is no clear consensus on the root causes of the rise in long-term joblessness among mature men in the U.S., it is evident that the demand for unskilled labour fell faster than the supply during the 1970s and early 1980s. This had two consequences: unskilled workers’ wages fell, and the least desirable workers had trouble finding work at any wage. Both were factors in the increased risk of homelessness among mature men.

   D. Deinstitutionalization

One commonly held view is that the observed numbers of the homeless rose at the same time as psychiatric hospitals instituted a policy of "deinstitutionalization," in which mentally ill patients were released from long-term care.

Although Rossi agrees that the policy of deinstitutionalization may have contributed to homelessness, he points out that this initiative started as far back as the late 1940s; most patients destined to be released from mental hospitals had already left them when the rise of homelessness began in the early 1980s.

On the other hand, Christopher Jencks believes that deinstitutionalization, originally intended to improve the life of mentally ill patients, did have the unexpected effect of increasing the homeless population. He points out that deinstitutionalization should not be considered as one policy, but rather as a series of policies; although each had the goal of reducing the number of residential patients, carrying out these policies proceeded at varying rates and for different reasons. In a sense, the policy was intended to move patients from one type of institution to another, by moving them out of mental hospitals. The aim was originally to move patients from long-term incarceration to more community-based care.

The policy was based on the idea that long-term hospitalization is more harmful than helpful to mentally handicapped patients. It was thought that many of these people could be better cared for on an out-patient basis, while patients with histories of episodic mental illness could be admitted and then released after treatment.

The first wave of deinstitutionalization in the U.S. started in the late 1940s and was completed by 1965. By that time, the population of institutionalized patients had dropped from 513,000 to 475,000.(25) The policy gathered momentum in the 1950s, with the advent of new psychotropic drugs, which enabled the efficient treatment of patients suffering from depression and schizophrenia.

Deinstitutionalization continued during the mid-1960s and early 1970s, when the federal government voted improved public assistance benefits to former mental hospital patients in the form of Medicaid, food stamps and other types of income support. This provided former patients with enough financial means to support themselves, and helped poor families to provide shelter for – and take care of – their disturbed relatives. At this point in time, deinstitutionalization was still accompanied by certain provisions to take care of these individuals. The more seriously mentally ill remained in institutions, as did some patients who had nowhere else to go or whose repeated admission and discharge incurred too great an administrative or financial burden.

According to Jencks, the process of deinstitutionalization began to go awry after the mid-1970s, when social reformers, concerned about the civil rights of mental patients, succeeded in curtailing the ability of mental health professionals to commit severely disturbed patients to institutions. Although this may have improved the quality of life inside the mental hospitals, it reduced the quality of life outside them. Once they had lost the power to commit mental patients without their consent, psychiatric institutions began to release many seriously disturbed patients. Some of these were incapable of taking care of themselves, or soon alienated themselves from friends and families who might have given them support. Thus, many of these recently released and severely disturbed patients found themselves without any means of shelter or professional care.

Jencks argues that in society someone has to be held responsible for every individual’s actions. Usually this means that an individual is responsible for his or her own actions; however, if the individual is incapable of this, society has to intervene, if necessary by isolating that individual. According to Jencks, when families are unable or unwilling to take care of family members who are mentally incompetent and constitute a danger to themselves and others, mental hospitals should have the power to commit those members.

During the late 1970s and early 1980s, federal and state governments were encountering greater public resistance to tax increases; fiscal austerity became the order of the day. Governments were pressured to find ways to control public expenditures and urged mental hospitals to trim their budgets. As a consequence, some psychiatric wards were closed and many chronic mental patients, even those unwilling to leave, were discharged. According to Jencks, state governments could have made arrangements to provide their former wards with some form of shelter but they felt that endowing mental patients with rights included endowing them with responsibility to fend for themselves. Unfortunately, the chronically mentally ill are seldom able to assume this responsibility.

The prevailing political climate was sympathetic to the idea of deinstitutionalization, and especially to any savings that might result from it. In reality, these expected savings proved to be elusive. As Jencks points out, "Deinstitutionalization saves big money only when it is followed by gross neglect. That is why neglect became common during the 1980s."(26)

Operating mental institutions is very costly; they must have all the medical, support and administrative services needed to offer patients continuous treatment, as well as the staff and facilities to provide custodial care for those who are disruptive or dangerous. In short, mental hospitals provide three basic but very expensive services: subsistence, supervision and treatment. All attempts by state hospital planners to economize by curtailing one or more of these types of service proved unsuccessful or even counter-productive. In most cases, they merely shifted the financial burden from one institution to another, at considerable cost to society.

In the early 1980s, a U.S. government administration sympathetic to the notion of deinstitutionalization raised the criteria for receiving federal disability benefits and undertook a periodic review of eligibility rolls that purged some 300,000 persons judged to be capable of working. Of these, almost one-third (100,000 persons) were considered to be mentally ill. According to Jencks, few of the reclassified persons found work and presumably some became homeless. At the same time, however, states reduced the adult population in mental hospitals and cut their own welfare rolls by making their former wards eligible for U.S. federal assistance. As a result, by the mid-1980s, the federal government was compelled to reverse its own welfare policy and the disability rolls grew back to roughly the same levels as had prevailed in the 1980s. Indeed, the percentage of working-age adults receiving federal benefits for mental disability was higher at the end of the 1980s than ever before in U.S. history.

According to Jencks, if the policy of deinstitutionalization and the courts’ curtailment of the power of mental health professionals to commit patients had not occurred, the adult population of state mental hospitals would have stood at 234,000 in 1990, instead of at 92,000. It follows that 142,000 persons who under the pre-1975 rules would have been sheltered in mental institutions were now sleeping somewhere else.

On any given night, some of these people were in psychiatric wards of general hospitals, a few were in private psychiatric hospitals, but many were in shelters or on the streets.(27)

   E. The Crack Cocaine Epidemic

Before the mid-1980s, alcohol was the drug of choice for many of the poor, because the other available forms of oblivion were far more expensive. Alcoholism is still a significant problem among the homeless; however, Jencks does not think that it offers a satisfactory explanation of the recent rise in homelessness, because the proportion of alcohol abuse among the homeless population has remained more or less stable over time.

Introduced in the mid-1980s, crack cocaine was much cheaper than alcohol and other "hard" drugs and offered an intense but short "high." Its low price and easy availability soon made it the most popular drug of the homeless.

Jencks relies primarily on urine sample tests to estimate the frequency of crack cocaine use among the homeless. In 1991, New York City used a large sample of shelter users for voluntary anonymous urine tests. Of the sample of single adults using general-purpose shelters, 66% tested positive for cocaine use. Of adults living in family shelters, only 16% did so. Jencks then compares these results with New York crime statistics to infer the prevalence of use of cocaine across the country. In 1990, of men arrested in Manhattan, 65% tested positive for cocaine use. In a similar survey of men arrested in seven large U.S. cities in 1990, 49% did so.(28) Jencks concludes that the prevalence of cocaine use is about the same among single adults living in shelters as among arrested individuals. He suggests that if the result is valid for metropolitan centres, about half of all individuals using shelters in New York City during 1991 had used crack cocaine within a few days of being tested.

Nationwide, a reasonable guess might be that a third of all homeless single adults use crack fairly regularly. If so, crack is now as big a problem among the homeless as alcohol.(29)

Jencks believes that use of crack cocaine may explain why homelessness increased even at a time when the overall level of unemployment declined.

Although there is no unequivocal proof of its contribution to homelessness, drug dependency – whether it causes homelessness or is merely the result of it – certainly keeps the homeless on the streets. Increased drug use makes an unskilled worker even less employable, further reducing disposable income and eroding the ability to secure shelter. Drug dependency can also alienate friends and family who might otherwise provide support and shelter, making it more likely that the individual will become homeless.

Although information on how the homeless obtain and dispose of their income is very sketchy, Jencks assumes that many of them would rather spend it by seeking temporary oblivion through drugs or alcohol than on paying for accommodation in a dangerous shelter. He estimates that between one-third and two-thirds of the homeless population may engage in some form of substance abuse.


The Canadian experience with homelessness closely parallels that of the United Kingdom and the United States. In the nineteenth century, homelessness was usually regarded as the result of a failure of character on the part of the homeless individual. Aid to the destitute was dispensed mostly by private charitable or philanthropic organizations on an emergency or dire need basis and usually in the form of food, clothing, heating fuel (wood or coal) rather than cash. It was felt that if assistance was dispensed too freely it would create dependency and erode the self-sufficiency of the beneficiary, thereby aggravating the problem. There was virtually no role for the public sector in the distribution of aid, although during periods of stagnant economic activity in Canada, social reformers and public officials would rely on the relocation of itinerants and the unemployed to unsettled lands in the North and West of the country.

In the latter part of the nineteenth century, the combined effects of rapid industrialization, urbanization and immigration gave rise to greater interest in urban poverty and its underlying causes and possible cures. Voluntary philanthropic organizations on both sides of the Atlantic influenced each other and experimented with model homes for the deserving poor; this activity led to the establishment of sanitary codes and public health regulations to deal with overcrowded and unsanitary housing.

Social problems brought about by uncontrolled immigration to the United States and Canada led to proposals for settlement housing to facilitate the assimilation of immigrants. Social unrest resulting from economic depressions and deplorable housing and working conditions highlighted the need for reforms aimed at improving public health, housing, and working environments.

In Canada after the passage of the Municipal Corporations Act of 1849, local governments began to assume social welfare responsibilities. From 1870 to 1900 the provinces had a far greater role, principally in establishing prisons and asylums, but also regulating the work of publicly subsidised private charities. This system focused on urban areas, mainly older cores of major cities, where the very poor people and relief institutions were concentrated.(31)

From 1900 to 1930, continued population growth and rapid industrialization and urbanization led social reformers to ponder the linkages between industrial growth, urban expansion, housing, sanitation and public health. Social initiatives designed to improve the urban population’s living and working conditions were eventually implemented, and social benefits such as workers’ compensation were introduced in Quebec (1909) and Ontario (1914).

The years following the First World War were also marked by social unrest resulting from stagnant economic activity and rising unemployment. Residential construction had been put on hold during the war years and the housing stock had deteriorated. In response, the federal government instituted a modest housing program to stimulate economic activity and foster employment in the construction trades.

During the 1920s, public spending on welfare grew by 130% in response to public clamour to improve the conditions of the poor. However, little thought had been devoted to housing and public welfare reform.

When the Great Depression struck Canada, its severity and duration overwhelmed the country’s political leadership and social institutions. Public policy responses, if any, were most often uncoordinated, inadequate and ineffective. In the absence of any organized public welfare system, people in need were often left to fend for themselves or rely on private charities. During this period, Canada’s Gross National Product dropped by 42% and official unemployment levels peaked at 26.6% in 1933. Massive job losses occurred in shipping, manufacturing and railway transportation industries; as well, agricultural activity stagnated.

There was little response from the public sector as a result of the social attitudes towards public relief and the lack of adequate financial resources. Local and provincial governments were unable to provide much aid because their tax revenues had declined so precipitously that a number of them had defaulted on their debts. Moreover, the federal government urged local governments to reduce relief expenditures.

Continued economic stagnation contributed to social unrest and eventually culminated in the Regina Riots of 1935. In the same year, endemic unemployment, homelessness and the resulting social dislocation eventually prompted the federal government to pass the Dominion Housing Act authorizing $20 million in loans and financing the construction of 4,900 housing units over three years. This initiative, designed to stimulate employment and economic activity in the construction trades, was followed by the federal Home Improvement Plan (1937) and the National Housing Act (1938) which provided funds to assist people to purchase their homes and for the construction of low-rent housing.

During the Second World War, the war effort and the armed forces quickly absorbed the idle workforce. Unemployment and homelessness practically disappeared as war production surged ahead.

Early in 1944 the Canadian government appointed Emergency Shelter Administrators to monitor overcrowded housing markets: migration to these areas was restricted. Municipalities used federal funding to build temporary shelters for homeless people which were intended to keep families "warm and dry…[with] sufficient space and essential facilities so that a good standard of health and morale [was] maintained.(32)

During this period, government involvement in economic and social spheres became commonplace as a result of the centralized wartime planning of industrial production, food rationing and the rise of Keynesian economics in academic circles and among government policy-makers. One result was a gradual change of attitude towards government intervention in general which laid the foundation of the modern welfare state during the post-war period. At this time, many of the components of the Canadian welfare state were introduced: unemployment insurance (1941), family allowances (1945), Old Age Security (1952), and the Canadian Pension Plan (1966).

Although it contained a number of provisions directed to the working poor, the housing legislation passed in the late 1940s was primarily designed as a macro-economic stabilization tool to support employment and activity in the construction industries. Further, because housing assistance was mostly directed to the middle classes, it was also a means of gathering political support for the government.

In 1945, the federal government created the Central Mortgage and Housing Corporation or CMHC (renamed the Canada Mortgage and Housing Corporation in 1969). The National Housing Act was amended in 1949 to provide funding to support the construction of public housing for low-income families, disabled persons and senior citizens. In 1954, the federal government began insuring mortgage loans to facilitate home ownership. In 1964, CMHC was authorized to provide loans to municipal and private non-profit corporations. Starting in 1973, legislation introduced the notion of "income mix" to avoid large public housing ghettos. In more recent years, municipal governments have taken more initiatives to address the shelter and income security needs of low-income residents. Not until the mid-1980s, however, did municipal officials acknowledge the growing problem of homelessness.


Homelessness and poverty in Canada continue to parallel the experience of the United States. Unlike the U.S., however, Canada has only recently attempted to determine the scale and scope of the problem and to discuss what defines homelessness and to measure its extent.

How homelessness is defined is a critical factor in estimating the size of the homeless population. The narrower or more exclusive the definition, the smaller the estimate will be, and vice versa.(33) The United Nations uses a very broad definition that includes two types of individuals:

  • those who have no homes and who live outdoors or in emergency shelters or hostels; and

  • people whose homes do not meet basic UN standards of providing adequate protection from the elements, access to safe water and sanitation, affordable prices, secure tenure and personal safety, and access to employment, education and health care.(34)

Canada uses a very similar concept – "core housing need" – to measure the adequacy of housing. To establish whether a household is in "core housing need" is a two-step process. First, one must determine whether households suffer from a lack of one or more of the three basic housing needs: adequate repair, adequate plumbing, and adequate space. Second, it must be asked if the household has the financial wherewithal to resolve its housing problems (affordability). A household is defined as being in core need if it is unable to rent a unit of suitable size and with adequate conditions in its community without spending more than 30% of its income.

In 1985, the CMHC estimated the number of Canadian households experiencing one or more housing problems and the incidence of core housing need among these households. Of a total population of 8.75 million private households, 33.8% experienced housing problems of one sort or another. Of those households, 44% met CMHC’s criteria for core need. Of these, 28% owned accommodation and 61% rented accommodation.

Core need households do have shelter, although it is inadequate, and thus are not defined as homeless. Nevertheless, the homeless are most likely to come from the core need population. These households are in a precarious housing situation and constitute an "at risk" group; the slightest deterioration in income or family circumstances may push them towards homelessness. Thus, the size and circumstances of the core need household can indicate the potential scale of homelessness.

The population at the lowest end of the housing spectrum, i.e., people without a fixed address and with no shelter, are transient and elusive and thus hard to count. The first serious attempt at enumerating the homeless on a national basis was carried out on 22 January 1987 by the Canadian Council on Social Development (CCSD) with the objective of identifying the causes of homelessness and developing strategies to eliminate the problem.

Enumeration was carried out through a survey in which agencies that provided temporary or emergency shelters or services to the homeless were asked to fill out a questionnaire. Of 472 questionnaires distributed, 283 were completed and returned.(35) The results placed the nightly shelter capacity at 13,797 available places. The overall occupancy rate averaged 77%. In total, it was estimated that 259,384 individuals spent at least one night in a shelter in 1986 and the average length of stay was 19.4 days. The largest group using shelters were men residing in men-only hostels (61%), followed by women (27.5%) and children under the age of 15 (ll.5%). The CCSD estimated that the total homeless population in 1987 varied between 130,000 and 250,000 individuals, representing between 0.5% to 1.0% of the Canadian population.(36)

The survey’s methodology suggests that the total population of the homeless was probably underestimated. Only 283 of 472 agencies provided completed questionnaires, a response rate of only 66%. Moreover, four categories of persons were not covered:

(a) persons who slept in shelters that did not participate in the survey, including about half the emergency homes for battered women and children in Canada; (b) persons who that night were out on the streets, sleeping in abandoned buildings, stairwells, parking garages, and public buildings or doubled up with friends or acquaintances; (c) persons in detoxification centres, maternity homes, or other special-needs centres that frequently assist people who have nowhere to go; and (d) persons sent to hotels or motels by social-service providers or accommodated in local jails because no other shelter beds were available.(37)

Many criticized this methodology as inadequate or ineffective for measuring the homeless population.

A second attempt to count the number of homeless in the country was carried out by Statistics Canada in 1991. Using the 1991 Census, Statistics Canada used a survey strategy very similar to that used by CCSD in 1987 and that had been criticized as inadequate. Essentially, census enumerators based in 90 soup kitchens asked the clients where they had spent the previous night. Unlike the CCSD, Statistics Canada has never published its findings because of the mediocre quality of the data.

No further attempt has been made to enumerate the number of homeless individuals on a national basis. Given the lack of reliable and representative data, the federal government has instructed its housing agency, Canada Mortgage and Housing Corporation (CMHC), to make homelessness a research priority. The agency is currently developing computerized research tools that will standardize the collection and management of data for admission to services for the homeless.

Only a few provinces have assembled relatively dependable databases on the homeless. So far, Alberta, Manitoba and Ontario and their capital cities have taken the initiative in gathering information on the size of their homeless populations; unfortunately, every provincial and municipal survey suffers from the same problems as the two national surveys. However, all confirm the greater variation in terms of gender, age and ethnic composition of today’s homeless population, compared with that of earlier times.


Homelessness remains a persistent phenomenon, but its characteristics have changed considerably over the years. Controversies continue regarding how the homeless should be defined and their numbers, but the changing composition of the homeless population is not in doubt. The most striking feature about current homelessness is its divorce from the urban economy and society.

In the past, people without shelter residing in the Skid Row areas of the city were rare in the urban landscape. These areas, with their populations of predominantly white and elderly men, actually formed an integral part of the urban economy, supplying a pool of cheap unskilled labour. With the mechanization and computerization of productive processes, however, Skid Row progressively lost this economic role. Deprived of even low-paying job opportunities, the inhabitants of Skid Row found themselves in a progressively precarious situation with respect to stable shelter.

With the anticipated decline in the elderly population, improvements in government entitlements, provision of public housing, and the continuing economic boom, many social scientists expected to see the gradual decline and eventual disappearance of homelessness. Circumstances have clouded this rosy scenario, however. The dearth of affordable housing, the poorly implemented policy of releasing mental patients, and the crack cocaine epidemic have all contributed to changing the nature and prevalence of urban homelessness, but it still exists. Indeed, it has become more visible and widespread and a more severe possibility for a wider range of individuals.

Although homelessness is essentially a problem linked to poverty among individuals or families, the multiplicity of factors involved make it difficult to overcome. No single strategy will be successful. The search for a solution must address issues as diverse as the supply of affordable housing, opportunities for employment, income distribution, mental and physical health, drug addiction, crime prevention, and law enforcement.

(1) Peter H. Rossi, Without Shelter: Homelessness in the 1980s, New York: Priority Press Publications, 1989.

(2) Michael Parkin, Economics, Addison-Wesley, 1980, p. 573.

(3) Rossi (1989), p. 7.

(4) Donald Bogue, Skid Row in American Cities, Chicago: Community and Family Studies Centre, University of Chicago, 1963.

(5) Howard Bahr and Theodore Caplow, Old Men Drunk and Sober, New York: New York University Press, 1974.

(6) Leonard Blumberg, Thomas Shipley, and Irving Shandler, Skid Row and its Alternatives, Philadelphia: Temple University Press, 1973.

(7) Given the methodological difficulties and opposing viewpoints involved, it is a very difficult and often controversial exercise to establish an accurate count of the homeless population (see Lyne Casavant’s "Counting the Homeless," Module on Homelessness, PRB 99-1E, Parliamentary Research Branch, 8 January 1999). Nevertheless, it is interesting to contrast the above studies with more recent attempts at enumerating the nationwide homeless population. A 1984 report put the U.S. homeless population at between 250,000 and 300,000 (U.S. Department of Housing and Urban Development, "A Report to the Secretary on the Homeless and Emergency Shelters," Washington, D.C., H.U.D., 1984). A 1986 study set the nationwide count at 350,000 persons (Richard Freeman and Brian Hall, "Permanent Homelessness in America," Cambridge, Massachusetts: National Bureau of Economic Research, August 1986). At the other end of the spectrum of population estimates, the National Coalition for the Homeless puts the figure somewhere between 1.5 to 3 million individuals.

(8) Ibid., p. 9.

(9) Ibid., p. 11.

(10) Ibid., pp. 13-29.

(11) Ibid., p. 29.

(12) Ibid., p. 31.

(13) Ibid.

(14) Households whose income is equal to or below the poverty level.

(15) Christopher Jencks, The Homeless, Boston: Harvard University Press, 1994.

(16) W. Tucker, "How Rent Control Drives out Affordable Housing," Cato Policy Analysis, No. 274, May 1997.

(17) Lawrence B. Smith, Anatomy of a Crisis: Canadian Housing Policy in the Seventies, Fraser Institute, 1977.

(18) Ibid.

(19) James D. Gwartney and Richard L. Stroup, Economics: Private and Public Choice, 6th ed., Harcourt Brace Jovanovich, 1992, pp. 72-73.

(20) Jencks (1994).

(21) W. Tucker, "How Regulations Cause Homelessness," Public Interest, Winter 1991, Issue 102, p. 78.

(22) Lawrence B. Smith, "Ontario Housing Policy: the Unlearned Lessons," in Home Remedies: Rethinking Canadian Housing Policy, C.D. Howe Institute, 1995.

(23) Tucker (1991), p. 78.

(24) Rossi (1989), p. 35.

(25) Jencks (1994), p. 20.

(26) Jencks (1994), p. 34.

(27) Ibid., p. 39.

(28) The cities are: Los Angeles, Chicago, Houston, Philadelphia, San Diego, Detroit and Dallas.

(29) Jencks (1994), p. 43.

(30) Gerard Daly, Homeless: Policies, Strategies, and Lives on the Street, Routledge, 1996, pp. 51-88.

(31) Ibid., p. 57.

(32) Ibid., p. 77.

(33) Lyne Casavant, "Definition of Homelessness" Module on Homelessness, PRB 99-1E, Ottawa: Parliamentary Research Branch, January 1999.

(34) Alex Murray, "Homelessness: The People" in Fallis and Murray, ed., Housing the Homeless and the Poor: New Partnerships among the Private, Public, and Third Sector, University of Toronto Press, 1990.

(35) Lyne Casavant, "Counting the Homeless," Module on Homelessness, PRB 99-1E, Ottawa: Parliamentary Research Branch, 8 January 1999.

(36) Murray (1990).

(37) Ibid. (1990), p. 21.