Defining and Re-Defining Poverty: A CCSD Perspective
This position paper briefly presents the Canadian Council on Social Development's perspective on poverty lines, with recommendations to Statistics Canada and the federal, provincial and territorial governments. The CCSD is a leading social research and advocacy organization with particular expertise in the study of poverty.
Multiple Measures of Low Income and Poverty
Unlike the United States and some other countries, Canada has no official, government-mandated poverty line, and Statistics Canada publicly cautions that its Low Income Cut-offs (LICOs) should not be seen as such. Nonetheless, the Canadian Council on Social Development (CCSD), the National Council of Welfare, most anti-poverty groups and, in practice, even government researchers have used the LICOs as semi-official poverty lines since the 1960s. That is because reliable estimates of rates of low income based on the LICO lines are released annually by Statistics Canada, along with the release of annual data on the incomes of Canadians.
That is about to change. Human Resources Development Canada (HRDC) has, through a federal/provincial/territorial working group, proposed a Market Basket Measure of poverty. (MBM), and Statistics Canada is currently examining methodological issues behind the implementation of such a basket. The expectation of governments is that MBM-based estimates of low income will be released with the annual Statistics Canada income data in the near future, perhaps in 2002 with the release of income data for the year 2000. Statistics Canada also intends to consult Canadians on the methodology of the MBM in the near future.
Statistics Canada officials have said they will continue to publish low-income rates based on the LICOs, and they may also continue to publish relative low-income rates based on the Low Income Measure (LIM), set at one-half of median income. LIMs are the usual basis for international comparisons of such things as child poverty, although they are used much less frequently in Canada than are the LICOs.
Because we are about to enter a world of multiple poverty measures, it is appropriate for the CCSD to discuss some of the issues involved. This position paper sets out some of the pros and cons of the different lines and outlines some of the important issues. Most of the substantive issues are discussed at much greater length in the CCSD's Canadian Fact Book on Poverty 2000 (Chapter 2). The paper, Low-income Trends in the 1990s, available on the CCSD's website, summarizes the most recent data.
What are we measuring?
If we are going to talk about persons living in poverty, we have to define who is poor and who is not. Defining a poverty line is necessary if we are going to analyze not only the incidence of poverty, but also the depth of poverty (how poor are the poor?) and the duration of poverty (how long are the poor, poor?). And the analysis must, of course, be done for different family types, persons of different ages, and the like. The CCSD has long highlighted the depth of poverty in its research, and new data from the Survey of Labour and Income Dynamics (SLID) have made detailed analyses of the duration of poverty possible.
Drawing any poverty line is arbitrary to some degree, even if there is conceptual agreement on what poverty means. But in this case, even conceptual agreement is lacking. Popular and expert opinions apparently clash on whether and to what extent poverty should be defined in absolute terms - inability to meet very basic needs - or in relative terms - distance from the community norm. In fact, upon closer examination, virtually all measures of low income are relative (except the most extreme), and they differ mainly in their underlying assumptions about how wide a gap should exist between "the poor" and fellow citizens within the same society. This gap can be defined in terms of income or in terms of consumption.
One example of an absolute poverty line is the World Bank's extreme poverty line of $1 U.S. per day which is applied to developing countries as a proxy for a minimum survival budget. It is basically irrelevant to understanding poverty in advanced, industrialized countries, since virtually all their citizens have incomes above that minimal level. The official U.S. poverty line is a somewhat higher survival budget, based on the costs of a very basic food basket with an arbitrary percentage added on, the total of which has fallen further and further behind average U.S. living standards. The U.S. line is generally regarded as a very bad measure by social advocates - who see it is as far too low compared to both actual living costs and average budgets - and by experts because of its lack of clear links to actual living costs today.
In Canada, the tradition of defining poverty in terms of very basic needs is most closely associated with the Fraser Institute, which argues that no one is poor if they can meet their "basic needs." This line attempts to calculate the cost of a basket of necessities including food, clothing, shelter, and some limited additional items. Notably absent from the Fraser Institute's basket are items which the great majority of Canadians take for granted, such as coffee, a daily newspaper, and cable TV. There is also no allowance for access to recreation or culture. This definition of poverty has been widely used to discredit the LICOs, which are based, in part, on the incomes of the poor compared to others in society.
So-called "absolute" poverty lines do, in fact, embody community norms and standards, at least to a limited degree. Even the extremely bare-bones poverty line of the Fraser Institute would look "generous" to many citizens of very low-income developing countries, or to average Canadians of a century ago. Indeed, the Fraser Institute itself acknowledges that "the basic needs poverty line is not absolute…to be meaningful, a poverty line has to be connected to the society in which people live." (Measuring Poverty in Canada, 2001)
Social policy advocates, including the CCSD, have strongly defended a relative definition of poverty, arguing that to be poor is to be distant from the mainstream of society and to be excluded from the resources, opportunities and sources of subjective and objective well-being which are readily available to others. Adam Smith himself closely embraced a relative definition of poverty, arguing that to be poor was to have to go without what was needed to be a "creditable" member of society. He noted that, in his day and age, a man needed a linen shirt if he was to appear in public "without shame." What defines being poor is not just being without a shirt (any shirt), but being without the means to be minimally respected in a particular society at a particular time.
Similarly, the Organisation for Economic Co-operation and Development (OECD) has argued that absolute poverty lines have little meaning in advanced industrialized societies, and that poverty should be seen conceptually not as deprivation of very basic needs, but as exclusion from the standards of living broadly available to others in the same society. As they note, "in order to participate fully in the social life of a community, individuals may need a level of resources that is not too inferior to the norm of that community." They add that, "from a normative perspective, it may be considered unfair for members of a community to benefit unequally from a general increase in prosperity."(OECD Employment Outlook, June 2001, p. 41)
It is very difficult to argue that relative well-being does not matter to people. How many of us would prefer to have an income of $30,000 in a society where the average was $50,000, compared to having an income of $29,000 in a society where the average was $25,000? It is simply untrue to say that people are indifferent to where they sit in the total income distribution compared to others, even provided their basic needs are met.
Research also demonstrates that many key social outcomes such as health and literacy are closely linked to the relative incomes of people, rather than the extent to which they suffer absolute deprivation. The Nobel-prize-winning economist, Amartya Sen, uses as a telling example the fact that blacks in the United States have much lower life expectancy than do citizens of the state of Kerala in India, despite having much higher absolute incomes. In advanced industrialized countries, the link between health outcomes and average GDP per person breaks down beyond a threshold level, and it is most closely linked to income gaps.
It is often argued - quite falsely - that little or nothing can be done to lower relative poverty. In fact, relative poverty rates differ a great deal among countries. Measured by the LIM (a poverty line using one-half of median income), poverty in the mid-1990s was almost 11% in Canada, compared to less than 5% in the Scandinavian countries and 16% in the United States. Relative poverty in Canada could be significantly lowered through changes to the tax and transfer system, increases to low wages, and higher rates of employment.
The central point is that virtually all measures of poverty are relative. Whether we define poverty in terms of a given level of income compared to the average (an income line) or in terms of the cost of a basket of goods and services compared to the average, we still have to settle on how great a distance we want to exist between "the poor" and others who live in the same society. This decision is explicit in relative income lines such as the LIM, but more implicit in the consumption-based approach.
All measures of poverty, whether they are based on income or a basket of goods and services, are also arbitrary, at least to some degree. It is really a matter of values how great a distance we are prepared to accept between the "poor" and the rest of society. The CCSD and most social welfare advocates support a relatively generous poverty line, because we recognize needs as social as well as physical. To be poor is to experience a significant degree of exclusion from the wider society, and not just to be deprived of very basic needs.
The Pros and Cons of Three Poverty Lines
Statistics Canada regularly publishes data on the proportion of Canadians falling below the pre-tax and post-tax Low Income Cut-offs (LICOs), which define a low-income household as one which spends a much higher percentage of its income than an average equivalent household on the necessities of life - food, shelter and clothing - and thus has much lower absolute and relative "discretionary" income than the norm.
The LICO line is calculated by adding 20 percentage points to the percentage of income spent by an average equivalent household on food, clothing and shelter. LICOs are calculated for different sized families and communities, and re-based periodically to take into account changes in household spending patterns. Currently, the average household spends about 35% of its pre-tax income on these three necessities, so a low-income household is defined as one which spends more than about 55% on necessities.
The post-tax LICO now highlighted by Statistics Canada is based on spending on necessities relative to after-tax income, rather than before-tax income. In 1998, the pre-tax LICO for a family of four in a large urban centre was $31,107 ($27,890 after taxes), and for a family of two, it was $20,603 ($17,705 after taxes). While the trend of the two LICO measures over time is basically the same, the proportion of low-income households is lower for the post-tax measure by about four percentage points. That is because the gap between pre- and post-tax income grows as one moves up the income scale due to the progressive nature of the income tax system. The CCSD has expressed concerns that this shift to post-tax LICOs may be falsely interpreted as a drop in poverty. As it now stands, the post-tax LICO is not adjusted for payroll taxes, but this is being reconsidered by Statistics Canada and has been recommended by the CCSD.
The CCSD has extensively used pre-tax LICOs as a poverty line in analytical works such as The Canadian Fact Book on Poverty because of their wide availability and because they are a meaningful measure of trends in the incidence, depth and duration of low income by family type and other characteristics. However, the CCSD has never viewed LICOs as the only valid poverty line. In fact, the Council has used LIMs, drawn attention to the living budget estimates of social planning councils, and in 1973, even developed its own poverty lines (a close approximation of LIM).
- LICOs provide a consistent time series, and many groups are familiar with them.
- LICOs are effectively both a needs measure (income compared to basic needs) and a relative income measure (since the LICOs reflect differences relative to spending by the average household and they are periodically revised to take into account spending patterns by the average family).
- Survey evidence based on Gallup polls indicates that over the period 1976 to 1997, pre-tax LICOs have consistently been close to what the public itself judges to be an adequate income. Eight surveys over this period show almost no difference between the LICO line and what respondents believe "is the least amount of money a family of four needs each week to get along in this community." Opinion polls show that the public tends to strongly favour a social inclusion approach to the definition of poverty. (Canadian Fact Book on Poverty, p.31).
- There is some evidence, as highlighted in the CCSD publication Income and Child Well-being, that the LICO line provides a meaningful approximate break-point in terms of child outcomes, such that children below the LICO line do significantly worse than those above the line. (That being said, this study confirms the existence of a gradient of child outcomes by family income, showing that children well above the LICO do better than those close to but slightly above the LICO line.)
- The conceptual basis of LICOs is mixed and thus, somewhat hard to explain.
- The three areas of expenditure on which LICOs are based are the most basic, but hardly exhaustive.
- LICOs are sound in principle, but they have been undermined by the Fraser Institute and others who want to abolish poverty by definition, rather than by increasing the incomes of the poor. Talking about poverty as measured by LICOs is often questioned by the media, distracting attention from what is really important, such as trends in poverty in terms of meeting needs, and trends in income inequality.
- LICOs are not set at the provincial or city level. While geographical variations in household spending were found to be quite minor in the past, LICOs have been criticized recently for failing to take into account large differences in housing cost between major urban centres. Given that rent is a major proportion of people's living costs, LICOs likely understate poverty in high-rent cities such as Toronto and Vancouver.
LICOs are familiar, widely used and provide a consistent and detailed time series. The CCSD believes that Statistics Canada should continue to calculate and publish both before-tax and after-tax incidence of low income based on the LICOs, and they should continue to periodically re-base the LICOs to changing family expenditure patterns. Post-tax LICOs should be adjusted for payroll taxes.
The Low Income Measure (LIM)
LIM is a purely relative poverty measure which is used in many international comparisons. It explicitly defines low income as being much worse off than average, and it is drawn at one-half the median income of an equivalent household.
The CCSD has done some analytical work using LIMs and historically, has put forward a definition of poverty based on a slightly different LIM measure than that calculated by Statistics Canada.
- LIM is simple to understand and captures the important inequality dimension of low income.
- LIM facilitates international comparisons and makes clear how Canada is doing in comparison to other countries.
- LIM does not tell us directly if the poor have sufficient income to meet their basic needs - however they are defined - but it does tell us that they are significantly distant from the social norm in terms of income available to buy goods and services in the market economy.
- It is conceptually unclear what kinds of regional differences should be allowed for in LIMs. Does being poor in Newfoundland mean that a family has less than half the median income of a similar household in Newfoundland, or in all of Canada?
- LIM tells us nothing directly about what a low-income household can buy.
- LIMs do not vary with the business cycle in the same way as LICOs. Poverty using the LIM tends not to increase in recessions, even though income falls for the working poor. That is because unemployment reduces median incomes and not just the incomes of low-income households.
Statistics Canada should publish LIM-based measures of low income to supplement other measures.
The Market Basket Measure (MBM)
Starting in 2002, Statistics Canada could provide data based on a new line called for by federal, provincial and territorial governments - the Market Basket Measure of poverty or MBM. This measure attempts to calculate the amount of income needed by a given household to meet its needs, defined not just in bare subsistence terms, but also in terms of what is supposedly needed to approach "creditable" community norms. Human Resources Development Canada views the MBM as falling somewhere between a subsistence standard of living and a more generous social inclusion basket. They note that "the MBM approach is based on the lack of necessities with the key being the interpretation of the word 'creditable' in today's context...it does not imply a standard of living as comfortable as that implied by the concept of social inclusion." (Human Resources Development Canada, Applied Research Bulletin, Summer-Fall, 1998). The basket which makes up the MBM has been defined by federal and provincial officials. Statistics Canada has not determined the components of the basket.
The MBM will be calculated for a four-person family, then adjusted for other family sizes. It will directly include: rental costs based on an average of rents for two- and three-bedroom apartments; food costs sufficient to ensure a nutritious diet; clothing and footwear costs; and transportation costs calculated as a transit pass in larger urban areas or auto-operating costs for those living elsewhere. There will also be an allowance for other items, such as recreation and entertainment, based on the spending of low-income families - defined as those at the 2nd decile, that is, 20% of families having less income. The MBM line will be compared to an MBM disposable income defined as after-tax income, minus alimony payments, child care expenses, payroll taxes, and health expenses.
A discussion paper on the methodology behind the MBM is expected from Statistics Canada in the fall of 2001. Until we see this paper, the exact design of the MBM is unclear. The current intention is to provide low-income estimates based on the MBM at the provincial level and by size of city, if possible. Thus, there will effectively be poverty lines for individual, large urban centres. (LICO-based measures for individual large urban centres may also be released in the future).
The MBM is controversial in that it involves subjective judgements - mainly by Human Resources Development Canada officials, to this point - on what to include in the basket, and it is less generous than some alternative, needs-based baskets such as those developed by social planning councils. Preliminary estimates show that the MBM will significantly lower the low-income rate compared to the traditional pre-tax LICO measure.
The MBM should not be rejected out of hand, just because it is based on costing a basket of goods and services. Many anti-poverty groups, such as the Social Planning Councils of Toronto and Winnipeg and the Social Planning and Research Council of BC, have followed the same approach of attempting to calculate needs-based budgets. These attempts to measure household budget needs have highlighted the grossly inadequate level of social assistance benefits and minimum wages when compared to the costs of rent, food, clothing, transportation, and so on.
The basic issue raised by the MBM is what to include and what not to include in the market basket. The MBM is not an absolute measure of poverty, since it is explicitly intended to provide a "creditable" standard of living and not just bare-bones survival. At the same time, the MBM is designed to fall short of a basket sufficient for social inclusion, such as those advanced by social planning councils.
Other key issues include the process for updating and revising the basket to take account of changing spending patterns. If the basket in the MBM is not periodically revised, it will tend to fall further and further behind consumption patterns of the mainstream in society. Technical issues such as the equivalence scale - that is, how to adjust the basket for families of different sizes - will also have to be resolved.
- The MBM may resonate with the public and with government officials, thereby drawing fresh attention to poverty issues.
- An "adequate" needs budget would highlight grossly inadequate social assistance benefits and minimum wages.
- The MBM could make more apparent the large gap in living standards between the poor and the rest of society.
- The MBM is not based on an adequate conceptual premise of social inclusion.
- The MBM could distract attention from relative poverty and income inequality.
- By artificially lowering the poverty rate compared to pre-tax LICOs, the shift to MBM may falsely suggest that poverty is falling.
- If the MBM is not periodically re-based to take into account the changing consumption patterns of the more affluent, it could quickly become very outdated and very low compared to median income, as has occurred with the U.S. poverty measure.
- To date, the MBM has been developed by government officials with little or no input from poor people or community-based organizations about what should be included.
The federal and provincial governments should consult widely on what items are to be included in the MBM, the distance that should be allowed for between low-income families and community norms, and how the basket should be updated. Community-based organisations and persons living in poverty should also be involved in helping to define items in the MBM basket and MBM lines at the level of specific communities. The process for determining the MBM should be made much more participatory and transparent. Consideration should also be given to providing a more broadly based body with authority on the MBM measurement. Finally, the MBM should explicitly include items which are central to social inclusion, such as participation in recreation and cultural activities.