A New Poverty Line

A NEW POVERTY LINE
YES, NO OR MAYBE?

I. SOME BASIC FACTS ABOUT POVERTY LINES1

Why Have Poverty Lines At All? Statistics Canada's Low Income Cut-offs The Search for Alternatives

Spending Habits in 1996 Differences in Spending by Income Class Spending Patterns Across Canada Other Sources of Information

III. SOME EXAMPLES OF MARKET BASKETS

IV. WHAT WILL HAPPEN IF WE CHANGE THE POVERTY LINES?

V. OTHER ISSUES RELATED TO POVERTY LINES

PARTING THOUGHTS

ENDNOTES

QUESTIONS FOR READERS: TELL US WHAT YOU THINK

AN INVITATION TO READERS

This discussion paper from the National Council of Welfare is designed to inform readers about some of the different ways poverty could be measured in Canada and also to give them a chance to pass along their views about poverty and poverty lines to government. There is a questionnaire at the end of the report that makes it easy to tell us - or the federal Minister of Human Resources Development or Statistics Canada or your own elected representatives - what you think.

The paper arises in part because of sharp and continuing differences of opinion about the meaning of the "low income cut-offs" or LICOs of Statistics Canada and in part because of an initiative begun by the federal, provincial and territorial governments last year to develop alternative poverty lines based on the cost of a "market basket" of goods and services.

The National Council of Welfare and most other social policy researchers in Canada have used the low income cut-offs for many years. We consider them to be a reasonable measure of poverty, but not the only reasonable measure. We are aware that some of our colleagues are very strong supporters of the LICOs and are very suspicious about the motives behind the decision by governments to develop market basket measures of poverty.

We believe it is worthwhile having a full and open debate about the way we measure poverty. It remains to be seen, however, whether market basket poverty lines will eventually emerge as a useful alternative or as a complement to the LICOs. The uncertainty is reflected in the title of the paper, A New Poverty Line: Yes, No, or Maybe?

The discussion paper is by no means the definitive word on the subject of poverty lines. It focuses on different ways of developing market basket lines, because it is governments that are promoting that particular approach. It does not deal with a number of other possible approaches. Neither does it deal with related and very important concerns about the unequal distribution of income and wealth in Canada.

The Council has already participated in several discussions on poverty lines in recent months with colleagues in the social policy field. The director of the Council attended a workshop on poverty lines sponsored by Human Resources Development Canada, another workshop held in Vancouver by the income security and labour market committee of the Social Planning and Research Council of British Columbia, and has discussed market basket lines with the group which developed "Acceptable Living Level" measures in Winnipeg. The work in Winnipeg was done by the Social Planning Council of Winnipeg, the Winnipeg Harvest food bank and seven low-income people working as consultants.

The National Council of Welfare sponsored its own roundtable on market basket poverty lines in Hull, Quebec, on January 30, 1999. Guests at the meeting included members of anti-poverty groups from across the country, people from non-profit agencies who have developed market baskets of their own and officials of the federal government.

Comments about poverty and poverty lines by the people who attended these meetings has helped us greatly in preparing the final version of this discussion paper. We hope that the people who read the paper will take a few minutes to tell us what they think about the way poverty is measured in Canada. The final pages of the paper contain a number of questions about poverty lines and additional space for readers to add comments of their own. We will publish the results of the survey later in the year.

I. SOME BASIC FACTS ABOUT POVERTY LINES

The debate in Canada on poverty lines goes back at least a decade and has produced much more heat than light. Most of the attacks on Statistics Canada's "low income cut-offs" or LICOs are ideologically motivated and ignore the four most basic facts about poverty and poverty lines:

! All poverty lines are relative.

! All poverty lines are arbitrary.

! Poverty lines are a research tool for measuring the incomes of groups of people, not a measure of individual need.

! Some poverty lines are better than others, but none of them is perfect.

All poverty lines are relative. Our view of poverty in Canada is profoundly different than it was a century ago, and it is profoundly different from current views of poverty in Greece, Peru, Nepal, Somalia and most other countries of the world.

A century ago, people in Canada would have considered someone poor if they did not have a horse or some other beast of burden for their work and basic transportation. Today, poverty is more likely to be measured in terms of reasonable access to buses or subways in cities or access to an automobile or truck in rural areas.

Poverty is also relative in the sense that every culture has its own ideas about what is needed to be part of the mainstream of society. Being able to afford meat products is considered necessary in many western cultures, but it would not be considered very important in a country where most people are vegetarians because of their religious beliefs. Warm winter clothing and central home heating are vital in a country like Canada, and anything but a necessity in Honduras or Tahiti.

All poverty lines are arbitrary. Some poverty lines are drawn statistically and some are drawn by measuring the cost of a "market basket" of essential goods and services. Both types of lines are equally arbitrary.

In lines which are drawn statistically, someone has to decide what raw data to use and what methodology to use to get from the data to the poverty lines. One measure that is used in some international studies of poverty is one-half of median income or the half-way point for family income in each country. The decisions to use one-half rather than one-third of median income and to use median income rather than average income are both arbitrary.

In lines which are drawn from market baskets of goods and services, someone has to decide what goes in the basket and what does not. That is not nearly as easy as it sounds. How many rolls of toilet paper would a family of four reasonably use every week? How much clothing, new or used, would that same family need to buy in a typical month? What is a reasonable budget for school supplies over the course of a year? The answers to these and a host of other questions determine the overall cost of the market basket.

Poverty lines are a research tool for measuring the incomes of groups of people, not a measure of individual need. The primary limitation of both statistical poverty lines and market baskets of goods and services is that they cannot cover all the exceptions to the rule. Poverty lines work best when they are used to measure the incomes of large groups of people. They work worst when people try to apply them to individuals.

Take the cost of housing, for example. Some people have mortgages and others do not. Some people live in government-subsidized housing. Some people live in well-insulated apartments where the rent includes heating, while others rent houses which are poorly insulated and have to pay sizable heating bills on their own. A couple with two young children might be able to manage with two bedrooms, but a couple with a teenage boy and a teenage girl would need three bedrooms.

The people who draw poverty lines have to make compromises to keep the number of lines down to a reasonable total. The idea is to have poverty lines that cover the most common types of families in different parts of the country. All poverty lines are based on the number of people in the family, and some of them also take account of the size of the community where they live. Most poverty lines do not vary with the type of housing or the age of the people in the household.

There will always be individual cases where family needs are not well covered by any given poverty line. That is the reason welfare programs use a "needs test" for each individual case. With a needs test, all the family's needs and all its sources of incomes are considered. Welfare entitlements can be increased to cover the added cost of prescription drugs for a child with asthma or additional transportation costs in a very remote area or the cost of child care to allow parents to work full-time in the paid labour force.

Some poverty lines are better than others, but none of them is perfect. No matter how hard the experts try, they cannot escape the first three facts about poverty lines. Poverty lines will always be relative, poverty lines will always be arbitrary, and they will always be suitable for research and unsuitable as a measure of individual well-being. The most we can hope for is a useful research tool that is easily understood and widely accepted.

Some poverty lines are better than others at measuring changes in poverty over time. That is a very desirable feature, because it allows researchers to see the impact of changes in economic conditions or changes in government policy. Statistics Canada's low income cut-offs are very good in reflecting the ups and downs of the economic cycle. However, poverty lines based on average or median income tend to change very little from year to year and are much less useful in measuring poverty over time.

Some poverty lines are better than others because they focus on disposable income or cash on hand. The low income cut-offs are based on income after government transfer payments such as welfare, unemployment insurance or the Canada Child Tax Benefit but before income taxes. Most market basket poverty lines are after-tax measures that reflect the cost of sales taxes as well as income taxes. That is because people actually purchase the items in the market basket, and they have to pay any federal, provincial or territorial sales taxes on taxable items when they make their purchases.

Some poverty lines are better than others because they are easily understood and therefore easily accepted. One of the problems that has dogged the low income cut-offs is the fact that the methodology used to derive the lines is not easily explained. Lines based on a market basket of goods and services are much more straight-forward and much more likely to be accepted as reasonable by the general public.

Why Have Poverty Lines At All?

Given all the limitations and shortcomings, why should Statistics Canada or anyone else go to all the trouble of developing poverty lines? The answer to that question goes to the heart of our existence as a political and economic democracy.

Canadians have a well-honed sense of fair play that dates from our beginning as a nation. In politics, we accept the principle of one person, one vote and the principle that governments derive their mandate to govern from the collective will of voters. In economics, we pride ourselves on the ideal of equal opportunity for all and in making it possible for everyone to share in the vast bounty that is Canada.

Simply put, poverty lines are one measure of how well our democracy is working. They delineate that minority within the population that stands apart from the mainstream of Canadian life because of meagre income.

The statistics that come from poverty lines are every bit as valuable as widely accepted statistics such as the unemployment rate and the Consumer Price Index, yet they always seem to instill fear in the hearts of governments. Governments are loathe to admit that poverty exists in a country like Canada, but everyone knows that poverty exists - with or without official government sanction.

Poverty lines also give us the ability to take concrete steps to reduce the risk of poverty and to measure how good a job we are doing. The watershed in social policy is between the people who are content simply to count the number of poor people and those who use poverty statistics to champion better programs and policies for disadvantaged Canadians.

The National Council of Welfare's biggest complaint with the very low poverty lines supported by the Fraser Institute of British Columbia, for example, has been that the institute's only apparent interest in poverty lines is to show that poverty is not a problem in Canada and does not warrant action by government.

Our Council has maintained from the beginning that poverty statistics are only a first step. Having identified groups of people who are poor or most likely to be poor, we follow through by putting forward policy options to ease the burden of poverty and by promoting our proposals through our reports and our representations to governments.

Our annual Poverty Profile reports track how well or how poorly Canada is doing in the fight against poverty and identify the groups of people who face the highest risks of poverty. Our other reports build on this data with detailed policy recommendations to governments.

Statistics Canada's Low Income Cut-offs

The National Council of Welfare has long used the low income cut-offs or LICOs of Statistics Canada as its measure of poverty. Statistics Canada has consistently maintained that it does not regard the LICOs as poverty lines, presumably because the federal government does not want to give official recognition to poverty. Most social policy groups in Canada have consistently disagreed with the position of the federal government and continue to use the LICOs as poverty lines.

Despite the long-running dispute over terminology, the low income cut-offs are by far the most widely used measure of poverty in Canada. The survey data and methodology used to generate the cut-offs are done by a federal government agency with an international reputation for high-quality work. Statistics produced using the LICO methodology are readily available to researchers inside and outside government year after year at modest cost. Coincidentally or not, the income levels of the LICOs are in the mid-range of the alternative poverty lines that have appeared from time to time.

The LICOs for 1996 are shown in Table 1 on the next page. There are a total of 35 individual lines that vary with the size of the family and the population of the area of residence. These particular lines are known as 1986 base LICOs because they were first drawn from spending patterns in Statistics Canada's Family Expenditure Survey for the year 1986. The LICOs are updated every year using the Consumer Price Index.

TABLE 1 STATISTICS CANADA'S LOW INCOME CUT-OFFS (1986 BASE) FOR 1996
Family Size	Community Size
Family Size	Cities of 500,000+	100,000-499,999	30,000-99,999	Less than 30,000	Rural Areas
1	$16,061	$14,107	$13,781	$12,563	$10,933
2	$21,769	$19,123	$18,680	$17,027	$14,823
3	$27,672	$24,307	$23,744	$21,644	$18,839
4	$31,862	$27,982	$27,338	$24,922	$21,690
5	$34,811	$30,574	$29,868	$27,228	$23,699
6	$37,787	$33,185	$32,420	$29,554	$25,724
7+	$40,640	$35,696	$34,872	$31,789	$27,668

The methodology used to calculate the LICOs is unintelligible to most people. Very roughly speaking, the cut-offs mark income levels where people have to spend disproportionate amounts of their incomes on food, shelter and clothing.

The methodology starts out with data from the Family Expenditure Survey about the spending habits of all Canadians, not just spending by low-income people. The survey estimates what people actually spent on different items during the previous year, not what they needed to spend. Poor people often have difficult choices to make because of their limited incomes. Actual spending on housing, for example, is largely a function of what kind of housing is available in a given area in a given price range rather than family needs.

The low-income cut-offs vary by area of residence, but not by specific metropolitan areas. All metropolitan areas with populations of half a million people or more have the same poverty lines. However, it is clear from other data in the Family Expenditure Survey that the cost of living sometimes varies greatly from one large city to another, often because of differences in housing costs. Housing prices in Vancouver, for example, have been well above housing prices in many other major cities in recent years. From time to time, the lower mainland of British Columbia goes through housing "booms" that push very high housing prices even higher or "busts" that see prices plummet almost overnight.

The low income cut-offs are based on income after government transfer payments such as the Canada Child Tax Benefit, Old Age Security pension, GST Credit, Employment Insurance benefits and provincial or territorial welfare payments but before federal, provincial or territorial income taxes are deducted. This sometimes makes it difficult to assess the impact of government programs which provide non-taxable benefits.

Statistics Canada regularly produces data on income after federal and provincial income taxes, and it also produces after-tax LICOs and after-tax poverty statistics. Unfortunately, the results are not directly comparable with the regular LICO statistics and wind up being more confusing than enlightening. For example, the poverty rate for all persons before income taxes and after government transfers was 17.9 percent in 1996 using the LICOs. The comparable poverty rate for all persons after income taxes and after government transfers was only 13.5 percent using the after-tax LICOs. That could lead people to the erroneous and absurd conclusion that reducing your disposable income by paying income taxes actually reduces poverty.

The Search for Alternatives

Statistics Canada published a discussion paper in 1989 on alternatives to the LICOs that included a technical review of many of the shortcomings of LICOs. The paper led to interest in a poverty line known as a "low income measure" or LIM. LIMs were equal to one-half of median income adjusted for family size, but there was only set of LIMs for all parts of Canada.

Many social policy groups were enthusiastic about LIMs at first. The methodology was simple and in line with some of the international research on poverty. Unfortunately, the big disadvantage with LIMs turned out to be that poverty statistics based on LIMs were relatively flat over time - that is, poverty rates during the worst part of the economic cycle were not very different from poverty rates during the best part of the cycle. Statistics Canada has published several years' worth of poverty data based on LIMs, but there has been little sustained interest in LIMs from social policy groups.

A number of social planning and research agencies have done market baskets for their own areas over the years. Some of the baskets could probably double as poverty lines, but others are more like guides to reasonable living standards. The lines developed by the Montreal Diet Dispensary reflect a bare-bones approach to daily needs, for example, while the basket of the Community Social Planning Council of Toronto includes items that go beyond what most people would consider necessities.

Each of the planning and research groups has chosen particular family types as the focus of its work. That often makes it difficult to compare the totals from one city to another. It also leaves a number of large cities and smaller population centres without any guidelines at all.

One researcher who has done extensive work on market basket measures of poverty is Christopher A. Sarlo of Nipissing University in North Bay, Ontario. Professor Sarlo has developed poverty lines for all provinces and all major cities with the support of the Fraser Institute, a right-wing think tank based in Vancouver.

Professor Sarlo's work has attracted the interest of other right-wing groups, some government officials and some media editorialists. The salient feature of his poverty lines is that they are far lower than the LICOs and almost all of the other lines used from time to time by social policy researchers and therefore produce poverty rates that are very low.

For example, the Sarlo food basket contains no coffee or tea. There are no health care items in the basket on the grounds that poor people should be able to get charity dental services from dentists in the community and they should be able to pick up free eyeglasses from the local Lions Club.

Finally, researchers at Human Resources Development Canada are taking the lead in efforts to develop "market basket measures" of poverty as a special project for the federal, provincial and territorial governments. The project is an attempt to develop a "consensus definition" of poverty as an alternative to existing measures, and it arose in part from the collective efforts of governments that went into the new Canada Child Tax Benefit.

A paper on the market basket measures or MBMs was published in March 1998 and distributed to a very limited number of people inside and outside government. A more public release in December 1998 prompted charges that governments were trying to get rid of poverty by lowering the lines rather than doing anything of substance. British Columbia's Minister of Human Resources bluntly told the federal government that the market basket approach to measuring poverty was not acceptable. "The Government of British Columbia's commitment is to reducing, resolving and eliminating child poverty, not redefining it," the Minister said in a letter to the federal Minister of Human Resources Development.

The March 1998 paper said the MBMs are not a final product, but many social policy groups believe that the eventual and inevitable result of the exercise will be some kind of market basket poverty line to replace the LICOs.

Some anti-poverty groups consider the push for market basket poverty lines a form of poor-bashing. They say that many of the statistical measures used by governments - including the unemployment rate and the Consumer Price Index - are difficult to understand and are fraught with methodological shortcomings. Yet the only statistical measures that governments really do not accept and really want to change are the low income cut-offs.

II. SPENDING PATTERNS IN CANADA

Before we look at different ways of developing market baskets poverty lines, we need to look at possible sources of information about the spending habits of people. First of all, we should know what kinds of goods and services people actually buy in order to get a better idea of the items that could go into the market basket. We should also know if there are significant differences in spending patterns from one part of the country to another, so we know how many different market baskets to produce.

Some of the answers to these questions can be found in the Family Expenditure Survey (FAMEX) of Statistics Canada. Some of the information can be inferred from data in the Family Expenditure Survey. Some of the information is best obtained from more specialized surveys of the cost of specific goods and services.

Spending Habits in 1996

Most people think of food, shelter and clothing as the basic necessities of life, but the list is actually much longer. A reasonable market basket of goods and services would also include a host of other items such as household cleaning supplies, basic telephone service, soap and other personal care products, dental and eye care that is not covered by medicare, basic transportation, and a few toys for any children in the family.

As soon as we get beyond the absolute minimum items needed to survive and not one calorie or one cent above the absolute minimum, the list becomes quite lengthy. We believe that the vast majority of Canadians, given a detailed list of possible goods and services to choose from beyond basic food, shelter and clothing, would not hesitate to add many other items to a market basket.

To get a better idea of the items that could go into the market basket, we asked Statistics Canada for a detailed breakdown of the spending patterns of the poorest 20 percent of the population, as reported in the Family Expenditure Survey for 1996.

The results are presented in Table 2 on the next page. The spending patterns in the table show what the poorest people in Canada actually spent on different types of purchases in 1996. The figures may or may not reflect what they needed to spend.

The left-hand column for each household size in the table gives the percentage of households which reported a particular kind of spending, and the right-hand column shows average spending by those households. For example, only nine percent of the one-person households reported spending on education in 1996, but those who did spend on education spent $1,452 each on average.

Because not all people spend money in all categories every year, the columns do not add up to the figures near the bottom of the table in the rows for total current consumption and total expenditures.

The poorest 20 percent of one-person households reported average total expenditures for all items of $10,688 in 1996. By way of comparison, the Statistics Canada low income cut-off for a single person living in a city of half a million or more was $16,061. The poorest 20 percent of households with two or more persons had total expenditures of $22,868 in 1996. The average household size was 2.54 persons. The low income cut-off in a large city was $21,769 for a family of two and $27,672 for a family of three.

Here is a more detailed commentary on spending in each of the categories in Table 2. All in all, it is clear from the table that the vast majority of spending by poor people goes for items that are widely considered to be essential items.

Food: The vast majority of the money that poor people spent on food was for food purchased in stores rather than food purchased in restaurants or at snack bars. The one-person households who reported spending on restaurant food spent $427 on average in 1996, and the larger households with spending on restaurant food spent $694 on average.

TABLE 2 SPENDING IN 1996 BY POOREST 20 PERCENT OF HOUSEHOLDS
	One-Person Households		Households of Two or More Persons
	Percent Reporting	Average Spending	Percent Reporting	Average Spending
Food	100%	$2,091	100%	$4,325
Shelter	100%	$4,059	100%	$6,131
Household Operation	100%	$753	100%	$1,507
Furnishings & Equipment	72%	$299	92%	$755
Clothing	92%	$394	99%	$1,048
Transportation	84%	$801	96%	$3,048
Health Care	83%	$350	95%	$671
Personal Care	99%	$225	100%	$540
Recreation	84%	$460	97%	$1,183
Reading	64%	$141	78%	$172
Education	9%	$1,452	36%	$972
Tobacco & Alcohol	61%	$739	79%	$1,067
Miscellaneous	81%	$439	92%	$663
Total Current Consumption	100%	$10,082	100%	$20,844
Personal Taxes	46%	$71	73%	$1,250
Insurance & Pensions	29%	$255	61%	$742
Gifts & Contributions	63%	$785	74%	$885
Total Expenditures	100%	$10,688	100%	$22,868
Low Income Cut-off For A Large City	$16,061		$21,769/$27,672

Shelter: About 76 percent of the poorest one-person households were renters rather than home-owners, and roughly half of the renters had utility costs in addition to rent. About 57 percent of the larger households were renters, and about two-thirds of them had utility costs in addition to rent.

Household Operation: The major items in the category are telephone charges, cleaning supplies, paper products and child care. The one-person households in the table spent an average of $419 a year on telephones (including $260 on basic charges), $94 on cleaning supplies, and $112 on items such as paper towels, toilet paper and paper tissues. The larger households spent an average of $628 on telephones (including $263 on basic charges), $212 on cleaning supplies and $223 on paper products and similar items. Only ten percent of the households of two or more persons reported child care expenses, but those who did paid an average of $990 in 1996.

Furnishings and Equipment: There were no clear spending patterns for items in this large and diverse category.

Clothing: There is no thing as "typical" spending on clothing because of the huge range of items for men and women, boys and girls, and babies or very young children, but the amounts spent seem quite consistent for both types of households. Average spending for the one-person households was $394 in 1996. Average spending for the larger households was $1,048 or $413 per person when the total is divided by the average household size of 2.54.

Transportation: Seventy-two percent of the one-person households relied on public transportation, and their average spending on public transportation was $343 in 1996. In the larger households, there was a considerable overlap, with 74 percent of the households using private transportation and 63 percent using public transportation. The spending on private transportation was by and large for gas and oil, maintenance and repairs, insurance and license fees rather than for the purchase of a car or truck.

Health Care: The major items in the category are prescription and non-prescription drugs, eye care, dental care, hospital and other care not covered by medicare, and health insurance premiums, including premiums for public health insurance in Alberta and British Columbia. Most households reported spending on pharmaceuticals, and the average amounts spent were $167 in one-person households and $281 in larger households.

The percentage of households which reported spending on the other items in the category was shockingly low, considering how important the items are. It appears that many poor people were doing without essential health care items simply because they could not afford to pay for them.

Among one-person households, 21 percent had eye care expenses averaging $170, 20 percent had dental expenses averaging $316, seven percent had hospital and other non-insured expenses averaging $225, and eight percent had health premiums averaging $337. Among households of two or more persons, 37 percent had eye care expenses averaging $214, 35 percent had dental expenses averaging $354, 19 percent had hospital and other non-insured expenses averaging $195, and 22 percent had health premiums averaging $500.

Personal Care: The major items in the category include soap, cosmetics, shampoo, deodorant, toothpaste, razors, disposable diapers and haircuts.

Recreation: The category covers an array of sports, toys, hobbies and other forms of entertainment, including home entertainment, cable television and spectator events. Among one-person households, half had home entertainment expenses averaging $194 and 46 percent had cable TV at an average cost of $315. Among larger households, 74 percent had home entertainment expenses averaging $392, 61 percent had cable TV at an average cost of $343, and 81 percent attended spectator events at an average cost of $331.

Reading: The category includes newspapers, magazines and books.

Education: Relatively few of the poorest households reported spending on education, although average spending was sizable. The category includes books and supplies and tuition fees.

Tobacco and Alcohol: Forty percent of the people in the one-person households were smokers and 52 percent of the larger households had at least one smoker. About 44 percent of the one-person households and 64 percent of the larger households reported spending on alcoholic beverages.

Miscellaneous: The category includes interest charges, banking fees, union dues and various forms of gambling. Sixty-four percent of the one-person households had gaming expenses of $256 on average, and 73 percent of the larger households had gaming expenses averaging $235.

Personal Taxes: The category covers income taxes, but few other taxes. The totals in the table are misleading, because they include income taxes paid for 1996 minus any tax refunds. Fifteen percent of the one-person households paid taxes on their 1996 incomes that averaged $778, and 51 percent of the households of two or more persons paid income taxes averaging $2,080.

Insurance and Pensions: The category includes life insurance premiums, premiums for Employment Insurance and contributions to the Canada or Quebec Pension Plans.

Gifts and Contributions: Gifts refers to items or money given to other people, and contributions includes donations to both religious and charitable organizations.

Differences in Spending by Income Class

It should come as no surprise that the spending patterns of poor people often differ sharply from the spending patterns of people in other income groups. Table 3 on the next page shows the percentages of total current consumption spent on different categories of spending by the poorest 20 percent of all households, the middle 20 percent and the richest 20 percent in 1996.

As might be expected, poor households spend proportionately more of their available incomes on food and shelter than households with higher incomes. Conversely, they spend proportionately less than other households on clothing, transportation and recreation.

What is surprising about the data in Table 3, however, is the consistency in several major categories of spending. Spending on household operation, furnishings, health care, personal care and reading as a percentage of total current consumption were more or less than same in all three income classes.

TABLE 3 SPENDING AS PERCENTAGE OF TOTAL CURRENT CONSUMPTION BY INCOME CLASS, ALL HOUSEHOLDS, 1996
	Poorest 20 Percent	Middle 20 Percent	Richest 20 Percent
Food	20%	18%	16%
Shelter	35%	25%	22%
Household Operation	7%	6%	7%
Furnishings & Equipment	3%	4%	4%
Clothing	4%	6%	7%
Transportation	12%	18%	19%
Health Care	3%	3%	3%
Personal Care	3%	3%	2%
Recreation	5%	7%	9%
Reading	1%	1%	1%
Other Items	8%	9%	10%
Total Current Consumption	100%	100%	100%

Spending Patterns Across Canada

The next logical question to ask is whether spending by poor people differs from one part of Canada to another. Unfortunately, that question cannot be answered on the basis of available data. The Family Expenditure Survey is not large enough to allow data to be broken down for poor people alone by region or urban centre. The sample sizes in the survey from individual cities are relatively small to begin with, and taking just the lowest 20 percent of the sample for each city would not produce reliable results.

The only available alternative is to look at variations in spending for all households based on the population of the area of residence or in spending for all households from one large city to another.

Table 4 on the next page shows the different categories of household spending by size of area of residence, from rural areas to cities of a million or more people. The first column of the table gives the total amount spent in dollars on current consumption. The rest of the columns show spending in each category as a percentage of total current consumption.

In terms of dollars spent, the big divide comes when the population hits 100,000. Areas with populations under 100,000 have total current consumption ranging from $30,510 to $31,697. Areas of 100,000 or larger fall into a higher spending group that ranges from $33,496 to $35,990. Coincidentally, a population of 100,000 is also the size chosen by Statistics Canada to start using census metropolitan areas. There were a total of 25 census metropolitan areas in the 1996 census.

The percentages of spending for each of the categories and populations reflect some interesting patterns. The percentages for different areas of residence are almost the same for food, household operation, furnishings, clothing, health care, and personal care. The big differences are in rented or owned housing, transportation and perhaps recreation and reading combined.

TABLE 4 SPENDING BY CATEGORY AS PERCENTAGE OF TOTAL CURRENT CONSUPTION, ALL HOUSEHOLDS, 1996
&NBSP;	Total in Dollars	Food	Rented Housing	Owned Housing	Household Operation	Furnishings & Equipment	Clothing	Public Transport	Private Transport	Health care	Personal Care	Recreation & Reading
Rural	$31 697	18%	13%	14%	7%	5%	6%	2%	21%	3%	2%	9%
Under 30,000	$31 106	18%	15%	15%	7%	4%	6%	2%	20%	3%	2%	10%
30,000-99,999	$30 501	17%	18%	18%	6%	4%	6%	2%	19%	3%	2%	9%
100,000-249,999	$33 496	17%	16%	17%	6%	4%	6%	1%	20%	3%	2%	9%
250,0001-499,999	$34 579	17%	19%	19%	7%	4%	6%	2%	19%	3%	2%	9%
500,000-999,999	$35 641	17%	17%	18%	6%	4%	6%	2%	17%	3%	2%	10%
1 million or more	$35 990	18%	20%	23%	7%	4%	6%	2%	18%	3%	2%	8%

TABLE 5 SPENDING BY CATEGORY AS PERCENTAGE OF TOTAL CURRENT CONSUMPTION, ALL HOUSEHOLDS, 1996
&NBSP;	Total in Dollars	Food	Rented Housing	Owned Housing	Household Operation	Furnishings & Equipment	Clothing	Public Transport	Private Transport	Health care	Personal Care	Recreation & Reading
St. John's	$36 348	18%	13%	14%	7%	5%	7%	2%	20%	3%	3%	7%
Charlottetown	$31 396	17%	18%	16%	7%	4%	7%	2%	17%	3%	3%	9%
Halifax	$34 694	16%	18%	16%	7%	4%	6%	2%	18%	3%	2%	9%
Saint John	$29 596	18%	15%	14%	8%	4%	6%	2%	20%	3%	2%	8%
Quebec	$30 821	20%	13%	17%	6%	4%	7%	1%	19%	4%	3%	9%
Montreal	$31 243	19%	18%	21%	6%	4%	7%	2%	18%	3%	3%	8%
Ottawa	$38 067	17%	20%	23%	6%	4%	7%	3%	16%	3%	2%	10%
Toronto	$39 273	17%	21%	23%	7%	4%	6%	3%	19%	2%	2%	8%
Thunder Bay	$36 468	16%	17%	15%	6%	4%	6%	2%	20%	2%	2%	10%
Winnipeg	$32 130	18%	17%	18%	7%	4%	6%	2%	16%	3%	3%	9%
Regina	$33 491	17%	14%	14%	7%	4%	7%	2%	19%	3%	3%	10%
Saskatoon	$33 912	17%	15%	16%	7%	5%	7%	2%	18%	3%	3%	9%
Calgary	$36 515	17%	17%	17%	7%	5%	7%	3%	15%	4%	2%	1%
Edmonton	$38 194	17%	14%	16%	7%	4%	6%	2%	18%	4%	2%	10%
Vancouver	$37 246	17%	22%	22%	6%	4%	6%	3%	17%	3%	2%	9%
Victoria	$34 803	16%	21%	21%	7%	4%	6%	3%	15%	4%	2%	9%

The 1996 Family Expenditure Survey does not present detailed data for all of the 25 census metropolitan areas, but it does have details for one or more cities in each province. The data are presented on the previous page as Table 5.

The table shows that total spending is not always a function of size, since some of the biggest differences are found among the eight largest urban centres: Toronto, Montreal, Vancouver, Ottawa, Edmonton, Calgary, Quebec and Winnipeg. Statistics Canada reported that total current consumption in Montreal in 1996 was $31,243 on average, spending in Quebec City was $30,821, and spending in Winnipeg was $32,130. All three figures are far below the totals for Toronto, Vancouver, Ottawa and Edmonton, and substantially below the total for Calgary.

As in the previous table, sharp variations in spending as a percentage of total current consumption were found in housing, transportation and recreation. Food costs also varied noticeably, from 16 percent of total current consumption in Halifax, Thunder Bay and Victoria to 20 percent in Quebec City.

Altogether, the table adds weight to the idea of developing market baskets for each large city individually.

Other Sources of Information

There are alternative sources of data that could be used for the food and shelter components of a larger market basket of goods and services.

For a number of years, Agriculture Canada produced a "nutritious food basket" of 64 items and priced the basket in 18 cities. It also produced and priced a Canada-wide basket based on data originally obtained in the 18 cities.

Even though the baskets were widely used by health and social service agencies, Agriculture Canada stopped producing them in 1995. Health Canada has expressed interest in reviving the exercise and has been exploring options for national and provincial or regional food baskets. The goal is to develop baskets with a variety of foods that people like to eat and which meet accepted standards of nutrition.

Health Canada has priced a revised national nutritious food basket for a family a four in Ottawa at $123 a week or $6,396 a year in a "final" report published in March 1997. The amounts were for a family made up of one man and one woman between the ages of 25 and 50, a boy aged 13 to 15 and a girl aged seven to nine. Different baskets were suggested for 23 age and sex groups, including pregnant and nursing women.

For housing costs, one possible source of information is a survey by Canada Mortgage and Housing Corp. on average rents in different cities and towns. The survey covers rents in privately owned apartments consisting of at least three units. It does not cover housing in duplex units and "granny" or in-law suites in private homes. It also does not cover public housing, subsidized housing or housing in institutions.

More than 150 areas of 10,000 or more people are covered in the CMHC survey, and information is normally available on the average cost of a bachelor apartment and the cost of one-bedroom, two-bedroom and three-bedroom apartments.

Given the variations in housing costs that we saw earlier in this chapter and the variations in food costs among some of the largest cities, the CMHC data and an updated version of the Agriculture Canada nutritious food basket would be logical choices in any overall market basket of goods and services.

III. SOME EXAMPLES OF MARKET BASKETS

Now that we have a more detailed understanding of the spending patterns of poor people and the variations in spending from one part of the country to another, we can take a look at some of the many different ways of putting together a market basket of goods and services.

The family type we use as an example in this chapter is a family of four living in the greater Vancouver area. We start with the Family Expenditure Survey data from the previous chapter and compare the data with four different approaches: a market basket measure or MBM being developed by researchers at Human Resources Development Canada, a less statistical variation put forward simply for the purpose of comparison by the National Council of Welfare, calculations by the Social Planning and Research Council (SPARC) of British Columbia and the bare-bones poverty line of Professor Sarlo. Details of the different approaches are summarized in Table 6 on the next page.

The column labelled "adjusted FAMEX, all areas" is a point of reference for the other examples rather than an actual market basket poverty line. It consists of data from the Family Expenditure Survey for the poorest 20 percent of households of two or more persons in all parts of Canada, as reported in Table 2 of the last chapter. The data were "adjusted" by dividing the dollar figures in Table 2 by the average household size of 2.54 and multiplying the results by four to get comparable figures for a family of four. We dropped the spending categories for education, tobacco and alcohol, and miscellaneous items, along with personal taxes, insurance and pensions, and gifts and contributions. The total of $30,519 at the bottom of the column is the sum of all the individual items in the column.

The next column shows the market basket measure for a family of four in Vancouver that is being developed by Human Resources Development Canada. There are three specific components of the basket: an updated version of the nutritious food basket of Agriculture Canada, housing costs from the Canada Mortgage and Housing Corp. survey of rents for October 1996, and a clothing item based on 75 percent of a clothing budget originally developed by the Social Planning Council of Metropolitan Toronto.

TABLE 6 MARKET BASKETS FOR A FAMILY OF FOUR IN VANCOUVER, 1996
	Adjusted FAMEX, All Areas	HRDC Market Basket	Less Statistical Basket	SPARC	Sarlo
Food	$6,811	$7,196	$7,196	$6,693	$5,901
Shelter	9,655	10,380	10,380	9,420	10,432
Utilities				245
Household Operation	2,373		744	771	985
Furnishings & Equipment	1,189		1,189		194
Clothing	1,650	2,200	1,650	2,540	1,178
Transportation	4,800		1,488	1,386	424
Health Care	1,057		866
Personal Care	850		850	603	648
Recreation	1,863		1,230
Reading	271		154
Other		5,638	1,287	5,608	42
Total	$30,519	$25,414	$27,034	$27,266	$19,803

Rather than do separate calculations for the other spending categories, the researchers at HRDC decided to have an "other" category equal to 60 percent of the cost of food and clothing combined - a calculation that works out to $5,638 in the table. Housing costs were not used in the calculation because they vary greatly from one part of the country to another. The figure of 60 percent is a rough approximation of the average cost of other items in different budget guides from the mid-1980s that were prepared by social planning agencies in Montreal, Toronto, Hamilton, Winnipeg and Edmonton. It is also roughly comparable to data for other categories of spending combined that are reported in the FAMEX survey for 1996.

The column labelled "less statistical basket" is an alternative calculated by the National Council of Welfare to get a comparable market basket that had more individual items than the HRDC market basket measure, but which also made use of some information from the Family Expenditure Survey.

The less statistical basket uses the same sources for food and shelter as the HRDC market basket measure, but it has different figures for the other categories of spending. The figures for furnishings and equipment, clothing and personal care are the same as the figures in the FAMEX survey. People would not normally buy or replace all their household furnishings and equipment in a single year, and the items and amount of spending would vary greatly among all poor families. Therefore, we decided to use the average from FAMEX as a reasonable "proxy" for furnishings and equipment. In the cases of clothing and personal care, spending patterns vary greatly with the age and sex of members of the household, so it probably makes more sense to use averages from FAMEX than to develop specific baskets of clothing or personal care items.

For all the rest of the categories except the other category near the bottom of the column, we made up individual baskets of goods or services to represent basic spending by poor people. Readers can judge for themselves whether our baskets are realistic or unrealistic.

The basket for household operation consists of basic telephone service of $252 a year plus the following items: 144 rolls of toilet paper, 52 rolls of paper towels, 12 boxes of facial tissue, 5,000 paper napkins, six rolls of aluminum foil, six rolls of plastic wrap, 288 loads' worth of laundry detergent, three jugs of bleach, three jugs of fabric softener, three cans of floor wax, 12 bottles of dish detergent, ten dish towels, 54 scouring pads, 12 bottles of all-purpose cleaner, 12 cans of scouring powder, 80 garbage bags, and 20 light bulbs. Also included were a broom, scrub brush, four cans of shoe polish, and five spools of thread.

The transportation basket is made up of bus passes for the two adults in the family. We also budgeted for 24 trips by taxi (one round trip per month) at a cost of $16 for each round trip. The basket for private transportation in an area without public transportation is roughly estimated at $1,900 a year. It consists of 1,500 litres of gasoline, the equivalent of two 8,000 kilometre checkups at a car dealership, automobile insurance and license fees. Alternatively, it could be calculated as 9,500 kilometres a year at 20 cents a kilometre.

For health care, the basket includes three dental checkups at $85 each, six courses of antibiotics at $50 each (one each for the adults and two each for the children) and one pair of glasses at $200. We assumed for the purposes of illustration that one of the children would be too young to need a regular visit to the dentist and that only one of the members of the family had poor eyesight. The health care basket also includes a thermometer and a variety of first aid supplies and non-prescription remedies for headaches, upset stomachs and colds.

The recreation basket is two individual YMCA memberships and basic cable TV, and the reading basket is an annual subscription to the Vancouver Sun.

Finally, we took five percent of the total of all the above items for "other" spending. Our assumption was that every family has special needs or priorities of its own. It makes more sense to have a miscellaneous category than to try to identify and quantify a host of other specific items of spending. At the same time, we think the miscellaneous category should be relatively small.

The next column in the table is the basket calculated by the Social Planning and Research Council of B.C. as its estimate of the cost of daily living in the Vancouver area. The basket in the table was developed specifically for a family made up of two adults, one five-year-old child and one infant.

The SPARC methodology reflects both of the approaches used in the two previous columns of the table. The calculation for food is a variation of the Agriculture Canada food basket, and the calculation for shelter came from Canada Mortgage and Housing Corp. data for the lowest one-quarter of the local housing market. The rest of the items came from a combination of individual baskets, statistical data from the Family Expenditure Survey or estimates originally developed by the Metro Toronto Social Planning Council.

The final column of the table represents the calculations published by Professor Sarlo in Poverty in Canada - 1994 and updated to 1996 by the National Council of Welfare using the Consumer Price Index. The total of $19,803 is far below any of the other alternatives shown in Table 6, primarily because of the items that Professor Sarlo left out of the calculations on purpose.

The basket for furnishings and equipment is only $194 a year, the budget for transportation is only $424 a year and the cost of writing materials shown in the other category was pegged at a mere $42 a year. There are no allowances at all for health care not covered by medicare, recreation and reading. On the other hand, the budgets for shelter, household operation, and personal care are not out of line with the figures in the other baskets.

Leaving aside the Sarlo market basket for the moment, the other three market baskets in Table 6 are all in the same general income ranges, and they all share a number of common features. Some of the components of spending are defined and priced very carefully, and others are general categories where the specific kinds of spending are left open to reflect the different priorities and needs of different families.

All things being equal, the National Council of Welfare would rather see more specific basket items and fewer other or miscellaneous items. That is not because the less statistical approach is any more valid or any less arbitrary. It is mainly because we believe that people can more easily relate to specific items of household spending, from toothpaste to monthly bus passes, than they can to general categories of goods and services such as "personal care" and "transportation."

Some groups which have worked with market baskets in the past say it is easier to defend market baskets where all the items are listed. The best line of defence, they argue, is challenging sceptics to say which items are not essential and should be removed from the basket. Unspecified miscellaneous items are always the most likely to be challenged.

We also like having as many items as possible specified in a market basket to remind people that being in the mainstream of Canadian life requires more than a few scraps to eat and a place to get away from the cold. It requires other types of "necessities" for reasonable living, from a daily newspaper to decent dental care to a few leisure activities.

That brings us back to the poverty lines of Professor Sarlo. They reflect a mean-spirited view of life that regards people as poor only if they can be shown to be visibly and strikingly different from the rest of society. The National Council of Welfare believes that poor people already have enough trouble making ends meet and should not have to be considered social pariahs as well before they qualify for assistance from governments.

IV. WHAT WILL HAPPEN IF WE CHANGE THE POVERTY LINES?

Before we offer support for changing the poverty lines, it seems reasonable to ask how the changes would affect our view of poverty. Some of our colleagues in anti-poverty and social policy groups are wary that governments might use the exercise on market basket measures as a "convenient" way of defining a large portion of poverty out of existence. Lowering the poverty lines invariably means lowering the number of people living in poverty. The National Council of Welfare believes these concerns must be fully and completely addressed by governments if they expect public support for any changes in the poverty lines.

More than anything else, moving to market basket measures would affect our perception of the distribution of poverty in different parts of the country. The low income cut-offs are based on the size of the area of residence and lump all the biggest cities in Canada into the same category. All the market basket measures we have seen involve poverty lines that were drawn to reflect the cost of living in individual cities.

We saw in Chapter II that actual living costs are much lower in Montreal, Quebec City and Winnipeg than they are in most other big cities, so the move to market basket lines would lead to much lower poverty rates in those three cities. Because of their large populations, lower rates in Montreal, Quebec City and Winnipeg would translate into lower provincial poverty rates in Quebec and Manitoba.

Table 7 on the next page shows how the provincial poverty rates for all persons would have changed in 1996 with the shift from LICOs to the market basket measures being developed by HRDC. At first glance, the most striking changes would appear to be the drop in poverty rates in all provinces except Newfoundland. On closer inspection, it is the differences in the drops from province to province that are the most interesting.

The most dramatic drop is in Quebec, where the poverty rate falls by 49 percent with the switch from LICOs to MBMs. The drops are almost as sharp in Manitoba and Alberta. Meanwhile, the poverty rate in Ontario falls by only 21 percent. Ontario, long regarded as the richest of the so-called "have" provinces, winds up with a poverty rate that is higher than the poverty rate in Quebec and five other provinces.

TABLE 7 POVERTY RATES FOR ALL PERSONS, 1996
	Before-Tax Low Income Cut-Offs	HRDC After-Tax Market Basket Measures	Percentage Change
Newfoundland	17.2	17.8	+4%
Prince Edward Island	12.6	9.6	-24%
Nova Scotia	18.1	14.9	-18%
New Brunswick	15.8	12.0	-24%
Quebec	21.2	10.8	-49%
Ontario	15.8	12.5	-21%
Manitoba	18.8	11.1	-41%
Saskatchewan	16.5	12.1	-27%
Alberta	15.8	9.2	-42%
British Columbia	17.6	13.9	-21%

Many of the other changes in poverty rates would probably be much easier to understand and to accept. Unfortunately, we have not yet seen the full range of statistics that could be produced using any given set of market basket poverty lines.

Based on our past work, the National Council of Welfare would expect to see a range of poverty rates by family type not unlike the current range. Single-parent mothers under 65 with children under 18 would still be the family type with the highest poverty rate by far. Married couples 65 and older would still be the family type with the lowest poverty rate. In fact, poverty among all groups of seniors could fall noticeably. Seniors tend to have incomes that are very close to the current low income cut-offs. Even a modest drop in the poverty lines under a shift to market basket measures could lead to a further decline in the poverty rates of seniors.

One other unknown is the way that poverty rates based on market basket poverty lines would change over time. The researchers at HRDC have yet to produce any poverty statistics for years except 1996. The National Council of Welfare would like to see data for an entire economic cycle before offering a firm opinion on any changes in the way we measure poverty.

Under the low income cut-offs, poverty rates tend to go up when the economy goes into a recession, and they tend to go down when economic growth resumes. Market basket poverty lines simply would not be credible if they produced poverty rates that were almost identical from one year to the next or if the poverty trends turned out to be strikingly different from the trends using LICOs.

When Statistics Canada experimented with its Low Income Measures a few years ago as an alternative to the low income cut-offs, it produced several years' worth of data using LIMs to allow researchers a chance to see whether the results were useful or not. The same should happen long before there is any move to market baskets as measures of poverty.

Historical data would also be a good safeguard against those who would minimize the extent of poverty in Canada by focusing on lower poverty rates resulting from any move to market basket measures. The National Council of Welfare would be the first to criticize anyone who mixed the "old" poverty rates with the "new" poverty rates in an effort to show that poverty was no longer a serious problem in Canada.

V. OTHER ISSUES RELATED TO POVERTY LINES

We have already shown some of the different ways of drawing market basket poverty lines and expressed some concerns about the poverty statistics that would arise from any new measures. In this chapter, we raise a number of other issues that we believe should be addressed by governments before they proceed any further.

How Many Poverty Lines? One of the big questions left to answer is how many lines should we have - or how many lines could we tolerate. It seems clear that the number would increase substantially in any move to market basket lines.

The low income cut-offs of Statistics Canada consist of one set of 35 different lines. The spending patterns discussed earlier in this report suggest that it would be reasonable to have separate sets of lines for each of the 25 census metropolitan areas in Canada and separate lines for all other areas of each province. That would give us 35 sets of lines for starters. Further research would be necessary to tell whether any of these lines could be combined. For example, it may be reasonable to have one set of lines for all the eight census metropolitan areas in Ontario aside from Toronto and Ottawa rather than eight different sets of lines.

Each set of lines would have to be adjusted for different household sizes. The current LICOs provide for household sizes ranging from one person to seven or more persons. Seven different household sizes and 35 different geographical areas multiplied together would give us a total of 245 poverty lines. That sounds like far too many, but it may not be unmanageable. From the requests we receive in our office on a regular basis, it is clear that the vast majority of callers are interested in the poverty lines for their home towns and really do not care about the lines elsewhere.

Adding New Areas. One of the Statistics Canada surveys used to get data on poverty every year is done in conjunction with the Labour Force Survey, which specifically excludes Indian reserves and the territories. The National Council of Welfare urges Statistics Canada to consider including all these areas in any new system based on market basket poverty lines.

Statistics Canada now does a Survey of Household Spending every year instead of the Family Expenditure Survey every four years. The new survey covers spending patterns in Whitehorse, Yellowknife, the new Nunavut capital of Iqaluit, and smaller communities in the territories on a rotating basis. We hope that Statistics Canada will take whatever additional steps are necessary to collect and publish poverty statistics for the North as well as statistics on spending patterns.

The National Council of Welfare also believes that Indian reserves should be included in all surveys that deal with poverty in Canada. Poverty among Aboriginal peoples has long been a problem, and we do ourselves no service by failing to measure the extent of the problem. Collecting data and calculating poverty statistics in reserve communities should be no more difficult than current efforts to produce information on poverty in small towns and rural areas.

Family Size and Type. Most poverty lines start out with a typical household - often a family of two adults and two children - and then make adjustments for other sizes of households.

The low income cut-offs make these adjustments in a simple and straight-forward manner, and the result is seven different sets of lines based on family size. The low income measures or LIMs developed by Statistics Canada took a different approach by distinguishing between the adults and the children in a family. Instead of one set of lines for a family of five persons, for example, the LIMs had four different sets of lines for the following household types: five adults, four adults and one child, three adults and two children, and two adults and three children (alternatively, one adult and four children).

We would hope that Statistics Canada never again tries such convoluted arrangements. There are probably no more than a handful of households in Canada that consist of four adults and one child, and it seems ridiculous to have unique poverty lines for such an untypical group.

A similar problem arises with the Agriculture Canada nutritious food basket, which varies with the sex of the adults in the family and the age of the children. We would hope that standard baskets could be developed that would relate exclusively to family size. To do otherwise would require separate poverty lines for a host of different family types, a requirement that we believe would be unworkable.

Adjustments to After-Tax Income. The poverty statistics from Statistics Canada are produced by comparing the incomes of different families in the Survey of Consumer Finances with the appropriate poverty line for each family and then projecting the results to the population as a whole.

With the current low income cut-offs, the comparison is between the poverty line and family income before income taxes and after government transfers such as welfare, Employment Insurance, Old Age Security and the Canada Child Tax Benefit.

Income after income taxes would be the logical starting point for calculating poverty statistics under market basket poverty lines. HRDC raises the idea of having Statistics Canada use income after income taxes and also after contributions to the Canada and Quebec Pension Plans and premiums for Employment Insurance. Both types of levies are required by law, both are paid by literally millions of Canadians, and both can make a huge difference in the amount of income a family has available to spend on other items. The HRDC proposal therefore makes sense to us.

Where the Department is on very shaky ground is its related proposals to deduct child care expenditures, maintenance payments, medically necessary health expenditures and work-related expenses from the after-tax income to be compared to the poverty line.

This may sound appealing and logical at first blush, but it would become a nightmare of mammoth proportions and could wind up discrediting all the work that has been done on market basket poverty lines.

Take the example of child care. Families which pay for child care out of their own pockets have less available income than families without children or families with relatives that provide child care free of charge or families which receive child care that is fully subsidized by government. Only ten percent of the households of two or more people in the lowest 20 percent of the population reported day care expenses in 1996, and the average outlay was only $990. Making adjustments in the poverty statistics for the small portion of the population with relatively modest out-of-pocket expenses makes no sense at all in our view.

The issue is not whether child care is a worthwhile expenditure or a logical consideration in the calculation of available household income. The issue is that poverty lines should apply to large groups of people within the general population. They should not be tailored to each and every possible variation in family circumstances.

Once you start making adjustments, it is almost impossible to stop. There are many other spending items that are equally important, but which should not be considered in the calculation of poverty statistics. Homeowners without mortgages tend to have much lower outlays on housing than homeowners who still have mortgages. People in some rural areas drive long distances to work and have higher than normal transportation expenses. People with allergies have special dietary needs that are more expensive than the food in the average food basket. People with employer-sponsored health insurance plans have much lower outlays on drugs and dental care and eyeglasses than people without insurance. People going to college and university have tuition fees and expenses on textbooks and school supplies that most other people do not have.

Updating Market Basket Lines. One of the other problems with the low income cut-offs is that the base year for the LICOs changes from time to time. The changes can cause a noticeable break or "notch" in the historical series of poverty statistics and make it more difficult to do comparisons over long periods of time. Between changes in the base year, the LICOs are updated using the Consumer Price Index.

With market basket lines, the items in the basket could be priced by Statistics Canada every year. If housing prices in Vancouver or elsewhere changed dramatically from one year to the next, the changes would be reflected in the poverty lines.

Some of the other items in a market basket of goods and services could be based on spending patterns in the new annual Survey of Household Spending. Rather than updating the basket based on the Consumer Price Index, Statistics Canada could simply use the cost of items from the most recent survey.

A more touchy issue is how to adjust the actual items in the market basket over the longer haul. Spending on leisure-time activities, for example, has changed dramatically during the last generation. Black-and-white television was a luxury item in 1949, but colour TV is the norm at all income levels in 1999. Eating habits have also changed over the years, with people eating less fatty food today and drinking more low-calorie beverages.

There is no perfect way of making sure that a market basket is kept up to date. On the other hand, Statistics Canada itself adjusts the items that it prices in its Consumer Price Index from time to time. It should not be impossible to make comparable changes in a market basket of goods and services.

An Official Measure of Poverty. One of the sillier debates about poverty lines arises because of the federal government's continuing refusal to give any official recognition to the existence of poverty in Canada. As a result, the National Council of Welfare and most other social policy groups simply regard the low income cut-offs as poverty lines. We could call them the not-much-money lines or the cut-off-from-the-mainstream-of-society lines or the having-difficulty-finding-enough-money-to-pay-the-rent-at-the-end-of-the-month lines, but we find it much more direct and much more honest to call them poverty lines.

Assuming that governments and ordinary Canadians reach some reasonable consensus, we should all agree to call them poverty lines.

Full Consultations with Interested Parties. Like many other ventures undertaken by governments these days, the work on market basket measures of poverty was started behind closed doors a number of months ago and continued in secret for months on end. We believe the secrecy was totally unnecessary and may ultimately make it harder for ordinary people to support changes in the status quo.

The National Council of Welfare would like to see governments consult widely and openly with people outside government who are interested in poverty. First and foremost, this means that people who are poor and anti-poverty groups everywhere in Canada should have a large say in how governments define and measure poverty. Poor people have a detailed understanding of poverty from their personal experiences, and it would be stupid for governments to ignore this valuable source of expertise. Beyond this, it would be wrong for elected officials or bureaucrats who are tens of thousands of dollars away from any definition of poverty to decide unilaterally whether the poverty line for a poor family of four living in a large city should be $31,862 a year before taxes or $25,414 a year after taxes.

We would like to see the experts at Statistics Canada play an active, day-to-day role in the research being done for the federal, provincial and territorial governments on market basket measures. Their expertise in survey methodology and data collection and processing is a natural complement to the expertise at Human Resources Development Canada and could ultimately determine whether any new measures of poverty are viable in the long term.

Once the best options for market basket measures are found, we would like Statistics Canada to publish poverty lines and detailed data free of charge for a full economic cycle to allow people to assess the usefulness of the new measures and to root out any anomalies in the poverty lines or the statistics that arise from them.

We envisage the entire process taking a minimum of three years. Then and only then should governments decide whether to use the new measures and abandon the low income cut-offs. Then and only then could ordinary people be satisfied that the search for new measures of poverty was begun to assist governments in the fight against poverty rather than allowing them another excuse for doing so little.

PARTING THOUGHTS

This discussion paper is an attempt by the National Council of Welfare to find out whether there is any inclination among Canadians to explore new ways of measuring poverty - more specifically whether we should consider moving from the low income cut-offs of Statistics Canada to poverty lines based on the cost of a market basket of goods and services.

The Council participated in consultations begun by Statistics Canada in 1989 in a search for alternatives to the low income cut-offs or LICOs. At the end of the exercise, Statistics Canada decided to continue using LICOs as its preferred way of defining the "low income" population. The Council and most other groups continued using LICOs as their preferred measure of "poverty."

The current round of debate about poverty lines was touched off by the federal, provincial and territorial governments in a search for alternative ways to assess the impact of the new Canada Child Tax Benefit on families with children. It remains to be seen whether the search will turn out to be any more successful than previous efforts.

We would like to see the debate on poverty lines pursued, resolved if possible within a reasonably short period of time, and then put aside. The outcome could be continued use of the low income cut-offs by most researchers. It could be a switch to a new measure of poverty based on the cost of a market basket of goods and services. It could even be having two sets of measures that complement each other. The Acceptable Living Level group in Winnipeg, for example, found that their guidelines were close to the low income cut-offs. That led them to the view that the market basket and statistical approaches validate each other and make both approaches more credible.

Whatever happens with the debate on poverty lines, the prime concern of the National Council of Welfare is finding the most effective ways of fighting poverty in Canada and promoting our proposals among ordinary people and governments at all levels. Counting poor people is always going to be unsatisfying work. It only makes sense if we take the next logical step and do our best to eradicate poverty in all its forms.

ENDNOTES

1. M.C. Wolfson et J.M. Evans, Statistics Canada's Low Income Cut-Offs: Methodological Concerns and Possibilities (Ottawa: Statistics Canada, 1989).

2. See the Montreal Diet Dispensary's Bugeting for Basic Needs anf Budgeting for Minimum Adequate Standard of living and the Guides for Family Budgeting published by the Social Planning Council of Metropolitan Toronto, now the Community Social Planning Council of Toronto.

3. Christopher A. Sarlo, �Poverty in Canada-1994� Fraser Forum (february, 1994),pp.18, 29.

4. Statistics Canada, Family Expenditure in Canada, 1996, Catalogue No. 62-555-XPB.

5. The percentages were calculated from data in Historical Table 1 - Family Expenditure in Canada, 1996.

6. Judith Lawn, �A Revised National Nutrious Food Basket: Final Report� (Ottawa: Sant� Canada, 1997), p.17. See also Stefa Katamay, �A Revised National Food Basket - Where to From Here?� (Ottawa: Health Canada 1996).

7. Construction of a Preliminary Market Basket Measure of Poverty, Report by the Federal/Provincial/Territorial Working Group on Social Deve;opment Research and Information (Ottawa: Human Resources Canada, 1998).

8. Michael Goldberg, Widening the Gap (Vancouver: Social Planning and Research Council of British Columbia, 1997).

9. Professor sarlo made a few minor changes in methodology for the second edition of his book Poverty in Canada; The result was a poverty line for a family of four in Vancouver of $19 628 in 1994 or $20 354 when the total is ipdated to 1996.

10.Construction of a Preliminary Market Basket Measure of Poverty, p.21, and unpublished data from Statistics Canada.

QUESTIONS FOR READERS: TELL US WHAT YOU THINK

Here's your chance to let us know - and to let governments know - what you think about the way we measure poverty in Canada. Take a minute to answer the questions below and add any other comments you like. You may also want to send a copy to Statistics Canada or the Minister of Human Resources Development. The addresses are at the end of the questionnaire.

1.	In your view, should governments promote the use of more than one set of poverty lines in Canada?	___ Yes ___ No
2.	Are the low income cut-offs or LICOs of Statistics Canada a reasonable measure of poverty?	___ Yes ___ No
3.	Would you prefer a market basket poverty line rather than a "statistical" measure of poverty?	___ Yes ___ No
4.	Should all governments in Canada have "official" poverty lines?	___ Yes ___ No
5.	Should governments guarantee that no one in Canada falls below the poverty line, once a poverty line is officially recognized?	___ Yes ___ No
6.

A NEW POVERTY LINE YES, NO OR MAYBE?

A NEW POVERTY LINE
YES, NO OR MAYBE?