Improving Social Security in Canada
Guaranteed Annual Income: A Supplementary Paper

Government of Canada

NOTE: This document was retrieved from the Human Resources Development Canada SOCIAL SECURITY REFORM (SSR) website and copied here for information purposes. The SSR site is no longer active.

[ Go to Canadian Social Research Links Home Page ]


Social Security Reform Discussion Paper released

In Quebec City on September 18, 1994 the Prime Minister outlined four key components of the government's jobs and growth agenda:
- reforming social security;
- ensuring a healthy fiscal climate;
- reviewing government programs and priorities; and
- strengthening the performance of the Canadian economy in investment, innovation and trade.

The Discussion Paper, Improving Social Security in Canada, released to the public on October 5, 1994, provides Canadians with a framework for participating in the reform of our social security system. The paper takes a close look at why the social security system is not working for many Canadians and for the country as a whole. It sets a direction for change and offers a range of options for redesigning federal programs in the areas of working, learning and security. These programs include Unemployment Insurance, employment development services, child care and federal support for post-secondary education and social assistance.

Supplementary Papers provide analytical detail

This paper is part of a series of supplementary papers which are being released to provide Canadians with more detailed information about the current system and the options outlined in the Discussion Paper. This material is intended to provide a deeper understanding of the issues and to encourage more informed participation in the debate.

All contributions to the discussion are welcomed and encouraged. It is only with the participation of all Canadians that we can design an effective, fair, flexible and affordable system, which will respond to the needs of Canadians today and in the future.

Table of Contents

What is a Guaranteed Annual Income (GAI)?

The Negative Income Tax (NIT)
The Universal Demogrant (UD)
In Summary

Context for a GAI in the 1990s

A Changing World Means Different Challenges
Measures of adequacy
Considerations for a GAI Design
Universality, Targeting and Responsiveness
How Much of the Current System Could be Replaced/Reformed?
Federal - Provincial Considerations

GAI Approaches

Overview of Options
Option 1 - An "Adequate" Universal Demogrant (UD)
Option 2 - A Negative Income Tax
Options Results
- Low income
- Winners/Losers
Work Incentives
Horizontal Equity

Future Directions

Making Work Pay
Replacing Parts of Social Assistance
Earnings Supplementation
Newfoundland ISP Proposal

Appendix A - Historical Background

The First GAI - Speenhamland
The Canadian Context - Social Minimum Without a GAI
The Revival of Guaranteed Income Concepts in the 1960s
The Early Seventies - Failure to Implement a GAI
Late Seventies and Early Eighties - the Shift to Incrementalism
The Mid-Eighties to the Present - The GAI Revived

Appendix B - Measures of Adequacy

Relative Measures of Adequacy

Low Income Measure
CCSD Poverty Thresholds

Absolute Measures of Adequacy

USA Poverty Lines
The Sarlo Poverty Lines

Comparing Measures of Adequacy

Appendix C - Social Assistance Incomes

1: What is a Guaranteed Annual Income (GAI)?

A guaranteed annual income (GAI) is a perennial centrepiece in any discussion of social security reform. Modern versions of the idea originated about 30 years ago and still command much interest and attention today. It has been a dominant theme in social policy debates in Canada since then.

Different people mean different things when they speak of a GAI. Common to all concepts is a floor income provided on a continuing basis, which may vary depending on household configuration, age and other sources of income. Beyond this, however, a GAI depends on the purposes it is intended to serve. For some, the purpose of a GAI is to increase the incentive to work while reducing the cost and complexity of the current system. For others, the goal is to establish a non-stigmatizing entitlement to an adequate floor level of income, regardless of work effort, for all members of society. As will be demonstrated, such different and conflicting objectives cannot be met by the same income guarantee scheme.

In the federal discussion paper, Improving Social Security in Canada, the point was made that converting Canada's income security programs into a single GAI is not a practical idea. This paper provides an analysis for understanding the difficulties facing the implementation of a GAI.

Modern proposals for a GAI have usually taken two basic forms that reflect these very different purposes and an infinite number of variations. The form usually favoured by people who place a high value on simplification and work incentives is the negative income tax (NIT). This is a payment by governments to persons or households below certain income level as opposed to positive income taxes which are paid to governments by persons with income above a certain level. The NIT was initially conceived by the American economist George Stigler, in 1946 (George Stigler, "The Economics of Minimum Wage Legislation," American Economic Review (1946) pp. 358-365.), as another way of achieving the objectives of minimum wage legislation, and refined by another American economist, Milton Friedman, in 1962 (Milton Friedman, Capitalism and Freedom (Chicago: University of Chicago Press, 1962).), into a full-scale alternative to all social assistance and income support programs.

The second form of GAI is the universal demogrant (UD). This is a payment to all persons regardless of income. It is usually favoured by those who see the GAI as a right of citizenship and whose purpose is to eliminate poverty and lead to more equal sharing of the economic benefits of society. This approach to a GAI received its classic description by another American economist, Robert Theobald, in his 1965 book, Free Men and Free Markets. (Robert Theobald, Free Men and Free Markets (New York: Anchor Books, 1965).)

The Negative Income Tax (NIT)

An NIT consists of three elements: the guarantee, the reduction rate and the break even income level.

NIT Elements for a Family of Four

Guarantee: $15,000

Reduction Rate: 27 percent

Break even income level: $55,555

The guarantee is the maximum benefit level for each family. It varies by family size and configuration. In this example, the benefit level for a family of four with no source of income other than the NIT, is equal to the guarantee of $15,000. (This is for illustrative purposes only. A description of the reasoning behind the parameters is provided in sections 2 and 3 where the options are developed more fully.)

The reduction rate comes into play when a family has sources of income other than the NIT. The reduction rate of 27 percent means that for each dollar of income other than the NIT, the NIT benefit is reduced by 27 cents. For example, if a family of four earns income of $5,000, its benefit level is reduced by 0.27 times $5,000 or $1,350. Its total benefit would therefore be $13,650 (i.e., $15,000 minus $1,350). As a result of the additional work income of $5,000, the family's total income increases by $3,650 (i.e., from $15,000 to $18,650). The $18,650 is made up of $13,650 of NIT and $5,000 of earnings.

The break even income level is the maximum income level at which NIT benefits can be received. It too varies by family size and configuration. In other words, for families with non-NIT sources of income greater than the break even income level, the NIT benefit is zero. In this illustration the break even income level is $55,555. (In NIT designs where the reduction is applied to all income from sources other than the guarantee, the break even will always equal the guarantee divided by the reduction rate (i.e., $55,555 = $15,000 divided by 0.27). If a threshold level of other income is exempted from the reduction rate, this formula will not apply.)

Under this system, it is possible to identify three groups: those receiving full benefits, those receiving partial benefits and those receiving no benefits. Families with no sources of income other than the NIT receive the full benefit of $15,000 (i.e., the guarantee). At the other extreme, families with non-NIT income over $55,555 receive no benefit. In between those extremes, families with non-NIT income between $1 and $55,554 receive a benefit equal to the guarantee reduced by 27 cents for each dollar of non-NIT income. (The formula is: benefits equals guarantee minus (non-NIT income multiplied by reduction rate).)

Under the NIT system, there is always an incentive for beneficiaries to earn more income or acquire it from sources other than the NIT benefit. However, unless the guarantee is set at the poverty threshold, the NIT cannot eliminate poverty on its own. Typically, the NIT guarantee will be lower than the poverty threshold by design, in order to maintain an incentive to obtain other income. The NIT can, however, reduce the poverty rate and the depth of poverty if the basic guarantee is above the minimum level of current last-resort social assistance programs.

Crucial to the NIT anti-poverty strategy is the belief that improved work incentives will lead to a higher labour supply, subject to the availability of jobs, and result in lower levels of poverty and lower levels of public expenditure. An important element of the strategy is to ensure integration of the NIT with the personal income tax system. This is to ensure that NIT beneficiaries do not pay taxes so the combined effect of the reduction rate and the income tax rate does not create a work disincentive. However, this goal is extremely difficult to achieve since the NIT benefit is based on the family unit, and personal income taxes and payroll taxes are based on the individual. Nonetheless, even if perfect integration is an elusive goal, keeping the sum of the NIT reduction rate and the personal income tax marginal rate, i.e., the effective marginal tax rate, as low as possible is an important consideration for the NIT.

The Universal Demogrant (UD)

Under a UD approach, all adult citizens would receive a tax-free cheque from the government adequate for their needs and those of their family. Income from all other sources would be taxed (rates would have to be set at levels sufficient to pay for the UD). Total disposable income would equal the UD plus after-tax income from other sources. Taxes on other income could either be levied at a flat rate or scaled to rise as incomes increase as under Canada's current progressive tax system.

For example, a UD could provide a non-taxable benefit of $20,000 to a family of four and would tax income from other sources at a flat tax rate of 50 percent. (The UD option in Section 3 has a progressive tax structure with three rates. This single rate structure is chosen here to simplify the illustration.) Under such a proposal, a family with earnings of $30,000 would receive a disposable income equal to the guarantee plus half their income from other sources or $35,000 (i.e., $20,000 plus $15,000).

In contrast to the NIT approach, which would provide net benefits only to households up to the break even income level, the UD provides benefits to households all the way up the income scale. The UD has no reduction rate or break even income level. One consequence is that UD expenditures are much higher than NIT expenditures. Another consequence is that the level of taxation must be increased to pay for the higher level of expenditure.

In the view of some economists, a GAI set at poverty thresholds and with a 100 percent tax rate on other income up to those thresholds would be the cheapest way to eliminate poverty. (e.g., J.E. Meade, The Structure and Reform of direct Taxation (London: Institute of Fiscal Studies, George Allen and Urwin, 1978).) However, it would remove any financial incentive to take a job paying less than the poverty thresholds. Moreover, the higher the tax rate on other income, the less the incentive to take a job paying above but close to the threshold levels. After-tax incomes would not be significantly higher, compared to the guarantee, until earnings were considerably above the threshold levels.

In Summary

There are two major philosophical orientations around the concept of the GAI. One orientation focuses on GAI as a mechanism to facilitate adjustment to economic change. It is concerned with a simplified and coherent program structure, work incentives and expenditure control. The second orientation focuses on antipoverty objectives. It is concerned mostly with the redistribution of income and the creation of entitlements based on citizenship.

With two such fundamentally different concepts, many of the objectives trade off against one another. Nonetheless, that does not mean that the models cannot be combined. In fact, the GAI proposals that have been made in Canada differ in their emphasis on the importance of the different objectives. The historical review (Appendix A) illustrates the balance struck over conflicting objectives in different proposals and concludes by identifying the large number of often conflicting objectives that are part of the GAI debate.

Section 2 provides the context for thinking about a GAI in the 1990s and sets the stage for the GAI designs developed in Section 3. These approaches are not meant to be options in the usual sense of the word, but rather generic illustrations of the two main concepts of the GAI outlined above. The purpose is to demonstrate clearly the strengths and weaknesses of each approach.

Section 4 concludes the paper by examining the role GAI approaches might play in social security reform.

2: Context for a GAI in the 1990s

A Changing World Means Different Challenges

In the years since the heyday of comprehensive GAI proposals in the early 1970s (see Appendix A), the environment for considering a GAI has changed considerably. Between 1943 and 1971, government programs and services proliferated to provide much of the infrastructure that Leonard Marsh projected as necessary to achieve his 1943 vision of a social minimum. Unemployment and health insurance programs were put in place. Social assistance programs were made available to all persons in need, and children's benefits were introduced to assist parents with the costs of raising their children. A guaranteed minimum income for seniors was created through a combination of pensions (Old Age Security and Canada/Quebec pension plans) and income-tested assistance (the Guaranteed Income Supplement and provincial supplements for seniors).

Another change has been the declining importance of employment income as a means of support paradoxically, at the same time the employed proportion of the population was increasing. From 1945 to the early 1970s, unemployment rates were generally low, and real earnings were rising. Most low-income families headed by persons under the age of 65 received most of their income from earnings. Poverty was more widespread than today, but dependence on government transfer payments was also lower. Since that time, higher rates of unemployment, the broader availability of, and decreasing stigma attached to receiving, government transfer payments and the growth of groups who face considerable difficulty earning adequate incomes from paid work, such as lone-parent families, have changed that picture.

In such a world, concerns about providing further disincentives to work by introducing a non-stigmatizing guaranteed income, particularly if it provides a more adequate income than current social assistance benefits, have become increasingly widespread.

At the same time, the fiscal situation of governments has worsened. Until to the mid 1970s, the federal budget was usually balanced over the business cycle, and the debt to GDP ratio was steadily falling from the very high levels reached during World War II. It was possible, because of high, and only briefly interrupted, economic growth, to introduce costly new programs, balance the budget and believe that the resources to introduce and enrich government programs would continue to increase. That, too, can no longer be assumed.

Measures of Adequacy

The concept of adequacy has also changed over time. What is considered adequate today must meet a more generous standard than would have been applied in the past. This is because, over long periods of time, the concept of adequacy is relative rather than absolute. (For example, we look back on 1967 as a year of prosperity and virtually full employment (the unemployment rate was 3.8 percent). Yet in that year, (using the 1978 low income cut-offs as a measure of poverty) the child poverty rate was over 40 percent - more than double the rate in 1992 when the unemployment rate was 11.3 percent and, as Table 1 indicates, the standard of adequacy was somewhat more generous. See: M.C. Wolfson and J.M. Evans, Statistics Canada's Low Income Cutoffs: Methodological Concerns and Possibilities (Ottawa: December 1989) p. 59.) This relativity is reflected in Statistics Canada's low income cut-offs (LICOs). The LICOs are not official poverty lines, but are often used as such by organizations such as the National Council of Welfare. The LICOs are based on a periodic re-evaluation of the overall proportion of income spent on food, shelter and clothing by the average family. This proportion has declined over time from 50 percent in 1959 to 34.7 percent in 1992, and the result is a poverty line that has increased by 62 percent in real terms over the last 30 years. The low income cut-offs are indexed to annual inflation.

TABLE 1: Proportion of Average Family Income Spent on Food, Shelter and Clothing

LICO Level is for a family of four in 1993 dollars, urban centres of 500,000 plus

Proportion - 50 percent
LICO Level - $19,123

Proportion - 42 percent
LICO Level - $24,870

Proportion - 38.5 percent
LICO Level - $28,243

Proportion - 36.2 percent
LICO Level - $30,645

Proportion - 34.7 percent
LICO Level - $31,007

Source: Statistics Canada, Income distributions by size in Canada, (Cat. 13-207 various issues)

Other measures of poverty, both relative and absolute, are also available for Canada.
Some of these measures are summarized in Appendix B.

Considerations for a GAI Design

Universality, Targeting and Responsiveness

A GAI delivered as a UD would entitle every individual and family to a benefit regardless of income. The benefit could vary by age, family size, family type or other demographic criteria. It would be non-taxable. The net additional cost of introducing a GAI (beyond the offset provided by the cost of the programs it would replace) would be recovered through the personal income tax system. (This is assumed for analytical simplicity. Costs could also be offset by increasing other taxes or reducing other expenditures.) The more progressive that system, the more targeted the net result.

In effect, the demogrant delivers benefits universally, and then tries to target the needy by adjusting personal income tax rates. The advantages are the elimination of stigma and the simplification of the system (if other income support and tax credits are eliminated) in terms of administration. However, the disadvantages are a much higher level of government taxation, negative work, savings and investment incentive effects, and the inappropriateness of sending out large numbers of cheques to middle and upper income families who then repay the benefits through higher taxes.

Since the social assistance system would be eliminated or significantly reduced under both the UD and the NIT (further discussion below), it is critical that the GAI design be responsive to changes in income and family structure. The demogrant is uniquely placed to achieve this since the maximum benefit is always delivered to the entire population, and the design can be made invariant to family type (further discussion below). The result would be significant administrative savings.

Because a GAI delivered through an NIT mechanism provides benefits only to those with incomes below the break even income level, it pays out less than a UD of equal generosity and recovers less by tax increases or spending reductions. However, this means that, to be responsive to fluctuations in disposable incomes during the year, a negative income tax system must be able to keep track of income fluctuations on a family basis so benefits are delivered when needed, and overpayments and underpayments minimized. (The Guaranteed Income Supplement to the Old Age Pension addresses the problem of responsiveness by allowing persons the option of applying for benefits on the basis of their anticipated income in the current year rather then their actual income in the previous year as reported on their income tax return. However, the anticipated income of seniors is generally far more predictable than the anticipated income of the working-age population. Consequently, the risk of overpayments and underpayments is much smaller.) Also, since a family income test is applied, the administrative structure must be responsive to changes in family structure. To do this, high administrative costs would have to be incurred.

The alternative is to pay the benefit at a uniform monthly rate based on the previous year's net family income as reported on the tax return, a method now used with the Child Tax Benefit. This system, however, is possible only when highly responsive social assistance payments continue to operate. If social assistance is to be eliminated, the GAI must be responsive.

How Much of the Current System Could be Replaced/Reformed?

GAI proposals advanced by various proponents at various times in Canada would replace all, or part, of the following types of programs.

Social assistance payments, such as the existing provincial/territorial and municipal social assistance systems, are currently provided to persons and families subject to a needs test of their income and assets. Most Canadian GAI proposals have suggested withdrawing federal funding for these programs under the Canada Assistance Plan. The 1971 Special Senate Committee on Poverty also proposed a reimbursement to provinces and territories for the full costs of social assistance payments to those not covered by its GAI. (See Appendix A.)

Unemployment Insurance provides benefits to eligible insured persons. GAI proposals, such as those of the Special Senate Committee on Poverty, the Macdonald Commission and the Wolfson Guaranteed Income/Simplified Tax have included changes to Unemployment Insurance to make it more of an "insurance" program by increasing entrance requirements, reducing benefit levels and/or durations, and removing non-insurance features such as regionally extended benefits. (See Appendix A.)

Demogrant payments are made to members of a population group such as parents of children or the elderly (such as Family Allowances before 1993 and Old Age Security).

Negative income tax benefits are provided to persons and families subject to an income test as with the Child Tax Benefit, the Guaranteed Income Supplement, Spouse's Allowance benefits and the Goods and Services Tax Credit.

Income tax credits and deductions are based on demographic characteristics such as the personal and married credits, the equivalent to married credit, the age credit and the former child credit.

Federal - Provincial Considerations

Any GAI would require fundamental changes to federal - provincial fiscal arrangements. Basic federal and provincial/territorial roles with respect to financing and program delivery would have to be thoroughly reexamined, and new arrangements negotiated. Achieving consensus would be very difficult.

GAI Approaches

Overview of Options

In this section, two illustrative GAI options are developed and analyzed:

- Option 1: An "adequate" universal demogrant (UD);

- Option 2: A negative income tax (NIT).

Both GAI options are directed to the working-age population. They do not change the system of benefits directed to persons 65 and over, since that system forms part of a separate policy review of seniors issues announced in the 1994 Budget and since a GAI for seniors already exists. However, refundable and non-refundable credits in the federal personal income tax system based on demographic characteristics would change under the GAI approaches presented here. Seniors would be affected by these changes to the personal income tax system. But for the purposes of illustration, seniors are assumed to continue to be subject to the existing tax system under both scenarios, even though such a scheme could not be implemented in practice.

Federal cost-sharing of income support benefits provided by provincial/territorial and municipal social assistance programs under the Canada Assistance Plan would cease under both GAI designs. In the UD option, the GAI would also replace provincial/ territorial and municipal social assistance in most cases (cash assistance only). In the NIT option, provincial/territorial and municipal social assistance would continue to operate, but a much smaller program would be required. In both cases, cost-sharing of social services under CAP would continue.

Since both options eliminate some personal credits, provincial tax revenues are increased. In the UD option, where the GAI replaces social assistance entirely, the additional provincial/territorial revenues are put toward the cost of the GAI. In NIT option, the additional provincial/territorial revenues are returned to taxpayers through lower personal income tax rates.

Both GAI designs would retain the Unemployment Insurance program, but entrance requirements would rise from 12-20 weeks of employment in the previous year to 26 weeks. The maximum duration of benefits would also be 26 weeks. Benefit rates would decline 15 percentage points to 40 percent of insurable earnings for persons earning more than one half of the maximum insurable earnings or persons without dependents, and decline to 45 percent for persons earning less than half the maximum insurable earnings and with dependents. Special benefits, such as those for sickness, adoption, maternity and parental purposes, would continue under current parameters as would Unemployment Insurance Developmental Uses programs.

With respect to fiscal implications, the UD requires higher expenditure levels than provided for in the fiscal parameters laid out in the February 1994 Budget and in Improving Social Security in Canada. Personal income tax rates are increased to pay the additional costs so the result is deficit neutral (i.e., there is no increase to the combined federal and provincial/territorial deficits). This tax increase enables a better illustration of the essential trade offs between cost, adequacy and work incentives under a design based on redistributing income and providing an adequate floor level of income support to all persons, regardless of their work pattern.

However, since proposals based on an NIT design traditionally have a greater concern for fiscal implications and work incentives than for adequacy, the second design is constrained to be more consistent with the fiscal parameters established in the February 1994 Budget and in Improving Social Security in Canada. The NIT does not result in a higher deficit level than the framework outlined in those documents and, to the extent possible, respects the expenditure and tax level in that framework.

In short, the first design will illustrate the cost of an "adequate" GAI while the second will illustrate what kind of GAI can be provided within the government's current fiscal framework. Both of the following GAI designs are illustrative and are put forward to provide information on the trade offs and issues involved. They are not intended as specific proposals.

Option 1 - An "Adequate" Universal Demogrant (UD)

The UD would provide the following income guarantees:

Universal Demogrant

All adults 18 to 64 - $7,000

First dependent child less than 18 in a lone-parent family - $7,000

All other dependent children less than 18 - $3,000

For example, a household consisting of a lone parent and one child under age 18 would receive $14,000, while a married or common-law couple with two children under age 18 would receive $20,000. Benefits would not be taxable.

As described previously, the Unemployment Insurance system would be significantly returned, the basic personal credit and married and equivalent to married credits in the personal income tax system would be abolished, as would the refundable Child Tax Benefit and the GST Credit.

While falling short of the standard of adequacy represented by the LICOs this design would provide an income for all families with children headed by persons under age 65. It would be above current social assistance levels in all provinces and territories except Ontario and would be over 70 percent of the commonly used measures of adequacy for such families. (It falls short, however, of the adequacy levels provided by social assistance for single disabled persons in all provinces and territories.) Because it falls short of full adequacy, some or all provinces and territories might want to "top-up" benefits through an income or needs-tested program similar to those now provided to top-up the OAS/GIS system for seniors. Aside from such small residual programs, provinces and territories would cease to make cash social assistance payments, and the federal government would cease its contributions to such programs under the Canada Assistance Plan. Provincial/territorial social assistance savings are used to pay for the UD.

Option 2 - A Negative Income Tax (NIT)

The NIT design would provide the following income guarantees:

Negative Income Tax

All adults 18 to 64 - $4,500

First dependent child less than 18 in a lone-parent family - $4,500

All other dependent children less than 18 - $3,000

For example, a household consisting of a lone parent and one child under age 18 would receive $9,000, while a married or common-law couple with two children under age 18 would receive $15,000. These guarantees represent less than one third of LICOs for a single adult and between 50 and 56 percent of the cut offs for families of two or more persons. They are also below current social assistance incomes in almost every province and territory. (See Appendix C.)

Benefits would be reduced at a rate of 15 percent for the first adult in the family, by six percent for the second adult or first dependent child under age 18 in a lone-parent family and by three percent for each of the first two children under age 18 in all other families. These reduction rates would be stacked. That is, a household consisting of a lone parent and one child under age 18 would have its benefits taxed back at a rate of 21 percent on the parent's income, (i.e., 15 percent plus 6 percent) while a married or common-law couple with two children under age 18 would have their benefits taxed back at a rate of 27 percent (i.e., 15 plus 6 plus 3 plus 3 equals 27 percent). The reduction rate would not exceed 27 percent, even if there were more than two dependent children in the family.

Under this design, the break even levels of income, i.e., the level of other income where no GAI benefit would be received, would be as follows for common household configurations:

NIT Parameters

One adult
Benefit Level - $4,500
Reduction Rate - 15 percent
Break even Income Levels - $30,000

One adult-one child
Benefit Level - $9,000
Reduction Rate - 21 percent
Break even Income Levels - $42,857

Two adults
Benefit Level - $9,000
Reduction Rate - 21 percent
Break even Income Levels - $42,857

Two adult-one child
Benefit Level - $12,000
Reduction Rate - 24 percent
Break even Income Levels - $50,000

One adult-two children
Benefit Level - $12,000
Reduction Rate - 24 percent
Break even Income Levels - $50,000

Two adults-two children
Benefit Level - $15,000
Reduction Rate - 27 percent
Break even Income Levels - $55,555

The benefits would be offset by all the program and tax changes outlined for the UD. The only exception is provincial/territorial social assistance. Since the NIT provides lower guarantee levels than the UD, a second income support tier is needed to enhance the adequacy of the benefits provided. In other words, a significantly reduced social assistance system would continue to exist, wholly framed and delivered by provinces and territories without federal cost-sharing, although the social assistance benefits provided would be much lower than they are currently and the number of people needing the system would be much smaller.

This design does not add to the deficit presented in the February 1994 Budget. It focuses on enhancing financial incentives to work, facilitating economic adjustment and simplifying the income security system for those under age 65.

Options Results (Source: HRDC analysis using Statistics Canada's Social Policy Simulation Database and Model (SPSD/M). These results do not account for behavioural effects (such as changes in work effort) or macroeconomic impacts.

Low Income

The analysis of the impact of both options on the incidence of low income is based on Statistics Canada's after-tax LICOs in order to capture properly the effects of changes to personal income tax. (The abolition of the basic personal credit, the married credit and the equivalent to married credit will increase taxes paid by all taxpayers, while enabling the level of the income guarantees to be higher than they would otherwise have been. Analyzing the effect of these changes on the basis of pre-tax LICOs would significantly overstate the benefits accruing to individuals and families in terms of disposable (i.e., post-income-tax) income.) On this basis, before the implementation of a GAI, 12.8 percent of families and unattached individuals had after-tax incomes below the Statistics Canada after-tax LICOs. After the implementation of the UD, the incidence of low income would decline to 9.7 percent. The implementation of the NIT would reduce the incidence of low income to 11.4 percent. (See Table 2.)

TABLE 2: Distributional Impact on Low Income

Rate of low income
Pre-GAI - 12.8 percent
Universal Demogrant - 9.7 percent
Negative Income Tax - 11.4 percent

Depth of low income
Pre-GAI - $6.3 billion
Universal Demogrant - $3.5 billion
Negative Income Tax - $4.2 billion

Source: HRDC Analysis using SPSD/M.

Because of the large reductions in Unemployment Insurance benefits in both models and because, in some cases, the provided guarantee levels fall short of those benefits now available through refundable income tax credits and provincial/territorial and municipal social assistance programs, both designs result in some households moving into poverty on an after-income-tax basis, and others climbing above the after-tax low income thresholds. In both cases, however, the latter group outnumbers the former, so the incidence of low income is reduced. Some or all of those who fall below the low-income thresholds might be raised back above them by provincial/territorial and municipal income or means-tested top-up programs which would supplement the GAI benefit. (No attempt has been made to estimate the effect of such top-ups on the rate or depth of low income.)

The depth of low income measures the difference between the low-income thresholds and the incomes of low-income families and unattached individuals. Before the implementation of the GAI options, the aggregate depth of low income was $6.3 billion. The UD reduces the depth of low income by 45 percent to $3.5 billion. The NIT reduces it by 33 percent to $4.2 billion. (See Table 2.)


In this section, winners, losers and neutrals are analyzed according to disposable income, province/territory and economic family type for each design. (For the purposes of this paper, families and unattached individuals are designated as winners as a result of the implementation of a GAI design if their annual post-income-tax income increases by $250 or more. They are designated as losers if, as a result of the implementation of a GAI design, their post-income-tax income declines by $250 or more. If, as a result of the implementation of a GAI design, post-income-tax income increases or declines by less than $250, the change is designated as neutral.)

Under the NIT, 42 percent of families and unattached individuals are winners, 51 percent are losers and 7 percent are neutral. Under the UD design, 59 percent of families and unattached individuals are winners, 37 percent are losers and four percent are neutral. (See Table 3.)

TABLE 3: Distribution of Winners/Losers By Income Category

Disposable Family Income of $0 to $20,000

Universal Demogrant
Winners - 77 percent
Losers - 20 percent
Neutral - 3 percent

Negative Income
Winners - 70 percent
Losers - 20 percent
Neutral - 10 percent

Disposable Family Income of $20,000 - $50,000

Universal Demogrant
Winners - 65 percent
Losers - 30 percent
Neutral - 5 percent

Negative Income

Winners - 41 percent
Losers - 52 percent
Neutral - 7 percent

Disposable Family Income of $50,000 and higher

Universal Demogrant
Winners - 32 percent
Losers - 65 percent
Neutral - 3 percent

Negative Income
Winners - 19 percent
Losers - 77 percent
Neutral - 4 percent

Total Disposable Family Income

Universal Demogrant
Winners - 59 percent
Losers - 37 percent
Neutral - 4 percent

Negative Income
Winners - 42 percent
Losers - 51 percent
Neutral - 7 percent

Source: HRDC Analysis using SPSD/M.

Disposable Income

Under the NIT design, winners outnumber losers in the $0 to $20,000 disposable income group. (Disposable income is defined as all sources of income (except the GAI) less the payment of federal and provincial/territorial income taxes and payroll taxes such as Unemployment Insurance premiums and Canada/Quebec Pension Plan contributions.) Losers outnumber winners in the middle and upper income groups. (See Table 3.) Under the UD design, winners outnumber losers for both the low and middle disposable income groups.

In either option, there are a significant number of losers with family incomes below $20,000 because, for many, the GAI does not offset UI losses. In the middle income range, i.e., $20,000 to $50,000, there are large numbers of losers because the tax increases (through rate increases or losses of the personal credits) necessary to finance the GAI are larger than any GAI benefits they might receive.

Under the NIT, families and unattached individuals who are winners gain an average of $2,500 in additional income. Families and unattached individuals who are losers lose smaller amounts than winners gain. Under the UD design, the biggest winners are in the middle income range with large gains occurring also in the high income range (mostly large families with children who do not currently use UI). Losses for families and unattached individuals depend on income with average losses exceeding $14,000 in the upper category.

TABLE 4: Size of Gain and Loss by Income Category

Disposable Family Income: Family here refers to economic family which may be either an unattached individual or a unit of two or more related persons.

Disposable Family Income of $0 to $20,000

Universal Demogrant
Winners - plus $4,490
Losers - minus $3,340

Negative Income Tax
Winners - plus $2,680
Losers - minus $2,130

Disposable Family Income of $20,000 to $50,000

Universal Demogrant
Winners - plus $5,710
Losers - minus $4,380

Negative Income Tax
Winners - plus $2,870
Losers - minus $2,030

Disposable Family Income of $50,000 and higher

Universal Demogrant
Winners - plus $4,990
Losers - minus $14,050

Negative Income Tax
Winners - plus $2,390
Losers - minus $2,640

Source: HRDC Analysis using SPSD/M.

Economic Family Type

Under the NIT, winners outnumber losers for all economic family types except unattached individuals and couples with no children under 18. On average, couples with children under 18 have the largest gains followed by lone parents with children under 18. Under the UD design, winners outnumber losers among all economic family types except unattached individuals. In both designs, families with children fare better than families without children under 18. (See Table 5.)

TABLE 5: Distribution of Winners/Losers By Family Type

Family Type: Family here refers to economic family which may be either an unattached individual or a family of two or more persons.

Family Type: Families without children

Universal Demogrant
Winners - 50 percent
Losers - 45 percent
Neutral - 5 percent

Negative Income

Winners - 37 percent
Losers - 55 percent
Neutral - 8 percent

Family Type: Families with children

Universal Demogrant
Winners - 73 percent
Losers - 25 percent
Neutral - 2 percent

Negative Income
Winners - 50 percent
Losers - 45 percent
Neutral - 5 percent

Source: HRDC Analysis using SPSD/M.

Work Incentives

The effects of either of the two GAI models on the amount of work supplied cannot be predicted with any precision. While experiments have been conducted in the United States and Canada, those participating knew that their benefits were not permanent and, consequently, they were not likely to change their behaviour as much or in the same manner had the GAI been ongoing. As a result, total hours worked fell by about five percent on average. The work reduction was largest for second earners in two-earner households and weakest for the main earner. Further, the negative work effect was higher the more generous the benefit level.

The GAIs discussed here are very different from those of the experiments because of the inclusion of a UI reform based on a purer social insurance model. This would improve the overall impact on work incentives (other than for part-year work). Whether this is enough to offset the negative effect of the GAI itself is not known.

The main aspect of the NIT and UD design that would help improve work incentives is the removal of high reduction rates for social assistance. Generally effective marginal tax rates for the social assistance population would drop from about 75 to 80 percent to around 50 percent for the NIT and 40 percent for the UD. In the case of the NIT, the lower rates would only occur if the provincial/territorial social assistance program top-up does not add yet another level of reduction rates to the system. (Since there is no family income testing in the UD option, work incentives could be significantly improved for a low income spouse or common-law partner since incomes are not added together for the purposes of calculating the benefit or for taxation.)

For the NIT, this improvement to incentives would be offset by three features. One is the higher benefits available to working poor and low income families at their current level of work effort. The second is the higher reduction rate on the NIT than on the Child Tax Benefit and the GST Credit for the middle income group (up to 27 percent compared to about 10 percent), and the third is a lower personal income tax threshold for all levels.

For the UD, the improvement of work incentives for social assistance recipients is offset by higher benefit levels at current levels of work effort, a lower personal income tax threshold and the sharp increase in the marginal rates of the personal income tax system (e.g., from 25 percent to 40 percent in the low bracket, 39 percent to 64 percent in the middle bracket and 45 percent to 72 percent in the highest bracket).

Horizontal Equity

With horizontal equity, the benefits received by families in different circumstances, but at similar levels of income, should reflect those different circumstances. Both proposals have major horizontal impacts.

The UD provides more benefits to families with children at every level of income, an effect similar to the Family Allowance program before the clawback, but at a much higher benefit level. The NIT on the other hand, reduces the income level at which families with children can receive benefits by about $10,000 (e.g., for a family with two children, the maximum income level under the Child Tax Benefit is about $65,000 and under the NIT, about $55,000).

Generally, except in the case of children, the UD benefits do not vary by family type. Since all adults receive the same benefit level and since children over 18 who are living with their parents receive adult benefits and the personal credits are eliminated, there is no difference in benefit levels for different family configurations. For example, single- and dual-earner families receive the same benefits at the same income level (unlike the current system where single earners claim a personal and a married income tax credit, and dual earners get more by claiming two personal income tax credits).

However, since the net impact of the UD depends on the combined effect of the transfer plus the increase in personal income taxes and since personal income taxes are individually based, the net benefits to any family will depend on the distribution of income between spouses. Whether this is desirable or not depends on one's point of view. (For example one consequence of family type neutrality is the equity issue known as "The Banker's Wife" problem, where a non-working spouse of a high income individual receives a maximum benefit.)

With respect to the NIT, there is also more neutrality to family type (for example single- and dual-earner families are treated the same) but, unlike the UD, there is a family-based income test. This means that two people reporting as a couple must add their incomes together to claim their benefits while two people reporting as singles do not, resulting in more favourable treatment.


The cost of the UD program is $146 billion. The reallocation of existing expenditures and refundable tax credits ($30.2 billion including the provincial/territorial expenditure on social assistance) and the elimination of the personal credits ($22.8 billion, including both federal and provincial/territorial effects) fall far short of financing the program. As a result, an additional $93 billion is required in personal income tax revenue through increased rates. (See Table 6.)

The cost of the NIT is more modest (i.e., $37.3 billion) and is financed entirely from the reallocation of expenditures and refundable credits ($21.1 billion, which does not include provincial/territorial social assistance as does the UD) and the elimination of the basic, married and equivalent to married credits (not including the increases in provincial/territorial tax revenues which are assumed to be returned to taxpayers through lower tax rates).

TABLE 6: Fiscal Impact ($ Billion)

GAI program cost

Universal Demogrant - $146.1 billion
Negative Income Tax - $37.3 billion

UI, CTB, CAP, GST tax credit reallocation
Universal Demogrant - $30.2 billion
Negative Income Tax - $21.1 billion

Elimination of personal credits
Universal Demogrant - $22.8 billion
Negative Income Tax - $16.2 billion

Increased tax rate
Universal Demogrant - $93.1 billion
Negative Income Tax - $0 billion

Source: HRDC Analysis using SPSD/M.


While the UD design yields favourable results in terms of winners and losers and substantially reduces poverty, it requires enormous increases in federal and provincial/territorial income taxes and in overall expenditure levels.

The NIT design produces far more losers than winners, particularly among families without children, and only slightly reduces the incidence of low income. However, it is affordable without additional tax rate increases and significantly reduces the depth of poverty.

Both options would probably have net negative effects on labour supply except for current social assistance recipients.

The results seem disappointing. The significant disruption associated with implementing any of these designs would make it difficult to advance the GAI as the centrepiece of social security reform.

The bottom line is that both these models are too expensive. In the UD approach, the additional cost is clear: $93 billion in additional personal income taxes is required. In the NIT, the expense is less clear because the option was designed to fit existing fiscal parameters. The result was a large number of losing families, many at low income levels. Fixing this problem would require billions of dollars, although not on the scale required for the UD. For example, an increase to the adult benefit of $500 would cost an additional $6 billion under the NIT.

Why then are GAIs so expensive? A GAI, by its very nature, and in the context of the current Canadian income distribution, provides more benefits to more people than existing systems of social support. To provide a reasonable level of income support, i.e., an adequate guarantee, and a low reduction rate to ensure an incentive to work, most working poor and middle income families become eligible to receive the benefits. The lower the reduction rate on the GAI benefit, the higher up the income scale the benefit goes and the more people are included. At the limit, the reduction rate falls to zero, the entire population benefits and costs soar, which is what happens under the UD model. Moreover, unlike the programs it replaces, a GAI is non-stigmatizing and easier to understand. Families who are eligible for existing benefits but either do not realize it or do not want to apply, are assumed to benefit from the GAI. A final reason is that the GAI provides coverage to groups currently not covered by many social programs such as adult children living with their parents and the self-employed.

The question is how to pay for these additional benefits. The NIT tried shuffling income among different parts of the population. The UD resorted to large tax increases. The consequences of both approaches have been documented. The trade-offs are very strict: no "surplus" funds arise as the result of the consolidation and simplification of programs that could be used to improve the system. What could the sources of this "surplus" be? Any excessively high benefits that some may have in the existing system from dipping into several programs at the same time are eliminated and this does not provide a significant source of new funds. Administrative savings could be possible in the UD design due to the elimination of social assistance but this would be very small compared to the cost of the program. As for the NIT, running a responsive benefit for a much larger population could actually result in increased administrative costs. The main potential source of "surplus" would be an increase in labour supply, leading to both lower poverty levels and lower benefit payouts. But, as discussed in the previous section, even if labour supply might increase for particular groups, the overall effect on labour supply is more likely to be negative than positive.

These results must also be viewed in the context of what a GAI does not do and what it could prevent if it absorbs the financial resources and overextends the capacity of the population to accept major change, i.e., measures to expand economic growth and employment, to increase the employability of persons with inadequate earnings, to provide required supports for persons with disabilities to function in society and in the labour force, to promote a better balance between work and family responsibilities or to fund preventive measures in areas such as child development and school-to-work transitions.

However, this does not mean that for more limited purposes, GAI mechanisms may not have value. The final section of the paper examines recent initiatives and proposals that merit further attention.

Future Directions

If a GAI, whether based on an anti-poverty or an economic adjustment vision, appears to offer too much disruption and insufficient benefits, is there still a useful role for GAI mechanisms in social security reform?

Before examining this issue, it is important to remember that some of the ideas underlying a GAI and the mechanisms for implementing it have already had a major impact on Canadian income security programs. As noted earlier, the level of benefits and the taxback structure of the current Ontario social assistance system are remarkably similar to those in the 1971 Special Senate Committee on Poverty proposal. The Child Tax Benefit and the GST Credit are significant negative income tax programs - the first restricted to families with dependent children and the second available to all adult tax filers. And, of course, the combination of the Old Age Pension, the Guaranteed Income Supplement and provincial/ territorial supplements for seniors provide a GAI for Canadians over age 65.

Provincial/territorial and municipal social assistance programs, combined with the GST Credit, the Child Tax Benefit and certain provincial/territorial benefits available to low income persons do provide minimum "last resort" incomes to persons under age 65. However, benefit levels vary widely, work disincentives remain substantial, a stigma is still attached to the receipt of social assistance, and there is a lack of coherence between social assistance and programs such as Unemployment Insurance. Further, the depth and incidence of poverty in Canada remains too high whatever the definition used.

Making Work Pay

A central dilemma facing a GAI continues to be the old principle of "lesser eligibility". How generous a guaranteed income can the state provide compared to what is available for doing low wage work?

The stronger the demand for labour at wages adequate for the needs of individuals and families, the less the risk of providing more generous income guarantees. However, even in tight labour markets, as in Ontario in the late 1980s, higher real social assistance benefits helped keep caseload levels from falling as those benefits approached or exceeded entry-level wages available to such groups as lone parents with young children.

Lower reduction rates on earned income, often advanced as a way to reduce work disincentives for people on social assistance, can instead lead some people into a situation where they supplement social assistance with part-time or part-year work or vice versa instead of moving off dependence on social assistance into year-round full-time employment and progressing up the job and wage ladder from that position. Further, lower reduction rates can lead to rising caseloads as the program covers a greater proportion of the population.

There appears to be a need for some mechanism to encourage people now unemployed and/or on social assistance to take entry-level, low-paying jobs. Encouraging this first step on the wage ladder by providing a living standard at least equal to what people would have been receiving on social assistance, is considered a major objective of policy.

This has led to the development of strategies to provide support and supplementation to people in low-wage jobs relative to their needs and those of their families, i.e., to make work pay. As with the two visions of a GAI outlined in this paper, fighting poverty while reducing the work disincentives embodied in public programs remain important goals. But to these goals has been added a concern to improve the conditions surrounding low-wage work in order to make working more attractive, i.e., making work pay.

Recent strategies to achieve this goal vary widely. Some are not yet off the drawing board, others are in the experimental phase and still others are operating on a provincial/territorial and national scale in Canada and the United States. While those discussed below involve various forms of income support based on NIT-type mechanisms, one important strategy to make work pay is the provision of ample opportunities for learning, education and training. Increasing the human capital of persons stuck in low wage jobs has a double effect. It enables them to compete for more skilled and better-paying jobs and it opens up their former positions to persons who previously were unemployed or working on only a sporadic basis.

Replacing Parts of Social Assistance

Short of replacing social assistance with a comprehensive GAI, some policy makers have proposed replacing social assistance benefits for specific groups with NIT-type programs which would also be available to low income working persons not receiving social assistance. In 1988, the Ontario Social Assistance Reform Commission (SARC) recommended that national income-tested programs be created for persons with disabilities and parents of dependent children which would replace social assistance benefits for these groups.

In 1993, as part of a general reform of Ontario social assistance, that province's Ministry of Community, Family and Children's Services proposed to replace the children's component of social assistance benefits with an income-tested benefit for all low income parents of dependent children, not just those receiving social assistance. (Government of Ontario, Ministry of Community, Family and Children's Services, Turning Point, 1993, pp. 16-18.) This would enable the province to link adult social assistance benefits to what could be earned at a full-time, year-round minimum wage job, avoiding the current situation where total adult and child social assistance benefits significantly exceed this level of earnings.

Because many more families would qualify for benefits under this proposed Ontario Child Income Program (OCIP) and since it was not eligible for federal cost-sharing under current Canada Assistance Plan rules, (it was an income-tested rather than needs-tested benefit) the additional cost of OCIP made it impossible for the Ontario government to implement the proposal.

However, it would have had a significant positive impact on making work pay. Earnings would exceed social assistance benefits at much lower wage levels for families with children and, instead of being taxed back at a rate of 75 percent on net earnings, children's benefits would be reduced at a much lower rate and be received over a broader income range, substantially reducing the financial disincentive to earn. The means of achieving this two-pronged attack would be a provincial income supplement for families with children designed, like the federal Child Tax Benefit, as a form of NIT. The idea of replacing social assistance benefits paid on behalf of children by an integrated federal - provincial child benefit is discussed further in the background paper, Income Security for Children.

Earnings Supplementation

The Government of Canada, the federal government in the United States and the Quebec government have taken a more direct approach to making work pay by directly supplementing the earnings of the low income working population.

In Canada, a Work Income Supplement of up to $500 is added to the Child Tax Benefit for families earning between $3,750 and $25,921 annually. This supplement is designed as a modest step toward making work pay. About 700,000 families qualify for the supplement.

In the United States, such supplements are paid through the mechanism of an Earned Income Tax Credit delivered through the personal income tax system. It is the goal of the U.S. government to enrich the credits to the point where, added to year-round full-time work paying the federal minimum wage, they would exceed the official American poverty lines for families of four or fewer persons. The credit is available to single adults and childless couples as well as to parents with dependent children and represents varying percentages of family earned income (up to a maximum) depending on the presence and number of children up to two. It is taxed back based on family income.

Quebec has a somewhat different version of earnings supplementation under its Parental Wage Assistance (PWA) program which is available only to low income workers with children. This program not only supplements the earnings of low income parents (once a minimum monthly earnings threshold is exceeded), but also partly reimburses actual child care and provides a special shelter benefit up to levels where earnings, combined with the supplements, provide an income which is at least comparable to social assistance.

Moreover, the Quebec personal income tax system has been adjusted so that all households with incomes below what they could receive on social assistance pay no provincial income taxes. Thus, the PWA program is a true NIT - at least so far as Quebec provincial income taxes are concerned.

However, in an attempt to make the program responsive to income and earning fluctuations during the year, PWA benefits are not paid automatically and must be applied for directly. Consequently, many parents whose earning levels would qualify them for benefits do not apply and the program reaches only a fraction of the intended target group.

Newfoundland ISP Proposal

An even more ambitious program, combining NIT features with an earnings supplement, was proposed in 1993 by the Newfoundland Economic Recovery Commission with the support of the Government of Newfoundland and Labrador.

The Income Supplementation Program (ISP) is designed to provide a minimum level of income security and improved incentives to earn income. At the same time, the Unemployment Insurance system in Newfoundland and Labrador would be substantially scaled back by increasing the number of weeks of work needed to qualify for benefits and reducing the maximum number of weeks that benefits could be collected. The intent is to change radically the incentive structure of the income security system from one which encourages working just long enough at the highest wages possible to maximize UI benefits, to one which rewards maximizing one's weeks of employment even at low wages.

The ISP would provide a base income guarantee, scaled by family size and composition, at levels roughly equal to those of current social assistance benefits. In addition, earnings would be supplemented up to a maximum. For higher family incomes, both the base income guarantee and the supplement would be reduced.

As stated earlier, since the scheme is designed to maximize self-sufficiency through earnings, the guarantee level provided falls short of commonly used measures of adequacy. The hope is that by putting a floor under incomes and supplementing earnings in addition to that floor, Newfoundlanders will seek out opportunities to earn while having a non-stigmatizing, secure source of income.

Implementing an ISP would be quite a challenge, and many practical aspects would have to be worked out to see if it could be feasible. Because of the high current levels of UI expenditure in Newfoundland and Labrador, the Economic Recovery Commission believes that this modest comprehensive GAI with a second tier of income-tested earnings supplementation can be paid for out of UI savings and by replacing the provincial social assistance system without taxation or deficit increases. It is unlikely that this would be the case elsewhere in Canada as the analysis of the NIT design in this paper indicates.


All three of these "targeted" GAI-type programs - an integrated income-tested federal - provincial/territorial child benefit, earnings supplementation and a combination of a base income GAI with earnings supplementation - merit further consideration. Each addresses the largely unmet income security needs of the working poor in a much more direct and targeted way than the current system. Each can be designed to reduce the incidence and depth of poverty among this group in a manner that is more acceptable than comprehensive GAI designs such as those analyzed in the previous section.

These approaches are not without flaws. An enriched integrated income-tested child benefit could be costly. Earnings supplementation, unless carefully designed, could lead to an increase in the proportion of jobs paying low wages. The Newfoundland ISP proposal would require major tax increases and/or many middle income losers to finance in other provinces and territories. It also faces considerable delivery and administrative challenges.

However, while the arithmetic of current Canadian income distribution makes achievement of a comprehensive GAI appear either politically or fiscally out of the question, the problems inherent in these more targeted approaches can seemingly be resolved through practical, acceptable remedies. Moreover, while directly improving the standard of living of their target populations, they would do so while encouraging rather than dampening incentives to become more self-sufficient through earnings.

Appendix A - Historical Background

The First GAI - Speenhamland

A precursor of a guaranteed annual income can actually be traced back to late 18th century England. The first known GAI was the Speenhamland system implemented from 1795 to 1834 in parts of England where wages had fallen below subsistence levels. The Elizabethan Poor Law had provided relief only to those who were unable to work or could not find work. The Speenhamland system provided a subsidy, in addition to wages, that was scaled according to the price of bread and family size. This income floor was provided regardless of work effort. But the subsidy was reduced at a 100 percent rate once earnings exceeded the income floor.

Since the system was financed by poor rates levied in the village parishes, it gave manufacturers in nearby towns not subject to those rates, an incentive to hire workers only for periods when they were most needed. At the same time, by meeting the subsistence needs of families, regardless of whether they worked or not, and offering no incentive for able-bodied workers to take work paying wages less than the Speenhamland rates, the system came to be seen as undermining the work-ethic. The Poor Law Commission, which reported in 1832, concluded that the only way to correct these abuses was to ensure that public relief should not pay those able to work benefits that were higher than the worst jobs society had to offer. This became known as the principle of "less eligibility".

This principle was quickly transferred to Canada and remained the key assumption behind social assistance policy in Canada well into the 20th century. While social assistance rates are now tied to an assessment of basic need rather than to the lowest wage rates in society, the principle of "less eligibility" is still reflected in the widespread view that people should not be "better off on social assistance" than they would be working full time at a minimum wage job.

The Speenhamland experiment is also a lesson in the unintended effects of government policy. By enabling employers to hold down wages and still retain a supply of reasonably healthy workers, and allowing them to evade the costs of providing for those workers and their families, a humane attempt to help the "working poor" resulted in drawing an increasing number of workers and their families into that situation. At the same time, by taxing back assistance at a rate of 100 percent on the earnings of workers who became employed, it provided no financial incentive to work at wages less than the subsistence level. The attempt to put a floor under the incomes of the poor did maintain a subsistence standard of living, but increased the numbers of those applying for assistance and living at the subsistence level.

The Canadian Context - Social Minimum without a GAI

In 1943, Leonard Marsh, a Canadian public servant inspired by the plan for a postwar welfare state advanced by Lord Beveridge in the United Kingdom, introduced the notion of a comprehensive, integrated social security system to protect Canadians against the economic insecurities of work and raising a family. Like the Beveridge Plan, this approach was based on the idea of a social minimum. The income security aspect of this social minimum was to consist of the following three tiers:

- social insurance programs for such target groups as the unemployed, the sick, the disabled and those retired from work on scales adequate to meet the minimum needs of a single individual or a married couple;

- universal family allowances payable to the parents of all children, regardless of parental income, to meet the minimum maintenance costs of a child; and

- means-tested social assistance for those exceptional cases not covered by social insurance.

Marsh's concept of a comprehensive social minimum assumed policies to ensure high levels of employment. He emphasized that "the first positive measure in providing social security ... is a program which will make work available, or in other words, which will offer wages rather than subsistence maintenance to the furthest extent to which it is possible." (Cited in Social Planning Council of Metropolitan Toronto, A Guaranteed Income: A New Look at an Old Idea (Toronto, 1986), p. 17.)

Marsh's proposed social minimum was never fully realized, but it provided the framework for much of Canada's current social security system. In such a system there was assumed to be no need for a GAI. Income security was to be provided largely through insurance programs. Employment was to be widely available and wages from work, with a floor provided by minimum wage laws in combination with family allowances, would ensure adequacy for working families. While social assistance would provide a minimum income floor, it was to be only for cases of exceptional need, primarily those not expected or able to work because of age, child-rearing responsibilities or severe physical or mental disabilities.

The Revival of Guaranteed Income Concepts in the 1960s

The ability of a low-unemployment economy, supplemented by social insurance, to eradicate poverty began to be challenged in the early 1960s. This was a time of high economic growth and low unemployment in most western countries. Books, such as Michael Harrington's The Other America in the United States and Richard Titmuss's Income Distribution and Social Change in the United Kingdom, (Michael Harrington, The Other America; Poverty in the United States (New York: Macmillan, 1962) and Richard M. Titmuss, Income Distribution and Social Change (London: George Allen and Unwin, 1962)), were pointing out that postwar prosperity and existing social insurance and social assistance programs had left large portions of society still in poverty. In part, this reflected inadequate provision for the income needs of groups not able or expected to work such as the elderly and persons with disabilities. However, the population in poverty also included large numbers of households whose head was working for wages at least part-time or for part of the year. Despite low unemployment and high economic growth, social assistance caseloads also began to rise in major cities in the United States at this time.

Canada shared in this rediscovery of poverty. In fact, the first attempts to define statistically and measure the incidence and depth of low income in Canada through Statistics Canada's low income cut-offs (LICOs) date from this period. Poverty came to be seen primarily as inadequate income.

Anti-poverty strategies began to focus on income maintenance and support programs while the redress of macroeconomic causes of poverty such as inadequate levels of full-time, year-round employment, assumed less importance. With this new perspective, a low debt and rising tax revenues generated by rapid economic growth and rising income levels, a guaranteed income began to seem, to many Canadians, an attractive and simple mechanism to deal with the problem of poverty.

The initial response, however, was not an attempt to design a GAI, but efforts to finish the agenda established by Marsh. This led to the introduction of the earnings-related Canada and Quebec pension plans in the mid 1960s, the gradual reduction of the age of eligibility for the Old Age Pension from 70 to 65 in the late 1960s, the creation of the Canada Assistance Plan in 1966 to encourage a rationalization of provincial/territorial social assistance systems and the major expansion of the Unemployment Insurance program in 1971.

But it was recognized that enriching social insurance and demogrant programs alone could not provide all Canadians with incomes above low income thresholds. Many people were poor because they did not have sufficient attachment to the labour force to benefit from social insurance, their labour force earnings were insufficient or they had not been able to save enough or establish pension entitlements large enough during their working years to provide for an adequate income in retirement.

Many Canadian policy analysts conceived of the GAI as a reform to the third tier of Leonard Marsh's integrated approach. Social assistance at the provincial/territorial level by the 1960s had become a maze of categorical programs for lone parents, the blind, the long term unemployed and the disabled that carried considerable stigma. Some were cost-shared by the federal government. Others were not. There was also a growing recognition of the social assistance system's tendency to trap people in dependency due to its asset tests and punitive tax rates on earned income. The way forward seemed to be to maintain the social insurance programs and the universal demogrants for children and seniors while providing a new income-tested income support tier to replace social assistance and provide incentives to work and save for retirement. The negative income tax appeared to be a workable model for putting these ideas into effect.

This agenda proceeded for seniors. The Guaranteed Income Supplement was added to the Old Age Pension to create a GAI for seniors. The result, combined with the expansion of eligibility for the Old Age Pension and the creation of the Canada and Quebec pension plans virtually removed seniors from social assistance and led to a marked reduction in poverty among those over age 65. There was little resistance to a GAI for seniors because it was built on top of universal social insurance programs (rather than replacing them) and, for this group, there were few concerns about work-disincentives. In fact, with the baby boom generation just beginning to enter the labour force in large numbers there seemed good reason to encourage people to leave the paid work force at age 65 and open up opportunities for younger workers. Some have proposed that a GAI for persons with disabilities could proceed on the same basis, although there is much concern that such an approach could result in the marginalization of persons with disabilities rather than their inclusion into the social and economic mainstream.

The same agenda, however, did not proceed in the same way for the rest of the population. Nonetheless, an important step was taken. The creation of the Canada Assistance Plan in 1966 provided federal cost-sharing to encourage the provinces and territories to establish comprehensive social assistance programs available to all simply on the basis of need that were national in scope since residency tests were prohibited. The federal government reimbursed the provinces and territories for 50 percent of the cost of establishing and maintaining social assistance programs that met these requirements.

Wide interprovincial variations remained in the level of income guaranteed under provincial/territorial social assistance programs, and guarantee levels also varied within provinces depending on household type and the presence of physical or mental disabilities. Moreover, as social assistance was still subject to an intrusive means test which included assets as well as income, it continued to carry a stigma not attached to social insurance programs such as Unemployment Insurance or income-tested programs such as the Guaranteed Income Supplement for the elderly. Nevertheless, the Canada Assistance Plan was a watershed in Canada's Unique Social Historyboth in terms of federal - provincial/ territorial cooperation and the impact it had on building a modem and comprehensive last-resort safety net. By including all Canadians in the safety net, the stage was set for what many hoped would be the next step - a comprehensive GAI. Serious consideration of a GAI quickly followed in the early 1970s.

The Early Seventies - Failure to Implement a GAI

Not only in Canada, but internationally, the late 1960s and the early 1970s were the glory years for comprehensive guaranteed income proposals, particularly those based on mechanisms inspired by the negative income tax. In the five years following President Nixon's 1969 proposal for a guaranteed income for families with children in the United States, the so-called Family Assistance Plan, no fewer than 10 other countries, from Norway to New Zealand, seriously considered comprehensive reform of their income support programs based on guaranteed income principles. Except for highly targeted categories of individuals, none of these schemes were implemented. (See Leslie Lenkowsky, Politics, Economics and Welfare Reform: The Failure of the Negative Income Tax in Britain and the United States (Lanham, Md: University Press of America, 1986) pp. 3-6.)

Canada followed this pattern but in three separate initiatives.

The Quebec Commission of Inquiry on Health and Social Welfare

The first official proposal for a comprehensive GAI in Canada was made in 1971, by a Quebec provincial commission (the Castonguay-Nepveu Commission) which proposed a three-tiered income security program for Quebec. A basic negative income tax would comprise the first tier, social insurance programs the second, with income-tested family allowances providing the third.

The negative income tax program, called the General Social Allowances Plan (GSAP), would replace Quebec social assistance with two tiers of benefits: one for persons deemed to be employable and a second for those deemed not to be employable. The benefit levels for the first tier would be set at 60 percent those of the second, under the assumption that these would be supplemented by earnings from employment. This design was intended to provide an incentive to work that was consistent with the principle of "less eligibility".

A Comprehensive GAI - The Special Senate Committee on Poverty

Later in the same year, the Special Senate Committee on Poverty, under the leadership of Senator David Croll, took a different approach, with a proposal for a uniform guaranteed income through a federal negative income tax program to cover most Canadian families living in need. No distinction in benefit or reduction levels was to be made between persons deemed to be employable and those unable or not expected to work. The main elements of the Senate Committee's proposal were:

- benefit levels would be set at 70 percent of poverty levels defined by the Senate Committee and would be reduced at a rate of 70 percent against other income;

- non-Canadians and single unattached Canadians under 40 years of age would not initially qualify for GAI benefits because "the appropriate solution for this group ... lies not in income maintenance but in opportunity programs - education, training, counselling and job placement"; (Special Senate Committee on Poverty, Poverty iii Canada (Ottawa: Queen's Printer, 197 1 ) p. 187.)

- family allowances, youth allowances and Old Age Security would be abolished (in effect, these programs would be converted from universal demogrant programs to selective income-tested programs.);

- the federal government would assume the full cost of all social assistance payments made under the Canada Assistance Plan by the provinces and territories, to provide for those not covered by the GAI;

- social insurance programs, such as UI and CPP, would continue but "be reconstructed as true insurance schemes and withdrawn from the welfare field entirely" (Ibid. p. 184.); and

- the GAI would be integrated with the tax system by a mechanism to offset any income taxes payable by a household with an income below the poverty lines with a tax credit of an equal amount. The personal income tax system would be reformed to place the payment of taxes on an economic family rather than on an individual basis.

Every couple with children and every person over 40 years of age would be guaranteed an income level equal to at least 70 percent of the poverty level in the first year of implementation. A social assistance system wholly funded by the federal government would provide for the needs of non-citizens and singles under 40 years of age. The estimated cost for the year 1967 was $645 million or about one percent of GNP more than existing program outlays.

The guaranteed income provided was $3,500 for a family of four or the equivalent of $15,200 in 1993, adjusting for inflation. The Senate Committee's poverty line, however, was designed to be indexed, not to inflation, but to average living standards. In addition to the large incremental cost, equivalent to over $7 billion in 1993, other objections were raised to this proposal.

It did not provide strong incentives to work. For every dollar of earnings by poor families and older individuals, GAI benefits would be reduced at a rate of 70 cents. (In the most generous provinces current reduction rates for social assistance are 75 percent - slightly above the Senate proposal.) The cumulative effect would be that employable recipients could not increase their disposable incomes significantly through work until their earnings were well above the guarantee level.

It was no simpler than the system in place. By eliminating two benefit programs for sub-populations and replacing them with a much larger guarantee payment for a broader category of households, a larger number of cheques would have to be delivered than before. Moreover, more people would have to file income tax returns to indicate their entitlement to benefits. The expansion of the Canada Assistance Plan to cover those persons who would be excluded from the GAI would complicate matters even further.

While the Croll proposal set a framework for a made-in-Canada GAI and remains the classic expression of a GAI aimed at guaranteeing an adequate income level, the problems associated with it prevented serious consideration by governments at the time.

The Social Security Review of the 1970s

In 1973, the federal government published its Working Paper on Social Security in Canada (the Orange Paper), which became the basis for the federal - provincial social security review of the 1970s. The review was established in part to develop a social security program to combat poverty by ensuring an acceptable minimum income for all Canadians.

After consideration of a number of options by federal - provincial working groups, the federal government proposed, in February 1976, a form of GAI inspired by the two-tiered system of Castonguay-Nepveu. The provinces and territories would administer both an income support program for those without income from employment and an income supplementation program for the working poor.

The federal government would pay two thirds of the cost of income supplementation, two thirds of that part of support payments equivalent to the maximum provincial/territorial supplementation benefit eligible for federal cost-sharing and half of the remainder. The levels of the support and supplementation guarantees were to be established by the provinces and territories. However, the maximum level of income supplementation benefits eligible for federal cost-sharing was set at $80 a month or $960 a year for a family of four in 1975 dollars ($2,830 in 1993 dollars). Eligibility for income supplementation would be restricted to families with dependent children and to individuals and childless couples aged 55 to 64.

Both of these components were designed to motivate persons to work, with a 70 percent benefit reduction rate for income support and a 35 percent reduction rate for income supplementation. The income security system for those over age 65 was to be left unchanged. (National Council of Welfare, Guide to the Guaranteed Income (Ottawa: March 1976) pp. 37-39.)

An analysis by the National Council of Welfare estimated that almost 450,000 households containing 1,600,000 people could benefit from income supplementation but stated: "There can be no denying that the proposed supplementation program will not lift most of these families out of poverty. The benefit levels are too low to accomplish this." (National Council of Welfare, Support supplementation: who will benefit? (Ottawa: November 1976) p. 32.)

The February 1976 proposals were not accepted by the provinces because of fears about their short and long-term costs in a context of high inflation, recent sharp increases in the federal deficit and a marked slowdown in the growth rate of real government revenues.

While the federal - provincial social security review was considering the design of a GAI in the early 1970s, the province of Manitoba, with financial support from the federal government, chose to go one step further by actually piloting the concept. The Manitoba Basic Annual Income Experiment (Mincome) made payments to over 1,000 Manitoba families over three years, beginning in 1975. The experiment was designed to evaluate work responses of employable recipients under a guaranteed income.

Analysis of the work incentive impacts of the program were consistent with the results of similar experiments in the United States, i.e., labour market participation did not increase as might have been hoped. In fact, it declined, but only slightly overall with most of the reduction coming from the lower earner in two-earner couples. Whether this is a success depends on what is expected from a GAI. If the focus is on poverty, this result can be interpreted positively, i.e., benefits can be increased without major changes in work patterns. If the focus is adjustment, the fact that simplification and theoretical reduction of work disincentives did not increase labour supply would seem to be a disappointment. (Derek Hum and Wayne Simpson, Income Maintenance, Work Efforts and the Canadian Mincome Experiment (Ottawa: Supply and Services Canada, 1991)). Finally, it is noteworthy that the level of income guarantees provided by Mincome, in 1993 dollars, are below current social assistance levels in that province. Therefore, the Mincome study tells us little about the potential work disincentive effects of a more "adequate" GAI.

Late Seventies and Early Eighties - The Shift to Incrementalism

Following the collapse of the Orange Paper process in 1976, reform to income security programs in Canada shifted its focus from comprehensive to incremental change. New programs based on the GAI idea, such as the refundable Child Tax Credit and the refundable Sales Tax Credit at the federal level and various supplements at the provincial level, aimed to improve, modestly, the situation of low-income households, particularly families with dependent children. However, these programs were additions to the current system and did not consolidate existing income support programs. Moreover, they made no pretence that the income guarantees they provided were in any sense adequate.

The result was to multiply significantly the number of programs available and to create considerable overlap between federal and provincial/territorial programs in the area of income maintenance and among income maintenance programs within levels of government. Unlike the de facto GAI for seniors, these programs usually did not provide high enough benefits to remove families from social assistance. (The APPORT program in Quebec, which supplements the earnings of the working poor is an exception to this generalization since it is designed to complement the social assistance system.)

On the other hand, the introduction of refundable credits provided much legitimacy to the negative income tax mechanism for delivering benefits to poor Canadians. Many considered these credits as a step toward a GAI.

The Mid-Eighties to the Present - The GAI Revived

A new generation of GAI proposals emerged in the mid 1980s. The most prominent of these was the Universal Income Security Plan (UISP) in the 1985 report of the Royal Commission on the Economic Union and Development Prospects for Canada.

Also, in the late 1980s Quebec and Ontario made major, but very different reforms to their income security systems which addressed the issues of adequacy, supplementation of the incomes of working poor families and work incentives raised by earlier GAI proposals.

Quebec White Paper on the Personal Tax and Transfer System (1984)

The reforms in Quebec were inspired both by the Castonguay-Nepveu Commission of the early 1970s and by a white paper on the provincial personal income tax and income support system published in 1984 which laid great emphasis on the very high marginal tax rates faced by the poor as they attempted to make the transition from social assistance to work. (Quebec, Department of Finance, White Paper on the Personal Tax and Transfer Systems (Quebec: 1984).) It also stressed the need to harmonize the income tax and income transfer systems so households with incomes below social assistance rates would not have to pay income tax. To do this, it proposed using common definitions of the family unit and of income in both systems and indexing each to inflation in the same way.

The Universal Income Security Plan (UISP) (1985)

The proposal of the UISP by the Royal Commission on the Economic Union and Development Prospects for Canada (the Macdonald Commission) marked a number of departures in thinking about a GAI and its relationship to social insurance programs such as Unemployment Insurance.

While past GAI proposals had attempted to improve the standard of living of the poor while seeking to minimize work disincentives, the UISP was consciously designed to provide an income floor which would make it politically and socially acceptable to embark on what would otherwise be controversial economic adjustment policies, including a significant scaling back of the Unemployment Insurance system, an expansion of adjustment programs, such as job training and mobility assistance, and the signing of a free trade agreement with the United States. This was an attempt to convert what had been a "passive" policy tool into a foundation for "active" measures designed to improve economic competitiveness.

Previous GAI proposals (with the partial exception of the 1971 report of the Special Senate Committee on Poverty) had assumed that Unemployment Insurance would remain in place. However, anticipating the conclusions of the Newfoundland and federal royal commissions on Unemployment Insurance which would report a year later, the Macdonald Commission saw many aspects of the Unemployment Insurance system, particularly regionally extended benefits, as crossing the boundary between social insurance and income supplementation. (Newfoundland, Building on our Strengths: Report of the Royal Commission on Employment and Unemployment (the House Report) (St.John's: 1986) and Canada, Commission of Inquiry on Unemployment Insurance (the Forget Report) (Ottawa: 1986).)

To restore the UI program to its social insurance roots, the Commission recommended stripping UI of its income support elements and replacing them with an explicit, income-tested supplementation program with benefits related to family size.

Concerned as it was with promoting economic adjustment, the Macdonald Commission deliberately made its UISP inadequate both in terms of commonly used measures of low income and existing social assistance levels in many provinces. Moreover, to avoid work disincentives it proposed to tax back benefits at a low rate of 20 percent.

The UISP also attempted to disentangle the roles of the federal and provincial/territorial governments in the income security system. In contrast to the proposal emerging from the social security review of the 1970s, the federal government would be 100 percent responsible for the administration and financing of income supplementation for those with earnings as well as for providing an income base for the non-earning poor.

This would leave the provinces 100 percent responsible for the financing and administration of a support tier which would top up the UISP to ensure that persons on social assistance would be no worse off than under the current system.

The UISP was designed to replace:
- the three existing child benefits;
- the spousal and equivalent to married exemptions in the income tax system;
- the Guaranteed Income Supplement and Spouse's Allowance for the elderly;
- federal contributions to provincial/territorial social assistance programs; and
- federal housing programs.

The UISP would provide full benefits to all families and to unattached individuals over age 35. The annual benefit level for a couple with two children would be about $10,000 in 1993 dollars, and would be indexed to growth in the Consumer Price Index.

The Macdonald proposal had many attractive features which improved on previous GAI scenarios. It provided for a streamlining of federal programs through eliminations and tax simplifications - one program would now stand where several had previously been. It also provided a more rational system of income supplementation to the working poor than Unemployment Insurance and recognized an obligation by government to assist those families whose earnings did not provide them with an adequate income.

The program was also intended to be fiscally neutral. Dollars saved from programs eliminated would be redeployed into the UISP. However, one independent estimate put the incremental cost of the program, as designed, at an additional $3 billion to $5 billion in 1985. (See Michael Wolfson, "A Guaranteed Income", Policy Options (January, 1986) p. 37. Wolfson estimated the proposal could be made fiscally neutral by raising the reduction rate to 25 percent.)

However, the UISP disappointed many advocates of the GAI concept because it did not promise to reduce substantially either the incidence or the depth of poverty.

Combining GAI and Tax Reform

The program design of UISP also failed to recognize that because the reduction rate of the benefit program was added to the existing income tax system, the cumulative marginal tax rates and work disincentive effects would be far more severe than intended. With a basic benefit for a family of four of approximately $10,000 in 1993 dollars and a reduction rate of 20 percent, the UISP would provide net benefits to families with income from other sources up to $50,000. (Not all these families would be better off than under the current system because they would lose their Child Tax Benefit and might have to pay higher taxes because of the elimination of the married exemption.) However, adding that 20 percent reduction rate on top of current federal and provincial/territorial marginal income tax rates would mean a 45 percent effective marginal tax rate for those earning up to $30,000 a year and a 59 percent effective marginal tax rate for those earning between this level and the $50,000 threshold.

In 1986, to avoid this problem and to attempt to provide income guarantees which were more adequate, Michael Wolfson, an economist at Statistics Canada, proposed a version of the GAI which, while retaining the disentanglement of federal - provincial/territorial roles in income security proposed in the UISP, involved a profound reform of the personal income tax system to complement the base federal income guarantee.

The most radical version of this Guaranteed Income/Simplified Tax (GI/ST) model would have abolished all federal child benefits, the Guaranteed Income Supplement for seniors, federal cost-sharing of provincial/territorial social assistance spending under the Canada Assistance Plan and the refundable GST Credit, reduced spending on Unemployment Insurance by 30 percent and made the basic Old Age Pension non-taxable. Within the tax system, it would have abolished the basic personal credit, the spousal and equivalent to married credit, the age and pension income credits and Unemployment Insurance premiums.

All these would be replaced by a federal basic income guarantee of $7,200 for a family of four headed by a person under age 65 and guarantees of $4,220 for seniors living alone and $7,000 for senior couples. (The guarantee levels for seniors were designed to mimic the existing GIS benefit structure for this age group.) The guarantees would not be taxed back so, in this respect, the design resembles the universal demogrant approach described in the introduction. This means that, unlike the negative income tax design of the Croll Committee under which families with children at income levels above the point where no benefit is paid would have the same disposable income as families without children and the same pretax income, the GI/ST design provided for horizontal equity between families with and without children at all income levels. (Families with children and high incomes would probably pay more in additional taxes than the benefits they would receive on behalf of their children under GI/ST but their net additional tax burden would be lower than for families without dependent children possessing the same pretax income.

Added to this would be a flat-tax personal income tax system which would tax all personal net income at a basic rate of 29.5 percent with a surtax of 16.5 percent on total income in excess of $40,000. (Net income is total personal income from sources other than the guarantee after subtracting income spent on the remaining deductions in the tax system such as for charitable donations, pension contributions and child care expenses.) Men and women would be treated as individuals within married couples in the tax system.

With a constraint of fiscal neutrality, Wolfson argued that provinces and territories could afford to top up the federal guarantees for persons not earning income because of disability, lone parenthood or high local unemployment rates. This income support tier which would be wholly financed by the provinces and territories would be set at $3,000 a year for each adult aged 18 to 64 and at $1,500 for each child under age 18. This benefit would be taxed back at a rate of 40 percent. This support tier would replace existing social assistance benefits. Families of four qualifying for maximum support and supplementation benefits would have a combined guaranteed income of $16,200 in 1986 dollars or just over $21,500 in 1993 dollars. This would still be over $5,000 below the weighted average Statistics Canada low income cut-off for a family of four in 1993 (And below the income available to a family of four on social assistance in Ontario in 1993. See National Council of Welfare, Welfare Incomes 1993 (Ottawa: Summer 1994) p. 30.) and would be available only to families with no other income. The basic federal income floor for families of four would equal about $9,500 in 1993 dollars.

The GI/ST proposal is the most comprehensive Canadian version of the universal demogrant type of GAI, although unlike the classic universal demogrant, it has separate supplementation and support tiers, with the support tier being a form of negative income tax.

Social Assistance Reform

While the federal government chose not to pursue either the UISP or the GI/ST proposals, Canada's two most populous provinces - Ontario and Quebec - moved in the late 1980s to reform their social assistance systems in ways resembling much earlier GAI proposals.

In Ontario, reforms moved in a direction similar to that proposed by the Croll report, i.e., higher benefit guarantees and reduced tax rates on earned income. Between 1986 and 1993, real social assistance benefits in the province increased by approximately 25 percent. By 1993 the income available to families with children on social assistance stood at between 72 percent and 82 percent of Statistics Canada's low income cut-offs (these are the 1992 base low income cut-offs. The National Council of Welfare used the 1986 base cut-offs.) in urban areas of Ontario with populations of 500,000 or more, and between 82 percent and 95 percent of the cut-offs for families with children living in less densely populated areas of the province. (Ibid p. 27. The lower figure in the range is for a married couple with two children. The upper figure is for a lone parent with one child.) The adequacy of benefits for single persons with disabilities was increased to similar percentages of the low income cut-offs. During the same period the taxback rate on non-social assistance net income was lowered to 75 percent. (Net income is income after deducting income and payroll taxes and child care and work-related expenses up to a fixed limit. Persons receiving social assistance in Ontario and other provinces are also permitted to earn small amounts of income a month before the 75 percent taxback rate is applied.) Previously, much higher rates had been levied on non-social assistance gross income.

These guarantee and taxback levels are very similar to those proposed by the Croll Committee in 1971 (70 percent of the poverty line with a 70 percent taxback rate.)

In Quebec, equally significant changes were made to the income security system in the late 1980s, but were less focused on benefit adequacy and more on harmonization between taxes and transfer payments and improved work and employability incentives. Quebec did not significantly increase its real social assistance rates for families with children. Instead, it roughly indexed them to inflation at levels between 50 percent and 60 percent of the low income cut-offs.

However, the provincial income tax system was changed so households with incomes lower than what they would receive on social assistance were not subject to provincial income tax. Payments to parents of newborn children were introduced, and the universal provincial family allowance for families with children was enriched. Social assistance benefits, among those deemed able and expected to work, were set at higher levels for those willing to participate in active programs to improve their employability (and for those actually doing so) than for those unwilling to participate in such programs. And a new income-tested earnings supplementation program, APPORT, was introduced to supplement the earnings and reimburse the child care expenses of working poor parents.


Policy makers have been motivated by many different and sometimes incompatible objectives in considering a GAI. The main objectives have been to:
- provide more money for the poor to contribute to eliminating poverty in society;
- provide more choice and less stigma. Basic needs would be met without the stigma of needs-tested social welfare or, potentially, the classification of the population into employable and unemployable groups;
- simplify the current system of programs and services. (Many of these programs and services would be integrated and harmonized into a single benefit that would require a smaller infrastructure to administer and would make it easier for the client to receive government benefits.);
- cost less than current programs and services;
- provide better incentives to work than the current system of taxes and transfers which imposes conditions and limitations on the receipt of benefits and the level of earnings; and
- allow for a faster and more efficient adjustment process to structural economic change. If all Canadians could be assured of a base income that doesn't penalize work effort, then the willingness and ability to adjust to changing economic circumstances would improve.

The historical review of the GAI proposals indicates that although any one of these objectives may be feasible, combining several objectives into the same program requires trading off one objective against the other.

Appendix B: Measures of Adequacy (This Appendix deals with selected measures only. For a fuller discussion, including an evaluation of the strengths and weaknesses of the various measures, see Wolfson and Evans, op.cit.)

Relative Measures of Adequacy

Low Income Measure

In addition to the low-income cut-offs (LICOs) Statistics Canada also publishes the low income measure (LIMs). These represent 50 percent of adjusted median family income where the adjustment reflects a judgment on how much income needs increase with family size and configuration. In calculating the LIMs, it is assumed that each additional adult after the first increases the family's needs by 40 percent while each dependent child's needs are assumed to be 30 percent of those of the first adult. The exception is in the case of a lone parent family where it is assumed the first dependent child adds 40 percent rather than 30 percent to the family's needs.

Adjusted family size is then determined by counting the first adult as 1, the second adult or the first child in a lone parent family as 0.4 and all other dependent children as 0.3. Thus, a married couple with two dependent children would have an adjusted family size of 1 plus 0.4 plus 0.3 plus 0.3 equals 2.

Adjusted family income is then determined by dividing family income by adjusted family size. The median adjusted family income is the adjusted family income where 50 percent of families have a smaller adjusted family income and 50 percent have a larger one. The LIM for a single person living alone is 50 percent of the median adjusted family income and the LIMs for all other family configurations are equal to this value multiplied by adjusted family size.

The LIMs are automatically adjusted each year to reflect changes in median adjusted family income.(Unlike the low income cut-offs the LIMs are not adjusted to take the size of the community in which the family resides into account.)

CCSD Poverty Thresholds

Another commonly used relative measure of adequacy is the income threshold calculated by the Canadian Council for Social Development (CCSD). It represents 50 percent of average adjusted family income. The CCSD assumes that the first additional person in the household increases the family's needs by two fifths, and each subsequent person adds an additional one third. Since the average size of a census family in Canada is just over three persons, the CCSD sets the poverty threshold for a family of three at 50 percent of average pretax family income.

Adjustments are then made for different-sized families on the basis of family income units which are scaled as follows: a family of one equals three income units, a family of two equals five income units, a family of three equals six income units, a family of four seven units and so on. Thus, for example, the poverty threshold for a family of one is three sixths of that for a family of three.

The CCSD thresholds are automatically updated each year according to changes in average family incomes, but are not adjusted for the size of the community in which the family resides.

Absolute Measures of Adequacy

Other analysts have attempted to construct absolute measures of adequacy based on the cost of essential food, clothing, shelter and other needs for varying family sizes and configurations. Two well known examples are those used by the federal government in the United States and those recently calculated for Canadian cities and provinces by economist Christopher Sarlo.

USA Poverty Lines

Unlike Canada and most other developed countries, the United States has an official set of poverty lines which are used to determine eligibility for a number of government programs.

These lines were based on the cost of a basic nutritious food budget for households of varying sizes in 1961. These amounts were then multiplied by three which was the average ratio of family food expenditures to post-tax family income in the United States at the time. These amounts have since been updated annually to match changes in the American Consumer Price Index.

The Sarlo Poverty Lines

These lines were initially calculated by Christopher Sarlo in 1988 for major Canadian cities and for each of the provinces. They reflect the actual cost of a basic nutritious food diet, basic shelter and clothing and the cost of a range of other goods and services deemed to be essential for a basic standard of living by Sarlo.

Comparing Measures of Adequacy

As the attached table indicates, the range of adequacy for common household types under these relative and absolute measures of low income in a given year varies widely.

1992 Thresholds of Pre-Tax Adequacy - in Canadian dollars

Sarlo's amounts are 1988 estimates for the City of Toronto in 1992 dollars.
U.S.A. amounts are adjusted for 1992 purchasing parities (U.S. $1 dollar equals Canadian $1.24)
LICO amounts are the weighted average of the population by community size

One adult
Sarlo - $7,983
U.S.A. - $8,857
LIMs - $12,148
LICOs - $14,615
CCSD - $13,419

One adult-one child or two adults
Sarlo - $10,903
U.S.A. - $11,330
LIMs - $17,007
LICOs - $17,830
CCSD - $22,365

Two adults-one child or One adult-two children
Sarlo - $14,126
U.S.A. - $13,871
LIMs - $20,652
LICOs - $22,213
CCSD - $26,838

Two adults-two children or One adult-three children
Sarlo - $17,936
U.S.A. - $17,775
LIMs - $24,296
LICOs - $26,843
CCSD - $31,311

Appendix C: 1993 Social Assistance Incomes (Social assistance income refers to provincial or territorial and municipal cash social assistance benefits plus other cash benefits available to persons receiving social assistance such as the refundable Child Tax Benefit and GST Credit.

Employable Adult - $4,522
Disabled Adult - $8,541
Lone Parent One Child - $12,986
Couple Two Children - $14,825

Prince Edward Island
Employable Adult - $8,180
Disabled Adult - $9,294
Lone Parent One Child - $12,773
Couple Two Children - $19,110

Nova Scotia
Employable Adult - $6,100
Disabled Adult - $8,637
Lone Parent One Child - $12,080
Couple Two Children - $15,111

New Brunswick
Employable Adult - $3,256
Disabled Adult - $8,238
Lone Parent One Child - $10,150
Couple Two Children - $12,151

Employable Adult - $6,316
Disabled Adult - $8,164
Lone Parent One Child - $12,607
Couple Two Children - $16,251

Employable Adult - $8,527
Disabled Adult - $11,725
Lone Parent One Child - $16,790
Couple Two Children - $22,334

Employable Adult - $7,236
Disabled Adult - $8,257
Lone Parent One Child - $11,386
Couple Two Children - $19,410

Employable Adult - $5,965
Disabled Adult - $8,512
Lone Parent One Child - $12,093
Couple Two Children - $17,382

Employable Adult - $5,608
Disabled Adult - $9,753 (Assumes recipient qualified for Assured Income for the Severely Handicapped (AISH) benefits
Lone Parent One Child - $11,281
Couple Two Children - $18,122

British Columbia
Employable Adult - $6,639
Disabled Adult - $9,318
Lone Parent One Child - $13,345
Couple Two Children - $17,374

Employable Adult - $8,121
Disabled Adult - $8,977
Lone Parent One Child - $14,841
Couple Two Children - $22,404

Northwest Territories
Employable Adult - $11,599
Disabled Adult - $13,104
Lone Parent One Child - $20,893
Couple Two Children - $25,307

Source: National Council of Welfare, Welfare Incomes 1993 (Ottawa: Summer 1994) pages 16 to 23.