John Richards is an editor of Inroads. For the last several years he has been co-editing, with Bill Watson of McGill University, a major series on social policy for the C.D. Howe Institute.
Operating in pre-electoral mode, the Liberals announced in the 1997 federal budget their intention to enrich the "platform" of tax benefits available for families with children. The provinces can now build on this platform additional anti-poverty programs reflecting their respective strategies. As an aside, this is a nice example of federalism functioning as it should. Ottawa is restricting its intervention to what it can do well. It is providing a relatively simple rules-based transfer of money, and leaving to the provinces the more administratively complex social policy additions to this platform. (1)
Some provinces--most notably British Columbia--favour the idea of incrementally building a generous guaranteed annual income for all families with children. Other provinces--including New Brunswick, Quebec, Saskatchewan and Alberta--prefer instead introduction of earnings supplements. Unambiguously, I think the second group is right and, in this article, I attempt to explain why.
What are guaranteed annual incomes and earnings supplements?
In the 1960s and 1970s, the terms "guaranteed annual income" (GAI) and "negative income tax" (NIT) were bandied about interchangeably in North American social policy debates. Proponents included a wide range of people--extending from advocates for the poor through to Chicago economists like Milton Friedman. In essence, the government would provide every person a guaranteed floor income sufficient to lift him/her and any dependents above poverty. Above some threshold of earnings, this supplemental income is taxed back.
By contrast, earnings supplements render employment more financially attractive by augmenting earnings, at a predetermined percentage, over a phase-in range of low earnings. By construction there is no income transfer without earnings. The maximum transfer occurs at the upper limit of earnings eligible for supplement. As with the GAI/NIT, this supplemental income is taxed back beyond some threshold earnings level.
Academics and policy analysts have debated earnings supplements for many decades, but only recently have they become a prominent component of social policy. The United States introduced an earnings supplement (the Earned Income Tax Credit) and has greatly expanded it in recent years. Ottawa introduced a small earnings supplement (the Working Income Supplement) in 1993, at the time of the overhaul of the child tax benefit. To date, Quebec is the province that has experimented most ambitiously with earnings supplements (the Parental Wage Assistance).
To give readers a more precise feel for how earnings supplements work, let me sketch one proposal being considered in Saskatchewan. The provincial government would pay a supplement of 25-35 percent (varying by number of children) over earnings of $0-$12 000. Above this phase-in range, the benefit would be taxed back at 25cents per dollar of earnings. Furthermore, so that it "feels" like earnings to the recipient, the province wants a quick-response administrative system that can pay supplements within 30 days of someone receiving the eligible earnings.
Keep in mind that roughly half of all single parents receive some form of social assistance during any year. Accordingly, it is impossible to consider the financial position of poor families with children independently of provincial welfare programs. Now, the purpose of welfare programs is to target funds to the "truly poor." Targeting means that welfare payments must fall sharply as payments rise. Typically, provinces reduce welfare payments by 80cents or more for every dollar earned. This means that parents on welfare face a prohibitively high marginal taxback rate on earned income. This 80-percent tax rate also applies to financial support from the non-custodial parent, and acts as a major disincentive to pursuing such income. (If the government takes 80cents of every dollar, why would a welfare mother exacerbate relations with the estranged father by pursuing child support?)
The great value of earnings supplements is to lower the effective marginal tax rate on earnings (over the phase-in range) among parents on welfare. (With a supplement rate of 30 percent and a welfare taxback rate of 80 percent, the net taxback rate falls from 80 percent to 50 percent.) The earnings supplements should also apply to payments from the non-custodial parent (i.e. the father), thus providing an important incentive for the welfare recipient (i.e. the mother) to pursue this income. For the minority of eligible parents who are not on welfare, an earnings supplement simply renders employment more financially attractive.
Earnings supplements are no panacea to the problems of poverty. They are not a substitute for good training programs. Nor are they a substitute for well-trained, professional social workers, helping people to get back on their feet and making the tough professional distinctions between those who should and should not be eligible for social assistance. Since earnings supplements entail a subsidy to low-wage employment, some object to them on the grounds that government should not be subsidizing nonunion "macjobs" and propose instead raising the minimum wage. My response to these critics is that they underestimate the importance of work experience in helping poor families. Poor people usually have low skills and, to be realistic, can only command low wages. Raising the minimum wage is usually an inferior option: it tends to force out of the labour market precisely those we want in.
On the causes of poverty
Let's step back from the current debate over provincial additions to the federal child benefit platform, and address a more fundamental question: why has the poverty rate in Canada not declined over the last generation?
In attempting to answer, I start with Robert Haveman and Barbara Wolfe, (2) two economists at the Institute for Research on Poverty at the University of Wisconsin at Madison. They are well known for their pioneering work with longitudinal studies that follow large samples of families over decades. Such studies use empirically testable models and generate answers--albeit tentative--to questions about the effects of the environment (including here both government policy and trends in civil society) and parental choice on outcomes of children. In studying the cause of poverty, the outcomes typically measured include success of children in completion of education, in avoiding out-of-wedlock pregnancy, and in avoiding unemployment or reliance on welfare.
Longitudinal studies offer the best empirical answers to questions about the causes of poverty. Nonetheless, modesty remains essential. As Haveman and Wolfe put it:
"Without exception [longitudinal studies] are constrained by both data and modeling limitations. While some of the estimates do derive from causal models, the typical study explores ... a limited number of parental characteristics or choice (e.g. family structure, income, and welfare dependency) and some aspect of children's attainment, controlling for as many other relevant factors (e.g. parental education or neighbourhood characteristics) as the data permit."
Implicit in this approach is the assumption that the coefficient on (or simulated effect of) a particular variable reflects its total effect on children's attainment; that the variable is an exogenous determinant of attainment, and as such is independent of other potential determinants. Because this assumption is often violated, but to widely varying degrees, the estimated relationships revealed in the studies must be interpreted with caution.
While other variables matter, and while clouds of uncertainty swirl around, I suggest two independent sets of variables can be distinguished in any attempt to explain the stubborn stability of poverty rates in Canada over the last generation.
First, changing technology and globalization have together reduced the supply of "good jobs" for those with little schooling. Over time the distribution of training among Canadian workers has been shifting up, but this upward shift has not taken place quickly enough. Accordingly, the supply of good jobs for low-education workers has fallen faster...