Dear Ms. Lankin and Mr. Sheikh,
You have taken on a very difficult task and I commend you for it. As I outline below, the social
assistance system faces a mixture of priorities that makes it difficult to agree on a course of action –
and you do not even know yet to what government you will be presenting your recommendations. I
hope that reason and compassion will guide you equally and you will come to a set of
recommendations that any government can adopt.
I have followed the Ontario system and the literature from across North America with great
interest over the last couple of years and have done a small amount of research myself. My expertise is
primarily in the area of effective tax rates; as such, the comments that follow are mostly directed at
Ontario Works, rather than the Ontario Disability Support Program, and mostly ignore the methods of
improving employment supports. I apologize for the length of the submission, but hope that the
information and framing here is helpful to you. I have bolded my references to your Discussion Paper
questions to help you organize the feedback and I have included a brief summary of my
recommendations at the end of the paper.
You are welcome to contact me if you would like any further information or further input in this
process. You are also welcome to publish this submission, but please do not publish my contact
Submission to the Commission for the Review of Social Assistance in Ontario – Robbie Brydon
The question of how to build a social assistance system is only relevant once its goals have been
set, so the logical place to begin is with Issue 4, Question 1: “What should be the expected outcomes
of social assistance?” In reviewing programs from across Canada and the United States, I have found
that social assistance (SA) programs generally have the following three main goals and the difference
among programs lies primarily in which of these are emphasized:
1. Minimizing the effects of poverty
2. Supporting unemployed individuals to achieve employment, where possible
3. Minimizing costs to the government and society
My recommendations in this submission will attempt to find a balance of these goals,
capitalizing on areas of overlap where they exist. Since the changes in SA policy beginning in 1995,
the system has been far more effective than its immediate predecessor in controlling costs, though
generally less effective in minimizing poverty. From a high of $630 per Ontarian in 1994 (all figures in
real 2009 dollars), costs were reduced to $360 in 2001 before gradually climbing back to $537 in 2010,
driven by rising ODSP caseloads and the recent recession1 (Brydon, mimeo). Though the Fraser
Institute argued that the current system reduced poverty rates, it is hard to attribute most of the
employment gains to the new social assistance system, as they implicitly do (LeRoy, 2004). Rather, it
seems that those gains were part of a broader Canadian and global expansion and that the one firm
conclusion available is that those who did rely on Ontario Works were in deeper poverty than they
would have been in the late 1980s and early 1990s. This state of affairs has been altered somewhat for
families in the past five years, due to a slew of additional child-based benefits, but remains true for
single individuals on OW. Thus, it seems to me that the present areas of weakest performance are in
poverty alleviation and employment transition and that the focus should be on improving these two
while limiting costs.
With these goals in mind, I would suggest that the Commission recommend that the
Government of Ontario evaluate the Social Assistance system based on three criteria:
1. A measure of low-income that takes into account both the prevalence and severity of income
Excludes the Social Benefits Tribunal and Automating Social Assistance Project costs.
The LIM (40%) currently used in the Ontario Government’s Poverty Reduction Strategy is not an ideal measure of
severity as it makes no distinction between an income at 39% and 10% of the median. The average low income gap is a
2. The employment rate of individuals at 6 months, 1 year and 5 years from date of entry to social
3. The total cost of the program.
Most of this data is already available (Issue 4, Question 2), but it will also be necessary to track
a large sample of former social assistance recipients to determine their employment rates. This survey
could also provide useful information on many other fronts, including highlighting successful (and
unsuccessful) employment programs. It may even create an additional measure for #1, which would
analyze income levels of (former) social assistance recipients in the medium-term. Disaggregated data,
including that on ethnicity or race, would be useful in determining which groups are struggling – or
succeeding – disproportionately and identifying supports needed or supports that should be replicated.
Issues 2, 3 and 5 are inextricably linked and must be dealt with jointly, to some extent. What
should benefits be, how should they be structured, what are the rules that will govern access and how
do they fit into the broader context of income security policies? To answer these questions, I will draw
on the goals outlined above.
Reducing the incidence of poverty is about increasing employment opportunities for clients,
their employability and their desire to seek out and take on opportunities. This has both “pull” and
“push” factors (demand for and supply of labour) and while the former shapes much of the landscape, it
is the latter that is most directly affected by the social assistance system. Structurally, the biggest
question is very similar to Issue 2, Question 2, but must actually consider a broader scope: What are
the costs of leaving welfare for work? This includes elements of Issue 5, Question 2 on interactions
with other programs and also the question of child care, which is the largest piece missing from the
Discussion Paper questions. These costs are significant and sometimes prohibitive and can mostly be
broken down into five categories:
1. Effective tax rates
2. Wage : welfare ratio
3. Child care cost
4. Cost of working (transportation, clothes, etc.)
5. Cost of leaving welfare (loss of benefits, risk of not getting back on when needed)
much stronger measure and a combined tool might use an index of that and the LIM (50%), each based at 10, and take the
product to give a measure with base 100 that could be easily followed.
Since the 2004 reforms and the introduction of the WITB, effective tax rates for individuals on
OW have been reduced greatly. Poschmann (2004) and Laurin and Poschmann (2011) examine
Marginal Effective Tax Rates (METRs) with a particular focus on low-income earners, but directly
exclude social assistance receipts. When these are included, the picture becomes very different.
Further, where marginal tax rates examine labour supply effects on the intensive margin (whether one
chooses to work more), Average Effective Tax Rates (ETRs) are more appropriate for examining the
extensive margin (whether one works at all). In research I have been conducting with St. Christopher
House in Toronto, based on the initial work of Drummond and Manning (2005), I have examined the
effective tax rates for welfare leavers with and without child care and housing subsidies in Toronto.
Using the characters of Jim, a single male on OW, and Sally, a single parent of a child under 6 on OW,
we get a very good understanding of the incentive system by looking at their “Disposable Incomes”
graphed against the hours they work in the charts below. To calculate Disposable Income, I have used a
constant rent amount equal to the shelter component of OW. Child care costs are estimated at either end
of the general spectrum – $5 per hour of work in informal care (Figure 1) and $8 per hour in formal
care (Figure 2) – though rates actually tend to be by day or half-day, giving parents less flexibility than
displayed here. Figure 3 compares Jim and Sally and all three figures show both minimum wage and
$15/hr possibilities3. Other potential costs of working are not included here.
Figure 1: Sally's Disposable Income, Informal Child Care at $5/hr
Annual Disposable Income
Min Wage, Subsidy
0 10 20 30 40 50 60
Calculations for $12 and $20 were also performed and Effective Tax Rates based on $1,000 annual increments are
available for each of these cases (and cases including rent-geared-to-income housing) from the author.
Figure 2: Sally's Disposable Income, Formal Child Care at $8/hr
Annual Disposable Income
Min Wage, Subsidy
0 10 20 30 40 50 60
Figure 3: Jim vs. Sally, Disposable Income, Informal Child Care at $5/hr
Annual Disposable Income
Jim, Min Wage
Sally Min Wage
0 10 20 30 40 50 60
As we can see, Jim’s income rises steadily as he works more. In fact, his effective tax rate
oscillates between 40% and 56% – not low, but low enough to see significant benefit to working.
Sally’s, on the other hand, depends mightily on the cost of child care. At $5 per hour, she does a little
better than Jim, moving between 37% and 52% – until her OW child care exemption expires and her
METR clears 100% (she is actually losing money by working more). At a higher child care rate, her
ETR sits between 65% and 80% before the reduction in child benefits and WITB swallows the rest of
her paycheque: at minimum wage in this scenario, the most she can keep in a year is just under $3,600,
or $300 a month! All of this, of course, changes drastically when her child care costs are tied to her
income through a subsidy and her effective tax rate stays between 39% and 52% all the way up to full-
time work (56% is the maximum if she earns $15 per hour instead of minimum wage).
Perhaps this exercise was more detail than necessary in this submission, but I wished to
highlight two things: that effective tax rates are on the high side, but not ridiculous for singles and that
they are a serious impediment for single parents. In making recommendations, the Commission must
take into account how to reduce this barrier. The cheapest and easiest solution is likely a provincial
child care system similar to what Quebec has developed – recent studies have estimated cost recovery
in the 40% range from the additional labour supply of parents and up to 105% if indirect tax effects are
included (Fortin, 2011). An alternative, or complement, would be a provincial extension of the WITB,
similar to that advocated by Stapleton (2009).
Reducing the depth of poverty for those reliant on the system, temporarily or permanently, is
ultimately about higher benefit levels. The challenge is to do this with minimal impact on the
provincial budget, the incentive to stay off welfare (also the concept of “fairness,” as presented in the
Discussion Paper) and effective tax rates, as outlined above. There are several courses of action, three
of which strike me as appealing possibilities in this vein:
1. Increase rates
2. Leave the base rate and introduce a second tier that is reward-based and/or time-limited
3. Scrap social assistance entirely and replace it with an income-tested benefit and an income-
tested disability benefit
To address the most challenging one promptly, Option 3 requires the Conversion of the Basic
Income Tax Credit to a Basic Income and the introduction of a major new source of revenue, such as a
carbon tax, to increase income redistribution. To be fiscally sound, it would likely require an income
level no higher than the current social assistance level, which does not address the current problem, and
cooperation at the federal level, which is not presently feasible. Given these challenges, I do not
recommend it as the primary option, though I do support mentioning it as a means of making income
redistribution fairer and improving work incentives (benefits would no longer be conditional on not
working, as they presently are).
Option 1 is the most effective at reducing the depth of poverty, but is also expensive. I predict
that increasing OW benefits to the point where they and tax credits total $10,000 per year for singles
and a similar move for families (approximately a one-third increase) could cost up to $2.3 billion per
year4. As a single move, this would likely exceed the Government’s target of reducing child poverty by
25% and would severely reduce the depth of poverty experienced in the province. Given that we have
lagged in addressing poverty, I personally would prefer this option to a path of no change; others in the
province would heartily disagree.
Option 2 is my preferred specification. In answering Issue 2, Question 1 – How should social
assistance rates be determined? – I begin by asking, “What is the least we will accept for an Ontarian to
have, given our resources?” The current answer for a single person is approximately $700 per year: if
you do not participate in the OW program or cannot produce the necessary identification, you are only
eligible for a handful of tax credits (if you file taxes at all). The costs of dealing with those individuals
then moves to other areas (health, emergency shelters, corrections) and generally jumps significantly. If
you do meet the OW participation criteria, the base level jumps to just under $8,000 per year – just
barely enough to survive on. I recommend setting the baseline here and make receipt at this level
unconditional on participation, which will increase some benefit costs, reduce very slightly monitoring
costs and possibly reduce costs in other jurisdictions, including relieving stresses on our emergency
support systems. At the same time as this, introduce a second tier of OW that is either participation-
related (as at present or more demanding) or time-limited: The lowest level of the Market Basket
Measure in Ontario or 60% of the minimum wage at 40 hours per week might be a reasonable
compromise between poverty and fiscal advocates.
If time-limited, clients would have access to this increased level for a period while they attempt
to find a job, six months is comparable to programs like EI. If they failed to do so, they would return to
the lower level and would not be eligible again either for an extended period of time (3-5 years) or
permanently. This creates a strong incentive to act, not quite the same as the one developed in the Self-
Sufficiency Project, but of a similar nature. It also avoids the cruelty under TANF in most of the United
States, where one is no longer eligible for any benefits after a certain period. A time-limited system
I use an extreme model, which is likely to overstate caseload increases that follow rate increases (Brydon, 2011). In this
case, my model predicts a 50% increase in caseload; combining this with the 33% rate increase gives a doubling of OW
benefits. An increase in administrative and employment benefits is not included as partial compensation for the conservative
would also allow the role of caseworker to switch from a combined enforcement and support to strictly
support, potentially improving effectiveness and efficiency in the system.
Alternatively, this could be introduced as a separate system which aims to prevent people from
coming onto OW, to fill some of the gap left by EI (Issue 5, Question 1). Under this system, eligibility
rules should be much looser (e.g., very high or no asset limits) and access would be simpler, allowing
those who have lost a job, but do not qualify for EI, to bridge the gap to another job without some of
the challenges brought on by the OW system (the investment of time in following rules and learning to
live on the very meagre resources).
There is much more that could be done and I am sure that the Commission will ponder a
number of solutions. However, given that I have used at least my share of space, I will conclude simply
by summarizing the recommendations made above:
1. Evaluate social assistance based on prevalence and severity of poverty; employment rates of
participants in the short- and medium-term; and total cost.
2. Reduce Effective Tax Rates for single parents on OW.
A child care program, similar to that in Quebec, is the most effective way to do this
A provincial extension to WITB could also address this problem
3. Increase rates for some or all OW recipients.
A “second tier” could be introduced that is either time-limited or participation-related to
reduce the overall burden of poverty and encourage leaving the system early.
4. Change role of caseworker from enforcement and support to support alone.
Time limits likely to be more effective than behavioural requirements.
I look forward to the Commission’s recommendations.
M.A. Candidate, Economic Policy
Brydon, R. (Mimeo). The Price of Reducing Poverty: Welfare Rates, Caseloads and Costs. Available
from the author on request.
Drummond, D. and Manning, G. (2005). From Welfare to Work in Ontario: Still the Road Less
Travelled. TD Economics Special Report. Available at: www.td.com/economics/special/welfare05.jsp
Fortin, P, Godbout, L and St.-Cerny, S. (2011). Economic Consequences of Quebec’s
Educational Childcare Policy. Early Years Economics Forum. Available at:
Laurin, A. and Poschmann, F. (2011). What’s My METR? Marginal Effective Tax Rates Are Down –
But Not for Everyone: The Ontario Case. C.D. Howe Institute. Available at:
LeRoy, S. (2004). Don’t Abandon Successful Welfare Reforms. Fraser Forum. Available at
Poschmann, F. (2004). Background Issues: Marginal and average effective tax rates in Ontario. C.D.
Stapleton, J. (2009). Time for a “Made in Ontario” Working Income Tax Benefit. Institute for
Competitiveness and Prosperity and Open Policy Ontario. Available at: