Memory-lane time today. Because I just realized it was 25 years this week that the Canada Millennium Scholarship Foundation began inking deals with provinces to give away what turned out to be about $3.7 billion in scholarships. And I think there are some lessons that the folks in Ottawa who are fleshing out the (IMHO) poorly-conceived “national” school lunch program announced a few weeks ago. 

The story of the Canada Millennium Scholarship Foundation is hard to explain to anyone who didn’t live through the 1990s. To start at the beginning: in the fall of 1997, the Government of Canada was still reeling from the massive cuts imposed in the 1995 budget; however at the same time, buoyed by a crap-load of new revenue from capital gains taxes (a welcome side effect of the early dot com boom), the feds were also for the first time in 30 years about to run a surplus. Meanwhile, two other issues were bubbling along. The first was growing concern about rising levels of student debt (it really was rising quite quickly back then), and the second was intense irritation in the Prime Minister’s Office for having got zero credit in Quebec for acting on a long-standing nationalist demand to cede much of the skills and training portfolio to exclusive provincial control. Which meant there was a constellation of forces lining up within the federal government: money to spend, a desire to annoy Quebec, a bunch of cowering line departments, and a rampant Finance Department that basically could do anything it wanted. Cue the Millennium Scholarship Foundation, whose early evolution I described back here (see also my colleague Joseph Berger’s work on the same subject here).

I’m not going to re-hash that story today; but I do want to talk about some of the Foundation’s program parameters and their effects on outcomes. In particular, I want to talk about what happens when a program has to simultaneously “work with provinces” and work from a set budget. Let’s start with the latter condition. Most transfer or welfare programs in Canada work on the basis that as long as a resident meets certain qualifications, they will receive some sort of benefit, regardless of how many other individuals meet the criteria. This makes budgeting a bit challenging: depending on the size of the client base (which is never stable) the amount spent will vary from one year to the next. 

So, what happens if you have a set budgeted amount to give out? In the case of the school lunch program, it’s $200 million/year; in the Foundation’s case, it was a $2.5 billion initial endowment (plus interest/capital gains on investments) which had to last ten years. The Foundation opted to set annual amounts of $300M per year for the first five years, growing to $350M per year later due to better-than-expected return on investments). Think this through for a second. It means you can’t adjust your budgets to growing or shrinking needs. It means you have to adopt a rationing mentality.

Now, this intersects with the second parameter, which is working with and through the provinces. If you’re going to sign deals with provinces to make your program work, you need to tell the provinces how much money you are giving them each year. So there’s rationing job #1: you have to divide your annual budget in a way that the provinces accept. By and large, that means dividing the money based on provincial population.

Now, the question is: once you’ve divided things up into annual provincial allocations, how do you spend it? In the Foundation’s case, this was pretty easy. The primary political challenge it had to overcome at the beginning was to get Quebec to sign an agreement. Fortunately, the province set some public conditions that were actually meetable, thanks to Liberal MNA Henri-Francois Gautrin who in May 1998 proposed a motion in the Assembly (passed with universal approval) which said the province could work with the Foundation if it met three conditions:

1) Quebec would receive funds equal to its share of population (this was a given, as I’ve already shown),

2) Quebec could determine which students got the award, and

3) The new program would avoid duplication with the provincial student aid program.

So the challenge was to construct a deal for all provinces that would meet the second and third Gautrin criteria. But the answer was pretty obvious: both criteria would be met by using provincial need assessment criteria to work out who would get the awards. The problem was that by handing out money on this basis—that is, giving it to the students with the highest need—you ran into the problem that the provinces were already giving these same people money through their own grant or loan remission programs. If the Foundation were to come along and started giving the same people money, what happened was that existing funds were displaced and—in theory at least—captured as savings by the provincial government.

That’s bad, right? Well, fortunately, this was the age of the Social Union Framework Agreement, which was born of the federal-provincial negotiations that accompanied the birth of the Canada Child Tax Benefit, which had exactly the same problem. Basically, by transferring money to low-income Canadians, it reduced their eligibility for provincial welfare benefits. The solution was to allow provinces to do precisely that but require them to re-invest any savings in a limited number of ways that were related to relieving child poverty. So for instance, one province could put all its savings into expanding housing benefits for low-income parents, another could put money into services for children with fetal alcohol syndrome, etc. It was the same thing with the Foundation’s money: one province put all its savings into expanding its own student grant program, another would put money into transfers to universities for the purpose of freezing tuition, etc. The net benefit to an individual recipient of a Millennium Bursary usually was a lot less than the amount on the cheque, but the net benefit to students generally (including non-recipients benefitting via re-investments), was always equal to the amount of the provincial allocation. With some prodding from student groups, even Quebec agreed to this.

The problem for the current-day designers of new national programs is that they dislike the optics of the Foundation/SUFA approach because the provinces end up getting a lot of the glory for their re-investment announcements. The preferred mechanism now seems to be to announce a program and then sign deals with individual provinces committing each to near-identical targets (e.g. the expansion of $10/day childcare by a certain number of slots) as was the case with the Early Childhood Education accords of 2022. But the thing is, that strategy only worked because Quebec basically did not have to do anything to get the money. It’s not clear that the same approach is going to work on an area like school lunches. The federal government is right now insisting that all the money it puts in goes to expanding current programs—or rather, adding new recipients—meaning (as Paul Wells waspishly put it) that the federal money will only go to the country’s second-poorest set of children. It does not—publicly, at any rate—foresee any possibility of allowing provinces to capture benefits and re-invest them in a related field. 

Were the Millennium Scholarships a “national” program? Well, the merit-based scholarships were because the Foundation administered those directly. But the main set of bursaries, the ones meant for students in need? Arguably, yes, in the sense that a federally-endowed organization was approving cheques to individual Canadians (technically, the provinces “decided” who got the award, but under the legislation, the Foundation’s board has to approve each bursary, so every year each province would send massive lists of students to the Foundation for approval). But, arguably not, in the sense that its’ design-logic from the get-go meant the program looked somewhat different in each province; both in terms of the way money was applied and in the way that re-investment was conceived (this Parliamentary committee hearing from 1999 gives you a flavour of how things looked in different parts of the country). 

Now the question is whether it is possible to design a “national” program which accepts the logic of provincial allocations and implementation but denies much variation in implementation? Will it be a national program if Quebec does not participate? Will it be a national program if Quebec gets money but uses it unconditionally and does little to nothing to build up its school lunch program (the province’s current position)? It’s hard to say. But my advice would be that the Canada Millennium Scholarship Foundation provides evidence of how a national-ish program can work. A return to the principles of SUFA and the Millenium Scholarship Foundation are probably a more durable route to success than what the feds seem to have in mind right now.