Public Safety Committee on June 5th, 2014

[Randall Garrison, NDP, Esquimalt—Juan de Fuca, BC]

Thank you very much, Mr. Chair, and thanks to both of our presenters today.

One of the things you've really done is.... We started off saying there are few of these in Canada. Now you're telling me there are few in the United States. There are very few at all.

I think Mr. Butler was talking about the quadfecta, and the quadri-loser state is kind of where we are here. The government on the quadfecta side tends to emphasize those possible gains. On this side we've been asking more questions about the possible quadri-loser side. So it's useful to think of it in that way.

Both of you said something about where you're really working on these projects is in the margins, where there's a margin available, like high recidivism or high crime rates, where you can get value out of reducing that.

In a situation like in Canada, where we have lower crime rates, lower rates of recidivism, doesn't that mean we probably have lower opportunities for these kinds of projects? I'm not expecting you to know the Canadian situation, but I'm forcing you to accept that premise and then say whether you believe that would be a possible difference between Canada and the United States.

[Prof. Jeffrey Liebman]

In just about every project we've been involved in, we've had to focus the projects on the highest-need and highest-risk individuals. If that is an early childhood intervention, you want to find the neighbourhoods where the most people are not showing up to kindergarten ready for school, and focus resources there. Or if it's a recidivism project, you want to find the folks who have the highest risk of going back.

It's often the case that if you just take the entire population you could potentially serve and try to design a project around that, it just doesn't work in terms of the economics. You really have to be very focused on identifying the highest-need individuals.

[David Butler]

The only thing I would say is that if what you're saying is true, the implications are that you really don't have to improve your programs. If you're saying that there really isn't an opportunity for marginal impacts, then the status quo is acceptable to you. I don't think you want to say that.

To me, it seems as if it's always in comparison to where you are, but I don't think you would say you don't think you can do any better, that you can't run better programs, you can't have better results. Canada is a great country, but is it utopia?

[Randall Garrison, NDP, Esquimalt—Juan de Fuca, BC]

No, I wouldn't say that. But we have a group called Circles of Support that works with high-risk sex offenders. It's a partnership with the government, but not a social impact bond. They're achieving 80% success rates for non-reoffending. It's spectacularly high.

[David Butler]

That's spectacular; you're right.

[Randall Garrison, NDP, Esquimalt—Juan de Fuca, BC]

So we do have some examples here, without using these impact bonds.

I want to go to a question on two things. Obviously government pays. If these work and they achieve a higher success rate, then the government pays more for these projects. My question is about the value that's being added there. Why couldn't government have run the same program and achieved those same good outcomes? Why is the government paying somebody else an interest rate up to 20% to achieve something it could have done itself? What's the value added there?

[David Butler]

I'm stuck on this one. We have a long history in the States of philanthropically funded, relatively small-scale programs that have looked very good and have been evaluated and have shown some positive effects. But government, for the most part, has not taken them up because of the fear that doing it at scale would not be successful. What this does is mitigate some of those risks associated with doing it at scale.

It's a legitimate risk to worry about, because we don't know that it's going to be as successful, when you have done something on a small scale and you expand it to a larger scale,.

Here, the upfront costs are covered by the investor. The government pays back only if it's successful. Then if it's successful, the government, and not the investor, reaps the rewards in the future of having a program that's been evaluated and that's been shown to work.

[Randall Garrison, NDP, Esquimalt—Juan de Fuca, BC]

Go ahead, Mr. Liebman.

[Prof. Jeffrey Liebman]

I think Mr. Butler answered that one well.

But I think you're absolutely right that if you have a terrific program, like the one you described a moment ago that's working great, that's not where you use this tool. This is a tool to apply to a place where you don't have the results you need and you need a better way to collaborate with the service providers to get better results. That's where you should aim this tool.

[David Butler]

I agree with that as well.

[Randall Garrison, NDP, Esquimalt—Juan de Fuca, BC]

I want to turn to the question of accountability.

The projects are working with people who are, by their nature, at risk in some way, especially if we're talking about the criminal justice field and dealing with people who obviously have extensive social needs. Who are the projects accountable to? If individuals feel they have not been well treated in the program or if they demonstrably have not been well treated, how is that dealt with?

In Canada, if it's a government project, we have accountability mechanisms. We have the correctional investigator, as an officer of Parliament, who investigates what goes on.

What happens with these projects in terms of accountability?

Mr. Liebman, perhaps we can start with you.

[Prof. Jeffrey Liebman]

That's a great question.

When you're setting up one of these projects, the government is choosing what the contract terms should be and whether or not to enter into one of these. You can build whatever accountability mechanisms you want into these contracts. In general the thing people are not trying to do is tell people exactly how to deliver the services. The idea is to give them some flexibility as long as they can achieve results. But in terms of provisions for protecting populations against bad outcomes, there is no reason why you can't build your traditional provisions into these projects, if that's a priority for you.

[Randall Garrison, NDP, Esquimalt—Juan de Fuca, BC]

When you're talking about traditional accountability mechanisms, those apply to government; they don't apply to charitable or philanthropic organizations. For instance, in our case we have a correctional investigator as an officer of Parliament. He examines only what's done by the public agencies. He doesn't have the ability to examine what's done by private or philanthropic agencies.

[Prof. Jeffrey Liebman]

I see.

We're using this tool in the United States mostly in areas where government is already procuring social services from non-profit service providers. Often, even in those circumstances, government will build in protections—the right to investigate or audit or whatever—and there is no reason why those kinds of provisions can't be part of a pay-for-success contract as well.

[The Chair, Conservative, Daryl Kramp]

Just briefly, Mr. Butler, do you have something?

[David Butler]

Yes, that's one of our responsibilities, frankly, as the intermediary on this project. We do a lot of monitoring of what's happening on the ground. We talk to the program participants. Fortunately, we haven't heard much of this, but if they have complaints about the service—if they don't feel they're being treated fairly—we have the responsibility of trying to do something about that, either through the city's department of corrections or the mayor's office, or wherever. That's one of the ways we build in that control.

[The Chair, Conservative, Daryl Kramp]

Fine, thank you, Mr. Butler.

Now, Mr. Payne, you have seven minutes, please.

[LaVar Payne, Conservative, Medicine Hat, AB]

Thank you, Chair.

Thank you to the witnesses for being here today, as well as being here virtually.

I find this area is really interesting, particularly because I see some innovation happening here. I think that's quite positive in terms of social finance. It seems we're bringing together a number of different sectors to, hopefully, create some better outcomes.

I wonder if both of you could comment on the flexibility of service providers, and on being able to tailor the program to specific individuals.

Mr. Butler, you go first.

[David Butler]

Our particular project is not tailored to individuals. It's essentially an intervention that is provided to and the same for everybody. It's a course in cognitive behavioural therapy that's designed to change thinking. You can take into account the needs of individuals, but there's a set curriculum that's provided in this program.

I do think, despite that, we've had to, in conjunction with the service provider, adjust this intervention to address the particular population we are working with in the jail. We've had to make some changes, but we also wanted to test something that was relatively simple and easy to measure the effects of, something that didn't have too many complicated program components, certainly for the first one on the block, and something that had some good evidence of success in the past.

[LaVar Payne, Conservative, Medicine Hat, AB]

Dr. Liebman.

[Prof. Jeffrey Liebman]

In some models being tested—and I'm thinking particularly of the Roca model, which is the one being used to work with the young men in Massachusetts—the model really is explicitly designed to meet the individuals where they are and to figure out which services they need. Some individuals need educational services before they're ready to get job training services. Other people are ready to go directly into a job. I think a number of the service providers involved in these projects are very good at figuring out which services go to which individuals.

I would say in some cases these projects themselves, because they tend to work though the government referring a fixed population to the service providers rather than having the service providers going out and doing the recruiting themselves, the individuals are causing the service providers to have to get better at figuring out how their existing model is the right solution and how they have to add to their existing model to be able to serve the whole population that they've been assigned.

We're seeing a lot of the kind of flexibility you're talking about, but I think this model is actually causing more of that to be created in the service providers, because they're on the hook for the outcomes, so they have to figure out innovative solutions that will get better outcomes.

[LaVar Payne, Conservative, Medicine Hat, AB]

Thank you.

Mr. Butler, in terms of your program, you talked about your two years on track to achieve your goals and to reduce recidivism by 10%, and you talked about a net impact evaluation. I wonder if you could describe this a little further for us. What kind of checklist are you using to make sure you're actually achieving the results you desire?

[David Butler]

Actually tracking the net impact, which is done by an independent evaluator and not by us, is relatively easy. They're going to take the data regarding the number of days in the program. Once people leave jail, they're going to track them for two years, count how many of them come back, and count how many days they spend in jail when they come back. They compare those numbers to their comparison group, which is the same-age population over the prior five-year period. So that's pretty straightforward.

But in a situation like that, how do you track whether you're on target to meet those goals? We're looking at whether the program is being implemented with fidelity, because we know that it has worked in the past, so it has to be done in the way it's supposed to be. Then we check the level of participation. How far along in the program are they getting? How many are completing it? How many are getting to step 3 or step 7 in the step-based program? Based on that, we have some analyses we do that tell us whether they appear to be on point to being successful. But, you never know. At the end of the day, the comparison could surprise us.

That's the way we do it.

[LaVar Payne, Conservative, Medicine Hat, AB]

Dr. Liebman, you also talked about early childhood services and homelessness as well. Do you have any details on those kinds of programs you're working with that would give us some indication as to what kinds of folks are involved? What kinds of outcomes are you expecting? What kinds of measurements are there to make sure you're meeting those targets?

[Prof. Jeffrey Liebman]

That's a great question. On the early childhood ones we're seeing two very different kinds of programs being developed in multiple states. One is really early childhood, like six months before birth through the first two years of life, where the goal is to make sure the mother has a healthy pregnancy, and the first outcome that is tracked is the share of births that are low birth weight with the hope of reducing health care costs by having fewer low birth weight births, and then lower medical expenditures in the first couple years of life because the children were born healthier.

So one set of projects is very early childhood, and it's primarily with these kinds of health outcomes in mind at least in the short term, and then the other type of early childhood intervention is for children right before they're going to enter school, trying to spread high-quality child care to low-income communities where the hope is that by making sure that the children arrive at school ready for school rather than behind other children, there will be lower what we call “special education expenditures” helping people catch up once they get to school.

What makes these two kinds of projects somewhat different from the recidivism ones is that some of the benefits in these projects are much longer term. You help a child when they're five or six years old and you hope they have a lot of good things happen to them, so that they're less likely to commit crimes when they're a teenager, less likely to be a high school dropout, more likely to have higher earnings. So you end up in discussions with governments about whether they'll pay for these longer term outcomes too based on short-term results, or whether they'll only pay for the savings that have actually been accrued during the first few years.

That's what's going on in early childhood. I don't want use up all your time. I could tell you about homelessness too, if you'd like, or we can go onto another question.