MR. KOJO NNAMDIFrom WAMU 88.5 at American University in Washington, welcome to "The Kojo Nnamdi Show," connecting your neighborhood with the world. Later in the broadcast, a carwash controversy in Arlington. A local government cracks down on charity carwashes, citing environmental concerns. But first, a new tool to address vexing social problems. District agencies and nonprofits have spent years trying to tackle teen pregnancy. And while the rate of unplanned pregnancies has declined citywide over the last decade, it's held stubbornly high in the lowest-income neighborhoods.

MR. KOJO NNAMDISo last month, D.C. unveiled a new experimental way to attack the problem, a Social Impact Bond, also known as a Pay for Success Bond. Here's the idea: Investors and banks provide upfront cash for nonprofits, experimenting with new ways to solve a given social problem, whether it's unplanned pregnancies or homelessness. If those interventions are successful, the government pays investors back with interest. If they fail, the investors lose their money and the government is off the hook.

MR. KOJO NNAMDIThese bonds could provide a new tool for investors to do good, while doing well. But some nonprofit veterans are skeptical. Joining us in studio is Rick Cohen. He is national correspondent for Nonprofit Quarterly. Rick, thank you so much for joining us.

MR. RICK COHENMy pleasure.

NNAMDIJoining us from studios at Harvard University is Jeffrey Liebman. He is a professor of public policy at Harvard's Kennedy School of government and director of Harvard's Social Impact Bond Lab. Jeff Liebman, thank you for joining us.

MR. JEFFREY LIEBMANIt's great to be with you.

NNAMDIJeff, New York City became the first in the U.S. to experiment with a Pay for Success Bond in 2012, partnering with Goldman Sachs to spend $10 million on rehabilitating prisoners. If Goldman can reduce the number of people who return to jail by 10 percent, it will get its money back plus interest. Now D.C. is looking to do the same with teen pregnancy. Can you explain, what are Social Impact Bonds and how do they work?

LIEBMANYes. This tool is designed to overcome the existing barriers to investing in preventative social services. When I was a budget official at OMB, almost every week someone would come in to see me and try to get me to fund their program and say, if you just fund my program, you'll save money down the road. And I'd say, how do I know? What's the evidence? And they would tell me some story about two people whose lives they'd transformed. And the conversation would end there, because there wasn't evidence.

LIEBMANWhat's happening with Social Impact Bonds, is the conversation keeps going. Because the provider is saying, here's the deal, they're saying to the government. You only have to pay us if this works. We think our program works. What we want is a contract that says, if we really can reduce the rate of recidivism or house the homeless or improve the outcomes for -- individuals can -- young children showing up to kindergarten, then you pay us. But if it doesn't work, you don't have to pay us. And that makes it possible for skeptical budget officials to go ahead and say, okay, we'll take a chance on these preventative programs.

NNAMDISo governments are paying banks and investors for solving problems. This money must be budgeted in advance. And the government could end up owing banks more than if they had just spent the money themselves. What do these bonds offer that traditional government spending doesn't?

LIEBMANWell, really what they're offering the taxpayers is a money-back guarantee. If the programs don't work, the government doesn't have to pay and the taxpayers don't have to pay. And so, on the one hand, it pays a return if it does work, but it doesn't -- the government doesn't have to pay if it doesn't work. And so, on average, it basically balances out. But it's, I think, also doing something else that's pretty fundamental here. When you're scaling up a successful social service, typically what we'd do is fund things one year at a time.

LIEBMANBut if you're really going to try to tackle a hard social problem and work with a service provider to go from, say, serving 100 people a year to serving 500 people a year, you want to do that on a multi-year basis. And these are four- to six-year contracts that actually build in enough time to do a successful job of scaling up social services. And that's next to impossible to do with conventional government funding, which typically operates year-to-year.

NNAMDIIf you'd like to join the conversation, you can call us with your questions or comments at 800-433-8850, or by sending email to kojo, K-O-J-O, @wamu.org. You can send us a tweet @kojoshow. Rick Cohen, this, on the surface of it, certainly seems like a win-win both for the government and for taxpayers. But are there other considerations, other factors that need to be taken into account?

COHENIt does sound like a win-win. It's a very attractive kind of concept that basically addresses a lot of the concerns that nonprofits and government really have. There are probably limitations. I mean, the reality is that -- and Jeff is the real expert on this -- that there are only a few of these in the country. I think four are actually running in the U.S. right now. Most of them -- there is a project on recidivism, which is the J.P. Morgan one, and so forth. There's a project in Utah dealing with early-childhood education.

COHENThey're based on a model that comes out of the U.K. involving prison recidivism at the Peterborough Prison in the U.K. The concerns that I think people have are more a matter of tempering down the exuberant expectations of what these can accomplish. When you hear language like money-back guarantee, people get a little concerned and say, what is this really involving? So you look at the J.P. Morgan investment in New York, and you realize, well, J.P. Morgan invested $9.6 million, expects a $2.something million return on its investment, which is a pretty hefty 21 percent return.

COHENIn addition, for that innovation, the Bloomberg -- philanthropies of Mike Bloomberg, is going to guarantee $7.2 million of that. So that leaves J.P. Morgan risking $2.4 million to get $2.1 million. That's not a bad rate of return. So there's a little bit of a concept of how much return to give to investors. A second, probably more fundamental issue, is that these projects typically go -- these investments typically go into areas where there's a proven, demonstrable track record on the part of both the provider and the concept that they're demonstrating.

COHENAnd if you basically say, let's put two or three or four years' of money into this one project, there's a question of, well, if it's so good and if it's so proven, why not just do this as -- governmentally, why not actually make this a program that is much broader, rather than waiting for the results of this one project and then saying, how do we scale it up? In the Peterborough example, actually, the U.K. said, this was a great idea. Why are we waiting for the Peterborough project to pan out? Let's make this a broader program.

COHENAnd they went ahead and made a broader governmental program that basically supplanted, in some ways, the Social Impact Bond in Peterborough. So there's a question of how much to invest in one project? How much to guarantee a private investor? Why to limit a really good idea to one area or one site, when it might be replicated more broadly? And why not look for ways of funding projects or nonprofits in concepts that aren't necessarily as strongly proven as the kind of things that a private investor would go do?

NNAMDIProfessor Liebman, obviously Rick Cohen is a reporter. He's asked all of the questions that I would have asked. You can go ahead and answer them in whatever order you choose to.

LIEBMANTerrific. Well they were excellent questions. You know, what I think Rick is rightly pointing out is this is one tool for innovating and solving social problems. But it's not the only tool. And there need to be, you know, philanthropies continuing to need to work with providers to invent new solutions. And government still needs to pay for social service directly.

LIEBMANBut what this tool is doing is, at a particular place in the development of an intervention, it's allowing us to scale up that intervention and to rigorously assess whether it's working, so that governments can make the right decision about whether to permanently fund something versus whether something perhaps doesn't work or needs to be improved before it's ready to be taken broader. So I think Rick's point about this being, you know, of tempering down the euphoria and saying, this is a tool, but one that has to be targeted at the right point in the social-innovation space, is exactly right.

LIEBMANHe also mentioned the Peterborough history in the U.K. And I think it's actually interesting to look at what happened there, because in some ways, it shows that they were pretty clever and foresighted in thinking about how to set up that Social Impact Bond. They set up a six-year project, but they didn't do it as all-or-nothing based on the six years of service delivery. They separated it out into three two-year cohorts. So people get served for the first two years and then their payment's made based on the outcomes for the individuals served during the first two years.

LIEBMANAnd then people get served during year three and four, and then payments happen again, and then similarly after five or six. And what's happened in the U.K. is they're not going to do the last two years. They're going to do the first two cohorts, but not the third one. And because they were clever enough to design this thing in these separable pieces, even though there's been policy development, policy reform -- the government is heading in a new direction more broadly on criminal justice policy -- the Social Impact Bond itself was robust to this, because it built in the possibility of stopping if it was no longer the right tool in the new environment.

LIEBMANAnd, you know, you wouldn't really want to lock in a particular intervention, you know, for six or eight years with no possibility of changing it if the world might change in the interim. And I actually think the U.K. government did a very good job in designing that Social Impact Bond such that when the policy environment changed, they were able to, in an orderly way, make it so there will be payouts and they'll be learning from the first four years, but the last two years aren't going to happen.

NNAMDIIn case you're just joining us, we're discussing Social Impact Bonds with Jeffrey Liebman. He's a professor of public policy at Harvard's Kennedy School of Government and director of Harvard's Social Impact Bond Lab. And Rick Cohen, national correspondent for Nonprofit Quarterly. If you have questions, give us a call, 800-433-8850. Does this sound like a good idea? How would you introduce innovation into the nonprofit sector? What do you think? Are businesses more effective than nonprofits at solving problems? 800-433-8850. You can send email to kojo@wamu.org.

NNAMDIRick Cohen, one point that Jeff Liebman has made more than once is that Pay for Success Bonds free up money to be spent over the long term on tough problems. Something that governments have a hard time doing these days. What do you say to that? Could this be an effective way to take more long-term approaches to solving problems?

COHENI don't think government is prevented from doing long-term approaches. I'm a former government official myself, which is why I'm actually attracted to the whole discussion here. And also I worked, by the way, I did financing for the Local Initiative Support Corporation, which I know you know well.

NNAMDISure, LISC.

COHENAnd the Enterprise Foundation. And one of the key issues here is not just the long-term funding, which is an issue that Jeff is correctly raising. I think one of the bigger issues is the front-end funding. And one of the attractions that I think of a Social Impact Bond, and potentially other kinds of mechanisms as well, is to provide nonprofits and government with that front-end capital to do the planning and development to actually engage in risky propositions. The challenge here is that the kind of examples of Peterborough in the U.K. and others are often dealing with proven examples.

COHENYou want to find mechanisms to help those less well-known examples, those less well-known nonprofits, find capital to take risk and experiment with areas that are not as well proven. And I think that's the gap in this mechanism. It's a mechanism that looks to be very strongly evidence-based, proven models, with strong government support to say, yep, we'll buy into this model, you know, provide potential payoff of some significance at the end of it. But the models that provide the financing for the less well proven, highly risky, not well capitalized groups, really need to be developed.

NNAMDII think we have such -- one such example. But we'll get to it after a short break. We'll take your calls, however, during the break at 800-433-8850. You can still call or send email to kojo@wamu.org. Should Washington D.C. bring in a business approach to reduce teen pregnancy? You can also shoot us a tweet @kojoshow. I'm Kojo Nnamdi.

NNAMDIWelcome back to our conversation about social impact bonds with Rick Cohen, national correspondent for Nonprofit Quarterly and Jeffrey Liebman, Professor of public policy at Harvard's Kennedy School of Government and director of Harvard's Social Impact Bond lab. I'm going to turn it over to our listeners for a while here.

NNAMDIWe received an email asking specifically about this, about the motivations of investors. Jeff Liebman, why would somebody as our emailer -- why would somebody want to purchase one of these bonds and what is the motivation of the financial institutions? What do they stand to gain?

LIEBMANWell, that's a great question. If you talk to the financial firms that are getting involved such a Goldman Sachs in the New York State one or Bank of America, Merrill Lynch in the -- I'm sorry, I just said it backwards -- Goldman Sachs in the New York City one and then the Massachusetts one in Bank of America and Merrill Lynch in the New York State one. What they tell you is that they have high-net worth customers who would like to put some of their portfolio to work in socially-minded investors.

LIEBMANAnd that the reason they're getting into the space is frankly completely in the firm's self interest. If their high-net worth individuals want to be putting some of their money to work to do good, they better be offering that product to those individuals if they want to keep them as their clients.

LIEBMANAnd so they say that they're responding to a market demand particularly among a new generation of high-net worth individuals who perhaps have made their money in a high tech or data-driven sector of the economy and would like to know that when they're making their -- when they're putting their money to work for good that they're doing it in a way where there'll be a rigorous assessment of whether it really had the impact they were hoping for.

NNAMDIAnd we got this tweet from Shane Farthing, Rick Cohen. "Is the social impact bond D.C. is implementing mostly filling accountability needs due to poor procurement oversight rather than a fiscal purpose?" I think what Shane may be asking in 140 characters is whether this program is being rolled out because of flaws in the existing nonprofit sector in D.C. or whether this is just about saving money.

COHENWell, I can't attribute motivations to what's behind the D.C. project. I would say that there are plenty of nonprofits that know this area well and do a good job. I think there's a little bit of an issue of where do they find capital to get the upfront money to actually start something or try something new. And that may be one of the motivations, but that's an issue for nonprofits in general.

COHENAnd that's an issue that I think ought to be attracting high-net worth individuals to invest actually not just in projects to get a guaranteed return like these very strong ones, but to risk their money on the projects that might be less guaranteed and more innovative in terms of their testing approaches that are not well documented yet at this point but really going through areas that are really needed.

NNAMDIHere is Jeffrey in Washington, D.C. Jeffrey, you're on the air. Go ahead, please.

JEFFREYHi. Now I agree with what he just said in terms of social bonds. And I think it's a great idea. I think that this should be a collective partnership between the community in which it's being served by the social bonds. And my suggestion is to build a group capital among individuals where there's a partnership where the individual community is educated through e-activism , i.e. social media, they even engage in the meaning -- dialogue with that particular investor to see what the needs are of the community and build a partnership around that.

NNAMDIWell, when you say the community, you seem to be thinking about a geographical space. What is the -- what if the community is pregnant -- teenage pregnant girls? What if the community are ex -- people who have been incarcerated and trying to avoid recidivism?

JEFFREYWell, that's a very good point. And I'm writing my dissertation on that very same topic, the impact of social capital and social activism.

NNAMDIOkay.

JEFFREYSo this -- so I think the problem is that, yeah, there's a lot of well-meaning programs, but the vast majority of the recipients of these programs are not educated as to what they were -- I mean, what the services are. I have a understanding of the power of a group to come together an effect they all change. It's one thing to pour money into a project, and by all means I support that, but the individual as a whole needs to come up with a means to control and (unintelligible).

NNAMDIOkay. Here's Rick Cohen.

COHENWell, I think that's a very valid question he's raising but I would raise it not from the question of educating but more from community involvement. One of the challenges in the social impact bond phenomenon, that I know Jeff is very aware of, is the idea that the availability of private capital will influence government prioritization.

COHENTo move to the areas where private capital, where those high net worth investors are interested in, it potentially take away from not necessary negative in terms of capital but take away attention from other areas where you don't have that private capital dictating or influencing thinking.

COHENSo there's a real issue for the community, whether geographic community or a constituent community to make sure that they are articulating their needs and that their issues don't get lost in the direction provided to government by private capital.

NNAMDICare to comment on that, Jeff Liebman?

LIEBMANWell, it's -- I think it's an interesting hypothesis but I think if we look at what's actually happening around the country, and we are working with eight governors and two mayors right now, our Harvard (word?) is, and so we're working with Republicans in South Carolina, Michigan and Ohio and Democratic governors in Colorado and Illinois and Connecticut and Massachusetts and New York.

LIEBMANAnd I think what you're actually seeing is a very wide range of social problems being tackled with this tool. We're seeing not only recidivism but also homelessness, troubled youth, early childhood, both really early prenatal interventions and also preschool interventions, preventive health care, diabetes prevention type of areas.

LIEBMANSo I think we are seeing this tool being used in a very wide range of policy areas. And it's not really I think the case that the private money is directing us to a very narrow set of social problems.

NNAMDIMany of these big complex problems are difficult to quantify. The question comes up about how outcomes are measured. For instance, over the past decade within the education arena, there's been a huge focus on testing and measured outcomes which makes it possible to assess success and failure. But it's also created incentives for different actors to cheat, whether it's teachers or principals, because they're assessed based on those benchmarks. Is that a concern with these new initiatives, Rick Cohen?

COHENThere's an old principle by the social scientist Donald Campbell who said that when you rely on -- the more you rely on quantitative indicators to direct your social policy, the more you're subject to those indicators being corrupted and the social processes beneath them being corrupted. And I think you see that in any number of spheres.

COHENI mean, you know, ranging from education to the recent VA -- the Veterans Affairs scandal regarding incentives give to VA hospitals to show that they were meeting quantitative measures about getting people appointments.

NNAMDI...secret waitlists.

COHENThe waitlists so that there's a danger here. And I know that Jeff and others are working on these kinds of issues of how do you make sure that the third party evaluations that will be done to prove outcomes are actually legitimate, that don't get corrupted by a process or corrupted by the need to show results, and actually can be relied upon? That's a tough, tough issue which I think to some extent narrows the range of activities that can be pursued by social impact bonds.

NNAMDIDoes it, Jeff Liebman?

LIEBMANThat's exactly right. So when we're working with a state to decide and help them figure out which policy area is a good fit for this, almost certainly the number one issue that crosses things off the list are issue areas where there isn't a good way to measure holistically what it is they're trying to accomplish with that policy. Because if you can only measure 10 percent of what they're trying to accomplish, you might get really good performance in that 10 percent but fail on the other 90 percent of what you're trying to accomplish.

LIEBMANAnd so when choosing policy areas to use this tool in we really have to limit it to things where we really think we have a good measure of what it is that is trying to be accomplished. And then, as Rick said, we are building in accountability unlike the VA where basically the same people who were -- they were trying to be held accountable for controlled the data. In all of these social impact bond projects we get an independent evaluator who is neither part of the government nor part of the investor team or the service provider team to be in charge of the data and assessing what happened.

LIEBMANAnd then in most of these states we add another level, an auditor, who looks at the work of the independent evaluator and checks the data to make sure that the outcome's really achieved. And only when an auditor certifies it does the government make payments.

NNAMDIHere is Shroda in Washington, D.C. Shroda, you're on the air. Go ahead, please.

SHRODAHello, Kojo. Thank you for having me. This is your fan and friend Shroda from M Street Village. We are a nonprofit organization that provides services for homeless and low income women.

NNAMDISure.

SHRODAAnd I thank you for this conversation which is wonderful. And I'm very interested in these thoughts about innovation for how we make our social services sustainable in the long run. I really strongly believe that we've got to move from a charity to an investment model in order to continue to provide these important services.

SHRODASo that said, my question really is along the lines of exactly what you all are discussing which is how would, for an organization like ours, where do we find the right benchmarks? So if we're serving a particular population of women who need, say, permanent supportive housing who may have disabilities or long term current conditions that require support, where do you find the right measurements and how do you benchmark yourselves in order to understand that what the success would be that would provide that return on investment?

SHRODAMy other question is, how does a nonprofit get the upfront funding under this -- using social impact bonds? You know, does it -- I'm not familiar enough with the way they look to understand how we could work with a government intermediary to get the initial capitalization.

NNAMDIShroda is from M Street Village. This is a continuation of a conversation that she and I had over two years ago. Jeff Liebman.

LIEBMANOkay. Well, let me start with the second half of the question which is the upfront funding. What I think -- I think the reason that service providers have -- are attracted to this model is that typically today when government funds service providers, it pays for slots in programs, sort of a certain number of people to be served.

LIEBMANAnd government typically doesn't give providers any money for their management overhead, their data systems, their infrastructure to actually innovate and grow. The government just says, we want to buy slots and programs. And often governments don't even pay the full marginal costs of delivering slots and programs under our current mechanisms. What they do is they pay .8 of that and say to the provider, go raise the other 20 percent from philanthropy.

LIEBMANWhat's happening in these social impact bond projects where successful social entrepreneurs are able to expand their projects is that they're pricing into these projects all of the resources that are needed to expand, not just the marginal cost of serving extra folks, but the management oversight that's going to be needed too to do that to scale appropriately. And the investors are providing the upfront operating funds such that at the beginning of the project when you need the funds for building your management capacity, hiring the new people to expand the funds are there.

LIEBMANAnd so it's I think -- and you get this commitment that the project's going to happen for four to six years, which is -- if you're really going to take that chance on expanding is something a provider really wants. And so I think it's really a much more predictable source of funding from the provider's side than many sources of government funding today.

NNAMDIRick Cohen.

COHENWell, I think that's only part of the story though. And one issue is the overhead question where that's a legitimate concern for all government-funded nonprofits. But then they ought to be fighting for better overhead and as a change in OMB regulations, that Jeff as a former OMB official certainly knows, to start guaranteeing nonprofits a better overhead rate. And he's exactly right that nonprofits shouldn't be cheated on the exact -- the real costs of delivering programs.

COHENBut in addition, there's even the upfront money that a group like M Street Village -- which I love M Street village by the way -- but there's the upfront money to even plan and think about this stuff which is often not there, even not even built into the financing of some of these innovative financing mechanisms. So recoverable grants, upfront funding, philanthropic capital in general is really needed. And many nonprofits find that really upfront money, which is preprogram money, hard to find and very difficult to even build into these projects particularly when they don't have a lot of walking around capital to work with.

COHENSo I think that the smaller nonprofits, the ones working on projects that are less quantitatively demonstrable and less evidence-based are going to find themselves a little bit out in the cold in this kind of model.

NNAMDIAnd for those who may have questions specifically about what social action bond was actually issued by the district government, we reached out to Social Finance, the organization that is running the program for D.C. And they informed us that it is still in process but they are in the final stages, so the bonds have not yet been issued here in the District of Columbia.

NNAMDIFinal question for you, Jeff Liebman. Over the next few years we'll begin to see the results of the nation's first social impact bonds. If teen pregnancy and illiteracy rates plummet, we'll be likely seeing a lot more of them. Or if they could turn out to be more expensive and less effective than hoped they could be phased out. What do you ultimately see as the future of pay for success bonds?

LIEBMANI think what you say is exactly right. This is an experiment. One possible outcome is we do two dozen really neat projects but they end up basically being one-off projects and we discover that there's a better way to foster the development of solutions to social problems. Another possibility is this tool really does continue to seem to be a unique tool for solving these problems. And we suddenly see, you know, instead of having, you know, a few -- 20 or $30 million of these, which we have right now, we could see 4 or $5 billion of this three or four years from now if everything goes right.

NNAMDISame question to you, Rick Cohen.

COHENWell, I can tell you what I'd like to see, which is I would love to see guys like Jeff and the people he works with thinking about how to develop this model not for the big proven nonprofits that have programs that you and I would invest in tomorrow because we know they're good, but for the areas where the evidence is a little bit unclear for the problems that are less well-defined and for the nonprofits that are not the big ones, to say how do we provide capital to those nonprofits that are often at the outside of the capital markets looking in.

NNAMDIIt's a work in progress. We'll be following it. Rick Cohen is national correspondent for Nonprofit Quarterly. Rick, thank you for joining us.

COHENMy pleasure.

NNAMDIJeffrey Liebman is a professor of public policy at Harvard's Kennedy School of Government and director of Harvard's Social Impact Bond Lab. Jeff, thank you for joining us.

LIEBMANYou're welcome.

NNAMDIWe're going to take a short break. When we come back, that car wash controversy in Arlington, a local government cracking down on charity car washes citing environmental concerns. What do you think? Call us, 800-433-8850. I'm Kojo Nnamdi.