Man in the middle
John Grossman brings government, providers, and money together for pay-for-success projects
By 
Photographs by Frank Curran
JOHN GROSSMAN calls himself an intermediary. It’s an unusual job description, but it’s one that captures what he does putting together complex deals that leverage money from the private sector to fund social service work for government. Grossman is the man in the middle, trying to address social ills by bringing together government bureaucrats, social service agencies, philanthropists, and corporate titans from major Wall Street firms.
The 48-year-old Grossman is the co-president and general counsel at Third Sector Capital Partners, a nonprofit company on the rise in the emerging field of social impact investing. Social impact investing is a catchall term that applies to investments designed to yield a profit as well as some form of social or environmental benefit. Third Sector’s niche is pay-for-success projects, where private groups put up the money for a program to address a specific social problem and get paid back by the government if the initiative is a verifiable success.
The approach is similar to what’s going on in health care. Instead of paying doctors and other health providers for each test or service they provide, insurers are starting to pay for successful outcomes. Health providers and insurers agree ahead of time on what it costs to care for various types of patients. If the providers can successfully treat their patients at less cost than the agreed-upon reimbursement, they pocket the difference. If providers fail to attain the desire outcomes — a patient gets an infection while being treated or has to return to the hospital for emergency care because of avoidable complications — they have to absorb the extra cost.
Grossman helped pull the deal together, coaxing all of the parties to agreement on a whole host of complicated issues. For example, the parties had to come up with a mutually agreeable way of defining recidivism success. They also had to put a dollar value on it. And they had to work out an arrangement for who would profit and how much.
Grossman is well-suited for the job. He’s a lawyer (BU), a finance guy (MIT’s Sloan School), and a veteran of state government, both at the attorney general’s office and at the Executive Office of Public Safety. He’s a good listener and a good communicator, both skills that come in handy when trying to bring diverse parties together.
When Grossman joined Third Sector in 2012, he was the fifth employee and worked part-time. Now he works full-time alongside 32 other people who are spread between offices in Boston and San Francisco. The employees are working on 37 projects as consultants, intermediaries, or a combination of both.
The business is growing quickly because of a convergence of interests. Government officials want concrete results for the taxpayer money they are spending. Social service providers are increasingly confident they can produce the desired results. And even though the profits aren’t spectacular, big players in finance are jumping into social impact investing because their clients are demanding it. The latest convert, Bain Capital, hired former Massachusetts governor Deval Patrick to head up a social investing arm of the company.
While in office, Patrick helped launch three pay-for-success ventures: the Roca project on recidivism, a homelessness project developed by the Massachusetts Housing and Shelter Alliance, and a Jewish Vocational Service initiative on adult education. The federal government is also investing in pay-for-success ventures, last year awarding Third Sector a $1.9 million, three-year grant to offer consulting services to state and local governments at discounted prices. Thirty-nine government agencies responded to Third Sector’s consulting offer.
“We were only able to choose seven of the applicants, so that speaks a little to the demand and increasing awareness of this,” says Grossman. “Several years ago we were knocking on people’s doors.”
I met Grossman in Third Sector’s offices on the 29th floor of the John Hancock Tower in space it rents from New Profit, a venture philanthropy fund and one of the investors in the Roca project. What follows is an edited transcript of our conversation.
COMMONWEALTH: How do you define pay for success?
CW: Government doesn’t do that already?
GROSSMAN: We certainly haven’t been measuring to date, and I would argue in many cases we haven’t been improving the lives of the people most in need. The war on cancer and the war on poverty were declared at the same time roughly. We’ve made phenomenal progress in the war on cancer; there are cancers that are now completely treatable and there are other cancers where we’ve extended life remarkably. The war on poverty by almost any measure we haven’t done much with. Our big picture hypothesis here is that the difference between those two is that we had a feedback loop in the war on cancer. We had experimental methodology where you could set out to do something, test it, and get an outcome. If it worked, you could develop it further. If it didn’t work, you try something else. In social programs, we basically fail to have that feedback. We do programs and we have no idea whether they work.
CW: But it seems like government is always testing a program with a pilot project.
GROSSMAN: We did it with Head Start in 1974 in Ypsilanti, Michigan [in the well-known Perry Preschool project]. But because it was government, we said everybody has to do it the way we did it in 1974, as opposed to constantly testing and iterating, which is what medicine does, which is what the private sector does. We’ve largely failed to do that in the social sector. Some people say when you test existing social programs now, only one in 10 are effective. If we could get that to 1.5 out of 10, that would be a 50 percent improvement. That’s an unbelievable change for the people receiving those programs and for taxpayers. I’m not telling you that pay for success is going to do that, but it is part of a larger movement that is asking about outcomes and seriously looking at performance measures. That’s the big picture for why we’re doing this.
CW: Why can’t government do that on its own?
GROSSMAN: Governing the state is hugely complicated and there’s only a certain amount of bandwidth that’s left for discretionary work. The state has made a lot of progress. Before the Romney administration came along, there was no reporting around performance. The Romney administration started something called Mass Goals, where each agency had a report card and three or four specific goals that they had to report on both to the chief executive — the governor — and to the public on a quarterly basis. Then the Patrick administration comes in and, like most new administrations, it said, well, that’s good, but we can do it better so we’ll start again. And that takes a while. The Patrick administration made some progress and did some good things, but I don’t think [the use of data] ever really got imbued in the system. I don’t think everybody at every level of management thought this was something of value to them. It became sort of a compliance thing rather than a management tool.
CW: But you need data to make pay-for-success work, right?
GROSSMAN: Roca, the provider we’re working with on this project, created a performance management system down to the level of the youth worker. They think the data they’re gathering helps them do their job better, so they’re not reporting because thou shalt. I mean there’s an element of that. If they don’t report it, their boss is going to say, where is it? But Roca thinks getting those performance metrics will make them better at doing their jobs. That’s the holy grail around any evidence-based management system. And they’ve done that really well. State government, I don’t think we have done that very well.
CW: How complicated was the Roca deal to put together?
GROSSMAN: It was complicated in the sense that there were only two other projects in the United States before us, one just by three weeks and the other by a year-and-a-half. So we had no go-bys. None of the details of those projects was public, so it was very complicated figuring out how we were going to do the project. We were also acutely aware that what we were doing was a model for future projects. We were setting a precedent, not just here in the Commonwealth but around the country. We were also doing something that’s complicated. It’s complicated to talk about measuring success and to attach financial consequences to that and to contract for that, none of which had been done before. Even if you and I were just to whiteboard the problem out without having to negotiate with someone, there’s a lot of thorny issues that you have to work through. Then you think about the strange bedfellows a project like this brings together – the funders include Goldman Sachs, the Kresge Foundation, Living Cities, the Laura and John Arnold Foundation, the Boston Foundation, and New Profit. And then there’s Roca, this Chelsea-based, sophisticated nonprofit; the Executive Office of Administration and Finance; and the Probation Department. Those are some serious cats that have to be herded to get to a consensus.
CW: When you negotiated the deal, wasn’t the Probation Department going through a federal corruption probe?
GROSSMAN: Yes. It was a sub-optimal time to try to do a project with the Probation Department.
CW: What was your role as intermediary?
GROSSMAN: While there was unanimity about wanting to build a good project, everybody speaks a different language and has somewhat different needs and different masters to whom they answer. It was our job to sit in the middle of all of those people and understand the needs of each of the parties and try to come up with solutions or solicit ideas for solutions and build consensus around them. It was very much an iterative process. It started without the funders at the table, building out the shell of a project with the government and with Roca. Then we tried to figure out what the funders needed. Our job was to play the other side’s role on almost every occasion. So I’d have one-on-one negotiations with the government and say that makes sense, I understand why you want that, but is it going to be fundable? And then you say to the funders, that makes sense, I understand why you want that, but is it going to be acceptable to the government or Roca, and how do we fix that? So that was our job.
CW: What makes you suited for that job?
GROSSMAN: At some point at the attorney general’s office I realized that what I was really doing was being an intermediary. My ability to succeed as the head of the computer crime division was really because I was not a technologist, not a detective, not a judge, not the attorney general, not a businessman. But I could speak and understand enough of each of their languages to be the connective tissue and bring them all together and explain one to another. At the time of the Big Dig tunnel collapse, they said we need somebody who can run the scientific side of this investigation and we need someone who understands criminal prosecution. I’m not a medical examiner, but I understand enough, ask questions, and seem to be able to translate and play that intermediary role, which sort of logically led me to end up in this job.
CW: What did the Patrick administration want out of the Roca deal?
GROSSMAN: The enabling legislation said the government could spend up to $50 million on pay-for-success projects. They ultimately decided to do three projects with the maximum exposure being $50 million, and maybe more depending on how the third project fleshed out. We had two other drivers on the Roca project. We needed to reach a critical mass of young people so that the outcomes would be statistically significant. That was one of the government’s non-negotiables. And we needed to figure out how much the state would save if Roca kept people coming out of jail from going back in. You come up with a ballpark figure that you start negotiations with.
CW: What is the state government’s exposure on the Roca project?
GROSSMAN: The max exposure from the government is $28 million.
CW: How do you measure success in terms of keeping people out of jail?
GROSSMAN: It relies on a control group. Roca gets more than 1,300 young people referred to it and it takes 929 into its programs. Then there’s a separate control group made up of young people who are leaving prison and are not referred to Roca. It’s a different population of young people but it looks like the people being served by Roca. You compare the Roca group to the control group. The premise is you’ll be able to say, but for Roca this result wouldn’t have happened. You’re controlling for all of the other circumstances in society, positive or negative, that might affect recidivism.
Roca youth worker Santos Mejia and Roca participant Kevin Sanchez.
CW: How do the government payouts work?
GROSSMAN: The project gets paid in three ways. The first and primary way is based on a reduction in bed days, which is our term for prison days. So we look at how many days the young people in the control group are incarcerated for and we compare that to the number of days that young people who are in Roca’s programming are incarcerated for. The government will pay out based on that comparison for the reduction in bed days. It’s a sliding scale that’s tied to the amount of government savings that occurs as a result of that reduction in bed days. The second way the project gets paid is based on quarters of employment. For every quarter that the young people in the Roca programming are employed, when compared to the counterfactual, there’s a payment the government will make for that. The third way is based on completion of certain tasks among the young people in the Roca population that prepare them to enter the job market.
CW: What do you mean by counterfactual?
GROSSMAN: That’s what happens in the absence of the intervention. In the case of Roca, the control group is the counterfactual.
CW: Does the government pay a fee for every avoided bed day?
GROSSMAN: No, there has to be a baseline level of impact, enough so that the government is comfortable that it’s statistically significant. After that, the government pays about $55 a day for each day and that number goes up until the point where it realizes the full potential of savings the government gets when someone isn’t incarcerated, which is about $145 a day. And that’s if we could close the wing of a prison or forego the building of a new prison. After that, if Roca continues to reduce the number of bed days among the young people assigned to it, the government shares in some of the savings, but not all of the savings. Ultimately, if the government hits the $28 million cap on success fees, the government will keep all of the savings and not pay any more to the project.
CW: How are those bed day fees calculated?
GROSSMAN: It’s a pretty complicated algorithm that was developed by the folks at the Kennedy School.
CW: What is the baseline impact that’s required before the government has to start paying?
GROSSMAN: It’s about 27,000 foregone bed days before the government pays anything.
CW: Do avoided bed days mean actual savings?
GROSSMAN: It’s important to realize it’s the opportunity for savings. I never want to oversell this. I’m reluctant to say this is going to redound to money in the taxpayer’s pocket. I think it’s going to make the taxes we all pay more effective, and stop at that.
CW: So you argue that basically any quantified outcome is a success?
GROSSMAN: I would argue that any quantified outcome is better than what we’re doing now, because now we don’t know what we’re getting. If you know what outcome you’re getting, you can decide if it’s worth it or not from a policy point of view.
CW: Does Roca get paid no matter what?
GROSSMAN: This project is unique in that Roca has what bankers would call skin in the game. Roca, if it costs a dollar to provide its service, gets paid 85 cents. They’ve agreed to put 15 cents at risk. It’s a total of $3 to $3.5 million that is at risk and they’ll only be paid that money if the project is successful. But they’ll also get some additional returns, some upside on that $3.5 million. There are not a lot of providers that are willing or can afford to do that. It sets a very high bar. There aren’t a lot of Rocas in the world.
CW: What quarter is the project in now?
GROSSMAN: The project is currently in quarter seven.
CW: Is Roca on track?
GROSSMAN: It’s really too early to say. I have no insight into what is going on in the counterfactual. What I do know is the management metrics that Roca uses, whether in this project or others, are primarily around attrition. I do know that the attrition numbers are at or below historic performance levels for Roca.
CW:How are people referred to the Roca program?
GROSSMAN: When we started, we were getting referrals from Probation and the Division of Youth Services (DYS). It turned out that there weren’t enough young people coming out of Probation and DYS who were eligible for the program. So we added six additional referral sources, which are Parole, the Department of Correction, and the Houses of Correction in the counties we’re operating in – Middlesex, Essex, Suffolk, Hampden. That took a fair bit of work by everyone to bring in those other referral sources.
CW: Why weren’t there enough referrals from Probation and DYS?
GROSSMAN: We don’t entirely know. The good news is that fewer people are going to jail in the Commonwealth, and that’s a reflection of a national trend. I don’t think we adequately weighted that trend when we did the 10-year look-back on incarceration rates. Also, the juveniles from DYS have to consent to be in the program. So DYS couldn’t give names directly to Roca. The consent has proven more difficult than we expected, so we’re not getting the referrals we expected from DYS.
CW: So the people coming out of prison are given no incentive to work with Roca? Roca has to convince them to participate?
GROSSMAN: The Roca model is that they go find young people and persuade them that they want to participate in Roca. Roca is given their name and date of birth and where they live, that’s all. For people who are coming from the prisons and the houses of correction, Roca is actually introduced to the people behind the walls so that they’re building a relationship with the people behind the walls.
CW: But the inmates are not required to join Roca?
GROSSMAN: It is voluntary. They have some people who are sent to Roca as a condition of incarceration, but the core model is that it’s voluntary.
CW: Why is Goldman Sachs, which is known for earning enormous returns on its deals, involved in this project?
GROSSMAN: There are seven pay-for-success projects in the United States, and Goldman is involved in four of them. They have a group called the Urban Investment Group, which invests both firm capital as well as money raised from clients to do double-bottom-line investment. Our project was the first where they invested client money. The projects they did before always involved firm capital.
CW: What’s a double-bottom-line investment?
GROSSMAN: Double-bottom-line is financial returns and social returns. People sometimes talk about triple-bottom-line, which would also involve environmental benefits.
CW: What type of people invest in these funds?
GROSSMAN: I think it’s a mix. The generation that is now coming into control of significant wealth wants to align their investing with their social goals. I think there’s a market opportunity for that, which is one of the reasons why a lot of banks are interested in impact investing in one way or another. Goldman has found a sweet spot, if you will, with pay-for-success contracts. They have clients who say they want to invest this way. I think there’s also definitely an element of publicity for Goldman. They want to be known as people who brought financial sophistication to bear for social good.
CW: Goldman is putting an $8 million loan into the Roca project. How does the firm get paid?
GROSSMAN: Money is drawn down from Goldman over a four-year period and over that period they are paid 5 percent on the loan on a quarterly basis. So they get 5 percent on their money, but they may not get the principal back. The $8 million is at risk. The first time some of that can be paid back is around quarter 18 — 4.5 years after launch. But it may take as long as the full six years before they get paid back, in whole or in part. And they may never get paid back.
CW: Does Goldman make anything else, like fees?
GROSSMAN: They do not get any fees.
CW: It seems hard to believe that Goldman Sachs is satisfied with a 5 percent return.
GROSSMAN: They get the 5 percent, plus they have the possibility of success payments at the end of the project.
CW: How do the success payments work?
GROSSMAN: After all the costs have been paid, after the senior and junior lenders are paid back, then the balance of the success payments are split using a complex formula between the senior lenders, the junior lenders, Roca, and the philanthropy and a little bit to Third Sector. There’s a negotiated formula that shares out that money.
CW: I understand the federal government invested $12 million in the project. What role does that money play?
GROSSMAN: The $28 million that the state could pay, the first $12 million of that could come from the feds. From a project point of view, that allows us to examine, in a couple of quarters if the project is going well, whether to extend it a couple years. The state still has $28 million. It could be $40 million if you include the federal money.
CW: Do you think the federal money and philanthropic money will be needed in the future as pay-for-success gains traction. In other words, will pay-for-success become more of a business proposition than a philanthropic proposition?
GROSSMAN: The hope is that the philanthropy will go away, but this is early days. This is the third project launched in the United States and the first statewide project.
CW: What does private sector involvement in these projects bring to the table?
GROSSMAN: The idea of this model is that the fiscal discipline that the private sector brings — the questions about outcomes, the due diligence, the rigor — is to the benefit of the public. It forces government, it forces providers, it forces all of us to ask whether we’re getting value for our dollar. What do we value as a society? What do we want to pay for? What is it that we’re paying for? And are we getting that? All questions that I could make a pretty good argument that we should be asking anyway, whether we’re doing a pay-for-success transaction or not. Ultimately, if we get good at this we’ll have less and less philanthropy in there and more and more pure commercial interests. But we may decide that some other project makes sense with philanthropy.
CW: How does it work for the philanthropies?
GROSSMAN: From a philanthropic point of view, if you did $5.5 million of philanthropy it would run 300-odd people through the Roca program. Here you take the $5.5 million and leverage it up so you’re running 929 people through the program. If nothing else ever happens, you’re getting triple the social return just from the people you’re running through Roca. You’re getting three times the benefits.
CW: Do the philanthropies get their money back if the project is a success?
GROSSMAN: They do in a way. They won’t get back money in their pocket. But they do get to redirect it, so at the end of the project if there is $5 million left, that $5 million will be allocated based on the instructions of the philanthropists when they first gave the money. Let’s assume they get half their money back and do another one of these programs. Now they run 460 people through the program. They get half of that money back and they run 240 people through the program. And then 115. All of a sudden, for the initial $5 million, which would have run 300-odd people through the program, you’ve run 2,000 young people through the program. So as a philanthropic proposition, it’s amazing.
CW: Have you ever thought about doing a pay-for-success venture with the MBTA?
Disclosure: Greg Torres, the publisher of CommonWealth and head of MassINC, is currently an honorary board member at Roca and a former board chairman. The Boston Foundation, which provided money for the Roca project, has also contributed money to MassINC. John Grossman is a member of the Massachusetts Criminal Justice Reform Coalition, an initiative that MassINC helped launch.