Social Impact Bonds for Dummies

We've been hearing about Social Impact Bonds, or "Pay for Success" programs for a few years now, but only recently have they entered the world of public education. Chicago, for example, launched one $17 million program last year for Pre-K, and last month Utah's United Way was happy to announce that Goldman-Sachs' Pay for Success program in Utah had yielded dividends.

The spread of Social Impact Bonds to the education sector raises all sorts of questions like "How are the fiduciary interests of a private investment firm balanced against social demands of education" or "What overseeing groups can best evaluate programs with a balanced view toward all involved interests."

Or, "What the hell is a social impact bond?"

On the ground, it looks kind of ridiculous, like a program that pays a Wall Street firm a bonus every time a kid is taken off of special ed rolls.

But how does that even work? How does the Wall Street firm get paid? With what money? How do you make money on an investment in something that creates no profit?

An Oversimplified Example

Here's the basic structure of a Social Impact Bond. Note: I am not an economist, banker, or investment counselor, nor do I play one on TV, so I may cut a few corners here.

My house is drafty. My windows leak and my heating bill is $10,000 a year.

My landlord goes to the bank. She says, "Banker, I would like a bond of $4,000 for new storm windows. I think they would reduce my annual heating bill by $3,000."

And the investor issues a bond for the program costs, in return for which he gets a healthy cut of the $3K saved by installing the new windows. My landlord's savings from the successful Stop Freezing My Butt Off Social Program become the bond holder's profit-- but only if our goals are met.

Typically a third party will come in to judge the result, making sure that I didn't just turn the thermostat down or it wasn't just a warm winter or my landlord didn't actually save $6K and hide it from the bondholder. Also, it's worth noting that bonds generally come with negotiated maturity dates, at which point the original loan amount is to be paid back. And remember kids-- bond holders are different from investors. An investor owns part of the company, but a bondholder is just a fancy debtor, and as such has legal priority for being paid back.

In this example, the government is, more or less, my landlord. For a more thorough explanation, we can look here. Here's the shortened version of their explanation:

In the classic... social impact bond, a government agency sets a specific, measurable social outcome they want to see achieved within a well-defined population over a period of time. ...The government then contracts with an external organization—sometimes called an intermediary—that is in charge of achieving that outcome. ... The intermediary hires and manages service providers who perform the interventions intended to achieve the desired outcome. Because the government does not pay until and unless the outcome is achieved, the intermediary raises money from outside investors. These investors will be repaid and receive a return on their investment for taking on the performance risk of the interventions if and only if the outcome is achieved.

Okay, Watch Carefully Now

From New York Times coverage of a SIB program that failed. “Social Impact Bonds offer a strikingly different way to pay for social programs. Governments, rather than tapping taxpayers, can turn to outside investors and philanthropists for funds, and reward them only for programs that work." If the program fails, the taxpayers are off the hook. If it succeeds, the bond holders are paid off with what would have been taxpayer savings of taxpayer dollars.

But the finances get muddier because in the couple of years we've been trying this, we've learned a useful insight:

“The tool of ‘pay for success’ is much better suited to expanding an existing program,” Andrea Phillips, vice president of Goldman’s urban investment group, said in an interview on Wednesday. “That is something we’ve already learned through this.”

But issuing bonds for existing programs means we'll have public and private money swimming in the same pool.

So Is This a Good Thing for Education, or Not?

There's huge cheating potential here, on both sides. The school system could pocket the SIB money and declare, "Damn, but the goal wasn't met. Guess we'll just keep your pile of cash and you get nothing." On the other hand, when a metric is as simple as moving students off the special education rolls, it's really easy to fake the results if you are so inclined. If that happens, all the money that used to be your special ed budget is now funneled straight to Goldman Sachs or whoever is holding the bond, and this whole set-up becomes one more way to turn public tax dollars into private corporate profit.

It all comes down to the third-party evaluator. That's the entity that is supposed to keep the whole game honest and determine whether the goals of the SIB-funded program have been met-- they will determine who gets a payday.

So all that's needed to keep this system honest, fair, and above-board is an entity that has the expertise to judge the program achievement but which has no interests in either side of the transaction. An independent overseer. You know, kind of like the SEC or the firms that were responsible for making sure that big Wall Street firms weren't peddling junk investments ten years ago.

It is, of course, up to government to make sure that such above-board groups are in place. So as Chicago runs forward with its SIB program, I'm sure that Rahm Emanuel will select third-party evaluators who have no ties to or interests in the investors who laid out the money in hopes of big fat tax-funded returns on their money.

You Begin To See the Problem

This is a hugely easy system to game, in part because the whole business is byzantine and twisty that by the time you get to discussion of the proper disinterested expertise of the overseers of the metrics for judging the success of the blah blah blah and now the general public thinks you sound like the grownups in a Peanuts cartoon.

That's not my only problem with this approach.

This is the kind of system that favors easily-measured results, so investments are liable to steered toward the program that is easier to attach to some easily moved metric, and not the program that is most necessary.

What difference does that make, you may ask. The money saved over here can be used to pump up the program over there. Except it can't be, because the money saved over here becomes financial returns for the bond-holders over there.

In the unlikely event that a Social Impact Bond program doesn't just become an exercise in cooking the books, the development of a more efficient or more effective school system, the local taxpayers reap no benefit from that because the cost of the district remains the same-- it's just that now taxpayers are paying off Goldman Sachs with their tax dollars instead of paying for an educational system.

Finally, when if this turns into an exercise in cooking the books, students pay the price. If our metric is getting 110 students off the special education rolls, we had better be damned sure that we don't end up with 100 students who are being denied the services they need so that Goldman Sachs can enjoy a healthy return on their bond. (Update: In fact, I've since found an account of how the above-mentioned Utah program did, indeed, lead to book cookery.)

I can understand the appeal of Social Impact Bonds in some situations. But every time we let the bulls and bears of Wall Street loose in the china shop of education, bad things happen. Maybe there's something magical about all this that I just don't get, in which case at a minimum, these guys are doing a lousy job of explaining themselves. But maybe I should just trust the guys on Wall Street at place like Goldman Sachs. After all, it's not like they've ever tried to screw us all over before, right?