Tag Archives: goldman sachs

The Promise of Social Impact Bonds

Over the past few weeks, social impact bonds have received a lot of attention.  That is because New York City has partnered with Goldman Sachs to run a pilot program aimed at reducing recidivism among inmates at Rikers Island prison.  But first, a little background on social impact bonds.

There are many social problems for which there is no clear-cut solution.  Homelessness, foster care, inmate recidivism, and other issues are often expensive to control.  Programs designed to address them are often part of large bureaucracies and susceptible to the same inefficiencies and perverse incentives endemic in other government agencies.  And, despite best efforts to fix the problems, they will only get worse as social programs move closer to the chopping block.

Non-profits supplement government efforts by addressing specific problems.  Halfway houses, community health clinics, shelters, low-income housing developments, and soup kitchens are all examples of non-governmental organizations serving the homeless.  Many have developed innovative approaches that are effective in achieving specific goals – i.e. placing homeless in low-income housing, job-training, etc. – but are constrained by a lack of capital.  Without access to greater resources, non-profits will never be able achieve scale.

So the government has lots of money, but not enough dynamic programs to fund.  In contrast, non-profits run effective programs on a small scale, but lack the money to expand.  This is where social impact bonds come in.

In a social impact bond, the government contracts an intermediary to put together a social impact bond to address a specific social problem.  The intermediary then identifies non-profits with promising potential and connects them with investors.  The investors provide multi-year funding to the non-profit, allowing them to scale their intervention.  In return, the investor is reimbursed by the government based to the program’s success.  A monitoring-and-evaluation firm is brought in to assess the impact, which is based on a pre-determined set of metrics.  If the intervention achieves the targets, the investor makes a return on the bond.  If not, it takes a loss.

The best way to explain the mechanics of a social impact bond is to provide a real-world example.  In the case of New York City and Goldman Sachs, the city government wants to reduce the number of repeat offenders, which cost taxpayers money in the form of prison costs, increased law enforcement, and lost productivity.  Here is how it works:

The Goldman money will be used to pay MDRC, a social services provider, to design and oversee the program. If the program reduces recidivism by 10 percent, Goldman would be repaid the full $9.6 million; if recidivism drops more, Goldman could make as much as $2.1 million in profit; if recidivism does not drop by at least 10 percent, Goldman would lose as much as $2.4 million.

It seems like a win-win situation, if investors see social impact bonds as a viable means of earning a financial return.  I am mostly in favor of any programs that place greater emphasis on “outcomes over outputs.” But this emphasis is hardly new in the international development community, which has seen a surge in rigorous testing for interventions after economists like Dean Karlan, Esther Duflo, and Abhijit Banerjee popularized the use of randomized controlled trials (RCTs) to determine the efficacy of different approaches.  And I have the same concerns about social impact bonds that I do about RCTs.

Tying financial returns to outcomes creates two potential problems.  First, it risks incentivizing the wrong things, a la “teaching to the test.” Often, these problems are extraordinarily complex and difficult to address, and rarely lend themselves to a timeline that works with an investment.  In microfinance, for example, most RCTs occur over 2-3 years, and have shown little improvement in the well-being of recipients.  I would argue it takes much longer than 2-3 years to realize the fruits of microfinance.  If that is the case, which timeline will be used for the social impact bond – the one that shows progress, or the one that doesn’t?

For some issues, this is not a concern.  Recidivism, for example, is cut-and-dry.  Chronic homelessness, however, is not.  For social impact bonds to be successful, they will require metrics that truly reflect the success of the program.

The second problem is that the interconnectedness of institutions can mask success.  Mark Rosenman of Caring to Change explains both of these problems (h/t Democracy in America):

Where does a nonprofit get the funding to provide the services from which they are to later show a monetized gain to government? How far out in time does the performance metric need to go before quantifiable economic value can be shown and the charity repaid its expenditures? What happens when a nonprofit is providing superb and highly effective services to individuals, but other institutions and variables deteriorate and affect its outcomes?

These are very real concerns that the international development community has been forced to confront (or avoid) in its work.

I am not as pessimistic as Mr. Rosenman or Mr. Steinglass.  I think that social impact bonds are a pragmatic and innovative solution to a very real problem.  In addition to capital, investors will bring human resources and technology to bear on the problem, which will infuse the sector with new ideas and perspectives.  Social impact bonds are still in their nascent stages, but, if they can figure out a way to effectively capture success rates and avoid the pitfalls of “juking the stats,” I see no reason why they can’t be a game-changer in the fight to address social problems.

Develop Economies’ Music Recommendation

Commodity Speculation, Rising Food Prices, and Goldman Sachs

Old habits die hard, and the motor patterns in my fingers that brought me to the Drudge Report so many times when I worked in a cubicle in Boston once again led me to page the other day.  Living up to its reputation for sensationalism, it featured a headline recently about the escalation of food prices around the world.  Unfortunately, while Drudge is usually over-the-top, rising food prices are no laughing matter.  In 2008, the rising cost of our daily bread led to food riots around the globe and massive destabilization in developing countries, most prominently in Haiti.  It alerted food-dependent developed countries to a glaring Achilles’ heel, spurring a land grab in Africa that (almost) comically culminated with the South Korean conglomerate Daewoo making a bid for half – yes, half – of the arable land in Madagascar.  So when the next food crisis hits, and hit it will, the developed countries with a foothold may think they are food-secure, until the hungry populations of the food-insecure countries serving as their respective breadbaskets see the fields of gold beyond the fence and decide to Mugabe it for themselves.  Unless, of course, the landowners (read: nations) deploy armed guards to protect these critical investments, resulting in rioting, bloodshed, and, inshallah, the toppling of governments.

And now, it could be happening again.  I sound like Drudge.

Theoretically, commodity prices fluctuate based on the principles of supply and demand.  When the demand for grain exceeds supply, prices go up.  In the movie Trading Places, Randolph and Mortimer Duke, the lovable racist WASPs, try to corner the market for Florida oranges.  They pay Clarence Beeks for an advance copy of the classified crop report, which will determine the price of oranges for the next trading period.  Akroyd and Murphy intercept the report and forge a new version, giving the impression that there will be a shortage of oranges due to a long winter.  On the trading floor, the Dukes’ trader buys as many orange futures as he can, under the assumption that they will become more valuable once the negative forecast for oranges is released.  The other traders see what is happening, and also buy, driving prices up and up.  Akroyd and Murphy begin selling at 120, until the crop report is released.  When the real crop report is released, which says that this year’s orange yields will be high, the price plummets, and Akroyd and Murphy buy all the futures they sold in the morning, becoming millionaires in the process.  This is how commodity trading works.

In reality, this isn’t always the case. Continue reading