As the drive for accountability in government spending increases, many are looking for ways to keep from paying the full price for programs that don't work.

In Washington, D.C., that desire has led to the first-ever environmental impact bond, issued this fall by DC Water, the city's water and sewer authority. The $25 million bond will pay for new, green infrastructure like rain gardens and permeable pavement to reduce stormwater runoff.

But if the projects don't work as expected, that's where the new financing structure comes in. Under the terms of the bond, which DC Water sold directly to Goldman Sachs Urban Investment Group and the nonprofit Calvert Foundation, the utility stands to get a multimillion discount on its total borrowing costs if the project doesn't meet a certain threshold.

It's essentially an insurance policy on the project's effectiveness. Here's how it works: After five years, the new infrastructure will be evaluated. If stormwater runoff isn't reduced by at least 18.6 percent, investors will owe DC Water a $3.3 million "risk share" payment. The payment represents a near-full refund of the 3.43 percent interest rate payments DC Water made during the first five years of the bond. After that, the bonds would likely be refinanced into 25-year bonds. DC Water would also drop green infrastructure projects and go back to so-called gray ones (like pumps and water tunnels) to reduce runoff.

So what's the incentive for Goldman Sachs and the Calvert Foundation to buy these bonds? If the reduction of stormwater runoff exceeds expectations -- if runoff is reduced by more than 41.3 percent -- the investors get a bonus payment of $3 million from DC Water after five years. The bonds would then still refinance into 25-year bonds.

Although the deal took two years to iron out, DC Water's CFO Mark Kim said it's a structure that could easily be copied by other utilities because it is still, at its core, a basic market transaction. This makes environmental impact bonds different from so-called social impact bonds or pay for success projects, which are not bonds at all but are negotiated contracts between a private financier and a government. These "bonds" finance certain projects that aim for an agreed-upon outcome, such as reducing recidivism among a certain prison population. The financier gets paid back only if the project outcomes are met after a certain period of time.

For those reasons, pay for success projects are very difficult to replicate. "We structured this as a debt instrument rather than a [pay for success] service contract, so it is very scalable, very transparent and very accessible," said Kim. "Utilities know how to issue debt. We've just structured the deal so that they can look and replicate."


While the environmental impact bond is getting interest from other governments, and was even held up by the White House as a model, it has its critics. Dan Kaplan, who manages a $4 billion debt portfolio for the King County, Wash., Wastewater Treatment Division, said he isn't convinced the environmental impact bond is a better deal because of the "exceptionally high interest rate" DC Water is paying the first five years of the deal. Typically, the shorter the terms of the bond, the lower the interest rate. Under a regular five-year bond, Kaplan said, DC Water would likely pay less than 2 percent instead of 3.43 percent.

Also, given that rain gardens and permeable surfaces aren't new, untested technology, Kaplan doesn't see the point in DC Water hedging its bets that the projects won't do their jobs. "If there's some new technology that needs to be tested and there simply aren't the resources within the utility to commit the personnel and technology to do it," he said, "then perhaps [this financing mechanism] could be a tool."

But Kim said comparing the bond's terms with a five-year bond's terms isn't an apples-to-apples comparison. Although the deal does refinance after five years, it is structured as a 30-year deal and therefore is assigned an interest rate comparable to the utility's typical long-term borrowing cost. In addition, Kim said, a typical five-year bond doesn't "provide a risk transfer or downside protection if green infrastructure does not work, which is the whole point of the deal."

Beth Bafford, investments director for the Calvert Foundation, said she hopes the DC Water deal spurs a new field of social investing that essentially splits the difference between a pay for success project and a traditional bond. Investing in the former means returns might not be realized. Investing in the latter is far less risky -- and less exciting.

"We've looked at a few pay for success deals," says Bafford. They are such uncertain , complex systems that it's "hard to determine what's causing the outcome. In the environmental space, you can measure it, look at it, it's more of a science. The hope is it'll help investors who are more risk averse get into the social contracting space."