International giving feels risky. It need not be.

Americans gave $15.1 billion to help poor people outside the US in 2014. That's a lot of money, but it's just 4 percent of the $358 billion that Americans gave away last year, according to Giving USA. This is unfortunate because the world's poorest countries are where the needs are greatest and where our dollars go further. One reason why Americans don't give more money overseas is that people in Africa and Asia are out of sight and out of mind. A second is uncertainty about whether money sent so far away is being spent effectively; skepticism about global aid programs is widespread.The Zurich-based UBS Optimus foundation has embarked upon an experiment to take some of the risk out of international giving. The foundation serves clients of UBS, the Swiss-based bank that is the world's largest wealth manager, with nearly $2 trillion under management; it says it is "dedicated to funding high-impact programs." To live up to that promise, the foundation has created and invested in a Development Impact Bond, a financing mechanism to raise money to do good in a poor country--but only if evidence demonstrates that agreed-upon outcomes are achieved. In this case, UBS Optimus is supporting efforts by an NGO called Educate Girls to keep girls in school in Rajasthan, India's largest state and one of its poorest. For now, about 40 40 percent of girls drop out before reaching the 5th grade.The structure of the Development Impact Bond is complicated, but the idea driving it is simple and potentially powerful: The work of Educate Girls will be evaluated by an independent, third-party nonprofit, called IDinsight, using a randomized control trial, the gold standard of evidence. Only if girls who get help from Educate Girls outperform girls who don't will money from UBS Optimus and its partners continue to flow to the effort."You have to deliver impact or there will be no return," says Phyllis Costanza, the foundation's CEO.This is not how most international development projects operate. Many do no rigorous evaluation at all.

“We don’t have solid evidence for most programs, about whether they work or not,” says Neil Buddy Shah, the CEO and founding partner of IDinsight, by phone from Rajasthan, where he's working on the Educate Girls evaluation.

Here's how the Development Impact Bond, or DIB, works. The UBS Optimus foundation will provide almost $300,000 to Educate Girls, which has run programs to get girls into school and keep them there since 2005. (It is one of 2.4 million NGOs in India, or one for every 400 people. Crazy.) Educate Girls will use the money to survey villages, identify girls who are not in school, encourage them to go and give them academic support, typically in the firm of three to five visits per week from volunteers, with the goal of helping about 18,000 girls. Meantime, IDinsight will track the progress of about 4.500 girls who are getting the help and another 4,500 who are not. If the program meets agreed-upon outcomes, which are tied to academic performance, the Children's Investment Fund Foundation (CIFF), one of the largest charities in the UK, will "buy" those outcomes and repay the investment from UBS Optimus, along with a 10 percent return. That money will then flow back into philanthropy. If the program fails, for whatever reason, CIFF won't have to pay. UBS Optimus assumes the risk.

Why go to all that trouble? The long-term goal, Costanza told me, is to steer more philanthropic dollars to the most effective development programs through mechanisms like DIBs.

"We want to see if there is a market for these types of vehicle—and to emphasize the point, about how important outcomes are to programs. There is so much money going to programs that aren’t working," she said. “I think that we can bring more money to the sector if we can create more vehicles like this. Particularly for people who don’t feel comfortable giving internationally. We’re trying to de-risk international giving.”

DIBs are designed to deliver both financial and social returns. The Global Alliance for Vaccines and Immunization (GAVI) has raised more than $5 billion from investors, who will be repaid by governments and foundations if targets are met. DIBs are similar in structure to Social Impact Bonds, which have succeeded at delivering educational outcomes in the UK. Goldman Sachs has invested in four Social Impact Bonds in the US, including one, designed to keep adolescent offenders out of prison, that failed to meet its objective and was shut down. But the social impact bond "proved to be a highly useful tool," as Andrea Phillips of Goldman and James Anderson of the Bloomberg Philanthropies, recently wrote.

Measuring impact is, of course, essential to these funding mechanisms. But separating signal from noise is a big challenge. “It’s really hard, normally, to disentangle whether the intervention you are studying is a responsible for the outcome you are measuring,” says Neil Buddy Shah, who, with his colleagues at IDinsight, has evaluated education, health, sanitation and finance projects in Asia and Africa.

Shah, who has an MD from the Albert Einstein College of Medicine as well as a BA and MPA from Harvard, wants to bring some of the rigorous tools of medicine to the development arena. Randomized clinical trials, which have been used to test medical interventions since the late 1940s, weren't applied to development projects until the 1990s; since they have been used only sparingly. Shah worked briefly as a researcher with the Abdul Latif Jameel Poverty Action Lab (J-PAL), an MIT-based network of researchers that along with New Haven-based Innovations for Poverty Action has pioneered the use of randomized evaluations to better understand how to fight poverty.

What's exciting about all this is the underlying commitment to measure, learn and improve--and ultimately to bring more money to groups that are helping the world's poorest people. UBS's wealthy clients, Costanza told me, can be persuaded to give more to poor children through the UBS Optimus foundation if they can be assured that their donations will be put to good use. "It could start to change the conversation around impactful philanthropy," she said.