At least as good as cash from a helicopter? A new standard for aid effectiveness

by Franck Wiebe, Chief Economist at the Millennium Challenge Corporation In the face of particularly senseless uses of foreign assistance, aid workers sometimes say “it would have been better to drop the money out of a helicopter” to convey how bad programs waste money.

Cash Transfers are less dramatic (and possibly less efficient) than throwing money from a helicopter, but CTs are increasingly accepted as a standard aid mechanism. Their beauty is their simplicity – simply give poor people what they need: more money so they can decide what they need most. Moreover, their likely impact on welfare is easily assessed, because the benefits can be quantified and tracked.

One CT in Zambia incurred administrative costs that cut the value of each aid dollar to beneficiaries to 73 cents. This reduction is large, but not unexpected – a World Bank review of CTs showed that such costs range from 11 to 63 percent. The net value of a cash transfer program to local households can be estimated, accounting for predictable local admin costs. Why not expect donors to demonstrate that their aid projects will be at least as good as a cash transfer?

For many aid projects, standard benefit-cost analysis (BCA) can be used to estimate impact and compare to CTs. The Millennium Challenge Corporation has used BCA for nearly 85 percent of its activities and posts its analyses online. BCA was standard practice at the World Bank, but no longer, and many other donors have never embraced the practice.

A farmer training project, for example, that aims to raise farmer productivity through technical assistance and access to credit poses few challenges to estimating the relevant costs, and the program logic should describe how those expenses will be translated into inputs, outputs, and ultimately higher household incomes. The net discounted value of additional income earned by participants in such a project in Zambia could be more or less than the 73 cents delivered through a CT. Should donors fund such projects without first determining that they are expected to provide more benefits than a CT? The logic of teaching a person to fish rather than giving them a fish only holds if it is more cost-effective to do so.

Distributional analyses enable us to make even better comparisons. The CT that provides 73 cents per aid dollar to local beneficiaries probably also targets the lowest-income households well (but imperfectly). Perhaps 40 cents go to families living under $1.25/day, 25 cents go to households under $2/day, and 8 cents is captured by non-poor households.

Does the farmer training project deliver similar benefits to the poor? The answer depends on the project design, and this information, invaluable for donors and government decision-makers, is usually not too difficult to estimate. A relatively effective farmer training project might generate 85 cents for each dollar invested, but if it targets richer farmers to enhance efficiency, as many do, those under $2/day might get less than the 65 cents expected from the CT. Of course, neither program is a growth-enhancing investment (both are transfers that cost more than they deliver), but why would donors fund “capacity-building” without first demonstrating that the project will outperform a CT?

Ultimately, enhancing aid effectiveness means putting more resources behind those interventions that put more money in the hands of the poor, and CTs should be considered an option that might outperform many of our traditional (and some traditionally ineffective) interventions. Of course, other projects might outperform CTs. As a matter of practice, MCC invests in public goods and services and rarely funds those that are not expected to generate at least a dollar of additional income for each dollar of aid, a higher standard than using CTs as a default. But, in the interest of aid effectiveness and as a starting point, donors could agree not to fund projects unless they can be demonstrated to be at least as good as a cash transfer. Is that too much to ask of aid?