The COVID-19 pandemic has upended life for populations across much of the world and prompted many country governments to bolster their social safety nets in response—often by providing direct cash transfers to individuals and households. Cash can help replace lost income when work is disrupted by public health measures, economic contraction, or caregiving needs. Cash payments can also help stimulate local economic activity and can be relatively quickly deployed—particularly when compared to other forms of support. For these reasons, a growing number of donors are looking to cash transfers as a means of advancing global development objectives.
Whether a COVID-induced expansion of cash transfers can set the stage for increased use of cash as a broader development tool remains to be seen.
Whether a COVID-induced expansion of cash transfers can set the stage for increased use of cash as a broader development tool remains to be seen. But with COVID’s fiscal impact likely to squeeze aid budgets into the future, identifying and pursuing approaches that deliver value for money will be as important as ever. A new study supported by the US Agency for International Development (USAID) adds to a growing body of research that shows that cash transfers (while not a silver bullet) can improve a range of development outcomes in a cost-effective way.
Back in 2015, USAID’s Development Innovation Ventures (DIV) launched the first of a series of pioneering studies to examine the cost-effectiveness of programming for development, using a technique known as cash benchmarking (it is the first bilateral donor to undertake such an effort). Cash benchmarking is distinct from cash transfer programming, but the approach relies on a comparison of the per-dollar results of “traditional” aid programs with those of cash transfers. The latest from this groundbreaking research endeavor is now out, and provides further evidence that cash may, in some cases, and for some development objectives, be more effective on a dollar-for-dollar basis than “traditional” development programming.
USAID deserves a lot of credit for its nascent efforts to understand more about the cost-effectiveness of some of its programs. I hope to see these efforts continue and expand. But a key next step for USAID should also be expanding its ability to pursue those approaches that offer the biggest bang for the buck. That may be rethinking the design of some of its “traditional” programs; it should also include expanding what it can do with cash. Whether due to perceived risk or limited experience with cash programming modalities, the US government has been slow to embrace the use of cash as a tool for furthering development (outside of humanitarian response settings). And, indeed, cash transfers won’t always deliver the most value for money, and they’re not the right tool for every development objective USAID pursues. But evidence is mounting that cash should, now more than ever, be on the menu for development.
Evidence is mounting that cash should, now more than ever, be on the menu for development.
Tell me more about cash benchmarking. What is it, and what has it shown?
Cash benchmarking is one way to evaluate cost-effectiveness. It’s an analytical exercise that compares the per-dollar impact of a “traditional” development program (one that delivers defined goods or services) with that of a program that gives beneficiaries cash (usually of a roughly equivalent per person value to that of the in-kind program).
Why compare with cash? Well, cash transfers have been shown to yield a number of positive individual or household level outcomes (e.g., higher school enrollment, improved food security and nutritional status, reduced poverty, increased earnings, reduced HIV infections) and, with their minimal administration costs, they’re just about the lowest cost way to help someone. It’s that low price point that makes cash such a useful benchmark; it shows what kind of results the lowest cost intervention can achieve. If a traditional aid program, with its costly procurement of goods and services and more intensive management needs, doesn’t achieve the same level of results, one can conclude that there are other ways to pursue those same objectives more cost-effectively. The goal of cash benchmarking isn’t to promote cash programming; instead, it seeks to set a value-for-money standard that ”traditional” programming should meet. In the most fundamental sense, cash benchmarking asks whether a designed program can do more for the poor per dollar spent than the poor could do for themselves. Sometimes the answer to that question will be yes, but sometimes it will be no. Knowing which is which can help donors—including USAID—make more efficient programming decisions and help steer them away from programs that end up detracting value of the aid funds appropriated by Congress.
The results of the first USAID-backed cash benchmarking study, published in 2018, were mixed. The study compared the results of a nutrition program in Rwanda with those of a cash transfer program finding that neither the program, nor a cash transfer of equivalent per person value, moved the needle on the key nutrition and child health outcomes of interest, though the program did boost savings and the cash transfer raised several household economic indicators. Only a more substantial cash transfer succeeded in improving the targeted health and nutrition outcomes, along with more substantial economic benefits. But that wasn’t the end of it (after all, one study only takes you so far). There were more studies to come.
The findings from a second study—released in early September—gives us new insights into the second of USAID’s cash vs. program comparisons. This one compares the per dollar results of a job training program in Rwanda with a cash transfer program. The training program appears to have been successful. Participants worked more hours, had more savings and productive assets, improved their business knowledge, and reported higher well-being compared to the control group. But it turns out that cash transfers were more effective. Those who got cash had more favorable outcomes almost across the board—for the same cost per beneficiary.
|Table: Comparing select primary outcomes – program vs. cash*
||Program vs. control
||Cost-equivalent cash transfer vs. control** (Lower estimate)
||Cost-equivalent cash transfer vs. control** (Middle estimate)
||Cost-equivalent cash transfer vs. control** (Upper estimate)
||Large cash transfer vs. control
|Productive asset value
|Per capita household consumption
*Statistically insignificant effects are represented as 0% change here. The others are significant at either p=5% or p=1%. See table 4 in the paper for more details.
**Due to the initial uncertainty of the actual per capita cost of the training program, three estimates of cost equivalency were developed. The lower estimate is the one that ended up closest to the actual program cost.
The data reported are a midline assessment, reflecting progress 18 months into a three-year program—six months after the first cohort wrapped up their training. The researchers will follow up again at the end of three years to measure longer-term outcomes. And, of course, one can’t draw firm conclusions about the relative impact and cost effectiveness of all job training programs vs. all forms of cash transfers from a single study in a single place. But it is an important contribution to an underpopulated evidence base.
Kudos to USAID for taking this on. What should the agency do next?
Continue and expand cash benchmarking research
US aid budgets have been under pressure for the better part of the last decade. And with the growing fiscal effects of the COVID-19 pandemic, that pressure is likely to increase. US foreign aid investments remain a critical (albeit extremely modest) component of the US budget—and a vital part of the US response to the pandemic (after all, in a global pandemic our fates are tied to those of the rest of the world). But in the months and years ahead, international affairs spending likely will need to compete with many other heightened demands on public resources. Ensuring value for money will be all the more critical for USAID moving forward—and the agency can build on its cash benchmarking foundation to double down on understanding the cost-effectiveness of its various interventions and seeking to maximize its per dollar results. Rigorous cost-effectiveness analysis hasn’t yet been widespread at USAID, but the agency’s cash benchmarking research is a good start that deserves continued support.
Do more costed impact evaluations of traditional programs
For cash benchmarking to take hold, USAID will need a better understanding of the impact-per-dollar of both cash transfers and its traditional programs. The agency’s first cash benchmarking studies were conducted in an innovative and rigorous head-to-head way. But research doesn’t have to be structured that way to be instructive. Impact evaluations of discrete cash or traditional programs can also provide useful comparative information, as long as they include cost analysis. To date, these have not been common at USAID, even as the agency has ramped up its evaluation capacity and outputs over the last decade. Establishing a cash benchmarking learning agenda that sets out a common costing methodology and helps the agency prioritize questions for impact evaluation could help build this side of the evidence base.
Synthesize the body of cost-effectiveness findings for greater usability
Like all evaluations, USAID’s cash benchmarking studies—and other relevant impact evaluations with cost analysis—are only useful to the extent they are used to inform program and policy decisions. To get there, the findings need to be made accessible: summarized, synthesized with those of other studies, and easy to access. USAID should take the growing body of evidence on the cost-effectiveness of various “traditional” aid programs—from its own evaluations as well as those undertaken by others, such as the World Bank and Britain’s former Department for International Development—and synthesize the findings in a way that provides rules of thumb about how the range of per dollar results for different interventions compares to summarized research on the cost-effectiveness of cash for similar targeted outcomes. A product of this type will need to find the right balance between simplicity for ease of use and fidelity to the nuance and assumptions that shape how findings can be interpreted. But, especially in combination with staff who can serve as “evidence brokers” to help interpret results, it could be useful for staff working on program design as well as for leadership charged with reviewing proposed high-dollar-value awards to steer the agency away from less cost-effective approaches.
Expand cash programming
The next step in pursuit of cost-effectiveness is to enable USAID to expand its use of cash transfer programming. Cash won’t always be the most cost-effective solution. But where a preponderance of evidence suggests cash transfers can achieve more results for the money than traditional programming choices for targeted outcomes, USAID staff should be able to choose cash-based approaches. While USAID has often employed cash transfers as part of its humanitarian response (the Office of Foreign Disaster Assistance funds over 80 cash or voucher-based programs a year), they have yet to become a regular tool in the agency’s development toolkit.
Reluctance to pursue more cash transfer programming likely stems from a combination of factors, including uncertainty about how they fit into USAID’s sector-specific spending directives, worries that recipients would spend funds on “vice goods” (e.g., alcohol) or spend less time working (or fear that overseers might harbor these concerns), and an element of path dependency, with staff inexperienced in cash for development programming. However, these concerns are likely overblown. Cash transfers have been shown to move the needle on various sector-specific outcomes, so—armed with this evidence—USAID could make a compelling case for how they align with sectoral spending requirements. As for fear of funding vice or laziness, multiple studies dismantle these concerns as unfounded.
Cash transfers certainly won’t be the right programming choice every time. Cash is only relevant for moving individual and household level outcomes—not for addressing critical areas like supporting good governance or financing global public goods. And sometimes, cash won’t generate higher results per dollar than some traditional programs. Furthermore, as USAID staff have noted, the agency sometimes has to balance cost-effectiveness alongside other considerations like US national security priorities or local priorities. And cash certainly isn’t a silver bullet—effects may not last and there can sometimes be negative effects on those who don’t get cash grants. But as budget pressures rise, USAID should be thinking hard about getting the most value from congressionally appropriated funds. If the evidence suggests cash transfers would be the approach most likely to deliver, that kind of programming should be an available option.
Create an agency strategy about when to program cash
Putting cash on USAID’s menu for development will require creating an agency strategy about when and under what conditions staff should consider this approach.
In practice, putting cash on USAID’s menu for development will require creating an agency strategy about when and under what conditions staff should consider this approach. A cash transfer strategy would provide essential decision-making guidance to staff and put a public stake in the ground that the agency values cost-effectiveness and sees cash transfers as a legitimate program approach to pursue it.
The strategy should provide broad guidance on when cash should be considered as a program option and principles—based on those that inform humanitarian cash-based programs—for how to employ it responsibly. The evidence synthesis described above, along with things like an assessment of local market conditions (are there things for cash to buy?) and the quality of the digital payment infrastructure would feed into the guidance on when and where to consider use cash. The strategy should discuss different types of cash transfer programs—conditional, unconditional, long-term, short-term—and the purposes each best serve (the kind of short-term, unconditional cash transfers used as the comparison arm in cash benchmarking studies are certainly not the only choice, nor always the right fit, for a cash transfer program). The strategy should also encourage the exploration of linkages with efforts to improve financial inclusion and outline expectations for evaluation of cash programs, both to add to the body of evidence and to demonstrate accountability to potential skeptics of the approach.
Support countries’ own cash-based social safety net programs
While the cash benchmarking studies focus on USAID-funded transfers, donor support for cash-based interventions doesn’t have to center around setting up and funding the transfers directly. In fact, for cash transfer programs to be sustained and scaled, they need to be owned, operated, and at least partly funded by local governments. USAID can help advance these efforts by supporting government-run pilots or helping build the enabling environment for cash transfers by supporting, for instance, expanded mobile money, improved digital payment infrastructure, and increased access to formal financial services.