Working Paper: NBER ID: w29086
Authors: Susanna B. Berkouwer; Pierre E. Biscaye; Eric Hsu; Oliver W. Kim; Kenneth Lee; Edward Miguel; Catherine Wolfram
Abstract: In response to the Covid-19 crisis, 186 countries implemented direct cash transfers to households, and 181 introduced in-kind programs that lowered the cost of utilities such as electricity, water, transport, and mobile money. Do cash or in-kind transfers generate greater welfare improvements? And, does a country’s financial infrastructure affect optimal aid disbursement? Through a parallel set of surveys in two urban regions in Africa—with comparable education, cell phone ownership, and electricity connectivity—we show that optimal government aid disbursement hinges on financial infrastructure. In line with economic theory favoring direct cash transfers, in a randomized experiment in Kenya 95% of urban recipients prefer mobile money over electricity transfers of a similar monetary value. But Kenya is an outlier with high mobile money adoption: this increases its value and reduces transaction costs of buying electricity credit. By contrast, in Ghana—where mobile money is less widespread and the transaction costs for buying electricity are higher—half of recipients prefer electricity transfers, and many are willing to forego significant value to receive electricity instead of mobile money. These results have several important policy implications. First, the optimal government policy in response to an economic crisis is not uniform: cash and in-kind transfers have different advantages that make each suitable for specific contexts. Second, the adoption of modern financial technologies will likely increase the efficiency of government cash transfer programs, even as in-kind transfers continue to be preferred in settings where mobile money uptake is slow. Finally, giving recipients a choice harnesses valuable local information that a policy maker may not have access to.
Keywords: financial infrastructure; cash transfers; in-kind transfers; COVID-19; government aid
JEL Codes: G23; O38; Q38
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| financial infrastructure (G20) | optimal government aid disbursement (H84) |
| adoption rate of mobile money in Kenya (E40) | preference for mobile money over electricity transfers (E41) |
| adoption rate of mobile money in Ghana (L96) | preference for electricity transfers (L94) |
| type of transfer (prepaid electricity) (L97) | electricity consumption (L94) |
| type of transfer (cash) (F16) | electricity consumption (L94) |
| local acceptance and integration of mobile money systems (E42) | effectiveness of cash transfers (F35) |
| type of aid disbursed (cash vs. in-kind transfers) (F35) | welfare outcomes (I38) |