In a recent Slate article, Jamie Zimmerman and Sascha Meinrath lament the exclusion of the poor from Kenya’s much-touted expansion of mobile money, exemplified by the growth of M-PESA. Citing our early empirical analysis of the adoption of M-PESA, the authors correctly note that in 2008, many of the poor did not use the service. Zimmerman and Meinrath infer from this pattern of adoption that poor households are made worse off by M-PESA’s success.
Moving beyond the hype that has sometimes surrounded M-PESA, and instead focusing objectively on the discernible impacts of this new technology, is welcome. And we share the authors’ skepticism about the ability of this or any other single innovation to meaningfully transform the lives of the poorest segments of the population. However, the logic behind the argument that the success of M-PESA somehow makes nonusers of the technology worse off is difficult to fathom. Moreover, there is little evidence to suggest this is the case.
One reason for the lower take-up among the poor was the transaction cost of sending and withdrawing money using M-PESA, but the examples cited by the authors are misleading. The fee schedule has a relatively large fixed cost component, so the proportional cost of sending a small amount is high—8 percent to 12 percent for transactions in the $1-$5 range. However, these costs fall with the size of the amount sent, averaging about 1 percent to 2 percent over the range of transactions we observe. For sure, there is plenty of room for fee-reductions, but importantly the fee structure is consistently competitive with all other money transfer mechanisms, and has in fact led to reductions in prices of other services.
This said, the initial promise of M-PESA was that it would represent “banking for the unbanked,” although in the early years it appeared this was not the case. But new data suggests that M-PESA has quickly reached the unbanked. Among the population outside Nairobi, during a period of four years when the prevalence of bank accounts remained relatively flat, the share of the unbanked who used M-PESA rose from about 21 percent in 2008 to 75 percent in 2011. While 96 percent in 2011 of those with bank accounts also use M-PESA, diffusion to fully three-quarters of the unbanked population in such a short period can hardly be construed as exclusionary.
Similarly, the share of non-Nairobi households with very low incomes who use M-PESA has also risen over time. According to our data, in 2008 fewer than 20 percent of the population outside the capital living on less than $1.25 per day used M-PESA, but by 2011 this share had steadily expanded to 72 percent.
In our analysis of data from 2008 to 2010, we identified a particular benefit of mobile money: It allowed households to better cope with risk. When bad things happened—illness, crop failure, job loss, and violence—households with access to M-PESA were able to get help faster, from more people, and in larger amounts. More anecdotal evidence also suggests that M-PESA can help households cope with natural disasters. For example, between 2007 and 2009, successive droughts struck the country, but the widespread deprivation and starvation associated with the first (before M-PESA had spread widely) were not observed as acutely during the second. Some commentators have attributed this to the fact that purchasing power was more easily directed to affected areas. Governments did not have to send trucks filled with food; families and friends simply sent money by M-PESA, creating enough demand to induce suppliers to deliver food where needed.
Abrupt reductions in available resources can be the first step of a descent into poverty, in which households can become trapped indefinitely. Conversely, by keeping them away from the precipice, M-PESA could conceivably have long-term impacts on poverty rates among otherwise vulnerable households.
Historically, the adoption of new technologies has been slow and incomplete, especially in Africa, where the low rates of adoption of agricultural technologies remain a puzzle. Mobile phones are unlikely to be a panacea for the complex myriad of development challenges that persist. But in Kenya their use in general, and as a means of engaging in the financial economy, has been transformational. The poor have been slower than the rich to ride this wagon, but they are increasingly on board. The idea that they are being left behind, and actually hurt by M-PESA, is not borne out in the data.
Sascha Meinrath and Jamie Zimmerman respond:
We’re grateful to Professors Jack and Suri for their thoughtful response, their updated data, and the optimistic picture they provide for M-PESA in Kenya. We agree with their assessment of the real and potential benefits of M-PESA (and mobile banking services generally) as a risk management and financial empowerment tool of the poor. However, their rebuttal misses the fundamental point of our article: Our contention is not that M-Pesa or mobile money tools in general are hurting the poor; but that maximizing the potential positive impacts that are so frequently touted (particularly for the poorest of the poor) requires a harder look at the systems that create and control access to these tools.
Unfortunately, the shortcomings of these systems have been too often ignored in all the pro-mobile hype. To be sure, technologies exist that would ensure universal access to mobile technologies at prices vastly more accessible to those living in poverty than what is currently on offer. Indeed, the crux of our argument is not even about M-Pesa or mobile money per se—it is that without a more serious examination of the systemic barriers to universal access to mobile tools, an inevitable and increasing digital divide between the connected (haves) and disconnected (have nots) will dampen the potential for mobile solutions to help eliminate global poverty. As we highlighted in the recent expert panel on this topic, engaging in a debate on these issues is a first critical step, and we thank the respondents for doing so.