1. Effective Altruism: It's the right time of year to be talking about charitable giving--most US-based charities take in about 50 percent of their annual revenue during the month of December. Here is GiveWell's list of recommended charities this year (NB: I'm on the board of GiveWell). Jennifer Rubinstein has a new essay about the "hidden curriculum" of effective altruism, as seen in Peter Singer's and Will MacAskill's books. There's always a hidden curriculum isn't there?
2. Evidence-Based Policy: Effective Altruism shares a curriculum, hidden or not, with evidence-based policy. At Stanford Social Innovation Review, Jennifer Brooks of the Gates Foundation has a post making the case for evidence-based decision-making. I suspect that prior to November 8th most readers of this newsletter wouldn't have thought the case needed to be made. One of Brooks' key points is the need for better data from rigorous evaluations so that there is evidence not just on effectiveness of a particular program, but information on how to improve other programs' performance. That just so happens to be one of the points in the conclusion to my shortly to be available book on the use of RCTs in development economics. You're running out of time to buy a copy for a holiday gift. It won't arrive until January regardless, but it's the thought that counts right? Oh wait--the whole point of effective altruism and evidence-based policy is that it's not the thought that counts.
3. African Bank Failures: It doesn't make the global news, but there have been a number of bank failures in sub-Saharan Africa in the last few months: Kenya, Mozambique, Zambia, and Uganda have all closed banks since the beginning of October. At FSDAfrica, Mark Napier looks at whether there's a trend to be concerned about. He forecasts a "rocky ride" for African banks, and lots of work for bank regulators, in 2017.
Read More1. Poverty Traps: Every year the Development Impact blog selects a few interesting job market papers and invites the authors to blog the papers (you have to believe that there is some randomization going on in the background and at some point David McKenzie et al. are going to publish something about the causal impact on citations and job offers). The most interesting to me so far this year is a paper by Arun Advani attempting to explain why, given that there's lots of inter-household lending in poor communities, and at least some opportunities for productive investment, so little informal lending seems to flow into productive investments and households stay poor.
Using theory and data from one of the Targeting the Ultra-Poor studies, Advani shows how lenders can be reluctant to help their peers make profitable investments because success will weaken the bonds that keep them in mutual support relationships. It's a useful lens to think about the limitations of informal finance and where the relative advantage of formal financial services may lie.
2. Pro-Poor Digital Finance: Last week, I posited this topic as a question. This week, in strong contrast to the piece I linked about Safaircom preying on poor women, a new paper from Tavneet Suri and Billy Jack argues that access to mPesa moved 194,000 households in Kenya above the $1.25 poverty line. They write, "Thus, although mobile phone use correlates well with economic development, mobile money causes it," which seems to me to be a remarkably strong causal claim. Meanwhile, the UNCDF has published the first in a series of toolkits for financial services providers hoping to develop pro-poor digital finance. And the Aspen Institute's Financial Services Program has launched the Non-Profit Leaders in Financial Technology (nLIFT) group to link groups working on pro-poor digital finance in the United States.
3. Agricultural Finance: Agricultural finance is hard and it always has been (see David Graeber's Debt for an intro to agricultural finance debt crises in ancient Mesopotamia). So it's not surprising how little use of formal or informal agricultural credit there is in sub-Saharan Africa despite the spread of microfinance and increasing use of modern inputs. This new paper finds that the only form of "credit" in wide use is output-labor arrangements, which fits nicely with the poverty trap model in Bangladesh noted above. Agricultural finance isn't all about credit--insurance is a big issue too. Here's a new paper looking at the "Samaritan's Dilemma" (moral hazard arising from the expectation of a bail-out by private charity or public aid) in agricultural insurance markets in the US which finds the dilemma exists and leads to farmers underinvesting in insurance and inputs. Like I said, agricultural finance is hard.
Read More1. The Case For Social Investment in Microcredit: Four years ago, at the suggestion of Alex Counts, I started working on a review of operationally relevant academic research specifically for practitioners. I finally finished it this month [sad trombone]. One of the reasons for the long delay was that the world kept shifting. Over the last 18 months it became clear that the need was not just to document the opportunities for innovation in microfinance but to specifically address whether additional social investment in microcredit was justified given the published impact evaluations.
So I ended up making the case for social investment in microcredit. I believe the case for additional social investment is strong—not despite, but because of, what we’ve learned from impact evaluations. Obviously there’s much more in the paper, but here’s the one sentence summary (there’s a one-page summary in the paper): Microcredit is a cheap intervention with modest but generally positive effects with a great deal of scope for evidence-based innovation that could materially improve impact. The kicker, though, is that the innovation required to boost microcredit’s impact is unlikely to happen without targeted social investment.
Please take a look and argue with me, publicly or privately, about it.
2. Digital Finance and Household Behavior: I lied, I admit it. A few weeks ago the faiV featured "the most interesting" papers from NEUDC. But the most interesting paper wasn't ready for circulation so I couldn't include it. It is now. Tomoko Harigaya studied what happened when savings groups in the Philippines were transitioned to digital finance tools--in other words, group leaders stopped taking cash deposits, instead directing members to make deposits themselves using mobile money. Members could now also make withdrawals without traveling to a bank branch. The result was a significant drop in savings deposits and savings balances and an increased reliance on informal loans. In other words, "convenience" went up and usage went down. The effects seem to be driven by those closest to bank branches ex ante, by the loss of positive peer effects and by increased salience of fees for transactions. Now there are some obvious ways to potentially counteract these effects but it is an important cautionary insight into how little we know about how digital stores of value and transactions affect household financial behaviors--and an especially important finding for the bank itself which would have seen it's funding costs rise from a program designed to reduce operational costs since it relied on deposits as a cheap source of capital.
Read More1. Demonetization in India: It doesn't seem like I'm the only one who's a bit confused by exactly what's happening in India and why this particular set of steps will yield the stated outcomes. Here's my current understanding: Last week, the government declared that 500 and 1000 Rupee notes would no longer be legal tender, effective immediately. Except that those notes could be exchanged for new notes until December 31 at banks and post offices. But only by people with official government ID. The purpose is to drive more of the economy into the formal sector and to clamp down on black market activity and corruption. Usually advocates of this sort of step talk about high denomination bills (which they say facilitates corruption by making it relatively easy--in terms of size and weight--to transport large sums) like $100 bills. But 1000 Rupees is roughly $15 and a new 500 Rupee note will be in use and other large denominations like 2000 Rupees will also continue to exist.
As you can imagine, when 86% of the currency in circulation by value has to be immediately exchanged, there are some problems. Of particular interest to faiV readers might be the effect on microfinance banks, which are not allowed (as of now) to accept or exchange the old notes. That apparently has caused repayment to plummet since people can't get their hands on legal notes to make their payments. There's also a surge in use of ATMs and people signing up digital finance systems. Of course, then there's the problem that roughly 30 percent of the population (a mere 300 million people) doesn't have official ID (not counting the additional millions who are short-term migrants and don't have their ID with them where they currently are). Lot's more to come on this story I'm sure.
2. Digital Payments and State Capacity: Dan Radcliffe of the Gates Foundation has a new paper (published by CGD) on the knock-on benefits of government-to-citizen digital payments infrastructure. Direct transfers have already shown significant benefits in terms of efficiency and effectiveness of social welfare programs. Radcliffe argues that other benefits also deserve attention, specifically "strengthening energy policy, food security, government transparency" and overall state capacity.
Read More1. Mentors for Microenterprises in Kenya: Brooks, Donovan and Johnson assign high profit microentrepreneurs to mentor newer entrants. That's a particularly interesting way to potentially change the trajectories of microfirms. The mentored firms see a significant jump in profits driven by learning how to cut costs but don't maintain the gains once mentorship stops.
2. Grants and Plans for Senegalese Farmers: Ambler, de Brauw and Godlonton give $200 grants and develop a farm management plan for smallholders. The grants boost production (by more than $200), but the gains seem to fade out, though higher stock of assets remains. Farm management plans don't have a measurable impact. I find this interesting for many of the same reasons as #1: figuring out how to boost profits of small enterprises is near the top of my list of urgent program/policy questions.
3. Seasonal Migration in India: Imbert and Papp use NREGA and choices about short-term migration to better understand why the large gap in earnings between rural and urban migration doesn't lead to more seasonal migration. They estimate that more than half of the income gap is consumed with higher living costs in urban areas, with the rest due to non-economic costs--like being away from home and "hard-living", (e.g. sleeping on the street). There are some interesting policy applications for the design of rural public works/income programs and the development of migration finance and support programs.
Read More1. The Future of Microfinance: The big news this week is the merger of two giants in the field--Grameen Foundation and Freedom from Hunger. The merger follows the relatively recent retirement of two giants of the field--the former leaders of Grameen and FFH, Alex Counts and Chris Dunford, respectively. As the announcement states there are clear ways that combining the two organizations may improve their impact, but it's also clear that consolidation is a result of the maturing of the microfinance industry, and I mean that in the Silicon Valley sense, not in the childhood development sense. Where is the industry headed? Reading between the lines of the announcement makes it clear that the combined organization sees the future as one where microfinance is just a utility, or perhaps one of a set of tools, not the center piece of anti-poverty interventions. That's probably right.
One reason why: it's easy for microcredit to go off the rails when it is the centerpiece. Here's an update on overindebtedness in Cambodia.
2. Household Finance: How should households be managing their finances? I don't mean the basics, like choosing a lower-cost over a higher-cost loan, but managing their finances over their lifetime to achieve their goals. Almost all of the the ideas and advice we have for households in developed countries is built on the life cycle model--a household's earning power starts out low but steadily grows, peaks in late middle-age and then declines. Households then should use financial services and tools to shift their income and smooth their lifetime consumption. This piece from the WSJ about what mistakes households in the US make at each phase of their life is particularly explicit about using the life cycle model and it's implications as a standard for whether households are managing their finances correctly or not. The problem with that view, of course, is that it's increasingly clear from financial diaries and other work that the life cycle model isn't an accurate representation of the many households' situation. And we don't have good advice for households' where volatility is the norm, not a blip, and where slow and steady doesn't win the race, or even work at all.
Read More1. News from Rwanda: An evaluation of the use of small-scale household solar panels in Rwanda finds that there are benefits but those are small and diffuse enough that subsidies will be needed to scale adoption. At the conference itself I learned that while 89% of Rwandans are "financially included" only 6% are "adequately served" according to recent FinScope data--a healthy reminder that heavy caveats are required when setting inclusion goals. The next step is to recognize (with a nod to James Scott) that in markets with high "inlcusion," under-served is a strategy not a condition. And while this isn't news about Rwanda, I learned about it in Rwanda: MFO is conducting garment worker financial diaries in southeast Asia which should help us understand a bit more of the difference between Blattman and Dercon's results in Ethiopia and Heath and Mobarak's results in Bangladesh.
2. The Cost of Volatility: One of the common findings from financial diaries work around the world is the prevalence of income volatility, perhaps most surprisingly among US households. In the US Diaries data we see a lot of the volatility coming from variations in amount earned per week in the same job. There are lots of reasons to suspect that volatile schedules and the income volatility that flows from it is bad for households, but how bad? A new field experiment hints that it's really bad. Mas and Pallais randomize wage offers to potential staff for a national call center and find that workers aren't willing to sacrifice pay for a flexible schedule, but are willing to give up 20% of their wage to avoid having a schedule set by the employer with a week's notice.
Read More1. Digital Identity: A few weeks ago we featured a paper on the general equilibrium effects of NREGA in India, which depends on a universal ID system. Next Billion takes a look at India's digital ID system and compares it with Pakistan's program.
2. Insurance (Is Hard All Over): When you read about attempts to launch microinsurance programs for developing countries, it can often seem like insurance markets work very well in developed countries. But insurance is hard no matter where you are, and may be getting harder due to climate risks and our human failings in thinking about large but rare risks. Here's a new brief from the Penn Wharton Public Policy Institute looking at how under-insured many American homeowners are and proposing some steps to get those people to buy insurance.
3. Shocks and External Validity: Typically conversations about the external validity of an impact evaluation focus on whether a finding in one place applies to a finding in another place. Here's a new paper by Rosenzweig and Udry looking at external validity issues in the same place but in different times, specifically at how important aggregate shocks can be when impact is likely to vary over time (as with agriculture or schooling). I'm not sure how big a problem not considering time variance is, but it is a good reminder to examine assumptions when applying findings from impact evaluations.
Read More1. The End of Cash?: Ken Rogoff has a new(ish) book arguing for the end of paper currency. In the New Yorker, Nathan Heller explores Stockholm, one of the most cashless cities on the planet. The move away from cash in Sweden was strongly influenced by high profile robberies of cash depots, making the insecurity and anonymity (for criminals) of cash much more salient. In Heller's piece, there are a few references to issues of privacy, regulation and insecurity of digital tools, but surprisingly little reference to digital payments in less developed countries, or issues in countries where government is less trusted and less trustworthy than Sweden.
2. Cashless in the USA: The US is a very long way away from cashlessness, but one of the primary mechanisms for movement in that direction is prepaid cards. In the last decade they have become increasingly popular alternatives to bank accounts, and as a mechanism for delivering government benefits like food stamps and unemployment and pay to workers without accounts. The Consumer Financial Protection Bureau has released new regulations for prepaid cards to increase consumer protections and bring prepaid cards more in line with credit cards. The regulations include limited liability for lost or stolen cards and new requirements that cards that allow overdrafts have to evaluate customers' ability to repay.
Read More1. Jobs! Jobs! Jobs!: For quite a few years now, my mental model has been that most poor households are "frustrated employees, not frustrated entrepreneurs." In other words, most people aren't held back from their entrepreneurial dreams by lack of access to credit, but they are held back in their dreams of having a job by the lack of jobs. That view is tied heavily to the fact that most microenterprises don't grow at least in part because the owners don't appear to be trying to grow them. This week Chris Blattman and Stefan Dercon released a new working paper about an experiment in Ethiopia where they were able to compare factory jobs to grants for self-employment. They find, among many other details, that those who randomly receive factory employment leave the jobs quickly and those who receive grants for self-employment tended to stay in self-employment and out of the industrial sector. There is a lot going on in this paper so it requires careful reading and some thinking, but it will definitely alter at least my confidence level in my priors.
But the discussion of the new Blattman and Dercon paper revived my memory (hat tips to Rachel Glennerster and Asif Dowla) of this Heath and Mobarak paper on the positive impact of factory work in Bangladesh so there's multiple updating going on for me this week.
2. But Wait, There's More Jobs! Jobs! Jobs!: Karthik Muralidharan and Paul Niehaus have a new paper based off of one of the world's largest RCTs, the roll-out of the new and improved NREGA guaranteed work scheme in India. They find that the program raised incomes of poor households dramatically, but that most of the gains comes from pushing up private sector wage rates, not from income from the program itself. Jonathan Morduch notes that the jump in wages was a factor in the ultra-poor program he studied in Andhra Pradesh not having much impact (many participants left the program to take jobs).
The Muralidharan and Niehaus paper also brings to mind this earlier paper from Breza and Kinnan looking at something similar--how the availability or unavailability of microcredit in India to fund self-employment had generalized effects by altering wage rates. That paper is one of the reasons I believe in the "frustrated employees, not frustrated entrepreneurs" thesis, so now my brain hurts.
Read More
