Press "Enter" to skip to content

Friday Links Posts

August 17, 2018 (Links edition #66)


No better way of kicking off than with a new flagship report on safety nets in Africa: Beegle et al just unveiled a great edited volume taking stock of performance, impacts, operations, institutions and financing of social assistance in the continent. Some select juicy findings: over 2010-2015, an average of 14 new safety net programs were introduced in Africa annually; Tanzania had the highest rate of safety nets coverage expansion in the world; coverage of Senegal’s flagship program is higher than in select middle-income East Asian countries; safety nets are not just about consumption — for instance, they increase ownership of livestock (+34%) and durables (+10%); administrative costs of managing safety net is 17% of program budget (but large variance exists); and over half (55%) of safety nets in Africa are externally-financed.

Hard times for pilots around the world. Canada’s Ontario province followed Finland’s footsteps in scrapping its widely-publicized UBI pilot: while Finland dropped it before finishing it, Ontario’s dismissal occurred even before commencement. This generated a flurry of reactions (e.g., here, here, here, here, and here). There are clear ethical issues with the move, with perspective beneficiaries having made decisions based on the expectation of receiving support. At the same time, both Ontario and Finland reveal the risk of calling “UBI” what is not so: for instance, Ontario targeted people with less than $34,000-$48,000 per year, so it wasn’t universal. I suspect that while labeling it UBI may be a marketing expediency to enhance visibility, the backlash from failures will be equally visible, and ultimately damaging, in terms of optics and substance for UBI.

The Indian state of Jharkhand is withdrawing its pilot in Ranchi to digitalize access to the food-based Public Distribution System (PDS). The pilot, covering over 52,000 beneficiaries, transferred food subsidy money into people’s bank accounts and then allowed them to buy ration from fair price shops at the market rate (other ‘DBT’ pilot projects were launched in Haryana, Andhra Pradesh and Karnataka). The state is now returning to its traditional PDS version due to technical and legal glitches. In fact, surveys show that about 97% of beneficiaries were against the pilot (apparently if people failed to use the transfer within the month, they were denied food the next one).

Speaking of the PDS, a brilliant IMF working paper by Coady and Prady simulates the effects of a UBI in India. The paper is generally supportive of UBI as an alternative to food and energy subsidies, although it points out a number of concerns in terms of policy priorities (e.g., vis-à-vis health and education), vested interests that may resist reforms, and the pace of introduction (starting from categorical programs and then scaling up). In particular, the current mildly progressive incidence of PDS implies that “… a sizeable percentage of existing PDS beneficiaries would lose from the [UBI] reform, including many low-income households”. This would include people in the poorest decile. Instead, the replacement of regressive energy subsidies “… would deliver unambiguous distributional gains”.

From simulations to the real world: Iran had a real UBI for a short spell of time in 2010-11. The JDE version of a paper by Salehi-Isfahani and Mostafavi-Dehzooei investigates the impacts of the Iranian UBI on labor supply. Using diff-in-diff and fixed effects methods, the paper finds no support for a negative effect on hours worked or labor force participation. Instead, they show evidence of positive effect for women (hours worked and participation) and self-employed men (hours worked). Since I mentioned UBI: Reich reviews Lowrey and Yang’s basic income books. Reading his lucid analysis, was amused by the following line: “… the two books cover so much of the same terrain that I’m tempted to wonder whether they were written by the same robot, programmed for slightly different levels of giddy enthusiasm”. In her volume, Lowrey’s quickly dismisses jobs guarantees on the basis of implementation complexity, but in a related commentary, Covert underscores that actually “… jobs guarantee is hardly any more utopian than a universal basic income”.

More on the US safety nets. A JLE article by Bastian and Michelmore documents a range of long-term impacts of the earned income tax credit (EITC): an additional $1,000 in EITC exposure when a child is 13–18 years old increases the likelihood of completing high school (1.3%), completing college (4.2%), and being employed as a young adult (1%) and earnings by 2.2%. Turning to food vouchers (SNAP), a CBPP report by Butcher and Whitmore-Schanzenbach shows that SNAP beneficiaries are mostly working families in low-pay and unstable jobs: establishing work requirements in such unfavorable labor market, they argue, would be hard to meet and administer. A similar argument – i.e., poor people work already – is forcefully made in an article by Schram et al. But how does US poverty compare to poverty elsewhere? An IGC piece by Lessing shows that while US destitution is clearly different to that found in developing countries, better social safety nets are important in all contexts.

Turning to OECD countries more widely, Causa and Hermansen offer some interesting insights on why equity is declining in those countries’ social protection systems. They argue that taxes and transfers are less effective at reducing inequality today than they were in the mid-1990s. This drop in effectiveness has been largely driven by the changing nature of cash transfers that favored efficiency over equity objectives: “… more targeted transfers did not generally deliver more redistribution because increased targeting was in most cases combined with decreased size”. Speaking of tax-transfer systems, The Economist had an interesting set of articles around the notion that capitalism needs the welfare state to survive: “… as liberals such as Beveridge realised, the best way to secure support for free markets is to give more people a stake in them. The welfare state must be seen as more than providing shoes and soup for the poor, and security in old age. In a democratic society it is also crucial to the case for capitalism” (h/t Ruslan Yemtsov).

A range of materials on labor markets-oriented matters. A new paper by Autor and Salomons finds that automation displaces employment and reduces labor’s share of value-added in the industries in which it originates (a direct effect). However, their framework cannot explain why labor shares fell more rapidly during the 2000s. A PNAS piece by Simon et al show that, based on survey data from Jamaica, enhanced channels for regular migration would displace irregular migration. Indeed, restrictions on family-based visas result in significant reductions in total migration. However, they also produce the largest reorientation toward unauthorized channels—an unintended consequence that even the highest rates of apprehension may not effectively eliminate. Baragwanath et al have a new working paper that defines Indian urban markets based on economic activity detected by satellite imagery, including based on contiguous pixels that have a minimum threshold of light intensity. Bonus: an interesting podcast by ESID’s King on researching urban poverty in Uganda.

A wave of human capital-related resources. Let’s begin with nutrition: how much income do countries lose because of child malnutrition? A new working paper by Galasso and Wagstaff estimates that, on average, per-capita income penalty from stunting is around 7%. The paper also identifies a package of 10 nutrition interventions, the gradual scale-up of which would have a rate of return of 17% and a benefit-cost ratio of 15:1. In Ethiopia, results from a JDS paper by Gebrehiwot and Castilla indicate that PSNP has not had the desired effect on household dietary diversity or child nutrition, suggesting that transfers may need to be paired with additional accompanying measures such as nutritional information. In India’s Andhra Pradesh and Telangana states, Aurino and Morrow found that, from an early age, children were able to describe and explain the food security challenges of their families and recognize the negative consequences of household food insecurity on their diets, health and education. A WD paper by Roy et al reveals a “hidden” ingredient of human capital: self-efficacy, or an individual’s self-belief that they can accomplish a given task and cope with life’s challenges). Examining self-efficacy among adolescent and young women in Jharkhand, India, they find that self-efficacy is an important correlate of the educational and employment aspirations of these women.

A couple of education-related materials. In a CGD policy note, Pritchett makes the case that too much information can be unhelpful. Using the analogy of the infamous Air France plane crash in 2009, he shows that India’s State Report Cards 2016-2017 included 977 distinct numbers reported, but none of those had any direct measure of student learning. In an NBER paper, Mbiti et al present results from a randomized experiment across a sample of 350 schools in Tanzania. In the study, schools were provided with (a) unconditional school grants, (b) bonus payments to teachers based on student performance, and (c) both of the above. At the end of two years, they found (a) no impact on student test scores from providing school grants, (b) some evidence of positive effects from offering performance-linked bonuses to teachers, and (c) significant positive effects on learning from providing both programs (the effect of joint provision being significantly greater than the sum of the individual effects).

In terms of health, a paper by Corman et al reviews the literature on the effects of prenatal care on birth outcomes; Ahmed et al show that, as result of antimicrobial resistance (AMR), economies will experience an increase in mortality, reduced productivity for labor, and increased health care costs (e.g., expected losses from AMR in 2015–50 may sum to $85 trillion in GDP and 28.3 million more people living on less than $1.90/day). Bonus on mental health: Singhal shows that in Vietnam, early-life exposure to war negatively affects mental health in adulthood.

Several pieces on gender. A JDS article by Tsaneva and Balakrishnan finds that in the first year of the NREGA program in India, women in recipient districts were less likely to experience depression symptoms. They provide suggestive evidence that one mechanism for such effect is greater economic security and independence. The impacts of the program on men’s mental health, on the other hand, are generally not statistically significant. Can conditions in CCTs negatively affect women? A short piece by Haenn explores the unintended intrahousehold effects of conditionalities in Mexico. She argues that Prospera frees up the time and money of its recipients’ husbands, while increasing women’s household responsibilities, hence reinforcing unequal gender dynamics within the family. BTW, in a WD article on Bangladesh and Ghana using the Relative Autonomy Index (RAI), Seymour and Peterman underscore the importance of understanding and measuring men’s and women’s perceptions of decision making within the household.

Let’s look at news on the humanitarian, fragility and crises front. A consolidated fact-sheet highlights the jaw-dropping trends of violence against aid workers in 2017: a total of 313 people were victims, out of which 139 were killed, 102 wounded, and 72 kidnapped. Christian and Barrett Summarize their paper questioning previous studies examining the relationship between food aid and conflicts (methodological note here), with Koren weighing in with an AJAE article claiming that it’s food abundance – not scarcity – fueling conflicts in Africa (h/t Jonathan Morduch). Rubin is out to solve a puzzle that I myself have been wondering about: while famines continue to strike, research remains scarce. Interestingly, the steepest decline in scholarly attention to famine coincided with Amartya Sen being awarded the Nobel Prize in Economics in 1998, in part for his work on famine. Why such decline? His JDS article offers three explanations: (i) the field of contemporary famine research exhibits limited interest in theory-building; (ii) the field is impeded by inaccessibility to key research sites; and (iii) the field is weakened by a small and dispersed research community.

A new ODI report by Twigg et al. identify five key challenges that need to be addressed in order to promote disability inclusion in disaster risk reduction and humanitarian action, relating to data, contextual understanding, institutions, representation and discrimination. It highlights the importance of rights-based approaches, together with improved standards and indicators, in overcoming these challenges. Steets et al have a comprehensive evaluation of the performance of humanitarian actors (WFP and partners) in applying humanitarian principles. The overall assessment is positive, although some concerns are highlighted in reference, for example, to quality of assistance and impartiality. The NGO community has produced several new resources: let me just signal out the NGOs position on cash coordination, the PCI’s implementation toolkit for cash transfers and vouchers, a review of food assistance programs in Guatemala, and a workshop report on lessons from the Hurricane Maria response. Finally, a couple of videos: in one, Dercon discusses the various aspects of aid effectiveness in ‘messy’ humanitarian situations, including the importance of cost-benefit analyzes, feedback loops, prioritizing the engines of inclusive growth, theories of change, and planning for humanitarian aid. In another CGD video, a high-level panel discusses flexible financing in fragile situations.

There has been an explosion of papers on evaluation methods. Dreze has a thoughtful piece defining and positioning evidence within the wider decision making factors: specifically, he makes the point that ‘evidence’ is more than RCTs, that ‘understanding’ is more than ‘evidence’, and that ‘policy’ is more than ‘understanding’; in a similar vein, a CGD working paper by Ravallion questions whether the randomistas should rule; Kowalski discusses how to examine external validity within an experiment; Kahn-Lang and Lang assess the promise and pitfalls of quasi-experimental difference-in-difference methods; Arkhangelsky and Imbens review the role of propensity scores in fixed effect models. A little treat: video and slides Duflo’s 1.5h NBER Summer Institute Master Lecture on machine learning and RCTs. (Bonus: Tomizawa and Masugi have a short Brookings piece on statistical capacity in low-income countries).

Let me conclude this extended, back-to-work edition, with something that might be compelling for our daily work: a set of inspiring lessons from Harford on Arpa, the little-known agency that founded the internet. Among them, the secret of the agency’s success was to constantly create “… a vigorous marketplace for ideas” – a principle that, it seems to me, is more and more pervading the development community.