When discussing the post-Covid future, it often boils down to this question: are social protection systems better prepared to deal with a pandemic today than they were in 2020? There is one way to find out: stress test! A new report edited by Johnson and Walker examines the state of adaptive social protection in South Asia. To do so, it elegantly lays out the risk profile of the region and stress-tests systems in Afghanistan, Bangladesh, Bhutan, India, Maldives, Nepal, Pakistan, and Sri Lanka. Results reveal an ample agenda to strengthen robustness and responsiveness, i.e., no country had an average performance exceeding an “emerging” or medium score (see p.48). Now, we need make social protection adaptive, but without overwhelming systems. So among the priorities, I really liked how the report didn’t call for tailoring the system to every possible risk, but rather to establish a nimble framework that sets out the basic rules, triggers and requirements to make a system scalable in times of need (p.228).
Interested in examples on how technology can help social protection (with of course several limitations)? Aiken and Ohlenburg tackle the issue through a handy set of five case studies, including social registries and satellite imagery in Costa Rica; selecting urban neighborhoods in DRC; web-based mobility data for food assistance in Chile; credit information for targeting in Colombia; and the digitized Integrated Social Assistance System in Turkey. Oh, an event on the paper is coming up on April 25 (h/t Tim Ohlenburg).
Since I mentioned targeting… Dietrich et al argue that not all PMT targeting errors are the same. A sole focus on “accuracy” is misleading as it assumes that societies are indifferent about who is being incorrectly classified. Instead, the authors use a social welfare framework and data from Tanzania and Malawi to weight targeting errors depending on the position in the welfare distribution. From this perspective, increasing accuracy may even cause welfare losses – and that such losses are underestimated due to bias in the data (label bias and unstable PMT weights). And this penalizes smaller households who are more likely to be incorrectly classified as non-poor.
More on Africa, with an “energetic” question: do social protection programs influence energy consumption of participating households? An article by Khundi-Mkomba et al shows that in Rwanda, the Girinka (asset transfers) and VUP (cash) schemes decreased the share of fuel expenses among beneficiaries. Such switch of spending in favor of food or health might, the authors conclude, “affect the overall household welfare”.
Human capital! What does the evidence say about the effects of cash transfers on human development? Das and Sethi just published a new (and succinct) systematic review of 44 studies on children’s education (enrolment, dropout, and completion rates), nutrition (stunting and wasting) and health (infant mortality). The review found that 73% of the studies present positive results, 16% detect no change, and 11% report an unfavorable outcome (see table 4, p.9-10).
Bonus on children, but from a migration perspective: a study by Yao and Walker evaluates the effects of Dibao and other cash transfer program on the welfare of “left-behind and migrant children” in China (Qinghai and Anhui provinces). Their conclusions put it bluntly: “[a]fter receiving benefit, the incomes of left behind and migrant children were not statistically different from those of their neighbours in the village or in the migrant community”.
From children to the whole population: Bashur proposes to test a basic income for peacebuilding in post-conflict settings, while the Basic Income Earth Network extended the call for papers for its Congress in Korea till April 30.
Let’s move to a couple of interesting labor markets resources: Seira et al reject the claim that universal health insurance may induce informality. In particular, their analysis on Mexico’s Seguro Popular finds no significant effects in reducing formal employment, suggesting the program wasn’t attractive enough to outweigh the benefits of having a formal job (btw, the Seguro social health insurance program was repealed in 2020) (h/t Florian Juergens-Grant). And ILO’s latest WESO report estimates that global coverage of social protection among workers is between 40.6 – 50.7%. Note that the lower bound rate is for “key workers”, or essential workers “that enable societies to function”, while social protection is there defined as eligibility and access to either pensions or paid sick leave (see p.92).
In need for a social protection training? TRANSFORM has a free online course on eight social protection modules, including Selection and Identification, Monitoring and Evaluation, Management Information Systems, Governance, Coordination, Legal Framework, Administration, and Financing.
A duet of development papers that caught my attention: Horn et al show that China’s role as “global lender of last resort” is growing, with its bailouts amounting to $240B across 22 countries over 2005-2021; and Sherman et al built a municipal-level human development index for Mexico using machine learning and satellite images.
Final assortment: the Swedish Pensions Agency released a report asking whether a new pension system is needed (yep, in Swedish…); West et al call for introducing either a guaranteed minimum income or reinstating the child tax credit to improve the mental health of Americans; a piece by Winder-Rossi and Razavi argues that 1.77 billion children lack access to child benefits (h/t Ian Orton); IDS reflects on 50 years of learning from Robert Chambers; and Pimpare has a talk (April 11) on how social workers and social work students can advocate for change.