Let’s kick off with a super race between three acronyms: PMT vs CBT vs FCS, which targeting method performs best in a low-income setting? A fantastic quantitative paper by Premand and Schnitzer examines the performance of proxy means test, community-based targeting and food consumption scores in Niger. They find that PMT is relatively more efficient at including the poorest and excluding the better-off based on consumption per capita (CBT and PMT are more closely aligned with subjective well-being measures). However, village inequality matters for results: in the case of CBT, inclusion errors are 14 percentage points higher in low-inequality than in high-inequality villages (errors are similar to other methods in low-inequality villages). Also, households assigned to formula-based methods (PMT and FCS) were up to 7.8 percentage points more likely to prefer repeating the method compared to households assigned to CBT (half the respondents expect committee members to benefit themselves). In contrast, legitimacy is found to be similar across methods in high-inequality villages.

From targeting to payments: what are we learning from digital technology in cash transfer responses to Covid19? Two gurus on the matter, Gelb and Mukherjee, present a range of findings: for example, digital payments have, despite some limitations, played a vital role (e.g., in Togo, Namibia and India); backup processes are essential (e.g., Namibia); digital campaigns can play a key role (e.g., South Africa, Brazil, and Pakistan); digital onboarding and screening can work up to a point (e.g., Turkey); lower transactions charges may not be sustainable after the emergency (e.g., Kenya, Rwanda); communication is central, even when scaling up existing programs (e.g., South Africa); and some concerns remain around data privacy (e.g., Pakistan, South Africa and Brazil).

Moving from a low-income to an even more extreme setting… how does the implementation of cash transfers look like in conflict situations? A thoughtful report by ICRC argues that cash is often best, but not always so. In particular, it provides a range of tips for the delivery of cash in contexts of armed conflict (see “cash wise” responses, need-for-speed and other considerations on p. 46-53), but it also warns against an “unthinking rush for cash” in humanitarian action. Such model, it is argued, may tilt action towards “cash-ready” environments and away from contexts where cash programming is challenging – that is, where markets are weak, delivery infrastructure is limited, or where highly volatile military situations make access to affected areas sporadic.

Conflict generates displacement, so what’s the latest on such theme? Sterck et al study the problem of debt among refugees receiving cash transfers in Kenya’s Kalobeyei settlement. In particular, they document that 89% of households are indebted towards retailers, and cash transfers are used as a form of collateral to guarantee debt repayment. The pressure exerted by creditors results in anxiety, high prices, and inability to select between competing retailers (see also accompanying blog) (h/t Paul Bance).

More on crises… but on anticipating them: can food crises be predicted with statistical forecasting? Using FEWSNET data for 21 countries, models developed by Andree et al provide good crisis predictive performance up to 12 months in advance. I really liked the way the paper balances trade-offs between false positives and false negatives. The implication is that “… when the costs of full-blown food crises are large and preventive measures have high returns, a forecasting framework with a high tolerance for false positives might be appropriate. Conversely, when resources for prevention are scarce, a greater tolerance for false negatives is more suitable” (p.20).

From food crises to the Covid jobs crisis: India’s NREGA employment guarantee scheme just got a sister idea — Jean Dreze’s DUET proposal for a wage subsidy program in urban India. How does it work? The state government issues ‘job stamps’ and distributes them to approved institutions; each stamp can be converted into one person-day of work within a specified period, with the approved institution arranging the work and the government paying the wages (statutory minimum) directly to the worker’s account; employees are to be selected from a pool of registered workers by an independent placement agency. Reactions? Ravallion’s reception of the proposal is rather lukewarm (“we need a broader discussion of (…) whether extra public employment and training, along with a government subsidy, is the best solution, or would even work at all”).

Let’s remain in the labor market realm: a great paper (and blog) by McKenzie reassesses the evidence on trainings! Since those programs are generally short in duration, he develops benchmarks for identifying their realistic effect size (i.e., an increase in profits of 4-5% due to trainings would compare favorably to regular education and capital investments). What did the paper find? Effects of trainings on firm profits are of 10.1% (15 studies) and on sales of 4.7% (17 studies); more tailored, alternative trainings increase those returns to 14.8% and 11.3%, respectively (estimates from 12 studies).

But do training and job-search programs work during a crisis? Michaelides and Mueser present experimental evidence on 4 US reemployment programs targeting unemployment insurance (UI) recipients during the Great Recession. Implemented in Florida, Idaho and Nevada, programs reduced UI spells, produced savings that exceeded program costs, and increased employment rates. However, in 3 cases programs had no effects post-intervention, while in one program (Nevada) produced substantial effects on employment and earnings in the subsequent year (apparently because job counseling services induced participants to undertake more effective job search) (h/t Aylin Isik-Dikmelik).

Looking for a gender perspective on social protection and the future of work? Bastagli and Hunt have a handy review how recent labor market trends may compound persistent challenges in undervaluation and unequal distribution of unpaid work between men and women (see section 3 for a nice discussion of all social protection components as well as services).

Let’s move to financing! The ILO has an excellent brief calculating current financing gaps for social protection: “… an additional $1.2T would be needed in 2020 to fully finance (…) a social protection floor in developing countries. This represents an additional investment of 3.8% of these countries’ GDP”. For LICs, the gap is almost 16% of GDP. Bonus on financing: Mexico issued world’s first sovereign ‘SDG’ bond, raising €750 million (h/t Cindy Paladines).

Final fireworks! The OECD has issued its state of fragility in 2020 report; a combo of thought-provoking articles by Slim tackles some very sensitive issues in humanitarian assistance: in one piece, he argues that humanitarian aid actors do not need to be neutral, and that neutrality is one of the ways in which to be humanitarians. After the storm the piece generated, a follow-up article restates that point by illustrating historical antecedents. And finally, Freeland has a highly critical blog about Schubert’s article claiming that Malawi’s social cash transfer program performs better than Lesotho and Eswatini’s universal pension schemes.