Can social protection help compliance with Covid confinements? Deiana et al show that in Italy, a food voucher program decreased mobility by 5% two weeks after lockdown. The effects persisted for two more weeks. How did they estimate it? Authorities allocated 8% of the program’s budget ($485M) to municipalities proportionally to their population; the rest was provided as a function of the difference between municipal and national per capita income in 2017. The paper then exploited the sharp kink schedule in the allocation of funds, which generated random treatment assignment in a neighborhood at a threshold point of $1,216.

Two health-related papers! Which programs can best prevent first pregnancy among adolescents? A rapid review by Estrada et al finds that among the most effective interventions are transfer programs, namely “…  school uniforms, school subsidies, or cash transfers as a way of increasing mainly school attendance or the opportunity cost of school absence”. A new article in The Lancet by Bauer et al argues that cash transfers should “… expand their focus to address longer-term mental health impacts of crises [on] susceptible young people (aged 15–24 years), including those with mental health conditions”.

From health to nutrition: in Myanmar, Field and Maffioli show that the combination of cash and nutrition information/trainings reduced stunting by 4.6 percentage point (13.5%) after 30 months. How? Mostly via better dietary diversity. Noteworthy that cash “alone” had no effects on stunting (the paper has a paywall – if you’d like to get a PDF just send me an email).

From nutrition to education: a paper by Gazeaud and Ricard on Morocco aligns with the global evidence base showing that conditional cash transfers improved school enrollment and attendance (in this case the dropout rate decreased by 1.3 percentage points on average); but learning needs more than cash, which in the study had small and even a seemingly negative effect on test scores at the end-of-primary school exams (h/t Dave Evans).

Some insights on intrahousehold dynamics… and from a country that hasn’t appeared much in recent research: a paper on cash transfers in Timor-Leste shows that benefits were largely spent on food and education, with decisions on their use mostly made by husbands and wives together. And cash helped “… informal businesses to reopen and re-establish trade vital for additional cash income” (h/t Anit Mukherjee).

Let’s take a look at targeting in fragile states: Bance and Schnitzer reflect on the conditions under which “lotteries” can be an appropriate method for beneficiary selection. The approach involves the transparent and random selection of beneficiaries in a public event. Where would it be desirable? Particularly conducive environments include post-conflict, volatile settings; contexts beneficiary data is lacking (no social registry, etc.); and where “everybody is poor”, with differences in welfare being nearly indistinguishable. Experimental evidence from DRC and Cote d’Ivoire shows that the approach holds promise. One of the interesting features was that the selection (extraction) turned into a positive event, a celebration that cemented social cohesion.

Speaking of volatile contexts… Barrantes and Caravani argue that in Uganda, part of the reason for Karamoja’s development challenges is its “exceptionalism” framing, and that a “normalizing view” is needed “… to positively include Karamoja as a contributor to overall socioeconomic development of the country”.

More on fragile states: a thought-provoking blog by Carrasco discusses whether humanitarian actors should engage in social protection when the upholding of humanitarian principles cannot be fully ensured. The discussion, which largely revolves on data management, is a variant of longstanding tradeoffs the humanitarian community faces when engaging in development: the latter needs at least some work with states, but most states have, as the blog puts it, “… low data protection infrastructure or standards”. Hence the dilemma of reconciling the adherence to humanitarian standards with growing calls for “localization” assistance… is there a middle ground to espouse humanitarian and ownership principles?

Bonus: what’s the state of data to inform humanitarian assistance? A new report from OCHA finds that “… [a]s measured through the Data Grids, just over 50 percent of relevant, complete crisis data is available across 27 humanitarian operations”. Double bonus: interesting event on Feb 17 (2pm CET) with IFRC’s Holt speaking about digitization of humanitarian assistance.

More on digital transfers! Gupta offers a framework with four layers or filters for exclusion on direct benefit transfers in India. And speaking of India… Afridi et al show that the 2021 budget allocation for NREGA public works is 34% lower than revised estimates for 2020-21 – this is disappointing since districts with good capacities have been able to use NREGA funds for employment recovery during the pandemic (in those cases, NREGA reduced job losses in rural areas during April-August 2020 by 3.1 percentage points overall or 7% over the pre-crisis employment rate). Plus… Wapling et al also discuss how NREGA itself has been able to include people with disabilities, and p.50-52 assess the (variable) extent to which such commitment occurs in practice.

From Asia to LAC: is it myth or reality that the political left spends more on social protection than other parties? A new paper by Gouvea and Girardi studies such partisan effects in Brazilian cities. What did it find? Based on a sample of 8,943 municipal elections between 2004-2012, the authors estimate that “… left-wing mayor tends to raise the share of social expenditures [including cash transfers] by 0.6 percentage points [or] a 1 percent increase in social spending per capita”.

Let’s turn to labor and skills (h/t Michael Weber)! Basu and Dimova revisit the causes of child labor in Ethiopia and found that this increased in households where the head is more risk averse (i.e., due to either low or uncertain returns to education, some parents favor child labor at the expense of schooling). On the other end of the lifecycle, Lee et al find that in Korea, attainment of ICT skills and participation in job-related training can help older workers stay productive.

In Bangladesh, Das estimates the effects of a on-the-job and classroom training to disadvantaged unemployed youth: the program increased labor market participation by 16 percentage points and earnings by 23%. The effect on employment declines after 22 months, but the effect on earnings is sustained as the program induces a shift from casual work to wage employment. Bonus on Bangladesh: Canessa and Giannelli estimate that floods increased women’s employment probability increases by 13 percentage points. Extra evidence on gender: Berniell et al show that in Chile becoming a mother implies a sharp decline in employment, working hours, and labor earnings; also, the birth of the first child produces a strong increase in labor informality among working mothers (38%).

Couldn’t share it last Friday since the paper was embargoed till the afternoon – but if you haven’t seen it yet, an excellent Science article by Egger and 25 coauthors assessed the impact of Covid in Burkina Faso, Ghana, Kenya, Rwanda, Sierra Leone, Bangladesh, Nepal, Philippines, and Colombia: their survey of 30,000+ households showed that 68% of them experienced an income drop (see also a handy IPA brief) (h/t Jeff Mosenkis).

Assorted regional news: Jamaica approved a new social pension program of $22/month, with the amount indexed to half of the minimum pension under the National Insurance Scheme (h/t Florian Juergens); and Chelwa examines 10 academic journals and finds that only ¼ of papers on Africa published over 2005-15 have an author based in the region (h/t Amber Peterman).

Let’s end with a historical gem: Ravallion found that border discontinuity in impact evaluation was mentioned in 1893 by German doctor Robert Koch for assessing the impact of water filtration on cholera.