Links Jan 26 – Comparative targeting performance in the Sahel; two papers on Ghana (on messaging and the effects of cash transfers on low-birthweight in high-temperature areas); assessing pandemic responses in Nigeria, Zambia and Saudi Arabia; two systematic reviews on cash transfers and violence and ‘cash plus’ on HIV; comparing disability benefits’ standards and practices; localization of humanitarian assistance in Afghanistan and Somalia; on testing social policies; and… take a rapid survey about the newsletter!

Welcome to the weekly links edition #300!

Let’s begin with a fresh look at a big question: an article by Schnitzer and Stoeffler evaluates the comparative performance of targeting methods using 9 datasets from 6 countries in the Sahel. Their results show that when evaluated against consumption, the median targeting error for community-based targeting (CBT) is 50%, while that of proxy means testing (PMT) is 39%. Yet PMT still performs better than a random allocation, that is, PMT provides 19% more resources to the poor than a random allocation versus 5% for CBT targeting; for food security, however, both methods are similar to a random allocation (table 2, p.8). Coverage matters: exclusion errors decline markedly with higher levels of coverage (see figure above), although after a certain threshold (around 50% of the population) the added value of targeting tends to diminish and broadly conform to a random allocation. Another important factor in shaping performance is the use of distributional-sensitive measures (i.e., the distance of the poor from poverty lines), such as the poverty gap and poverty severity: because the widespread and deep poverty, when those indices are used “none of the methods perform significantly differently from each other and from random or budget-neutral universal delivery of benefits”. What about administrative costs? Contrary to mainstream assumptions, variable costs tend to be relative low, i.e., for individual or combined methods they range between 0.5% and 5.5% of total transfers (see table above).

More from Africa… and be careful with messaging! A paper by Aurino and Wolf examines a large-scale randomized trial on SMS-nudges to parents. Text messages included information, reminders, and suggestions of practical, non-academic activities to engage with children’s education. For parents with some level of education, the program generated positive effects on child schooling and social-emotional skills (see figure 1, p.41). But among parents with limited education, “… the program backfired, exacerbating educational inequality [and leading to] undesired cognitive processes such as decreases in self-efficacy and educational aspirations, or increases in proboy bias”. Bonus on Ghana: LaPointe et al used data from the LEAP 1000 cash program and found that it protected infants from low birthweight in contexts of climate temperatures greater than 30°C.

Some further articles on COVID-19: a paper by Osabohien et al analyzed data from Nigeria’s longitudinal baseline phone survey and found that social protection reduced the probability of a household being affected by food insecurity during the pandemic. Siachiwena has a blog on COVID-19, social relief, and the informal sector in Zambia. And this one is really interesting. And Brik found that the Saudi government scaled up its social protection system in response to COVID-19 (86 implemented measures), but these largely neglected non-citizens and migrant workers.

Let’s go global: can cash transfers reduce violence? A systematic review by Machado et al sheds new light on the issue: based on 48 studies, their analysis shows that the evidence is strongest for (what they classify as “type I” grade) for cash plus reducing IPV and child maltreatment; cash for work lessening physical, emotional and sexual IPV, and physical child maltreatment. They also detected strong effects (Type II) that CCT or UCT interventions reduced suicide and transactional sex, and for start-up grants ameliorating emotional IPV.

But that’s not the only systematic review this week! Another one by Rogers et al assesses the effects of “cash plus” or bundled interventions for adolescents in Africa to reduce HIV risk. The reviewed 58 studies indicate more encouraging results for dimensions like mental health and psychosocial effects, as well as HIV prevention knowledge and testing (see table 4 for a comprehensive overview of results across 12 dimensions). Yet cash plus was less effective in improving metrics related to HIV incidence and sexual risk behaviors.

Interesting in the latest on disability programs around the world? ILO has a great report contrasting standards and practice – for example, see discussion on calculating benefits (flat rate in the case of non-contributory benefits, more diversified for contributory schemes, p.8-9), their frequency (mostly periodic, p.11), replacement rates (54-58% on average, p.13-14), and adequacy of transfers (28.7% of minimum wage, p.18).

Two thought-provoking pieces on “localization” of humanitarian assistance: Levine and Pain identify and discuss how to avoid ten traps in Afghanistan (“have your discussions started thinking about possible solutions very early? Are discussions on intervention design dominated by people who are experts in solutions because they have applied them in other contexts? If so, you are almost certainly caught in the trap”, p.9). And a blog by Diini points out how prevailing aid approaches don’t empower local actors in Somalia.

Assorted mix: Lokshin and Umapathi argue that “an untested social policy could be more damaging than crashing the A380”; MerciCorps has a Lebanon Crisis Update; Marcucci et al have a blog on anticipatory methods for migration; Horton et al propose a methodology to estimate the costs of unhealthy diets, preceded by a handy review of various measures of health and nutrition (p.3-8); and ISSA is having a webinar on linking research, policy analysis and innovation in social security on Jan 31.

And let me conclude with huge thanks for your readership over the past 7 years. If you have a couple of minutes, please consider taking a quick survey about the newsletter – would love to hear your feedback and suggestions!