Written by Ugo Gentilini.

A step towards reconciling thorny dilemmas: a brilliant paper on social assistance by Brollo et al quantifies trade-off between targeting methods, expanding coverage, poverty impact, and fiscal costs. So much to unbundle here: the analysis reiterates that while poverty-based methods could be difficult to implement and present social and political challenges, also “categorical targeting (such as say child benefits or social pensions) … comes with substantial leakage to higher-income households” (see figure 7, p.18 for an application to India). This implies that whatever the criteria, expanding coverage among people living in poverty inevitably generates – intentionally or not – “leakage” of transfers to the non-poor. More to the point, the analysis found that higher coverage of the poor actually requires higher population coverage and cannot be pursued in isolation: “… no country achieves high coverage of poor households without having high total coverage”. In other words, to cover all people in poverty (extreme or bottom quintile, see graphs above), data from 84 countries shows that between 75-78% of the total population would need to be covered (and coverage of the poor increases at a slower rate than coverage of the population, see table 1, p.15). In turn, if we need more coverage of the population to have more poverty impact, this entails higher budgets (the authors identify spending as the key driving factor in poverty gap reduction, see figure 3, p.12). But how much is needed? All considered, there is a hefty price tag for covering about 90% of the poor (and 70% of the population), i.e., roughly 3% of GDP. Ironic how the destinies of “poverty targeting” and “universality” are intertwined.

Gender! Groves et al evaluate a trial testing the effects of a conditional cash transfers (based on school attendance) on intimate partner violence among adolescent girls and young women in South Africa (a monthly $10 to students and $20 to parents). Over the intervention’s four years, receipt of cash transfers had a protective effect on risk of physical IPV, and a “marginally significant effect” on sexual debut, likelihood of having a sexual partner or having more than one sexual partner (see table 3 for pooled effects). The impact on IPV declined over time — compared to year 1, the effect of cash in year 3 dropped by 22% in Year 3, while impacts post-intervention (after year 4) were not statistically significant. See also the qualitative analysis, where “young women described how receiving cash enabled them to engage in healthy relationships and in some cases, to leave violent relationships”.

Bonus on gender: Cookson et al review 53 publications and identify literature gaps around gender and social protection, including in terms of conceptual framing, institutional features, benefits offered, and overall performance (i.e., equity, inclusion and resilience metrics).

Speaking of gaps… an “evidence gap map” of cash transfers and cash-plus programs by Pasha et al scanned 709 impact evaluations and 33 systematic reviews. Main results indicate that the evidence balance is tilted towards conditional designs; there is an emphasis on health and consumption outcomes; relatively little studied are population groups like smallholder farmers, migrants, and people with disabilities; and with evidence mostly stemming from rural areas and “stable”, non-emergency settings.

But hey, here is a specific area where there appears to be literally zero evidence: a systematic review of 43 studies by Gabrielli et al found that “[n]o studies have yet assessed the effects [of social protection] on breast and cervical cancer incidence, survival, or mortality”. (Yet there is plenty of evidence on risk and mediating and protective factors, such as accessing Papanicolaou screening or effects on overweight, see figure 3 mapping them out).

More on a specific form of (mental) health: a paper by El-Enbaby et al found that Egypt’s conditional cash transfer Takaful does not significantly impact anxiety levels of recipient mothers. Measured by the generalized anxiety disorder 7-item scale (GAD-7), main results are summarized in table 2 and visualized in figure 4. What can explain such null impact? For once, the size of transfers is modest, including representing only 12.7% of income of households living under the poverty line. Also, perceived financial security that come with the CCT may be eroded by a level of uncertainty surrounding administrative decisions to discontinue participation.

Buckle up, this is a sobering one: based on interviews with 26 officials and focus group discussions, an article by Langnel and Tweneboah-Koduah argues that the LEAP program in Ghana has become heavily politicized. After an initial period of data-driven decisions, political influence started to be exerted via partisan composition of the District Level Implementation Committees, which select eligible communities after a district-level targeting. DLICs deliberate selection of participants based on political affiliation (as well as including friends and relatives), and even requiring beneficiaries to hold party membership cards.

Let’s get a positive boost: a great, open access edited volume on informality and its future was just launched by Chen et al. I only had a chance to read two chapters so far, both intriguing, i.e., the one by Kanbur et al (“we have a complex web of schemes with their own histories and logics. It is this system that we have to work with, not some abstract first best world, albeit it might provide general guidance on direction of movement”) and the contribution by Alfers and Juergens-Grant (“if one important stream of financing—financing from those who profit from the work of informal workers—is precluded, universal social protection is likely to remain highly constrained and probably not very universal”).

On automation… and insurance!! A paper by Brollo suggests that US states with more generous unemployment insurance were better able to mitigate the negative labor market effects of automation. The study, which focuses on the period 2000-2007, finds that protective effects were particularly pronounced on wages instead of employment in general (see figure 1, p.11), although insurance was very important in mitigating job losses among workers with lower levels of education (figure 2, p.12). There are also effects on poverty, where one additional robot per thousand workers increased the poverty rate by 3%: here UI didn’t help much, but social assistance (TANF) offset such impacts where transfer generosity was relatively higher (figure 3, p.13).

Moving to another high-income country, a paper by Baudisch and Neuenkirch compared Germany’s  VAT rate reduction during the pandemic to cash transfers such as those implemented by a cohort of peers like Canada, Denmark, Japan, and the US. The paper concluded that cash transfers were more cost-effective at boosting consumer spending (see figure 2, p.24). Why? Because they were “more comprehensible, salient, and actionable… in a dynamic environment with high uncertainty induced by unclear future economic prospects”.

And since I mentioned taxes… let’s look at two financing materials. One is on climate: a review by Hopper et al highlights the gulf between the large body of literature linking social protection to climate change and the actual funding of social protection within the climate finance architecture. (And if you are wondering… right, there are no clear figures yet on the volumes of climate-related social protection financing). The other material is an article by Kentikelenis and Stubbs is quite critical of the IMF’s recent track record on social protection, claiming that “… 15 of the 21 countries studied here experience a decrease in fiscal space over the course of their IMF programs”.

The dessert is served: a blog by Coudouel et al discussed the WBG, WFP and UNICEF partnerships on adaptive social protection in the Sahel (h/t Deborah Kirby); recordings are now available of the CSW68 sideline event on “Power, practice and potential: social protection and adolescent girls” (and see Palermo’s slide deck here), while here are those of another session in the same event – “Why universal social protection is essential for inclusive development, peace and security in the present and future”; and on May 16 – my birthday – tune in for a seminar on measuring disability related costs for social protection, followed on May 23 by an webinar on Lebanon’s Social Registry (DAEM).

Social Protection Programmes: 
  • Social assistance
    • Social transfers
      • Cash transfers
        • Cash plus
        • Conditional cash transfers
  • Labour market / employment programmes
    • Passive labour market policies
      • Unemployment benefits
        • Unemployment insurance (contributory)
Social Protection Building Blocks: 
  • Policy
    • Coverage
    • Expenditure and financing
  • Programme design
    • Targeting
  • Programme performance / impact analysis
Social Protection Approaches: 
  • Adaptive social protection
Cross-Cutting Areas: 
  • Climate change
  • Disability
  • Gender
    • Gender-based violence
  • Health
    • COVID-19
  • Labour market / employment
    • Informality
  • Global
  • Global
The views presented here are the author's and not socialprotection.org's