Weekly Social Protection Links: 04 December 2020
Let’s start with a smile! Does money make people happier? McGuire et al review the impact of 38 cash transfer studies on measures of subjective well-being. They find that after about 2 years, the average effect of CTs is 0.10 standard deviations in an well-being index; the size of cash benefits matters both in absolute (left hand graph above) and relative terms (right graph); unconditional transfers have a similar effect (~0.04 SDs) than those with conditions; there is also evidence of tapering out over time, with average effect reaching zero after about 7.5 years from program launch.
Since I mentioned generosity, a paper by Haushofer et al estimates that the provision of large unconditional cash transfers (equal to $1,076) had greater psychological (and consumption, asset holdings, and revenue) effects than a five-week psychotherapy program (and had the same effects as combining cash and training) one year after the interventions. The study also finds that that weekly transfers have a substantially larger impact than lump-sum transfers on both consumption ($15 vs $7) and revenue ($51 vs 19); but such difference doesn’t emerge for asset holdings, subjective well-being, and violence.
Speaking of psychology, can the type of narratives undermine beneficiaries’ motivations and thereby the effects of cash transfers? An experiment by Thomas et al shows that cash beneficiaries in Nairobi were more motivated to build their skills (and their stigma was reduced) when the program’s narrative involved “empowerment” of individuals or communities as opposed to “poverty alleviation” focusing on people’s deficit in circumstances or abilities.
Same region, same intervention, but different theme: in rural Togo, Briaux et al found that unconditional cash transfers provided in the first “1000 days” had a protective effect on child’s growth; increased mother’s (+4.5pp) and children’s (+9.1pp) consumption of animal source foods; reduced the financial barrier to seeking healthcare for sick children (−26.4pp), improved the odds for women of delivering in a health facility (+10.6pp), and lowered the chance of giving birth to babies with birth weights (<2,500 g) (−11.8 pp). However, cash had no impact on stunting and child morbidity (h/t Amber Peterman).
Does it matter if transfers are provided by the state or other actors? As part of a wider paper on whether evaluation results are generalizable, Vivalt estimates that government-run programs (including 13 cash transfers, see p.27-28) have smaller effects than academic or NGO-managed schemes – in other words, there are trade-offs between scale, sustainability, and impact!
A quintessential government-run program is NREGA public works in India: between April and August 2020, the scheme was expanded by 22%, with over half (55%) of the “job cards” issued in the ‘high’ outmigrant districts. Narayanan et al reflect on such experience and argue that 100 days of work does not offer adequate protection (in fact select states have requested 200 days/household). Delays and rejections in wage payments continue to remain serious problems, while barriers to accessing employment, especially among return migrants, tend to be significant.
Two bonuses on India: Prakash’s model predicts that a 1-percentage point increase in the employment quota for Scheduled Castes increases the likelihood of obtaining a salaried job by 0.6-percentage points for it rural caste members. And here is a great review of two books on India’s 21 months of emergency in 1975-77 (h/t Martin Ravallion).
Gender! Wijesekera reflects on gender-based violence and social protection, while a JMP by Etcheverry shows that Uruguay’s 2014 subsidy facilitating women’s gradual return to work after maternity leave increased the likelihood of employment by 17% (h/t Dave Evans).
More on work tied to safety nets: in the US, Gray et al show that work requirements in SNAP food vouchers led to 52% decline in program participation, especially among homeless and people too poor to comply. BTW, in the US Chetty et al built an amazing database tracking economic activity at a granular level in real time using anonymized data from private companies (they report daily statistics on consumer spending, business revenues, employment rates, etc. by ZIP code, industry, income group, and business size).
Social pensions can be effective crisis response mechanisms! For instance, Bottan et al found that during Covid19, Bolivia’s Renta Dignidad (which provides $50/month to seniors >60, reaches 1/3 of Bolivian households, and costs about 1% of GDP) increased the probability that households had a weekly stock of food by 25% and decreased the chance of going hungry by 40% (h/t Florian Juergens).
And since we are talking Covid… new research on urban informal sector workers is out! Roever and Rogan discuss results from the first round of the WIEGO-led Crisis and the Informal Economy Study in 12 cities. The analysis shows a massive drop in earnings (e.g., in India’s Ahmedabad city, 96% of street vendors reported zero earnings during the height of the pandemic restrictions). Also, coverage by government emergency cash and food transfers was limited and uneven: on average only 40% of workers were covered… including 1% in Dakar and Dar es Salam, and none in Accra.
More on crises, with a duet of new humanitarian papers: Barbelet and Wake review the concept of “inclusion” in humanitarian action and call out an “over-reliance on technical fixes”, narrow definition of needs and broader issues like diversity in the sector. And a paper by Nimeh et al examine how to connect humanitarian assistance and social protection in a refugee context (I liked the trajectory illustrated on p.14) and concludes that “… [r]ecognising the inherently difficult political economic repercussions surrounding its provisions, social protection shadowing is proposed as a pivotal element to come to a development led approach”.
What’s new on skills and labor market institutions (with Michal Weber and Indhira Santos)? Two papers on China! Fang et al estimate the effects of minimum wage increases in China: these raise the wages of otherwise low-wage workers by less than half (41%) of the minimum wage growth and lead to a 2 to 4 percentage point reduction in the probability of being employed. In another article, Zheng et al show that about half of online job-seekers in China are two or more years overeducated, resulting in 5.1% pay penalty.
Moving to Europe, what’s the state of self-employment in the UK eight months into the pandemic? A survey by Blundell and Ventura found that 1 in 5 self-employed workers will likely leave self-employment. In Germany, Hennecke and Pape show that when fathers become unemployed, paternal child care and housework significantly increase in the short run on weekdays… but not on weekends. And in Sweden, Caliendo et al found that within one year, training programs reduce the probability to have a prescription for cardiovascular and mental health drugs by about 6-8% (and a reduction of sickness absence by about 20%): the impacts are likely due to the adoption of a more rigorous daily routine. Unemployment benefits sanctions cause a short-run deterioration of mental health, possibly due higher stress levels, but this tapers out quickly.
Final assortment! Ravallion uploaded his lecture notes on “Growth, Poverty and Inequality” (part 3.2); USAID just released a new “evidence framework” for external feedback; and VoxDev is pulling together a series of Wikipedia-style literature reviews in development topics.
Ugo Gentilini is from the World Bank’s Social Protection & Jobs global practice. The Social Protection Links newsletter, issued every Friday, distills and discusses a selection of curated resources on the topic, from academic articles to podcasts. The blog is republished on socialprotection.org each week, offering knowledge on social protection to helps you stay on top of it — succinctly, regularly and frequently. Previous editions can be found here.
To sign up to the newsletter or share materials, you can contact Ugo by email ([email protected]), Twitter (@ugentilini) or LinkedIn.