Plenty of vibrant discussions on the role of cash transfers in the ‘graduation’ agenda…
Banerjee et al are back with a new NBER paper on the classic graduation model (a package of assets, training, coaching, and savings). They explore two variants: whether the transfer of assets only would generate similar impacts, and whether access to a savings account and a deposit collection service would generate comparable impacts. Neither outperforms the holistic package. Similarly, a CSAE paper by Sedlmayr et al assesses graduation variants in Uganda--the full package of transfers and training, only the transfers, transfers with only a light-touch training and just attempting to boost savings. They find that cash only was less effective than the more integrated interventions.
Pritchett weighs in (again) on such model: in a new CGD paper, he estimates that gains in graduation interventions are about 40 times smaller than letting people migrate to a high-productivity destination. His second-best option? Broad-based domestic economic growth, which he argues is “many-fold” larger in impact on poverty than direct individual interventions.
When interviewed by Tyler Cowen, Chris Blattman mentioned that the type of cash transfers used for graduation “… are much more of a temporary acceleration than they are some sort of permanent solution to poverty”. Picking up on this, Berk Ozler calls for being even-handed about findings that support our priors and those that don’t do so.
New work on universal basic income spanning implementation, simulations, and politics.
About a year ago, India released its much-awaited Economic Survey featuring a chapter on universal basic income – or better, a quasi-UBI since it excludes the top 25% (see here). This is now the subject of Carnegie’s Khosla new book, which points to a range of upsides and bugs with the ES proposal. You can also check out a book online event featuring Justin Sandefor and Rinku Murgai (listen in particular to her insightful remarks at min 50 and 60).
A new IFS working paper by Harris et al. examined the potential of a VAT-financed UBI in four African countries. The paper found that broadening the VAT base (eliminating most exemptions and tax reductions) and using the revenue to fund a UBI would be a way more efficient of achieving redistribution than a VAT. A UBI would reduce extreme poverty and inequality – even if only 75% of the additional revenue was used for this purpose.
The latest European Social Survey (ESS) included a question about UBI. An analysis by Vlandas reveals that the largest number of respondents, or about 45%, are in favor of the scheme while 15% strongly oppose it. But when it comes to political matters, a new paper by Gourevitch and Stanczyk claims that proponents of UBI got it all wrong. It’s not that a UBI would empower workers (“the power of saying no”): workers would first need to get organized as influential political agents and then, as a consequence, they may be to lobby for a UBI. Build such political constituency in the changing world of work is, they argue, a daunting challenge.
What about crises and resilience?
A new toolkit by O’Brien et al complements previous extensive lit reviews and case studies shock-responsive social protection. This time they dig into how to connect disaster risk management and humanitarian assistance, including in a step-by-step fashion. Another toolkit was released by WRC, this one dedicated to making cash transfers more sensitive to gender-based violence (GBV). Materials include focus group discussion/interview tools, accompanying guidance to assess and mitigate potential risks, post-distribution monitoring forms, protocols to assess and address GBV survivors’ needs, case management services, and more. Three case studies on Jordan, Niger, Somalia provide insights on practical applications of the materials.
A package of early humanitarian response and safety nets is about 30% more efficient than typical humanitarian aid in Africa. A super interesting USAID study by Cabot Venton updates and deepens a previous 2012 paper. The study uses Household Economy Analysis based on price, rainfall and production data between 2001 and 2016, and models avoided income and livestock losses for 15 million people living in 54 livelihood zones in the Horn. Findings? Relative to typical humanitarian assistance, an early humanitarian response would save an estimated US$2.5 billion in humanitarian aid costs over a 15-year period. Safety nets save US$3.5 billion over the cost of a late response, or an average of US$231 million per year. A combined, resilience-building scenario (early humanitarian response + safety nets) could save US$4.3 billion, or an average of US$287 million per year. In other words, every US$1 spent on safety nets or resilience programming results in net benefits (savings) of between US$2.3 and US$3.3, respectively. Check out also the infographic and case studies on Ethiopia, Kenya and Somalia.
Bonus: Elliott examines the prospects for US food aid reforms (more cash and less monetization); extra bonus: a 150-page report by Babu et al evaluates the 2015/16 emergency response in Malawi.
On the past and future of social protection…
One of the most fascinating questions in social protection is how history and path-dependency regimes influence the shape of current welfare models in low and middle-income countries. Hickey et al have an awesome WIDER paper exploring the issue in Africa. They show that countries with the highest levels of commitment to social assistance are driven more by domestic political imperatives than by external pressure. Importantly, this process has been strongly informed by long-standing ideas around deservingness and the role of the state.
A great presentation by Nancy Birdsall sheds light on the ‘strugglers’ — those not poor enough, not yet middle class, but with high aspirations. The strugglers constitute about 60% of the population in many large middle-income countries – and will keep doing so in 2030. How to deal with it? Her slide 28 touches upon an array of social insurance, automatic stabilizers, targeted and universal options.
Can a social protection system be at the same time inequality-reducing and poverty-augmenting? Yes, it can. A new working paper by Davalos et al examines the redistributive effect of fiscal policy on income distribution and poverty in Albania. Based on the CEQ methodology, the authors found that the fiscal system played a positive role in reducing inequality. Yet, taxes and contributions had a moderate poverty-increasing effect, particularly VAT taxes. This effect is somewhat compensated by direct pro-poor government transfers (e.g., Ndihma Ekonomike program), but which are not large enough to offset fully the negative impact on the taxation side.
Ridao Cano and Bodewig have an amazing report – ‘Growing United’ – on the EU’s convergence trajectory. They provide an in-depth examination of the skills divide, bottlenecks for firms and innovation, and elements social contracts that could be upgraded. On safety nets, they propose a combination of an expanded set of innovative public works, including for services, while relaxing some constraints on minimum guarantee schemes.
Speaking of public works…
Gehrke and Hartwig’s review of the evidence suggests that public works programs are only preferable over alternative interventions if they generate substantial investments among the target group, if there is clear evidence that private-sector wages are below equilibrium wages, or if the public infrastructure generated by programs has substantial growth effects.
Work requirements are widely debated in the US, too. In a House Committee testimony, Hahn discusses research on and implications of potential expansions of work requirements in Medicaid, SNAP and other programs. Her main point is that “work requirements create cumbersome administrative processes that affect both the people who are eligible for assistance and the government agencies providing it”.
Since I mentioned SNAP… this month saw the release of two contrasting papers on the program: a report by Wheaton and Tran found that the program lifted 8.4 million people out of poverty in 2015, reducing the poverty rate from 15.4 percent to 12.8 percent (a reduction of 17 percent). Also, SNAP reduced the poverty gap (the aggregate amount of additional income required to remove all poor families from poverty) by $35 billion (21 percent) in 2015. However, another report by Waxman et al question SNAP’s adequacy: the average cost of a low-income meal is $2.36, or 27% higher than the SNAP maximum benefit per meal of $1.86. In other words, SNAP per meal benefit does not seem to cover the cost of a low-income meal.
Final assorted fireworks!
Goldstein and Evans have a great round-up of papers from Oxford’s CSAE conference on economic development in Africa. Among the social protection-related materials: a secondary school stipend (with a condition on not being married) in Bangladesh led to higher education and no change in fertility or female labor force participation (Tanaka and Otzuka) is it better to get cash transfers quarterly or monthly? In Nigeria it doesn’t matter for nutrition, consumption or investment outcomes (Bastian et al); and bank cards for cash grants in South Africa increase women’s autonomy, which in turn augments female labor force participation by 62% (Biljon et al).
A new evaluation by Courtin et al investigate the impact of the old CCT in New York City. The program, called Family Rewards, supported 2,377 families between 2007-2010 and led to overall modest effects. Specially, it helped improve in health insurance coverage, increased the use of preventive dental care, and enhanced parents’ qualitative perception of their own health and levels of hope. However, the program was consequently phased out, although a new experiment was eventually set up in Memphis.
Finally, an ACF paper shows how seasonal cash transfers helped mitigating malnutrition in Burkina Faso while Azeem et al have a JDS article showing the (positive) effects of cash transfers in Pakistan’s Punjab.
This monthly column draws from a weekly newsletter – to join send me an email: (firstname.lastname@example.org). Previous editions of the column can be accessed here and here.