Syndicate content

Pushing the Envelope

Laura Ralston's picture

Giving Cash Unconditionally in Fragile States

2012 Spring Mtgs - Close the Gap There have been many recent press articles, a couple of potentially seminal journal papers, and some great blogs from leading economists at the World Bank on the topic of Unconditional Cash Transfers (UCTs). It remains a widely debated subject, and one with perhaps a couple of myths associated with it. For example, what is cash from UCTs used for? Do the transfers lead to permanent increases in income? Does it matter how the transfers are labelled or promoted? I am particularly interested in whether UCTs could be a useful instrument in countries with low institutional capacity, such as fragile and conflict-affected states (FCS).

Why UCTs in FCS? UCTs present a new approach to reducing poverty, stimulating growth and improving social welfare, that may be the most efficient and feasible mechanism in FCS. A recent evaluation of the World Bank’s work on FCS recognized, “where government responsiveness to citizens has been relatively weak, finding the right modality for reaching people with services is vital to avoiding further fragility and conflict”. Plus there is always the risk of desperately needed finances being “spirited away” when channeled through central governments. UCTs may present a mechanism for stimulating the provision of quality services, which are often lacking, while directly reducing poverty at the same time. As Shanta Devarajan’s blog puts it, “But when they (the poor) are given cash with which to “buy” these services, poor people can demand quality—and the provider must meet it or he won’t get paid.” We should explore more about this approach to tackling poverty: where and when it has worked, what made it work, and whether we can predict whether it will work in different contexts.
 

There is already a large literature on conditional cash transfers (CCTs) that I will not attempt to cover here, so let us instead consider the potential benefits for making transfers unconditional, rather than conditional. First is feasibility. Many FCS countries simply do not have the service infrastructure in place to condition transfers. For example, conditioning transfers on attendance in school is only feasible if there are enough places in schools to accommodate all the potential students, without a detrimental impact on the quality of education. However, as Devarajan suggests, providing unconditional transfers may catalyze service delivery by private providers, and provide the right incentives to these providers to maintain quality. Second is targeting. By making transfers unconditional we may be able to increase the likelihood that we reach the recipients who would most benefit from the transfer – not just the ones who face the lowest costs of reaching conditions, or who would have met conditions anyway. Often, the poorest households live furthest away from schools and healthcare centers. Conditioning the transfer on specific outcomes may mean the poorest are less likely to receive the transfer and miss out on its long-term benefits. Third is transparency. By making the transfers unconditional the process of allocating the transfers will be simplified, increasing the scope for the use of automatic electronic transfer mechanisms (e.g., the use of m-Pesa in Kenya or pre-loaded ATM cards for Syrian refugees in Lebanon). Electronic transfers create a paper trail enabling better tracking of the cash to recipients and onto purchases, which could be useful for reducing leakage, detecting corruption and learning about the consumption of the recipients.

Furthermore, there are arguments for re-thinking the targeting of transfers. Recent studies have shown inaccuracies in the use of proxy indicators to target poor people. In Indonesia, a proxy-means test mimicking the government’s standard practice incorrectly excluded 52% of truly poor households and incorrectly included 20% of non-poor households, based on their consumption levels. In addition, these methods of identifying the recipients are often costly and reduce the amount left to distribute. The standard features of CCTs, targeting and conditionality, have been estimated to account for 60% of the administrative costs of PROGRESA, 49% of the costs for RPS in Nicaragua, and 31% for PRAF in Honduras. Perhaps it is worth considering universal UCTs in specific geographic locations – for example, rural areas with a high proportion of the population below the poverty line - where the lower costs of universal transfers outweigh the benefits of more accurate targeting. There may also be a case for making transfers universal in fragile situations where any kind of targeting may exacerbate existing tensions.

Despite the potential benefits of UCTs, there are concerns. Will the recipients of transfers use them for productive purposes? Will UCTs be the most efficient and economical way of using aid to reduce poverty? Fundamentally we would like to see the recipients use the transfers to choose investments that give them the best chance of escaping poverty (and hope that this will coincide with maximizing their welfare). However, this is a really difficult problem to solve, and I would bet that many of us well-educated working professionals are not even fully optimizing or on our own production possibility frontiers! Yet, there has been recent evidence that recipients of UCTs use them productively. New research has found that the recipients use them to make investments and improve their monthly earnings, and in another paper focusing on transfers to teenage girls in Malawi, UCTs had an important impact on lowering marriage and pregnancy rates, helping to empower these women. While these papers provide only a partial picture of the potential outcomes of UCTs, their initial positive results suggest that it is an area the World Bank and other development agencies should explore further.

And UCTs need not be implemented in isolation: finding ways to equip recipients with information on their investment choices are important parallel interventions. As Richard Thaler points out, “in some cases (giving UCTs) will work great. When people have a good sense of how to make use of that money.” He is an advocate for creating nudges to help people make the right decision and advises the British Government, which now has its own “Nudge” unit. Simply labelling UCTs with a productive investment choice may be possible approach. A recent paper found that a cash transfer labelled as an education support program was sufficient to increase school participation even though no conditionality was imposed. In Poor Economics, the authors note, “It is not easy to escape from poverty, but a sense of possibility and a little bit of well-targeted help (a piece of information, a little nudge) can sometimes have surprisingly large effects”.

What do you think? Can UCTs provide the resources and, assisted with “a little nudge”, can they work to reduce poverty, spur growth and improve service delivery in FCS?
 

Comments

Well done, Laura - great synthesis on UCTs. So, I know you're a fan of (or at least a skilled user of) large-n studies, so are there any large-n studies on UCTs in the works on your end? I'm all for shifting some WBG aid from projects to UCTs, maybe partially conditional on its recipients providing some feedback on the use of those UCTs.

Submitted by Laura on

Thanks Alison - when you say large-n studies do you mean large sample sizes or meta analysis of existing studies? For the two UCTs referenced in the post, Haushofer and Shapiro have a sample of 1440 households, while 535 individuals in the Blattman, Fiala and Martinez study. As yet, to my knowledge, there have been no meta-analysis of the impacts of UCTs.

Submitted by alexandre marc on

Laura thanks, it is great that you bring this issue in the open for discussion. I always thought that there was a naive enthusiasm about UCT and its benefit in FCS or any country with weak institutions. The idea that you put money in hand of people and that it will automatically empower them and they will require the State to behave in a better way is certainly naive especially when there are no services to speak of provided to communities and that the community is already paying for what ever services offered. At the same time we have so few instruments that allow communities to sense some form of beneficial presence of the State in situations of conflict and social and political tensions. As you say the issues ar two fold: how can you distribute cash benefits fairly and equitably when you have no real institutions and how can you avoid political capture of the process when there are so few checks and balances. Also we know that war economies have their own logic and that an unsecure environment creates all sort of incentives. urchasing weapon will be the first priority of a community that feel unsafe (and it might be good in teh short term for teh community). Racket is extremelly high in situation of conflict and I would imagine than jihadist and local milicias would love to see the community they control by night with a bit cash in their hands. There is nothing more market oriented than a war economy and the cost of going through road blocks or for purchasing essential goods would immediately go up with cash in the communities. So this makes me at the best sceptical of UCT in FCS unless this is really thought through extremelly carefully.

Submitted by Laura on

Dear Alex, Thanks for your insightful comments. Indeed, there will be considerable risks that must be thoroughly thought through before implementing any type of development policy in FCS, and with UCTs you rightly question: (a) whether the cash would increase the risk of violence if they are spent on weapons, and (b) whether the cash may make recipients more vulnerable to extortion. I think for (a) we would need more information on the preferences of the recipients to know if they would really spend it on weapons, and whether if security could be provided in another way this preference would still prevail, i.e., with a parallel policy on security could recipients be nudged towards more economically productive and less defensive investments? On (b) investigating ways to make transfers secure and less vulnerable to extortion is a must, but not making communities better off, in case they may then become more vulnerable does not seem the right argument for in action.

Submitted by Benim on

A nice literature review and I guess the question at the end of the long blog is "can a nudge, get people to budge?". Well, I am sure it can't hurt but by the time you get to making that decision you need a system in place to deliver the cash to the right people. While the writer is good at telling us about the "what" part of the story, it would be useful to hear a bit more about the "how" in specific contexts. How does one overcome resistance to what appears like throwing money out the door? How does one sell the idea of universal UCTs in specific locations? Even refugees engender resentment, so what can one practically do to allay fears by others? As always there are winners and losers - empowering one community or group may be seen as disempowering another. A bit more on how these and other challenges were actually overcome would be great.

Submitted by John M. Anderson on

Laura, many thanks for this very interesting overview. While noting that I am an expert neither in UCTs nor tax policy, one initial thought that came to mind is the similarity between a UCT and a tax rebate, such as the earned income tax credit (EITC) in the US. Through the EITC, people earning low incomes are credited back a portion of the taxes they paid. This seems to me to be related to a UCT - there is the condition that you have to have earned an income, I suppose, but other than that the government is effectively giving people an unconditional transfer to spend however they'd like based on one poverty indicator (salary level). There is of course the difference that the tax credit was previously the individual's money in the first place; but come April 15 in the US, this still represents a transfer from government to individual. You mention relevant concerns raised about UCTs, such as whether people will use them productively; in terms of politically selling the idea of UCTs, perhaps the experience of these kind of tax rebate programs in developed economies could be considered.

Submitted by Laura on

Dear John, Thanks for your comments - they seem to speak to Benim's comment about how to make UCTs politically feasible, and is a concept worth further consideration. The constraint I envision in FCS will at least in part be how to track poverty levels of recipients. Perhaps with electronic cash transfers through mobile phones it will be possible to both track transfers and collect information on income levels (through electronic expenditures) so that in the future a contributory welfare system can be further developed.

Submitted by A. D. Lee on

The Economist contrasted UCTs and CCTs in their article 'Pennies from Heaven' (26 Oct 2014). Your thought-provoking post reminds me of their analysi. They conclude, however, that "UCTs work better than almost anyone would have expected. They dent the stereotype of poor people as inherently feckless and ignorant. But CCTs are usually better still, especially when dealing with the root causes of poverty and, rather than just alleviating it, helping families escape it altogether". They also look at a radical form of UCT to poor households of Kenya and Uganda identified by satellite imagery with no involvement of government. More at:
www.economist.com/news/international/21588385-giving-money-directly-poor-people-works-surprisingly-well-it-cannot-deal
www.givedirectly.org

I just like the helpful information you provide for your articles.
I'll bookmark your blog and test again here regularly. I'm quite certain I'll learn many new stuff right here. Good luck for the following!

Submitted by Deepak Maheshwari on

India probably leads the world in terms of the sheer diversity and the number of schemes, projects and programs for extending some sort of social welfare or the other. However, the administrative cost and leakages are extremely high.

While the biometric ID Aadhaar with 600 million enrolments already can be an enabler in enhancing the efficiency of targeting, the unconditional transfers can make the whole process even more efficient as was seen in some field trials as well. See http://www.guystanding.com/files/documents/Jhabvala_Standing_EPW_2010.pdf

Submitted by Vivek on

I think this article risks confusing the benefits of UCTs at large with UCTs in the specific context (FCS). Perhaps a better approach would be to enquire whether UCTs can address the challenges inherent in FCS.

Two FCS specific challenges are security and divisions within the population. The 2011 World Development Report suggested sequencing different efforts and reforms over time - my takeaway from this is the vital important of security as an enabler for other efforts - security comes before "service delivery." I don't see how UCTs contribute to solving the security challenge.

Some of the OECD's Principles for Engaging in Fragile States speak to the second challenge- specifically, "avoid pockets of exclusion" and more generally the "context as the starting point" and "do no harm" principles. Again, my takeaway of the underlying idea is that it is very easy to make things worse by favouring (or appearing to favour) one side over another. It's not clear to me whether UCTs can help solve this challenge, and I have very strong doubts about the mechanisms that are suggested:
"Perhaps it is worth considering universal UCTs in specific geographic locations – for example, rural areas with a high proportion of the population below the poverty line - where the lower costs of universal transfers outweigh the benefits of more accurate targeting. There may also be a case for making transfers universal in fragile situations where any kind of targeting may exacerbate existing tensions."

Immediate thoughts about that quote:
1. How do you define "universal"? Are refugees (plus internally displaced persons and migrants) excluded? How do you stop people from moving around and generally defrauding the system?
2. How do you deliver this cash transfer service, sorry, enable individuals to 'buy' this service? (Only partly tongue-in-cheek)
3. Overcoming the challenges of 1 and 2, wouldn't a truly universal transfer just result in inflation (of assets/land if not goods)?
4. Would benefits be only transitory? Will there be any sustainable improvements? (Apart from setting up the cash transfer delivery mechanism ;-)

For the last point, looking at existing efforts (which aren't cash transfers, but similar) might shed some more light on this issue. For instance, the Afghan NSP impact evaluation found no lasting economic effects (there were apparently positive non-economic effects plus temporary effects from the cash injection into communities).

In conclusion, I'm inclined to think that UCTs provide a strong set of incentives for fraud which would undermine their effectiveness in many FCS. A truly universal UCT seems like a potential solution, but I suspect its effects will disappear through inflation of goods and asset prices.

Submitted by Laura on

Thanks Vivek for your comments. One of the arguments for UCTs in FCS is that it is a more direct way of reaching the poorest in these countries. Security and developing cohesion/mitigating divisive populations are two other very important goals in FCS, and UCTs are not directly asserted to address these goals. However, given the correlation between poverty and fragility, perhaps UCTs may help alleviate the causes of conflict.

Re. inflation - yes. it's worth further thought. Fundamentally it will depend on the laws of supply and demand - if there are big jumps in demand for particular goods or assets there could be price increases, but soon supply should respond and increase.
Also it's hard to imagine that given the likely size of transfers and heterogeneity in demand preferences, that such price changes would occur even in the short run, but for sure, before implementing an UCT, it would be very important to study consumption patterns and the flexibility of supply. Related to this, that is why I suggested possible "nudges", as well as understanding what recipients would use cash on. We might want to consider complementary activities to increase the chance that these cash transfers are not transitory but are invested and lead to long term increases in income

Submitted by John Edward on

Thank you for this post Laura. It's come at a perfect time as I am researching the efficacy of UCTs versus CCTs in FCSs for a masters thesis. Your links and references have been invaluable. Thanks a million for this brilliant synthesis.

Add new comment