The ethics of cash transfer targeting

[Update as of 16 November 2015: this post has been edited to accurately reflect GiveDirectly’s role in the study.  They did not design the randomization scheme and had not seen the results before the paper was released.  Paul Niehaus from GiveDirectly has also informed me that future research with their clients will no longer involve randomization at the household level, and that they are offering cash transfers to the comparison group from the earlier study.]

Is it ethical to give cash transfers to some poor people while their equally poor neighbors get nothing?  Johannes Haushofer, James Reisinger and Jeremy Shapiro just released a new study of this program design with GiveDirectly.  They found that people who received nothing were less happy than they’d been before the program started.  Anke Hoeffler has taken them to task for this, essentially arguing that these negative effects are so predictable that it’s unethical to study them, no matter how clever the research design.

As a basic point of research ethics, I agree that a study should never be designed to knowingly decrease participants’ wellbeing.  However, this wasn’t the goal of this project.  The new paper builds on earlier work by Haushofer & Shapiro (2013), which randomized access to cash transfers at the household level in order to examine both direct effects (on recipient households) and spillover effects (on their non-recipient neighbors).  Ex ante, the expectation was that the  neighbors might still benefit from the program, as recipient households shared resources through informal insurance networks.  While the prospect that the program would create jealousy or unhappiness among the neighbors might seem obvious in one sense, other studies have indicated that transparent eligibility criteria for cash transfers can mitigate unhappiness among non-recipients, so it wasn’t clear that household-level randomization would make non-recipients either economically or psychologically worse off.  Clinical equipose still applied.

As it turned out, the 2013 study found that the spillover group didn’t see any economic benefit from the program.  This might explain part of the 2015 findings – people might be unhappy not only because their neighbors received cash grants, but also because they failed to share their resources.    Finding a negative outcome doesn’t make the research design inherently unethical, but researchers now have an additional datapoint to consider when thinking about how future studies of cash transfer programs might be designed.

Given this new evidence, could there ever be a good reason to distribute cash transfers to fewer than 100% of eligible poor people in a town (as a long-term program design rather than a short-term research project)? It’s not clear to me that there is, although not for the reasons that Hoeffler points out.  Most countries haven’t got the funding to offer a basic income guarantee to all citizens, so distributional questions are an important aspect of program design.  Say a program only has the funding to provide a certain level of grants to 50% of eligible people.  Should the grants be allocated to all the poor people in 50% of towns to avoid conflict among neighbors, perhaps at the cost of conflict between towns (c.f. Bates 1974)?  Or might it be preferable to reduce the value of the grants by 50% and provide them to all eligible people in each town, with the possibility that this makes the amount too low to make a meaningful difference?  Note that this latter solution is a formal version of what Haushofer & Shapiro believed might happen anyway – if recipient households are assumed to share their large grants with their neighbors, it’s simply a less efficient way to give everyone a small grant.  It would be fascinating to see a study comparing these two targeting schemes.

Cash transfer programs aren’t just for low income countries

Eduardo Porter had a fantastic article in the NYT last week about the myth that welfare programs make their recipients lazy and entitled.  He highlights recent research from a team of MIT and Harvard economists which finds that cash transfer programs in low income countries don’t discourage people from working, and connects this to other studies which find the same result for American welfare programs.  In particular, most of the ostensible success of the 1990s welfare reforms were attributable to the strong economy, and poverty increased again with the recessions of the 2000s.  Meanwhile, pushing people off welfare probably led to worse outcomes for children who grew up in poverty.

If the evidence base for cash transfers in low income countries is so strong, should we expect to see the same effects in high income countries?  My prior on this is that we should, and there seems to be an increasing amount of evidence supporting this position.  Aside from the study that Porter mentions on the Mothers’ Pension Program, which took place in the early 20th century, I’ve found two relatively more recent studies that evaluate the use of cash transfers in North America. One looks at a town in Manitoba where poor residents were given basic income grants for four years in the late 1970s.  People with no other sources of income were given grants up to 60% of the poverty line, and people with some outside income received smaller grants on a sliding scale (the precise value is not specified in the study).  Evelyn Forget analyzed administrative data from the town some years later, and found that grant recipients experienced a range of benefits.   They were less likely to be hospitalized for work-related injuries, car accidents, domestic abuse, or mental illness.  Children’s test scores increased, even as their dropout rates decreased, and more adults went back for continuing education.  While there was a small decrease in hours worked, this mostly came from mothers of young babies and teenagers, who are arguably investing in other types of human capital by raising children or staying in school.

The second study tracks a group of children in North Carolina who were members of the Eastern Band of Cherokee Indians, and whose families began receiving an extra US$4000 per capita each year after a casino was build on their land in the mid-1990s.  Researchers found that the grants lowered rates of behavioral and economic problems among treated children, and improved their relationships with their parents.  It also increased personality traits that are correlated with financial success later in life, like conscientiousness and agreeableness.  (The researchers don’t discuss the grants’ impact on children’s incomes or educational achievement in this paper, although I assume they’ll do in future work if they have the data.)

Do cash transfer programmes foster dependency?

At the FAO’s blog, Ben Davis argues that they don’t.  Some key quotes, drawing on recent research on several African projects:

Along with the increase in productive activities, the cash transfers programmes have given households more flexibility with their time. In most countries of sub-Saharan Africa, low paying casual agricultural wage labour is an activity of last resort, when households are desperate for cash.  In Zambia, women in beneficiary households reduced their participation in agricultural wage labour by 17-percentage points and 12 fewer days a year. Both men and women increased the time they spent on family agricultural and non-agricultural businesses. … As one elderly beneficiary said, “I used to be a slave to ganyu (low-paid wage labour) but now I am free.”

Cash transfer programmes also have allowed beneficiary households to better manage risk. Fieldwork in Kenya, Ghana, Lesotho, Zimbabwe, Ethiopia and Malawi has found that the programmes increased social capital and allowed beneficiaries to ’re-enter’ existing social networks and/or to strengthen informal social protection systems and risk-sharing arrangements.

One of the important points that I take from this is that the idea of “welfare dependency” as transfer-induced withdrawal from the labor market is overly simplistic.  Selecting a livelihood strategy rarely comes down to a binary decision to work or depend on the state, be it in rural sub-countries in Kenya or poor neighborhoods in Chicago.  Instead, people  choose from a portfolio of livelihood options, often combining various sources of income at the same time.  These might include agricultural production, self-employment, waged labor, salaried labor, support from family or friends, and support from the government.  Davis’ point suggests that state-provided transfers don’t substitute for all the other livelihood choices here, but rather give people enough of a buffer that they don’t have to resort to the most poorly paid or abusive options quite so often.

Do politicians substitute cash transfers for other public goods?

The results from Twaweza’s latest Sauti za Wananchi poll in Tanzania are out, and they include some interesting questions about public support for cash transfers.  There’s a good write-up at the CGD blog.  In short, people were less supportive of cash transfers than one might have expected – and the more they learned about the transfers, the more likely they were to say that they would prefer the government to spend money on other public goods.  This is all the more surprising given that the transfers provided through the Tanzania Social Action Fund have been found to have a wide range of positive social impacts.

I came across similar types of skepticism when I was speaking with people in Ghana about the LEAP program this summer.  These were informal conversations with friends and casual acquaintances, so obviously not representative of Ghanaian public opinion generally, but they still had an interesting range of variation.  A number of people voiced the standard objection that cash transfers would make recipients lazy and entitled.  When I pointed out that research in other low income countries has shown that this isn’t generally the case, they often suggested that Ghana might be the exception.  (Regardless of one’s priors on this matter, though, I’m fairly sure that receiving US$10 – 20 a month isn’t encouraging many people to drop out of the labor market.)

One reaction that I hadn’t expected was what I’ve come to think of as the public goods critique, similar to the Twaweza results.  Several people agreed that cash transfers might be useful, but suggested that this could lead to lower investment in other types of public goods.  Tony Hall at LSE has argued this explicitly in the case of Brazil.  The underlying concern here seemed to be that the current donor enthusiasm for cash transfers would give governments a way of ducking their responsibilities to provide public goods – giving poor citizens small amounts of money before sending them off to fend for themselves.  It’s certainly true that cash transfers are administratively much simpler than maintaining functioning public education or healthcare systems.  A promising area for future research, I think.

The persistence of extreme poverty in the US

Vox had a thought-provoking article last week on people in the US who live on less than $2 per person per day.  It’s adapted from the book of the same title by Kathryn Edin & Luke Schaefer.  They find that millions of Americans (including up to three million children) have months where they live on next to nothing, frequently going hungry and facing homelessness as a result.  The entire article is essential reading.  Here are some of the main points:

  • There’s no clear demographic pattern among people who face extreme poverty.  As Edin says, “It’s racially and ethnically diverse, it’s regionally diverse. You see both married and unmarried couples in this situation.”  However, poverty tends to be worse in rural areas and in the South, where fewer services are available.
  • People want to work, but find it difficult to hold down jobs.  Edin notes that “almost all of these households actually do have workers… You still see these pretty lengthy spells in extreme poverty, but these people are in and out of the low-wage labor market. Seventy percent of them have had a worker in the low-wage labor market in the past year.”  Schaefer adds that “it’s very hard to find a job. The unemployment rate has been very high for low-educated workers for a long time. These folks are at the back of the line.”
  • Social services and family support networks rarely help.  Many eligible households either don’t apply for TANF, or (in some cases) have been told mistakenly that they’re not eligible.  Only a million people in the entire country receive TANF at present, although 15% of the population or 45 million people live below the poverty line.  In addition, few people seem to receive much assistance from their families.
  • People come up with creative ways to access even small amounts of money if they can’t work.  Selling plasma is one of the most common, followed by cashing out food stamps (which cuts the value of the stamps by about half), collecting scrap metal for redemption, and doing sex work.  Selling sex can be a way to access housing or food as well as cash.
  • And my own takeaway: While the availability of formal employment is different, overall this is quite similar to what extreme poverty looks like in countries around the world.  People find various ways of making claims on others in order to access food, shelter, and clothing, or the cash to buy the same.

Perhaps what’s most striking in this context is that, while we do have a wide range of social safety nets, none of them are designed to address this type of poverty. Nor has half a century of prolonged economic growth done much to reduce it.  I came away from this article thinking that it’s one of the strongest claims I’ve yet seen for the value of a universal basic income grant.

Going the last mile in ending extreme poverty

Brookings had an excellent blog series last month on going the last mile to end extreme poverty.  The posts were adapted from a book of the same name which was just released.  I was most struck by Raj Desai’s article on the role that welfare programs played in ending extreme poverty in today’s high income countries.  As he pointed out, when the US, UK, and Germany adopted major welfare programs in the late 1800s and early 1900s, they had the same GDP per capita (in real terms) as India, Indonesia and China do today.  Welfare subsequently played a major role in eliminating extreme poverty in these countries.  So why haven’t today’s middle income countries done the same?

His answer is worth quoting at length.

The earliest social protection programs in Western Europe were of the contributory variety—financed out of taxes on wages—as a way of preventing social conflict. Otto von Bismarck’s pension, sickness insurance, and worker compensation programs were, after all, created to pre-empt a Social Democratic victory in Germany. These systems, combined with the political changes taking place in the continent, would lead to dramatic expansions in social protection in later decades. Even as industrialization initially caused poverty, it also created rising wages for workers. Eventually, organized working classes formed a strong alliance with the fast-growing, urban middle class. This political coalition sought policies that would protect itself from economic cycles—especially after the economic turmoil of the 1920s and 1930s—that would eventually result in the postwar welfare system. Indeed, the durability of these welfare programs may be attributed to the participation of the middle class, which shaped program design: welfare programs provided universal coverage so that the middle class itself was not excluded from benefits.

Many of today’s developing countries have followed a very different path. Labor tends to be less organized and have weaker relative bargaining strength. Much of the labor force remains in the informal sector and is left out of any contributory schemes, which tend to have limited scope. Social protection is more reliant on a fragmented system consisting of a large number of non-contributory programs financed out of general revenues. More importantly, the preferences of both governments and donors are mainly for programs that target particular sub-populations to achieve cost efficiency.

Consequently, an opposite political dynamic appears to be playing out in developing countries. Instead of middle class “buy-in” resulting in broader and more comprehensive programs, targeted and fragmented programs are inhibiting median-voter support for social protection and leading to middle-class exit. The consequences are familiar to designers of targeted social protection—their vulnerability to shifts in political winds, their susceptibility to abuse or capture by elites, and their occasional failure to outlive the aid programs that may finance them. The overall result is that the demand for comprehensive welfare programs in middle-income countries remains weak.

One of the best succinct descriptions of the political economy of social protection that I’ve yet come across, and an interesting consideration thinking about novel social protection delivery mechanisms like GiveDirectly’s potential partnerships with regional governments in Kenya.