Syndicate content

Confusing a treatment for a cure

Berk Ozler's picture
A treatment is an instance of treating someone, say, medically. A cure ends a problem. Sometimes, the treatment is a cure. Other times, it just keeps the problem under control without curing it: if you remove the treatment, the problem comes back…
Our dog Dizzy has cancer: an MRI showed multiple tumors in his brain. He also needs to have some of his molars removed, as severe dental disease has rendered the simple act of eating painful. But, he cannot go under anesthesia because the toxicity of the cancer drugs have left his liver in a very delicate state, making even quick dental surgery risky.
So, we give him antibiotics and painkillers. These drugs clear things up very quickly. After a week, we stop the drugs: within 3-4 days, the pain is back, he is not eating, and we’re back on the drugs. The drugs are palliative and the vet does not like him to be on them continuously. Even if they don’t cause any long-term damage (which is not obvious), they will not remove the underlying cause. He needs surgery, which is the cure to that particular problem. What he has is a treatment – really, a patch – until, hopefully, his liver function returns to normal so that he might have the dental surgery (and be eligible again for chemotherapy).
What should we think when we read headlines like “More evidence that giving poor people money is a great cure for poverty”? To me, it should indicate that the recipients of the cash transfers escape poverty and stay that way barring significant negative shocks. Instead, what it increasingly seems to mean is that poor people don’t use the money to drink or smoke. Setting aside the question of ‘so, what if they did?’ (which was asked by Jishnu Das here), is the bar we want to set for cash transfers to poor people really that low?
Look, I get that there are many people who hold these views that are ill-informed, outdated, or worse, prejudiced: that cash in the hands of poor people will be wasted. It is also more than unfortunate that such views affect charitable-giving, donor behavior, and policy. GiveDirectly has done a great service trying to debunk this myth, channeling your donations to households in Kenya and Uganda to spend how they themselves see fit. The new Evans and Popova paper is also a contribution that is long overdue. One would hope that all of this evidence would be sufficient to put this issue to bed. But, if the migration debate is any indication, the critics will just come up with other reasons to remain defiantly opposed.
In the meantime, we will do an injustice to poor people everywhere and to ourselves if we are satisfied with cash transfer programs as our best tool to fight poverty. Cash transfers may well be the best item in our social protection toolkit, but they remain unproven as a policy for promotion...That poor people are less hungry, have more assets, consume more, and are happier are almost tautological while they are the recipients of large cash transfers. An important question is whether they’re back to square one when the benefits are removed. (I abstract here from the ‘basic income support’ debate, which may well be a sensible policy for countries that can afford and implement it: this is not, however, what advocates of cash transfers are referring to…)
For people to escape poverty as a result of any policy, they have to build some capital that would change the trajectory of their future earnings: this could be physical capital, human capital, or social capital. If transferring cash alone is not altering the accumulation of at least one type of capital for beneficiary households/communities, then poverty reduction is unlikely. The evidence that simple cash transfers increase earnings (and consequently consumption, health, nutritional status, etc.) in the long run is currently scant. David McKenzie blogged here about the five-year effects cash grants to small business owners in Sri Lanka, showing that giving a man a fish seemed to have long term effects on business survival and profits, but giving a woman a fish did not. Furthermore, they found no significant effects on consumption, health, or education as a result of these improvements (see comments in the blog linked, with the caveat that statistical power was low). When they did see consumption effects, this time in Ghana, it was because the grants were used towards consumption, not investment.
In fact, it’d be perfectly consistent with forward-looking investment behavior to see consumption declines in the short-run in response to large cash grants. If the grant is enough to make a lumpy investment but additional funds are required to cover running costs, households may liquidate other assets and savings, go into debt, and decrease consumption in the short-run. That’s why I don’t get too excited when I see the increases in HH assets and consumption in the beautifully designed GiveDirectly study by Haushofer and Shapiro, because the follow-up is within 12-18 months of the large transfers: it’d be hard to make transfers that large disappear in that kind of timeframe.
One study that did show uniformly large gains after four years is the study of NUSAF’s Youth Opportunities Program by Chris Blattman, Nathan Fiala, and Sebastian Martinez, which Chris blogged about here. Setting aside the fact that the intervention in Northern Uganda is not a simple cash transfer program, the authors themselves argue that the apparent success of this program could be due to the fortunate collision of several favorable factors (decently functioning markets, a fairly educated and young target group, etc.). I am part of a team studying more or less the same intervention in a nearby country for a different target population that faces tougher odds and, not surprisingly, our preliminary findings paint a gloomy picture (you’ll have to wait for the paper for a little while longer).
[While Chris Blattman has been guilty of writing blog posts, such as “Dear governments: Want to help the poor and transform your economy? Give people cash” and “End poverty by giving the poor cash?”, he has also written a post titled “Why cash transfers to the poor are not the next big thing,” in which he has pretty much outlined what I said about cash transfers above: (i) the bar we’re setting for these programs is not very high; (ii) if your only problem is a credit constraint, they can be great; and (iii) they are mostly palliative. So, I repeat his arguments here just in case his inconsistent messages have somehow confused the audience. And because others are carrying the symbolic flag in traditional and social media even after he has recently signed off the topic here.]
And, the lack of evidence of long-term effects does not stop with poverty reduction. In our Malawi cash transfer experiment, we have shown significant improvements in various outcomes among adolescent girls and young women during and immediately after the cash transfer program (see here and here for two examples). But, ongoing analysis of longer-term data (two years after the program ended or approximately five years after baseline) mostly indicates convergence between the treatment and control groups (again, working paper coming later this year). Like Dizzy’s achy tooth, the pain comes back soon after the treatment is removed…
So, here is the challenge I have in mind, which sets the bar at the other end of the spectrum. Let people come forward with their best ideas forward for sustainable, long-term poverty reduction. These can be ultra-poor poverty programs; simple cash transfer programs; training programs; combinations thereof, etc. They all get the same amount of funds and put their best foot forward. The most promising ideas (according to our latest knowledge) are entered into the challenge and all get independently evaluated for their effectiveness per dollar spent. The final outcome is poverty reduction at the community level after at least five years – ideally longer. We can agree on how to measure poverty (money-metric, multi-dimensional, etc.). We should have done this with the Millennium Villages Project, not against a counterfactual of no intervention, but compared to our best alternatives. That ship has sailed, but we can still do this one.


Submitted by D A Phillips on

Sorry folks but this is all too obvious! One major issue with cash transfer programs is whether and how they fit into the design of even the most basic social safety net designed by the Government of the country. In this context is the size of the transfer consistent with the funds available for a safety net now or in the medium term future to the Government from taxation?? If the answer is NO then the cash transfer program is unsustainable and will also undermine Governmental capacity building. These programs are justifiable as emergency assistance - they have little to do with economic development. Do them to help the poor but don't do them as development assistance. (see my book "Development Without Aid" )Anthem Press 2013)

Submitted by Berk Ozler on
I think many would disagree that cash transfers have little to do with economic development. Countless studies show that they reach both of their goals of reducing current poverty and increasing investments in the human capital of children. This post was about whether they help poor households' ability to stay out of poverty if/when the transfers are removed...

Submitted by D A Phillips on

Berk - OK I don't want to demean the research - but look - if CTs have no sustainable effect on HH poverty then presumably they cannot sustain investment in human capital either. But my point was about the consistency of external (CT) interventions with Government efforts to develop their own sustainable safety nets.

Submitted by Berk Ozler on
That doesn't sound correct: accumulating human capital among children through schooling years has little to do with why CTs don't increase earnings of the parents.

Add new comment