30 January 2013

Against cash transfers in South Sudan

A few quick thoughts on this recent article republished in the Sudan Tribune by the Auditor General of South Sudan, Steven W├Ându.

He makes 4 arguments against cash transfers

1. There is no clear definition of poverty in South Sudan
2. Targeting would be impossible
3. Delivery would be impossible
4. Cash transfers won't increase productivity

I'd suggest that the first two issues could be resolved by instituting a universal child benefit. No poverty targeting to worry about. Delivery systems could reasonably quickly be created by mobile phone companies if the government committed to start pushing serious money through this system. And finally that cash transfers surely will increase productivity down the road if they have an impact on child malnutrition, but that the productivity effects should not be our highest concern when there are people who are literally starving.

Update: I should credit Joe Hanlon with the child benefit idea (here and here), and also note the Blattman vs Banerjee debate on the subject from 2011.


Laura said...

The first one seems almost wilfully bizarre - surely when your poverty rate is as high as that in South Sudan your poverty line is 'most people', to the extent that it could well be cheaper to do a universal benefit, as the savings from excluding some people would not be offset by the cost of the targetting.

Michael Eddy said...

Three responses to the child cash transfer idea:

1. Birth spacing - A child cash transfer puts perverse incentives on having children rapidly, thereby creating negative health consequences for both the child and mother. Birth spacing in South Sudan is already dangerously low. This would only aggravate the problem. Implementing mechanisms to mitigate the adverse consequences may be prohibitively costly & difficulty to implement.

2. Family size - Another perverse incentive. Households are likely to be myopic and heavily discount the long-term costs. This is essentially a financial nudge toward larger family sizes. Is this desirable from a social welfare perspective? Doubtful. Again trying to monitor and mitigate this consequence is likely to be costly & largely incomplete.

3. Inflation: South Sudan has notoriously thin markets. Inflation has long been a problem. In a open economy with low transport costs, arbitrage would . But this hardly describes South Sudan. In Mexico (a place with far thicker markets than South Sudan), cash transfers were compared to in-kind transfers. Cash created inflation, further diluting the purchasing power of the HH. In-kind transfers lowered prices for the food items, creating more purchasing power for the HH. Together, in-kind transfers generated 11% more welfare for households than cash. Paper: http://www.povertyactionlab.org/publication/price-effects-cash-versus-kind-transfers

Of course, Child Cash Transfers can't be evaluated in isolation, but must be evaluated in terms of its redistributive efficiency in comparison to other social protection programs. Many have talked about community-driven development as another potential option, looking toward Indonesia, Afghanistan and even Sudan as potential models. CDD, of course, has its own limitations as well.

Michael Eddy said...

Three responses:

1. Family Size - The rational response for myopic households that heavily discount the future (and who wouldn't if you're near starvation), would be to over-produce children. Essentially, this is a strong financial nudge to have more children. Would this be optimal from a welfare standpoint? Doubtful. Monitoring and implementing mechanisms to mitigate this risk would likely be costly and easily circumvented.

2. Birth Spacing - Limited time between births is already a serious problem in South Sudan with significant health consequences for the mother and child. This would be a perverse incentive that would only aggravate existing problems. Again, monitoring and implementing mechanisms to mitigate this risk could be difficult.

3. Inflation - South Sudan is characterized by notoriously thin markets and poor arbitrage. Injecting a cash transfer has the potential to affect local markets by quickly increasing demand, in an environment where the supply response by traders will likely be slow to catch up. In Mexico, a J-PAL study shows that cash-transfers actually resulted in 11% less welfare impact than a similarly sized in-kind transfer because of price effects. South Sudan has far thinner markets than Rural Mexico. ( http://www.povertyactionlab.org/evaluation/price-effects-cash-versus-kind-transfers )

Of course the re-distributive efficiency and welfare impact of a child cash transfer cannot be assessed in isolation, but only alongside other social protection programs. Another option, also being discussed in South Sudan is a community driven development model based on successful examples from places like Afghanistan, Indonesia and even South Kordofan. This, of course, comes with its own host of limitations. Nevertheless, proponents of the child cash transfer should reasonably discuss how they will address the above three unintended consequences.

rovingbandit said...

Thanks Michael, good points.

On the first two I don't have a strong rebuttal, except to say that this is an empirical matter, and the effect may well be small, especially if you already have a high fertility rate.

On the last one, I'm sceptical - the summary opens with the line - "The cash transfer program caused prices to increase by 3.9 percent, although this effect is not statistically significant". So a big welfare impact driven by a statistically insignificant change?

Yes markets are thin in South Sudan, but we have to also consider what is the more efficient way of transporting food and other basic items around South Sudan - market traders or the UN and NGOs? I would bet any day that market traders can transport goods much much cheaper, and so comparing an in-kind vs cash transfer, I would expect the overhead costs to be much lower for the cash transfer, meaning you could push those savings to the recipients, to compensate for price changes. Finally, I think that any price effect would be temporary and be smoothed out as suppliers adjust to the increased demand.

Finally, the JPAL summary doesn't mention the welfare loss to producers of having lower prices for their goods.

Do you have any good links for those CDD programmes? I'm pretty sceptical of those, and the RCTs I've seen (DRC, Sierra Leone) have not been glowing.


Anonymous said...

Nice respond in return of this difficulty with solid arguments and telling the whole
thing concerning that.

Here is my blog :: atlaslm.com

Post a Comment