Some nuances of conditional cash transfers

Professor Clive Thomas stimulated much discussion last week when he proposed the idea of a US$5000 per year cash transfer to Guyanese families. The fact that the proposal is open ended stimulated a lot of heated discussions on social media. Several claims were made: these payments will stimulate the money supply and therefore engender inflation; many would become welfare “parasites” and so on.  I am sure Professor Thomas is well aware of the vast literature on conditional cash payments that exists in the field of development economics. The old sage is also cognizant of the acute income and wealth inequality in Guyana. He sees these transfers as perhaps the only direct mechanism through which the most vulnerable people will share in the oil revenues that will go directly to central government. There will not be broad-based employment in the oil and gas sector any time soon.

Guyana’s present per capita GDP (a proxy for average income) is just around US$4500. After 2020 the per capita GDP will see a level shift as the average will rise to a new higher steady-state by virtue of the fact the economy has the new oil and gas sector. However, the volatility of the price of oil will cause much fluctuations in economic growth around the higher average income. This means there will still not be an abundance of financial resources. The amount of US$5000 per year seems reasonable, but what is most important is the number of families who will receive the cash transfer.

Guyana has approximately 210, 000 households. It is highly unlikely that every single family can receive a cash transfer of US$5000 before 2025. That would require about US$1.1 billion a year. It is unlikely the government will realize this amount of oil revenues by 2023. However, even if Guyana earns an average sustained billion dollar (plus) revenues after 2025, when ExxonMobil significantly increases production, this is still a significant amount of aggregate transfers given the demands for public services and infrastructure. Moreover, given the volatile nature of the expected revenues, a certain pay out of US$1.1 billion per year will bring to the forefront the problem of fiscal sustainability. It also raises the issue of fairness. Should a rich and poor family alike get an equal amount of money?

However, what if about 10,000 of the poorest families in Guyana are identified for an annual cash transfer of US$5000? There will have to be an agreed upon cut off level of income that indicates deep poverty. The threshold level raises a major moral conundrum, however. Let’s say we agree that a family living on US$5 per day is in poverty. But what if we increase the poverty cut off level to US$6 per day for a family and found that the number of families in poverty doubles? What justification can we bring to say that US$5 and not US$6 per day is the threshold for receiving the cash transfers?

For now, let us assume we can identify 10,000 poorest families in Guyana where personal income data is scarce. The data scarcity makes the policy susceptible to vote buying and patronage games by politicians. The last serious study of poverty levels in Guyana was done in 2006. It might require a major administrative cost to just identify the people and monitor the transfers. It is for this reason, the few countries which offer cash transfers provide a small flat payment to all citizens. These payouts are usually small and cannot address adequately the inequality and poverty problems.

It would cost US$50 million per year for cash payments to 10,000 families in extreme poverty. This amounts to about 1% of expected 2023 GDP. With respect to projected 2023 oil revenues, it amounts to just over 10%, which is still quite high. Presently, the Guyana government spends a paltry average of 4% of GDP on education. Once an agreement is reached, these payments are a sure liability for government, but the oil revenue supporting them is susceptible to wild swings as historical prices suggest. If excessive commitments are made – such as targeting many more than 10,000 families for electoral purposes – the country may have to borrow in some periods to meet them. I think everyone agrees that borrowing for conditional cash payments is a bad policy. 

Once the poorest families are identified, the money must go towards the female head of household. There is an overall tendency for women to do a much better job at managing the family funds. It is simply not true the poor women focus on mainly the short-term (have a high time preference). They tend to have higher saving propensity, whether in commercial banks or box hand. This is well known to those of us who grew up in rural villages. The funds must be deposited at a local bank in the woman’s name. Poorer men tend to have higher time preference, which is a fancy term used by economists. It simply means the man is much more likely to partake in short-term social ills and self-serving consumption instead of investing in longer term family needs. There are, of course, deviations from the general trend, but on average women do a better job in managing family finances.

Would the policy of cash payments incentivize people to permanently depend on government handouts? I doubt it. There is no conclusive evidence this is the case. The most I have seen are a few anecdotes about remittances causing people to throw back and lime instead of seeking work. Some studies have found that people in the Caribbean use remittances for home improvements, education and consumption. There is opportunity here to conduct a study on whether remittances reduce people’s willingness to work. Remittances provide what economists call a natural experiment, since we cannot have controlled lab experiments in social and economic life.

However, a sizable number of development economists use randomized controlled trials (RCTs) to evaluate various micro-based policy proposals. The idea is to randomly assign the poor into control and treatment groups. The method has been popular these days, but it also faces serious criticisms. I also believe it raises some serious moral dilemmas. The RCT method is based on the notion that the individual’s outcome in life is a matter of personal choices. This is the epistemological notion of micro-foundation in mainstream economics. There are however many downward causations, whereby the social and macro determine individual behaviour as much as the choices people make. For example, we can develop a very good randomized trial of cash transfers, but the group-level strategic voting still ultimately determines the overall political economy of the country. Essentially, then, not even RCTs can hold all other factors constant in economic life.

Nevertheless, I agree the methodology could be helpful in evaluating some short-term policy effectiveness.  It should not however be generalized for the very long term since structural economic change will change individual behaviour. Also, we should ask for how long privileged academics like myself can randomly assign the poor into treatment and control groups?

Mr Nigel Hughes raised a central caveat in the discussion. Would Guyanese politicians use conditional cash transfers for electioneering and vote buying? The reader can decide, but in recent memory much campaigning was pinned around the school voucher programme.



Around the Web