Credit where credit is due. The World Bank recently published a report on the Philippines Pantawid Pamilyang Pilipino Programme scheme showing that, among young children excluded from the scheme, stunting rates increased by an average of 11 percentage points. In contrast, among beneficiary children aged 6 to 36 months, stunting fell by 9 percentage points. This is an extremely powerful finding demonstrating how poverty-targeting can unintentionally harm children. In fact, the deterioration in stunting among non-beneficiary children outweighed the benefits to recipients, suggesting a possible overall adverse impact of the programme on child stunting in the Philippines, writes Stephen Kidd.
The Pantawid scheme is a conditional cash transfer programme targeting the poorest families with children in the Philippines, reaching around 23 per cent of households. It selects beneficiaries using a proxy means test (PMT) and, while it is one of the best-targeted schemes found in low- and middle-income countries, around 45 per cent of eligible families are still excluded. It is among these families – the so-called ‘exclusion errors’ – that much of the harm is being caused.
The report’s authors – Deon Filmer, Jed Friedman, Eeshani Kandpal and Junko Onishi – speculate that the increase in stunting among non-beneficiary children may have been caused by increases in the price of perishable protein-rich foods – such as eggs, fish and chicken – of between 5 and 7 per cent. This was driven by beneficiary families using their higher incomes to purchase protein-rich food for their young children. However, there was no rise in the prices of rice and sugar.
The authors present good evidence that, due to the price rises, non-beneficiary families were obliged to substitute protein-rich foods with cereals, resulting in a less nutritious diet for their children. So, while the health of beneficiary children improved, it deteriorated among those children excluded from the programme. In fact, the negative impact on children was greater in those villages with a higher proportion of beneficiaries, possibly because more cash entered these villages, thereby resulting in even higher inflation.
There may, of course, have been other factors explaining the increase in stunting. For example, the authors note that the cost of some health services – in particular antenatal care – had increased in Pantawid villages as a result of the programme, thereby creating access barriers for non-beneficiaries. A lower availability of health care may, therefore, have also contributed to a worsening of nutritional outcomes.
The paper does, however, appear to contain at least one key error. The authors seem to believe in the effectiveness of the proxy means test and, as a result, assume that non-beneficiaries are consistently better-off than beneficiaries (since they all had higher PMT scores). Yet, in reality, due to the inaccuracies inherent in the PMT methodology – and, as noted above, exclusion errors in the Pantawid scheme are around 42 per cent – it is highly likely that many of the non-beneficiaries were just as poor as the beneficiaries. Therefore, not only did the PMT exclude families living in poverty, it almost caused significant harm to their children.
These results come hot on the heels of other research – again published by the World Bank – which has shown that the Pantawid scheme has generated an increase in child labour: some beneficiary children have been obliged to work so that they can pay the costs of attending school and avoid being sanctioned by the programme.
The World Bank’s paper throws yet another nail in the coffin of poverty-targeting. We now know that, not only does poverty-targeting usually exclude the majority of households living in poverty from social protection, it can also cause significant harm to their children. Benefits for ‘the poor’ truly are poor-quality benefits.
One recommendation from the researchers is for the Pantawid scheme to be offered on a ‘universal’ basis, at least in those villages where coverage is relatively high. In effect, they are moving towards recommending a universal child benefit which, in contrast to the current poverty-targeted scheme, would ensure that all children can access social protection while simultaneously protecting them against the risk of stunting. However, a much better approach would be for the universal child benefit to be offered nationwide and not just in sub-sets of villages. While such a scheme would cost more than the current system of poor relief, it would be much more effective. As I’ve discussed elsewhere, countries need to decide whether they want a Rolls-Royce or second-hand Lada national social protection system. If the Philippines wants the former, it will have to be willing to invest.
So, can we now expect the World Bank to listen to its own research and question its commitment to poverty-targeted social assistance which, as I’ve argued elsewhere, is anti-poor anyway? Furthermore, will the World Bank reverse its opposition to universal child benefits which, as we’ve recently seen, has hit children hard in Kyrgyzstan and Mongolia? Let’s hope so, but my advice, based on past experience, is: don’t hold your breath.