Icon Our Work

Thailand’s Old Age Pension: an example of the importance of good quality analysis


Development Pathways

I recently come across a draft report by the World Bank on Thailand’s old age pension system. The report states it is for consultation, so I thought I’d take the opportunity to provide a few thoughts and helpfully point out a key problem with the analysis.

In 2009, Thailand implemented a large-scale expansion of its tax-financed “social” pension. While the scheme had previously been targeted at the poor, in 2009 it was extended to all those who were not receiving another form of government “formal” pension, thereby establishing universal pension coverage for all over-60s.

 Currently, the pension scheme provides beneficiaries with between 600THB and 1,000THB per month (*approximately US$19 to US$32), with transfers increasing by age. The programme is relatively low cost, at around 0.59% of GDP, which is much cheaper than most universal pensions in developing countries; for example, it is well below the levels of spending in countries such as Brazil, Bolivia, South Africa, Mauritius, Namibia and Lesotho, all of which invest more than 1% of GDP in their schemes (see Pension Watch for comprehensive information on the costs of pension programmes across the world).

 The paper does have some very useful analysis. It indicates that the inclusion of the poor in the programme has improved dramatically as a result of the move to universal coverage: around 90% of the poorest decile receive the programme, compared to less than 50% in 2008 (see the table below from the World Bank paper which shows coverage across the wealth deciles, with 1 being the poorest decile).

 WB-Thai-Grapholdimagefor Thailandsblog
The paper also indicates that the pension has had significant impacts on poverty. The poverty rate among the elderly fell by 23% between 2008 and 2009 – from 14.1% to 10.9% – and the World Bank estimates that the social pension contributed to 72% of this poverty reduction. In fact, among recipients of the pension, the poverty rate in 2010 fell by a massive 43% compared to 2009. The World Bank team concludes: “Social pensions appear to be a very effective tool for reducing poverty among the elderly in Thailand.”

 So, it’s a little schizophrenic of the World Bank to then recommend that the pension should, once more, be targeted at the poor. It appears to contradict the evidence that they themselves present on the relative ineffectiveness of the pension when it was poverty targeted, since more than half of poor older people were excluded. So, why does the World Bank make the recommendation to return to a pretty ineffective form of pension?

 Unsurprisingly, it appears that their proposal to target the pension is driven by ideology. The World Bank’s main concern is to reduce the costs of the Thai pension – which, as we’ve shown, is already low cost – and, implicitly, the tax burden on the better-off. It appears to be yet another example of Tea Party social protection.

 However, while we can understand and appreciate different ideological positions, we are concerned that the World Bank team used questionable analysis to make its case. They undertook simulations to assess the effectiveness of different targeting options, comparing the current universal targeting with poverty targeting. Yet, they assumed perfect targeting for the poverty targeted option, an entirely unrealistic assumption.

By using the assumption of perfect targeting, the World Bank was – unsurprisingly – able to demonstrate that the poverty-targeted option would be more “effective” than the universal option in eliminating the poverty gap in Thailand (in effect, they assumed universal targeting among the poor!). Yet, this result is meaningless given that it is common for poverty targeting to have very high exclusion errors, usually more than 50% (see here for an explanation). The World Bank appears to recognise this and, therefore, they recommended introducing proxy means test targeting. Yet this will not help since it is established that the proxy means test methodology has high errors and is relatively arbitrary in its selection of recipients (see here for further information).

In effect, the World Bank is employing unrealistic assumptions in its policy influencing in Thailand and, if successful, its advocacy could be very detrimental to the interests of millions of elderly Thais (and their families). Indeed, reducing the size of pension could also have negative impacts on the economy, as it would lead to a reduction in consumption, a key driver of economic growth (the Thai pension was made universal during the global recession to boost demand and stimulate the economy).

Debates on social security need to be taken forward on the basis of good quality analysis and evidence. Supporters of inclusive social security have no problem in admitting that the schemes they promote are more expensive than schemes targeted at the poor. But, neoliberals also need to accept that inclusive schemes will always be more effective in reaching and benefiting the poor, given that they have exclusion errors that are virtually zero: just compare the UK’s inclusive National Health Service with the USA’s much inferior poverty targeted Medicaid scheme. As we’ve pointed out before, debates on targeting are really just ideological debates: let’s not pretend that they’re anything else.

*Conversion rates taken from www.ft.com


Your email address will not be published. Required fields are marked *