Proxy indicators: What are they and why do they matter?
What are proxy indicators and what purpose do they serve? Targeting is a critical aspect of finding the most-needy beneficiaries for poverty reduction programmes. Given the recent interest in revamping Sri Lanka’s largest poverty reduction program, a discussion on measures to reach the poorest of the poor is important. But how can poverty reduction programs such as Samurdhi be more effective through better targeting?
Cash transfers are an important component in poverty reduction programmes. Evidence from cash transfer programmes around the world, particularly in Latin America, suggests that if adequate in value and well targeted, such transfers can help lift the very poor out of poverty. But this leaves an important question: how do we identify and target the poorest of the poor?
The Proxy Means Test (PMT) has been used to improve targeting efficiency of cash transfer programmes in many countries, including Bangladesh, Rwanda and in Sri Lanka too. The PMT assesses eligibility on the basis of a score obtained with a weighted set of indicators which are easy to verify, such as household assets, housing conditions and demographic characteristics. What this means is that the PMT generates a score for the household based on easy to observe characteristics such as education, occupation of adult members, ownership of durable goods, and the location and quality of their dwelling. The indicators used to calculate the score and their weights are derived from statistical analysis of data from detailed household surveys.
These are verified by collecting information through visiting the house or by written verification by the potential beneficiary. Eligibility is then determined by comparing the household’s score against a predetermined cut off. So, one of the advantages of PMT is that it does not measure income, which can be very difficult in Sri Lanka given the size of the informal sector. Since the PMT uses more easily verifiable indicators of standard of living, it might be a more appropriate method of assessing eligibility for the beneficiaries of cash transfers. Another important aspect of PMT is the involvement of the community in selecting beneficiaries. The measure takes advantage of local knowledge on the identity of the most vulnerable households/ individuals in the village.
This allows us to identify ‘pockets of poverty’ at the Grama Niladhari and Divisional levels – the lowest administrative levels – which may not be reflected in national level poverty data. This is a particularly relevant point in the context of Sri Lanka’s post war recovery and economic growth strategies. Whilst official statistics on poverty levels record a steady decline in monetary terms, as is reflected in the decrease in the poverty headcount ratio to 6.7 per cent in 2012/2013 from 15.2 per cent in 2007/2008, (Household Income and Expenditure Survey), there are large geographical disparities in the poverty figures. For example, the poverty rates in the Northern districts of Mullaitivu (28.8 per cent) and Mannar (20.1 per cent), the Uva province’s Monaragala district (20.8 per cent) and the Eastern district of Batticaloa (19.4 per cent) all show a much higher poverty headcount ratio than the national average.
The poverty ratio was also higher in the estate sector (10.9 per cent), as compared with the non-estate rural areas (7.6 per cent) and the urban areas (2.1 per cent). While designed to benefit the poorest of the poor, in practice, the Samurdhi programme does suffer from poor targeting, reflected in benefits for many non-poor while at the same time excluding more than half of those in the poorest decile. Official data indicates that some of the richest households still receive Samurdhi benefits: 3.1 per cent of households in the richest decile and 5.8 per cent in the second richest, whilst only slightly less than half of households in the poorest decile (47.8 per cent) were Samurdhi recipients (HIES 2009/2010). Thus improving the selection methods of the programme, requires a thorough review of the beneficiary selection criteria.
PMT is a method worthy of consideration for this challenge. This is not to say that PMT as a targeting criterion has no drawbacks. Since the formula is only a prediction, there is a possibility for inaccuracies at the household level. These inaccuracies could be the result of the PMT being insensitive to quick changes in household welfare or availability of disposable funds. The formula and results may also seem arbitrary to some households and communities. However, well-targeted and adequately funded social transfers could be the key to raising the standard of living of the poorest and most vulnerable families. Programmes across the developing world have shown these transfers to improve living standards of both the poor and near-poor, and to help tackle the root causes of poverty and deprivation. With a more effective selection method, we can ensure that this programme provides the most benefit it can to those who need it most.
(This is the final instalment of a 4-part series by the Centre for Poverty Analysis (CEPA) looking at poverty-related policy interventions in Sri Lanka)