The Multidimensional Poverty Index: Another underestimate of urban poverty

Yet another global study has understated the scale and depth of urban poverty, by failing to appreciate the differences between rural and urban contexts.

18 June 2014
The Multidimensional Poverty Index only has one indicator to assess housing quality – whether floors are made of dirt, sand or dung. So these houses in Manila would not be classified as sub-standard (Photo: Mark Edwards/IIED)

The Multidimensional Poverty Index only has one indicator to assess housing quality – whether floors are made of dirt, sand or dung. So these houses in Manila would not be classified as sub-standard (Photo: Mark Edwards/IIED)

The Global Multidimensional Poverty Index 2014 (MPI), launched on 16 June, says 85 per cent of people suffering from 'multidimensional poverty' in low- and middle-income nations live in rural areas. But it reaches this conclusion by failing to measure real deprivation in living conditions for urban populations. The key problem is that the study uses the same set of indicators in rural and urban areas. This is absurd, as the following examples show.

You can be poor without having a dirt floor

One of the study's indicators of deprivation is that a households has a floor made of dirt, sand or dung. This is the only indicator used for housing conditions. But in the informal settlements in cities of Africa, Asia and Latin America, many buildings have two or more floors. Clearly the households on the upper floors cannot have dirt floors, but they can still be poor. 

But use this criterion to assess poverty in urban areas, and you will score many households as non-deprived, despite the fact they live in poor quality dwellings and lack space, both inside and out. Nor does the MPI consider other deprivations that contribute to urban poverty, such as the constant risk of eviction and the lack of the rule of law.

Poverty viewed from an urban perspective could include indicators such as no room on the house plot to dig another pit for the latrine and no access to land to grow crops or graze livestock. It should include illegal occupation of the house plot and land sites at high risk from extreme weather (as these are often the only sites low-income groups can occupy in urban areas). These would score urban deprivation above rural deprivation.

Having a radio does not mean you're not deprived

Another of the study's indicators is a list of assets that includes telephones, fridges and radios. A household is scored as deprived if it has no more than one item on the list. This measure of wealth may work for rural populations but it ignores the range of extra costs urban populations face on a daily basis. 

People in urban centres generally have to pay for water, fuel and food. Many also must pay to use a toilet and washing facilities. This is less the case in rural areas. In cities, many low-income groups buy, build or rent housing on the city periphery because land is cheaper. But this also means time-consuming and expensive journeys to work or to access schools and other services. 

It is also likely that, compared to rural people, a far higher proportion of low-income urban dwellers have to pay rent for their accommodation. The validity of using the 'assets' listed as indicators of deprivation is not clear; having a radio is hardly an indicator that a household is not deprived.

To oversimplify, the main issue here for much of the rural population is distance to services whereas for much of the urban poor population it is about access to services. There are health care services and schools nearby but the urban poor do not have the right nor the money to use them.

What's 'improved' in rural areas may be inadequate in urban

Another of the study's indicators of deprivation is that a household does not have 'improved' provision for water and sanitation. Again, context is key.

Urban centres generally have much larger populations and higher densities than rural populations, and so need different technologies to reduce health risks in provision for water, toilets, and sewage. The importance of household solid and liquid waste collection also increases in urban areas, as there is nowhere close by to dispose of household waste. 

So it doesn't make sense to use the same definitions for what constitutes good or 'better' provision. Pit latrines that can work fine in rural contexts are usually problematic in urban areas because they cannot be emptied (especially in large settlements with no roads) or there is no space to dig a new pit. For water, a shared standpipe or protected well may work well in rural contexts when relatively few people use it – but not when several thousand have to share it, as in many informal urban settlements. Many households have rural and urban components to their livelihoods or incomes. 

We need better LOCAL data

What is so much needed is better local data for rural and urban populations. The examples above show that it doesn't make sense to use the same set of indicators for all aspects of rural and urban deprivation. To address deprivation in both rural and urban areas, local governments need data on exactly who within their jurisdiction lacks good quality and safe provision for water, sanitation and basic services and where they live.

This requires a very considerable change to the whole data collection system. The national sample surveys of the kind from which the Multidimensional Poverty Index is constructed do not serve this purpose well.

One way forward would be for census authorities to share their data with local governments in a format that helps them to identify who faces what deprivations. Censuses can be far more useful to local governments as they should provide data for each neighbourhood and building and so provides the level of detail needed for local decisions, which national sample surveys do not. The disadvantage of censuses is that they are expensive and usually held only once every 10 years but local governments can set up systems that keep key statistics updated between censuses.

Policymakers and researchers repeatedly underestimate urban poverty. Our efforts to eradicate poverty are little served by another international dataset that simply draws on the (inadequate and often inaccurate) data available from national sample surveys.  

David Satterthwaite (david.satterthwaite@iied.org) is a senior fellow in IIED's Human Settlements Group.

For more detail on the scale of urban poverty and of the effectiveness of measures to reduce it, see two books by Diana Mitlin and David Satterthwaite, both published by Routledge.