Home / Opinion / Spatial clues to track poverty

Indian academia, media and policymakers have been involved in a long drawn debate on poverty for many years. In recent times, this debate has degenerated into a pointless argument of who can be identified as poor. The very concept of poverty is ambiguous—poor for one person may not be so for another. More importantly, simply defining poverty need not mean that the poor person or household can be identified and targeted well. All this is well known, but the lack of data and analytical abilities have further exacerbated the problem. A surfeit of poverty lines have been created by those at the World Bank, Planning Commission, Suresh Tendulkar Committee, and now, the Rangarajan Committee. All of them make sense from some perspective, but each is flawed in its own way.

Now a very different method of addressing poverty is being proposed. To do so, we change the question. Instead of asking who are the poor, we ask where they are. What we propose is a new way of identifying and addressing poverty. Imagine rather than targeting poor households, India started to target poor areas. Issues of leakages would become less important, one would not need to wait for reliable identification mechanisms, nor would we need to come to a consensus on who is poor and who is not. We will just have to find the areas.

All of this is possible only if we have a precise method of identifying poor areas. Government surveys, such as those conducted by the National Sample Survey Organisation, by their very nature can only identify poverty levels at the state or regional level. India needs to identify poverty at a much lower level of aggregation. There have been attempts to estimate poverty at the district and even block levels. But these are not enough for the task at hand. We need to get even closer—down to the village or habitation level—if targeting is to work well.

Is this even possible? No doubt such surveys would be extremely costly, but the bigger problem is they would be fraught with errors as their scale would need to be similar to the census. Perhaps it is scale that has deterred innovation in poverty measurement in many years. But there is another reason—the refusal of economists to talk to geographers.

Just as poverty has an economic dimension in terms of low incomes and expenditures, it also has a spatial dimension. The poor tend to use less space—both in their place of residence and work. They also tend to use less infrastructure that those who are better off. These locational and spatial clues are difficult to measure the way economists measure their own evidences. But it is possible to measure spatial poverty at a much lesser cost and on a larger scale.

Through the National Remote Sensing Centre the government puts out high resolution imagery of India in the public domain. The Geographical Information System (GIS) services provides a host of images on roads, habitations and so on. The US government’s National Oceanic and Atmospheric Administration has been providing images of night lights in the public domain. The advantage of such non-conventional data is that it allows analysis at a very fine level—we use an area of slightly less than 1 sq. km as our standard metric of analysis.

For Bihar, therefore, for each cell in the grid, we estimated spatial poverty by calculating the inverse of road length and habitable space normalized by population as well as population density. An index comprising these three elements was created. This Spatial Poverty Index (SPI) gave us a value of how ‘poor’ an area is for each of the cells in the Bihar GIS grid.

It was found that high poverty areas, as per the SPI, not only have fewer roads and inhabitable area per person but also use significantly lower energy on a per person basis. We also find that the conventional urban and rural distinctions do not hold in reality—they are just an artifact created by the government and its census.

In other words, urban and rural poverty are very similar in that the poor use less space and infrastructure in both urban and rural areas, even the quantum of this usage does not differ as much on a per person basis.

This is only the beginning, and it is done using only public data; far more is possible, and better policy can be designed.

Laveesh Bhandari, director, Indicus Analytics.

Catch all the Business News, Market News, Breaking News Events and Latest News Updates on Live Mint. Download The Mint News App to get Daily Market Updates.
More Less
Subscribe to Mint Newsletters
* Enter a valid email
* Thank you for subscribing to our newsletter.

Recommended For You

Trending Stocks

Get alerts on WhatsApp
Set Preferences My ReadsWatchlistFeedbackRedeem a Gift CardLogout