The Planning Commission estimates that poverty levels in India have been falling sharply since 1983 and, based on the mixed recall period, around 21.8 per cent of the population (238 million out of a total population of 1,093 million) suffered from deprivation and were below the poverty line in 2005.

In terms of the rural-urban spread, 21.8 of rural Indians are poor (170 million people) while the figure for urban India is 21.7 million (68 million). Arvind Panagariya has somewhat different numbers, and argues that poverty numbers fell from 48.4 per cent in 1977-78, to 43 per cent in 1983-84, 39.1 per cent in 1987-88 and to 32.6 per cent in 1999-2000.

These, however, could be vast under-estimates going by the results of the World Bank's latest study on global poverty. To arrive at its estimates, the World Bank uses a new poverty benchmark based on purchasing power parity.

The World Bank uses an income cut off of Rs 21.6 per day in urban areas and Rs 14.3 per day in rural areas (at 2005 prices) to define who is poor. For 2005, based on these income cut offs, the World Bank estimates that 42 per cent of Indians had incomes that placed them below the poverty line.

In other words, the World Bank estimates add around 200 million more Indians to the category of 'poor'. These additional 'poor', needless to say, are not eligible to the various benefits available under different government schemes for the poor, but it adds a totally new dimension to the poverty debate in the country.

NCAER has been collecting income and expenditure data since 1985 using large sample surveys. Applying this Planning Commission expenditure cut off for poverty for each and every state (at an all-India level, this works out to Rs 538.6 for urban areas and Rs 356.3 for rural areas), gives roughly the same poverty levels as the Planning Commission's.

According to the NCAER data, 21.7 per cent of rural households and 18.7 per cent of urban ones were poor in 2004-2005, adding to a total of 20.8 per cent at the all-India level (214 million poor out of a total of a population of 1,027 million).

If you plug in the higher World Bank income cut-offs used to define the poor, the NCAER data shows that around 29.6 per cent of the population is poor -- 33 per cent of the population in rural India and 21.6 per cent in urban India fall under the category of poor.

But the numbers apart, the real question is whether we should continue to define poverty in terms of calories and the food/income required for this, or whether poverty lines should take into account other aspects of consumption lifestyles?

After 1985, for instance, there is ample evidence that even the 'poor' are buying a host of manufactured products. The World Bank estimate adds 200 million to the 'poor' over the Planning Commission estimate. But if the 'poor', based on either of the income cut offs used, can afford to own manufactured goods, can we still consider them 'poor'?

Papers written by SL Rao in 1992 and Rakesh Mohan and others in the Economic and Political Weekly in 1997 analysed the NCAER survey numbers and asked precisely these questions. NCAER's large surveys of consumption by households in five income categories started in 1985 and by 1994, there was a sufficient trend that could be analysed.

At least some of the households counted as below the poverty line are owners of manufactured consumer goods. The bottom-most income category in the NCAER surveys was large and included the 'poor'. Not counting the 'poor' as owners would mean that the remaining households in the NCAER lowest-income category must own such goods. That gives an exceptionally high penetration of ownership in those residual households in NCAER's lowest income category.

Ownership of (select) consumer durable goods in the NCAER survey among the non-poor households is significantly higher than among 'poor' households. At an all-India level, 33 per cent of non-poor households own colour television sets, 25 per cent have telephones, 22 per cent have refrigerators, 19 per cent own cellular phones, nearly 7 per cent have cars and 2 per cent own credit cards.

In contrast, 8 per cent of 'poor' households own colour television sets, 4 per cent have telephones, 3 per cent have refrigerators, 3 per cent own cellular phones, and hardly 1 per cent have cars and credit cards each. But these 'poor' do own them.

Economic liberalisation since Rajiv Gandhi "broad-banded" the licensing system in 1986 has added to the supply of manufactured consumer goods. Competition and the drive by manufacturers to expand sales have led them to redesign products to make them affordable by households in the lowest income categories.

Innovations to products like the "Velvette" shampoo in sachets priced at one rupee have extended themselves to a variety of consumer goods, both of the non-durable and the durable kinds. Tata's Nano, the Ginger hotel chain and low-priced airline seats are the sign of how all marketers have inexorably extended the markets for their goods and services to low income groups.

The NCAER survey data led to the understanding that Indians love to own and consume manufactured goods. A closed economy for over forty years led to the misinterpretation that Indians were non-materialistic and not consumption oriented. That myth has now been shattered.

There are other facets of poverty that also need to be kept in mind for any meaningful discussion on the subject. A book published by NCAER last year, (How India Earns, Spends and Saves by Rajesh Shukla) shows that the degree of deprivation is a function of education (highest education level in a household) and the major source of income of the head of the household.

While nearly 26.7 per cent of non-poor households have at least one graduate, just 8.5 per cent of 'poor' households qualify under this attribute. A higher percentage of 'poor' have just a primary school education (25.5 per cent) or are illiterates (7.2 per cent) as compared to just 10.4 per cent and 3.8 per cent respectively in the case of non-poor households.

The major source of household income also varies significantly across 'poor' and non-poor households. Unskilled labourers constitute the largest segment (62 per cent) of 'poor' households; in contrast, this group accounts for 26 per cent of the non-poor households.

Those earning salaries account for 21.7 per cent of non-poor households whereas just about 4.4 per cent of 'poor' households earn their living through salary/wages.

It is time for us to review our definition of poverty to include not merely food intake in calories but also the purchase and ownership of some manufactured goods like soap, shampoo, radios, phones, small television sets, etc, as well as various socio-economic indicators like education and occupation.

This might enlarge the number of households who are 'poor' but enables a better understanding of how even the 'poor' would like to live.

S L Rao is a former Director-General of NCAER; Rajesh Shukla is a Senior Fellow at NCAER.