Those who have been reading various newspaper reports on the latest results from the National Sample Survey (for its January to December 2003 round) are certain to be a confused lot.

For while it is true the survey shows positive results as far as consumption growth is concerned, the numbers are very different from those that come from the National Accounts or GDP figures that are regularly put out by government agencies such as the CSO.

While the official annual per capita income was around Rs 21,000 in 2003-04, according to the National Accounts; the NSS throws up a figure of around Rs 8,500 for the same period! So where's the rest gone?

Clearly you need to add in the savings rate (around 27 per cent) and another 15 per cent or so to account for government expenditure, but you still don't come anywhere near the National Accounts figures.

Economist and Business Standard columnist Surjit Bhalla, who has studied this at length, says this is not a peculiarly Indian problem and that consumption surveys the world over capture less and less of overall consumption each year -- the Survey to National Accounts ratio for India, Bhalla's calculation was 61 per cent for 1983, remained around the same in 1993, fell to 55 per cent in 1999, and is 49 per cent in the latest round.

So, anyone building up a hypothesis on spending habits in India on the basis of NSS data would do well to keep this in mind.

But if you assume, perhaps correctly, that the Survey has no inherent bias or that this is even distributed across various income and geographical (that's rural/urban) groups, the results are interesting.

While the annual growth in consumption in urban areas has been higher than that for rural areas in the past 15 years that I have data in front of me for, there seems to have been a turnaround in recent years, and from July 2000-June 2001 onwards, it is rural income that is growing faster--between two periods, July-June 2001-02 and January-December 2003, rural per capita consumption expenditure grew by 11.5 per cent while the urban one grew by 9.6 per cent.

This change, of course, is brought out by the NCAER's Great Indian Middle Class (2004) even more powerfully.

Compare the change in growth for various income classes in rural and urban areas between 1995-96 and 2001-02 with the projected incomes for the end of the decade in the case of households that have an annual income of between Rs 90,000 and Rs 200,000.

Between 1995-96 and 2001-02, the number of such urban households grew by 6.5 per cent as compared to 5.7 per cent for rural areas.

By 2009-10, however, this changes dramatically, and while the number of urban households has been projected to grow by just 4.1 per cent, the number of such rural households is projected to grow by 11 per cent -- since both rural and urban areas had broadly similar numbers of households in this income category in 2001-02, the difference in future growth projections cannot be attributed to a lower rural base.

So, don't be surprised if, over the next year or so, you see marketers upping their spending on developing rural networks. Going by the NCAER projections, while around half those households earning between Rs 90,000 and Rs 200,000 lived in rural areas in 2001-02, by the end of the decade, this will rise to over 60 per cent.

And as far as the country's rich, or those that earned over Rs 10 lakh per annum, are concerned, 22 per cent already lived in rural areas in 2001-02, and this is projected to rise marginally by 2009-10.

And, apart from their large share in demand for washing soaps and detergents (50-70 per cent), rural markets account for around 40-50 per cent of the demand for scooters and motorcycles and televisions, and this share is likely to rise even further.

While it is true that rural income levels are still smaller than urban ones, as the NSS brings out, clearly, several rural markets are almost as prosperous as urban ones.

Rural Kerala, for instance, had a per capita consumption of Rs 981 per month in 2003, a figure that's higher than that for urban Karnataka (Rs 960) and just marginally lower than urban West Bengal's Rs 991 per month or the all-India urban average of Rs 1,022 per month.

Apart from the faster growth in rural consumption, what's equally important from a marketer's point of view is the change in consumption habits in rural India.

So, while the urban consumption of cereals fell from 23 per cent in 1983 to 17 per cent in 1993 to 12 per cent of the total urban consumption in 2003, that for rural areas also fell, from 37 to 27 to 21 per cent.

Expenditure on education rose dramatically, from 2.1 to 3.9 to 6.3 per cent of the total expenditure in urban areas in 1983, 1993, and 2003, respectively, and it rose from 0.8 to 1.5 and 2.9 per cent in the same period in rural India.

Medical expenses rose from 2.2 to 4.5 and 5.9 per cent in urban areas but rose from 2.5 to 5.3 to 7 per cent in rural areas. While there are differences from item to item, in a broad sense, rural demand patterns are the same as they were in urban areas 8-10 years ago.

Postscript: While policy makers, especially of the UPA-type that are constantly looking for "aam aadmis" to pamper, would probably look at the unemployment as well as the poverty data the survey throws up, other interesting statistics that the survey throws up include the sharp decline in the number of people living in katcha houses, from 30 per cent of rural families in 2001 to 21 per cent in 2003.

The number of families using kerosene as the primary source of cooking, interestingly, fell by 21 per cent between 2001 and 2003 in rural areas (33 per cent in urban areas)--it's unlikely, however, that supplies of subsidised kerosene fell equally dramatically!