Rural Urban Economic Disparities in India: Database and Trends

The increasing gaps between the per capita income and consumption expenditure in urban and rural areas are major areas of concern in the strategy of inclusive growth, currently being followed in  India. The 'approximate procedures' used by the Central Statistical Organisation to work out the RU break-up of NDP in different sectors underestimates the urban vis-à-vis rural income. Similarly, the National Sample Survey system has a build-in bias to exclude households at the upper end of the scale, resulting in significant underestimation of urban consumption expenditure. The factors responsible for this systematic bias are likely to get strengthened over time so that the actual RU disparity would be greater than what is revealed through official statistics.

1. Introduction 

 Income and consumption expenditure in urban areas being significantly above that of rural areas and the gaps increasing over the past few decades is a matter of serious policy concern in India. The issue has become all the more important because of a relatively faster decline in population growth in urban areas due to higher levels of income, education, health care and family planning services, greater exposure to media and modernity etc. and a low rate of RU migration. This has also meant high growth in rural workforce on the face of a declining share in total income. All these make an analysis of the trend in rural urban economic inequality extremely important.

 The present note begins by overviewing the data situation pertaining to income and consumption expenditure for urban and rural areas, while highlighting the problems of temporal comparability. It then discusses the trends in rural urban inequality over the past four decades and their implications.    

2. Data on Income for Rural and Urban India

The Central Statistical Organisation (CSO) provides income estimates for rural and urban areas for the select years that have been used as the base, for computing the national income series. These are known to be generated through approximate procedures and therefore turn out to be not very robust. The rural and urban areas are not economically closed sub-systems, much less than the Indian states. Further, the changes in RU boundary demarcation are much more frequent than that in state boundaries. The conceptual issues at state level have to an extent been resolved through adoption of similar methodologies for estimating net value added (NVA) in different sectors and cleaning up of the data, received from State Statistical Bureaus, by CSO on a regular basis. This, however, is not the case for rural urban income data.

The methods followed for apportioning the net domestic product between rural and urban areas are different for different sectors. The CSO, in its note detailing out the methodology clearly stipulates that the apportionment “is primarily dependent on estimates of working force for each economic category”. The number of workers, used as the basis for determining the rural and urban shares in the total income, in many of the sectors is obtained from the preceding Population Census. These, therefore, pertain to a past year and not the year for which income figures are being generated. Understandably, this overestimates the rural income figures as the actual growth in rural workforce during the intervening years always works out as low compared to that in urban areas, largely due to reclassification of rural settlements as urban and RU migration. 

There are three other factors in the methodology (CSO, 2007), responsible for the over estimation of rural income. For several sectors, for which wage data for rural and urban areas are available like fisheries, unorganised manufacturing, construction etc. through National Sample Surveys (NSS), the apportionment is based on total wage bill, computed by multiplying the average wage with the number of workers. This would overestimate rural income as the wage bill as a percentage of NVA is likely to be higher in rural areas. It may be argued that the share of urban income in many of these sectors would be larger than the share in wage bills because of the higher labour productivity (in urban compared to rural units). This is because the impact of other factors like capital assets, technology and management would be far greater in urban than rural areas. The assumption that the ratio of wages to NVA is the same in rural and urban areas tends to give higher share for rural areas than what would emerge from direct calculations.

In case of a few other sectors wherein separate wage data for rural and urban areas are not available, RU apportionment is done simply based on the number of workers, as reported in the last Population Census. This is the case for the public segment of the services sector like health, education etc. One can argue that several of the urban jobs in these sectors are at supervisory/professional levels and would record higher productivity or carry higher emoluments than the average of the rural wages. Assuming uniform labour productivity at two or even three digit level of classification would imply serious overestimation of rural income. Similarly, for agriculture and animal husbandry sectors, the RU bifurcation is done based on the number of cultivators/tractors and livestock population respectively. This again ignores the differences in input productivity across rural and urban areas.

There is, thus a systematic bias in the methodology leading to relative overestimation of rural income. Given that the urban rural gaps between the growth in workforce, the ratio of wages to value added and productivity of the factors of production are likely to increase over time, the degree of overestimation of rural income over urban income would go up in future years.

The problem is far more significant in case of consumption expenditure. It is well documented in the literature that respondents in urban areas have definite reasons for understating their consumption levels than in rural areas because of the progressive taxation system and certain social security benefits being linked with their economic wellbeing. Besides, sampling procedures result in larger exclusion of households at the higher end of the scale. The number of such households being higher in urban centres, the problem would be more serious here. This, combined with relatively greater anonymity and difficulties of locating a household  or having access to to its head in urban milieu, results in greater probability of exclusion of better off households. Finally, non response or non compliance is far higher among the urban rich than the others (Deaton, 2003). All these are responsible for significant underestimation of urban expenditure. Furthermore, the importance of these factors would increase over the years, leading to much greater degree of RU inequality than captured through NSS data.

3. Trends in per capita SDP and consumption expenditure

The income share of rural areas in total GDP at current prices can be seen as declining sharply from 62 per cent in 1970-71 to 48 per cent in 2004-05. This decrease can not be attributed merely to the falling share of rural population or workforce. To an extent, this is due to the faster growth in urban per capita income compared to that in rural areas, particularly in recent years. It is indeed true that the categories - agriculture and non-agriculture - do not coincide with rural and urban areas and yet there is considerable overlap. Now, the growth in non-agricultural GDP was 8.11 percent and 7.22 per cent during 1993-99 and 1999-2004 respectively. Both these rates are significantly higher than the corresponding rates for agricultural GDP that were 2.8 per cent and 2.6 per cent. Since 2003-04, the non agricultural GDP has grown at over 12 per cent as against agricultural growth of less than 3.0 per cent only.

The income share of rural areas has gone down in three sectors (a) Electricity, Gas and Water supply, (b) Financing, Insurance, Real Estate and Business Services and (c) Community, Social and Personal Services, implying that here the NVA have grown sluggishly, compared to their urban counterparts. The growth rate of rural income, on the other hand, has been faster in (a) Mining and Quarrying, (b) Manufacturing (c) Transport Storage and Communication and (d) Trade, Storage and Communication.  The significant fall in the urban share during the period from 1999-00 to 2004-05, especially in case of the last two sectors is due to change in the method of determining the urban share. Notwithstanding all these, the rural urban disparity in per capita income has gone up sharply over the three and a half decades. The ratio of urban to rural per capita income was 2.45 in 1970-71 and remained at a low level of 2.30 during eighties and early nineties but went up sharply to 2.7 and 2.8 in early and mid years of the present decade (Graph 1).

Graph 1

All-India Average Per Capita Income

1970-70 to 2004-2005

 An overview of the trend based on the data on per capita consumption expenditure from NSS at current prices gives a similar trend. The growth rates in both have picked up slowly during seventies and eighties (Graph 2). The rise, however, has been particularly very steep in urban expenditures in the nineties and subsequent years. It may be mentioned that figures up to the year 2004-05 are based on the large sample data of NSS while that for the subsequent years - shown in the Graph in dotted lines - are from the small sample. The trend of growing inequality however emerges clearly from both the data sets.

Graph 2

All-India Per Capita Consumption Expenditure (Rs.)

1972-73 to 2007-08

• Login or register to post comments
• 3103 reads
• Calendar
• Printer-friendly version
• Email this page