A major new source of data on rural households is the recently completed NABARD-sponsored National Financial Inclusion Survey or NAFIS 2016-17. This is a large sample survey covering 40,327 households in 245 districts of 29 States. The definition of rural for this survey is the RBI classification of Tier III to Tier VI centres, that is, all places with population up to 50,000. This survey is going to be repeated every three years and could potentially be a useful source of information on rural India.

The first striking thing about this survey is its ambitiousness in terms of data collection. The objective of NAFIS is to provide a “comprehensive overview of the status of rural households in terms of sources of livelihoods, economic status of households including income, consumption expenditure and assets (p. 32).” Except for a pilot survey, the NSSO has never attempted to collect data on income and expenditure whereas NAFIS provides estimates of income, expenditure, saving, investments and asset ownership. In this first Note on NAFIS, I shall focus on income.

According to NAFIS, on average, a rural household had a monthly income of Rs 8,059 in the year preceding the survey. The average income was slightly higher for an agricultural household (Rs 8,931) and lower for a non-agricultural household (Rs 7,269). This is mainly because a large number of wage labour households would fall in to the latter category. These estimates suggest that an average rural household obtained an income of about one lakh rupees a year. Interestingly, the Situation Assessment Survey (SAS) of 2013 found that the average income of an agricultural household was Rs 6426 in 2012-13. So, even without adjusting for inflation, the increase over three years (2012-13 to 2015-16) is only around two thousand five hundred rupees a month. Adjusting for inflation would, of course, lower the increase in income. Further, for an agricultural household, income from crop cultivation averaged Rs 3,140 a month and that from wage labour another Rs 3,025 a month.

While I welcome this Report, as there are very few surveys of household income in India, at the same time, I have to caution against relying on these numbers. The reason for the lack of surveys on household incomes in India is that data on income are very difficult to capture, as income is a derived variable and has to be derived from a detail cost accounting (Bakshi, 2010). In order to estimate income, we need to document a range of economic activities and the costs and returns of each. This has been attempted by the Project on Agrarian Relations in India (PARI) through its village-level surveys.

Drawing on the PARI experience, there are at least two glaring problems with the NAFIS data on incomes. The first pertains to the reference period. The survey was undertaken in January to June 2017 and the reference year for agriculture is July 2015 to June 2016, and for all other information the reference year is the preceding 12 months. It is not clear from the published data as to how incomes for 2015-16 from crop production were added to other incomes, say from business or services, for the period after June 2016. For example, a household interviewed in January 2017 would have been asked to report income from business from the previous year (Jan to December 2016). More importantly, a household interviewed say in March 2017 would have to recall data on crop expenditure for a period ending nine months prior to the interview, that is, for the year ending June 2016. In short, in the data on household incomes, there can be a problem of recall, especially for crop incomes, as well as of matching the reference period for different sources of income.

The second and more easily understood problem is that of the limited information collected in the questionnaire, particularly on crop incomes. To illustrate, in Table 6 of the questionnaire, for each crop, the information collected is total production, quantity sold, price and value, and total input cost in rupees as well as cost of hired labour. In short, the respondent has to give a lump sum amount spent on inputs for a specific crop. This raises a lot of questions, such as valuing of own inputs like manure or home grown seeds. It also does not discuss how inputs common to more than one crop (such as costs of irrigation) are collected and included. And there is no mention of several other costs such as of credit or rent or depreciation. The questionnaire does not meet the standards for calculation of costs already set by the Commission on Agricultural Costs and Prices. There is a high probability of underestimation of costs in the approach used by NAFIS, implying that the incomes from crop production reported above are likely to be over-estimated.

References

Bakshi, Aparajita, 2010, Rural Household Incomes, Ph.D thesis submitted to the University of Calcutta.

NABARD All India Rural Financial Inclusion Survey 2016-17, available at
https://www.nabard.org/auth/writereaddata/tender/1608180417NABARD-Repo-16_Web_P.pdf