Why India’s new income survey won’t address key questions around poverty and inequality
The reliability of the household consumption surveys conducted by the National Statistical Office (NSO) for 2022-23 and 2023-24 for the measurement of poverty and inequality is a matter of debate. More so with regard to their comparability with past estimates of consumption expenditure. But these surveys are also important sources of data for inputs to several other macro indicators such as the National Accounts and the Consumer Price Index (CPI), both of which use these for updating the base year.
Meanwhile, the National Statistical Office (NSO) has decided to carry out a survey of household incomes. As part of the exercise, last month it released a Brief Report on Pre-testing of Schedule of Household Income Survey for comments and suggestions.
Unlike the consumption surveys, which have been in existence since the inception of the National Sample Survey Office (NSSO), there have not been any income surveys conducted. But there are income estimates for households available from the India Human Development Surveys for 2005 and 2012 and for 2022, results of which are awaited.
To be sure, NSSO has previously attempted to carry out an income survey, with pilot exercises being conducted to examine their suitability and reliability in the Indian context. The earliest attempts were made in the 9th round (1955) and the 14th (1958-59) along with the usual NSSO consumption surveys, but these were experimental in nature and were not taken forward as full-fledged surveys.
The last serious attempt was the pilot survey done from September 1983 to December 1984 as part of the 38th round in Haryana, Maharashtra, Orissa, Tamil Nadu and Uttar Pradesh along with Delhi and Calcutta. The results confirmed the difficulty of obtaining data on income from household surveys. For the purposes of comparison and testing the validity of the estimates, NSSO conducted the surveys with three similar samples. The first had only questions on income, the second had consumption and savings while the third had all the three.
In general, it was observed that income estimates were 30-40% lower than the corresponding sum of consumption and savings in rural areas. However, the difference was noticeably lower at 7% in urban areas. It was around 12% for metro cities. Also, it was lower for salaried households compared to non-salaried.
Within the non-salaried, it was larger for cultivator households. The dominance of salaried households explains the lower differences in urban areas compared to rural areas where cultivators are present.
On inequality, it was substantially higher when measured using incomes than using consumption and savings or consumption alone. Notably, income paid as income tax was insignificant even in metro cities.
The current pre-testing report confirms the apprehensions expressed in the 1983-84 survey. As high as 95% of the respondents were uncomfortable disclosing their incomes and a majority refused to answer questions on income tax. Similarly for jewellery. The pre-testing report also suggests a general trend of under-reporting of income compared to expenditures.
Despite these difficulties, the attempt to conduct an income survey is laudable. It may not accurately capture the income of all households in the country. It is easier for salaried households to report incomes.
But getting estimates of income for a large majority of households in the informal sector is a problem, more so for farmers and the self-employed who rarely maintain a book of accounts. For such households, the estimates would be less reliable. An increase in pluri-activity (multiple sources of income) and piece-rated or gig-work would create further difficulties.
The estimates on income will certainly add to our understanding of the nature and pattern of income distribution in the country after 2011-12. Especially given the serious comparability issues with the consumption surveys. But these may not provide an answer as to what happened to poverty in India, partly because these are the first estimates of income distribution and therefore no previous estimates exist. But also because of the absence of a poverty line to be used with income surveys.
It is likely that the incomes estimates would point to inequality levels larger than what is reported by the consumption surveys. This may be uncomfortable to those who believe India to be the fourth most-equal country. Nevertheless, it will allow for a much-needed discussion on the level of inequality in the country.
The author is associate professor at Jawaharlal Nehru University and visiting fellow at the Centre de Sciences Humaines, New Delhi.
Post Comment