It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts." That's Sherlock Holmes (Arthur Conan Doyle if you prefer) in "A Scandal in Bohemia". As most people know, the National Sample Survey Organisation (NSSO) collects data on consumption expenditure. However, for many policy questions, one would like to have data on distributions of income. Since income recall is perceived to be unreliable (deliberately or inadvertently), NSS doesn't collect it. The only organisation that does it is the National Council of Applied Economic Research (NCAER), through the National Survey of Household Income and Expenditure (NSHIE). But the last time NSHIE was undertaken was in 2004-05. However, there's the August-November 2014 survey of People Research on India's Consumer Economy (PRICE). This is known as an ICE (India's Consumer Economy) survey, with 20,195 households spread over 21 states and 67 districts. Rajesh Shukla used to be with NCAER and was one of the driving forces behind NSHIE. He is one of the driving forces behind PRICE now. Therefore, estimates from NSHIE in 2004-05 have been made comparable with those from ICE in 2014.

First, everyone debates inequality in the distribution of incomes and a standard aggregate measure of inequality is the Gini coefficient. Gini coefficients usually cited are from NSSO, based on consumption distributions and there is no disagreement that inequality in the distribution of income will be more than inequality in the distribution of consumption expenditure. NSHIE showed a Gini coefficient of 0.466, but ICE shows a Gini coefficient of 0.386. Hence, measured by the Gini coefficient, income inequality has declined between 2004-05 and 2014. This is a descriptive statement, not a prescriptive one. Second, aggregate measures of inequality are always misleading. They blur out what's happening with the distribution. Shares of quintiles or deciles are better, they tell us about the shape of the Lorenz curve. Between 2004-05 and 2013-14 (shares are determined on the basis of the full financial year), the share of bottom 20 percent in household income increased from 5.2 percent to 6.6 percent, while share of top 20 percent declined from 52.7 percent to 46.1 percent. I have given you figures for Q1 (bottom 20 percent) and Q5 (top 20 percent). But this pattern is repeated for Q2 (21-40 percent), Q3 (41-60 percent) and Q4 (61-80 percent) too. That is, shares of Q1, Q2, Q3 and Q4 have all increased, while that for Q5 has declined. The reason is simple. As one goes up the quintile ladder, annual growth in household income becomes lower and lower. It is 8.77 percent for Q1 and 2.86 percent for Q5. Stated imprecisely, annual income of the rich has increased. But that of the poor has increased relatively more.

Third, compared to Census 2011, there have been improvements. I am not going to deluge you with data. Take households with electricity as an example, up from 66.7 percent in 2011 to 87.3 percent in 2014. Or households with tap water, up from 21.6 percent in 2011 to 36.6 percent in 2014. Fourth, the debt/income ratio is almost the same across quintiles. (I hadn't expected this.) For Q1, the debt/income ratio is 25.8 percent, while for Q5, the debt/income ratio is 23.8 percent. The debt/income ratio is almost neutral across quintiles. The average debt of Q1 is Rs 22,400 and the average debt of Q5 is 107,000. What's different is the source of debt and the purpose of debt. Fifth, I haven't mentioned rural/urban differences so far. Let's bring that in. Thirty nine percent of Q5 urban households are indebted. The figure is 6 percent for Q1. Fifteen percent of Q5 rural households are indebted. The figure is 21 percent for Q1. Stated a bit imprecisely, in urban India, the richer you are, the more likely you are to be indebted. In rural India, the poorer you are, the more likely you are to be indebted. But the relatively rich borrow from the formal credit market, while the relatively poor borrow from the informal market. This is along expected lines. Sixth, why do people borrow? In rural India, the most important reasons, in descending order of importance, are agricultural loans, medical emergencies, consumption and children's education. In urban India, the most important are medical emergencies, business requirements/expansion, children's education and the purchase of automobiles. This vindicates the assumption that exogenous medical shocks can often push a household below the poverty line.

Seventh, urban India is sliced into metros (population more than five million), boom towns (2.5 to 5 million), niche cities (1 to 2.5 million) and other urban towns (less than one million). On an average, 47.7 percent of urban households and 52.3 percent of rural households have surplus income, defined as an excess of income over expenditure. Within the urban segment, guess where surplus income tends to be greater? Metros (probably expected) and other urban towns (probably unexpected). Surplus income tends to be lower in boom towns and niche cities (I hadn't expected either). Eighth, as with urban, there is a slicing of rural across developed rural (top quartile, defined through an index), emerging rural (2nd quartile) and under-developed rural (3rd and 4th quartiles). Would you have guessed that the average annual household income for developed rural is a little bit more than for niche cities? There are many more nuggets and I have given you no more than a flavour. ICE is worth milking much more than it has been.