How many Indians are poor depends on the line.

Pick a line on the horizontal axis and read up: the curve gives the share of Indians who consume less than that each day. The curve is steep in the middle because most Indians live within a few dollars of each other; a small move in the line there moves hundreds of millions of people across it.

A PPP dollar is not a dollar at the exchange rate. It is the amount of rupees that buys in India what one dollar buys in the United States, so it already allows for lower Indian prices. The World Bank uses the same lines for every country: $3.00 for extreme poverty, $4.20 and $8.30 for the typical lines of lower-middle and upper-middle income countries.

Each dot is one national consumption survey. The World Bank groups them into spells it treats as comparable: the three rounds from 1993-94 to 2009-10 used the same questions; 2011-12 asked about some items over a different recall period; the HCES of 2022-23 was redesigned, with more items asked over several visits. A line joins dots only inside a spell. The drop between spells mixes real change with the change in the questionnaire, and nobody can fully separate the two.

The recall period matters because people remember a week's vegetables better than a month's, and a year's clothes differently from a month's. A longer list of items and better recall both record more spending, and so fewer people below any line.

The NSS consumption survey of 2017-18 was completed but never released, and the next survey came only in 2022-23. For the years between, the World Bank fills the gap with a statistical model, anchored on the surveys either side and moved by growth in national accounts and other data. A year's estimate is aligned to the calendar year, so it does not equal the survey reading for the overlapping year.

The model cannot see what happened to the poor in particular: a drought, a lockdown or a rise in rural wages moves real lives but shows up only through the national average. For those years we do not know India's poverty rate; we have a well-made guess.

A falling share can still mean a rising count when the population grows; the bars show the people. The poverty gap adds depth: it averages, over everyone, how far below the line each person falls (zero for the non-poor), as a share of the line. Two countries with the same headcount can have very different gaps.

The Gini index runs from 0, where everyone consumes the same, to 100, where one person consumes everything. Consumption is far more equal than income or wealth: the rich save much of what they earn, and household surveys rarely reach the very rich at all. A low consumption Gini says little about inequality of income or wealth, which this source does not measure.

Line all Indians up by how much they consume a day, split the line into ten equal groups, and add up what each group consumes. If consumption were equal, each tenth would have 10%. The consumption guide shows the same survey in rupees, rural and urban, by fractile.

Each measure answers a different question. The $3.00 line asks who is destitute by a global standard. The $4.20 and $8.30 lines ask who is poor by the standards of countries a little or much richer than India. The societal line rises with the median, asking who has fallen behind their own country. India's national line, last set by the Tendulkar committee and last applied to 2011-12, uses Indian prices and consumption baskets. Multidimensional measures count deprivations beyond money: the UNDP and Oxford index counts homes without electricity, clean fuel, schooling or sanitation and leaves consumption out, while the World Bank's measure adds such deprivations to the $3.00 line.

India has had no official national poverty line since 2011-12. The Rangarajan committee proposed one in 2014; the government did not adopt it.