Encouraging News about Poverty

For all those following the debate on poverty trends since the reforms of 1991, there is encouraging news from the National council of Applied Economic Research. Its latest consumer survey for 1993-94 shows that:

  • The proportion of low-income households has diminished after the economic reforms of 1991, a continuation of the declining trend evident since 1985-86 when it conducted its first survey.
  • Social mobility is substantial, and people are moving up the income ladder from the bottom to the middle, and from the middle to the top.

Critics will regard this as too good to be true after all, haven’t the mini-surveys of the data shown a worsening of poverty in the first 18 months of reform? True, but the mini-surveys do not yield any data beyond the end of 1992, where as the NCAER has data up to 1993-94. Even reformers will readily admit that poverty must have worsened in the austerity phase of reform, but argue that the trend would have been reversed subsequently. Besides, doubts have arisen about the quality of mini-surveys based on thin samples—they show poverty worsening, but also show that the proportion of people eating two square meals a day is going up, an this paradox casts suspicion on their reliability. The NSS of 1993-94 covers a full sample and should provide a more authoritative picture, but until its results arrive it is important to consider data from other sources. The NCAER is one such. Its sample of 300,000 is even larger than that of the NSS.

Table 1 (Two views of poverty) is a compare of poverty trends as measured by the NSS (using the Lakdawala expert group’s methodology) and by the NCAER. The NCAER defines low-income households as those with an annual income below Rs 12,500 at 1989-90 prices. Given that on average we have 5.7 persons per household, the NCAER definition corresponds roughly to the poverty line in urban areas, but is well above the poverty line in rural areas (where the same income has much greater purchasing power because of lower rural prices).

The two sources tell very different stories. The NSS story is that there was a very steep fall in poverty in the two years preceding economic reform, followed by a steep rise subsequently in the first 18 months of reform. By contrast, the NCAER story is that poverty has diminished gradually, with no wild fluctuations, both before and after 1991.

The NSS data suggest that poverty fell by a whopping five percentage points in 1989-90, which means 45 million people were raised above the poverty line in one single year. I find that difficult to swallow. The worsening of poverty by 5.2 percentage points in 1992 looks more plausible, yet cannot be easily reconciled with the NSS claim that the number of persons eating two square meals a day increased from 88.3 per cent in1989-90 to 92.3 per cent in 1992. So I suggest that mini-survey data should be treated with great caution.

The NCAER data are all from full samples, without intervening mini-surveys. They show income groups, from 65 per cent in 1985-86 to 59 percent in 1993-94. A good check for credibility would be to see whether the NCAER reported a worsening of poverty in 1991-92 (which surely must have occurred, given high inflation and low GDP growth).But the NCAER has no data for this year.

Both sources give very similar ratios for urban poverty, and both agree these have not changed much since 1989-90. Given their overall divergence, this convergence on the urban poverty ratio may sound surprising. But it is not entirely expected, since the NSS poverty line are roughly the same for urban areas. The two lines are very different in rural areas, which may be one reason for the enormous difference in rural trends they shoe. Low-income families on the NCAER definition covered 67 per cent of rural families in 1989-90, double the 33.7 per cent of poor families as defined by the NSS. So the NCAER and NSS re talking about two very different groups in rural areas—the NCAER definition covers, by NSS standards, almost as many non poor as poor households.

The NCAER data have some limitations. The organisation asks respondents about their income over a full year, and it is not easy for casual worker or the self-employed to recall this accurately. The NCAER says respondents generally report take-home income, leaving out payment s in kind, provident fund or ESI contributions, and perks like an office bus. So its data under –report income by anything ranging from zero to 25 per cent.

Finally, the household data of the ANCAER do not tell us enough. It makes a huge difference to living standards whether a family earning Rs 12,500 has tow members or eight, and the NCAER makes no distinction between the two. So. The NCAER findings too should be treated with caution. The results of the 1993-94 NSS survey will be mire authoritative. Still, until these results become available, the NCAER findings provide useful clues.

They also tell an interesting story about income mobility. The NCAER divides households into three categories — low income, lower middle income, and middle/ high income (see the second Table).

Between 1985-86 and 1993-94, the proportion of low-income households declined by 8 percentage points. In the same period, the proportion of middle/high income groups rose 7 percentage points. So virtually as many households have moved up from the middle to the top as from the bottom to the middle. Indeed, some must nave moved from the bottom to the top.

This is an encouraging picture of social nobility. It shows that the rise of the consuming class (reflected by a boom in consumer durables) represents the promotion of the bottom and middle categories. Critics say, incorrectly, that the consumer boom shows a bias against the poor. In (‘act, the new consumers are promotees from the ranks of the poor.

Mobility is evident in rural as well as urban areas. The proportion of low income households in rural areas declined by 12 percentage points in the reference period, against 5 percentage points in urban areas (from 42 per cent to 37 per cent). The rural proportion of middle/high income households more than doubled from 5 per cent in 1987-88 to 12 percent in 1993-94.

Is there movement down the income ladder as well as up? The data do not give us an answer. However, an earlier NCAER study in 1984 by Dr I Z Bhatty looked at consumption patterns in a set of 3,12 8 families in 1970-71 and went back to the same families in 1981-82. The result looked like a triumph for socialism, with the rich getting poorer and poor getting richer. The biggest jump in consumption (127 per cent) was for the lowest income decile, the second biggest jump (60.2 per cent) was for the second-lowest decile. The progression continued steadily with the richest three declies actually suffering a decline. Indeed, the richest deciles suffered a decline of no less than 42.03 per cent, representing a decay of the old feudal class.

Many economists refuse to accept these findings at face value. But no matter what qualifications we make, the survey seems to establish that India exhibits considerable income mobility, down as well as up. If this was true in the 1970s, when poverty decreased very slowly, it should also be true of more recent times.

The recent consumer surveys of the NCAER, from 1985-86 to 1993-94, show that upward mobility at any rate continues, though it had decelerated in the difficult initial reform years.

Rural poverty continues to be much too high and the top income category much too small for comfort. The pace of change still leaves much to be desired. But, NCAER is at all to be believed, the is in the right direction.

Recent consumer surveys show that upward mobility at any rate continues, though it had decelerated in the difficult initial reform years, says Swaminathan S Anklesaria Aiyar

Two Vies of Poverty

Poverty Ratio (%)
Low Income Groupo (%)
Rural Urban Total Rural Urban Total
1983 49.0 38.3 46.5 1985-86 73 42 65
1987-88 39.1 40.1 39.3 1987-88 69 40 61
1989-90 33.7 36.0 34.3 1989-90 67 37 59
1990-91 35.0 37.0 35.3 1992-93 65 38 58
1992 41.7 37.8 40.7 1993-94 65 37 57
(Note data refer to-consumption while NCAER data refer to income. Its low-income group is defined as households earning below Rs 12,500 annually at 1989-90 prices. The NCAER underestimates income by up to 25 per cent since it records take-home rather than total entitlements.

Two Vies of Poverty

Up the Income Ladder
Per Cent of Household
Low income Lower middle income Middle high income
Overall Rural Overall Rural Overall Rural
1985-86 65 73 25 22 10 5
1987-88 61 69 28 25 11 5
1989-90 59 67 27 24 14 9
1992-93 58 65 25 23 17 12
1993-94 57 65 26 23 17 12
(Low income is defined as up to Rs.12, 500 per year per household; lower middle income, is Rs 12,500-Rs 25,000 per year; Middle/high income; is above Rs 25,000 per year. All these are at 1989-90 prices)

What do you think?