|
VIEWPOINT: KAUTILYA
Lifting
The Poverty Veil
India
loses Pravin Visaria just as his handiwork becomes official
By
Jairam Ramesh
The
sheer fragility of life is so very striking. Pravin Visaria, one of India's
finest economists, suddenly passed away, virtually unnoticed, just as
Yashwant Sinha was presenting his recent budget. Visaria was director
of the Delhi-based Institute of Economic Growth. He also chaired the governing
council of the Centre's National Sample Survey Organisation (NSSO). It
was in this capacity that he influenced the debate on India's poverty
numbers. He would have gone to his final abode vindicated by Sinha's words
in the budget speech that "poverty has fallen from 36 per cent in
1993-94 to 26 per cent or less now".
|
POPULATION
BELOW POVERTY LINE
|
| (1999-2000
Survey) |
30-day recall |
7-day recall |
| Andhra Pradesh |
15.8 |
13.8 |
| Assam |
36.1 |
30.6 |
| Bihar |
42.6 |
36.7 |
| Gujarat |
14.1 |
12.8 |
| Haryana |
8.7 |
7.8 |
| Karnataka |
20.0 |
16.6 |
| Kerala |
12.7 |
11.1 |
| Madhya Pradesh |
37.4 |
34.8 |
| Maharashtra |
25.0 |
22.6 |
| Orissa |
47.2 |
43.4 |
| Punjab |
6.2 |
5.3 |
| Rajasthan |
15.3 |
13.9 |
| Tamil Nadu |
21.1 |
19.3 |
| Uttar Pradesh |
31.2 |
28.8 |
| West Bengal |
27.0 |
23.4 |
| ALL INDIA |
26.1 |
23.3 |
Figures are in percentage and relate
to rural and urban areas combined
Source: Planning Commission |
Our poverty numbers are derived from sample surveys
carried out by the NSSO on consumer expenditure. The poverty line is the
monthly expenditure incurred in getting a daily calorie intake of 2,435
kilocalories in rural and 2,095 kilocalories in urban areas. The Planning
Commission computes state-specific poverty lines and ratios. The national
poverty line is worked out implicitly and the all-India poverty ratio
obtained as the weighted average of statewise poverty ratios.
Every five years full surveys on 1,20,000 households
are carried out. In the intervening period, "thin" samples of
around 20,000 households are surveyed. The "thin" samples are
not considered robust enough to be used to draw any firm conclusions.
Only the full-year surveys are considered authoritative; after 1993-94,
the next full survey was carried out during July 1999-June 2000.
In 1993-94 as in all previous surveys, respondents
were asked questions on spending on food items based on a 30-day recall,
that is, what they had spent a month prior to the date of the survey.
This 30-day recall period is unique to Indian surveys and goes back to
an experiment done by P.C. Mahalanobis way back in 1954 in 76 West Bengal
villages covering 1,254 households. Visaria believed that this would tend
to underestimate spending on food since people cannot be expected to reliably
recall their spending patterns after 30 days had elapsed. Poverty would
thus get overestimated. He held the view that a seven-day recall period
would be more realistic, particularly for food items, which for most Indians
remains the bulk of the consumption basket. In fact, he got the "thin"
sample surveys between 1994-95 and 1998-99 to include the seven-day recall
period question as well and from the results that were obtained Visaria
went public in July 2000 saying that the seven-day recall period yields
significantly lower estimates of poverty than through the use of the 30-day
recall period.
Thus, in 1999-2000 the seven-day recall period
was used for the very first time for a full survey year. Presumably respondents
were asked questions based on a seven-day recall period first and then
were asked questions based on the 30-day recall. This was unlike in the
"thin" sample surveys where different households were asked
questions based on seven-day and 30-day recall periods. This was certainly
the case for the period July-December 2000. What happened later is a bit
unclear but it transpires that the NSSO issued guidelines to change the
sequence. It is possible, as some experts fear, that the data, at least
for the first half of 1999-2000, would have got contaminated. What may
have happened is that a respondent would have replied on a seven-day recall
and then simply multiplied by four and given the answer for a 30-day recall.
Also, for clothing, footwear and durable goods, the 1999-2000 survey used
a 365-day recall unlike in 1993-94 when a 30-day recall was used. Puzzlingly,
the seven-day recall yields a poverty estimate which is only about 3 percentage
points lower than that based on a 30-day recall. In earlier surveys this
difference was at least five times more.
But Visaria himself was most insistent that
the 1999-2000 survey was well designed and supervised and felt that just
because the results of 1999-2000 did not appear to fit preconceived notions,
we should not dismiss the numbers outright. He thought that the 1999-2000
survey yielded "better" poverty estimates. And he said this
without having any liberalisation axe to grind.
What does one make of the poverty numbers themselves?
All that can be said is that we measured something in 1993-94 and got
36 per cent. We measured the same thing but by a different methodology
and got 26 per cent. Even when we compare like for like, that is estimates
based on a 30-day recall, we cannot straightaway conclude, as Sinha did
in his budget speech, that poverty has fallen. The data are simply not
comparable. Poverty may well have fallen and there is other evidence that
it indeed did in the 1990s but the only way forward is to carry out a
full-year survey in 2002-03 on the same basis as 1993-94 and get a better
fix on the numbers. Or we use the 1999-2000 methodology as a reference
point for the future. It is also time that individual states themselves
carried out large-scale poverty and well-being surveys at regular intervals.
(The author is with
the Congress party. These are his personal views.)
|