The point in my news analysis of June 14, was to question the validity
of the Finance Ministry’s claim that poverty had declined by 4.2
percentage points in the period 2000 to 2004. The official claim was scrutinized
at two levels: (i) I had questioned the lack of consistency between the
government’s figures on the composition of GDP growth (which suggested
increasing poverty) on the one hand and the claim of sharp reduction in
poverty on the other. (ii) I had also questioned the credibility of the
claim in terms of a flawed estimation procedure for poverty reduction.
In his response (DT June 15) Dr. Ashfaque H. Khan, an official of the
Finance Ministry failed to respond to (i) above, and pitched his defence
in terms of (ii) only. In this rejoinder I shall propose that his claims
of a credible estimate with respect to the comparability of the large
sample of the year 2000 with the small sample of the year 2004, are incorrect.
However, regardless of this debate on the veracity of numbers, an important
proposition needs to be made with respect to the right of citizens to
question the official truth on the basis of logic. Dr. Khan begins with
an aggressive tone when he asserts in effect that no one has the right
to question official fudging. I should like to remind him that the right
of a citizen to question the official truth on the basis of reason, was
not granted by the Ministry of Finance but is a fundamental right that
was won by Socrates in the year 326 BC when he drank the portion of hemlock,
rather than desist from questioning the official truth. I shall therefore
persist in exercising this right as a citizen of Pakistan and the world,
regardless of the threats and bluster of officialdom.
In the context of the government’s poverty reduction estimate
of 4.2 percentage points, the issue is really one of illegitimate approximation,
with respect to the principles of economic science. It is clear that this
figure emerges largely from the differences in sample design between the
PIHS sample of 2000-01 and the much smaller sample of the year 2004. Consider.
(1) It is incorrect to say as Dr. Khan claims, that the 19th April to
6th May data in the small sample survey of this period in 2004 is being
compared to the same period for the year 2000. The reason is that the
official poverty line of Rs.748.6 per adult equivalent per month (subsequently
inflated to Rs.848.798) for the base year 2000-01 was estimated on the
basis of annual data and therefore had averaged out the seasonality factor
in the expenditures of the poor. The data for 2004 however is based on
the harvesting period only when the poor can spend more from harvesting
wage income. Consequently poverty levels in the year 2004 would be understated
compared to those in the year 2000, thereby yielding poverty reduction
over the period as a statistical bias rather than a description of reality.
(2) It is also incorrect to say that the “methodology” with
respect to provincial representation is the same in the large and small
samples respectively. The reason is that in the case of the small sample
survey, which Dr. Khan admits covers only 5046 households (one third of
the large sample), the stratification at the provincial level would be
so small as to have no statistical significance. By contrast the large
sample survey would be meaningful in terms of provincial coverage. This
is the point. The small sample survey is likely to leave out a much larger
proportion of poor people than the large sample survey.
The problems of comparability between the large and small sample survey
results therefore are inherent in the differing degrees to which the two
samples are representative of the underlying population. Therefore a poverty
reduction estimate (4.2 percentage points) based on such a flawed comparison
is misleading. In the Oxford English Dictionary the meaning of the word
fudge (as a verb) is as follows:
“To deal with something in an inadequate way, specially so as to
conceal the truth, or mislead.” The New Oxford Dictionary of English,
Oxford 1998, Page 740. What Dr. Khan has done in proffering the official
truth on poverty reduction qualifies precisely for the word fudge (verb).
Regardless of the fact that Dr. Khan’s claims on poverty reduction
do not stand up to scrutiny, if he wishes his claims to transparency to
be believed then he should do the following: (i) Let the Federal Bureau
of Statistics place the raw data for both samples prior to its cleaning
by another agency, on the FBS website. (ii) Specify the basket of goods
on the basis of which the price index was used for the year 2000-01 and
also the official estimation procedure for the inflation factor. (iii)
Specify the official estimation procedure for the poverty line figure.
(iv) Specify the number of households covered for each province in the
2004 sample, and the number of households covered in each province for
the April-May period in the 2000-01 sample. If such information were made
available on the website, independent economists and competent civil society
institutions could do the estimations themselves and be able to judge
whether the Finance Ministry’s figures are right or wrong. In any
case what is even more important for resolving this important policy question
is for the PIHS to be allowed to conduct the July 2004 survey independently,
with integrity and without fear or favour.
|