Re: HDI\PPP Michael,Ulhas and Michael

[Daniel Davies <dsquared@xxxxxxxxxxxxxxxx>]

I am sure I would not be the only one who would be horribly disappointed if this was the last of your posts in this series.

 

dd

-----Original Message-----
From: PEN-L list [mailto:PEN-L@xxxxxxxxxxxxxxxx]On Behalf Of Paul
Sent: 29 July 2004 01:49
To: PEN-L@xxxxxxxxxxxxxxxx
Subject: HDI\PPP Michael,Ulhas and Michael

[See what happens with some encouragement - soon I'll be overposting!  I'll try to make this the last.]

1)      Uhlas writes:

Paul was trying to show how PPP numbers overstate the
economic growth in the developing countries. I am not
sure I understand how he has reached that conclusion.

For India, from 1992 to 2001, the GNI increased by 64% when calculated by the World Bank "Atlas" method (non-PPP).  But for the same period GNI increased by 91% using PPP!  For all low and middle income countries taken together the difference is even more extreme (22% growth vs. 44%!).  It is not just that India is made to look less poor via developed countries (which would be a one time distortion).  It is also that India (and poor countries as a whole) are made to look like they are also closing the remaining gap (a statistical "bias" because High Income countries are not affected).  [All stats from the WB's WDI.]

Furthermore, the discrepancy between the two methods grows - by as much as 4 or 5 times - during the neo-liberal period (tell me off-line if you want the chart for India, the change is dramatic).  So, there is a "bias in the bias" which shows neo-liberalism as a great success.  As I explained previous posts, this is "built-in" since the PPP model recalculates the numbers drawing from a neo-classical General Equilibrium model where market prices are assumed powerful and beneficial.  But in fact the GNI/PPP are just numbers from a model driven by assumptions...although they are presented as if they are statistics.  Then the statistics are used to prove correct the assumptions from the free market model.  [BTW there are various flavors of PPP-type models.  The Bank has chosen the most extreme version that has the most free-market assumptions.  The other versions of PPP, logically, produce lower numbers.  See below for examples.]

2) Michael Perelman writes:

If we were go[ing] to try to make some sort of quantitative measure of a human
development index, I think I [would] try to get a handle on how people at the bottom
fared rather than looking at averages.

I couldn't agree more.  One catch is that - in another little noticed development - the World Bank has withdrawn its support for calculating income distribution figures.  A quick look at the World Development Indicators shows that in most cases the last calculation is a decade old (and not capturing the radical changes of our era).  One imagines that they will soon not be published at all.  Income distribution is being replaced with the Banks own "poverty" measure.  Their poverty measure combines the (flawed) PPP with a (flawed) measure of poverty.  The two flaws combine to show great reductions in the number of the poor - a great theme of the World Bank publications recently (see comments about the Wade article below).

This is why I often suggest people use numbers like the infant mortality rate so that the plight of the poor isn't erased by progress at the top (also these numbers are more accurate than most and respond quickly to changes).

3) Michael Lebowitz writes:

In relation to questions raised by Paul on HDI, etc, a friend has directed me to a recent piece by Robert Wade in New Political Economy. I assume it's in the following issue:

Volume 9, Number 2, June 2004

Looks interesting. (It will take me a long while to read it, people interested in an electronic copy should contact me off-line).  Robert Wade (now at LSE) is often an insightful open minded liberal who is well informed (including several years working for the World Bank).  The article is mostly about the World Bank's claims that world inequality and world poverty are diminishing.  Although Wade does not spend much time on the statistics question he does make the point that the conclusions depend almost entirely on the particular choice of measurement.

I am not truly familiar with the literature on the statistical issue (anyone out there who is?).  Among the beleaguered non-mainstream economists who do write on these issues in a technical manner, I don't know anyone who has translated these issues into an applied context.  But here are some links:

a)      For a non-neoclassical critique: Columbia University economist and philosopher team "How Not to Count the Poor": http://www.columbia.edu/~sr793/count.pdf

b)      Other authors agree with neo-classical models but point out that PPP version used by the Bank (and hence internationally) has extreme free market assumptions (e.g. assuming no substitution bias in General Equilibrium models).  One such author is Steve Dowrick of Australia Nat. Univ. and has written on such issues with Brad deLong (familiar on the internet as no great critic of the World Bank).  Dowrick shows that relaxing just one of the extreme neoclassical assumptions produces a different PPP index that reverses the Banks conclusions about poverty and distribution. See
http://ecocomm.anu.edu.au/people/info/dowrick/world-inequ.pdf
and
http://acsr.anu.edu.au/staff/ackland/papers/global_poverty.pdf

C)      An "atypical" World Bank staffer is Branko Milanovic (now left the Bank?) produced a paper that accepts the PPP but revises the way other categories are shown.  This too changes some conclusions:
http://www.worldbank.org/research/inequality/pdf/Maksense4.pdf

Paul