Subj:    More assumptions in economics

Date:   28/12/02



File:     D:\Current articles\The Wealth of Persons 28 Dec .doc (553984 bytes) DL

Time (32000 bps): < 5 minutes




Dear Professor Pogge


Thank you for all the work you did with Sanjay Reddy and Howard Nye on PPP.  I

think this is the most sensible (and so important) piece of work on development

statistics for a long time.   


I'm working on a broad range of problems in economics, which I think you may be

interested in.    It stems from my reaction to the paper "Growth is good for the

poor" which I thought had several fundamental mistakes.     I have been amazed

to find that economists ignore them.   


The implications of the flaws in PPP are wide, as perhaps you know.    They

invalidate the findings of the World Bank on a range of issues including growth

and poverty, and on the claimed success of their policies.     So in fact do

other considerations.    One is the unreliability of data;    the other is the

fact that per capita income statistics do not measure the average income gain

during a period, let alone the average financial gain during a period, the

average lifetime income gain, the average financial lifetime gain or the average

welfare gain.     These are facts which economists and economic theorists have

not faced, though some have raised concerns which imply them.    


My own training is in classics (including philosophy) and experimental

psychology.    I've also lived with a typical family in a Bangladesh village. 

  One of my first reactions to the World Bank work on growth and poverty was to

think "the results might be due to poor people dying earlier in some countries

than others".    Economists have not thought about this problem  -   average

consumption goes up if hungry people die earlier (and the proportion of poor

people goes down)  -  in any depth.    


I don't know how much you know about the politics of this situation, but it is

fairly alarming.    Last year I spoke to Eric Swanson, who is the head of the

World Development Indicators project at the World Bank, about the differential

mortality problem in the context of AIDS.    He said that people were better off

after the plague because the density of people was less on the land;  that he

was thinking on a macro level, about the long term;   and that they could find

out how many people died rather than had raised incomes but he didn't think it

would be a good idea to say it.    Last September I had a long conversation with

Robert Johnston, Chief of Statistical Services at the UN.    I told him some of

what Mr Swanson had said, and he laughed, saying "well, Larry Summers tried to

do that with AIDS when he was at the Bank".      We had a chat for 70 minutes,

during which he said that it was suspect that in World Bank results no country

had growth without poverty reduction.   I asked him what he meant, and he said

that he thought they wanted to show a certain result.    I was left with the

impression that the research was selective.     He also said that many

statistics emerge with only "source:  World Bank" when the way the statistics

had been obtained was impossible to find out.   


I seem to be the only person who wants to publicly challenge the idea of

reducing the proportion of poor people.     I also seem to be the only person

who says that per capita income does not tell us average gain.    There are

several inferences involved in economists' conclusions about average gain.   

Your work on PPP has helped me to see perhaps the last piece of the jigsaw.    


I have sent various versions of the document I attach here to philosophers (Dan

Hausman, John Broome, Dan Little, Alex Rosenberg) a political scientist and

expert on statistical methods (Gary King at Harvard), a writer on assumptions in

economics (Uskali Maki at the Erasmus Institute in Holland) and the secretary of

the Official Statistics Section of the Royal Statistical Society (Ray Thomas of

the Open University in the UK).    From conversations, Dan Hausman thinks that

what I have to say is important and John Broome agrees that average gain can't

be inferred from a rise in per capita income;  Gary King says it is undeniable

that without information on mortality rates we don't know average gain.    Ray

Thomas, after reading some of what I've written, thinks that what I say about

development statistics and poverty is right.   


My document makes various fundamental points about economics:


1)  per capita statistics are not utilitarian:   the definition of "utility" in

economics is different from that in utilitarianism (an original point I make but

entailed by what John Broome has written)

2)  "poverty reduction" is a misnomer for current official statistics (entailed

by Sen's criticisms of the headcount measure of poverty)

3)  statistics which jumble up adults and children do not measure either wealth

or poverty.   


The default option (which assumes zero effect of demographic change on

statistical averages and proportions of poor people, and so an exact

correspondence between these and consumption gains) is untenable.     The fact

that economists are allowed by their training to assume it is valid is

ludicrous, partly because it allows them to include China (with its one-child

policy) in analyses of policies and poverty.     There are some very serious

issues for economic theorists to take on board.    I also note in my document

the comment of Angus Deaton:  "Because poverty counts come from the survey data,

while growth measures come from the national accounts, and because they are

evidently measuring different things, there is no consistent empirical basis for

conclusions about the extent to which growth reduces poverty. That economic

growth, as measured, has at best a weak relationship with poverty, as measured,

means little more than would a finding that growth in China had failed to reduce

poverty in India"   (Counting the world's poor: problems and possible solutions,

Research Program in Development Studies, Princeton University, revised version,

December 2000).


There is also the problem for economists that world growth has obviously not

been good for the world poor.


I would be interested in any answers you may have tin the following areas:


1)  Table 4 seems to show that the proportion of world poor has "fallen" due to

the change from 1985 to 1993 prices.  How does this square with the idea from

Martin Ravallion that the poverty lines did not change overall?

2)  There may be some selection bias due to the availability of income

statistics.    Countries which are undergoing crises may not be able to provide

statistics for those periods.   

3)  The PPP system may distort the real purchasing power of the local currency

purely because it excludes internationally-traded goods.   So it may ignore the

lack of purchasing power of the well-off for those goods.   This may show the

rich in poor countries to be more rich than they really are compared to people

from rich countries.   

4)  I would be interested to know any other reasons to think that the rich in

poor countries may have had their purchasing power underestimated.  

5)  In recent decades maybe the changing consumption patterns of the "emerging

nations" and not just that of the rich nations may have influenced the pattern

of goods and services taken into account by the PPP system.   

6)  Countries which trade more may have more foreigners buying luxury goods and

services.    So the prices of luxury goods and services taken into account by

the PPP system may change accordingly (up or down).

7)  There will be more significant statistical effects (in studies on growth and

poverty) where luxury goods have more of a share of the PPP calculation and the

luxury goods change more in price.   Perhaps there are identifiable examples. 


My own conclusion is that a monetary measure of poverty is both unnecessary and

insufficient for the purpose.    Consumption expenditure is not fundamentally a

measure of poverty, in the sense of inability to achieve a minimal level of

physical and social functioning.    In agricultural economies, prices fluctuate.

   I think that the problems of dealing with these fluctuations may be

insurmountable.    The numerical value on the coins is fundamentally irrelevant

to whether people have a minimally acceptable standard of living.    Food-based

PPPs are certainly in the right direction, but I don't really see the point of

converting calorific content to money and then converting money internationally.

   For malnourished people (who are the majority of those classed under poverty

lines for all poor countries) I suggest that survival rates are the most

sensible measure.  


I have done this work in my spare time.     There is more:   a review of the

Millennium Goals as well.    Insiders such as Robert Johnston and Robert Mayo of

the FAO, among others, tell me that the goals were chosen by politicians rather

than by social scientists.   


I have tried to ring you at Columbia, but the staff would not tell me where you

were!     The sections of my document about PPP are being revised, because the

implications are filtering through to my analysis.   As it stands, I've come up

with some axioms about national (or notional) PPP rates.     The

logico-mathematical relationships are quite complex, and it needed a detailed

reading of your paper to see that third-country prices influenced

individual-country poverty analysis.  


I appreciate your having brought to the debates a measure of sanity which has

been lacking.  


Happy new year!


Matt Berkley