The wealth of persons: Summary

Introduction

These pages are designed to promote thought about national and international statistics claiming to show welfare trends - aggregate welfare trends for real people.

The inquiry began in spring 2000, with the publication of the document “Growth is Good for the Poor” from the World Bank. To my astonishment, it became clear that two flaws - the life-length flaw and the inflation flaw (see below) - were hardly ever discussed, were not requirements of the economic method, and applied to all large-scale studies claiming to show trends in poverty, inequality and wealth. They apply to all large studies from economists claiming to measure the impact of policies on poor people - as does the problem that adults need more food than children.

One aspect of demographic change is changes in life length.

All measures of average wealth, and income ratios between rich and poor, and income gradients between rich and poor, and of poverty, will go in the “wrong” direction if poor people live longer. The effect of this demographic shift - or its reverse - on economists’ statistics is not known. The tradition in economics is to count most people’s survival as a “negative” influence on average income, with the erroneous implication that people had worse outcomes. (Most people have below-mean incomes).

There is a second problem about life length for economists who wish to infer aggregate welfare gains to people from statistics about the economy: they have assumed that an extra year of life is of zero value to people. This is a second fundamental error in the theory of welfare economics. The first problem is concerned with the statistical effects of differences in life length. The second problem is not a scientific problem, but a moral one - in other words, a matter of opinion. The question here is not “what are the mathematical effects” but “what do we think a year of life is worth in money terms?”.

Suppose you compare two countries where life length differs, and you know the annual economic position of the people is the same. The overwhelmingly dominant theory of welfare economics in 2003 says that the people in both countries have equal welfare. Supposing economic welfare is all there is to welfare, so the statistics really do tell you about annual “happiness” or whatever we call it. Which country would you prefer to live in?

If life is better than death, you are better off in the country where you live longer. So the theory of welfare economics is defective. Can it be remedied? Well, not really - only by introducing a valuation of lives which is clearly subjective. Why is it subjective? Because if in country A people live a bit longer, and in B people have a bit more “happiness” or “welfare” per year, where people are better off overall is a matter of personal opinion. It can’t be anything else.

Quite a common theoretical idea in economics is to measure lifetime income. But that values a poor person’s life at less than a rich person’s.

The problem of how to value a year of life, which has no objective answer, is a clear case where you can’t use numbers objectively to assess welfare.


These pages document some flaws in the use of economic statistics on poor people - and people in general. Some of the flaws are in the theory behind practical work, and some are in the theory of welfare economics.

The expensive and complex solution to measuring outcomes for the world’s poor people is to use economic data. That solution would be so complex that it is obscure to outsiders, and thus very vulnerable to distortion in public relations exercises by governments. Reliable data do not yet exist even for the income part of the equation, let alone for the expenditure part.

The cheap and simple solution is to measure how long they live.

Notes: This is work in progress, and most articles are in draft form. Part of the reason is that the issues are simple in isolation, but complex in combination. More detailed material will appear later. I apologise in advance to economists and statisticians for any distress which my remarks may cause. Others are far more knowledgeable, wise and adept than I am. My chief virtue, if any, is inquisitiveness. See below for a brief introduction to two fundamental issues in economics.

The World Bank and the life-length flaw

The headcount index is, in theory, the proportion of people in poverty. The World Bank has consistently claimed that a faster fall in the proportion of poor people - what is called in the jargon “poverty reduction” - shows that poor people have done better. Does it?

Here are the views of two people. The underlining is mine.

1. “A seminal paper by Sen (1976) drew attention to the undesirable properties of this measure, such as the fact that when a poor person becomes poorer the headcount index of poverty will not increase (indeed, if the person dies, the index will fall!) .”

- Martin Ravallion, co-designer of the methodology for the World Bank global poverty counts, 1996.

That sentence is available on page 2 of the original and page 6 of the electronic document at http://econ.worldbank.org/files/13469_wps1615.pdf .

2. “Many millions of people around the world are going to die from AIDS, with untold misery and deprivation. It would be a terrible thing if the Bank dealt with this only in terms of its effects on income poverty, which it already shows some signs of doing. ...(... If per capita income were to rise as people died, would that make the impact of AIDS somehow less severe?) My own view is that the Bank should be backing away from its current too concentrated focus on the income headcount numbers...”

- Professor Angus Deaton of Princeton University, “Counting the World’s Poor: Problems and Possible Solutions”, December 2000.

The document was written at the request of the Chief Economist of the World Bank, Nicholas Stern. The passage above is at the end of the document, available at http://www.wws.princeton.edu/~deaton/downloads/worldpov3b.pdf .

Economists and averages

If average income rises, does that mean that people had economic gains? Not necessarily.

[The change in average income] is a function of [the average income rise or fall] and [demographic change].

There is thus perhaps a need for economists to separate out these components of statistics on average incomes.

At present, we do not know the effects of changes in life length on average income statistics.

However, we do know that in the past certain events have caused poor people do die in large numbers.

[The rise in the average] is not the same as [the average rise for people] during a period.

A rise in the average is consistent with an average fall for people, and vice versa.

In respect of social scientists’ statements about average financial gains or losses, it is a failing of welfare economics that it counts as “better” for real people any rise in the average, which does not follow. A consequence of the same underlying flaw is that welfare economics counts all decreases in income ratios between rich and poor, other things being equal, as showing better consequences for poor people. This does not follow either.

In respect of social scientists’ statements about welfare gains and losses, it is a second failing of welfare economics that it counts longer life as of no value.

The value of life is a matter of taste.

It is a matter of opinion.

It is a matter of philosophy.

It is a matter of arbitrary choice.

Therefore, any measure of welfare which combines life length with a measure of well-being of people at different times would be arbitrary and subjective.

That is one reason why you can’t just add up numbers and say how well or badly people’s lives have gone.

How do you combine a measure of the quality of life with its quantity?

You can’t do it without putting in your own idea of the value of a year of life.

There can be no numerical solution to the life-length flaw in welfare economics which is not subjective.

That flaw is not the only serious flaw in the theory of welfare economics. The theory behind all large studies of “poverty” or “inequality” in most countries has another flaw: the inflation flaw. Unless you know about price changes for items which poor people buy, you do not know from income statistics whether they gained or lost. Large international studies from economists on most countries in the world do not take price changes for poor people into account. They therefore cannot tell us about economic gains or losses, or about welfare gains or losses.

In the case of hungry people, who are according to official statistics the majority of those targeted by the World Bank, their economic success is self-evidently equivalent to nutritional success. Nutritional success for hungry people can be measured in years. It is a simpler, more transparent, cheaper and less corruptible measure of outcome.

Email: matt@mattberkley.com