Lies and Statistics
Author: Greg Price. Greg Price is a training consultant in professional services.
Date: 7 January 2006
Publication: The Australian Financial Review
The way income is distributed within society is not going to change,
writes Greg Price.
You might think policy makes a difference. But judging by the work of
Victor Yakovenko, there are factors that override government policy,
at least in terms of income distribution. Yakovenko is professor of
physics at the University of Maryland and late last year spoke at the
Econophysics Colloquium at the Australian National University.
Yakovenko has published a paper tracking income distribution in the
United States from 1983 to 2001. He analysed the distribution of
nominal dollars (unadjusted for inflation) using tax data, and found
it has stayed very close to the same exponential curve. That's
astonishing considering the survey period covers four presidents,
several recessions, variable rates of inflation, two Federal Reserve
governors and varying policies in areas such as tariffs and trade.
Over the same period, Yakovenko has tracked average income and gross
domestic product per capita, finding that they both roughly doubled in
nominal dollar terms, and that their upward trends followed the same
steady upward slope. Factoring in the consumer price index, he finds
that 85 per cent of the increase in nominal income was due purely to
Yakovenko has also tracked income inequality using an index called the
Gini coefficient. A Gini of zero indicates an even distribution of
income and a coefficient of one indicates maximum inequality (one
person in a population has all the money). He finds that inequality
has tracked very close to an index of 0.5 for individual income and
0.375 for household income. This correlates with the indices for other
developed economies such as Britain and Australia. Developing
economies have higher and varying rates of inequality. Only Soviet
Russia had lower inequality (albeit with much lower real wealth).
Yakovenko believes he has uncovered values for income distribution and
inequality that are signatures of a developed economy. They are
equilibrium values. That is, a developed economy tends to run itself,
at least in terms of wage and salary income. Policy will at best make
a dent in the equilibrium distribution. For example, his recent
Australian study shows some clustering around values that he believes
represents welfare cut-off points.
The distribution of income from stocks and capital gains is even less
amenable to management. Income distribution among the rich (above five
times average income) follows an equation known as a power law, which
gives wildly unpredictable results. The number of people living off
stocks and capital gains rose from 1 to 3 per cent of the US
population over Yakovenko's survey period, peaking in 2000 with the
stockmarket boom and then falling. The amount of total income in this
stock bubble also steadily increased, from 4 per cent in 1983 to 20
per cent in 2000. But this was not enough to disturb the shape of the
income distribution curve among wage and salary earners.
Politics, they say, is the art of the possible. If so, then
politicians (and their economic advisers) can stop trying to find ways
to significantly change the distribution of income. It's not possible,
at least, not in a mature free-market economy.