[PLUG-TALK] Expansion, growth, the world, and time: redistribution
keithl at kl-ic.com
Fri Mar 25 08:03:17 PDT 2011
There are three things to reply to here. I will split the response
three ways, but all have to do with the nature of actual, choice
making human beings, as opposed to boring hypothetical statistical
units. The first is an essay I wrote in January for another list
about distribution of wealth:
It is certainly debatable about what the distribution of wealth
ought to be. Such debates should be informed with observation,
and one observation about the behavior of people is that many
of the measurable things they do tend to follow a mathematical
description called "power law". Power as in a variable taken to
some power or exponent, not as in who gets to tell who what to do.
Height, intelligence, years of education, etc. tend to follow
normal distributions, bell curves. Other measures, like wealth,
friendships, book writing, book sales, patents, number of photos
on Flickr, modules in the Linux kernel, and long rambling emails
on mailing lists, tend to follow a power law, where a few people
contribute (or take) a lot, and a most people contribute little
or nothing. For large data sets, like Flickr, or product sales
on Amazon, or video rentals from Netflix, the curves are
surprisingly smooth, and the median and mode quite different
from the mean. You can read about this in Clay Shirky's "Here
Surprisingly, this even holds true in massively redistributive
economies. In college, when I looked at tables of Gini
coefficient (a measure of unequal distribution of income),
I was surprised to learn that countries like Albania and the
USSR had higher Gini coefficients than the US. But those are
countries where the median person lived ten to an apartment,
while the powerful have expensive cars and country homes;
upper middle class here, ultra rich there (until the next purge).
Chopping off the head of the curve often just lowers the whole
curve - it reestablishes around the same averages and medians.
Bringing up the lower part of the curve may encounter resistance -
some people prefer other activities to contributing photos to
Flickr (like writing books, or book-length emails).
There is a large component of sustained personal obsession in
making some disparities occur. The "sustained" is the important
part - the richest people I know got there because granddad was
obsessed, dad was obsessed, and son was obsessed, their whole
lives. Ten thousand bucks in 1911 became a billion in 2011.
Luck helps, smarts helps, distorted incentives "helps", but
100 years of savvy compounded accumulation is essential,
requiring an unwavering and single-minded focus that would
make most of us very unhappy.
I also learned that much disparity is related to age, and
compounding investment. Person A starts working at 7-11 at
18 and buys two packs of cigarettes a day. Person B also
works at 7-11, but saves 5 bucks a day by not smoking,
diligently investing for 7% compounded growth. Person A dies
of lung cancer at 60, with a few assets sold to pay for medical
expenses. Person B dies at 83, still saving 5 bucks a day from
their social security check, and leaves a fortune of two million
dollars. Small behavioral differences, accumulated, can have
large outcomes over time. The behavior is rare enough that it
still makes headlines: "Janitor leaves millions to school."
Is that an argument for a widely disparate distribution of
wealth? No, just an explanation for part of it. Given that
knowledge, programs can be designed to leverage off the
observation to improve what can be improved. Perhaps, when
everybody is not smoking and saving $5 a day, billions of
people will die millionares. And there will still be a
few ultra rich people, using their trillions to fly their
personal spaceships to Mars.
Many of the rest of us will still resent the rich. That is
one of those bell curve distributions. A few will attempt to
assassinate them, with guns or with slander. That's power law.
Keith Lofstrom keithl at keithl.com Voice (503)-520-1993
KLIC --- Keith Lofstrom Integrated Circuits --- "Your Ideas in Silicon"
Design Contracting in Bipolar and CMOS - Analog, Digital, and Scan ICs
More information about the PLUG-talk