The 80/20 rule
I do not understand people concerned with inequality of wealth distribution.
So what if 20% of Americans have 80% of "wealth"? The world-wide average of the GDP share is 82% for the top 20th percentile. About 20% of peapods contain 80% of peas. 20% of beer drinkers consume 80% of beer. And so on.
This is the Pareto principle (generalized as the p/(1-p) law) that has great universality, like other empirical laws related to power distribution (Gutenberg-Richter’s law, Beford's, Zipf's). If you do not like such laws, you better start with peas in peapods and discover precisely why such laws apply in all kinds of systems: from WWW sites to protein bank networks to citation graphs. Take any "natural" network and you will find a power law distribution with exponential cutoff. What is there to protest? It is like the popular uprising against the normal distribution.
I know a few physicists studying networks, and their common belief is that this behavior is caused by some basic properties of stochastic processes in nature, like the classical Simon's network-growth model. The sheer ubiquity of such behavior suggests the enormity of the misguided ambition.
BTW, it seems that 80% of such concerns are generated by 20% of people...
PS: Terry Tao had a nice post
http://terrytao.wordpress.com/2009/07/03/benfords-law-zipfs-law-and-the-pareto-distribution/
illustrating why the three universal statistical laws must be compatible with each other.
So what if 20% of Americans have 80% of "wealth"? The world-wide average of the GDP share is 82% for the top 20th percentile. About 20% of peapods contain 80% of peas. 20% of beer drinkers consume 80% of beer. And so on.
This is the Pareto principle (generalized as the p/(1-p) law) that has great universality, like other empirical laws related to power distribution (Gutenberg-Richter’s law, Beford's, Zipf's). If you do not like such laws, you better start with peas in peapods and discover precisely why such laws apply in all kinds of systems: from WWW sites to protein bank networks to citation graphs. Take any "natural" network and you will find a power law distribution with exponential cutoff. What is there to protest? It is like the popular uprising against the normal distribution.
I know a few physicists studying networks, and their common belief is that this behavior is caused by some basic properties of stochastic processes in nature, like the classical Simon's network-growth model. The sheer ubiquity of such behavior suggests the enormity of the misguided ambition.
BTW, it seems that 80% of such concerns are generated by 20% of people...
PS: Terry Tao had a nice post
http://terrytao.wordpress.com/2009/07/03/benfords-law-zipfs-law-and-the-pareto-distribution/
illustrating why the three universal statistical laws must be compatible with each other.