Redistribution Policies and Economic Growth

This post is a little off the beaten path for me, but I heard a presentation today that bugged me so I decided to dig a little deeper. Scott Hodge of the Tax Foundation spoke at the Economics Club of Indiana luncheon re: taxes, equality and the future of America . He popped up a chart of two "societies", one of which had a comparatively even distribution of wealth and one of which had a more skewed distribution. He polled the audience to ask which society they would rather live in and revealed that the even distribution society was Romania in the 1950s and the skewed society was the Forbes 400 list.

He then concluded that more even distributions mean that everyone will be equal but poor, but skewed distributions show societies that reward risk taking and provide economic growth. Now, I've never been a flat-taxer, and while I don't like paying taxes I do favor progressive taxation. But I'll throw all of those ideas out the window if it means a better quality of life. So, his claim got my attention - are redistributive policies and/or societies correlated with a lower quality of life as measured by income or economic activity?

Below is a quick chart taken from the CIA's World Factbook that shows the GINI Coefficient (i.e. a measurement of income distribution) against per capita Gross Domestic Product. Hmm? Clearly, there are some societies at the extremes (i.e. high per capita GDP with a low GINI Coefficient and vice versa). But what about those societies that get some of both (e.g. the United States has a per capita GDP of about $48,000 and a GINI Coefficent of about 45, Hong King and Singapore are the two dots just above the U.S.)? The correlation between per capita GDP and the GINI Coefficient is -.42. Interesting info, but clearly a society can have a) very little per capita GDP and both a high and a low GINI Coefficient and b) relatively high per capita GDP such as the United States and a moderate GINI Coefficient. It reminds me of the mantra in my econ stat class - don't confuse correlation with cause/effect.