There are two big problems with this study. First, their data selection is skewed and relies heavily on countries with longstanding cultural problems (besides

There are two big problems with this study. First, their data selection is skewed and relies heavily on countries with longstanding cultural problems (besides inequality) affecting their economic growth. Just look at the list of countries constituting the data for the third chart. About half of the countries are from Latin or South America too, so they haven't controlled for geographic similarities either. http://www.imf.org/external/pubs/ft/fandd/2011/09/Berg.htm http://www.imf.org/external/pubs/ft/fandd/2011/09/Berg.htm Second, their perspective throughout the study is "top down" and makes the bold claim that data showing weak correlation is grounds for supposing strong causation. The editorialized headline from motherjones.com also makes the same mistake. I'm not too familiar with regular IMF publications so I have no idea if politicization is common in them, but this certainly seems like a classic case of choosing the data that confirms what you already believe. EDIT: If you want to see a more complete dataset go here and sort descending by the "UN Gini as a percentage" column (the same metric used by this study): http://en.wikipedia.org/wiki/List_of_countries_by_income_equality http://en.wikipedia.org/wiki/List_of_countries_by_income_equ... The top of the list (which the authors of the study draw from liberally) mostly consists of countries with severe economic and cultural problems in many areas. The rest of the list doesn't have a clear one dimensional trend. Some countries (e.g. Greece) with severe problems have a greater level of equality than the US, while other more stable and consistent economies (i.e. Scandinavian ones) also have greater equality. As an example, if the authors of the study wanted to apply their methods to measure the effect of melanin on a country's growth they could just as easily conclude that a population with dark skin results in less growth. While the absurdity of the correlation=>causation fallacy are apparent in that case, it is less so when you cater to widely held beliefs that are easier to support without looking at the data thoroughly.