Superpost: Thoughts on inequality

The GINI Coefficient is a measure of inequality of income distribution. The more unequal a society distributes its wealth, the higher the GINI figure will be. Let me decribe this further.

Let's say that countries X and Y each produce an income worth 100 dollars a year. Let's also say that both countries are 10 people big. Country X distributes all 100 dollars to 1 person. Country X would have a high GINI figure. Country Y, on the other hand, equally distributes its money (all 10 people receive 10 dollars). Country Y would have a low GINI fingure.

The NY Times blog, Economix, has a post about the implications of having a high GINI index. On the one hand, long-standing conventional wisdom was that "efforts to reduce income inequality would be punished by lower economic growth". On the other hand, "Richard Wilkinson’s new book, 'The Impact of Inequality,' summarizes evidence from comparative studies that violence is greater — and trust and cooperation lower — in more unequal societies."

What does this have to do with the efficiency argument noted above? The post author notes that higher violence rates may lead to a higher need for private gaurds and private gaurds are an inefficency. They then post the graph on the right (I circled Miami).






There are a few pieces of important information provided here. One, Miami has an oddly high GINI index (over 63%, which is similar to the indices of countries in the developing world), which is to say, Miami has a highly unequal distribution of wealth.

Two, Miami also has a large number of private gaurds as a percent of total employment. Again, according to the post authors this represents an ineficiency (and this makes sense since guards don't "produce" much, and they definitely don't "produce" anything beyond what law enforcement does).

Three, the dispersion of the points on the graph show a pattern of points being lower on the left and higher on the right. Typically, this indicates that there is a relationship between the two variables. Here, it shows that, for US cities, a higher GINI index correlates with more security.

If you have read this far congratulations. But, there's more. One, doing an analysis with only two variables present is never the best design. Two, the notes on the statistics say that the effect of one of the variables explains 64% of the variation in the other. That means that 36% of the variation is being caused by other things (and this is quite possibly lower than reality since only two variables are included in the model).

The point I am trying to make is that understanding this information is difficult. It is, however, important to understand statistics. The media ignoring complicated models (or uncomplicated one like this one) does not serve anyone.

I am also saying that Miami is scarily unequal.