Recently, Krueger, an economist for the Obama administration, put forward some research to tackle the issue of income inequality. The National Review and Scott Winship an Economist at the Brookings Institute jumped all over research for admitting that the simple model of how the Gini Coeffiicent, income dispersion in society, affects intergenerational earnings elasticity may contain omitted variable bias. The model addresses the question of whether societies with higher income inequality have less intergenerational wealth mobility. This is an important question because America sees itself as a country that provides equality of opportunity not equality of results. Those who have argued against government redistribution in the face of rising inequality in the US have often used the presence of sufficient income mobility as a foundation for their rationale.
The critique in the National Review is weak on a couple of levels. First, the admission of potential omitted variable is of course
standard intellectual honesty from a research perspective. The question of how significant that bias is depends on whether the relevant variables omitted from the analysis are highly correlated with the Gini Coefficient, if they are not then their coefficient estimates are reasonable. Second, using R-square as a justification for variable inclusion or ommission of a relevant variable, as the writer for NR does, is a undergraduate level no-no. Scott Winship, the critic at the Brookings Institute, promises a better model. Well, we are waiting. I am personally sceptical he is going to produce one.
Finally, those who think structural unemployment (like the Fed) and income inequality (like the National Review) doesn’t matter, I present two counterfactuals to your beliefs.
According to the Economist, 1/3 of the 3 million people who dropped out the labour force are now enrolled in post-secondary
education. Ivy League schools are hitting all time high rejection rates.
What is the market telling you?
The critique in the National Review is weak on a couple of levels. First, the admission of potential omitted variable is of course
standard intellectual honesty from a research perspective. The question of how significant that bias is depends on whether the relevant variables omitted from the analysis are highly correlated with the Gini Coefficient, if they are not then their coefficient estimates are reasonable. Second, using R-square as a justification for variable inclusion or ommission of a relevant variable, as the writer for NR does, is a undergraduate level no-no. Scott Winship, the critic at the Brookings Institute, promises a better model. Well, we are waiting. I am personally sceptical he is going to produce one.
Finally, those who think structural unemployment (like the Fed) and income inequality (like the National Review) doesn’t matter, I present two counterfactuals to your beliefs.
According to the Economist, 1/3 of the 3 million people who dropped out the labour force are now enrolled in post-secondary
education. Ivy League schools are hitting all time high rejection rates.
What is the market telling you?
RSS Feed
