« Help, Help! I’m being repressed!
R: solving a small system of equations fast »

Inequality Granger-causes growth

February 18th, 2011

So say Andrews, Jencks and Leigh. For what that’s worth.

This entry was posted on Friday, February 18th, 2011 at 5:06 pm and is filed under growth and progress, inequality.

4 Responses to “Inequality Granger-causes growth”

  1. Tweets that mention The Ambrosini Critique » Blog Archive » Inequality Granger-causes growth -- Topsy.com Says:
    February 18th, 2011 at 5:50 pm

    [...] This post was mentioned on Twitter by Economics, Will Ambrosini. Will Ambrosini said: Inequality Granger-causes growth: So say Andrews, Jencks and Leigh. For what that’s worth. http://goo.gl/fb/GLpRr [...]

  2. Scott Sumner Says:
    February 19th, 2011 at 5:51 pm

    I’ve never understood why people find Granger causality to be interesting. Isn’t it likely that stock market booms lead to more inequality, as capital gains are skewed toward the rich. Would anyone be surprised to learn that higher stock prices predict faster growth in the following year? Would anyone then assume causation from stocks to growth? I haven’t read the paper, so perhaps they control for that problem.

  3. pushmedia1 Says:
    February 20th, 2011 at 12:35 pm

    Yeah, they use Piketty and Saez’ top-income tax data, so this could be an issue. I think they only control for education and tax rates. I think the idea with Granger causality (btw, they never use that term in the paper) is there is a sort of argument from exhaustion: if you control for everything you can think of and you still get granger causation then, well, it must be real causation.

  4. Scott Sumner Says:
    February 21st, 2011 at 2:38 pm

    Pushmedia1, Yes, but people need to think about what they are assuming. It’s obvious that stock prices predict growth. If stock booms are associated with greater measured income equality, that will also predict growth.

    It’s not a question of whether people went through the motions and did everything the right, way, the question is whether the hypothesis being tested actually implies any causation.