Unemployment insurance and unemployment - John Cochrane
The issue in looking at reserach I discussed here, was whether state boundaries can be used effectively to note effects of federal policy changes. The answer was yes, if the error sources are properly accounted for. John, bless his heart, read the report and demonstrated the relevant methodology, and indeed the authors did do a good job.
What did they do?
Pretended to be banker bots and isolated the correlations across county borders between states. They can then allocate the correlation spectra, in proper order, and, within the given precision, identify material changes after a change in federal law. I never mentioned Marcus Hagedorn, ,Iourii Manovskii, Kurt Mitman who did the original research using state boundaries. Sorry, they did a great job.
What about discontinuities, the border is not Euler smooth? Well the human bots handle that just fine by insuring local, finite log additivity across the boundary, implicitly estimating Phi, actually. The estimate then generates adequate estimates of the two other transcendentals, just the way nature does it. But humans can use their judgement and decide that Newton's grammar can be used for a good approximation.
No comments:
Post a Comment