James Hamilton Estimates a Phillips Curve

I guess I just have a problem with regression based on YoY % changes in variables. The alternate specification for the regression is to look at log changes in CPI (the index, that is). So that the next period log change is a function of the unemployment rate and prior changes in the inflation rate. If you do this regression with 6 lags on each variable (including the present value), none of the unemployment variables are significant. If you do the unemployment in terms of differences, only the second lagged difference is significant.

A better approach is to just model core inflation and even then, use an average of different output gaps (GDP and unemployment). I model MoM changes in core inflation on two of its lags, the lagged inflation gap (relative to 10 year inflation forecasts from SPF), an output gap factor, and a central bank factor.

Reading an account of the hyperinflation in Germany during the 1920's, I was struck by the fact that foreign exchange and the stock market provided the best hedges against the inflation.

If the overhang of government debt is going to keep inflation above 2%, why isn't that showing up in long term treasury yields? The spread between long term and short term yields has actually fallen.

Simon,
The market and I clearly disagree. If the market is right, we are going to have low inflation.

Simon,

One may as well ask why 2-5 year yields so high 2 years ago when we were about to have deflation.

Yancey - That's a slightly different question because its retrospective. We know what actually happened. 2 years ago IIRC it was rational to expect inflation. The fed agreed with the bond market and they both pushed rates up, right?

Right now the key question is surely whether the fed would collaborate with the treasury and help monetize the debt. Right now it looks like the market believes the fed's assurances that that won't happen.

I think a better way of looking at it is seeing measured inflation as a symptom. There are a variety of things that can cause it. Anything that artificially raises the cost of labor will increase measured inflation as well as raise unemployment. Eventually as technology increases this will fix itself through economic growth.

Anyway, I suspect the reason most recessions are deflationary is because during recalculation, you have to switch to liquid assets in order to make your switch.

Anyway, this leads me to a prediction: If the health insurance mandate passes the house, we will have a recession within 2 years of the mandate becoming effective. If this is a employer mandate, it will be an inflationary recession. If it is an individual mandate, it will be a deflationary recession.

@Simon:

"If the overhang of government debt is going to keep inflation above 2%, why isn't that showing up in long term treasury yields? The spread between long term and short term yields has actually fallen."

1) Possible reason, the fed is buying t-bonds (is it? I don't know.)
2) Due to the change in reserve requirements US banks are legally having to buy more T-bonds.
3) Due to the increase in bank reserves, US banks are buying more T-bonds, because they are legally required to.
4) China is experiencing an even worse inflationary risk than we are.
5) Now that Morgage backed securities and other used to be AAA rated debt isn't considered as safe anymore, institutions that are legally or contractually required to buy AAA rated securities are back to buying T-bonds.

Anyway, uing T-yields to try to predict inflation isn't a very good way of doing things, as the T-bond market isn't exactly a free market. Lots of institutions are required to buy T-bonds in one way or another.