Kocherlakota and the math problem

Yet again.
One can, if mandated by Congress, create a relationship between inflation, unemployment and the yield curve. But, if you want that relationship to show common sense, use fixed signal to noise, not minimum noise, when optimizing the fit.

For example, after a period of over expansion, the production spectrum, we say, is expanded relative to inventory flow. Inventories are tight (have large variance relative to inventory level). So the Fed wants to signal a deflation, which it does by extending the period of its shortest target, and raising the rate. The Fed needs to signal a subtle shift in transaction rates, because we want the real goods to slow their transaction rates and bring inventory variances back to norm. Same with too few skilled employees, the variance of employees to level is too much.

So run the VAR model with a different norm, use constant SNR, or in a closed economy, constant variance in inventories. If we did this, then monetary policy will always give a positive definite result and match the real yield curve without alias problems. Essentially we are "compressing" the yield curve to fit into a specific SNR measure, we compress it by emphasizing precision on the most active terms, and reducing precision on the less informative terms.

The Congressional mandate relates employment, inflation to short term rates. When the variance of surplus labor rises, then lengthen the integration window to bring SNR back. Those integration limits get the sample period where there is low hanging fruit. Do that for both variables, and weight them as you please. The result is a change in both period and yield as the computed result.

I would suspect the math would be even easier by using nominal growth vs actual growth; rather than the Congressional mandate.

Go back and read some Zero Hedge posts where Morgan Stanley plays the curve alignment game, getting into a race to twist the curve closer to maximum entropy.

The real zero bound, or if we wish, our ignorance limits, will come from quantum effects, when actual growth is so below potential that the central banker should not even exist. At some point the underlying production spectrums cannot find solutions among the finite solution set. That is default time.

Kocherlakota's main point is that the math is out of balance, and cyclic. He is right.