Saturday, October 9, 2010

Notes on the yield curve

What I look for is underneath the yield curve, the unquantized version under infinite dimensionality; hydraulic macro. I look at it as a normal frequency spectrum. What do I see?

The spectrum really peaks toward zero as it moves toward the long term, shaped like a production spectrum. At ideal equilibrium, the curve is Bell shaped, peaking around ten years.

Quantizing that curve put most of the long term into one 'bit'. The amount of the long term curve that trails over, across the Y axis, through zero, that portion is very small in good times, the curve is either peaked out near 7 years or we have rapid gains from sudden discovery and we are peaking close to 5. When the curve is flatter, more of it spills into negative territory.

This is standard, square integrable, 'signal processing' view.

No comments: