Sunday, November 7, 2010

A simple hydraulic model of QE

Ben and friends are making a bet that the economy will absorb a certain NGDP change over the next eight month period.  The  US Treasury makes money if Ben's view is more correct than the bond market's view, Ben loses otherwise. What if Ben loses, say he loses $50 billion in the trade?  Where are those losses?  They were not absorbed into the economy, they are most likely being passed around as gains to traders and profits to bankers.

When Ben loses, then he is making the curve steeper than the bond market expects, risk off.  If Ben wins then the curve is flatter, risk on.

What is the on/off frequency? Average about one quarter, since the target is two quarters.     Nyquist tells me that traders guess the best at a period twice the sampler.  So the banker curve is going to rock up and down, mostly on a quarterly basis (the sampler). If the curve spends too much time in risk off [steep] then  traders jump in to grab obvious gain, forcing risk on for a while.  If Ben is making money, then the one year yield should jump a bit, making the curve spend more time in risk on.

I love this game!  We need this game on the market.

No comments: