Interest payment crisis in DC

The US Treasury suffer a bit of volatility in interest payments.  The capital gains tax surge is drying up and the two year rate is now .53%. The ten year is at 2.47%.  The central banks fees to member banks compounds to .5%. So the debt cartel is now back in actions collecting Keynesian debt service premiums. During the swing in tax income, the interest costs vary from 7 - 13% of the federal budget, which rarely carries more than a few days cash.

The actual one year bond is .09%, so Treasury pays, over the whole yields curve, about 1% of the economy in debt cartel fees. The debt service crisis comes right where the Keynesian cycle says it should come, right after the mid term elections.

Using the right log

The debt carter counts log 2 from the reserve account to the two year:
(1+.0025)^2 . Using the natural log results in a 17% error.  The next log up from 2 is base 5. So the cartel target is 2.52%.

Anyway, the 20% error bounds after two years when using the natural log tells us how good the Keynesian model is.  Rarely are they good for longer than 2 years. The relationship etween logs comes from computing the natural exponent to some base.  It is (1+x/N)^N as N goes to infinity.  Compounding in five years chunks makes a 1/5 as a multiplier in x, and that gets e^log(1/5) = 1/5, or the finite base 5.  So, as N goes to infinity, and you are compounding five year terms, then just use five  1/5, as the base.

So, the debt cartel needs to solve for some mix of ten and two year purchases that match the one year rate, compounded yearly.  The normal method os to convert everything to natural log, then solve for a relationship between the mix of ten and two years.  The particular value that satisfy the relations come fro external conditions placed on choices. If we assume minimum redundancy, then the external condition is simply to minimize the number of trades. This leads to a -pLog(p) condition. That assumption is the same assumption in the Poisson distribution, no trades happening two or three at a time. It is also the assumption in the general Shannon-Nyquist sampling rate theory. That latter theory can be generalized to any partition using a prime number.

So, queuing theory is fundamental. The system limits flow by adding unit circles, and this keeps the queue length down but causes the network to grow by Nlog(N).