Macro model

The Fisher equation (ex post):
 Interest charges = 
  real growth (R)+ inflation (I).

First, problem:

We are getting the best equilibrium measuring over the complete monetary cycle, one generation.  We assume the transactions over the conserved index space is finite, fixed because people don't change, technology does. If the population change assumed zero (bad assumption) then the real growth in transactions over the complete sequence is zero.

Take the innovator who bets inside information, the ability to do it better and sooner. The innovator returns the borrowed money sooner than expected, actually depresses interest rates, though he borrowed slightly to excess in the beginning. But it will appear as a market making loss to the currency issuer, within tradebook error.  This is corrected.
Ultimately, there will be exits, transactions that could not compete will drop out, yielding a gain to the currency issuer. The economic dimensionality remains, small and finite.  Even when leisure expands, we assume human consumption transactions are fixed (bad assumption). Real transactions over the complete sequence is fixed, finite.

Fisher says something different then we normally assume.

Fisher describes the requant process, the process by which new innovation causes exits from the system, losses at the start become gains at the end, for currency issuer.  R is really the current  currency losses (currency gains to the economy) and I the residual arbitrage left in the savings to loans queue structures. Th two queues should under go a step bysteo isomorphism, squeezing the losers out.  Real and inflation then become the two sides of the Hamiltonian, potential and kinetic. The shock to R dissipates into I. which then has rebound elasticity.

Now, great, this leads to the quantum mechanical solutions to problems like fiscal price theory, as discussed by  Cochrane, here.

In fiscal theory we have two queues, taxes in and payment out, they have o match (and you should use state and local systems). We know we have high trade book error, we know our complete sequence is one generation. We ave the tax sequence, we have the payment sequence.   

Dimensionality? Try finding an optimum huffman code for the sequences, or guess at six dimensional.  Find the random generating process for each, make them isomorphic to each other with market making queue entries. Over the sequence the two queues should be isomorphic, (one queue structure is a uniform ratio of another unifor queue structure). One is a compressed encoding of the other.