Keynes and the variable 'potential output' in economics

Potential output is mainly an accounting fiction used to adjust the national accounting identity. It exists because economists have no clue about the cost and mechanism of re-pricing. So absent any sound knowledge of repricing costs, economists use the recessions as market points to readjust their potential output variable.

I can make an argument that it is more efficient top defer re-pricing costs until the synchronous period of adjustment, on presidential election cycles. But Keynes can't, he had no knowledge or education in the concept of spectral analysis.  So he named 'animal spirits' and Freidman called it the 'pricing puzzle'.  And essentially they simply compute potential output as a recursive Markov variable. That is fine, both of them admitted the uncertainty of their models.

Krugman and the rest never admit the uncertainty, they simply make up some story, then claim supporting evidence, most of which is rigged evidence derived from the prior assumption. The expectation function these Keynesians use is simply fiction with no proof of soundness.  Keynesians use the function to dump any process they cannot explain.

Then these Keynesians are clueless about the cost of money to the federal government.  A simply calculation of outstanding debt and interest payments reveal that DC pays about 2.5% per year for money, as of  a couple of years ago. I quit doing the calculation when printing became normal. But if we assume DC can print the money, then the absolute floor of the interest rate will be 1%, because that is the cost of running the central bank. Without the Fed economists doing the accounting at the central bank, the paper is worthless, so 1% is an absolute minimum. Yet Larry Summers comes up with Zero, based on no actual facts.

Kling (PSST) and Stiglitz (Screening)  have some strong clues about a better model, and Jim Hamilton a strong clue about repricing scarce resources. The Krugman trade theory helps. Then there is Roger Farmer's work. It is mainly trying to estimate the total number of possibilities in the economy, then finding separable grouping to organize that estimate. Otherwise known are the entropy estimation, what I call the finite log estimation in an integer base. The quantity theory of money is totally screwed as of this moment, but it  really should be the same form as the Plank's curve with money velocity replacing temperature; both being a band limit on spectrum, at maximum entropy.

Getting at a better model:

The better model is  the minimal redundant, finite economic network, then add inefficiencies, all of which is outlined in Wiki. That is where young economists want to go, dump the dark ages and move on.