Geoffrey,

Thanks very much for your valuable input.

>We would all like to emulate mathematically the flow of credit
>transactions, but after only a couple of minutes' thought we give up,
>discouraged by the complexity of the subject. In physics I suppose that
>weather systems present the greatest problems for the scientists, but
>economic systems must rival them for complexity.
>
>Following the example of too many economists you have decided to model the
>"money supply", but that is a very arbitrary category. It is part of the
>overall credit supply which falls into two parts, the direct credit
>supply, where the lender and the borrower deal direct with one another,
>and the indirect credit supply, where the ultimate lender and the ultimate
>borrower are separated by an intermediary, a bank. We define a bank as a
>body which accepts loans and makes loans, the two being balanced except
>for that proportion of the amount loaned by the bank which must by law or
>prudence by provided from the banker's own funds.

It is important to say, that I do not make these simulations, because
I think they exactly emulate our system.

I want to find out a correct 'measurement systems' for monetary
wealth. Since such a system must also work in my particular system,
I check for contradictions with the simulations. In our case a
random economy does not go into a steady state as measured
by 'quantity of money' and I can show how methods to correct
this will alter the monetary wealth distribution.

So in a way, I exactly do what you want: questionning the
measurement foundation, which is at the end in any case a
function of the implemented bookkeeping. Any statistics
is sitting on top of bookkeeping. That's why i am so much
interested in the details and foundations of bookkeeping
(and so much puzzled by the fact that economics does not
really care about it).

>To model only the indirect credit supply (popularly called "the money
>supply") is to ignore the interaction between the two supplies, and the
>fact that the distinction is accidental, arbitrary, with little real
>significance. There is also enormous double-counting. But the basi problem
>is that credit can pass from one category to another at a moment's notice.

Our strategy in this respect is to give units of accounting
based on their origin of asset/liability creation. This is
quite revolutionary, but I think very powerful. So we try to
follow this passing along categories by following its path
in the space-time graphs. There it can be seen, whether its
unit of account was changed (for example by pair recreation)
or not.

>The flow is not random, but is a function of two factors, the interest
>rate curve, and the capital base of the banks. As with all economic
>factors there is also feedback, causing a permanent oscillation and
>instability.

We approximate all non-randomness by the bias brought in with
the transfer potential. So we have already a handle to put this
into it, however it is hard to measure. (In a way we could say: give
me the wealth distribution, I then can deduce directly the
transfer potential and therefore derive all the bias deviating
from random transfers. But this is pushing the model too far.)

>In the light of these influences on the "money supply" did you not
>oversimplify the model, and use assumptions which make sure that the model
>predicted what observation has long established, that the rich tend to get
>richer and the poor poorer, if the state does not regulate the process?

All the assumption is focussing on monetary wealth and
assuming random transfer. That's actually the funny part:
I do not have to assume any 'richer will get richer because
of this and that process'. Randomness does it already!

>Monetary economics is full of follies. We have had thousands of
>"economists" watching intently the day by day movements in M0, or M2, or
>M3, or M4, or M5, but only very few of them have had even a slight notion
>of what it all means. High interest rates were supposed to reduce the
>"money supply," but the empirical evidence was that they did the exact
>opposite. As this was against theory, the phenomenon was ignored.
>Other economists studiously study the growth rate of GDP. None of them
>noticed that the easiest way to raise GDP is to raise interest rates. GDP
>is a measure of the cost of production of the annual supply of goods and
>services. Interest is a cost like any other, so an increase in interest
>charges automatically raises the nominal value of GDP.
>
>What our Gang would like of physicists is help to expose these
>unscientific madnesses of woolly-minded economists, not to bolster them by
>prostituting the science of physics and mathematics even more that has yet
>been done. We need more of them to be as vitriolic as the late Frank
>Engledow whose comment on the activities of the likes of Nicholas Kaldor
>was "Economists use multi-linear regression analyses every parameter in
>which is an assumption."

I hope I made clear above that I agree and my path right now is
to do this by bookkeeping analysis. I did the Random transfer more
out of the beauty of the physical analogies. The more important work
is new approaches to explain and understand bookkeeping - in our case
by using time-mapped descriptions in space-time graphs.

I am still stunned on how isolated and arbitrary bookkeeping
is persued. In I way, I almost lost the faith in any economical
motivation if it can be so easily manipulated. Enron should still
teach us lessons (at least not for me).

>Kaldor was strangely weak on simple arithmetic, as you can see by reading
>his inaugural lecture as professor of economics at Cambridge, given in
>1966. His interpretation of statistics was hilarious. He left behind him a
>trail of ruined economies, culminating with the British.

Well at least some young economists are starting to revolt.

>Bill Ryan has been ferocious with you. We have all been subject to his
>fire at times, but we nevertheless have a great tendresse for him as he is
>a wonderful gadfly and has a much better understanding than the
>run-of-the-mill economists. He was a bit naughty on the matter of the
>"surplus" being the excess of assets over liabilities. The businessman
>balances his accounts by showing the "surplus" as a liability to himself.
>So the accounts do balance.

His fire is OK, if it would be precise. Damming approaches
based on thorough analysis is fine, but without its very poisonous.

All the best,
Dieter Braun.





>I cannot go into detail further at the moment as I am tied up with
>commitments and travels for a while.
>
>Geoffrey Gardiner
>----- Original Message -----
>From: <mailto:mail@...>Dieter Braun
>To: <mailto:gang8@yahoogroups.com>gang8@yahoogroups.com
>Sent: Monday, May 12, 2003 3:21 PM
>Subject: Re: [gang8] Bookkeeping-Mechanics
>Hi Geoffrey Gardiner,
>
>Thanks for your long response.
>
>I first wanted to wait a bit to tune into the
>sprachspiel of this list since - as you might see
>from the private response by William B. Ryan,
>you get dismissed fast on the basis of spurious reading.
>Economy's lack of experimental tests do not force
>people to be more humble.
>
>I am waiting on an overview paper on bookkeeping
>mechanics to be published in some weeks and do a
>more detailed introduction on that then.
>
>First a short remark on statistical mechanics, also
>called statistical thermodynamics. It is not a statistical
>technique in the way statistics is used in the examples
>you give. It does not experimentally count the gas particles
>to find the gas equation. It is based on energy and entropy
>principles and the law of big numbers. The Boltzmann distribution
>which we again find in the system of random monetary transfers
>is the corner stone of statistical mechanics.
>The paper does four things:
>(a) It analyzes a random economy on the monetary level of bookkeeping.
> So no depreciations, no capital, only a bookkeeping memory of who
> has given and who has received. We show that such a system does not
> yield a steady state and violates quantity theory.
>(b) We try to remove the monetary inflation of the model by boundary
> conditions. Your mentioned "a boundary condition "a constant number of
>assets"
> is only one of many we test. The boundary condition leaves a
> specific fingerprint on the monetary wealth distribution (gaussian,
> exponential, pareto, boltzmann distributions).
>(c) We can remove the monetary also by imposed systematic transfers
> with a transfer potential, finding the same results as in (b)
> without imposed boundary conditions, but specific transfer potentials.
> The imposed transfers can be either viewed to mimick a bias in
> the random transfer depending on the monetary wealth, or as an
> imposed tax system with a specific wealth progression table.
>(d) The system of (c) shows very simple analytical results and resembles
> the botzmann-statistics of statistical mechanics.
>All the best
>Dieter Braun.