[From Rick Marken (2000.08.03.1500)]
Bruce Nevin (2000.08.03.1236 EDT)--
One thing I learned at the conference, to my surprise, is
that "leakage" is an undefined term.
I think leakage is quite well defined. Leakage (as I said on
the net and at the meeting) is the proportion of income being
received by the aggregate consumer that is _not_ being
returned to the aggregate producer as payment for Q' (the
goods and services being produced).
It "explains" a discrepancy where aggregate demand (money
spent by the aggregate consumer) is less than the aggregate
cost of goods and services (money which, by being spent, is
available to the aggregate consumer for purchases).
It doesn't really "explain" this discrepancy. I'd say leakage
_is_ this discrepancy.
However, it is nothing more than a label for that
discrepancy, reified as an explanatory principle.
Now you're talkin'. Leakage is a label for the discrepancy
between aggregate consumer income and aggregate consumer
outflow (purchase of Q'). The concept of leakage is "reified"
as a variable in the model; it is a disturbance to the money
available to the aggregate consumer (P'Q') for purchase of Q'.
The model doesn't specify the measurable causes of this
disturbance but TCP suggests two possibilities. One, called
"alpha leakage", is identified with the "personal savings"
and "undistributed corporate profits" measures in the
_Statistical Index_. The cause of alpha leakage is assumed to
be maldistribution of income. The other, called rho leakage,
is identified with the monetary policies of the Fed. It's
hard to find measures of how much money the Fed has in it's
vaults so the best available measure of rho leakage is
the discount rate, which is supposed to be (according to
the Fed itself) proportional to the amount of consumer
income being taken out of circulation.
In particular, we can't legitimately say that the discount
rate *is* leakage.
The discount rate is one cause of leakage; it is not the
leakage itself.
This is a problem because leakage is such a central concept
to TCP's analysis and RSM's model derived from it.
I don't see what the problem is. I agree that leakage is hard
to measure. To measure it we would have to identify the money
that is being received as income (wages and profits) and is
never being spent on goods and services. Most of this money
is probably going into stocks and bonds (such as the bonds
the Fed sells to take money out of the economy). But it's hard
to tell which money is being "parked" in these instruments and
which is being left there to be drawn down for future consumption.
Anyway, what I wanted to get at the meeting (but never got) was
suggestions for improving the model, including pointers to sources
of data that might be associated with variables in the model. All
I got, however, was told that the model had problems. No one said
what the problems were or, more importantly, explained how these
problems might be fixed.
What I really wanted (and still want) to know is how any economic
model can explain the data attached below as a GIF. The data show
the incredibly strong _positive_ relationship between discount rate
and inflation rate over time. My H. economicus model accounts for
these results; the model corrects for a leakage disturbance to the
variables it is controlling (PQ'-P'Q' and P'Q') by decreasing
production of Q' _and_ by increasing the price of goods and
services (P'). If it is assumed that discount rate causes leakage
(the higher the discount rate the greater the leakage as consumer
income is moved into bonds) then the model accounts for the
positive relationship beteen discount rate and inflation. The
fact that H. economicus predicts this surprising result while
conventional economic theories (as far as I know) predict the
_opposite_ (and, thus, deny the data itself!) seems like pretty
strong evidence that the H. economicus model is on the right
track. But if it's not, I would sure like to know _why_ it's
not and get pointers to the existing economic models that make
the right predictions.
Best
Rick
(Attachment data.GIF is missing)
···
--
Richard S. Marken Phone or Fax: 310 474-0313
MindReadings.com mailto: marken@mindreadings.com
www.mindreadings.com