[From Rick Marken (970324.1000)]
In a private post, Mike Acree told me that he thought the macro
economic theory developed by T. C. Powers in "Leakage" was inconsistent
with the theory of individual behavior developed by his son in B:CP. I
don't know why Mike thinks this is the case but I told him that I had
come to the opposite conclusion and that I would mention the reasons why
on the net. So here they are:
TCP views the _aggregate_ economy as a machine made of individuals who
cooperate to produce goods and services for themselves. The aggregate
producer and the aggregate consumer are (in the ideal case) the same
people. This view of the economy seems to take it for granted that the
individuals who make up the aggregate are autonomous control systems,
acting (in cooperation) to produce the input perceptions (of goods and
services) that they want and need.
This view of the individuals who make up the aggregate economy seems
quite different than the views of other economic theories, which seem
to view individual producers as entities who must be proded, by the
"push" of investment and/or the "pull" of incentives (wages and
profits), to work and produce. In TCPs view, individual producers
produce because they are also consumers. So the aggregate producer
produces in order to get what it needs (as the aggregate consumer). This
view implies that work (production) is what people do naturally in order
to get what they need; they don't have to be proded into working. If
people are not working it is because there is no work (that is part of
the aggregate producer) available to them.
In TCP's model of the economy, the aggregate producer pays _itself_ (the
aggregate consumer) exactly what it costs to produce the goods that are
made available to itself. AIdeally, the aggregate consumer should use
this income to purchase all the fruits of its production. But, in fact,
the aggregate consumer returns only about 93% of its income to itself
(the aggregate producer). This is what TCP calls "leakage". The result
of leakage is a reduction in growth rate.
I believe leakage operates like reduced gain in a control system. The
aggregate producer/consumer is the control system; it is trying to
produce input for itself (goods) at a rate that matches its "demand".
The aggregate producer/consumer maintains production by spending on
capital investment (which is ALWAYS 20% of its income) and wages. These
production costs are paid for by the income from consumtion.
But the aggregate consumer is paying itself only 93% of what it _could_
pay. This limits the production capacity of the aggregate producer; it
has 7% less money available for capital investment and wages than it
should have. This means that the aggregate producer cannot produce at
the rate it COULD (or would LIKE TO) produce; so the aggregate producer
is not producing what it _could_ produce to meet its needs as aggregate
consumer.
The central assumption of this model is that the amount paid to the
aggregate consumer (measured as GNP + undistributed corporate profits
(U) + personal savings (S)) should match the amount spent by the
aggregate consumer (GNP). Of course, these don't match to the extent
that U+S >0. "Leakage" refers to the fact that U+S has always been
0 and, indeed, has averaged 7% of total consumer income since 1900 or so.
Mike Acree has suggested that the leakage seen by TCP is an artifact;
consumer income (GNP+U+S) cannot be spent right away so it is no
surprise that GNP+U+S at time t doesn't equal GNP at time t+some time
later. Mike suggested that consumer income at time t should actually
match consumer spending (measured as GNP) at time t+1 year or so.
I computed the relationship between consumer income and spending this
way and found that, indeed, GNP+U+S at time t is almost always VERY
close to GNP at time t + 1 year. If this little lagged anaysis holds up
then the whole idea of "leakage" goes right out the window; it would
suggest that all consumer income is spent within one year after it is
recieved. This would require a MAJOR change in TCPs data analysis. But I
don't think it invaliates TCP basic systems approach to looking at the
behavior of the aggregate economy.
I suspect, however, that the results of this lagged analysis are,
themselves, an artifact; they result from the fact that the average rate
of growth of the economy brings the GNP at year t+1 up to the level that
happens to match last year' GNP+U+S. One hint that this might be the
case is the fact that for several of the years for which data is
available, the lagged analysis gives a measure of spending for year t+1
that is quite a bit _larger_ than the income on year t. This could be
seen as a result of "pent up demand" from previous savings but the
numbers don't add up; more is spent during these _over spending years_
than is accounted for by the "underspanding" of previous years.
Anyway, as Bill Powers said, TCP's work is just a start. But I think
it's a good one. And I think his ideas about the behavior of the
aggregate economy are far more consistent with the PCT model of
individual behavior than is any other economic theory that I am
aware of.
Best
Rick
