[From Bill Powers (2008.12.05.0444 MST)]
This is a letter that needs to be conveyed to as many people who make
economic decisions as possible. OUR ECONOMIC SYSTEM CONTAINS
DESTABILIZING FEEDBACK LOOPS THAT CAN DESTROY IT. WE NEED TO FIGURE OUT
HOW TO REMOVE THEM AS QUICKLY AS POSSIBLE.
This is a true time bomb. It is perfectly obvious, and it is to my shame
and that of everyone who understands the dynamics of control systems that
it was not noticed, publicized, and corrected long ago. It is very simple
and we are watching it operate every day that this recession
deepens.
Its cause is some set of policies or principles that are thought to be
necessary to maintain the viability of a business, but which, when
generally adopted, have the effect of exaggerating swings in the market
and, if widespread enough, throw the market into a state of dynamic
instability that feeds on itself. Increases in market activity cause a
piling-on effect which drive the increases even further and induce more
frenzied market activity. The same underlying relationships work the
other way, too: when the market peaks and starts downward, this cause the
enthusiasm to wane and the market activity to slow down, and the slowdown
causes an even more dampening effect, which makes the slowdown
accelerate. Whichever way the market tends to change, the change is
exaggerated by this feedback effect. The initial result when the amount
of feedback is small is that the economy displays
“boom-and-bust” cycles of relatively small amplitude, which die
out after a time. When the degree of this effect becomes large
enough, the swings start to get larger and can enter a region in which a
runaway effect occurs. Then the only way to stop the growing oscillations
is for something in the system to be damaged enough to reduce the
feedback effect below the fatal threshold of sensitivity.
One obvious destabilizing effect we are seeing right now is that
companies faced with falling sales are laying off parts of their
workforces to reduce costs and preserve their net income. As the number
of employed persons decreases, the total income of the average consumer
falls, so there is less money to spend on goods and services – the very
spending that is the source of income for the companies who are laying
off workers (and each others’ customers). So we have the situation in
which the attempt of companies to reduce expenses and restore their
income to profitable levels has the effect of reducing, not increasing,
the total net income of all companies. That leads to an increase in the
attempts to reduce expenses, and so on. *If enough companies follow the
same policy, there can be only one outcome: total collapse.*This is probably not the only destabilizing feature in the basic
relationships that make up our economic systems. It is only one obvious
feedback loop with the wrong sign. This is known technically as a
positive feedback loop, one in which small changes lead to larger
changes, in contrast with the negative feedback loop in which
small perturbations are negated by feedback effects and are immediately
smoothed out. In popular usage “positive feedback” means
encouragement and approval, but if encouragement occurs only when one is
doing better and is withdrawn when one begins to do less well, the seeds
of true positive feedback have been planted; at the first sign of doing
less well, approval will decrease, and the resulting lessening of
encouragement will cause performance to decrease even more, and so on to
disaster. What works on the way up continues to work on the way down.
Changes in either direction are exaggerated by positive feedback, and the
result is instability. If there is enough instability, the system will
destroy itself.
The cure for positive feedback and its destructive effects is to ferret
out and remove all the positive feedback loops. There should be little
resistance to doing this, because what will be revealed is the clear fact
that actions intended to improve matters are actually making them worse.
Any sane person will cease such actions even if their undesirability is
not immediately explained or understood. Once we step back and look at
the overall relationships, the instability that is caused by a wrong
response to problems becomes obvious. Then the main issue is seen
clearly: what do we need to change to prevent this sort of destabilizing
feedback from occurring?
Stabilizing the economic system does not require telling people what to
buy or what to sell, or for how much, or to whom. It requires
intervention only when a reaction to a change in economic activity,
either upward or downward, results in amplification of the change rather
than smoothing it out. Just why this amplification occurs has to be
studied so the correct adjustment can be made. It could be only that
there is too much reaction to a change. It could be that the reaction is
too much delayed, so what would have been a stablizing reaction if it had
occurred right away becomes destablizing when applied too late. Or it
could be that the intuitive reaction to a problem is exactly the reaction
that will make it worse, showing that there is a mistaken understanding
of how one part of the system affects other parts.
Reaching an understanding of instability in the economic system is no
longer an abstract goal of only academic interest. It is a necessity for
our survival. It requires a massive investigation and the efforts of many
competent scientists who know how to analyze and model large systems – a
Manhattan Project that will finally unravel the mysteries of how the
economic system works. We need to understand exactly what kind of
technical public-sector interventions can be brought to bear when the
free market goes into free fall, without infringing on rights or imposing
micromanagement on individual entrepreneurs and consumers. The answer, at
times, could be as simple as a slowing of changes similar to what the
stock market imposes when panic threatens an individual stock. Sometimes
it could require rulings about policies which, if followed by enough
transacting parties, threaten the entire system. Some existing
regulations might prove to cause more of the problems they were intended
to lessen.
The only way to find the answers we need is to construct a working model
of the economy, which requires studying it in sufficient detail to
construct that model, and then, once consensus is reached, using the
model first to reproduce the economic events that are so troublesome to
everyone, and then looking for the minimum changes required to prevent
them from ever happening again.
Please forward this letter wherever you think it might do some
good.
Bill Powers
