In economic systems there are many contingent decisions, which take the place of the electronic controlled switches. For example, interest rate changes may be made with the decision to change rates being triggered by a measurement of the inflation statistics. Of course, here we have an example of feedback around the controlled switch, for if the inflation rate moves, the switch may be invoked again. Decisions to alter the economic rules may be made as a result of observations or measurements on the economic system's resulting variable. This then results clearly in a variable structure system.
On the stock market it is common to use computers to assist trading decisions. The computers are becoming ever more sophisticated in taking measurements of the prices and other indicators to inform their automated buying and selling decisions. Again, this is a classic case of embedding controlled switches in a feedback environment, and we should not be surprised when sudden events happen and the system becomes unpredictably chaotic.
There is a strong temptation to search for an immediate cause when a stock market crash occurs. It is a thesis of this paper that such events can be a natural consequence of the dynamics of the system being considered, and that any attempt to ``improve'' the system by fixing the immediate cause of a crash may have no effect, and may even make matters worse.