The author begins this section by noting that he hopes to provide clarity on such "topically interesting problems" as "the effects of saving and investment on the rate of interest" as well as "the effects of general changes in money wages." But he observes that it is difficult to determine the correct answers to these questions. The reason for this difficulty, he explains, involves the phrase he placed at the beginning of the previous section—essentially used as a subtitle of the chapter—namely that "the temporary equilibrium system is liable to be imperfectly stable."
As part of his discussion, the author reviews the results of his earlier analysis of stability in exchange. He summarizes these results as follows:
In order for a system of multiple exchange to be perfectly stable (and the temporary equilibrium system is simply an extended system of multiple exchange), the following conditions must be satisfied. A rise in the price of any commodity must make the supply of that commodity exceed the demand (a) if all other prices are given, (b) if some other prices are adjusted so as to preserve equality between demand and supply in their respective markets, (c) if all other prices are so adjusted.
He describes this last condition as being "indispensable." Without it, "the system is not stable at all, but will break down at the slightest disturbance." Assuming this condition is met, either of the other conditions could fail to hold, and the system would still be "stable in the end ... but we have to be prepared for its working to show queer anomalies."
When the author applied these stability tests to static systems, he "found no significant reason to suppose that they gave any particular trouble." Hence, his analysis treated them as perfectly stable. In the current chapter he addresses the question of "What happens when we apply the same tests to the dynamic system—or rather to the system of temporary equilibrium?"
His plan for answering this question is to try "to construct a particular case of the temporary equilibrium system" in such a way that its formal properties match those of the static case. This particular case will then be perfectly stable. He will then compare the particular case with the general (imperfectly stable) case, in order to "see whether there is anything in the general case likely to upset its stability—and if so, what the disturbing element is."
