From the point now reached, several roads open out, several which look like being worth exploring. I shall do no more than take a few steps on some of them.
I began, in Chapter II , by suggesting that a typical economic process would have a rising and then (probably) a falling phase. The profile of labour inputs (in particular) would be of this shape, though there might well be a kick-up at the end, when the 'machine' was requiring additional maintenance. The Simple Profile, with which we have so long been working, is a rather poor approximation to this economic shape. One of the things which we should like to know is the extent to which our results would be affected if we had used a more realistic profile. The working-out, obviously, would have been more complex; but is there anything essential which would have been changed?
I have made no attempt to calculate a Traverse, from one non-Simple Profile to another. It would be a formidable task; but I expect that in some cases it could be done. All I can offer are some rather suggestive results in the matter of convergence.
With a Simple Profile, as indicated in Chapter VII (and proved in sects. 7-8 of the Appendix), there is assured convergence of a Full Employment path, so long as the constructional input coefficient is greater than the utilizational (a0 > a1). This does not mean that a Traverse, from any initial position, can then be carried through smoothly, with every process that has once been started being carried to completion as planned. We have already seen, in our discussion of Lengthening, that that is not so. 1 All it means is that when a Late Phase has been reached, so that all live processes are of the same Simple type with the same coefficients, a smooth convergence to equilibrium (in the absence of