085 Econometric business cycle research
Turning, first, to the mathematico-economic part of the work, I think this may be characterised as the construction of a scheme for the utilisation of business cycle theories. It consists in indicating the logical structure of the business cycle mechanism. A graphic representation, given in Chart I, may serve as a starting point. This scheme shows, in each (vertical) column the list of phenomena (vari-ables) included: A, B, C … In each (horizontal) row the course of time is represented; i.e. the consecutive dots represent one phenomenon at consecutive unit time intervals. Denoting these by a suffix, the dots represent, e.g. A1, A2, A3, A4, etc. The extent of each of them, if A is a measurable phenomenon, could be plotted in a third dimension, e.g. perpendicular to the plane. We shall not, however, go into that now. Any definite theory tells us how a given change at moment t in A acts on other phenomena at other moments. Suppose the theory is that it acts on B without lag and on C with a lag of one time unit, i.e. A (t) acts on C(t+I). This is indicated by the arrows from A(t) to B(t) and from A(t) to C(t+I). If, e.g. changes in C are assumed to work on D and A, both with a lag of two time units, this will again be indicated by arrows. A change in A(t) may be said to be a „first“ or „direct„ „cause“ to a change in C(t+I), and a „second“ or „indirect“ cause to a change in D(t+3). All the arrows repeat themselves as long as the model’s structure is supposed to remain the same. The more details are considered, the greater the number of arrows. The totality of arrows may be „listed“ in two ways, viz. (i) according to the variable from which they start, or (ii) according to the variable at which they end. In the first listing all „effects“ of changes in one variable on others are grouped together; in the second all „causes“ of changes in one variable are put into one group. Both lists describe, however, the same mechanism. The latter corresponds to what will, in this paper, be called the system of elementary equations. Each equation indicates how changes in one variable depend on the „causing“ changes in other variables.