In recent years many business cycle models have been constructed. Generally speaking, they are based on (i) the interrelation between investment and income (resulting from the relation between income and saving and the equality of saving and investment); and (ii) the interrelation between the level of income or the rate of change of income and investment decisions. The most controversial of the assumptions underlying such models are those concerning the determinants of investment decisions. The rate of investment decisions is assumed in some theories to be determined by the rate of change in income (or output) and in some by the level of income. Indeed, the problem of determinants of investment belongs probably to the least explored subjects of modern economics.
It is for this reason that I propose in this paper to reverse, in a way, the usual course of constructing a business cycle model. I shall assume the interrelation between investment and income (based on the relation between income and saving, and the equality of saving and investment). But I shall make no assumption about the determinants of investment decisions. I shall assume instead, in agreement with facts, that a business cycle is in existence, that its period is no less than 8 years, and that the time-lag between investment decisions and actual investment is on the average no more than 1 year. On this basis we shall try to find out something about the determinants of investment. It will be seen, in fact, that this slender empirical foundation will enable us to apply the correlation analysis to our problem.
1. We assume a closed system and a balanced government budget. In such a system, saving S is equal to private investment I
S = I (1)
I is conceived of as measured in real terms, and S stands thus for 'real saving'.