190

A TREATISE ON MONEY

BK. Ill

as determining them. It was left to Bagehot {LombardStreet, chap, v.) to complete the story by emphasisingthe extent of the Bank of England’ s power, though“the Bank of England has no despotism in the matter”,to determine what the market conditions should be.

It is obvious that this too is an important aspectof the question. But it is by no means obvious howit is connected with our first strand, and I know ofno author who has attempted the synthesis. More-over—superficially at least—it seems to pull in theopposite direction. For the object of raising Bank-rateis to attract gold or to prevent the loss of gold, sothat its effect is to increase the basis of credit abovewhat it would have been otherwise. It may beobjected that the higher Bank-rate can only be madeeffective if the Central Bank reduces its other assetsby more than it increases its stock of gold, so thatthe effect on balance is to decrease the aggregate ofcredit. 1 But it is questionable how exactly or in-variably this corresponds to the observed facts.

3. The third strand of thought is the one whichcomes nearest to what seems to me to be the essenceof the matter. It also winds through many previousdiscussions, but seldom or never in a clear or distinctfashion. This strand of thought conceives of Bank-rate as influencing in some way the rate of investment,or at least the rate of some kinds of investment, and,perhaps in the case of Wicksell and Cassel, as influen-cing the rate of investment relatively to that of savings.

The simplest way of putting the point, as I conceiveit, is to say that to raise the Bank-rate discouragesinvestment relatively to saving, and therefore lowers

1 This is probably the underlying assumption of the traditional doctrine,t.e. that high bank-rate simultaneously reduces the aggregate superstructureof credit whilst increasing that part of its basis which consists of gold. Cf.Withers, Meaning of Money, pp. 276, 277. But I do not remember to haveseen the accuracy of this assumption tested by reference to statistics.According to my theory, it would not be true invariably but would requirevarious special conditions for its fulfilment.