The General Theory of Employment, Interest, and Money

Nature and Necessity of Interest
, that it is not certain that the sum saved out of a 
given income necessarily increases when the rate of interest is increased; whereas no one doubts 
that the investment demand-schedule falls with a rising rate of interest. But if the 
Y
-curves and the 
X
-curves both fall as the rate of interest rises, there is no guarantee that a given 
Y
-curve will 
intersect a given 
X
-curve anywhere at all. This suggests that it cannot be the 
Y
-curve and the 
X
-
curve alone which determine the rate of interest. 
In the second place, it has been usual to suppose that an increase in the quantity of money has a 
tendency to reduce the rate of interest, at any rate in the first instance and in the short period. Yet no 
reason has been given why a change in the quantity of money should affect either the investment 
demand-schedule or the readiness to save out of a given income. Thus the classical school have had 
quite a different theory of the rate of interest in volume I dealing with the theory of value from what 
they have had in volume II dealing with the theory of money. They have seemed undisturbed by the 
conflict and have made no attempt, so far as I know, to build a bridge between the two theories. The 
classical school proper, that is to say; since it is the attempt to build a bridge on the part of the neo-
classical school which has led to the worst muddles of all. For the latter have inferred that there 
must be 
two
sources of supply to meet the investment demand-schedule; namely, savings proper, 
which are the savings dealt with by the classical school, 
plus
the sum made available by any 
increase in the quantity of money (this being balanced by some species of levy on the public, called 

---

92
'forced saving' or the like). This leads on to the idea that there is a 'natural' or 'neutral' or 
equilibrium' rate of interest, namely, that rate of interest which equates investment to classical 
savings proper without any addition from 'forced savings'; and finally to what, assuming they are on 
the right track at the start, is the most obvious solution of all, namely, that, if the quantity of money 
could only be kept 
constant
in all circumstances, none of these complications would arise, since the 
evils supposed to result from the supposed excess of investment over savings proper would cease to 
be possible. But at this point we are in deep water. 'The wild duck has dived down to the bottom—
as deep as she can get—and bitten fast hold of the weed and tangle and all the rubbish that is down 
there, and it would need an extraordinarily clever dog to dive after and fish her up again.' 
Thus the traditional analysis is faulty because it has failed to isolate correctly the independent 
variables of the system. Saving and investment are the determinates of the system, not the 
determinants. They are the twin results of the system's determinants, namely, the propensity to 
consume, the schedule of the marginal efficiency of capital and the rate of interest. These 
determinants are, indeed, themselves complex and each is capable of being affected by prospective 
changes in the others. But they remain independent in the sense that their values cannot be inferred 
from one another. The traditional analysis has been aware that saving depends on income but it has 
overlooked the fact that income depends on investment, in such fashion that, when investment 
changes, income must necessarily change in just that degree which is necessary to make the change 
in saving equal to the change in investment. 
Nor are those theories more successful which attempt to make the rate of interest depend on 'the 
marginal efficiency of capital'. It is true that in equilibrium the rate of interest will be equal to the 
marginal efficiency of capital, since it will be profitable to increase (or decrease) the current scale 
of investment until the point of equality has been reached. But to make this into a theory of the rate 
of interest or to derive the rate of interest from it involves a circular argument, as Marshall 
discovered after he had got half-way into giving an account of the rate of interest along these lines. 
For the 'marginal efficiency of capital' partly depends on the scale of current investment, and we 
must already know the rate of interest before we can calculate what this scale will be. The 
significant conclusion is that the output of new investment will be pushed to the point at which the 
marginal efficiency of capital becomes equal to the rate of interest; and what the schedule of the 
marginal efficiency of capital tells us, is, not what the rate of interest is, but the point to which the 
output of new investment will be pushed, given the rate of interest. 
The reader will readily appreciate that the problem here under discussion is a matter of the most 
fundamental theoretical significance and of overwhelming practical importance. For the economic 
principle, on which the practical advice of economists has been almost invariably based, has 
assumed, in effect, that, 
cet. par
., a decrease in spending will tend to lower the rate of interest and 
an increase in investment to raise it. But if what these two quantities determine is, not the rate of 
interest, but the aggregate volume of employment, then our outlook on the mechanism of the 
economic system will be profoundly changed. A decreased readiness to spend will be looked on in 
quite a different light If, instead of being regarded as a factor which will, 
cet. par
., increase 
investment, it is seen as a factor which will, 
cet. par
., diminish employment.