The General Theory of Employment, Interest, and Money

e
o
= 0 or if 
e
w
= 1, output will be unaltered and prices will rise in the same proportion as 
effective demand in terms of money. Otherwise they will rise in a smaller proportion.
II 
Let us return to the employment function. We have assumed in the foregoing that to every level or 
aggregate effective demand there corresponds a unique distribution of effective demand between 
the products of each individual industry. Now, as aggregate expenditure changes, the corresponding 
expenditure on the products of an individual industry will not, in general, change in the same 
proportion;—partly because individuals will not, as their incomes rise, increase the amount of the 
products of each separate industry, which they purchase, in the same proportion, and partly because 
the prices of different commodities will respond in different degrees to increases in expenditure 
upon them. 
It follows from this that the assumption upon which we have worked hitherto, that changes in 
employment depend solely on changes in aggregate effective demand (in terms of wage-units), is no 
better than a first approximation, if we admit that there is more than one way in which an increase 
of income can be spent. For the way in which we suppose the increase in aggregate demand to be 
distributed between different commodities may considerably influence the volume of employment. 
If, for example, the increased demand is largely directed towards products which have a high 
elasticity of employment, the aggregate increase in employment will be greater than if it is largely 
directed towards products which have a low elasticity of employment. 
In the same way employment may fall off without there having been any change in aggregate 
demand, if the direction of demand is changed in favour of products having a relatively low 
elasticity of employment. 
These considerations are particularly important if we are concerned with short-period phenomena in 
the sense of changes in the amount or direction of demand which are not foreseen some time ahead. 
Some products take time to produce, so that it is practically impossible to increase the supply of 
them quickly. Thus, if additional demand is directed to them without notice, they will show a low 
elasticity of employment; although it may be that, given sufficient notice, their elasticity of 
employment approaches unity. 
It is in this connection that I find the principal significance of the conception of a period of 
production. A product, I should prefer to say, has a period of production 
n
if 
n
time-units of notice 

---

144
of changes in the demand for it have to be given if it is to offer its maximum elasticity of 
employment. Obviously consumption-goods, taken as a whole, have in this sense the longest period 
of production, since of every productive process they constitute the last stage. Thus if the first 
impulse towards the increase in effective demand comes from an increase in consumption, the 
initial elasticity of employment will be further below its eventual equilibrium-level than if the 
impulse comes from an increase in investment. Moreover, if the increased demand is directed to 
products with a relatively low elasticity of employment, a larger proportion of it will go to swell the 
incomes of entrepreneurs and a smaller proportion to swell the incomes of wage-earners and other 
prime-cost factors; with the possible result that the repercussions may be somewhat less favourable 
to expenditure, owing to the likelihood of entrepreneurs saving more of their increment of income 
than wage-earners would. Nevertheless the distinction between the two cases must not be over-
stated, since a large part of the reactions will be much the same in both. 
However long the notice given to entrepreneurs of a prospective change in demand, it is not 
possible for the initial elasticity of employment, in response to a 
given
increase of investment, to be 
as great as its eventual equilibrium value, unless there are surplus stocks and surplus capacity at 
every stage of production. On the other hand, the depletion of the surplus stocks will have an 
offsetting effect on the amount by which investment increases. If we suppose that there are initially 
some surpluses at every point, the initial elasticity of employment may approximate to unity; then 
after the stocks have been absorbed, but before an increased supply is coming forward at an 
adequate rate from the earlier stages of production, the elasticity will fall away; rising again towards 
unity as the new position of equilibrium is approached. This is subject, however, to some 
qualification in so far as there are rent factors which absorb more expenditure as employment 
increases, or if the rate of interest increases. For these reasons perfect stability of prices is 
impossible in an economy subject to change—unless, indeed, there is some peculiar mechanism 
which ensures temporary fluctuations of just the right degree in the propensity to consume. But 
price-instability arising in this way does not lead to the kind of profit stimulus which is liable to 
bring into existence excess capacity. For the windfall gain will wholly accrue to those entrepreneurs 
who happen to possess products at a relatively advanced stage of production, and there is nothing 
which the entrepreneur, who does not possess specialised resources of the right kind, can do to 
attract this gain to himself. Thus the inevitable price-instability due to change cannot affect the