A Model of the Trade Cycle Author(s): Nicholas Kaldor Source: The ...

1940] A MODEL OF THE TRADE CYCLE 85accumulate and the I curve will shift upwards; and this tendencyis likely to be reinforced by new inventions. For the same reason,the graduaI decumulation of capitaI, in so far as it causes realincome per unit of activity to fall, willlower the S curve.! Thesemovements cause the position of A to shift to the right and thatof. G to shift to the left (thus separating Band C ànd bringingA and C nearer to each other), involving a graduaI improvementin the level of activity (Stages IV and V). This will proceeduntil A and G fall together (the two curves again become tangential),when a new criticaI situation is reached; the positionbecomes unstable in an upward direction, since I >S on eitherside of the equilibrium point; an upward cumulative movementwill follow which can only come to rest when position B is reached(Stage VI). Thereafter the curves gradually return to theposition shown in Stage I, and the cyclical movement isrepeated.7. The necessary and sufficient assumptions under which thecombined operation of the saving and investment functions inevitablygenerate a èyclical movement which nowhere tends tocome to rest, can therefore be set out as follows :-(l) The "normal value" of ~;, valid for normal levels ofactivity, must'be greater than the corresponding value of~~.(2) The "extreme values" of ~~, valid for abnormally highor abnormally low levels of activity, must be smaller than thecorresponding values of ~~.(3) The level of investment at the upper equilibrium paintmust be sufficiently large for the I(x) function to JaZZ (in time)relatively to the S(x) function; and at the lower equilibrium pointit must be sufficiently small for the I(x) function to rise (in time)relatively to the S(x) function.In other words, the position ofzero' net investment must fall within the limits set by the levelsof investment ruling at B + C and A +. C, in Stage III andStage VI, respectively.If condition (1) did not obtain, equilibrium at C (which is infact the " normal " equilibrium position) would be stable, instead1 It is possible that even if net investment is negative, real output per headshould gradually rise in time (on account of the introduction of superior or more" capitalistic " processes of production during the depression), as a result of whichS would tend to rise rather than falle But this makes no difference so long as theI curve rises faster than the S curve.