[T]he [Quantity] Theory [of Money] has often been expounded on the further assumption that a mere change in the quantity of the currency cannot affect k, r, and k',--that is to say, in mathematical parlance, that n is an independent variable in relation to these quantities. It would follow from this that an arbitrary doubling of n, since this in itself is assumed not to affect k, r, and k', must have the effect of raising p to double what it would have been otherwise. The Quantity Theory is often stated in this, or a similar, form.
Now "in the long run" this is probably true. If, after the American Civil War, that American dollar had been stabilized and defined by law at 10 per cent below its present value, it would be safe to assume that n and p would now be just 10 per cent greater than they actually are and that the present values of k, r, and k' would be entirely unaffected. But this long run is a misleading guide to current affairs. In the long run we are all dead. Economists set themselves too easy, too useless a task if in tempestuous seasons they can only tell us that when the storm is long past the ocean will be flat again.
In actual experience, a change in n is liable to have a reaction both on k and k' and on r. It will be enough to give a few typical instances. Before the war (and indeed since) there was a considerable element of what was conventional and arbitrary in the reserve policy of the banks, but especially in the policy of the State Banks towards their gold reserves. These reserves were kept for show rather than for use, and their amount was not the result of close reasoning. There was a decided tendency on the part of these banks between 1900 and 1914 to bottle up gold when it flowed towards them and to part with it reluctantly when the tide was flowing the other way. Consequently, when gold became relatively abundant they tended to hoard what came their way and to raise the proportion of the reserves, with the result that the increased output of South African gold was absorbed with less effect on the price level than would have been the case if an increase of n had been totally without reaction on the value of r.
In agricultural countries where peasants readily hoard money, an inflation, especially in its early stages, does not raise prices proportionately, because when, as a result of a certain rise in the price of agricultural products, more money flows into the pockets of the peasants, it tends to stick there;--deeming themselves that much richer, the peasants increase the proportion of their receipts that they hoard.
Thus in these and in other ways the terms of our equation tend in their movements to favor the stability of p, and there is a certain friction which prevents a moderate change in n from exercising its full proportionate effect on p.
On the other hand a large change in n, which rubs away the initial friction, and especially a change in n due to causes whig st up a general expectation of a further rchatne in the same direction, may produce a more than proportionate effect on p. After the general analysis of Chapter I. and the narratives of catastrophic inflations given in Chapter II., it is scarcely necessary to illustrate this further,--it is a matter more readily understood than it was ten years ago. A large change in p greatly affects individual fortunes Hence a change after it has occurred, or sooner in so far as it is anticipated, may greatly affect the monetary habits of the public in their attempt to protect themselves from a similar loss in the future, or to make gains and avoid loss during the passage from the equilibrium corresponding to the old value of n to the equilibrium corresponding to its new value. Thus after, during, and (so far as the change is anticipated) before a change in the value of n, there will be some reaction on the values of k, k', and r, with the result that the change jun the value of p, at least temporarily and perhaps permanently (since habits and practices, once changed, will not revert to exactly their old shape), will not be precisely in proportion to the change in n…
Keynes (1923), "A Tract on Monetary Reform".pdf