The General Theory of Employment, Interest and Money

By John Maynard Keynes

Chapter 21




So long as economists are concerned with what is called the theory of value, they have been accustomed to teach that prices are governed by the conditions of supply and demand; and, in particular, changes in marginal cost and the elasticity of short-period supply have played a prominent part. But when they pass in volume II, or more often in a separate treatise, to the theory of money and prices, we hear no more of these homely but intelligible concepts and move into a world where prices are governed by the quantity of money, by its income-velocity, by the velocity of circulation relatively to the volume of transactions, by hoarding, by forced saving, by inflation and deflation et hoc genus omne; and little or no attempt is made to relate these vaguer phrases to our former notions of the elasticities of supply and demand. If we reflect on what we are being taught and try to rationalise it, in the simpler discussions it seems that the elasticity of supply must have become zero and demand proportional to the quantity of money; whilst in the more sophisticated we are lost in a haze where nothing is clear and everything is possible. We have all of us become used to finding ourselves sometimes on the one side of the moon and sometimes on the other, without knowing what route or journey connects them, related, apparently, after the fashion of our waking and our dreaming lives.

One of the objects of the foregoing chapters has been to escape from this double life and to bring the theory of prices as a whole back to close contact with the theory of value. The division of economics between the theory of value and distribution on the one hand and the theory of money on the other hand is, I think, a false division. The right dichotomy is, I suggest, between the theory of the individual industry or firm and of the rewards and the distribution between different uses of a given quantity of resources on the one hand, and the theory of output and employment as a whole on the other hand. So long as we limit ourselves to the study of the individual industry or firm on the assumption that the aggregate quantity of employed resources is constant, and, provisionally, that the conditions of other industries or firms are unchanged, it is true that we are not concerned with the significant characteristics of money. But as soon as we pass to the problem of what determines output and employment as a whole, we require the complete theory of a monetary economy.

Or, perhaps, we might make our line of division between the theory of stationary equilibrium and the theory of shifting equilibrium¾meaning by the latter the theory of a system in which changing views about the future are capable of influencing the present situation. For the importance of money essentially flows from its being a link between the present and the future. We can consider what distribution of resources between different uses will be consistent with equilibrium under the influence of normal economic motives in a world in which our views concerning the future are fixed and reliable in all respects;¾with a further division, perhaps, between an economy which is unchanging and one subject to change, but where all things are foreseen from the beginning. Or we can pass from this simplified propaedeutic to the problems of the real world in which our previous expectations are liable to disappointment and expectations concerning the future affect what we do to-day. It is when we have made this transition that the peculiar properties of money as a link between the present and the future must enter into our calculations. But, although the theory of shifting equilibrium must necessarily be pursued in terms of a monetary economy, it remains a theory of value and distribution and not a separate 'theory of money'. Money in its significant attributes is, above all, a subtle device for linking the present to the future; and we cannot even begin to discuss the effect of changing expectations on current activities except in monetary terms. We cannot get rid of money even by abolishing gold and silver and legal tender instruments. So long as there exists any durable asset, it is capable of possessing monetary attributes and, therefore, of giving rise to the characteristic problems of a monetary economy.


In a single industry its particular price-level depends partly on the rate of remuneration of the factors of production which enter into its marginal cost, and partly on the scale of output. There is no reason to modify this conclusion when we pass to industry as a whole. The general price-level depends partly on the rate of remuneration of the factors of production which enter into marginal cost and partly on the scale of output as a whole, i.e. (taking equipment and technique as given) on the volume of employment. It is true that, when we pass to output as a whole, the costs of production in any industry partly depend on the output of other industries. But the more significant change, of which we have to take account, is the effect of changes in demand both on costs and on volume. It is on the side of demand that we have to introduce quite new ideas when we are dealing with demand as a whole and no longer with the demand for a single product taken in isolation, with demand as a whole assumed to be unchanged.


If we allow ourselves the simplification of assuming that the rates of remuneration of the different factors of production which enter into marginal cost all change in the same proportion, i.e. in the same proportion as the wage-unit, it follows that the general price-level (taking equipment and technique as given) depends partly on the wage-unit and partly on the volume of employment. Hence the effect of changes in the quantity of money on the price-level can be considered as being compounded of the effect on the wage-unit and the effect on employment.

To elucidate the ideas involved, let us simplify our assumptions still further, and assume (1) that all unemployed resources are homogeneous and interchangeable in their efficiency to produce what is wanted, and (2) that the factors of production entering into marginal cost are content with the same money-wage so long as there is a surplus of them unemployed. In this case we have constant returns and a rigid wage-unit, so long as there is any unemployment. It follows that an increase in the quantity of money will have no effect whatever on prices, so long as there is any unemployment, and that employment will increase in exact proportion to any increase in effective demand brought about by the increase in the quantity of money; whilst as soon as full employment is reached, it will thenceforward be the wage-unit and prices which will increase in exact proportion to the increase in effective demand. Thus if there is perfectly elastic supply so long as there is unemployment, and perfectly inelastic supply so soon as full employment is reached, and if effective demand changes in the same proportion as the quantity of money, the quantity theory of money can be enunciated as follows: 'So long as there is unemployment, employment will change in the same proportion as the quantity of money; and when there is full employment, prices will change in the same proportion as the quantity of money'.

Having, however, satisfied tradition by introducing a sufficient number of simplifying assumptions to enable us to enunciate a quantity theory of money, let us now consider the possible complications which will in fact influence events:

(1)  Effective demand will not change in exact proportion to the quantity of money.

(2)  Since resources are not homogeneous, there will be diminishing, and not constant, returns as employment gradually increases.

(3)  Since resources are not interchangeable, some commodities will reach a condition of inelastic supply whilst there are still unemployed resources available for the production of other commodities.

(4)  The wage-unit will tend to rise, before full employment has been reached.

(5)  The remunerations of the factors entering into marginal cost will not all change in the same proportion.

Thus we must first consider the effect of changes in the quantity of money on the quantity of effective demand; and the increase in effective demand will, generally speaking, spend itself partly in increasing the quantity of employment and partly in raising the level of prices. Thus instead of constant prices in conditions of unemployment, and of prices rising in proportion to the quantity of money in conditions of full employment, we have in fact a condition of prices rising gradually as employment increases. The theory of prices, that is to say, the analysis of the relation between changes in the quantity of money and changes in the price-level with a view to determining the elasticity of prices in response to changes in the quantity of money, must, therefore, direct itself to the five complicating factors set forth above.

We will consider each of them in turn. But this procedure must not be allowed to lead us into supposing that they are, strictly speaking, independent. For example, the proportion, in which an increase in effective demand is divided in its effect between increasing output and raising prices, may affect the way in which the quantity of money is related to the quantity of effective demand. Or, again, the differences in the proportions, in which the remunerations of different factors change, may influence the relation between the quantity of money and the quantity of effective demand. The object of our analysis is, not to provide a machine, or method of blind manipulation, which will furnish an infallible answer, but to provide ourselves with an organised and orderly method of thinking out particular problems; and, after we have reached a provisional conclusion by isolating the complicating factors one by one, we then have to go back on ourselves and allow, as well as we can, for the probable interactions of the factors amongst themselves. This is the nature of economic thinking. Any other way of applying our formal principles of thought (without which, however, we shall be lost in the wood) will lead us into error. It is a great fault of symbolic pseudo-mathematical methods of formalising a system of economic analysis, such as we shall set down in section vi of this chapter, that they expressly assume strict independence between the factors involved and lose all their cogency and authority if this hypothesis is disallowed; whereas, in ordinary discourse, where we are not blindly manipulating but know all the time what we are doing and what the words mean, we can keep 'at the back of our heads' the necessary reserves and qualifications and the adjustments which we shall have to make later on, in a way in which we cannot keep complicated partial differentials 'at the back' of several pages of algebra which assume that they all vanish. Too large a proportion of recent 'mathematical' economics are merely concoctions, as imprecise as the initial assumptions they rest on, which allow the author to lose sight of the complexities and interdependencies of the real world in a maze of pretentious and unhelpful symbols.


(1)  The primary effect of a change in the quantity of money on the quantity of effective demand is through its influence on the rate of interest. If this were the only reaction, the quantitative effect could be derived from the three elements¾(a) the schedule of liquidity-preference which tells us by how much the rate of interest will have to fall in order that the new money may be absorbed by willing holders, (b) the schedule of marginal efficiencies which tells us by how much a given fall in the rate of interest will increase investment, and (c) the investment multiplier which tells us by how much a given increase in investment will increase effective demand as a whole.

But this analysis, though it is valuable in introducing order and method into our enquiry, presents a deceptive simplicity, if we forget that the three elements (a), (b) and (c) are themselves partly dependent on the complicating factors (2), (3), (4) and (5) which we have not yet considered. For the schedule of liquidity-preference itself depends on how much of the new money is absorbed into the income and industrial circulations, which depends in turn on how much effective demand increases and how the increase is divided between the rise of prices, the rise of wages, and the volume of output and employment. Furthermore, the schedule of marginal efficiencies will partly depend on the effect which the circumstances attendant on the increase in the quantity of money have on expectations of the future monetary prospects. And finally the multiplier will be influenced by the way in which the new income resulting from the increased effective demand is distributed between different classes of consumers. Nor, of course, is this list of possible interactions complete. Nevertheless, if we have all the facts before us, we shall have enough simultaneous equations to give us a determinate result. There will be a determinate amount of increase in the quantity of effective demand which, after taking everything into account, will correspond to, and be in equilibrium with, the increase in the quantity of money. Moreover, it is only in highly exceptional circumstances that an increase in the quantity of money will be associated with a decrease in the quantity of effective demand.

The ratio between the quantity of effective demand and the quantity of money closely corresponds to what is often called the 'income-velocity of money';¾except that effective demand corresponds to the income the expectation of which has set production moving, not to the actually realised income, and to gross, not net, income. But the 'income-velocity of money' is, in itself, merely a name which explains nothing. There is no reason to expect that it will be constant. For it depends, as the foregoing discussion has shown, on many complex and variable factors. The use of this term obscures, I think, the real character of the causation, and has led to nothing but confusion.

(2)  As we have shown above (Chapter 4), the distinction between diminishing and constant returns partly depends on whether workers are remunerated in strict proportion to their efficiency. If so, we shall have constant labour-costs (in terms of the wage-unit) when employment increases. But if the wage of a given grade of labourers is uniform irrespective of the efficiency of the individuals, we shall have rising labour-costs, irrespective of the efficiency of the equipment. Moreover, if equipment is non-homogeneous and some part of it involves a greater prime cost per unit of output, we shall have increasing marginal prime costs over and above any increase due to increasing labour-costs.

Hence, in general, supply price will increase as output from a given equipment is increased. Thus increasing output will be associated with rising prices, apart from any change in the wage-unit.

(3)  Under (2) we have been contemplating the possibility of supply being imperfectly elastic. If there is a perfect balance in the respective quantities of specialised unemployed resources, the point of full employment will be reached for all of them simultaneously. But, in general, the demand for some services and commodities will reach a level beyond which their supply is, for the time being, perfectly inelastic, whilst in other directions there is still a substantial surplus of resources without employment. Thus as output increases, a series of 'bottle-necks' will be successively reached, where the supply of particular commodities ceases to be elastic and their prices have to rise to whatever level is necessary to divert demand into other directions.

It is probable that the general level of prices will not rise very much as output increases, so long as there are available efficient unemployed resources of every type. But as soon as output has increased sufficiently to begin to reach the 'bottle-necks', there is likely to be a sharp rise in the prices of certain commodities.

Under this heading, however, as also under heading (2), the elasticity of supply partly depends on the elapse of time. If we assume a sufficient interval for the quantity of equipment itself to change, the elasticities of supply will be decidedly greater eventually. Thus a moderate change in effective demand, coming on a situation where there is widespread unemployment, may spend itself very little in raising prices and mainly in increasing employment; whilst a larger change, which, being unforeseen, causes some temporary 'bottle-necks' to be reached, will spend itself in raising prices, as distinct from employment, to a greater extent at first than subsequently.

(4)  That the wage-unit may tend to rise before full employment has been reached, requires little comment or explanation. Since each group of workers will gain, cet. par., by a rise in its own wages, there is naturally for all groups a pressure in this direction, which entrepreneurs will be more ready to meet when they are doing better business. For this reason a proportion of any increase in effective demand is likely to be absorbed in satisfying the upward tendency of the wage-unit.

Thus, in addition to the final critical point of full employment at which money-wages have to rise, in response to an increasing effective demand in terms of money, fully in proportion to the rise in the prices of wage-goods, we have a succession of earlier semi-critical points at which an increasing effective demand tends to raise money-wages though not fully in proportion to the rise in the price of wage-goods; and similarly in the case of a decreasing effective demand. In actual experience the wage-unit does not change continuously in terms of money in response to every small change in effective demand; but discontinuously. These points of discontinuity are determined by the psychology of the workers and by the policies of employers and trade unions. In an open system, where they mean a change relatively to wage-costs elsewhere, and in a trade cycle, where even in a closed system they may mean a change relatively to expected wage-costs in the future, they can be of considerable practical significance. These points, where a further increase in effective demand in terms of money is liable to cause a discontinuous rise in the wage-unit, might be deemed, from a certain point of view, to be positions of semi-inflation, having some analogy (though a very imperfect one) to the absolute inflation (cf. Chapter 21 below) which ensues on an increase in effective demand in circumstances of full employment. They have, moreover, a good deal of historical importance. But they do not readily lend themselves to theoretical generalisations.

(5)  Our first simplification consisted in assuming that the remunerations of the various factors entering into marginal cost all change in the same proportion. But in fact the rates of remuneration of different factors in terms of money will show varying degrees of rigidity and they may also have different elasticities of supply in response to changes in the money-rewards offered. If it were not for this, we could say that the price-level is compounded of two factors, the wage-unit and the quantity of employment.

Perhaps the most important element in marginal cost which is likely to change in a different proportion from the wage-unit, and also to fluctuate within much wider limits, is marginal user cost. For marginal user cost may increase sharply when employment begins to improve, if (as will probably be the case) the increasing effective demand brings a rapid change in the prevailing expectation as to the date when the replacement of equipment will be necessary.

Whilst it is for many purposes a very useful first approximation to assume that the rewards of all the factors entering into marginal prime-cost change in the same proportion as the wage-unit, it might be better, perhaps, to take a weighted average of the rewards of the factors entering into marginal prime-cost, and call this the cost-unit. The cost-unit, or, subject to the above approximation, the wage-unit, can thus be regarded as the essential standard of value; and the price-level, given the state of technique and equipment, will depend partly on the cost-unit, and partly on the scale of output, increasing, where output increases, more than in proportion to any increase in the cost-unit, in accordance with the principle of diminishing returns in the short period. We have full employment when output has risen to a level at which the marginal return from a representative unit of the factors of production has fallen to the minimum figure at which a quantity of the factors sufficient to produce this output is available.


When a further increase in the quantity of effective demand produces no further increase in output and entirely spends itself on an increase in the cost-unit fully proportionate to the increase in effective demand, we have reached a condition which might be appropriately designated as one of true inflation. Up to this point the effect of monetary expansion is entirely a question of degree, and there is no previous point at which we can draw a definite line and declare that conditions of inflation have set in. Every previous increase in the quantity of money is likely, in so far as it increases effective demand, to spend itself partly in increasing the cost-unit and partly in increasing output.

It appears, therefore, that we have a sort of asymmetry on the two sides of the critical level above which true inflation sets in. For a contraction of effective demand below the critical level will reduce its amount measured in cost-units; whereas an expansion of effective demand beyond this level will not, in general, have the effect of increasing its amount in terms of cost-units. This result follows from the assumption that the factors of production, and in particular the workers, are disposed to resist a reduction in their money-rewards, and that there is no corresponding motive to resist an increase. This assumption is, however, obviously well founded in the facts, due to the circumstance that a change, which is not an all-round change, is beneficial to the special factors affected when it is upward and harmful when it is downward.

If, on the contrary, money-wages were to fall without limit whenever there was a tendency for less than full employment, the asymmetry would, indeed, disappear. But in that case there would be no resting-place below full employment until either the rate of interest was incapable of falling further or wages were zero. In fact we must have some factor, the value of which in terms of money is, if not fixed, at least sticky, to give us any stability of values in a monetary system.

The view that any increase in the quantity of money is inflationary (unless we mean by inflationary merely that prices are rising) is bound up with the underlying assumption of the classical theory that we are always in a condition where a reduction in the real rewards of the factors of production will lead to a curtailment in their supply.


With the aid of the notation introduced in Chapter 20 we can, if we wish, express the substance of the above in symbolic form.

Let us write MV  =  D where M is the quantity of money, V its income-velocity (this definition differing in the minor respects indicated above from the usual definition) and D the effective demand. If, then, V is constant, prices will change in the same proportion as the quantity of money provided that ep ( = (Dpd) / (pdD)) is unity. This condition is satisfied (see Chapter 20 above) if eo  =  0 or if ew  =  1. The condition ew  =  1 means that the wage-unit in terms of money rises in the same proportion as the effective demand, since ew  =  (DdW) / (WdD) and the condition eo  =  0 means that output no longer shows any response to a further increase in effective demand, since eo  =  (DdO) / (OdD). Output in either case will be unaltered. Next, we can deal with the case where income-velocity is not constant, by introducing yet a further elasticity, namely the elasticity of effective demand in response to changes in the quantity of money,

ed  =  ¾¾¾¾
This gives us
¾¾¾¾  =  ep × ed  where  ep  =  1 - ee × eo(1 - ew);
so that  e  =  ed - (1 - ew)ed × eeeo
    =  ed(1 - ee eo + ee eo × ew)
where e without suffix (=  (Mdp) / (pdM)) stands for the apex of this pyramid and measures the response of money-prices to changes in the quantity of money.

Since this last expression gives us the proportionate change in prices in response to a change in the quantity of money, it can be regarded as a generalised statement of the quantity theory of money. I do not myself attach much value to manipulations of this kind; and I would repeat the warning, which I have given above, that they involve just as much tacit assumption as towhat variables are taken as independent (partial differentials being ignored throughout) as does ordinary discourse, whilst I doubt if they carry us any further than ordinary discourse can. Perhaps the best purpose served by writing them down is to exhibit the extreme complexity of the relationship between prices and the quantity of money, when we attempt to express it in a formal manner. It is, however, worth pointing out that, of the four terms ed, ew, ee and eo upon which the effect on prices of changes in the quantity of money depends, ed stands for the liquidity factors which determine the demand for money in each situation, ew for the labour factors (or, more strictly, the factors entering into prime-cost) which determine the extent to which money-wages are raised as employment increases, and ee and eo for the physical factors which determine the rate of decreasing returns as more employment is applied to the existing equipment.

If the public hold a constant proportion of their income in money, ed  =  1; if money-wages are fixed, ew  =  0; if there are constant returns throughout so that marginal return equals average return, ee eo  =  1; and if there is full employment either of labour or of equipment, ee eo  =  0.

Now e  =  1, if ed  =  1, and ew  =  1; or if ed  =  1, ew  =  0 and ee × eo  =  0; or if ed  =  1 and eo  =  0. And obviously there is a variety of other special eases in which e  =  1. But in general e is not unity; and it is, perhaps, safe to make the generalisation that on plausible assumptions relating to the real world, and excluding the case of a 'flight from the currency' in which ed and ew become large, e is, as a rule, less than unity.


So far, we have been primarily concerned with the way in which changes in the quantity of money affect prices in the short period. But in the long run is there not some simpler relationship?

This is a question for historical generalisation rather than for pure theory. If there is some tendency to a measure of long-run uniformity in the state of liquidity-preference, there may well be some sort of rough relationship between the national income and the quantity of money required to satisfy liquidity-preference, taken as a mean over periods of pessimism and optimism together. There may be, for example, some fairly stable proportion of the national income more than which people will not readily keep in the shape of idle balances for long periods together, provided the rate of interest exceeds a certain psychological minimum; so that if the quantity of money beyond what is required in the active circulation is in excess of this proportion of the national income, there will be a tendency sooner or later for the rate of interest to fall to the neighbourhood of this minimum. The falling rate of interest will then, cet. par., increase effective demand, and the increasing effective demand will reach one or more of the semi-critical points at which the wage-unit will tend to show a discontinuous rise, with a corresponding effect on prices. The opposite tendencies will set in if the quantity of surplus money is an abnormally low proportion of the national income. Thus the net effect of fluctuations over a period of time will be to establish a mean figure in conformity with the stable proportion between the national income and the quantity of money to which the psychology of the public tends sooner or later to revert.

These tendencies will probably work with less friction in the upward than in the downward direction. But if the quantity of money remains very deficient for a long time, the escape will be normally found in changing the monetary standard or the monetary system so as to raise the quantity of money, rather than in forcing down the wage-unit and thereby increasing the burden of debt. Thus the very long-run course of prices has almost always been upward. For when money is relatively abundant, the wage-unit rises; and when money is relatively scarce, some means is found to increase the effective quantity of money.

During the nineteenth century, the growth of population and of invention, the opening-up of new lands, the state of confidence and the frequency of war over the average of (say) each decade seem to have been sufficient, taken in conjunction with the propensity to consume, to establish a schedule of the marginal efficiency of capital which allowed a reasonably satisfactory average level of employment to be compatible with a rate of interest high enough to be psychologically acceptable to wealth-owners. There is evidence that for a period of almost one hundred and fifty years the long-run typical rate of interest in the leading financial centres was about 5 per cent, and the gilt-edged rate between 3 and 3½ per cent; and that these rates of interest were modest enough to encourage a rate of investment consistent with an average of employment which was not intolerably low. Sometimes the wage-unit, but more often the monetary standard or the monetary system (in particular through the development of bank-money), would be adjusted so as to ensure that the quantity of money in terms of wage-units was sufficient to satisfy normal liquidity-preference at rates of interest which were seldom much below the standard rates indicated above. The tendency of the wage-unit was, as usual, steadily upwards on the whole, but the efficiency of labour was also increasing. Thus the balance of forces was such as to allow a fair measure of stability of prices;¾the highest quinquennial average for Sauerbeck's index number between 1820 and 1914 was only 50 per cent above the lowest. This was not accidental. It is rightly described as due to a balance of forces in an age when individual groups of employers were strong enough to prevent the wage-unit from rising much faster than the efficiency of production, and when monetary systems were at the same time sufficiently fluid and sufficiently conservative to provide an average supply of money in terms of wage-units which allowed to prevail the lowest average rate of interest readily acceptable by wealth-owners under the influence of their liquidity-preferences. The average level of employment was, of course, substantially below full employment, but not so intolerably below it as to provoke revolutionary changes.

To-day and presumably for the future the schedule of the marginal efficiency of capital is, for a variety of reasons, much lower than it was in the nineteenth century. The acuteness and the peculiarity of our contemporary problem arises, therefore, out of the possibility that the average rate of interest which will allow a reasonable average level of employment is one so unacceptable to wealth-owners that it cannot be readily established merely by manipulating the quantity of money. So long as a tolerable level of employment could be attained on the average of one or two or three decades merely by assuring an adequate supply of money in terms of wage-units, even the nineteenth century could find a way. If this was our only problem now¾if a sufficient degree of devaluation is all we need¾we, to-day, would certainly find a way.

But the most stable, and the least easily shifted, element in our contemporary economy has been hitherto, and may prove to be in future, the minimum rate of interest acceptable to the generality of wealth-owners. If a tolerable level of employment requires a rate of interest much below the average rates which ruled in the nineteenth century, it is most doubtful whether it can be achieved merely by manipulating the quantity of money. From the percentage gain, which the schedule of marginal efficiency of capital allows the borrower to expect to earn, there has to be deducted (1) the cost of bringing borrowers and lenders together, (2) income and sur-taxes and (3) the allowance which the lender requires to cover his risk and uncertainty, before we arrive at the net yield available to tempt the wealth-owner to sacrifice his liquidity. If, in conditions of tolerable average employment, this net yield turns out to be infinitesimal, time-honoured methods may prove unavailing.

To return to our immediate subject, the long-run relationship between the national income and the quantity of money will depend on liquidity-preferences. And the long-run stability or instability of prices will depend on the strength of the upward trend ofthe wage-unit (or, more precisely, of the cost-unit) compared with the rate of increase in the efficiency of the productive system.