Jump to ContentJump to Main Navigation
Keynes's VisionA New Political Economy$

Athol Fitzgibbons

Print publication date: 1990

Print ISBN-13: 9780198283201

Published to Oxford Scholarship Online: November 2003

DOI: 10.1093/0198283202.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 19 October 2018

The Vision Translated

The Vision Translated

Chapter:
(p.133) 8 The Vision Translated
Source:
Keynes's Vision
Author(s):

Athol Fitzgibbons

Publisher:
Oxford University Press
DOI:10.1093/0198283202.003.0008

Abstract and Keywords

Considers the differences between Keynes's economic theories and the Keynesian economics that evolved in the second half of the twentieth century. The difference is between a non‐mechanistic system (Keynes) and a mechanistic one.

Keywords:   econometrics, Keynesian economics, macroeconomics, Thomas Malthus, mathematical economics, David Ricardo

The crisis [in macroeconomics] results from the problem that Keynesian macroeconomics has never been anchored in the normal paradigm of theoretical economics . . . That is one side of a crisis that has been smouldering for many years. On the other side the [neo‐classical] paradigm . . . fails to explain facts that everybody can see.

J. Tobin, in A. Klamer, The New Classical Macroeconomics

Between Keynes and the Keynesians who inherited his theories there lay a wide divide. His philosophy they did not dream of; their philosophy he knew and rejected. Yet the Keynesians had to bridge the divide, and to translate his ideas into their mechanics. The practical consequence was the Keynesian period from about 1945 to 1970, which, to judge by its results only, has probably been the most successful of all, with low unemployment, low inflation and reasonable growth. Yet, cut off from Keynes's vision, Keynesian economic theory was inconsistent and intellectually indefensible.

Keynes's theory was soon replaced by Keynesianism, and Keynesianism was in turn replaced by neo‐classical economics, which regarded Keynes's theory as a useful but inaccurate snapshot of an economy that was best portrayed by differential equations. But neo‐classical theory, which praised the usefulness of Keynes's theories and yet contested their validity, was itself between two poles of thought. Monetarism went to the other pole when it drew its formal theory from classical economists, without paying obeisance to Keynes. There has been a tendency to move away from the economics of Keynes, but because of the growing disillusionment with monetarism there has also been a new interest in his method. However, this new interest in Keynes will be unproductive unless it is recognized that he departed from (p.134) conventional thought at an epistemological point outside economics. ‘Post Keynesianism’, wrote Robert Solow (1979:344), ‘seems to be more a state of mind than a theory.’ That indeed was how Keynes understood it, and ‘our minds have not yet met’ was one of his phrases, although by a state of mind Solow meant a state of prejudice or partiality, which is not what Keynes meant.

Ricardo had understood economics as the science of equilibrium. After Keynes, the neo‐classical economists came to understand it as the mechanics of a transitional path to equilibrium, but they still had a notion of an economy being definable by a set of equations. In Keynes's economics there is no general mathematical theory because neither the variables nor the parameters behind the equations can be comprehensively defined. ‘As soon as one is dealing with the influence of expectations and of transitory experience,’ he wrote immediately after the General Theory, ‘one is, in the nature of things, outside the realm of the formally exact’ (XIV:2).

Since the difference between Keynes and the Keynesians has already been analysed in two challenging books by Paul Davidson (1972, 1982), I will briefly indicate where I depart from him. Davidson targets the conventional Keynesians, because they combine Keynes's economics with classical economics. I argue that none of Keynes's followers, not even Davidson, have adopted Keynes's theory of probability or his moral economic science. I agree that the conventional Keynesians misunderstood Keynes because they saw him through classical eyes, but the neo‐Keynesians also distorted Keynes, because to really follow Keynes's theory of uncertainty they would have to abandon Ricardo and Marx. Both the neo‐Keynesian socialists and the conventional Keynesians are separated from Keynes by the same chasm, which is deeper than Davidson acknowledges, of Keynes's method, which comes from his idealism. Therefore, unlike Davidson, I am not directly interested in the particular assumptions made by the conventional Keynesians, and my main theme is not their defects but Keynes's own method.

(p.135) Hume and Ricardo

The idea of economics as a moral science originated not with Keynes, but (probably) with the person whom he regarded as a paragon among economists, Thomas Malthus. ‘If only Malthus, instead of Ricardo, had been the parent from which nineteenth‐century economics proceeded, what a much wiser and richer place the world would be today’ (X:101). Malthus, personal friend and intellectual antagonist of Ricardo, spent the first part of his adult life contending that there was a tendency for an excess supply of population, and the second arguing that there was a tendency for the demand to be inadequate to employ the population. Although he is best known for the first part, being the author of the Essay on Population which inspired the evolutionary theories of Charles Darwin, Keynes described Malthus's second theory, concerning inadequate demand, as more far‐reaching. For in order to establish that demand could be systematically inadequate, Malthus developed a new idea of economic science. Malthus, as Marx put it, sinned against science.

‘The science of political economy’, Malthus wrote in the preface to his Principles, ‘bears a nearer relation to the science of morals and politics than to that of mathematics.’ It does not matter that Ricardo, who was the person being accused here of mathematical economics, never wrote an equation. The real question was the degree of abstractedness that is appropriate to economics. Keynes understood the methodological issue in Malthus's terms almost exactly:

According to Malthus's good common‐sense notion prices and profits are primarily determined by something which he described, though none too clearly, as ‘effective demand’. Ricardo favoured a much more rigid approach [than Malthus's effective demand], went behind ‘effective demand’ to the underlying conditions of money . . . and looked on Malthus's method as very superficial. But Ricardo, in the course of simplifying the many successive stages of his highly abstract argument, departed, necessarily and more than he himself was aware, away from the actual facts; whereas Malthus, by taking up the tale much nearer its conclusion, had a firmer hold on what may be expected to happen in the real world. . . . When one has painfully escaped from the intellectual (p.136) domination of these [Ricardian] pseudo‐arithmetical doctrines, one is able, perhaps for the first time for a hundred years, to comprehend the real significance of the vaguer intuitions of Malthus. (X:88)

Malthus's moral science of economics closely resembles the principles of Burke's doctrine of expediency in politics, and his criticisms of Ricardo recall Burke's criticisms of the arbitrary abstractions of the liberals. In essence, Malthus put principle and pragmatism ahead of scientific abstraction; Keynes said that Malthus's approach to economics was through the ‘best of all routes’, namely, through moral and political philosophy.

The moral science was cast more in negative than in positive terms. ‘Certain eminent persons’, Malthus said, ‘fear of destroying the simplicity of a general rule’ and ‘overlook any circumstances that may interfere with the generality of the principle’ (Malthus 1951 edn.:12). He held that the principles of political economy are cast at the wrong level to be truly general and invariant; that there is no general theory of political economy; and that theories, being only partial, must be interpreted according to the changing circumstances, rather than the circumstances abstracted from according to the theory; and he implied that the choice between these theories required a knowledge that was not to be found within political economy. The Ricardian method, he held, overlooked the problem of unemployment and deficient demand in the economy because it violated these principles, which were the principles of pragmatism.

Malthus focused, as much as did Keynes, on the surface phenomena of economic life, and looked for no deeper meanings in economic abstraction. Although like Keynes he recognized the need for formal structures of economic thought, he too stressed the dangers of long chains of economic logic.

I certainly am disposed to refer frequently to things as they are, as the only way of making one's writings practically useful to society, and I think also the only way of being secure from falling into the errors of the taylors of Laputa, and by a slight mistake at the outset arrive at conclusions the most distant from the truth. (Malthus to Ricardo, quoted in X:97)

Keynes made similar criticisms of Ricardo, but the closest parallel was Keynes's complaint that the classical economics of von Hayek was ‘an extraordinary example of how, starting with a (p.137) mistake, a remorseless logician can end up in Bedlam’ (XIII: 252).

Ricardo identified economic science and economic equilibrium. Ricardo is often said to have been the first scientific economist; he was also the first to analyse the economy as though it were in permanent equilibrium. Prices were held by him to be proportional to the money supply, and there was a natural rate of interest, dependent only on real forces and not on financial conditions. When Ricardo said that changes in the money supply ‘would not permanently alter the rate of interest’, the word ‘permanently’ meant ‘according to the theory’. For as Keynes noted, Ricardo believed that the surface of economic life was sub‐rational and outside the proper scope of economic science.

For Ricardo expressly repudiated any interest in the amount of the national dividend, as distinct from its distribution. In this he was assessing correctly the character of his own theory. But his successors, less clear‐sighted, have used the classical theory in discussions concerning the causes of wealth. Vide Ricardo's letter to Malthus of October 9, 1820: ‘Political Economy you think is an enquiry into the nature and causes of wealth—I think it should be called an enquiry into the laws which determine the division of the produce of industry amongst the classes who concur in its formation. No law can be laid down respecting quantity, but a tolerably correct one can be laid down respecting proportions. Every day I am more satisfied that the former enquiry is vain and delusive, and the latter only the true objects of the science.’ (GT:4)

Ricardo's belief that economic science required economic equilibrium led him to criticize the older classical economic theory, which had accounted for both equilibrium and the path to it, the theory that Keynes himself held in the Tract. In this theory, eventually prices would be proportional to the money supply, and eventually the interest rate would disengage from monetary influences and reach its natural level. In the meantime, although there is change, there are the reference points of equilibrium, beacons in the river of change, which are all that a science of economics requires. This theory was not only older than the Ricardian, but it had impeccable scientific credentials, because its author was the person whom Keynes regarded as the first classical economist of all—David Hume, philosopher, now in his manifestation of political economist: (p.138)

Hume a little later [after Locke] had a foot and a half in the classical world. For Hume began the practice among economists of stressing the importance of the equilibrium position as compared with the ever‐shifting transition towards it, though he was still enough of a mercantilist not to overlook the fact that it is in the transition that we actually have our being. (GT:343)

Although Hume believed that money had ‘chiefly a fictitious value’, he also acknowledged, which Ricardo did not, that a decrease in its amount would lead to a temporary deficiency of demand. ‘There is always an interval before manners be adjusted to their new situation; and this interval is pernicious to industry when gold and silver are diminishing’ (Hume 1955:40). In this, Hume may have been true to his own method, for as Keynes once stressed, in the Treatise on Probability, Hume was not deceived by the superficialities of mathematicians into thinking that the law of large numbers can solve the problems of causation. But Ricardo held it to be ‘erroneous view of Mr Hume’ that money can call forth goods (Sraffa 1951–73, V:524), and Alfred Marshall and his disciples followed the Ricardian assumption of permanent equilibrium.

Classical economics was therefore a two‐headed beast, with a Ricardian version that assumed permanent equilibrium and Hume's method, which assumed only an eventual equilibrium. Keynes especially objected to the particular assumption of the Ricardian system, its tendency to assume away the transition. However, he also objected in a more general way to both versions of classical economics.

Mathematical Economics

‘Human society, when we contemplate it in a certain abstract and philosophical light, appears as a great, an immense machine, whose regular and harmonious movements produce a thousand agreeable effects’. So wrote Adam Smith (1759:316), and economists still tend to think of the economy as being like a machine. There is a set of levers, such as government policy, expectations, etc., called the independent variables; and a set of outputs, such as prices, called the dependent variables. A theory of the economy is understood to mean writing down the (p.139) blueprints of the machine, which may be why a theory is referred to as a ‘model’ of the economy. The same metaphor applies to neo‐Keynesian economics, although it is not quantitative, and we cannot clearly read the dials on the machine.

When Keynes said that economics was not a natural science, he meant partly that the economy could not be specified as a comprehensive blueprint, because the mechanical metaphor does not apply. ‘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 organized and orderly method of thinking out particular problems. . . . This is the nature of economic thinking’ (GT:297). By this ‘method of thinking’ Keynes meant not inferential thinking, but probabilistic thinking, where strict inference does not always work. Although Keynes's opposition to mathematical economics, or more strictly to abstract economics, has been painted as idiosyncratic, it was part of a well defined philosophy, which is particularly crucial to understanding the General Theory.

In the first place, because Keynes regarded the economy as an unbounded and yet interrelated swirl of events, he did not think that it was meaningful to distinguish between the dependent and independent variables of the system, between the output and the levers. There are no truly independent variables, because everything influences everything else, and no matter how general the economic theory it cannot pick up all the interactions. In theory, we make the division between dependent and independent variables, but it is only a mental division. ‘The division of the determinants of the economic system into the two groups of given factors and independent variables is, of course, quite arbitrary from any absolute standpoint. The division must be made entirely on the basis of experience.. . . ’ (GT:247). Therefore he held that it was not possible to isolate the effects of changing one of the variables of an economic system, because the supposedly independent variables were all interrelated, although in different ways at different times.

Second, according to Keynes, economics cannot deal with the real workings of the economy, because it does not deal with the real atomic causal units, which are judgements of value and phantasmagoric fears and desires that are unquantifiable. In his essay on Edgeworth, Keynes says that ‘the atomic hypothesis (p.140) which has worked so splendidly in Physics breaks down in Psychics’, meaning in mathematical economics; and in the General Theory he warns that the independent variables of the theory are only aggregations, and that ‘these again would be capable of being subjected to further analysis, and are not, so to speak, our ultimate atomic independent elements’ (GT:247).

Rather than dealing with causa essendi, or causae verae, as Keynes also called them, economics can deal only with categories that are consolidated beyond the point where true causation can apply. The variables of economic theory are ‘fallible indexes, dubious approximations at that, with much doubt added as to what, if anything, they are indexes or approximations of’ (X:262). The economic machine which the economists typically have in mind is never the real machine, the workings of which cannot be represented, because economics is always expressed in the wrong terms for true causation.

Finally, the unboundedness of economic theory and the impossibility of true causation make the economy so complex that in principle it is not representable by mathematics. Because economic categories are only approximate and analogical, while the influences upon them are complex and unbounded, mathematical reasoning cannot apply. Given these insubstantial relationships and fluid categories, the best that mathematical economics can do is to illustrate the complexities of the economy by demonstrating its own limits.

It is, I think, of the essential nature of economic exposition that it gives, not a complete statement, which even if it were possible, would be prolix and complicated to the point of obscurity but a sample statement, so to speak, out of all the things which could be said. (XIII:470)

I do not myself attach much value to [mathematical] manipulations of this kind; and I would repeat the warning . . . that they involve just as much tacit assumption as to what 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. (GT:305)

Probably this is why Keynes declared it important that a (p.141) theory should not be quantified, because a quantified theory ‘loses its generality and its value as a mode of thought’. To quantify a theory is to tear the theory from its context, when what is important is the relation between the theory and the context. ‘Progress in economics consists almost entirely in a progressive improvement in the choice of models’, which is to say progress occurs through new and more revealing orderings of the swirl of economic data. The role of economic theory was to present to the intuition material that had been organized for assimilation. Keynes believed economic theory should be ‘lean’, its role being to convey only a general insight. He had the opposite to the usual positivist idea, which represents intuition as the handmaid of science, mainly providing hypotheses for testing. In his philosophy, intuition is paramount over science as a criterion of truth.

Hence the extreme complexity of the actual course of events. Nevertheless, these seem to be the factors which it is useful and convenient to isolate. If we examine any actual problem along the lines of the above schematism, we shall find it more manageable; and our practical intuition (which can take account of a more detailed complex of facts than can be treated on general principles) will be offered a less intractable material upon which to work. (GT:249)

Similarly, Keynes did not conceive of economics as a giant theory unified in all of its aspects, with economic galaxies built out of the economic atoms. If economics is not a natural science, based on real atomic units, it cannot be unified. ‘Between two different ideas we would examine,’ said Locke, ‘we cannot always find such mediums as we can connect one to another in all parts of the deduction’ (Locke 1690, Bk. 4, iii (4)). Keynes's macroeconomics seemingly drew its assumptions out of thin air, or at least without reference to any micro theory, because it was based on a probabilistic logic.

Keynes accepted that there were two entirely different theories of economics, one at the micro level in terms of demand and supply, and another at the macro level involving money. He did not try to reconcile the two accounts, his only conclusion being that it was necessary to have a correct subdivision of economics.

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 (p.142) route or journey connects them, related, apparently, after the fashion of our waking and our dreaming lives. . . . 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. (GT:292–3)

The obvious objection to Keynes's philosophy of action is that there will be different intuitions held by different people, so that relying on intuition means that the advance of knowledge will be halted. Keynes certainly did not agree that progress would be halted, but if we recall the Treatise on Probability, we do not have any choice. We draw upon intuition not only for our hypotheses, but also to interpret the tests and very meanings of theories, because exactitude is beyond us. If this means that we must for ever live in a twilight of probability, then, to transpose the theme in the Treatise on Probability, it is best to recognize our limitations and act upon them instead of representing to ourselves that our methods of knowledge are more powerful than they actually are.

Keynes's opposition to mathematical economics, or more strictly to long chains of abstract reasoning, should not be dismissed as a character quirk or a peevish reaction to his own failure to develop a mathematical theory of economics. His opposition long pre‐dates his interest in formal economic theory. He satirized mathematical economics a decade before the General Theory in his early essays on Marshall and Edgeworth, the latter of whom he regarded as the leading mathematical economist and econometrician of his day:

All his intellectual life through he felt his foundations slipping away from under him. . . . Edgeworth knew that he was skating on thin ice; and as life went on his love of skating and his distrust of the ice increased, by a malicious fate, pari passu. He is like one who seeks to avert the evil eye by looking sideways, to escape the censure of fate by euphemism, calling the treacherous sea Euxine and the unfriendly guardians of truth the kindly ones.1 (X:262)

We can trace Keynes's opposition to mathematics in the social sciences generally back to the Treatise on Probability, where he expressed his preference for systems of words rather than symbols, because words contain a power of metaphorical suggestion. ‘I . . . have not the same lively hope as Concordet, or (p.143) even as Edgeworth, “éclairer les Sciences morales et politiques par le flambeau de l'Algèbre”.’ Going back even further, mathematics presumably would not apply in an organic unity, where the secondary ‘reflex values’ would make any analysis intolerably difficult. Right or wrong, Keynes's opposition to mathematical economics was part of a consistent philosophy which attempted to bring formal thought to bear on probabilistic experience.2

A common criticism is that Keynes followed Alfred Marshall's method of partial equilibrium economics, instead of general equilibrium—meaning that Keynes theorized about a subsection of the economic machine, whereas modern general equilibrium economics examines the whole machine with all of its feedback loops. However, this division between partial and general again assumes that the economy is a machine rather than an organic unity, and Keynes did not think himself that he had the same method as Marshall. Keynes approved of Marshall's saying that economics can be understood as a set of principles, rather than as a set of theories:

‘Ricardo and his chief followers did not make clear to others, it was not even quite clear to themselves, that what they were building up was not universal truth, but machinery of universal application in the discovery of a certain class of truths. While attributing high and transcendent universality to the central scheme of economic reasoning, I do not assign any universality to economic dogmas. It is not a body of concrete truth, but an engine for the discovery of concrete truth.’ (Marshall, quoted in X:196)

But Keynes regretted that Marshall did not actually follow his own precept, and he regarded Marshall's theories as deterministic (X:199). In the above quote, Keynes noted, Marshall was complaining about socialist economics, and he did not apply the moral to his own method.

The Celestial System of Economics

The main point is that Keynes did not believe that the whole economic system could be represented in principle by a set of simultaneous equations. A mathematical system that expressed (p.144) his economic theory would have to be suited to modelling Chaos as well as allowing the variables to change their meanings in mid‐stream. ‘The material to which it [economics] is applied is, in too many respects, not homogeneous through time’ (XIV:296). I have stressed the point because Keynesian economics was discredited when it was translated into a mathematical system which was found to be defective.

Keynes attributed the general idea of ‘a whole Copernican system, by which all of the elements of the economic universe are kept in their places by mutual counterpoise and interaction’ (X:205), to Alfred Marshall. ‘Just as the motion of every body in the solar system’, Marshall had said, ‘affects and is affected by the motion of every other, so it is with the elements of the problem of political economy.’ But unknown to Keynes, in 1871, one year before Marshall wrote that sentence, Léon Walras had developed an economic theory which actually was inspired by a treatise on celestial mechanics. Unlike Marshall, Walras conceived of a single embracing economic theory which could cover the whole economic universe.

A corollary to this unbounded celestial theory is Walras's Law, which says that, if all but one of the markets in the economic universe is known, then the last one, whatever it might be, can be mathematically deduced from the others. This law is in direct contradiction with Keynes's theory, because, whereas Keynes said in the General Theory that the rate of interest depends only on what happens in the money market, Walras's Law means that a macroeconomic theory cannot be cast in terms of just one market, because everything in the economic universe influences everything else. According to Walras's Law, the money market could be left out of economic theory and the interest rate could be explained by the equations of all of the other market orbits. The question was not whether everything in the economy could influence everything else, which all sides agreed, but whether a scientific economics could embrace all the effects.

J. R. Hicks introduced Walras's Law into macroeconomics as part of a criticism of Keynes. Hicks said, soon after the General Theory, that the money market can be regarded as the last equation, which is simply omitted from the mathematical solution of the economic system; and, QED, the interest rate can (p.145) be accounted for in a theory without a money market (Hicks 1936/1982:92).

The ordinary method of economic theory would be to regard each price as determined by the demand and supply equation for the corresponding commodity or factor, the rate of interest as determined by the demand and supply for loans. If we work in this way, the equation for demand and supply of money is otiose—it follows from the rest; and fortunately, too, it is not wanted, because we have determined the whole price system without it. (Quoted by Keynes in XIV:204)

If Walras's Law is true, then, as Hicks implied, Keynes's money rate of interest is only an empty formalism. The interest rate was really set directly by the demand and supply of loans; but Keynes had left loans out of his theory, making the loans market the one orbit to be deduced from all of the rest. It was logically correct, but artificial, to say that the rate of interest was determined in the money market when it was really the price of loans.

But we could equally well work in another way . . . we could allot to the rate of interest the equation for the demand and supply of money. If we do this, the equation for loans becomes otiose, automatically following from the rest. . . . The latter is the method of Mr Keynes. (Quoted by Keynes in XIV:204)

Hicks assumed that Keynes was formulating a theory in the celestial science of economics, whereas Keynes believed that there cannot be an all‐embracing economic science, that science is limited by the wealth of possibly relevant detail. Of course, if it were valid to abstract from the detail, to combine great market systems into even greater market constellations, then a celestial system of economics would be more feasible, and so the issue was what degree of consolidation is admissible in economic theory. Hicks had said that the interest rate is determined in the first instance by the demand and supply for loans, but ‘loans’ is a great congeries of many different things, and in ‘Alternative Theories of the Role of Interest’, which was written in defence of his monetary theory of interest, Keynes simply dismissed Hicks, because ‘the meaning of . . . “loans” is not defined’ (XIV:204). He proceeded to analyse the meaning of ‘credit’, on the grounds that Hicks and other economists had used a portmanteau word (p.146) to confuse distinct terms, and thereby advance a classical theory of interest.

Hicks treated the moving constellation of interest rates at the centre of the heavens as a single point, but Keynes saw within it a single star much brighter than the others. Keynes's method was to bring that star into his theory, while keeping open his intuition as to whether other stars might come at times to cast a strong influence. In his theory there was not just the one interest rate but a spectrum of rates of return on many assets, of which one was more important than the rest.

There is no reason why their rates of interest should be the same for different commodities—why the wheat‐rate of interest should be equal to the copper‐rate of interest. For the relation between the ‘spot’ and ‘future’ contracts, as quoted in the market, is notoriously different for different commodities. This, we shall find, will lead us to the clue we are seeking. For it may be that it is the greatest of the own‐rates of interest (as we may call them) which rules the roost . . . and that there are reasons why it is the money‐rate of interest which is often the greatest. . . . (GT:223)

Abstracting from money returns, there were own rates of interest on wheat, houses, copper and other assets, all interrelated in an elastic way by differentials which were changeable and which Keynes did not try to explain, because he was concerned only with the more important variables. The significance of the own rate of interest on money, which equalled the liquidity advantage of holding money, was that liquidity preference did not fall as other assets' own rates fell in attractiveness during a slump. Therefore the money rate of interest set a more or less elastic minimum below which other interest rates could not fall.

It is only in this sense that the money rate of interest in the General Theory represents the rate. It is the most strategic rate, but not the generic rate of interest. A strategic rate implies that the surface economy is real; it is only the most important interest rate in the circumstances, and it does not try to explain everything. The generic rate of interest implies that the hidden underlying economy is real; it is the general rate relevant for the scientific analysis of regular or long‐run economic behaviour. It embraces all the rates of interest because it is supposed to capture their essential meaning.

(p.147) The Conventional Keynesians

The pre‐modern philosophy that underlies the General Theory is not easily communicable; Keynes's critics did not understand the point of his Treatise on Money, wrote Harrod in his biography, but I have suggested that neither did Harrod fully grasp Keynes's idea of indeterminancy. To become assimilable, the General Theory had to undergo a perverse translation of its vision into the methods of natural science.

Keynes, following a suggestion from Harrod, declared the classical theory to be one equation short because it did not take account of income as a variable influence upon savings. The classical economists saw the economy as a plane when in fact it was a cube. If this were Keynes's only objection to the classical theory, then the obvious response would be to put a third line into the cube, and make the classical theory mathematically determinate. Keynes, who had also stressed that the classical theory of interest was a nonsense theory beyond redemption, did not take that step; but Hicks did. It lead to the IS–LM system, the Keynesian theoretical structure, the conventional version of Keynes (Hicks 1937).

We must differentiate between Keynes's monetary theory of interest, which has been supplanted and forgotten, and the Is–LM or Keynesian theory of interest, which has taken its place. The Keynesian theory of interest is partly Keynes and partly the classical theory of interest. Classical theory had presented interest as being determined in one market—the savings and investment market—and the Keynesian theory of interest combined this classical theory of interest rates with Keynes's own monetary theory of interest. In the IS–LM Keynesian theory, the interest rate is in equilibrium if the demand for money equals the supply of money, and if savings equals investment. It was not Keynes's theory, because he anticipated and rejected the Keynesian theory by insisting that the interest rate was unrelated to savings and investment:

I am substituting demand and supply analysis for liquidity instead of that for savings. Marshall you say ‘thought interest was determined by the schedule of marginal efficiencies and the schedule of the propensity to save’; he forgot that incomes could change. I am saying something (p.148) totally different from this when I say that interest is determined by the demand and supply for liquidity. (XIII:550)

His argument continued to the effect that, if there were a change in the propensity to save, there would be no reason thereby to expect a change in the demand and supply of liquidity, and so the rate of interest would not change. The equilibrium of savings and investment would be ensured not by a fall in interest, but only by a fall in employment and incomes.

The General Theory was not a simultaneous equation theory, because its line of causation ran in only one direction. Given expectations, the money supply determined interest rates; given expectations again, interest rates determined investment; and investment spending then set the total demand for commodities in the economy. The line of causation was therefore as follows:

                      The Vision Translated

It was, as Keynes thought it should be, a very lean theory, but it had to be intuitively supplemented by a particular feedback loop, which Keynes called ‘backwash’. If demand increased in the economy, there would be more need for money to finance transactions, and this money would be siphoned off from the financial and capital markets. The capital markets would then find money more scarce, and the rate of interest would rise, partly offsetting the initial increase in demand. The extra feedback loop would therefore look like this:

                      The Vision Translated

Keynes did not try to represent this feedback loop formally, because he believed that in reality backwash would be inconstant and would fluctuate in force. However, he had to debate against Ralph Hawtrey and D. H. Robertson, who both argued (XIV:12 and 231) that there must be a fatal flaw in the monetary theory of interest if the feedback effect could not be formalized. By formalizing the feedback Hicks ended that line of criticism, and the theory that Keynes had said earlier (XIII:550) was ‘totally (p.149) different’ to his own he now found unobjectionable. ‘I found it [the IS–LM theory] very interesting and have nothing really to say by way of criticism’ (XIV:79). However, Keynes added that the feedback loop ‘need not raise the rate of interest’, even if it was likely to, and shortly after he reaffirmed the monetary rate of interest (XIV:201). Keynes had not abandoned his theory or his method in favour of Hicks.

The IS–LM theory did illustrate the backwash effect, but it was taken out of context and its significance greatly overstated. The theory encouraged the neo‐classical economists to extend their mathematical systems to absorb Keynes, and, as Hicks said later (1976:140–1), these extensions never got to the point. However, the neo‐classical Keynesians were misled not so much by IS–LM, as by their tendency to concentrate on the mechanics of any theory, instead of its spirit. By reducing the General Theory to simple equations and slogans, they popularized the theory; but when it was disconnected from Keynes the theory was illogical, and vulnerable to a classical counter‐attack.3

For IS–LM was an inconsistent combination, and the theory fell into theoretical disrepute, because a long‐run equilibrium rate of interest that depends on the money supply does not make sense.4 If there is such a thing as fullest equilibrium, then the equilibrium interest rate should be independent of the money supply. Start with long‐run equilibrium and let the money supply change. Abstract from problems of subnormal dispersion the problem of who gets the extra money—the rich, the poor or the enterprising; assume that all changes are eventually worked out through the economy; assume that there are no other changes in the meantime, and that everyone knows that there are no other changes; then the eventual interest rate will be unaffected by the money supply. Multiplying the money supply a hundredfold should not affect the underlying situation, as Ricardo has said. The only question is whether it is meaningful or profitable to think in such abstract equilibrium terms. The IS–LM system was a bastard theory, as Joan Robinson called it, because it drew upon two opposing philosophies to formulate a theory that was consistent with neither of them. The significance of this was that the IS–LM theory, with all its contradictions, became official Keynesianism, the rationale to regulate the macroeconomies of the world.

(p.150) After the Keynesians

When economists realized that the IS–LM Keynesian system did not make sense, they first turned to the economics of Hume, which Don Patinkin had refurbished (Patinkin 1965). Patinkin assumed that there is a long‐run equilibrium in the economy which is unblurred by uncertainty and independent of money, but he recognized nevertheless that the economy is not usually at this equilibrium, that we have our being in the transition. He believed that Keynes had taken a theoretical snapshot of the transition, without realizing that the economy was on its way to a definite equilibrium according to mathematical laws.

This new version of an old classical theory treated money as just another commodity and assumed that there is a natural rate of interest which must be reached for full economic equilibrium. It was recognized that there would normally be financial movements around the natural interest rate, and unemployment was said to arise from the time the economy took to get to its natural level, but in the long run the real economy was independent of the amount of money. But while Patinkin believed that he had generalized Keynes's General Theory, he had really taken a step away from its new spirit; for a theory that assumes away uncertainty is not Keynes's, and Patinkin had reintroduced the underlying real economy, hidden from unscientific eyes.

Axel Leijonhufvud set out to construct a theory based more closely on Keynes, but, again, cast in the language and method of classical economics (Leijonhufvud 1968). As Leijonhufvud put it, the important difference between Keynes and classical economics is that Keynes did not assume ‘Walras's auctioneer’; meaning that Keynes did not assume that buyers and sellers know the equilibrium price of anything. He believed that Keynes had understood the economy as a cybernetic system, in which information is not just one more saleable commodity in the system, but is something that influences the structure of the system. However, if there is ignorance about equilibrium prices, then prices will not immediately adjust to bring demand and supply together, and quantities in the economy, such as (p.151) investment and employment, will change as well. Unemployment is therefore due to ignorance and uncertainty.

There is a Keynesian flavour to this theory, and in particular there is a parallel with Keynes's Treatise on Probability, with which Leijonhufvud was evidently unacquainted. The quantity adjustments which he says arise out of a cybernetic economic system with partial ignorance are reminiscent of how investment is influenced by the objective chance that arises out of Keynes's veil of probability, philosophic ideas which I say have been transposed to Keynes's economics. It is also consistent with Keynes that Leijonhufvud's emphasis is on method and not on the errors of capitalism—he is critical of attempts to fit Keynes into a deterministic clockwork system, and in his account uncertainty arises out of the nature of knowledge in general, rather than out of the capitalist class struggle.

Nevertheless, Leijonhufvud nowhere acknowledged that Keynes primarily wanted to reformulate economics in terms of a completely different method, and critics have pointed out that he tried to pour his new Keynesian wine into the old classical bottle. ‘We have repeatedly referred to the need to discover a basis on which a synthesis of “Keynesian” economics and “Classical” value theory can be achieved’, he writes (Leijonhufvud 1968:393), and his theme was that the neo‐classical theory should provide the method and Keynes's theories should be the superstructure. ‘We have consistently viewed Keynes' contributions to economic theory as part of an overall effort to extend the use of the (largely received) tools of general value theory’ (p. 333).

Consequently, Leijonhufvud was unwilling to recognize those aspects of Keynes's theories that come most directly from Keynes's own method, such as the influence of speculation and the monetary theory of interest. For example, Leijonhufvud wrote of the monetary theory of interest: ‘This notion is a bit disturbing . . . the classical factors of “Productivity” and “Thrift” must still be considered’ (p. 174). Later in his book Leijonhufvud denied that Keynes had a monetary theory of interest, and said that Keynes believed in a natural rate of interest, and that Keynes only changed his terminology. ‘Although Keynes relinquished the natural rate terminology of the Treatise, his position underwent no fundamental change’ (p. 349). But the difference between a monetary interest rate and natural rate of (p.152) interest is fundamental, because whether there is a hidden economy underlying the surface hinges on it; and Keynes did not say that he was abandoning the terminology of a natural rate, but the concept of a natural rate of interest.

I am now no longer of the opinion that the concept of a ‘natural’ rate of interest, which previously seemed to me a most promising idea, has anything very useful or significant to contribute to our analysis. It is merely the rate of interest which will preserve the status quo; and, in general, we have no predominant interest in the status quo as such. (GT:243)

By putting uncertainty back in the classical bottle, Leijonhufvud puts bounds on the economic system. There is still a river of change, but it only flows between the banks of equilibrium, and these reference points in Leijonhufvud's theory do not exist in Keynes. His attempt to reformulate Keynes led to a theory of Knightian economics rather than Keynesian economics, an economics in which uncertainty is contained.

These neo‐classical versions of Keynes, IS–LM and the theories of Patinkin and Leijonhufvud, were compromises between logic and common sense, because they did not adequately develop their own assumption that uncertainty and indeterminancy are limited or can be assumed away. If there is no uncertainty, the principle of indifference applies. A new classical economics exploited that assumption to the full with its ideas of a natural rate of unemployment, rational expectations and an optimum quantity of money. For if there will be an economic equilibrium at point X in the future, then we cannot be too far from equilibrium today, being separated only by calculable costs of transition. The effects of subnormal dispersion in the economy can be assumed to cancel out, and the river of change becomes a narrow stream, no wider than the costs of economic information.

If the principle of indifference does not apply, if deviations around an economic mean do not cancel out, if we can make no meaningful quantitative estimate about the likelihood of a war or the price of copper ten years hence, if we cannot overcome uncertainty by any method however onerous and costly, if we simply do not know the future—then Keynes's economics would come into its own again. Yet if the principle of indifference and all that goes with it does apply, it is logical that this principle (p.153) should be exploited fully. Keynes did not argue that Ricardo was illogical in his own context, but to the contrary, he faintly praised Ricardo, because Ricardo alone could follow his logic so unwaveringly.

Ricardo offers us the supreme intellectual achievement, unattainable by weaker spirits, of adopting a hypothetical world remote from experience as though it were the world of experience and then living in it consistently. With most of his successors common sense cannot help breaking in—with injury to their logical consistency. (GT:192)

In the Treatise on Probability, Keynes tried to steer a middle course between Hume and Laplace, between reaching no probable conclusion and reaching probable conclusions too superficially. I believe that the misinterpretation of his middle course in probability has led to the loss of his middle ground in macroeconomics. In any event, the shade of Ricardo has returned. It came with the neo‐Keynesians, when they treated the process of transition in a capitalist economy as sub‐rational. The new classical economics has brought back Ricardo in another form, on the phantasmagoric wings of Laplacian method.5

The approach of the conventional Keynesians was useful and pragmatic in the sense that it made macroeconomic management a politically comprehensible and definite task, the adjustment of the economy from one hypothetical equilibrium to another and better one. However, unlike Keynes, they made no allowance for human folly and ignorance in their theories, and they implicated Keynes's system with their own theoretical inconsistency and philosophical shallowness. Despite their apparent success, despite their reputed pragmatism, the Keynesians had no valid defence against the monetarist logic. The conventional Keynesian school was open to the criticism that its full employment policies could only aggravate a natural level of unemployment, while its central economic management, without reference to a vision or a moral concept, was anti‐democratic and elitist.

(p.154) Econometrics*

Econometrics is the statistical measurement of economic quantities, and mathematical economics is economic theory in mathematical form, without necessarily specifying the quantities. I have suggested that Keynes's attack on mathematical economics in the General Theory was an indispensible part of the Theory. This runs against the received wisdom, because the validity of mathematical economics is not a serious issue among economists, the general view being that whatever can be said in words can also be said in symbols with greater precision and generality. However, economists do have reservations about econometrics, which they think is scientific and yet suspect. For example, although Patinkin simply dismisses Keynes's criticisms of mathematical economics, he is supportive of his criticisms of econometrics (Patinkin 1982).

Consequently, while Keynes's objections to mathematical economics have not been seriously analysed, his objections to econometrics have been analysed at length. Yet Keynes had two objections to econometrics which have not been separated out in the literature. Because these two principles have not been identified, his objections to econometrics have been understood only in particular terms, such as to specification bias, or simultaneous equation bias, whereas the matter can be put in a more simple and general way. The theoretically more fundamental objections to econometrics reduce to his objections to mathematical economics; his practical objections to econometrics arise from his objections to Laplacian methods, that is, to the frequency theory of probability which assumes away the limitations of the principle of indifference. Keynes's objections to econometrics reduce to his objections to classical economics.

Keynes opposed false abstraction and the principle of indifference, but his opposition was not to induction or quantification as such, as his discussion with Harrod of Tinbergen's econometric methods clearly shows. Harrod wrote to Keynes, expressing the (p.155) view that ‘Tinbergen may be doing very valuable work, in trying to reduce this part of the theory to quantitative terms’, to which Keynes replied:

There is really nothing in your letter with which I disagree at all. Quite the contrary, I think it most important, for example, to investigate statistically the order of magnitude of the multiplier, and to discover the relative importance of the various facts which are theoretically possible. (XIV:299)

Keynes's point here was that economics is properly not a pseudo‐natural science, with constant co‐efficients and definite independent variables. He objected to Tinbergen's econometrics first because Tinbergen was trying to interpret economics as a natural science in which the rules of mechanical inference without intuition would apply.

My point against Tinbergen is a different one. In chemistry and physics and other natural sciences the object of experiment is to fill in the actual values of the various quantities and factors appearing in an equation or a formula; and the work when done is once and for all. In economics that is not the case, and to convert a model into a quantitative formula is to destroy its usefulness as an instrument of thought. (XIV:299)

The fundamental objections to econometrics can be subsumed under the same categories as the objections to mathematical economics in the previous sections, and merely reinforce my argument that Keynes believed the economic system cannot be given a general mathematical representation.

1. The variables tested by econometrics are not in fact independent:

Must we push our preliminary analysis to the point at which we are confident that the different factors are substantially independent of one another? This is not discussed. Yet I think it is important. For, if we are using factors which are not wholly independent, we lay ourselves open to the extraordinarily difficult and deceptive complications of ‘spurious’ correlation. (XIV:309)

Likewise, there is no valid distinction between the dependent and independent variables (XIV:310). Nothing acts purely on anything else; everything is in a swirl of causes.

2. The relationships being tested are analogical only because we are not dealing with the real atomic units. Therefore the (p.156) background to the variables is not constant but is complex and changing:

There is first of all the central question of methodology—the logic of applying the method of multiple correlation to unanalysed economic material, which we know to be non‐homogeneous through time. If we were dealing with the action of numerically measurable, independent forces, adequately analysed so that we were dealing with independent atomic factors and between them completely comprehensive, acting with fluctuating relative strength on material constant and homogeneous through time, we might be able to use the method of multiple correlation with some confidence for disentangling the laws of their action. . . .

In fact we know that every one of these conditions is far from being satisfied by the economic material under investigation. (XIV:285–6)

3. Econometrics cannot bring all of the explanatory variables into the tests because there would be neither meaning nor limit to the process. It is impossible because economics is the field par excellence where objective chance applies.

If it is necessary that all the significant factors should be measurable, this is very important. For it withdraws from the operation of the method all those economic problems where political, social and psychological factors, including such things as government policy, the progress of invention and the state of expectation, may be significant. In particular, it is inapplicable to the problem of the business cycle. (XIV:309)

From the modern perspective, there is not much point in measuring equations if it is invalid to have the equations in the first place, and so we do not easily think that econometrics can be helpful when mathematical economics is not. We are however implicitly overlooking the limits of probabilistic analysis and according intuition and the commonly accepted facts of behaviour a lower status than science, which Keynes does not. For example, Keynes thought it was valid to measure the income multiplier, but invalid to go beyond a bare statement of the multiplier relationship.

When this is recognized, Keynes's comments do not support those economists who portray his attitude to econometrics as negative; to the contrary, he attributed a more positive role to induction than do the positivists. He did not agree that the role of (p.157) statistical investigation is only to refute a theory, because he held that induction can add ‘weight’ to the probability of theories; and he argued extensively against Hume that the repetition of what seems the same experiment can add weight to a theory, since the facts are never quite the same, and a slightly new test is always being undertaken. Keynes did not say, as has been suggested, that an economic theory is never testable. Hume had opposed induction because the mere multiplication of instances proves nothing—if all eggs are alike we learn no more from eating one hundred eggs than from eating one. Keynes replied that all eggs are never alike—‘Hume should have tried eggs in the town or the country, in January and in June’ (VIII:243). The independent variety in the data meant that probabilistic conclusions can be valid, although of course all induction is still based on some intuition about how alike the eggs really are. The general point concerning induction, to draw rather freely upon the Treatise on Probability, is that econometrics and quantifiable modelling are valid if the principle of indifference is valid and the processes of objective chance are frozen. If econometrics is to be applicable, there must be some reason for thinking that the probability of an event being due to an omitted variable is as great as the likelihood of that event not occurring.

Put broadly, the most important condition is that the environment in all relevant respects, other than the fluctuations in those factors of which we take particular account, should be uniform and homogeneous over a period of time. . . . For the main prima facie objection to the application of the method of multiple correlation to complex economic problems lies in the apparent lack of any adequate degree of uniformity in the environment. (XIV:316)

This is a prima facie objection, not to Hume, but to the Laplacian assumptions that the Treatise on Probability argues are less fundamental (VIII:55). Therefore Keynes proposed a practical modification which would be more likely to yield some uniformity and homogeneity of the background material, to make the principle of indifference more applicable and arrest the distortions of objective chance. He suggested that the periods to be tested should be subdivided in such a way that in each one something like uniform conditions could be assumed. ‘If they are, then we have some ground for projecting our results into the (p.158) future’ (XIV:316). He drew this suggestion from Wilhelm Lexis, whom Keynes credits in the Treatise on probability with being one of the first statisticians to understand the limits of the principle of indifference (VIII:403 ff.).6

The subdivision of econometrics into intuitively uniform periods would make economics more difficult because the number of observations would be smaller in each period, but the point that Keynes is making is of quite another order. The Laplacians were right, Hume notwithstanding, in believing that it was possible to obtain knowledge from induction and observation. Their error was in taking a statistical short‐cut and overlooking the necessary role of intuition. Intuition must organize the theory that we formulate, and intuition must also interpret the results of that theory. Tinbergen was following the methods of natural science because he was systematically displacing intuition, and all that it implied, with natural science and all that it implied. The fact that he was doing so by a statistical method that was invalid in itself was of course worse.

Therefore econometrics can be valid in a rough way, but there was a basic issue in econometrics concerning which Keynes and Harrod disagreed. Keynes did not agree with Harrod that econometrics would be a good outlet for a lot of workers ‘not outstandingly inspired’ who wanted to find systematic work to do. The true econometrician, according to Keynes, would have ‘vigilant observation’—‘passionate perception’ into the facts—and would practise a rare art.

The object of a model is to segregate the semi‐permanent or relatively constant factors from those which are transitory or fluctuating so as to develop a logical way of thinking about the latter, and of understanding the time sequences to which they give rise in particular cases.

Good economists are scarce because the gift for using ‘vigilant observation’ to choose good models, although it does not require a highly specialized intellectual technique, appears to be a very rare one. (XIV:297)

Keynes regarded econometrics as an art form and based on the intuition. In the end, Harrod too was excluding the role of intuition from econometrics, and Keynes was asserting it.

(p.159) Notes to Chapter 8

(1.) Keynes concluded his biography of Edgeworth by depicting him sitting aloft in a heron's nest, ‘so as it were he dwelt always, not too much concerned with the earth’ (X:260).

(2.) Patinkin among others suggests that we should not take Keynes's opposition to mathematical economics at face value, because Keynes attempted to express himself mathematically but failed. ‘It may have been Keynes’ lack of success with such formal model building in the Treatise that led him to a more critical attitude. . . . The General Theory reveals an ambivalent attitude toward the role of mathematical analysis in economics' (Patinkin 1982:226).

(3.) When an economic theory is deterministic, there needs to be a technical explanation of unemployment. The conventional Keynesians were not able to resolve among themselves what this mechanism might be. Hicks said that Keynes had assumed a liquidity trap, or minimum floor for interest rates in the economy. Because the liquidity trap meant that interest rates could not fall sufficiently, investment would be limited, and there would be unemployment.

I agree with Patinkin and Leijonhufvud that Keynes had not ever assumed a liquidity trap at all, and although he did say in passing that such a thing was a possibility, he immediately dismissed it as unlikely. ‘But whilst this limiting case might become practically important in future, I know of no example of it hitherto’ (GT:207). Hicks needed to invent the liquidity trap to translate Keynes into mathematics; it was his way of reconciling the classical method with some of the flavour of Keynes's monetary theory of interest. If there was a liquidity trap, the rate of interest could still be determined exclusively by monetary conditions, but as a special case.

However, it became apparent that the liquidity trap theory, which fell a long way short of Keynes's idea of the economy as a flux with a principle of indeterminancy at its centre, could not explain unemployment after all. Franco Modigliani showed that, whether or not there was a liquidity trap, there would still be full employment if wages were able to adjust to make labour demand equal to labour supply. Modigliani followed Hicks's method, but in an effort to make the IS–LM theory more logical he showed that a liquidity trap is a superfluous assumption. According to his logic, unemployment was due to rigidities in the labour market, such as trade unions, which stopped wages from falling adequately when labour demand fell. Given the method to be adopted, it was a more logical theory, but it was a further step away from Keynes, who, as Patinkin and (p.160) Leijonhufvud again show, did not think that high wages were responsible for unemployment. Nevertheless, and despite the clear evidence of scholarship, Keynesian economics came to be known among economists as the economics of the liquidity trap and rigid money wages, and the economics of uncertainty and change were forgotten.

(4.) Although Keynes referred to ‘equilibrium’ at a number of points in the General Theory, I support the usual view that he did not understand the word in the conventional sense. He sometimes did adopt the usual meaning of the word when discussing classical economics (as in GT:80), but evidently only for the purposes of the argument, because in XXIX for example he said, ‘In a world of uncertainty . . . a position of final equilibrium, such as one deals with in static economics, does not properly exist’. Where the word ‘equilibrium’ appears with respect to his own theories, it sometimes means a position where demand is equal to supply but price could nevertheless change. For example, there can be an ‘unstable equilibrium’ (GT:269) when demand equals supply but prices race towards either infinity or zero. More generally, Keynes often meant by ‘equilibrium’ an intellectual organizing principle, determinateness without implying stability. This is how D. H. Robertson understood Keynes (XIV:98); for examples see GT:248, XIV:104.

(5.) ‘Ricardo probably deserves chief credit for launching the [quantity] theory’ (Fisher 1911:25).

(6.) As opposed to the economists, Weatherford interprets Keynes not as a critic of statistical method, but as someone who has attempted to rationalize that method, by defining conditions under which the principle of indifference would apply. ‘Keynes has at best achieved a technical revision which eliminates (at least some) logical contradictions and other absurdities from the implications of the Principle [of indifference]—he has not fundamentally altered its theoretical position as the source of all numerical probabilities nor justified its basic assumptions’ (Weatherford 1982:83).

Notes:

(*) This section may be omitted on a first reading. Its purpose is to show that Keynes's reservations towards econometrics, which have excited scholarly attention, are an aspect of his general economic method, as I have described it in the previous sections of this chapter.

(1.) Keynes concluded his biography of Edgeworth by depicting him sitting aloft in a heron's nest, ‘so as it were he dwelt always, not too much concerned with the earth’ (X:260).

(2.) Patinkin among others suggests that we should not take Keynes's opposition to mathematical economics at face value, because Keynes attempted to express himself mathematically but failed. ‘It may have been Keynes’ lack of success with such formal model building in the Treatise that led him to a more critical attitude. . . . The General Theory reveals an ambivalent attitude toward the role of mathematical analysis in economics' (Patinkin 1982:226).

(3.) When an economic theory is deterministic, there needs to be a technical explanation of unemployment. The conventional Keynesians were not able to resolve among themselves what this mechanism might be. Hicks said that Keynes had assumed a liquidity trap, or minimum floor for interest rates in the economy. Because the liquidity trap meant that interest rates could not fall sufficiently, investment would be limited, and there would be unemployment.

I agree with Patinkin and Leijonhufvud that Keynes had not ever assumed a liquidity trap at all, and although he did say in passing that such a thing was a possibility, he immediately dismissed it as unlikely. ‘But whilst this limiting case might become practically important in future, I know of no example of it hitherto’ (GT:207). Hicks needed to invent the liquidity trap to translate Keynes into mathematics; it was his way of reconciling the classical method with some of the flavour of Keynes's monetary theory of interest. If there was a liquidity trap, the rate of interest could still be determined exclusively by monetary conditions, but as a special case.

However, it became apparent that the liquidity trap theory, which fell a long way short of Keynes's idea of the economy as a flux with a principle of indeterminancy at its centre, could not explain unemployment after all. Franco Modigliani showed that, whether or not there was a liquidity trap, there would still be full employment if wages were able to adjust to make labour demand equal to labour supply. Modigliani followed Hicks's method, but in an effort to make the IS–LM theory more logical he showed that a liquidity trap is a superfluous assumption. According to his logic, unemployment was due to rigidities in the labour market, such as trade unions, which stopped wages from falling adequately when labour demand fell. Given the method to be adopted, it was a more logical theory, but it was a further step away from Keynes, who, as Patinkin and (p.160) Leijonhufvud again show, did not think that high wages were responsible for unemployment. Nevertheless, and despite the clear evidence of scholarship, Keynesian economics came to be known among economists as the economics of the liquidity trap and rigid money wages, and the economics of uncertainty and change were forgotten.

(4.) Although Keynes referred to ‘equilibrium’ at a number of points in the General Theory, I support the usual view that he did not understand the word in the conventional sense. He sometimes did adopt the usual meaning of the word when discussing classical economics (as in GT:80), but evidently only for the purposes of the argument, because in XXIX for example he said, ‘In a world of uncertainty . . . a position of final equilibrium, such as one deals with in static economics, does not properly exist’. Where the word ‘equilibrium’ appears with respect to his own theories, it sometimes means a position where demand is equal to supply but price could nevertheless change. For example, there can be an ‘unstable equilibrium’ (GT:269) when demand equals supply but prices race towards either infinity or zero. More generally, Keynes often meant by ‘equilibrium’ an intellectual organizing principle, determinateness without implying stability. This is how D. H. Robertson understood Keynes (XIV:98); for examples see GT:248, XIV:104.

(5.) ‘Ricardo probably deserves chief credit for launching the [quantity] theory’ (Fisher 1911:25).

(6.) As opposed to the economists, Weatherford interprets Keynes not as a critic of statistical method, but as someone who has attempted to rationalize that method, by defining conditions under which the principle of indifference would apply. ‘Keynes has at best achieved a technical revision which eliminates (at least some) logical contradictions and other absurdities from the implications of the Principle [of indifference]—he has not fundamentally altered its theoretical position as the source of all numerical probabilities nor justified its basic assumptions’ (Weatherford 1982:83).