4

‘Animal spirits’
Rethinking economic theory

LOOKING BACKWARD, we know Keynes as the man who wrote the General Theory. But Keynes himself, living his life forward, naturally thought, while in the throes of composition, that the Treatise would be his magnum opus. During the first nine months of 1930, he was occupied in an overlapping way with the sittings of the Macmillan committee, then with the meetings of the EAC’s committee of economists, and all along with the final revisions of the Treatise for publication in October. As we have seen, there was much interplay between these different discussions of economic policy – and of theory too.

Keynes wanted his analysis in the Treatise to inform current debates about practical policy. ‘We quite realise we are getting the fruits of your research presented to us in a form in which we can understand it,’ Lord Macmillan had said in February. Keynes responded that his work had ‘been read now by some of the principal economists of Cambridge, who did not all start sympathetic to it, but they are now satisfied, I think, that it is accurate’.¹ He was referring in particular to his colleague Dennis Robertson (ultimately Pigou’s successor in the chair of Political Economy at Cambridge); and also to Ralph Hawtrey, who saw the proofs of the Treatise before he gave his own evidence to the committee and who continued to write lengthy private criticisms, draft after draft, nailing down Keynes’s inconsistencies.

The Treatise was due to be published before the Macmillan committee was due to report. ‘Therefore it will be exposed to the hostile criticism of the world for an appreciable time,’ Keynes assured Macmillan.² Meanwhile Keynes himself had no doubts about the soundness of its analysis: that if total investment was less than total savings then losses and unemployment must be the result. ‘This, of course, is a difficult theoretical proposition,’ he told Governor Norman in May 1930. ‘It is very important that a competent decision should be reached whether it is true or false. I can only say that I am ready to have my head chopped off if it is false!’³

Some of these confident expectations, however, did not survive for long. In September, with publication now imminent, Maynard told Florence Keynes that his seven years of work on the Treatise was now done, but with some dissatisfaction now intermingled with his relief: ‘Artistically it is a failure – I have changed my mind too much during the course of it for it to be a proper unity.’⁴

Try as he might, he could not kick this terrible habit of his. The great economist Friedrich von Hayek was currently a colleague of Gregory and Robbins at the LSE, the citadel of economic orthodoxy. Hayek devoted a two-part article for the LSE house journal Economica to a critique of Keynes’s big book, only to find that, after the second instalment of this long review appeared in February 1932, ‘he told me that he had in the meantime changed his mind and no longer believed what he had said in that work’.⁵ Likewise, while Hawtrey continued dutifully poring over the Treatise, its author, far from digging in to defend it, was speaking by May 1932 of ‘working it out all over again’.⁶

Whatever had happened to produce such an extraordinary retreat? Or was it perhaps an advance, this time in a more promising direction?

The criticisms of Robertson and Hawtrey had eventually sunk in. They were listened to precisely because each was so close to Keynes: personally as well as intellectually. They were seen as accomplices in his daredevil escape from a prison that Keynes subsequently labelled ‘classical economics’. Since he meant by this ‘a compact and coherent pre-war theory of economics to which most senior economists still subscribe’, most of the profession were inmates. Orthodox economics assumed that the system reached its own equilibrium through the effect of interest rates in reconciling the level of investment to the amount of saving available – through flexible prices, of course. ‘Before the war we were all classical economists,’ Keynes asserted. ‘I taught it myself to Robertson, undoubting and unrebuked.’⁷

Robertson had for years been Keynes’s closest collaborator. Like a couple in the terminal stages of a marriage, they spent years and years arguing about money (in their case, the theory of money and credit and banking and prices). The apparatus of the Treatise thus owed a lot to Robertson, since he too believed that there was no reason to assume that saving and investment must harmonise. Robertson too thought that there might well be a mismatch, with investment deficient, savings excessive. And Robertson too thought that the Treasury View was therefore nonsense since public works could turn unused savings into valuable capital assets.

Robertson was no monetarist. He did not think (as Hawtrey did) that it was simply the expansion of credit, or monetary policy, that would prove effective. Robertson did not accept that cheap money alone could do the trick in the midst of a depression. When he had appeared before the Macmillan committee, he had been pressed by Keynes on this point, arguing that surely entrepreneurs would borrow more when interest rates fell low enough. ‘But I still think there is a difficulty of the lenders coming forward,’ Robertson responded – pointing to a paralysing lack of confidence when even an interest rate of 1 per cent might fail to stimulate investment.⁸

We can well recognise such a situation today, with the dramatic drop in interest rates in 2008–9. We are all aware of how, none the less, investors can remain risk-averse, while lenders become anxious not to tie up their money. A modern economist will probably want to shout ‘liquidity preference’ at this point. For, since the General Theory, the idea has become familiar that, when a slump has undermined confidence, low interest rates may not in themselves be enough to revive investment. People may simply prefer the liquidity of holding their wealth in money rather than accepting what they see as the miserably low – but to them appallingly risky – returns on the available investments. It is an insight that is as relevant today as eighty years ago.

It was Robertson, however, not Keynes who glimpsed this possibility in 1930, as they later reminded each other. The paradox is that, not only did Robertson abstain from writing his own revolutionary theoretical work: he could not, in the end, even accept the message of his friend’s General Theory in 1936. The breach between them was really psychological as much as economic. Keynes told Robertson at this point that ‘in truth, you are the only writer where much of it is to be found in embryo and to whom acknowledgements are due’, and he perceptively added: ‘So unlike me! I, perhaps, am too ready to take pleasure in feeling that my mind is changed; you too ready to take pain.’⁹

Hence it was Keynes who picked up the ball and ran with it. In the Treatise, to be sure, he had written about the ‘bearishness’ of the public. But in subsequently debating this with Robertson, Keynes went further. He needed to dispel the widely held impression that, when he talked about saving exceeding investment, he meant that ‘hoarding’ produced tangible ‘idle savings’, sitting in bank accounts or elsewhere. This was not Keynes’s view (though Hayek thought it was). So Keynes sought to clarify his meaning by explaining that ‘hoarding’ was not an actual process but a psychological motive. And instead of talking about ‘bearishness’ to explain the public’s reluctance to make investments, he started talking about the ‘propensity to hoard’ instead.¹⁰ Under fire, he was refashioning his conceptual tools. As early as the end of 1931, Keynes told a graduate student in Hayek’s department at the LSE (Nicholas Kaldor, later himself a professor at Cambridge): ‘I must be more lucid next time’, and explained that he was now ‘endeavouring to express the whole thing over again more clearly and from a different angle’.¹¹

A lot of trouble stems from the way that the Treatise defines saving and investment. Rhetorically, it suits Keynes nicely to maintain that they need not be equal. It is the implicit assumption that they are bound to be equal, achieving equilibrium automatically and painlessly, that he wants to expose. In this respect, the vision is the same in the Treatise or the General Theory, for in Keynes’s view, it is investment that is active, saving passive in this relationship. He wants to show how the dynamics of the economy depend on the extent to which enterprise is or is not rewarded (in building cities, draining fens, etc.). If expectations are cheated in outcome, entrepreneurs make losses. Thrift is thus in the dock, not enterprise.

Then the Treatise takes a false step. It makes the dramatic but vulnerable claim that the losses are equal to the excess of saving over investment. The problem is that the rhetorical effect needs to be squared with mathematical logic. Here the Treatise has to make its ‘fundamental equations’ add up by using a peculiar definition. Income is defined as expected income, including expected profit, but excluding ‘windfall’ gains or losses. If expectations are fulfilled, savings are indeed equal to investment. Both have essentially the same definition: the residual part of income that is not consumed.

How, then, can savings and investment be different? Because, even if the expected profit fails to materialise, the Treatise still counts it as part of ‘income’. And since this notionally inflated ‘income’ minus actual consumption equals savings, it follows that ‘savings’ too must be notionally inflated by exactly the same amount. Income is obviously different from actual receipts because it includes this fairy gold – profits that entrepreneurs initially expected to make but fail to receive in a falling market. So it is indeed the case that business losses (the shortfall between expected ‘income’ and actual receipts) must be the same as the excess of ‘savings’ (the part of ‘income’ that is neither consumed nor profitably invested). But neither this slice of ‘income’ nor these ‘excess savings’ actually materialise. Both are really measures of unfulfilled expectations.

Robertson and Hawtrey, each in his own way, tried to get this across to Keynes. Hawtrey summed up by saying that of course the Treatise was correct in claiming that the excess of savings over investment equalled business losses, precisely because ‘the excess savings are the losses made by the entrepreneurs and have no other existence whatever’.¹² So what Keynes had spent so much energy in maintaining was not so much a truth as a truism. Both he and Hawtrey, it should be remembered, had studied mathematics as undergraduates. As Keynes told his students in one of his lectures at Cambridge: ‘The whole of mathematics is a truism and truisms help to clear one’s mind.’¹³

His own mind became considerably clearer by 1932. We can now see that he was moving beyond a monetary understanding of the way the economy works, assuming that all adjustments depend essentially on changes in prices. ‘Classical’ economics – really Marshallian orthodoxy – said that an infinitely adjustable price mechanism will deliver equilibrium via interest rates.

Suppose, however, that such adjustments work not through changes in prices, but changes in total output itself. Keynes had previously suggested such possibilities in the real world, when prices are sticky, but in theory he had always insisted that sufficiently low interest rates will do the trick. It was actually Hawtrey who was more alive to the possibility that the level of output itself might respond first. Indeed Hawtrey sketched a model of the economy in which deflation and inflation (or reflation) result in cumulative changes in the level of total income and output; but neither he nor Keynes saw the full significance of where this might lead (though, again, a modern economist might mutter about a multiplier effect).

A lot of these discussions would have been simpler, as Keynes later admitted, if he had known earlier of an idea introduced by the Swedish economist Gunnar Myrdal and developed by his compatriot Bertil Ohlin. This was the simple but compelling distinction between ex ante (how things are intended) and ex post (how they actually work out). Of course Keynes could have used this decades earlier in his academic work, to explain that his theory of probability rested on ex ante judgements, made in advance, not on ex post outcomes.

We are surely entitled to anticipate his own awareness of this distinction in the interests of clarity. For what Keynes was more clumsily saying in March 1932, when he conceded the argument over the definition of ‘savings’ and ‘income’ to Robertson, Hawtrey and Hayek, turned on exactly this point. Yes, he admitted, in adopting the conventional definition, ‘the sense to which I have now bowed the knee’: the Treatise had talked about both saving and income ex ante, from the viewpoint of expectations.

The key point, to which Keynes adhered, was that the expectations had not been realised. So he now accepted the ex post definitions, which could be called simple-minded or common sense. Savings were thus actual realised savings and always equal in the end to the prior investment. But Keynes warned that ‘the implications of this use of language are decidedly different from what “common-sense” supposes’ – because, ex post, saving ‘always and necessarily accommodates itself’ to investment, passively. Thrift does not determine enterprise. Instead, enterprise actively determines thrift, through a change in the total level of income. In fact, saving is ‘no longer the dog, which common sense believes it to be, but the tail’.¹⁴

Here was the seed of the Keynesian revolution in economic theory. It was not yet a flower. But the ground had been cleared for the seed to take root. We can now see why Keynes later said that ‘in recent times, I have never regarded Hawtrey, Robertson or Ohlin, for example, as classical economists’.¹⁵ He meant it as a tribute when he wrote for publication in 1937: ‘I regard Mr Hawtrey as my grandparent and Mr Robertson as my parent in the paths of errancy, and I have been greatly influenced by them.’¹⁶ This is consistent with what Keynes told Robertson in a private letter. ‘The last thing I should accuse you of is being classical or orthodox,’ he wrote. ‘But you won’t slough your skins, like a good snake!’¹⁷

There was another reason for Keynes to slough off an old skin. This was the influence of his younger colleagues in Cambridge. At the beginning of 1931, their small discussion circle, or ‘Circus’, began meeting regularly to discuss the recently published Treatise. The core members were Austin Robinson, soon to be assistant editor of the Economic Journal and much later to be promoted to a chair in the Economics Faculty; his left-wing wife Joan Robinson, who had the singular distinction of ultimately succeeding her husband in the same chair; Piero Sraffa, subsequently famous for his edition of the works of the classical economist David Ricardo; Richard Kahn, a twenty-five-year-old Fellow of King’s College for whom Keynes had secured temporary employment in 1930 as joint secretary of the EAC’s committee of economists; and James Meade, the future Nobel laureate, then a novice economist of twenty-three, in Cambridge just for one academic year to learn his trade before taking up a teaching Fellowship in Oxford.

The Circus thrashed out their problems between themselves, usually reporting back to the busy, famous, peripatetic Keynes through Kahn, whom he knew best. Schumpeter was aware of the Circus, impressed that he himself had nothing like it, and later suggested that its members, especially Kahn, had been given insufficient credit. This was a Harvard view, already more professionally sharp than 1930s Cambridge. The fact is that the Circus and Keynes alike had a deplorably weak sense of individual intellectual property by the standards of the modern research culture. They all borrowed from each other without generating feelings that anyone had been robbed, still less talking of plagiarism. As Austin Robinson later put it, Keynes ‘had a wonderful memory for arguments, but no memory for their authors’.¹⁸

Meade’s position was at once marginal and pivotal. He was the new boy; he only stayed in Cambridge for a year; he had to return to his post in Oxford before October 1931; yet he firmly believed that he then took with him the essentials of the theory of effective demand. This claim, moreover, withstands scrutiny because of what he contributed to one of the most famous conceptual breakthroughs in the history of economics.

This was the ‘multiplier’ concept, as Keynes called it after he had later appropriated it. It had its origin in a short paper that Kahn had prepared for the EAC’s committee of economists in 1930. This was an exercise addressed to two awkward questions, still obviously relevant today, about any public works proposal. First, how much new employment can it actually be expected to generate? Secondly, how to pay for it?

Kahn focused on the first question. It had been tackled in 1929 in ‘Can Lloyd George Do It?’, with the argument that, as well as the workers actually employed on new roads, indirect employment would provide a further boost. ‘The fact that many workpeople who are now unemployed would be receiving wages instead of unemployment pay would mean an increase in effective purchasing power which would give a general stimulus to trade.’¹⁹ This sounds like common sense and it is not wrong. But it might prompt a vague notion that this sort of cumulative prosperity can be infinite in its repercussions. That sounds – and is – too good to be true.

Kahn’s achievement was to show that the effects are finite and that they can be specified. The process itself might have an infinite number of repercussions, but these will always sum to a finite figure, which can be calculated. The key is what proportion of extra new income in the hands of a newly employed worker will be passed on in the form of spending – the more the better for multiplying the stimulative effect.

If, say, half is spent, then this expenditure in turn is subject to the same arithmetic – half of half (or a quarter) will be spent, and then half of that (an eighth), and so on. The final arithmetic here, dividing by 2 each time along the infinitely long chain of spending, will in aggregate multiply the original investment by 2, meaning that, in this case, it will have twice the leverage of the initial stimulus package.

Kahn’s was a great achievement. Every subsequent attempt to specify the effects of a stimulus package in a determinate way rests on this model. If half is spent, the multiplier is 2, which is simple and elegant for demonstration purposes. It is also, in fact, the sort of thing that early enthusiasts for the multiplier excitedly told each other, including Keynes, who based his own later calculations on a multiplier of 2, and expected the multiplier to be, if anything, higher than that in the United States in the 1930s. Actually, modern estimates of the multiplier are lower, generally in the range 1.25 to 1.75. The principle of the multiplier, however, is the real discovery. For this analytical tool is as indispensable for those who use low estimates to argue against stimulus as it is for those who argue in favour.

In 1930, when first offered to the EAC’s committee of economists in a primitive form, however, the tool was spurned. It needed to be sharpened and refashioned before its utility was acknowledged. This is where the reflective, analytical mind of Meade came into play. As a result of the Circus discussions, Kahn incorporated a new idea into his famous ‘multiplier’ article, as published in mid-1931 in the Economic Journal, under Keynes’s editorship. This further concept was called ‘Mr Meade’s Relation’ – a usage which had been common among the members of the Circus (though when Keynes attended one of their meetings and heard of it for the first time, he apparently looked around the room for this unfamiliar member of Meade’s family).

Mr Meade’s Relation tells us to add up, at each stage, the amount that is not passed on in increased consumption. This is essentially a generalisation of the multiplier relationship, so that it follows each alternative process, either of spending or non-spending (saving). Kahn was focused on the extra employment generated by increments of spending; Meade was also looking at the extent to which either prices or output will rise in response to an initial investment – through increments of saving.

What are these ‘leakages’, as Keynes was later to call them? Personal savings seem to us the most obvious (though overlooked initially by Meade). ‘Savings on the dole’ had often been mentioned, meaning the relief to government and any other finances that had previously supported unemployed workers now back at work and supporting themselves. To this Meade adds increased import costs, since this is a leakage out of the pipeline of repercussive spending at home. And he adds any increase in unspent profits – real profits this time, constituting real savings put aside out of the new prosperity. These are all parts of the original investment that are not passed on in spending – and must therefore, by subtraction, be classed as saving. (This only seems odd if we forget that saving is passive, as distinct from investment, which is actually a form of spending on capital goods.) So income minus consumption equals saving, in the common-sense terms to which Keynes ‘bent the knee’ after abandoning Treatise definitions.

Meade’s conclusion is another mind-clearing truism. For these unspent fractions of non-spending, in their parallel long series, inexorably sum to unity. Mr Meade’s Relation thus indicates a source of saving that must exactly equal the original investment.

This highly significant conclusion can be put in two ways. One is to point out, as he and Kahn did at the time, that this gave the answer to the question: how to pay for public works? The answer is still valid: that, in the end, they pay for themselves out of the increased economic activity that they stimulate. The savings that are eventually generated are passive, and therefore permissive in providing sources that allow (but do not require) active investment to be accomplished.

The only decision involved is to make the investment in the first place. Public works indeed normally require borrowing the funds that will initially finance them – pump-priming is one analogy (though not a term used by Kahn or Meade or Keynes). The borrowing, admittedly, creates a debt to be settled in the future. But the allegation that this creates a burden of debt upon the next generation ignores the fact that they will simultaneously be made more prosperous by exactly the process that the borrowing initially financed.

Another way of expressing Meade’s point is even more fundamental. For if this is true of an increment of investment for public works, it is surely true of all investment. What Mr Meade’s Relation demonstrates is that saving will always be brought into equality with investment via a change in the level of total output or income. This is, in fact, the general process that brings an equilibrium between saving and investment. Little wonder that, forty years later, Meade wrote (in an echo of words that Keynes had used himself): ‘Keynes’s intellectual revolution was to shift economists from thinking normally in terms of a model of reality in which a dog called savings wagged his tail labelled investment to thinking in term of a model in which a dog called investment wagged his tail labelled savings.’²⁰

It is not altogether clear how quickly the new thinking of the Circus was appreciated by Keynes at the time. Joan Robinson inimitably thought that he was very slow to catch on and that ‘there were moments when we had some trouble in getting Maynard to see what the point of his revolution really was’.²¹ In the summer of 1931, admittedly, Keynes was still giving public lectures in Chicago denying that saving and investment were necessarily equal, so he plainly had not yet jettisoned the Treatise definitions altogether. Yet some of the book’s examples, almost despite its formal analysis, were leading the Circus into new paths.

Take the banana parable. This had first been produced by Keynes for the Macmillan committee. ‘Let us suppose a community which owns nothing but banana plantations which they labour to cultivate,’ he began. They produce bananas, they consume bananas, and nothing else. What they do not spend on bananas, they save; and the investment in the production of bananas exactly equals this saving. ‘Into this Eden,’ Keynes continued, ‘there enters a thrift campaign urging the members of the public to abate their improvident practice of spending nearly all their current incomes on buying bananas for food.’ For whatever reason, increased saving is not matched by increased investment – why should it be? Bananas are then produced in the same quantities as before; they will not keep; they have to be sold; but thrift has reduced the amount the public will pay for them; so the bananas are sold at lower prices.

‘The only effect has been to transfer the wealth of the entrepreneurs out of their pockets into the pockets of the public,’ claimed Keynes, always the champion of enterprise rather than thrift. ‘The only thing that increases the actual wealth of the world is actual investment,’ he continued, showing that the entrepreneurs, as innocent victims, cannot save themselves by making their workers victims too, via wage reductions and unemployment. Such a response, however rational by individual employers, has the effect of further reducing the purchasing power of the community below the level at which production is profitable. Admittedly, a cartel might help them avert a situation where everyone starves to death. Otherwise, he said, only two options remain: ‘the thrift campaign falls off, or peters out. Or investment is increased.’²²

Actually, the banana parable proves too much. Although it could be used to illustrate a downward multiplier effect, through decreased increments of spending, it still assumes that, if income is reduced, spending is also reduced by that whole amount. It overlooks the fact that savings will surely also be reduced under such circumstances – or, conversely, that savings will also be increased to some extent under circumstances where income is increasing. Others were ahead of Keynes in seeing the relevance of this point.

Keynes, however, did come to recognise the significance of the fact that changes in income will probably not produce exactly equal changes in consumption. He called it ‘the psychological law that, when income increases, the gap between income and consumption will increase’. For if net personal savings increase in this way with increased prosperity, these savings will surely need to be matched by some commensurate increase in investment in order to sustain demand as a whole. A reorientation in Keynes’s thinking during 1932 was prompted by his dawning realisation that orthodox theory had long ignored ‘the theory of the demand and supply for output as a whole’.²³

These were important stepping stones towards his own attempt to produce such a theory. As Keynes later put it: ‘The novelty in my treatment of saving and investment consists, not in my maintaining their necessary aggregate equality, but in the proposition that it is, not the rate of interest, but the level of incomes which (in conjunction with certain other factors) ensures this equality.’²⁴ It follows that the level of output is itself the equilibrant, moving the economy between different positions, of which full employment is only one.

Then what role is left for interest rate? Not, any more, the key role that it had played in the Treatise and indeed all orthodox theory: that of the great equilibrant. There was now a gaping hole that Keynes needed to fill. He already had in his head his new theory of output as a whole when (as he once explained the process to Harrod) ‘appreciably later, came the notion of interest as being the meaning of liquidity preference, which became quite clear in my mind the moment I thought of it’.²⁵ All those old discussions of ‘bearishness’, and ‘hoarding’ and ‘the propensity to hoard’, suddenly made a new kind of sense. He saw that everything turns on the psychology of the public about holding money itself, and thus on their reluctance to invest their savings when confidence collapses.

We can pinpoint when Keynes reached this crucial stage in his thinking – which was significantly earlier than used to be supposed. In the autumn of 1932, once every week for one university term, Keynes was giving his first lectures in Cambridge for two years. They were packed to the doors. Colleagues attended as well as undergraduates; but the graduate students kept the best notes. ‘It was as if we were listening to Charles Darwin or Isaac Newton,’ one American student recalled, no doubt mindful that these too were local heroes who had instigated far-reaching intellectual revolutions.²⁶ Keynes paid his audience the compliment of talking about the most interesting subject he could possibly imagine: his own new book, now in preparation.

The course was called ‘The Monetary Theory of Production’. The notes of the graduate students for his fourth lecture, on 31 October 1932, show that Keynes led up to the conclusion that ‘in itself the rate of interest is an expression of liquidity preference’.²⁷ So output as a whole was not in theory equilibrated by the interest rate; instead investment was the motor of the economy and the level of output the equilibrant. And at the end of term Keynes found a name for this far-reaching new analysis, after his friend Piero Sraffa had shown him the rediscovered correspondence to Ricardo from the combative Thomas Malthus, more than a century previously. Keynes decided to salute Malthus as yet another brave Cambridge pioneer by purloining his term ‘effective demand’ to describe his own theory of output as a whole.

Keynes had seized on the essentials of the theory of effective demand by the end of 1932. Thus girded, he went into battle; so when he almost immediately re-entered the ongoing polemics over economic policy, he did so backed by his new arguments. Until this point his practical proposals, although they might ring a bell with some Americans, were addressed to peculiarly British conditions, not readily applicable across the Atlantic. The United States was hardly the victim of the Gold Standard; the ‘modus operandi of Bank rate’ seemed foreign; the Treasury View was a construct of another people’s Treasury. In 1933, by contrast, Keynes was ready to speak directly to the problems of Britain and the United States alike, with a new message which had a universal significance.

The author of the Treatise had been calling for stimulus measures for years. But, in theory, the Treatise had argued that flexible prices ought to deliver full employment. Keynes acknowledged time and again, to the Macmillan committee and elsewhere, the general principle that interest rates (or monetary policy) would do the trick. He said that cheap money could be relied upon to stimulate investment and thus economic recovery. The practical difficulty for Britain was simply that, under the Gold Standard, cheap money was not an option. Alternatively, it could be said that the British problem was sticky wages, which refused to respond to the credit squeeze of dear money by falling to an internationally competitive level. Unemployment was in this sense voluntary – collectively, workers chose to accept it rather than accept wage cuts. One way or another, these classical economic adjustments of the price level had failed to materialise; and Keynes had therefore opted for a series of second-best palliatives like tariffs and, above all, public works.

Keynes’s maxim about the long run had particular pertinence under these conditions. It continued to provide a relevant test, but the conditions themselves changed. In 1931 Britain had gone off gold; the pound was no longer overvalued against the dollar; the special case of the Treatise was no longer operative.

The theoretical premise for his support for tariffs thus disappeared. Indeed Keynes at once abandoned this particular argument (though he still showed a hankering for home-grown solutions, as we have seen). Moreover, his special case for public works likewise disappeared, since the Bank of England no longer needed to maintain dear money. In fact, bank rate (or base rate), which had hovered around 5 per cent in the late 1920s, fell by June 1932 to 2 per cent and was to remain at that level until the eve of the Second World War. This was as low as the Bank of England had ever set its rate since its foundation in the late seventeenth century. Only in 2009 have we seen an even lower rate.

But cheap money did not do the trick – at least, not in Keynes’s eyes. British unemployment remained historically high. Keynes consistently took the yardstick of ‘normal’ unemployment as about 5 per cent. The official figures in January 1933 showed a peak of 23 per cent.

If the interest rate could no longer be held responsible, perhaps the root of the problem, after all, might be uncompetitive wages in Britain. This was broadly the conclusion that Hubert Henderson had reached in 1930–1. At that point, when Keynes was accused by his old comrade of evading the remedy of ‘an assault on wages’, he could only respond that ‘I twist and turn about’ in seeking alternatives, because an increase in investment rather than a reduction in wages still seemed more promising.²⁸ There is indeed a good argument for saying that high costs, including labour costs, made British prices uncompetitive in world terms, even if Keynes himself – like Pigou – was reluctant to see wage reductions as a practical remedy. But this too changed once the whole world spiralled into Depression, with unemployment rising in practically every country. How could all countries be simultaneously uncompetitive with each other? How could wage cuts, in all countries simultaneously, help the overall situation?

Orthodox economists, like Pigou, could well persist in the view that ‘gadgets’ and ‘devices’ were still in practice necessary. What now distinguished Keynes was that, for the first time, he approached the problem with a theory that challenged orthodoxy. Monetary policy was no longer Keynes’s theoretical remedy. Yes, it could do the trick of checking inflationary exuberance, rather like pulling on a piece of string. But this was its only trick. Faced with the actual depressed, deflationary problem of the 1930s, it was like pushing on a piece of string. Rock-bottom interest rates failed to deliver the necessary stimulus.

Keynes now moved with such confidence because of intellectual conviction. The members of the Circus shared his thrilling sense of discovering inherent truths rather than inventing tactical arguments. The new book that he was writing was plainly not going to revise the Treatise but supersede it altogether. As he later told the story to Harrod, although there was ‘an immense lot of muddling and many drafts’ necessary to make it fit for publication, it was the change in his thinking during 1932 that produced ‘the moments of transition which were for me personally moments of illumination’ – glimpsing ‘particles of light in escaping from a tunnel’.²⁹

The theory of effective demand now explained that an equilibrium had indeed been achieved between saving and investment. But it was an equilibrium at less than full employment. Individual action was powerless to achieve results that could be secured only through collective action. Hence the new relevance of a stimulus package, whether at home or abroad, in one country or – preferably – the whole world. Although the term equilibrium had a comforting ring, suggesting that the economy was in balance, the real implication was sinister. The economy was at rest. It was not self-righting.

Here was Keynes’s revolution: one prompted by his engagement with real-world economic policy debates but transcending them with an analysis that changed the paradigm. The analysis, moreover, was already operational three years before the General Theory was published. His subsequent toils in drafting the book were not wasted, not unduly slow, not particularly easy, not unnecessary in providing the tight formal demonstrations that would convince the ‘fellow economists’ to whom the book was addressed. But his work here was more than that of embellishing a few technical or mathematical propositions with suitable rhetoric, applied afterwards for effect – if anything, the other way round. ‘The precise use of language comes at a late stage in the development of one’s thoughts,’ said Keynes in one of his university lectures. ‘You can think accurately and effectively long before you can so to speak photograph your thought.’³⁰

From the beginning of 1933, after nearly a year and a half of relative public silence, Keynes energetically began to publicise his new thinking. There was a new fiscal emphasis, consistent with his own shift away from monetary policy. In Britain, the National Government’s attempts to balance the budget at the bottom of a slump were his immediate target. In a radio broadcast, he insisted that ‘you will never balance the Budget through measures which reduce the national income’. Cuts in government spending, especially unemployment benefits, were thus self-defeating. ‘Look after the unemployment, and the Budget will look after itself’ was the new watchword. Indeed he went so far as to say that ‘the Chancellor of the Exchequer would be long-sighted if he were to take rather an optimistic view, and give us perhaps in his next Budget rather more relief than is strictly justified by the facts actually in sight’.³¹

For those who automatically associate the name of Keynes with budget deficits, these remarks may seem curiously tepid. None the less they are virtually the only endorsement he ever gave to running a fiscal deficit on current account. What he actually wanted was still loan-financed capital projects, for the obvious reason that this was a direct stimulus to investment. This was naturally the thrust of his proposals in his important and well-publicised pamphlet ‘The Means to Prosperity’, the American edition of which, in April 1933, incorporated his first published exposition of the multiplier.

Keynes was now, of course, in a position to give more robust answers to two long-standing objections. He could specify the cumulative effect of public works upon employment; and he could explain where the money was to be found. ‘Just as an initial impulse of modest dimensions has been capable of producing such devastating repercussions, so also a moderate impulse in the opposite direction will effect a surprising recovery,’ he wrote in the American edition of ‘Means to Prosperity’. He acknowledged the unfamiliarity of the new ideas that he was seeking to popularise, but insisted that the old theories were irrelevant: ‘Many people are trying to solve the problem of unemployment with a theory which is based on the assumption that there is no unemployment.’³²

Persuasion seemed more crucial than ever. Keynes’s own task in persuading inside opinion was a prelude to the persuasion of outside opinion. The psychology of the public fed into market confidence, with effects that were often self-fulfilling. Roosevelt’s inaugural address in March 1933 was famous for the claim that there was nothing to fear but fear itself. Perhaps too much time has been spent in scrutinising the possible influence of Keynes on Roosevelt; perhaps instead the influence of Roosevelt on Keynes needs more recognition. At any rate, in the following month, Keynes furnished a congenially similar message in a British context, as usual criticising Neville Chamberlain as Chancellor of the Exchequer. ‘Unfortunately the more pessimistic the Chancellor’s policy, the more likely it is that pessimistic anticipations will be realised and vice versa,’ said Keynes. ‘Whatever the Chancellor dreams, will come true!’³³ The new President’s dreams were more like his own.