1The 2008 financial crisis has revived debates on the interaction between the financial and the real spheres of the economy, the danger of a deflationary spiral and, consequently, the self-adjusting capacity of the economic system. Comparisons have been made with the 1929 financial crisis, and the way this turned into the Great Depression sheds new light on appropriate economic policy in both the short and the long run in a monetary economy. The title of Skidelsky’s recent book on Keynes, Return of the Master , conveys a general sense that the time of Keynes has come again. Conversely, when one keeps in mind the enormous expansion of the U.S. money supply since the later 1970s (especially of M2), criticism of central bankers after the 2008 episode echoes Friedman’s argument that the Fed, to a great extent, created itself the depression of the 1930s. Recent history might prove Friedman right in his plea for a monetary policy strictly conducted in terms of a money-supply rule, and for the creation of an institutional structure that would prevent the endogenous creation of money. And so, for both the interventionist and the non-interventionist wings, Keynes and Friedman still appear relevant today, even though they are generally represented as being worlds apart on the self-adjusting capacity of the economic system and the proper role of the state.
2As so often happens with great authors, things become more complicated when one goes back to their original writings. Garrison responds to the provocative question “Is Friedman a Keynesian?” by stating that “opposing answers can be defended with no loss of academic respectability” [Garrison, 1992: 131]. It is usual to insist on Keynes’ and Friedman’s very different political perspectives, which are barely reconcilable. Others remember the Friedman of “we are all Keynesians, now” [Friedman, 1968a],  and Modigliani’s response “we are all monetarists” [Modigliani, 1977, 1].  Keynes and Friedman can even be viewed as taking sides together against the “monetarism mark II” [Tobin, 1981: 35] advanced by the New Classicals. As Dostaler puts it: “All this is somewhat confusing, and makes it necessary to return to the founding fathers.” [Dostaler, 1998: 318] He therefore set out to analyze “Friedman’s reading of Keynes in the realms of politics, methodology and economics” [ibid.: 319]. As a synthesis of previous analyses, mixed results are then obtained: following their strong divergences on the political side, their Weltanschauung, “it is at the level of the social vision that the distance is greater” [ibid.: 326]. Conversely, most of their convergence appears to involve their methodological positions, although this should not be overdone: there is a shared distrust of econometrics, and there is above all a common Marshallian background.
3In comparing Keynes and Friedman this paper offers a slightly different approach, since it focuses on the specific issue of full employment. From the outset the article is limited to consideration of Keynes’ and Friedman’s respective conceptions of full employment, and in turn their respective advocacy for the full employment target which is to be achieved. This analytical perspective will shed light upon the theoretical rationale of the employment policies they advocate. The results that we obtain might appear at first sight surprising, especially in regard to discussion of the ‘fine-tuning’ of the economy. We show that Keynes and Friedman share the same distrust for short-run manipulation, and that both are very concerned with the importance of economic stabilisation. Both develop long-term perspectives and call for the use of rules, even if Keynes’ own rules would be better labelled ‘discretionary rules’. This is not to overlook their clearly opposing views on the self-adjusting capacity of a free-market economy, or that their political visions still sharply contrast. There remain divergences which would be open to reconciliation, especially their respective confidence in the ability of the state to pursue clear policy objectives in an effective and efficient manner. It would be misleading to emphasise their convergence on the proper employment policy to be pursued while ignoring the theoretical foundations of their respective positions.
4Consequently, the next step of our inquiry is to investigate the theoretical underpinnings of measures which they advocated in common. The theoretical positions from which they call for stabilisation are in fact quite different. Keynes for example does take account of uncertainty. That is basically the reason why it is an absolute necessity for the state to develop a long-term view, and seek to stabilise the economy through management of the long-term expectations of private actors. In contrast Friedman deals with a risky economy: long-term perspectives and rigid rules would prevent the state from yielding to the temptation of anti-cyclical devices that would prove counter-productive, since the economy is basically stable in the long-run. Here the basic rationale of their respective pleas are in very sharp contrast, turning on their divergent understanding of the knowledge possessed by private individuals on the ‘state of the world’.
5Methodologically the approach adopted here runs the risk of being anachronistic. Keynes was giving political advice from the early 1920s, a long time before the ‘Keynesian revolution’. His final plea dates from 1945. This was about the time that Friedman started his ‘counter-revolution’, his activity continuing until the last edition of Capitalism and Freedom in 2002. Moreover Friedman did not confront the work of Keynes himself, but instead that of his heirs. There is obviously a significant length of time between Keynes’ first advocacy and Friedman’s last plea, and so a great caution is required in dealing with the specific historical circumstances of this or that argument.
6Secondly, as far as Keynes is concerned we shall restrict our investigation to the time of the General Theory and beyond. With Friedman, we will resolve the chronological difficulty outlined above by largely setting to one side his criticism of Keynesianism, concentrating instead on his positive analyses.
7Finally, as a working hypothesis we will attribute consistency to both authors. As with all great authors, we shall assume when reading their texts that Keynes and Friedman meant exactly what they wrote and that, despite occasional amendment, their thinking should be considered to be consistent. In other words, we endorse Hoover’s view that “continued relevance is a fundamental rational for history of economic thought” [Hoover 1995: 677].
8The paper is organised as follows. In the first part of the paper Keynes’ full employment target is compared to Friedman’s natural rate of unemployment. We then investigate their respective employment policies. We conclude by discussing the theoretical underpinnings of the positions they advocate.
1 – Keynes and Friedman on the full employment target
9If our ultimate purpose is to compare Keynes’ and Friedman’s respective views on appropriate employment policy we should first consider the analysis of their respective conceptions of full employment. On the one hand Keynes argues that as soon as ‘full employment’ is achieved public works can do no more. On the other hand, Friedman considers that the economy naturally tends towards its ‘natural rate of unemployment’. To what extent do Keynes and Friedman take structural unemployment into account? Are Keynes’ ‘full employment’ and Friedman’s ‘natural rate of unemployment’ similar, or perhaps identical, constructs?
10Let us start with Keynes’ conception of full employment. As is well-known, the General Theory offers two definitions of full employment, one in terms of aggregate demand (to be found in Chapters 3 and 20 of the General Theory), and the other in terms of labour market functioning (in Chapter 2).  In the first case full employment is defined as the maximum level of aggregate demand, “a situation in which aggregate employment is inelastic in response to an increase in the effective demand for its output” [Keynes, 1936, in CW 7: 26]. In the second, full employment corresponds in Chapter 2 as “the equality of the real wage to the marginal disutility of employment” [Keynes, 1936, in CW 7: 15]. What is crucial is that full employment here does not mean zero unemployment, “both ‘frictional’ and ‘voluntary’ unemployment being consistent with ‘full’ employment thus defined” [Keynes, 1936, in CW 7: 15-16]. Voluntary unemployment is in turn defined as follows:
In addition to ‘frictional’ unemployment, the [second classical] postulate is also compatible with ‘voluntary’ unemployment due to the refusal or inability of a unit of labour, as a result of legislation or social practices or of combination for collective bargaining or of slow response to change or of mere human obstinacy, to accept a reward corresponding to the value of the product attributable to its marginal productivity.
12Keynes is well aware that market imperfections exist in the labour market, what Leijonhufvud [1968: 93] later called a ‘refusal to cooperate’. This distinction between voluntary and involuntary unemployment has not convinced all commentators, including even Richard Kahn, who did so much for the Keynesian Revolution.  However, Keynes continued using this distinction after the General Theory, even if his terminology has changed. During World War 2 Keynes dealt explicitly with “structural unemployment” [Keynes, 1943, in CW 27: 354] in seeking to evaluate each component of unemployment. And so he argues that:
13“Unemployment is due to:
- the hard core of the virtually unemployable (100 000);
- seasonal factors (200 000);
- men moving between jobs (300 000);
- misfits of trade or locality due to lack of mobility (200 000); and
- a “deficiency in the aggregative effective demand for labour.” [Keynes, 1942, in CW 27: 305]
14Conversely, Keynes’ full employment target does not entail a zero rate for increases in prices. In Keynes’ economy prices rise before full employment is achieved because of what he calls “bottle-necks” [Keynes, 1936, in CW 7: 300]. Perfect elasticity of supply is a simplifying assumption. So, there are a number of “positions of semi-inflation” [Keynes, 1936, in CW 7: 301], successive semi-critical points preceding the “final critical point of full employment” [Keynes, 1936, in CW 7: 301].  Indeed, in Chapter 21 of the General Theory Keynes uses a disaggregated analysis and takes account of the heterogeneity of both labour and supply curves in individual industries, so that prices rise before full employment is attained.  In particular, two critical assumptions have to be lifted so that bottlenecks can be taken into account: “(1) That all unemployment resources are homogenous and interchangeable in their efficiency to produce what is wanted, and (2) that the factors of production entering into marginal cost are content with the same money-wage so long as there is a surplus of them unemployed.” [Keynes, 1936, CW 7: 301] Scarcity will appear in some sectors before others, and wage bargaining will be more favourable to the workforce as employment rises. Keynes’ concept of full employment therefore entails neither a zero rate of unemployment nor a zero rate of price-increase. Quite autonomously, Keynes’ economy has an inflationary bias. To put it the other way around, at a zero rate of price increases in the entire economy, there would not be just structural unemployment.
15Let us now turn to Friedman. His famous definition of the natural rate of unemployment runs as follows:
The ‘natural rate of unemployment’ … is the level that would be ground out by the Walrasian system of general equilibrium equations, provided there is embedded in them the actual structural characteristics of the labor and commodity markets, including market imperfections, stochastic variability in demands and supplies, the cost of gathering information about job vacancies and labor availabilities, the cost of mobility, and so on. 
17This definition is taken up in his Nobel lecture, in which Friedman adds that “the ‘natural rate of unemployment’ …is not a numerical constant but depends on ‘real’ as opposed to monetary factors—the effectiveness of the labor market, the extent of competition or monopoly, the barriers or encouragements to working in various occupations, and so on” [Friedman, 1977: 458]. The natural rate of unemployment is determined by ‘real forces’ in the labour market, for example minimum wage rates or the market power of the labour force. That is not to say that the natural rate is given: “On the contrary, many of the market characteristics that determine its level are man-made and policy made” and it “will change from time to time.” [Friedman, 1968b: 9-10] As with Keynes, many commentators have thought Friedman’s concept of the natural rate of unemployment dubious, De Vroey arguing for example that “Friedman’s definition should simply be put aside” [De Vroey, 2001: 130] because of its internal inconsistency. 
18As with Keynes, the specific issue of ‘technological unemployment’ is set aside by Friedman. His natural rate of unemployment also includes both what Keynes called frictional and voluntary unemployment. The critical point to be underlined here is that the natural rate of unemployment does not correspond to any particular rate of inflation, since a vertical Phillips curve in the long-run implies a natural rate of unemployment consistent with any given rate of price-increase. In particular, full employment in Friedman’s sense of the term is perfectly consistent with a zero rate of price-increase.
19Both Keynes and Friedman acknowledge labour market rigidities, even if technological unemployment is a specific concern for neither of them. For both of them full employment is consistent with structural unemployment. However, does Friedman’s natural rate of unemployment entail the same level of employment as Keynes’ concept of full employment? To put it differently, does the zero-inflation rate of unemployment include only voluntary, structural unemployment, in Keynes’ understanding of the term? As shown above, Keynes does take into account heterogeneity of labour and supply, so that a perfectly elastic supply price before full employment is merely a first approximation. Prices begin to rise in step with effective demand before full employment is attained. Hence, true inflation is reached only when an increase in effective demand no longer has repercussions for employment. At Friedman’s natural rate of unemployment, which is consistent with a zero-inflation rate, there would still be involuntarily unemployed resources, according to Keynes’s full employment target. That is, Friedman’s natural rate includes an involuntary component of unemployment in Keynes’ sense of the term. Consequently, Keynes’ target entails a higher level of employment than that of Friedman.
20Even if it is not the basic purpose of this paper, it should be noted that Keynes can in no way be portrayed as the advocate of inflationary policies. As Leeson  clearly demonstrates, Keynes was a reflationist (in favour of those price increases associated with recovery at the end of the slump) rather than an inflationist. During the recovery price rises are interpreted as a signal of future profits. As such, this profit-inflation eases recovery. However, an inflationary policy is sooner or later followed by the reversal of the trade-cycle. Indeed, true inflation corresponds to a very weak state of confidence that in turn renders the marginal efficiency of capital more fragile. “This opposition to the ‘vicious spiral’ of rising prices and wages was a persistent theme of Keynes’s writings.” [Leeson, 1999: 499] According to Keynes, the state should be very cautious in seeking full employment: this requires “planning” [Keynes, 1937, in CW 14: 386] and the cooperation of all private actors all the bottlenecks of the economy are to be dealt with. 
21Their respective views on the target for the level of full employment in the economy no longer seem so divergent once the ‘inflationary bias’ in Keynes’ economy is incorporated in the analysis. A much sharper contrast between our two authors lies in their opposing views on the need for authorities to gather information so that they might ‘know’ their full employment target. Keynes takes some trouble to measure his full employment target. During World War 2, when discussion turned to peacetime employment policy, Keynes actively participated in debates on the minimum attainable post-war rate of unemployment. As a consequence, the need for the authorities to have “absolutely up-to-date information”, the “best and latest information” [Keynes,  1980: 409-410] was a crucial issue for Keynes. In passing, it might be noted that this insistence on the necessity to both collect and disseminate information in respect of the private sphere had always been crucial for Keynes, and even predates the General Theory.  For Friedman the reverse is the case. His prime argument is formulated as early as 1962 in Capitalism and Freedom: “In fiscal policy as in monetary policy, all political considerations aside, we simply do not know enough to be able to use deliberate changes in taxation or expenditures as a sensitive stabilizing mechanism.” [Friedman, 1962: 78, emphasis added]. Here, he is probably playing on Keynes’ 1937 phrasing that “we simply do not know” [Keynes,  1973b: 114] so that he might oppose state intervention. His second argument runs as follows: “One problem is that [the monetary authority] cannot know what the ‘natural’ rate is. Unfortunately we have as yet devised no method to estimate accurately and readily the natural rate of …unemployment.” [Friedman, 1968b: 10] Moreover, “the ‘natural’ rate will change from time to time” [Friedman, 1968b: 10].  Trying to ascertain what it might be is therefore a pointless exercise.
22While they had contrasting conceptions of the way in which full employment should be defined, this does not really present an insuperable difficulty; but they part company one each other on the ability of the state to orchestrate full employment. In contrast to Friedman’s claim that the natural rate of unemployment cannot be pegged, since it is determined by the ‘real’ functioning of the economy, Keynes consider that it is vital for the state to establish and evaluate its full employment target. Keynes is also deeply involved in the debates over the evaluation of structural unemployment. The reason for this involvement is obvious: if the state is supposed to prevent demand-deficiency unemployment, it ought to know precisely what its target is.
2 – Keynes and Friedman on appropriate employment policy
23Having established the views of Keynes and Friedman on full employment targets, we now turn to the analysis of their respective arguments for an appropriate strategy. Far from the stereotypical image one can get of our authors, it can be shown that they share some basic concerns, especially with regard to the stabilisation of the economic system. This is not to deny some issues on which Keynes and Friedman are definitely at odds, such as for example the ability of the state to conduct itself, or the proper instruments of monetary policy efficiently. If the debate were set in terms of rules versus discretion, where would they stand?
2.1 – The relevance of the budgetary weapon
24The issue of private expectations is at the core of Keynes’ plan: first, how to reverse a condition of pessimistic long-term expectations; second, how to maintain these expectations at their full employment level. These two issues must be clearly separated if we are to understand properly fiscal policy as advocated by Keynes. The first, which Keynes dealt with at the time of the ‘Keynesian Revolution’, concerns the cure for unemployment whenever it appears. The second, on which Keynes concentrates after the General Theory, mainly concerns the prevention of unemployment.
25With respect to the cure for unemployment, the key role of public works is to “break the vicious circle” [Keynes, 1933, in CW 21: 158]. In the language of the General Theory, ‘pump-priming’ aims to raise the marginal efficiency of capital. Targeting the pessimistic expectations of both investors and entrepreneurs, the state initiates public works to falsify private expectations and thus boost private investment; and, importantly, not as a direct substitute for the latter.  It must be emphasised that Keynes’ purpose is to boost the market for new capital goods— the production of consumption goods will follow on. Thanks to state intervention, investors are faced with an increase in aggregate demand and they are compelled to revise their expectations of future prices and profits. In other words, public works aim to modify private agents’ long-term expectations.
26The second issue, crucial to our investigation, is the prevention of unemployment. Keynes is somewhat cryptic when he states in the General Theory that “a somewhat comprehensive socialisation of investment will prove the only means of securing an approximation to full employment” [Keynes, 1936, in CW 7: 378; emphasised added]. But Keynes returns to this issue during World War 2 when he takes part in the debate over post-war employment policy. In this debate “what Keynes proposed is the creation of long-run stable government policies to help maintain business confidence in the future” [Bateman, 2003: 81].  Here again Keynes concentrates his attention on investment, rather than on consumption. With reference to Friedman’s permanent income hypothesis, it is worth noting that Keynes does not believe that households would quickly adapt their consumption-demand if taxation in general (and not only national insurance contributions) varied from year to year. He considered that “people have established standards of life” [Keynes, 1943, in CW 27: 319]. Hence “a remission of taxation on which people could only rely for an indefinitely short period might have very limited effects in stimulating their consumption” [Keynes, 1943, CW 27: 319]. Secondly, Keynes strongly opposed deficit budgeting:  “The capital budgeting is a method of maintaining equilibrium; the deficit budgeting is a means of attempting to cure disequilibrium if and when it arises” [Keynes, 1943, in CW 27: 352-353], the latter being a “particular, rather desperate expedient” [ibid.: 354]. Deficit budgeting through the Revenue or Ordinary Budget  is likely to involve short-term devices and temporary government spending that would not greatly influence private expectations. What is more, it might be possible that “government manipulations of tax and interest rates could only increase the state of uncertainty and the instability of behavioural relations” [Kregel, 1985: 45]. In contrast, capital budgeting  entails recurrent programmes and long-lived capital schemes, that are much more effective in fixing private expectations at the full employment target. As Dos Santos Ferreira puts it: “The ‘socialisation of investment’ involves the idea that the state assumes control of a coordination failure between consumers’ and entrepreneurs’ inter-temporal decisions.” [Dos Santos Ferreira, 2000: 290]
27The Capital Budget is the basic institution that must be established to effect the socialisation of investment. It aims to “balance and stabilise the Investment Budget for the national economy as a whole” [Keynes, 1945, in CW 27: 409]. Keynes’ plan is to prevent large fluctuations in private investment through long-term public capital expenditure. This is where we find the rule aspect of Keynes’ programme: “Emphasis should be placed primarily on measures to maintain a steady level of employment and thus to prevent fluctuations.” [Keynes, 1943, in CW 27: 323]. For Keynes, about the two-thirds of total investment would be involved, which seems at first sight quite comprehensive [Keynes, 1943, in CW 27: 322]. But “by ‘socialisation of investment’ Keynes did not mean nationalisation” [Skidelsky, 2011: 11] since it “need not exclude all manners of compromises and of devices by which public authority will co-operate with private initiative” [Keynes, 1936, in CW 7: 378].  As soon as one looks carefully at Keynes’ writings on policy it becomes plain there is no foundation for the exclusive association of Keynes with advocacy of ‘fine-tuning’ for the economy, nor with the idea that state intervention is a substitute for private initiative. The Keynes of World War 2 is mainly concerned with the implementation of a long-term employment programme capable of fixing expectations at full employment by the operation of definite rules. But this does not preclude small adjustments, what Keynes calls in another note the “proposals for adjusting its tempo to unforeseen changes” [Keynes, 1943, in CW 27: 357].  Here is to be found the opportunistic or discretionary side of his scheme: the state should not be prevented from judicious policy adjustments, so that Keynes’ programme would be better known as one based upon ‘discretionary rules’, “the means of ensuring stability in the long-term investment programme coupled with proposals for adjusting its tempo to unforeseen changes” [ibid.]. Hence, “Keynes alleged ‘short-termism’ applied not to the goals he chose for society but to the need for constant adaptation of policy measures to keep on the right path” [Peacock, 1993: 20].
28How might successful long-term employment policy be achieved? As we have seen, the first condition is to obtain information on private capital expenditure, “the continuous current collection both of the statistics of current performance and of prospective plans by the private sector” [Keynes, 1945, in CW 27: 409]. Here is the answer to Friedman’s opinion that “we do not know enough”. Secondly, Keynes opposes “wasteful and unnecessary enterprises” [Keynes, 1943, CW 27: 321]. On the contrary, he hoped that “capital expenditure would, at least partially, if not wholly, pay for itself” [ibid.: 319-20]. Hence such expenditure is supposed to be profitable even if only in the very long-run, and even if indirect benefits have to be taken into account. Here is the answer to Freidman’s caricature that the General Theory advocated “filling-holes type of make-work” [1962: 81].
29Let us now turn to Friedman’s criticism of fiscal policy. Over time he modified his criticism. In “Monetary and Fiscal Framework for Economic Stability”, the Friedman of 1948 is not opposed, as a matter of principle, to fiscal policy. Here “deficits or surpluses themselves become automatic consequences of changes in the level of business activity” [Friedman, 1948: 251]. While he then agreed with the role of the government budget as an ‘automatic stabiliser’, he did oppose fine-tuning of the economy, since delays added to price rigidities might “render extremely uncertain” [Friedman, 1948: 253] the efficiency of such a proposal. The Keynesian flavour of this scheme is striking, echoing for example Meade’s advocacy during WW2 of social security plans in line with the Beveridge Report.
30Later, in Capitalism and Freedom , Friedman’s argument against fiscal policy was reinforced. Friedman now insisted that, because of lags and delays, unfortunately “the balance wheel is unbalanced” [Friedman, 1962: 76]. Indeed, because of their ‘stop-and-go’ character and time-lags, governmental expenditures might actually be pro-cyclical. Hence “far from being a balance wheel offsetting other forces making for fluctuations, the federal budget has if anything been itself a major source of disturbance and instability” [Friedman, 1962: 77]. So far, Keynes would at least partially share Friedman’s objection. Indeed, Keynes himself acknowledged that “a fluctuating volume of public works at short notice is a clumsy form of cure and not likely to be completely successful” [Keynes, 1943, in CW 27: 326]. In Keynes’ view, that ill-timed measures might prove pro-cyclical cannot be excluded in principle. That is why Keynes lays so much emphasis on the long-term issue.
31Friedman then radicalized his 1969 assault on fiscal policy. Indeed, he now argues that:
I believe that the state of the government budget matters; matters a great deal—for some things. The state of the budget determines what fraction of the nation’s income is spent through the government and what fraction is spent by individuals privately. The state of the government budget has a considerable effect on interest rates.
33Even if, so far as we know, the term was not used by Friedman the argument clearly refers to the crowding-out effect associated with the ‘Treasury View’ during the Great Depression. For Friedman, “the state of the budget by itself has no significant effect on the course of nominal income, on inflation, on deflation, or on cyclical fluctuations” [Friedman, Heller, 1969: 51]— its sole effect being the distribution between public and private expenditures. By this stage Friedman’s position has become so radicalised that he definitely parts company with Keynes.
34We shall draw conclusions on state intervention in general below. For the time being we can draw some conclusions on the efficacy of fiscal policy. Despite the Keynesian flavour of Friedman’s position in 1948 where he acknowledged ‘automatic stabilisers’, Friedman the elder is characterised by a radicalised dismissal of the efficacy fiscal policy. Some of the arguments he advanced would not have been entirely denied by Keynes. Indeed, Keynes’ basic aim is to fix long-term expectations at their full employment level. From this perspective short-term devices are considered to be ‘small’ adjustments subordinated to long-term schemes. That is why measures relying on deficit budgeting and concentrating on counter-depression measures would in the short run prove relatively ineffective.
2.2 – Monetary policy
35There are two issues to be distinguished when dealing with Keynes on monetary policy. On the one hand, monetary policy might be impotent because of excessively pessimistic expectations of initial recovery following a protracted slump. On the other, the maintenance of a stable and very low rate of interest is critical if recovery is to be sustained. Keynes does in fact attribute efficacy to monetary policy—but this depends on particular characteristics of time and place:
I am far from fully convinced by the recent thesis that interest rates play a small part in determining the volume of investment. It may be that other influences, such as an increase in demand, often dominate in starting a movement. But I am quite unconvinced that low interest rates cannot play an enormous part in sustaining investment at a given figure.”
37Keynes concentrates on the long-term rate of interest because, being long-lived, capital assets are much more sensitive to long rates than short rates. However, this obviously involves the management of the entire rate structure. At the end of the thirties, when the first inflationary pressures appeared after the protracted slump, this insistence on a low and stable long-term rate of interest led Keynes to oppose forcefully dear money policy aimed at curbing a boom and hence avoiding inflation. For him, “the long-term rate of interest must be kept continuously as near as possible to what we believe to be the long-term optimum. It is not suitable to be used as a short-period weapon” [Keynes, 1937, in CW 21: 389]. Keynes at that time advocated planning and control rather than a rise of the rate of interest, since the latter amounted to “playing with fire” [Keynes, 1937, in CW 21: 389]. As a result he favoured what could be called in Friedman’s terminology a ‘monetary rule’: a stable long-term rate of interest that slowly decreased as the marginal efficiency of capital declined. How might such a ‘rule’ be implemented? First of all, a high degree of confidence has to be established with regard to current rates before attempting a further decrease, since the fall in government securities takes time to completely influence the entire interest rate structure. Moreover, if the Treasury were in charge of borrowing on long-term capital schemes, this would make the “management of the market generally” [Keynes, 1942, in CW 27: 280] easier.
38The counterpart of the budgetary policy advocated by Keynes can thus be formulated as “a policy of aiming at a long-term equilibrium rate, which helps to maintain average capital expenditures at the right figure, even if this requires a progressive change in the standard rate from time to time, a change which is more likely to be in the downward than the upward direction” [Keynes, 1944, in CW 27: 377]. We have shown above that fiscal policy is based on long-term plans, short-term devices being subordinated to the latter. There is symmetry with respect to monetary policy. On the one hand there is the general rule of a very low and stable long-term rate of interest. But on the other hand, there are also necessary adjustments to the rule when required by particular circumstances. So again Keynes’ monetary policy would be best called a ‘discretionary rule’.  In the simplest of the terms, the fundamental object of this rule is to manage the liquidity-preference of private actors, giving them confidence in their expectations.
39Let us now turn to Friedman. In his AEA Presidential Address he insisted that monetary policy “cannot peg interest rates for more than very limited periods” [Friedman, 1968b: 5]. Indeed, the relevant variable is the real rate, determined by the real forces of the economy. Monetary policy will be efficacious only in the short-run, during the lapse of time when private agents fail not correct their price-expectations. As soon as expectations are corrected, as soon as the bargained real-wage is the correct one, the expansionary effect of the policy disappears. In the long run, the sole effect of an expansive monetary policy is inflationary. Consequently, the objective assigned to monetary policy is to “provide a stable background for the economy – keep the machine well oiled to continue Mill’s analogy” [Friedman, 1968b: 13]. Therefore individuals “can proceed with full confidence that the average level of prices will behave in a known way in the future” [ibid.]. Friedman’s argument against discretionary monetary policy relies on an argument similar to the fiscal argument discussed above: “We do not know enough to be able to recognize minor disturbances when they occur or to be able to predict either what their effects will be with any precision or what monetary policy is required to offset their effects.” [Friedman, 1968b: 14] Hence, the monetary authority should “avoid sharp swings in policy” [ibid.: 15]. The monetary rule he then advocates to anchor private expectations consists in a fixed rate of growth of money. As argued in Capitalism and Freedom, “the precise definition of money adopted, or the precise rate of growth chosen, makes far less difference than the definite choice of a particular definition and a particular rate of growth” [Friedman, 1962: 54]. For Friedman, the best that monetary authorities can do is to provide a stable monetary framework to anchor private agents’ expectations of inflation. Friedman’s employment policy amounts to the stabilisation of the economy on its long-run growth-trend, which is consistent with the emphasis placed today on the need to damp fluctuations detrimental to collective welfare. But to base a monetary rule on a rigid money supply growth presupposes a highly stable velocity of money-demand –a conjecture that hardly fits into Keynes’ monetary economy. The most important practical difficulty in implementing such a rule is to maintain a ratio between the monetary base (currency plus bank reserves) and the money supply (currency plus bank deposits). To this end, Friedman argued as early as 1948 in favour of “a reform of the monetary and banking system to eliminate both the private creation or destruction of money, and discretionary control of the quantity of money by central-bank authority” [Friedman, 1948: 247). The purpose here was mainly to avoid the creation of endogenous money by the banking sector. Obviously, favouring rules over discretion does not preclude direct intervention in institutions. 
40What are the lessons to be learnt at this stage of our inquiry? It can hardly be argued that Keynes and Friedman are completely at odds, especially when debate involves the fine tuning of the economy. Both of them to some extent advocate rules, and favour long-term and known-in-advance measures. In particular, Keynes would have shared some of Friedman’s arguments against ill-timed short-term actions. Both would use monetary policy to stabilise the economic system. Yet our authors part company on the practical conduct of these rules. Indeed, on the one hand Keynes’ target is to keep the long-term rate of interest as low as possible. To this end, the monetary authority is supposed to “feed the market” [Keynes, (1942) 1978: 416] with securities having different maturities in order to saturate the liquidity-preferences of the public. Since the necessity for small adjustments is acknowledged as a matter of principle, Keynes’ rule is actually a ‘discretionary rule’. In line with his general ‘Middle Way doctrine’ as emphasised by O’Donnell , Keynes is quite pragmatic in his advocacy. On the other hand, Friedman’s rule involves fixed growth of the quantity of money, and public spending financed by bonds is definitely rejected because of the associated rise in the interest rate on the loanable funds market. In sharp contrast to Keynes, Friedman is highly dogmatic in the application of his rule. He wants to see public authorities tied hand and foot.
41Keynes’ and Friedman’s pleas appear similar insofar as they are formulated in terms of rules. But to understand their respective targeting of the rate of interest or the money supply rate of growth we should consider their respective views on the stability of the monetary side of the economy. Beyond these contrasting positions regarding what counts as the ‘real’, tending either to pragmatism or dogmatism, there remain diverging conceptions of the rate of interest. Indeed, their arguments are based upon opposed theoretical foundations: whereas for Keynes state intervention is precisely intended to influence private expectations, quite the opposite is true for Friedman: he consider that once private agents realise that they have been manipulated, the temporary effect of an expansionary policy disappears.
3 – Keynes and Friedman on the working of a monetary economy
42We now turn to the ultimate purpose of this inquiry, namely the theoretical underpinnings of the employment policy advocated respectively by Keynes and by Friedman. In our view, the explanation of their divergence over the state’s ability to behave efficiently in targeting full employment, despite similar arguments in terms of stabilisation, is to be found in their respective treatment of uncertainty, and the manner in which individuals are supposed to possess knowledge regarding the ‘states of the world’. Before analysing the working of a monetary economy, let us start with the relevance of Friedman’s ‘adjustment process’ to Keynes’ analytical framework.
3.1 – Friedman and Keynes on the ‘adjustment process’
43Friedman’s  point of departure in his study of the adjustment process that returns the economy to its previous natural rate position following expansionary policy—the rate of inflation excepted—is an unanticipated change in aggregate demand which leads first of all to an increase of output—and thus of employment—and prices.  This happens exclusively because of errors of perception on the part of both employers and employees. This is because “employees and employers have access to current information regarding prices of the goods and services they sell but are subject to misperceptions regarding prices of the things they buy” [Garrison, 1984: 121]. As soon as the perceived real-wage corresponds to the existing one, both workers and employers adjust their respective supply and demand, and the economy comes back to the vertical long-term Phillips curve. The sole means of obtaining a permanent gap between the ‘market’ and the ‘natural’ rate of unemployment is to permit an inflationary spiral. Hence, “only surprises matter” [Friedman, 1977: 456].
44What does Friedman’s ‘adjustment process’ tell us about the divide between Keynes and Friedman? Firstly, Friedman postulates full employment, so that there is obviously no room for discretionary policy to fill a supposed deflationary gap. If the analysis stopped here the question of our authors’ respective standpoints would degenerate into tautology: active state policy is considered to be efficient or inefficient quite independently of whether the economy is supposed to have already attained its full employment level. Clearly, Keynes would agree with Friedman on the following point: if effective demand continues to rise when the economy is already at its full employment level, a further rise in employment is not excluded as a matter of principle, but would be only temporary—so long as entrepreneurial misperceptions are not corrected.  As is now well-known, Keynes’ argument in favour of public spending in a state of underemployment relies in no way on money illusion. 
45It is not merely by chance that, when Keynes takes account of erroneous expectations, over-optimism is attributed only to the employers. In Friedman’s system, the employment-level can—temporarily—rise only because of misperceptions of the labour market. Ultimately, it is labour market functioning that determines the employment level. In Keynes’ monetary framework, it is quite otherwise: “The volume of employment is uniquely related to a given level of real wages—not the other way round.” [Keynes, 1936, in CW 7: 30, emphasis added]. This derives from the monetary character of the economy, “for there may be no method available to labour as a whole whereby it can bring the wage-goods equivalent of the general level of money-wages into accordance with the marginal disutility of the current volume of employment” [Keynes, 1936, in CW 7: 13].  Here is to be found the first theoretical underpinnings of Keynes’ and Friedman’s respective views on appropriate employment policy, namely their treatment of the labour market. For Friedman, and also the Classics targeted by Keynes in Chapter 2 of the General Theory, the employment level is directly determined by the real wage level. Even if the labour force bargains over the money-wage, it is as if wage-bargains were to bear directly on the real wage—except in case of transitory misperceptions. That is, a rise in employment is directly obtained through a fall in real wages. Instead, in Keynes’ monetary economy, labour market functioning is merely the by-product of what happens on the goods market. Real wages will move according to changes in the employment level. The labour force cannot bargain with anything but money wages. In our view, this divergence in treatments of the labour market is evidence of differing conceptions of the general functioning of the economy, and in particular of the interactions between real and monetary spheres. In what follows below, we try to go a step further and to make explicit why Friedman is led to postulate full employment in his analysis of the adjustment process.
3.2 – Keynes’ “monetary economy” versus Friedman’s “neutral economy”
46In Friedman’s economy expansionist policy is—temporarily—efficient only when unanticipated, thus leading to faulty perceptions. For him, discretionary state intervention is destabilising per se. By contrast, in Keynes’ economy expansionary policy is more efficient when anticipated, since it is precisely intended to manage the private sector long-term state of expectations. To explain this somewhat paradoxical result, our hypothesis is that Keynes and Friedman actually deal with similar economies: in neither of these economies is money a ‘veil’ since it might be used to insulate oneself against unexpected events, but all the same quite different kinds of uncertainty prevail.
47Let us start with Friedman’s conception of uncertainty. This issue is explicitly tackled for the first time in “In Defense of Speculation” . In this plea against market instruments, such as the holding of ‘buffer stocks’, Friedman analyses why people would accept payment of a premium to insure themselves against price fluctuations. The analogy made in this paper refers to casinos, lottery games and the like, so that ‘states of the world’ are fully known in advance, just like the distribution of probability. In Knight’s usage , ‘risk’ rather than ‘uncertainty’ prevails here. In such a context Friedman is right in arguing that speculation ensures the marketability of commodities, rather than being destabilizing. This issue of uncertainty is taken up in “The Optimum Quantity of Money” .  Here, Friedman’s conception of uncertainty is applied to his money demand function. The fifth characteristic of Friedman’s monetary framework runs as follows: “The society, though stationary, is not static. Aggregates are constant but individuals are subjects to uncertainty and change. Even the aggregates may change in a stochastic way, provided the mean values do not” [Friedman, 1969: 2], what he calls “individual uncertainty” [ibid.: 3]. Consequently, in a Friedmanite world people would of course hold money to protect themselves against these stochastic changes. Indeed, besides the transaction-purpose:
A second reason for holding money is as a reserve for future emergencies. In the actual world, money is but one of many assets that can serve this function. In our hypothetical world, it is the only such asset. This reason corresponds to the ‘asset’ motive for holding money.
It should be noted that both reasons depend critically on characteristic (5) of our economy, the existence of individual uncertainty.
49Again, as the distributions of probability are fully known by individuals there is no uncertainty as such, but only risk. Even if present in the money demand function, the rate of interest actually plays the role of a borrowing cost, an opportunity cost for individuals who insure themselves against the risk that the value of their assets may vary because of random error. The motive of money demand involved here is actually the precautionary-motive in Keynes’ meaning of the term. There is no liquidity-risk, so that Friedman ignores “the essence of uncertainty” [Davidson, 1974: 91]. More precisely, Friedman’s agents “are not Keynesian ‘liquidity preference’ speculators between money and bonds. They are Fisherian speculators between goods, or equities in goods, and money” [Tobin, 1974: 81]. Individuals might hold money, but definitely not for a liquidity motive.
50Now, in Keynes’ system, “no sampling from the future is feasible to ascertain probabilities for future alternatives, so there is no way uncertainty problems can be reduced to problems involving risk” [Weintraub, 1975: 532]. In an economy where uncertainty rather than risk prevails, in which “we simply do not know” as Keynes [1937, in CW 14: 114] famously puts it, money is not a good substitutable for all assets held, including semi-durable goods. As the distribution of probability is not fully known, there is now room for a liquidity risk. As stated by Hayes [2006: 21], “liquidity has value only because the future is unknown, and its value increases with our fear of what might happen that we cannot prevent or insure against”. The speculative-motive is defined by Keynes as “the object of securing profit from knowing better than the market what the future will bring forth” [Keynes, 1936, in CW 7: 170]. To put it very briefly, Keynes’ speculators do not face risk. In a way, they know better than ‘the’ market what the future will be, and then play against the market.
51These contrasted conceptions of uncertainty lead our authors to hold quite different conceptions of the role played by the rate of interest, and thus adhere to quite different monetary rules. For Friedman, “the price of money …is the inverse of the price-level—not the interest rate. The interest rate is the price of credit” [Friedman, Heller, 1969: 75]. Public spending financed by bonds will simply crowd-out private investment: “If the federal government runs a large deficit, that means the government has to borrow in the market, which raises the demand for loanable funds and so tend to raise interest rates.” [Friedman, Heller, 1969: 50] For Keynes of course, quite the reverse is true: the rate of interest is par excellence a monetary phenomenon that equilibrates a stock market of financial assets and not a flow credit market. As Leijonhufvud clearly shows, in a world where newly produced capital goods are adequate substitutes for existing financial assets the rate of interest is not merely a borrowing cost.
52What Keynes’ and Friedman’s economies have in common is the use of monetary assets to protect against unexpected events. To this extent, the monetary side of their theoretical worlds seems similar. Yet Friedman basically deals with actuarial risk, whereas Keynes’ concern goes far beyond that and is basically directed to liquidity-risk. Hence liquidity preference appears to be a ‘blind spot’ in Friedman’s framework. To highlight this point, it seems useful to come back to a classification of different types of economy used by Keynes in his work prior to the General Theory. Indeed, in the first drafts of this work, Keynes distinguishes three different kinds of economy: a real wage economy – also called a cooperative economy; a neutral economy; and finally an entrepreneur economy – also called a money-wage economy. In the first one, “the factors of production are rewarded by dividing up in agreed proportions the actual output of their cooperative efforts” [Keynes, 1933, in CW 29: 77]. In a neutral entrepreneur economy, the factors of production are rewarded in terms of money. However, “there is a mechanism of some kind to ensure that the exchange value of the money incomes of the factors is always equal in the aggregate to the proportion of current output which would have been the factor’s share in a co-operative” [Keynes, 1933, in CW 29: 77]. Finally, in an “entrepreneur economy” (also called a “money-wage economy”) the factors of production are rewarded in terms of money, but the “mechanism of some kind” that ensures that the actual product will be bought back is lacking. The key characteristic of a monetary economy is then the following:
Money is par excellence the means of remuneration in an entrepreneur economy which lends itself to fluctuations in effective demand. But if employers were to remunerate theirs workers in terms of plots of land or obsolete postage stamps, the same difficulties could arise. Perhaps anything in terms of which the factors of production contract to be remunerated, which is not and cannot be a part of current output and is capable of being used otherwise than to purchase current output, is, in a sense, money. If so, but not otherwise, the use of money is a necessary condition for fluctuations in effective demand.
54If the factors of production were paid with old stamps they would be able to hoard stamps and thus to defer purchasing power for an indefinite period. In this sense, stamps would then become ‘money’ insofar as they will be used some day as a means of payment.
55Friedman’s economy does actually correspond to a “neutral” entrepreneur economy as defined by Keynes. There is no necessity to hold money except to prevent oneself against risk. There is no basic impediment to full employment and thus no role that the state is supposed to play, except to keep the economy ‘well-oiled’. By contrast, in Keynes’ “entrepreneur economy”, where uncertainty prevails, private agents have to form expectations, but they also need to have confidence in those expectations. Here a key role for the state is to control the long-term state of expectation by means of a long-term employment policy.
56Friedman’s “Theoretical Framework for Monetary Analysis”  offers a common background that both monetarists and Keynesians should support. His money-demand function is a good example of an analytical proposition that aims to encompass both the ‘old’ quantitative theory of money and the Keynesian ‘absolute liquidity preference’. If it had been successful, Friedman’s undertaking would have shifted the quarrel between the interventionist and the non-interventionist wings of economists on to empirical ground and reduced the debate among them to the estimation of this or that parameter. In turn, no insuperable gap would remain in the field of economic policy. Yet the above investigation has shown that the matter at hand is much more complicated than Friedman seems to believe.
57What about the divide between Friedman and Keynes over appropriate employment policy? Our authors do share some true points of agreement. Neither of them believes in the efficiency of consumption schemes, and neither of them would rely on inflationary remedies. In their analysis of the labour market they take account of structural unemployment in a similar manner, even if Keynes’ full employment target entails a lower number of unemployed workers than Friedman’s natural rate. More generally, both Keynes and Friedman develop long-term schemes and rules-based frameworks. So their respective arguments do not appear to be too distant at a formal level. On the other hand, when one looks at the theoretical reasons for these arguments, strong disagreements show up. For Friedman, state discretionary intervention is—temporarily—efficient only when unanticipated. The ultimate reason why he favours long-term rules is his basic distrust of all collective rationality. In sharp contrast, for Keynes the true purpose of state intervention is to manage long-term expectations. An active policy is more efficient when anticipated, so that short-run devices would be of little help. Keynes develops long-term rules at both the fiscal and monetary level in order to encourage private actors to assign a high weight to their anticipation of full employment. Since he is highly confident in the capacity of the government to tend to collective efficiency, Keynes does not exclude small adjustments to his rules—depending on circumstances and advisability. Hence the elaboration of what can be called ‘discretionary rules’.
58As a result, if the debate were formulated in terms of rules versus discretion, both Keynes and Friedman would take sides in favour of rules; but for opposite reasons. In Friedman’s theoretical world, money is only required for transactions and precautionary purposes; there is no liquidity preference as such because there is no uncertainty as such. Here, expectations cannot be destabilizing since the ‘states of the world’ are always fully known. There is no room for a collective body in possession of knowledge superior to that of private individuals. For Friedman, an employment policy based on a rigid monetary rule is enough to support the self-adjusting capacity of a “neutral economy”. By contrast, in Keynes’ world, current supply will not necessarily be recovered if private agents prefer to stay ‘liquid’ because of their distrust in their own expectations. Indeed, “Keynes held that the process of expectation formation in our economy is potentially destabilizing” [Minsky, 1983: 97]. Hence the key role played by the state as a collective body to provide private actors with better knowledge regarding future ‘states of the world’. So Keynes’ “entrepreneur economy” also requires rules, but these ‘discretionary rules’ are an absolute necessity to coordinate private individual expectations. On the ultimate issue of the state, viewed as a collective body which might possess knowledge superior to that of individuals, and that might tend towards collective efficiency, the gap between our authors could not be larger.
Université de Mulhouse – Groupe de recherche sur l’Apprentissage, l’Innovation et les Connaissances dans les Organisations (GRAICO). Courriel : firstname.lastname@example.org
However, the complete quotation runs as follows: “It was the experience that led me to tell a reporter for Time Magazine that ‘in one sense, we are all Keynesians now; in another, no one is Keynesian any longer’. We all use the Keynesian language and apparatus; none of us any longer accepts the initial Keynesian conclusions.”
Again, the complete quotation is the following: “There are in reality no serious analytical disagreements between leading monetarists and leading non monetarists. Milton Friedman was once quoted as saying, ‘We are all Keynesians, now’, and I am quite prepared to reciprocate that ‘we are all monetarists’ – if by monetarism is meant assigning to the stock of money a major role in determining output and prices” [Modigliani, 1977: 1].
Denial of the ‘equivalence principle’ between these two definitions of full employment as the absence of involuntary unemployment turns out to be the core of the investigation by De Vroey .
See for example: “The distinction between ‘voluntary’ and ‘involuntary’ unemployment, while important conceptually as a basis for the Keynesian system of analysis, has not proved to have any practical significance, either in terms of statistical measurement or in terms of targets or objectives.” [Kahn, 1976: 27].
“Thus instead of constant prices in conditions of unemployment, and of prices rising in proportion to the quantity of money in conditions of full employment, we have in fact a condition of prices rising gradually as employment increases.” [Keynes, 1936, in CW 7: 296]
In Keynes’ words: “…in general the demand for some services and commodities will reach a level beyond which their supply is, for the time being, perfectly inelastic, whilst in other directions there is still a substantial surplus of resources without employment.” [Keynes, 1936, in CW 7: 300]
If one remembers Friedman’s claim that he, like Keynes, belonged to the Marshallian rather than to the Walrasian tradition, his definition of the natural rate of unemployment in Walrasian terms is puzzling. As Tobin already noticed, “we know little about the existence of a Walrasian equilibrium that allows for all the imperfections and frictions that explain why the natural rate is bigger than zero and even less about the optimality of such an equilibrium if it exists” [Tobin, 1972: 6]. The first explanation is straightforward and relates to tactics: Friedman would have tried to convince his opponents, like Patinkin or Tobin who were Walrasians albeit Keynesians, on their own theoretical ground. Hence, he would have “enhanced his policy-revolution by embracing the language of his opponents (IS-LM, econometrics, income-expenditure, money-demand), a language that he was often sceptical about” [Leeson, 2000: 129]. Another explanation of the natural rate of unemployment defined in terms of Walrasian general equilibrium implies that Friedman’s views on the respective methodologies of Marshall and Walras requires specification: “The Walrasian approach is ‘concerned with abstract completeness’, in contrast to the Marshallian approach which is ‘concerned with the construction of special tools for special problems’.” [Leeson, 2000: 114] As a consequence, the natural rate hypothesis would correspond to two separate issues: on the one hand, the existence of the natural rate as a fact that can be measured, on the other, the adjustment process around the natural rate of unemployment.
De Vroey ’s criticism goes beyond the relevance of Friedman’s natural rate of unemployment within a Walrasian general equilibrium system. What he points out is the internal inconsistency of Friedman’s concept. For him, the narrative that explicitly discusses unemployment contrasts with the model behind this narrative (to be found in Friedman’s Price Theory ) that actually features full employment. Indeed, Friedman’s result “can be viewed as a standard case of Marshallian disequilibrium, featuring a deviation of the market-day equilibrium values …from their normal or long-period equilibrium values” [De Vroey, 2001: 133]. Hence “departures from the natural rate of employment rather than from the natural rate of unemployment should be the real object of Friedman’s model” [De Vroey, 2001: 133-134].
In 1937 Keynes opposed a general expansionary policy when the average unemployment rate lay above 12% in Great Britain. To rationalise such an argument, it must be kept in mind that at that time unemployment was much lower in the South that in the North. Hence the application of the analysis carried out in Chapter 21 of the General Theory in terms of bottlenecks: “We are more in need today of a rightly distributed demand than of a greater aggregate demand.” [Keynes, 1937, inCW 21: 385] At that time, Keynes argued for “boom control” [Keynes, 1937, in CW 21: 390] taxation, the postponement of investment projects and the promotion of imports. Such a plan requires firstly that a Public Investment Board be established, “an authority whose business is not to launch anything at present, but to make sure that detailed plans are prepared” [Keynes, 1937, in CW 21: 394]. Second, this also requires the cooperation of private actors, for example in efforts to locate enterprises in distressed areas and encouraging labour mobility—what we would now call supply-side policies. Last, the cooperation of trade unions is also required, especially as far as incomes policy is concerned.
Indeed, one can see him arguing: “The Government, in my opinion, should make it a deliberate act of policy to break down that secretiveness, that failure to secure the collection of knowledge which is important to industrial society as a whole.” [Keynes, 1927, in CW 19: 644]
Coming back to this issue in 1996, Friedman clearly shows that the question at stake is not empirical. Indeed: “As the coiner of the term, I am disturbed as its widespread misuse and misunderstanding. The natural-rate is not a fixed number. It is not 6% or 5%, or some other magic number … The natural-rate is a concept that does have an empirical counterpart—but that counterpart is not easy to measure and will depend on particular circumstances of time and place.” [Friedman, 1996; quoted in Leeson, 2000: 122]
In Keynes’ words: “Unfortunately the more pessimistic the Chancellor’s policy, the more likely it is that pessimistic anticipations will be realised and vice versa. Whatever the Chancellor dreams, will come true! We must begin by resuscitating the national income and the national output.” [Keynes, 1933, in CW 21: 184]
Bateman goes further than an emphasis on uncertainty, already acknowledged by Kregel , O’Donnelll  or Skidesky [1998, 2000]. According to Bateman, the early Keynes adheres to the Cambridge theory of the trade-cycle and takes into account both uncertainty and expectations. He then rejects this theory until the turning point in 1933. Keynes’ treatment of expectation derives from his practice as an investor and his involvement in policy-making during the early 30s, rather than being “an intellectual by-product of his work as a philosopher” [Bateman, 2003: 78]. One would not necessarily follow Bateman when he maintains that “what happened to Keynes’s thinking on economic policy was a complete volte-face” [Bateman, 2003: 81). His marked emphasis on uncertainty can for example already be found in The End of Laissez-Faire . And it can be easily argued that “despite variations in language the idea of planning (in his sense) was present in his writings from at least 1922 onwards, and …planning was as prominent an element in his writings as it was before 1939 as it was afterwards” [O’Donnell, 1989: 311]. On this issue, see also O’Donnell .
Deficit budgeting corresponds to “the policy of collecting in taxes less than the current non-capital expenditures of the state as a means of stimulating consumption” [Keynes, 1945, in CW 27: 406].
The Revenue or Ordinary Budget relates to current spending, the ordinary expenses and revenues of public authorities.
The capital budget is defined as “a compilation of and budgetary forecast of all capital expenditure under public control, including local authorities and public boards” [Keynes, 1945, in CW 27: 405].
However, as Moggridge notes [2002: 116], “it is not the case that Keynes ‘always’ opposed … greater public ownership”. Keynes himself stated that “the whole purpose of nationalisation is to allow considerations of general and social policy to take their proper place in price policy” [Keynes, 1944, in CW 22: 465].
See also for example: “The main task should be to prevent large fluctuations by a stable long-term programme. If this is successful it should not be too difficult to offset small fluctuations by expediting or retarding some items in this long-term programme.” [Keynes, 1943, in CW 27: 322 ; emphasis added]
See for example, among others, Howson  and Moggridge and Howson .
See also the following quotation: “No dogmatic conclusions should be laid down for the future about the rates of interest appropriate to different maturities, which should be fixed from time to time in the light of experience.” [Keynes, 1945, in CW 27: 396-397].
The fact that contemporary Taylor rules target the rate of interest is clearly explained by a velocity of money demand no longer thought to be stable, mostly because of extensive securitisation and a massive growth of broad money.
We deal here with Friedman’s  Nobel lecture rather than his 1968 article. Indeed, as shown by Garrison , Friedman’s position has altered since then. Imperfect information has been extended to the perceived real wage, allowing for either a decreasing or an increasing real-wage during a—transitory— expansion.
Even if full employment is attained, “for a time at least, rising prices may delude entrepreneurs into increasing employment beyond the level which maximises their individual profits measured in terms of the product …i.e. they may underestimate their marginal user cost in the new price environment” [Keynes, 1936], CW 7: 290]. No doubt once erroneous expectations are corrected entrepreneurs return to their full employment position.
It might be noted that the lack of money illusion is an argument put forward by Keynes during World War 2 in opposing inflationary spirals: “Everyone, including the trade-unions, has become index number conscious. Wages will pursue prices with not so lame a foot. And this new fact means that the old-type laissez-faire inflation is no longer to be relied upon.” [Keynes, 1940, in CW 22: 120]
It might be noted that here money-wage bargains then determine the relative real wage rather than the general level of real wage. On the relative-wage argument, see Hoover .
Reproduced as Chapter 1 of The Optimum Quantity of Money and Other Essays .