John Maynard Keynes: saving marginalism

Promoting the national economy divides workers

Does Keynesianism represent an alternative to austerity? Mike Macnair begins by looking at John Maynard Keynes's actual theory

This article is an expanded version of my talk at the CPGB school, ‘Fundamentals of political economy’ (January 22). It is expanded because when I came to edit the transcript, I found that a large and necessary chunk of argument was missing due to the time constraints at the school. I have also provided more of an introduction than I gave at the school.

The increased length means that the article will appear in two parts. This first part consists of the introduction and engagement with Keynes’s actual arguments. The second part will consider the ‘golden age’, around 1950-70, in which ‘Keynesianism’ appeared dominant, whether a return to this period is a realistic or desirable policy, and how we can propose reforms within capitalism without seeking a ‘Keynesian’ policy and government.

In 2008 a financial bubble partially burst. The response of states worldwide was (in effect) to print large amounts of money to be poured into the banks to bail them and the rest of the financial system out (the technical measures involved were considerably more complex, but the details are irrelevant for present purposes).

In spite of these measures, the crash had an immediate effect of tightening the availability of credit. Since businesses depend on credit for their everyday operations, there was a - brief - sharp decline in the productive economy before the temporary stabilising effects of printing money kicked in. In Britain, there was a rash of short-time working and wage cuts in the private sector, though not massive redundancies; the government and the banks agreed to avoid a wave of foreclosures on mortgages which were (and remain) in default, in order to avoid a crash in the price of houses and commercial property: ie, keep at least this bubble semi-inflated.

The next step was that, since states were required to print money, and are able to do so with effects other than merely producing hyperinflation (because states borrow too), the problem shifted to the relative credit-worthiness of states. This problem began with Iceland and Ireland, is ongoing with Greece, and threatening in relation to Portugal and Italy. Another global financial crash, looming around the end of 2011, was averted (at least temporarily) because the European Central Bank agreed to print still more money to bail out European banks.

Middle class savings

The capitalist class is a small social group. The ‘Occupy’ movement’s ‘1%’ slogan if anything overstates its size. This minority is only able to rule over the society through the support of much larger social strata: the middle classes, both ‘classical’ petty bourgeoisie (small businesses, including the genuine self-employed)[1] and employed middle class (managers, bureaucrats, etc).

This support is material and financial and not merely ideological and political. To be middle class is to be (slightly) above the working class, to continue to be so in old age, and to be able to pass that social position on to your children. To achieve these effects, the middle classes have to save. A small business, and a fortiori self-employment, is not marketable at the capitalised value of its current income stream, because its future income depends very sharply on the personal characteristics of the owner-manager. So merely keeping the business going or selling it off will not (usually) provide sufficiently for old age or the kids.

Before capitalism, savings took the form of gold and silver hoards (withdrawing money from circulation), small land acquisitions, and local lending based on interpersonal trust. In capitalism, a large part of these savings are instead mobilised by financial institutions (starting from state borrowing and markets in transferable state debt) into large accumulations of money capital. Today the level of financial intermediation is great: most state debt and company shares are held by institutions; but these institutions still depend, in the last analysis, on a mass of middle class small savers, who are less likely to make large destabilising withdrawals than large capitals.

Printing more money to bail out the banks saves both big capitals and small savers from immediate loss of their capitals/savings. But it also dilutes the relative value of the savings. This is true even if it does not produce immediate consumer price inflation: to retain their value, savings have to keep pace with capital asset prices, not with consumer prices.

From these circumstances flows a necessary narrative for capitalism: that the costs of crisis must fall primarily on wages and the working class. To say that the costs must fall on capital (which is in the end true) would be a suicidal degree of altruism for any individual capitalist; to say that they must fall on small savers and not on big capitals would destroy the social base of capital, not merely in the sense of causing middle class disaffection, but also because, if the returns through finance are bad enough, middle class savings can return to hoards, small land acquisitions and local interpersonal lending, which would destabilise the financial system.

In the present crisis - more exactly, the aftermath of the 2008 crisis - this necessary narrative is expressed as ‘austerity’. States, the capitalist media and politicians claim, have been ‘living beyond their means’ by undue provision of welfare to the working class (and employed middle class, but these are less mentioned). The value of wages and in particular of the ‘social wage’ - collective welfare provision - must be reduced.

The particular form of the narrative responds to recent conditions. Its underlying content is an old, 19th century dogma of bourgeois economics, already present in Malthus, and persistent in the marginalists: that economic crisis and unemployment are caused by excessive wages, and real wages must fall to restore the conditions for capitalist equilibrium.

The responses of the left (in a very broad sense of ‘left’) to the ‘austerity’ turn have been almost across the board dominated by the counter-narrative of Keynesianism. From the Labour leadership and ‘let us cut more slowly, let us have a plan for growth’, through the majority of the Labour Party, through to the far left. Even the Socialist Workers Party, effectively abandoning its own history of criticising Keynesianism, is advocating Keynesian economic stimulus programmes and is doing so particularly in relation to Irish politics.

But how much sense, and perhaps more importantly what sort of sense, does this counter-narrative make?

Alternative to capitalism

The problem starts in a sense with a fundamental question: what is the real alternative to capitalism? The answer is conscious cooperation, among our different productive activities, as opposed to private, individual decision-making.

Marx’s Capital is in part a polemic against the Proudhonist case for cooperatives linked by a market as an alternative to capitalism.[2] In the first part of Capital volume 1 it is shown, among other points, that both a return to individual artisan production and cooperatives linked by a market (as opposed to linked by cooperation on a scale wider than the individual productive enterprise) will naturally and automatically develop into capitalism. By the mechanism of market competition itself, small differences in productivity are amplified, and this generates a process in which artisans - or cooperatives linked by a market - turn into groups of exploiters and exploited.

This is, in fact, also a description of what happened to artisan production in the late medieval and early modern towns, and what happened to peasant farming in late 15th through middle 16th century England.[3] The original equal, free, independent producers, by virtue of marginal differences in productivity, turn into capitalists on the one side and wage-workers on the other.

More immediately, cooperatives under capitalism - ie, leaving in place capitalist control of the financial system, which coordinates production, and the form of judicial corruption called the ‘free market in legal services’ - wind up as an analogue of the old ‘putting out system’, in which artisans carried on production on a household scale, but merchants controlled access to raw materials, finance and the marketing of the produced goods: what Marx called the “formal subsumption of labour to capital”. A significant number of the producer cooperatives of the 1850s-60s wound up in this position.

Hence, the realalternative to capitalism is cooperation across the level of the whole material division of labour. That is, we need to link in relations of conscious cooperation primary materials producers, workers in manufacturing industry, in transportation, warehousing, distribution and retail, in planning activities (the lower parts of management), in all forms of education and training, and so on. There are, of course, activities in a capitalist society that a socialist society could get rid of: eg, advertising, finance practised by a large part of the multiple competing institutions, and a lot of related legal services. But in order to have something other than capitalism, without its tendency towards extremes of rich and poor, recurrent crises and ultimately world wars, there must be conscious cooperation across the whole division of labour and democratic methods of decision-making for questions which affect us all.

Karl Kautsky argued in The class struggle,his extended explanation of the German Social Democratic Party’s 1891 Erfurt programme, that the national scale - at least in large nations - was sufficient for a cooperative-order alternative to capitalism.[4] That idea came down to us in modern times via Stalin as the idea of ‘socialism in one country’.[5]

The problem is that the material division of labour is played out on an international scale, was already on an international scale in the 1890s, and has been on an international scale since the later middle ages. For example, early enclosure in late medieval England was characterised as ‘sheep eating up men’. What that meant was that grain production was being replaced by sheep on grassland. The wool was being produced for sale to the Netherlands, where putting-out merchants were engaged in wool cloth production. The merchants were selling the products across the whole of western Europe and into the Islamic world. From the Islamic world they were getting back goods not produced in western Europe, and goods that the Islamic merchants were buying in from east Asia. An example: spices. Not a simple luxury: spices were used to preserve foods (for example, pickled herring, produced by industrial methods by early Dutch capitalism), and by increasing the life of foods, the same amount of food could feed more people.

So the material division of labour is beyond the scale of the nation, and has been since the earliest beginnings of capitalism. But the left puts forward Keynesian demands on the nation-state, which by their naturerequire that the working class should support the nation-state against rivals. Hence, on the one hand, these demands divide the working class along the lines of nationality: so that the Greek workers have to find a solution on their own. On the other hand, they imply unity with the national bourgeoisie against an imagined ‘cosmopolitan’ capital - or, immediately, ‘Eurocrats’.

This character actually flows from the nature of Keynes’s original project. It is not particularly surprising from the dominant elements of the left. Labourism has always been a nationalist and class-collaborationist project, and continental social democracy became such a project during and after World War I. ‘Official communism’ has been defined since 1935 by the projects of socialism in one country, national roads to socialism and the people’s front, as well as party monolithism. So it is unsurprising that these trends should find Keynesianism attractive. The SWP was founded as an enemy of social democracy and Stalinism, but, as it has sought to take over the role of the old ‘official’ CPGB, it has gradually been moving ideologically towards ‘official communism’.

The fact that Keynesian perspectives have become the only ones put forward as an alternative to capitalist ‘austerity’ - except by very small minority groups on the far left and unorganised individuals - is a deep political obstacle to the working class organising itself independently of the capitalist class and taking its own independent political initiatives. It is equally a deep obstacle to cooperation at the level at which the division of labour exists. As long as working class politics is cantonised between different nation-states, there is a drive to accept the infernal logic that British competitiveness requires us to defend the City of London, to accept lower wages, and so on.

Now it might nonetheless be the case that we have to do this. This sort of argument has been put forward by, for example, Dave Douglass in the letters column of the Weekly Worker: that if we do not support national protectionist demands, we are saying that ‘nothing can be done before the revolution’. Or the argument which the leaders of the SWP use for raising Keynesian demands: ‘We have to apply the policy of the united front: that means moderate demands, militant action.’ That is, we have to accept the demands which are being put forward by the right wing of the labour movement, but then say that in order to win them you need all-out strike action. The upshot of these and similar arguments is that the idea of a real alternative to capitalism is silenced because the left, which might put forward an alternative, instead puts forward Keynesianism.

To address this issue I will take a step back. I will look briefly at Keynes’s own arguments; and equally briefly at their real support, the so-called ‘golden age’ between 1950s and 1970s when Keynesianism was a widespread ideology in the economics profession. How far did the Keynesian theory actually describe what was happening in terms of state economic management in that period? I can then come back to the question, what does this policy mean now?


Over Christmas I spent some tedious time slogging through Keynes’s General theory of employment, interest and money.[6] It was tedious because Keynes’s actual arguments are logically incoherent. This is because in the first place he specifies his definition of things like wage, interest, capital and so on, on the basis that nothing can be a valid definition unless it can be expressed in, or form part of, a differential equation. If you cannot do Δx Δy equations with a definition, then he says (to paraphrase) that definition cannot be either true or false (pp40, 153-53). We are to create definitions on the basis of the antecedent (and unargued) postulate that it has to be possible to describe the economy through a system of differential equations.[7]

The consequence of this, together with the nature of Keynes’s underlying political-economic project, is that he actually shifts definitions between chapters. When he is doing one set of equations he has one definition for - say - the wage; when he is doing another set of equations he has another definition for the wage. So he shifts his analytical foundations and usage of ‘wage’, ‘equilibrium’, ‘capital’ and so on from chapter to chapter of the book and you have to keep tracing backwards and forwards in order to try and see if there is logical coherence. My conclusion was that there is not.

The underlying project of Keynes’s book is to save marginalist economics. In some sense it is to save capitalism from itself, but it is actually narrower: Keynes criticises marginalist economics, in order to save it.[8]

Marginalism (to express the point in a highly simplified way) is related to the issue of expressing economic claims in the form of differential equations. The proposition is that the underlying explanatory power of economics has to be the explanation of marginal changes in price, etc; and that this can be done in the form of differential equations. Hence we postulate definitions of economic phenomena which will enable differential equations to describe an equilibrium state of capitalism.

The marginalists’ postulated definition of the wage is from two angles. One is the marginal product of labour, meaning that the amount which is produced by the addition of one extra worker fixes the wage. If you have x workers employed and you move to x+1, the wage is fixed by the marginal product of that one extra labourer. The second is the marginal disutility to the worker of working one extra hour. Marginalists provide analogous definitions of capital, but this is not material here.

Now these definitions of the wage implicitly postulate a priori that the wage could fall indefinitely (or, for that matter, it could rise indefinitely). There is no material floor. This is a conception of economics without natural limits. The marginalist definitions of the wage presuppose that workers could in principle work for more than 24 hours a day on zero calories.

The logical conclusion from this approach was one defended by Arthur C Pigou, against whom Keynes polemicises. In Pigou’s view there is no such a thing as involuntary unemployment, because the work would be offered if only the unemployed were prepared to accept lower wages. OK, it is a bit hard for them if they accept work for which pay is only 10% of the rent which they have to pay and leaves nothing whatsoever for food; tough, the marginalist would say, that is just how it is; and, of course, in the free market if workers accept work for less than subsistence costs, as enough of them starve to death (Malthus’s solution to the problem of unemployment), lack of demand will force the prices of subsistence goods down.

Keynes, writing in the early 1930s, wants to save the general methodology of marginalism from Pigou’s conclusion that there is no such a thing as involuntary unemployment; because in the early 1930s the fact that there is involuntary unemployment is blindingly obvious. It was plainly untrue that people were unemployed because they were skivers and it was ridiculous to suggest it. Hence at the time of Keynes’s writing, marginalism was under severe ideological pressure in the form of ‘respectable people’ and fellow academics. Marginalism was regarded not even as flat-earthism, but as analogous to a ‘scientific theory’ whose central conclusion is that the sun rises in the west.

Hence, part of what Keynes does is multiply shifting redefinitions, which enable him to give what are not really marginalist answers, without directly interrogating the underlying marginalist assumptions. There are episodes in General theory where Keynes slips towards the labour theory of value or admits that labour inputs actually seem to be more predictive than marginalism; but then he shifts back again (pp213, 215, 252).

Marginalism makes the assumption that time does not exist and then reinstates it. This is part of the point of the centrality of the calculus. The method is one of ‘comparative statics’: you have a series of distinct points in time, and you draw a curve on the basis of an equation which models the movement between these distinct points in time. But the distinct points in time are pre-given, because the method is in origin about the method of calculating orbits, which, although they are not actually perfectly stable, nonetheless are sufficiently stable that they appear to move, yet at the end of the day remain the same.

Marginalism is, analogously, looking at the economy on the basis that it moves, yet at the end of the day remains the same. This is not unique to Keynes, because all the marginalists do it: you may introduce time and uncertainty at specific points, but at the foundation, in the mathematics used to test claims, you are still analysing a static system, the economy in equilibrium; so that the partial introduction of time and change inherently involves methodological inconsistency.[9]


The second aspect of the project of the General theory is that Keynes offers a defence of a very traditional idea: that you can deal with the problem of unemployment in a slump by using public works. By hiring more people in public works, you get more people employed overall, and then they are going to buy more food, housing, clothes … and that is going to push the economy forward. The marginalists claimed that this was impossible. On the assumptions of marginalist economics, the free-market economy is always running at equilibrium, meaning here the most efficient possible level. Hence if you put money into public works it is just going to divert money from private investment, so that what you gain on the swings with employment on public works, you lose on the roundabouts by the loss of employment in the private sector.

From there in turn, Keynes polemicises against the British ‘treasury view’ that government should always run a balanced budget. He says, no, you can run a deficit budget under slump conditions, when there is less than full employment, and by doing so you stimulate the economy and move it back into boom conditions. The flip side is that when you are in boom you should not run a deficit budget.

Unambiguously, Keynes’s project is also methodologically and politically nationalist. He is perfectly explicit that all his arguments rest on the assumption of a closed economy (pp11-12, 120, 264-65, 270). At the end of the book he relaxes that assumption. But he then says that if the states individually behave as though they are closed economies, aiming for full employment, then the reasons for conflicts between states will be mitigated, because states and capitals will no longer be searching for market share overseas (chapter 23).

In the preface to the 1936 German edition he says: “… the theory of output as a whole, which is what the following book purports to provide, is much more easily adapted to the conditions of a totalitarian state than is the theory of production or distribution of a given output produced under conditions of free competition and a large measure of laissez-faire” (pxix). Obviously that is intended to commend his book to German academic economists as of 1936, under Nazism.

Again, towards the end of the book he explicitly commends the mercantilist authors of the 18th century as against Smith, Ricardo, and so on (chapter 23). He says that at the end of the day the wealth of the nation depends on the full employment of capital and labour; and the full employment of capital and labour may well involve controls on the movements of capital and labour at the level of state borders.

In sum, the General theory displays a set of logical inconsistencies which are connected with Keynes’s attempt to save marginalism from itself, and with Keynes’s commitment to what is in substance a nationalist mercantilist policy.

Positive argument

There are two core elements of Keynes’s ideas. The first is the proposition that the money wage is ‘sticky downwards’: that is to say, workers do not like wage cuts and they tend to go on strike in response to attempts to impose them. But the real wage is not sticky downwards: that is, if the value of wages is reduced by inflation, the working class will not be able to organise effectively to resist, and may not even perceive that they are being screwed over by their employers. So step one is: Keynes says he agrees entirely with the marginalists that when there is a recession, real wages have to go down; but the way to achieve this may be to generate a bit of inflation, while leaving the money wage as it is (pp10, 14, 17, 232-33, 237, 270, 284).

Secondly, Keynes argues what is in the last analysis a variant of the old ‘underconsumptionist’ theory of capitalist crises, but gives it a different, psychological, explanation. This is essentially that, the better off you get, the lower the proportion of your income you are prepared to spend: consumption does not go up at the same rate as income. This produces a gap in consumption demand, particularly in relation to rentiers, the people who live off dividends and interest.

This gap has the consequence that, if there are more people in employment at lower levels of income, there will be a multiplier effect, because this consumption gap will be filled up (or partially filled up). Once full employment is achieved, if you carry on engaging in deficit spending, then you will produce genuine, serious and problematic inflation. But in the interim period, if the economy is not at full employment, then the multiplier effect means that relatively limited state spending will get more people in employment and that will increase demand by an amount which is more than or considerably more than the spending itself represents. (The Keynesians debated at length in the 1950s-70s what full employment is, and what the trade-off is between levels of employment and levels of inflation.)

This proposition is pretty clearly at a certain level straightforwardly true, and it is true on Marxist premisses as well. Of course, many Marxist economists object to ‘underconsumptionist’ theories of crisis and it is true that underconsumptionism cannot explain the cyclical return of crises. The cyclical return of crises started in Britain in 1760 and - as Engels (or possibly Marx) pointed out in the Anti-Dühring -the underconsumption of the broad masses has been present since the beginning of class society in at some point in the second millennium BCE.[10] The cyclical return of crises, in contrast, is a specific phenomenon of capitalism.

Nonetheless, there is a real gap in effective demand, and this demand gap has to be filled by capitalist investment. The objection to underconsumptionism is that in growth and boom periods it is filled by capitalist investment. Hence Keynes’s argument that the gaps could be filled in slump conditions by state investment is a partial truth. Nonetheless the underlying argument is logically incoherent. This is most obvious insofar as Keynes’s policy proposal is to increase effective demand among the relatively poor by increasing employment, while at the same time reducing effective demand among the relatively poor by reducing the real wage through inflation. Otherwise he would not assert, repeatedly, his agreement with the marginalists that in a recession the real wage has to fall.

The argument for the left putting forward Keynesian demands is essentially that after 1945 there were major reforms, at least ideologically represented by the ‘Keynesian consensus’, which had the effect that working class life was better (clearly when the left puts forward Keynesian demands it does not mean it is in favour of wage cuts through inflation). The same point in a different way was made by Werner Bonefeld in his talk at the ‘Fundamentals of political economy’ school: He said that ‘Don’t rely on the state’ is not a serious policy for the unemployed, etc, who are dependent on benefits. If ‘Don’t not rely on the state’ means taking away benefits, that is not just brutal, but also profoundly disadvantageous to everybody, because it tends to drive down wages.

Suppose, then, we agree that it is no good just saying, ‘We want revolution’, and we have to propose slogans for reform under capitalism. The real, underlying case for the particular form of slogans represented by Keynesian proposals is: ‘We know from the golden age of the 1950s-70s that Keynesianism can work’. The question is, do we? It is not at all clear that we do. I will discuss this question and the consequences in the second part of this article. Is a return to the ‘golden age’ feasible? And, if not, what sort of reform proposals should the workers’ movement propose?



1. ‘Genuine’, as opposed to those working under sham arrangements, whereby they are ‘self-employed’ to allow the employer various advantages: eg, Autoclenz v Belcher [2011] UKSC 41.

2. J Harrison Marxist economics for socialists (London 1978) develops this point.

3. Artisan production: the discussion in Capital Vol 1 has not been contradicted by subsequent evidence. Peasant farming: J Whittle The development of agrarian capitalism (Oxford 2000).

4. www.marxists.org/archive/kautsky/1892/erfurt/ch04.htm.

5. E van Ree, ‘Socialism in one country: a reassessment’ Studies in East European Thought No50, 1998, pp77-117 draws out the links between Kautsky’s conception and Stalin’s.

6. Royal Economic Society edition, Basingstoke 2007.

7. I do not mean by this to reject in principle any use of differential calculus in political economy. The point is merely that if you postulate that claims about political economy are validated by the applicability of calculus and hence must have a form which can be expressed in the use of calculus, you will be driven either to create a closed system, which manifestly has no relation to reality (marginalism), or to logical incoherence (Keynesianism). This is not a uniquely Marxist criticism of mainstream marginalism: it is, in fact, sared by the Austrian-school version of marginalism.

8. Keynes calls marginal economics ‘classical’ economics. This can lead to confusion. For Marx and hence the Marxist tradition ‘classical’ political economy meant Adam Smith, Ricardo and other authors who accepted versions of the labour theory of value; later authors who tried to find an inherent contribution to capital Marx called ‘vulgar’. For Keynes, who studied under Marshall, the labour theory of value was merely a mistake and the early marginalist authors - Jevons, Walras, Marshall - were ‘classical’.

9. Compare A Freeman, ‘The psychopathology of Walrasian Marxism’; and G Carchedi (ed) Marx and non-equilibrium economics Cheltenham 1996, chapter 1.

10. www.marxists.org/archive/marx/works/1877/anti-duhring/ch25.htm.