SUPPLEMENT: Imperialism and the state (Part III)

Capitalism produces huge wealth and huge inequalities, both of which play a role in recurring crises

(A PDF of this supplement can be found HERE)

Part I of this series introduced the issue: understanding imperialist dynamics requires a theorisation of the place of the state in capitalist political-economic dynamics. And it began my attempt at this analysis by considering ‘stateness’ as such, in abstraction from particular class forms - and from there, the state as a presupposition of the elementary forms of capitalist economy: commodity exchange and money.

Part II began by addressing the institutional forms which yield subordination of states to particular ruling classes - and, in particular, of capitalist states not merely to capital in general, but to particular capitals, especially through the early-capitalist transformation of warfare and arms production. It continued by developing the way in which one aspect of the means of this dependence - the debt-financing of capitalist states - originated in their revolutionary origins, by constituting the state as a joint-stock firm. That in turn constitutes the appearances of a ‘general rate of profit’, abstract capital and money as fructiferous.

In this part we push further into state responses to, and impact on, capitalist dynamics: in particular, the tendencies of capitalism to create radical polarisation of rich and poor, and recurring cycles of boom and bust, producing poverty in the midst of (and caused by) the production of material plenty. But a first step on this road is the last step from the previous part: understanding that state costs are analogous to corporate debt charges, rather than to the distribution of profits.

State costs

The costs of the state are a debt charge on the capitalist economy. Corporate dividends may be paid or not, depending on trading conditions. Interest on corporate debt, in contrast, has to be paid rain or shine. It should already be apparent from the previous sections that there is a core of state expenditure which has the same character as interest on corporate debt.

The absolute core is the cost of the maintenance of the state’s powers of internal coercion and tax-raising, without which nothing else in the state could exist. Next to this core, for the reasons already given, is keeping down interest and refinancing capital repayments on the state debt. Next in turn are the costs of maintaining the state’s relative position in the global hierarchy of states (ie, military expenditure going beyond pure ‘internal security’), since the state’s position in the global hierarchy both potentially impacts on the state’s existence, and more immediately impacts on its ability to raise money in capital markets and the interest it will be charged on loans.

‘Libertarians’ commonly believe that the only other fixed element of state costs is that of justice, police and prisons - the ‘nightwatchman’ state.1 The analysis in the previous article of the material requirements of state power under capitalism indicates that the fixed element is considerably wider. While - for example - expenditure on infrastructure can be cut in hard times, the effect is to diminish the state’s military capability relative to other states, as well as to handicap the broader economy. Or, for another example, later 19th and 20th century shifts towards welfarism in both Germany and Britain reflected abrupt realisations by state cores of the unhealthiness of urban working class recruits to the armed forces.2

State expenditures can be ‘downsized’ - just as companies can, when push comes to shove, restructure their debts. But the process is painful both to the internal political legitimacy of the state and to its global standing relative to other states - and in extremis to its existence as a state. State actors are thus motivated as far as possible to avoid or mitigate the necessity. There are two specific aspects of this, which are significant to the phenomenon of imperialism: the problems of polarisation and of cycles.


Polarisation can be dealt with relatively briefly. The polarisation of society into a small group of capitalist and rentier ultra-owners and a much larger group of proletarian non- or minimal owners is a very striking feature of capitalism in general (already remarked on before Marx) and the growth of polarisation has been one of the most striking features of the ‘deregulatory turn’, financial globalisation, etc since the 1980s.

A tendency to polarisation is, in fact, not unique to capitalism, though capitalist polarisation is much more rapid, extensive and thorough-going: both classical antiquity and feudalism displayed secular tendencies for ownership rights to become concentrated in small groups of families, with the converse relative impoverishment of the small proprietors.3 The underlying drive to polarisation comes from the institutions of heritable and alienable private property: a society with these institutions will naturally over time amplify small inequalities given by individual ability and health, and by luck and the fertility of land, resulting in the accumulation of inherited advantages in a minority, which partially or wholly exclude the majority. Capitalism is, in fact, in part a product of polarisation among peasants and artisans under feudalism; it is also able to tolerate much more extensive polarisation than prior forms of class society, because of the relative flexibility of wage labour, compared to household production or slavery.

Polarisation is a natural product of private property, but far more powerfully a natural product of capitalism. It is nonetheless a problem both for the capitalist class considered as a whole - as capitalists and as a ruling class - and for states. The problem for capital as capital is that severe polarisation in a closed economy implies constrained demand in ‘department 2’ (consumer goods), since, the smaller the elite, the less its ability to consume; and this in turn implies (later) constrained demand in ‘department 1’ (producer goods).

This ‘underconsumptionist’ point is not open to the usual objection that speculative investment in department 1 raises demand for labour and therefore demand in department 2 (which is the usual mechanism of recovery from crisis), since my point here presupposes that the tendency to polarisation is given completely free rein and runs to extremes, so that ownership of land and of money are monopolised in a very small number of families, with the result that speculative new investment in department 1 is obviously irrational for each of these families. However, since this point refers to a closed capitalist economy, it does not in itself imply a mechanical Zusammenbruch until (a) pre-capitalist forms of petty family production are wholly and on a global scale eliminated as an economic factor, and (b) trade unions are suppressed and all forms of state, etc, redistribution eliminated. In reality under these conditions the unemployed would starve to death more or less immediately, and this would render the labour market so tight that redistribution would occur through a rising wage share; and in any case political breakdown would occur long before this point, for the reasons given below. In a nation-state, however, the visible element in crises, of insufficient domestic demand (existing because profitability necessarily involves the production of a surplus over wages), does provide an obvious motivation for state efforts to increase exports from both department 1 and department 2.

The problem for capitalists considered as a ruling class is that extreme polarisation tends to undermine the legitimacy of the class system and of the inheritance of property, which - as we have seen - is already in tension with the objective approximate equality of human adults. Even if this does not result in rebellion and revolution, it will unavoidably result in enhanced levels of criminal victimisation, and so on.

For states the problem is much more immediate. In the first place, a state necessarily claims to represent the whole society, not simply as an enforcement agency for the ruling class. Hence, for the reason just given, polarisation tends to undermine political legitimacy well before it directly impacts on the legitimacy of private property and inheritance. Secondly, as has already been indicated, serious polarisation tends to undermine the military capability of the state, by reducing the pool of physically and educationally qualified recruits to the armed forces and civil bureaucracy. Thirdly, and perhaps most immediately, polarisation tends to reduce the relative tax take. On the one hand, it is hard to extract tax from an impoverished mass population; on the other, it is also hard to extract tax from the small group of the rich owners, who by virtue of their highly-concentrated wealth can deal almost as equals with the state. That is, they have access to legal, political and, if push really comes to shove, military means of resisting the extraction of tax or negotiating it down.

Private-property owning displays an objective dynamic towards social polarisation, and this dynamic is very much more forceful in capitalism. But the capitalist ruling class has (long-term, collective) interests in restraining polarisation, and so a fortiori does the state. The result is that - as far as they can - states will intervene to reduce polarisation within their own territories.

The extent and form of anti-polarisation measures are, obviously, affected by the direct intervention of the proletariat, in the form of trade union organisation, mass campaigns and political parties (in the ‘cold war’ period they were affected by the geopolitics of the presence of a bloc of states perceived as ‘socialist’). But anti-polarisation measures were in fact already created by early capitalist states in the effectively complete absence of permanent proletarian organisations and direct proletarian intervention in politics: for example, in the 16th-18th century Dutch Republic; and in the same period in the English Old Poor Law.4

This intervention can take a variety of forms: through direct redistribution (taxes and welfare), regulatory interventions in labour markets (Factories Acts, etc) and in capital markets (limited liability and other regulation of corporations to protect small investors; competition legislation, etc) and so on.5 All these forms, however, have costs to capital. The costs of direct redistribution are obvious. In relation to regulatory interventions, the costs arise because, if other states do not adopt similar regulatory interventions, capital operating in those states will be more able to externalise pollution, injury, etc costs onto labour and onto the petty bourgeoisie, and will hence obtain a price advantage over capital operating in the state which has adopted the measures.

These costs imply a clear and unambiguous interest in states and ‘their’ associated capitals in obtaining external sources of income and, conversely, externalising costs onto other states.


The question of cyclical crises is inherently more complex; and, though the proposition that cycles are a necessary feature of capitalism is a distinctive feature of Marxist political economy (and, in a slightly different form, of Schumpeterian theory), the causal mechanisms involved were not consistently stated by Marx and are violently debated among Marxist students of political economy.6 As with the average rate of profit, a lengthy ‘digression’ into this issue is required. The problem is that some theory of cyclical crisis is needed to grasp both state responses to the cycle, and the effects of these, which are significant both for the global hierarchy (colonialism, etc) and for geopolitical shifts; and the competing ‘classical Marxist’ and 1970s accounts of the cycle yield different probable outcomes.

The underlying empirical phenomenon is this. From c1760 a recurrent economic pattern appeared in Britain, in which a financial crash disrupted credit, producing a period of recession/depression in the material economy, followed by initially slow growth, which accelerated, slowed, displaced into a financial bubble, followed by a financial crash.7 The periodicity of the cycle and the depth of the crisis were quite variable, but the phenomenon and its ‘sigmoid sawtooth’ pattern (gradual and accelerating growth, slowdown in material growth and asset price bubble, crash) were stable, observable phenomena. By the mid-19th century similar phenomena were observable in the US, and by the later 19th century elsewhere, with some tendency to global synchronisation. The 1914-18 war and its aftermath sharply disrupted the periodicity of cycles and reduced international synchronisation, producing in the ‘Roaring 20s’ a period in which, in the US, claims were made that the cycle had been overcome, while recessionary features persisted in most of Europe; then in the aftermath of the 1929 crash a period in which permanent stagnation/depression could be imagined; then recovery of a sort through rearmament and war.

Post-1948 the trauma of 1914-45 and, in particular, the needs of the ‘cold war’, produced a period in which cycles, though they did not disappear, were managed by high levels of state intervention, so that the periodicity shortened and the ‘sawtooth’ form was replaced by what was called in Britain ‘stop-go’, as the state took forms of fiscal, monetary and regulatory action to cut off both the peak of the boom, before a full bubble developed, and to stimulate the economy before the nadir of the recession was reached. One result was the growth of a powerful ideology that the cycle could be managed by technical means and/or that it did not inherently involve financial crisis. Since the 1970s, however - whether because many of the technical means have been abandoned as involving too many concessions to labour at the expense of capital or because they ran up against real limits inherent in capitalism as such8 - the ‘classical’ pattern of sigmoid-sawtooth cycles involving financial crisis has tended to re-establish itself.

Pro-capitalist academic accounts of crises tend to project the phenomenon back into the remote past and explain it by some feature of ‘human nature’, obfuscating the peculiar regularity of cyclical crisis in capitalism.9 Among Marxists, on the other hand, theory commonly does not address the early regularisation of the business cycle in Britain in the later 18th century (it is taken for granted that the cycle starts with steam-driven industry in the early 19th). Nor has it yet addressed the recent shift back towards a more or less regular sigmoid-sawtooth pattern, routinely involving financial crisis which impacts on the ‘real economy’, as such. This is partly because both the deep political-military-economic crisis of 1914-45 and the peculiar features of the cold war period were taken as representing secular changes in capitalism: the first ‘imperialism, the highest stage’; the second ‘state monopoly capitalism’, ‘late capitalism’ or various other forms.

Marx passed comment on crises in a substantial number of different places, but never - since Capital remained unfinished - gave a fully systematised account. The ‘standard’ classical Marxist accounts of the cycle make it rest in the last analysis:

The polemics of the 1970s added the ‘profits squeeze’ thesis of Andrew Glyn, Bob Sutcliffe and others (drawn mainly from the British experience of the 1950s-70s), as well as considerable elaboration on the specific arguments.

All of these schools (including the widely rejected ‘profits squeeze’ approach) have some basis in comments by Marx, and all of them refer to real phenomena. But the polemics between the different schools have largely had the effect of undermining each of them.

This, in turn, has pushed theorists in three directions (not counterposed to each other). One is to emphasise the significance of ‘fictitious capital’: ie, movements in financial and credit markets.10 This moves the Marxist account towards ‘Austrian school’ accounts of the cycle, in which the cycle is produced simply by excessive growth of money and credit in the boom period, leading to the need for a crash.11 The second is to emphasise overinvestment in the boom period in ‘fixed capital’, especially land.12 This moves the Marxist account towards the ‘Georgist’13 account, in which the cycle is wholly driven by rent movements and land value speculation.14

The third direction is to emphasise the combination of the different features highlighted by the different schools among the ‘classical Marxists’ and the writers of the 1970s.15 By different routes, David Harvey, Hillel Ticktin, and the Regulation school of Michel Aglietta and others all remove the category ‘crisis’ from the short business cycle to more serious dislocations of the regime of exploitation. Though this approach is plausible in theories created in a period when short cycles were muted (in the imperialist centres) by state intervention, it loses the link of ‘crisis’ to its original meaning: the sudden catastrophic loss of confidence in the downward leg of the sawtooth form of the cycle. But the regular return of ‘crisis’ in this latter sense is precisely what must be explained as special to capitalism, since occasional major dislocations of the regime of exploitation, caused by internal contradictions, are found in all forms of class society;16 and, as ‘cold war’ regimes were gradually dismantled, so the regular return of cyclical ‘crisis’ in the finance-triggered and sawtooth form has reasserted itself.

Pre-1914 theories of crisis were linked to the Zusammenbruchtheorie of a general breakdown of capitalism, which would trigger revolution.17 So, in a different way, were inter-war Bolshevik and other (left) Marxist accounts.18 The consequence is that these theories were designed to generate not just recurrent crisis, but deepening recurrent crisis ending in Zusammenbruch. In my opinion the Zusammenbruchtheorie involves a misconception of the form of decline of particular class orders, which misses out the role of the state as ‘substituting’ for the economic role of a declining ruling class - found in later antiquity and later feudalism as well as visible in 20th-21st century capitalism. As a result, it seems to me that what should be expected to happen in the long term, as contradictions accumulate, is not Zusammenbruch, but increasing statisation.19 The debate in the 1970s was animated by debate over the ‘official communist’ and left-Labour, semi-Keynesian ‘Alternative Economic Strategy’ and - arising from this - the problem for the ‘orthodox Marxists’ of explaining the degree of control of cyclical crisis in the imperialist countries in the cold war, ‘long boom’ or ‘golden age’ period in face of semi-Keynesian and Sraffian arguments.

Both these contexts help, I think, to explain a fundamental feature of the discussion. This is that it has assumed that, because boom and crisis have effects at the level of national (or global) economic aggregates, their causes must lie at the level of national (or global) economic aggregates. In a sense, this idea is already present in Marx - since, in the first place, the reproduction schemas of Capital volume 2 are already large aggregates and, secondly, if there was an actual general rate of profit formed through a ‘transformation’ through the price mechanism, as is argued in Capital volume 3, only aggregate causes could create aggregate effects.

The latter problem is exacerbated by the fact that (as the ‘temporal single-system interpretation’ school points out) von Bortkiewicz’s ‘correction’ to Marx’s transformation, widely accepted by Marxist economists, tacitly imports Say’s Law (which Marx explicitly rejected).20 Even Harvey sees crisis as producing a return to ‘equilibrium’,21 where the reality is that capitalist growth and boom is just as much a ‘far from equilibrium’ condition as crisis and slump. The assumption of the real possibility of equilibrium outcomes (even of Keynesian multiple equilibria) more or less inevitably produces the result of orthodox bourgeois economics: that the business cycle is caused by ‘external’ ‘shocks’. But even if these ‘external’ ‘shocks’ are somehow internalised, they must have aggregate causes.

However, Xavier Gabaix has demonstrated that even in a neo-classical framework it is possible to generate crisis from the ‘granular’ effects on individual firms, if (as is in fact the case) the distribution of firm size in the economy studied follows a power law. That is, a ‘shock’ producing a boom in a particular large firm or sector generates effects across the economy as a whole; and a ‘counter-shock’ knocking down this firm or sector will precipitate a downswing with effects across the economy as a whole.22

The case is considerably stronger for Marxist theory, if (as I argued in Part II that we must) we discard the actual formation of a general rate of profit through price mechanisms, and if we do not illicitly import Say’s Law or equilibrium solutions. In the first place, the neoclassicals’ assumption that downward inflexibility of wages is an irrationality which exacerbates market ‘corrections’ is mere ideology, since there is a real subsistence floor to wages.23 The result is that wage costs form a real floor to prices: a firm can keep running at a loss by drawing on capital, but not if it cannot pay the subsistence costs of its workers.24

Secondly, commodities must be use-values as well as exchange-values. Hence, even if aggregate demand completely abstracting from use-values could be unlimited (which is, of course, questionable), there are real limits given by population on the amount of demand for, say, rail journeys from London to Birmingham, ready meals, mobile phones or any other single commodity. If the price of a commodity falls very dramatically relative to wages and other commodities, there will be a very large rise in demand for that commodity; but it will still be subject to natural limits. There can, therefore, be real sectoral overproduction, which is not caused by ‘underconsumption’, but simply by overaccumulation/overinvestment in the sector.

Thirdly, the quantity of land - access to which is essential to all production processes - is subject to natural limits. The quantity of money is not subject to natural limits in the same sense, but in another: that if the quantity of money is not limited, it will not function as a store of value and, if it will not function as a store of value, neither will it function as a means of exchange (as is seen in episodes of hyperinflation). Hence money is not a mere numéraire, as it is in marginalist accounts (and the marginalist idea that there is an irrationality called “money illusion” is - again - merely ideology).

The net consequence is that there is more reason under elementary Marxist assumptions than under marginalist assumptions to expect ‘granular’ causes - overproduction and/or a falling rate of profit in a large firm or sector - to have aggregate cyclical effects. Of course, to say so does not, as yet, escape from the external quality of the neoclassical writers’ ‘shocks’. Like Harvey’s unranked range of obstacles to capitalist growth, the point in itself does not predict the regular return of crisis, which has been the distinctive feature of capitalism since the 1760s.

To approach this issue we need to distinguish conditions for the existence of the cycle (causae sine qua non) from the motor of the cycle (causa causans). The clearest example of conditions for the existence of the cycle is underconsumption. If consumer demand was unlimited, aggregate growth could be uninterrupted and crisis would therefore have to be exogenous, so limited demand in department 2 is a necessary condition of crisis. But - as Engels points out in the Anti-Dühring - the restricted consumption of the masses is present in pre-capitalist as well as capitalist societies.25 So this cannot be the motor of the regular cycle which first appears in capitalist society. The same is true of land rent and land value speculation, which forms the basis of the Georgist theory: if the quantity of land was unlimited, or the land was everywhere nationalised, but capitalism remained, one of the usual aspects of the cycle would disappear (or, more probably, be shifted onto speculation in other forms of fixed capital); but, again, land value speculation undoubtedly long precedes in history the modern development of a cyclical return of crisis.

The case of credit money and the financial markets is slightly different. Limited-scale customary credit money certainly antedates capitalism, but central financial markets and the accompanying highly elaborated financial intermediation do not. The form of the crisis phase of the cycle is, indeed, a crisis of credit and financial markets, and the regular cycle does not and cannot appear until these markets are fully in place. Moreover, severe ‘financial repression’ (state monopoly of finance) can displace the domestic component of the cycle, and under the moderate financial repression of the ‘Bretton Woods’ regime the role of financial crashes in the cycle was sharply attenuated. The developed financial system is therefore not merely a background sine qua non condition to the cycle, like underconsumption, but is intimately involved in the process.

But under the Bretton Woods regime the cycle as such did not disappear, but took an ‘artificial’ form, as governments induced recessions early to prevent the bubble phase appearing.26 Moreover, irregular crises in financial markets antedate the establishment of the regular cycle.27 It is therefore probably still correct to regard financial markets and the (over-) expansion of credit money as a condition of the cycle rather than as its underlying motor.

The underlying motor of the cyclical movement is the search of investors for average or above-average profits, within the framework of an economy which (a) has already acquired a capitalist state and, therefore, developed financial markets, so that there is a transmission belt which generalises sectoral difficulties; and which (b) has gone beyond a merely interstitial role in a predominantly pre-capitalist immediate environment, so that ‘rational choice’ profit-maximising has become dominant among investment motives.28 Its form at the next level up is movement of the sectoral rate of profit (in the lead sector or group of sectors, in the sense of ‘sector’ discussed above).

Now a ‘lead sector’ may seem to be the result of an ‘external shock’ or Schumpeterian innovation. But this is misconceived, because it presupposes a priori both (a) an equilibrium starting point and (b) universal capitalism in all sectors of the global (or closed national) economy. The equilibrium starting point is merely Say’s Law.29 The supposition of universal capitalism is obviously false even today and was a fortiori false when the regular cycle began. On the contrary, capitalism came into the world sector by sector and on an international scale, starting with shipping, and continues to displace pre-capitalist forms of production sector by sector and on an international scale.

A lead sector is thus one which either extends the reach of capitalism at the expense of prior forms of production (as, for example, shipping or capitalist agriculture) or in which a sub-sector or individual industry or firm employs new technology to invade a sector conducted under a prior form of capitalist production (as, for example, railways in relation to canals and horse-drawn road carriers, or steamships in relation to sailing ships).

Within sectors, as opposed to across the economy as a whole, a general rate of profit is formed by price competition. The newness of the new sector or industry therefore initially yields a high rate of profit, because the average socially necessary labour time in the sector includes large amounts of lower-productivity production. The result is a further inflow of capital into the new sector. The resulting capital investments pull the rest of the economy upwards towards boom. Firms in the sector with a low organic composition of capital are either forced to copy those with a higher organic composition of capital (OCC), are bankrupted, or are driven into niche markets. The result is a sector with a (broadly) uniform, high OCC. As long as the sector remains internally competitive, the sectoral rate of profit in the lead sector now falls.

Contrary to both neo-classical assumptions and those of the ‘transformation procedure’, capital cannot move smoothly out of the lead sector. At a superficial level, the reason for this is that the fixed capital involved (land and machinery) has been paid for and, whether this has been done by incurring debt (either in the form of debt stricto sensu or equity) or by reinvestment of retained profits, there is a necessary expectation that it will retain its value and make a return. To decapitalise therefore involves the proprietors and creditors accepting a capital loss.30

The consequence therefore is that individual firms in the lead sector necessarily attempt to overcome the fall in the rate of profit in the first place by an intensified struggle for market share. One aspect of this is cost-cutting (which creates the ‘real appearance’ that supports the wage-push theory of crisis). But, since this is part of a struggle for market share, the gains made are ‘cashed’ in price competition more than in a rise in absolute returns; the net effect is therefore merely to reduce the wage share and therefore effective demand in department 2 (which creates the ‘real appearance’ that supports the underconsumptionist theories of crisis). The other and more fundamental aspect is centralisation of capital through takeovers and mergers.

It is the more fundamental aspect because there is another obstacle, below the level of capital losses to investors, which obstructs the movement of capital out of the former lead sector. The high OCC in the lead sector represents as its converse a fundamental improvement of the productivity of labour in this sector, and the rest of the economy reorganises itself on the assumption that this level of productivity will continue. For example, for the UK to have returned from capitalist agriculture to peasant agriculture when the capitalist agriculture sector ceased to be profitable in the later 19th-early 20th century would have entailed large-scale reallocation of labour from other sectors to agriculture. The same is true mutatis mutandis of other productive sectors.

In a sense, the underlying contradiction is this. Within sectors the mechanism of price competition, allocating prices in the last analysis on the basis of average socially-necessary labour time, can allocate profits to the producers with the highest productivity of labour, but between sectors it cannot not do so, because more socially-necessary labour is being performed in lower-productivity sectors: the point made above in relation to the issue of the ‘general rate of profit’. The nearest approach to a cross-sectoral ‘general rate of profit’ is the rate of return on government securities, which expresses itself to the productive firm as rent, interest and tax burdens on profits. If the firms in the lead sector are not to be crushed by these burdens, a non-market redistribution of the social surplus product in favour of this sector is necessary. This task can be performed by monopolisation and cartelisation; or by state subsidy; or both. In seeking to increase market share, especially by acquisitions, firms are thus instinctively seeking monopoly arrangements.

The closer the sector gets to monopolisation, the more the allocation of prices, capital investments and returns to the sector become conscious political decisions, rather than market decisions. In the case of utilities and public transport facilities this is obvious. But it is also present when - for example - British Leyland or General Motors represented themselves to the relevant states and their ‘public opinion’ first as “national economic champion” and then - as the economic position of each firm worsened - as a provider of scarce employment, which therefore could not be allowed to fail. There is a tendency for such sectors/firms to be starved of new capital investments, because the monopoly charge is perceived by other economic actors not as socially necessary production, but an aspect of the ‘tax burden’: again, obvious in British and US utilities, but also seen in private near-monopolies like US Steel or British Leyland.31

To actually reach this outcome takes more than a single cycle. In the absence of its arrival, the capital values of firms in the lead sector or sectors have to fall. This is true even where the lead sector has not reached its global limits, as long as it has reached a limit - given, for example, by obstacles created by states. The crisis phase is the moment at which this necessary loss is accepted by investors.

But this phase is preceded by a bubble phase, whether the bubble is in financial market assets, land or both. The reason for the bubble phase is that, where the firms in the lead sector or sectors have reached the point that they are struggling for market share in a high-OCC environment, the demand of this sector or sectors for investment goods falls. But a large part of the social surplus product has to be invested. Until the capital loss in the lead sector or sectors has been generally accepted by investors, no new lead productive sector can emerge to fill this investment demand gap. The share of investments in speculative activity in financial assets and land therefore necessarily increases, producing the bubble phase.

The crisis phase therefore takes the form of a collapse of the speculative asset bubble. This, in turn, has general effects on the economy beyond the bubble itself and beyond the lead sector. This is because capitalism requires credit money, and credit money is itself a speculative asset (a form of ‘fictitious capital’) and necessarily intertwined with capital markets. The bursting of the bubble therefore necessarily produces a general shortage of credit affecting all economic actors.

The crash-slump phase itself tends to produce centralisation of capital and polarisation. The reason is that the general shortage of credit favours firms which momentarily do have access to liquidity buying up or buying out other firms. It is important to be clear that this is not a matter of the most efficient or highest-productivity producer defeating less efficient or lower-productivity producers (a process which does occur in the growth-boom phase). The question of who has access to liquidity at the moment of crash is completely arbitrary: a low-productivity firm may happen to be sitting on a pile of cash or have good government connections at the moment of the crisis, while a high-productivity firm may be ‘caught short’ after taking on large debt commitments to fund new investment. The effects of the crash and the ensuing credit shortage also tend to fall heavily on the middle classes and small rentiers. The loss of savings value in the crash, and bankruptcies due to the shortage of small business working credit, tend to push sections of the middle classes into the proletariat.

To the extent that losses in the crash phase actually fall on investor interests, the crash/slump phase ‘clears’ both the bubble values and a part of the ‘overinvestment’ in the lead sector. By doing so it creates space for a renewed phase of (initially gradual) expansion in this or another lead sector. If, however, investors are to be protected from losses in the absence of full monopolisation, this can only happen through state rescue, providing a diversion of resources to investors from other activities and transfer to the particular capitals which are in difficulties.

State intervention may take very diverse forms: direct subsidy or bailout; tariff protection; non-tariff protection through ‘quality’ and other regulations; limited liability; printing more money; forcible action against other states to open up new markets; the use of secret services for industrial espionage; industrial sabotage, or ‘exposures’ designed to discredit foreign competitors; and so on. But every form is for the benefit of some particular capitals and at the expense of others.32

We have now, therefore, necessarily returned to the relation of the cycle to the state, in the form of an interest of particular capitals in state rescue. But this interest is not that of capital in general - except in the very limited sense of state action to keep the financial system afloat in the immediate crash phase, when there is a risk of a complete collapse of money. The reason for this limit is that every rescue of particular capitals is at the end of the day at the expense of other capitals. Why is there a more extensive interest of the state in bailing out particular capitals?

We have, in fact, already seen an accumulation of reasons in the course of the argument so far. First. the capitalist state is dependent through state borrowing on its larger particular creditors, who are, of course, investors not just in the state, but also in other capitalist operations and would therefore bear losses if these are allowed to collapse. Second, it is also dependent through forms of corruption on particular capitals in proportion to their ability to bribe public officials (whether this bribery takes direct forms or those of paying campaign contributions, backing media through advertising, and so on). Given that it is the lead sector in which the problem turning the cycle from boom to bust develops, the firms involved are a priori likely to be large, with a correspondingly large ability to bribe officials.

Third, a state is in the last analysis a military actor, and its military capability under the conditions of post-medieval warfare is dependent on the continuation of those productive activities within its territory that directly or indirectly bear on military capability and/or bear on the ability to stand a siege. However, capitalism is from the outset international (passim above) and the logic of the cycle, as tending to the monopolisation of particular sectors, tends to the global monopolisation of these sectors. This, in turn, implies that, if the global monopolist is headquartered outside the particular nation-state and organises an international material division of labour, the state will lose its independent productive capacity in the sector, with adverse consequences for direct military capacity and/or ability to stand siege.

Fourth. the crash and slump phase of the business cycle necessarily involves a loss of revenue to the state (everyone is by now familiar with this from media talk about deficits). But the state has much more difficulty in reducing its expenditures, which are in large part analogous to the fixed debt of a capitalist firm. Margaret Thatcher’s failure to reduce public expenditure and the high level of doubts and wrangling in the British state over the Con-Dem coalition’s cuts in 2010-15 illustrate the point. A related point is that, fifth, the capitalist class has an interest qua ruling class, and the state has an interest qua state, in restraining polarisation. But the crash and slump phases of the cycle tend to increase polarisation.

The net effect is that each capitalist state has a clear interest qua capitalist state in intervening to support ‘its’ capitals, when they are in difficulty as a result of the business cycle. But the state interest in intervening cannot stop there. In the first place, as we have already seen, every intervention is directly or indirectly at the expense of other capitals. This will naturally create a demand that if the state is to support firm A or sector X it should also support firm B or sector Y. Secondly, if (to take a 17th-18th century example) the Dutch Republic intervenes to support its ‘overcapitalised’ herring industry, the British state will be pressed by local capitals also to intervene to support its - as yet nascent - herring industry.

State intervention is thus not exogenous to market processes (the neoliberal view) or the result of an unexplained “territorial logic of power” (Giovanni Arrighi) or “accumulation by dispossession” (Harvey). It is endogenous: the product of a relation between particular capitals and ‘their’ states, which is necessary to capitalism as an economic form.

State intervention

Given what has already been said, the effects of state intervention can be treated quite summarily. This intervention is directed against the ‘negative side’ (from the point of view of the capitalist quasi-nation-state) of capitalist dynamics: the tendencies to polarisation and to recurrent crises, leading to global-scale monopolisation. To the extent to which state intervention is successful, it will slow down these dynamics within the state’s home territory33 - an important limitation. Hence, the absence of the integration of the role of the state in the political economy helps to account for the notorious tendencies of Marxists to over-predict both crises and social polarisation in the imperialist centres, which have been exploited by the right wing of the workers’ movement since Eduard Bernstein to deny that the tendencies to polarisation and recurrent crisis exist at all.

The converse, however, is the ‘positive side’ of capitalist dynamics. This is the tendency of capitalist development to revolutionise the forces of production - or, to put it another way, to increase the productivity of labour. State intervention against crisis and polarisation necessarily blunts the incentives in the capitalist order which produce this dynamic - again, within the state’s home territory. Counter-polarisation measures increase the wage share, whether directly or by supporting forms of petty proprietorship, or by reducing competition in the labour market. They also tend to reduce the savings rate. Forms of subsidy and protection reduce the bankruptcy risk and, on the other side, are always at the expense of innovators. The net result is that, the more successful the state interventions, the more capitalists’ investments in the state’s home territory will be safer (because there is both less bankruptcy risk and less risk of either political or criminal takings arising from polarisation); but they will also be lower-return. Older technologies will be preserved and persist: water transport versus railways in the early 19th-century Netherlands, old technology and relatively small-scale organisational forms in the British textiles and engineering industries in the late 19th-20th centuries, and so on.

This result has an important further implication: it will tend to produce inward investment from investors in countries whose home-country regime is higher-risk; and outward investment from investors in the country with higher levels of state intervention. This was visible in the 18th century Netherlands, the later 19th-early 20th century UK and more recently the US.

At the same time, as I have already observed, both counter-cyclical and counter-polarisation measures, as well as protection for ‘defence-related industries’, are necessarily at the expense of other particular capitals. It is rational for both states and the particular capitals connected to them to externalise these costs as far as possible on capitals connected to other states. State motivations for intervention in the capitalist market therefore inherently produce some level of geopolitical competition between states - this was already true in the late medieval Italian city-states.


  1. Some libertarians, indeed, would rule out the third head (costs of maintaining relative position in the global hierarchy) on the basis that these could be saved by an unambitious foreign policy: eg, I Eland The empire has no clothes Washington 2004.↩︎

  2. For Germany, see N Dörr, L Grawe and H Obinger, ‘The military origins of labor protection legislation in imperial Germany’ Historical Social Research/Historische Sozialforschung Vol 45, pp27-67. For Britain, see Martin Pugh State and society London 1994, pp123‑24, citing the 1904 Report of the Inter-Departmental Committee on Physical Deterioration; L Margaret Barnett, ‘Diet and nutrition’ in FM Leventhal (ed) Twentieth-century Britain: an encyclopedia New York 1995, p226.↩︎

  3. For classical antiquity see GEM de Ste Croix The class struggle in the ancient Greek world New York 1981 (the burden of the analysis of the whole book). On feudalism, W Kula in Economic theory of the feudal system (London 1976) gives arguments to expect a tendency to polarisation; while CT Bekar and CG Reed in ‘Land markets and inequality’ European Review of Economic History Vol 17 (2013), pp294-317 explore the particular case of the medieval English peasantry.↩︎

  4. See J de Vries and A van der Woude The first modern economy Cambridge 1997, pp653-64; L Charlesworth Welfare’s forgotten past London 2009. State welfarist measures were also present in the Italian city-states: see, for example, B Pullan, ‘Support and redeem: charity and poor relief in Italian cities from the fourteenth to the seventeenth century’ Continuity and Change Vol 3 (1998), pp177-208.↩︎

  5. In pre-capitalist societies the dominant form of redistributive arrangements tends to be religious, and religious redistribution persists in capitalism, though it tends to be dwarfed by state operations. But my present concern is with state measures against polarisation, as affecting the interaction of state and markets and inter-state relations.↩︎

  6. There is far too much literature for full review or citation here. A convenient - albeit not fully up-to-date, and inevitably contentious - summary of the disputes is in Simon Clarke’s Marx’s theory of crisis (London 1993).↩︎

  7. From c1760: J Hoppit, ‘Financial crises in eighteenth-century England’ The Economic History Review No39 (1986), pp39-58, draws the necessary distinction between financial panics, driven by wars and similar events, and cyclical crises; this contains data for the point at which the cycle becomes regular. Since this point is prior to the ‘industrial revolution’, it seems most plausible to explain it by the fact that, after the Seven Years War, British global ascendancy and the impossibility of a general restoration of feudal social order were both clear, with the result that British financial markets (in the broadest sense: ie, including bill-discounting, etc, not just organised stock and commodity markets) attracted a sufficient share of total savings to create both the investment conditions for the ‘industrial revolution’ and the conditions for the regular cycle.↩︎

  8. It is unnecessary at this stage of the argument (and may be absolutely unnecessary) to decide between these two lines of argument.↩︎

  9. Eg, CM Reinhart and K Rogoff in This time is different: eight centuries of financial folly (Princeton 2009) achieve the result by treating state default as primary, which allows medieval royal defaults to be incorporated in the analysis. GA Akerlof and RJ Shiller in Animal spirits (Princeton 2009) proceed directly to psychological explanations. At least these works avoid the feature, common to a lot of rightwing web commentary (not worth citing), of including as evidence for the explanation of modern crises such poorly attested events as the reported financial crisis of AD33 (see T Frank, ‘The financial crisis of 33 AD’ American Journal of Philology Vol 56, pp336-41) or the Roman currency debasement of the 3rd century AD: P-I Prodromidis, ‘Another view of an old inflation’ (2006): www.kepe.gr/pdf/D.P/dp_85.pdf. The abstracts of papers by Colin Elliott on indiana.academia.edu/ColinElliott/Articles, indicate some of the difficulties in using the 3rd century debasement as evidence for modern economic arguments.↩︎

  10. Eg, M Perelman Marx’s crises theory: scarcity, labour and finance (New York 1987); L Goldner, various papers: home.earthlink.net/~lrgoldner; N Potts ‘Surplus capital: the ultimate cause of the crisis?’ (2010) Critique No35‑49.↩︎

  11. There is a clear overview in RW Garrison, ‘The Austrian theory of the business cycle in the light of modern Macroeconomics’ Review of Austrian Economics No3, 1990.↩︎

  12. This is particularly developed by David Harvey in The limits to capital London 2006 (and later editions).↩︎

  13. Referring to Henry George (1839-97).↩︎

  14. See George’s Progress and poverty (1894) Boston 2006 (book 4, chapter 4 and book 5, chapter 1). There is further discussion of the crisis theory in several of the essays in RV Andelson (ed) Critics of Henry George (Oxford 2004). See also FE Foldvary ‘The business cycle: a Georgist-Austrian synthesis’ The American Journal of Economics and Sociology Vol 56 (1997) pp521-41. This has the distinction of having predicted in 1997 that “the next major bust, 18 years after the 1990 downturn, will be around 2008” (p538).↩︎

  15. Most clearly in D Harvey The limits to capital (London 2006), chapters 7, 10 and 13.↩︎

  16. Cf the premodern ‘crises’ referred to in note 9.↩︎

  17. This is quite strongly visible in some of the texts collected in RB Day and D Gaido (ed) Discovering imperialism (Leiden 2012).↩︎

  18. See RB Day The ‘crisis’ and the ‘crash’ New York 1981; and R Kuhn Henryk Grossman and the recovery of Marxism Urbana 2012.↩︎

  19. As I argued in Part I. See also the outline sketch in ‘Marxism and freedom of communication’ Critique No37 (2009), pp571-74.↩︎

  20. Eg, A Freeman ‘Marx without equilibrium’: mpra.ub.uni-muenchen.de/1207/1/MPRA_paper_1207.pdf. There is more detail and a different point of view in AB Trigg ‘Marx, Say’s law and commodity money’ Contributions to Political Economy Vol 39 (2020), pp23‑41.↩︎

  21. D Harvey The limits to capital London 2006.↩︎

  22. See ‘The granular origins of aggregate fluctuations’ (2009): papers.ssrn.com/sol3/papers.cfm?abstract_id=1111765.↩︎

  23. The same point is made in a different way by RC Allen in Enclosure and the Yeoman Oxford 1992, chapters 11-14: even under very strong pressure from mass unemployment, the wage paid to employed workers must pay for enough calories, etc for them to be able to do the physical work involved in their jobs.↩︎

  24. This is, of course, subject to (a) state subsidies to low-wage capitals in the form of welfare payments to employees; and (b) the possibility of not paying full subsistence costs, where labour-power is primarily produced in the peasant sector. Chinese and other employers who use migration controls to avoid paying wages do, in fact, pay workers’ day-to-day subsistence costs in the form of housing and canteens. But they fraudulently misrepresent at the time of hiring that the wages on offer are above subsistence costs, and externalise the costs of reproduction of the labour force.↩︎

  25. MECW Vol 25, p272.↩︎

  26. CM Reinhart and K Rogoff This time is different: eight centuries of financial folly Princeton 2009.↩︎

  27. As well as the Dutch tulip bubble in the 1630s, and South Sea and Mississippi Company bubbles around 1720, Julian Hoppit, in ‘Financial crises in eighteenth-century England’ (The Economic History Review No39, 1986, pp39-58) points to a number of 18th century British financial crises triggered by war events and not resulting in generalised recession. Cf also RC Mueller The Venetian money market: Banks, panics and the public debt, 1200-1500 (Money and banking in medieval and renaissance Venice Vol 2) Baltimore 1997, chapters 4-6.↩︎

  28. By “gone beyond a merely interstitial role” I am referring to the fact that it first appeared in Britain, not the Italian city-states or the Netherlands, and to the significance of 1763; “immediate environment” refers to western Europe and the ‘Atlantic’ economy, in spite of the fact that globally capitalism was very far from dominant. Cf Marx in the 1872-75 French edition of Capital Vol 1, cited by Kevin Anderson in ‘The “unknown” Marx’s Capital, Vol 1’ Review of Radical Political Economics No15 (1983), pp73-74.↩︎

  29. The point is made by Shigeto Tsuru in ‘Business cycle and capitalism; Schumpeter versus Marx’ Annals of the Hitotsubashi Academy (1952), pp134-47.↩︎

  30. This aspect of the process is helpfully fixed on by ‘temporal single-system interpretation’ school writers’ insistence that inputs should be accounted at historical costs.↩︎

  31. See M Perelman Marx’s crises theory: scarcity, labour and finance (New York 1987) on US steel; see also the Ryder Report British Leyland: the next decade London 1975 (available free from the National Archive Online at nationalarchives.gov.uk).↩︎

  32. A bailout by simply printing more money on a globally generalised scale, leading to a general inflation, still (temporarily) transfers resources to the loss-makers, redistributing such losses to those who are for some reason not in a position to increase their prices. Printing money at nation-state level drives the local currency down, redistributing resources towards exporters and away from importers, and externalising the losses to the exporters of countries whose currency goes up. If applied by a state which is not the sponsor of the global reserve currency it also risks either tipping into hyperinflation or leading to state default. In neither case does printing more money avoid the redistributive character of state bailouts as a response to crisis. If the wage share is above subsistence levels, it can be driven down. But, although in a sense this reduces the burden on the capitals not bailed out, it still involves costs to a particular group of capitals - in this case those in department 2 experience a narrowed market.↩︎

  33. ‘Home territory’, as opposed to colonies and semi-colonies. The picture in colonies is more complex: for example, the British in India intervened to support peasant proprietorship (and traditional landlordism) against market imperatives in agriculture. But India was not just any colony, but the military keystone of British imperial operations in Africa and Asia. See J Darwin The empire project; the rise and fall of the British world-system, 1830-1970 Cambridge 2011.↩︎