It is common to counter this analysis with the idea that the concept of global capital is a pure abstraction, without any real existence, which cannot be used in an argument concerning the problem of the profitability of any kind of production. According to this argument, each capitalist or each nation worries very little about ‘global capital'. It would therefore be absurd to attempt to define an economic totality on the basis of a sum of antagonistic entities.
The fact that capital, as a global phenomenon, lives only in a divided form does not imply that it does not exist at the global level. The totality of the thieves in one city live in constant competition with each other and the law of their environment is none other than that of the strongest thief. This in no way changes the fact that this totality exists in itself, and that it has its own interests (to extend the example, with regard to the police). The fact that it cannot exist for itself, that is with a collective and unified consciousness of its interests and acting in relation to those interests in no way changes the problem. Global capital is always a sum of antagonistic capitals. But it exists nonetheless with general laws which act exclusively at its own level, with its own phenomena (world war, world crises) which impose themselves on each of its fractions and over which no fraction of capital has a real control.
Under capitalism the domination of capital extended itself to the whole of the planet a long time age. Every commodity today contains labour and primary materials from every corner of the world. In such a state of affairs, it is first of all the reality of global capital which determines the reality of each of its parts, and not the other way around.
Massive military production is a phenomenon which, because of its origin, its development and its effects, concerns the world capitalist economy. To attempt to judge its nature outside the concept of global capital is quite simply a renunciation of any analysis.
In order to analyse the problem of the reproduction of capital, Marx does not hesitate to put himself on this terrain:
"In order to examine the object of our investigation in its integrity, free from all disturbing secondary circumstances, we must treat the whole world as one nation and assume that capitalist production is everywhere established and has possessed itself of every branch of the economy." (Marx, Capital, Book 1, p.581, London 1970)
This has nothing to do with any so-called theory of ultra-imperialism. It is a definition of the only objective terrain on which the fundamental phenomena of developed capitalism can be understood.
Certain individuals nonetheless accept the concept of global capital (a concept which can be applied both at the world level and at the level of national capital), without accepting the possibility of production which is simultaneously profitable for one capitalist or a fraction of global capital but not profitable for global capital as a whole. This is an objection which might be encapsulated by the American witticism: ‘What's good for General Motors is good for the United States'. It is thus imagined that if military production is productive for one capitalist, it is also productive for global capital.
Just as in the case of the first objection, here again the contradictory reality of capital is being ignored. The fundamental contradiction of the capitalist system, that which underlies all its contradictions, is that which opposes the ever-increasingly universal, socialised character of the production process, to its private, fragmented, property relations, on the basis of which the process takes place. In the first example, only the fragmented, divided aspect of capital is taken into consideration, while the global nature of capital is ignored. In this case, it is the inevitable antagonisms between capitals which is shunted aside, giving way to a harmonious vision of global capital without contradictions between its parts or between its parts and the whole. The conditions in which a particular capitalist realises his profits is obviously tied in the long run to those of capital as a whole. But things can develop in such a way that this tie becomes temporarily attenuated, to the point of making ‘what is bad' for the United States appear as ‘what is good for General Motors'! ( The classic example is that of the capitalist who produces consumer goods necessary for the working class: each general wage increase is for him, a supplementary condition for the realisation of a market, whereas it constitutes a threat to the rate of profit of capital as a whole.)
The inevitable disharmony which reigns in the private property system can only increase with the decline of the system. Thus, this phase gives rise to the development of state intervention with the aim of containing, through forced centralisation, the ever-increasing tendencies of disintegration.
There is then nothing surprising in the fact that the same criterion of profitability in the capitalist system gives different results when it is applied to a particular capitalist than when it is applied to global capital.
We must therefore explain why and how it is that the state, as the representative of national capital, maintains the capitalist sellers of armaments, while the production of armaments is ‘unproductive' for the national capital (leaving aside for the moment the possibility of transferring the costs to another state). And, first of all, it is necessary to define in a precise manner the marxist criterion of ‘productive' and ‘unproductive' labour.
Productive Labour and Unproductive Labour
Marx's answer can be summarised by the famous formula: "Labour which directly creates surplus value, that is to say, which expands capital, is productive." 
This formulation is specific to the capitalist mode of production. It is distinct from that which applies to ‘the question of productive labour in general', when we study: "the labour process under its most simple aspect, common to all its historical forms, as a process between man and nature." 
From this general viewpoint, historically, all labour is productive from the moment that it ends up in a product and that product corresponds to any human need. "In the labour process, therefore, man's activity, with the help of the instruments of labour, effects an alteration, designed from the commencement, in the material worked upon. This process disappears in the product; the latter is a use-value, nature's material adapted by a change of form to the wants of man. Labour has incorporated itself with its subject: the former is materialised, the latter transformed. That which in the labourer appeared as movement, now appears in the product as a fixed quality without motion. The blacksmith forges and the product is forging.
If we examine the whole process from the point of view of its result, the product, it is plain that both the instruments and the subject of labour, are means of production, and that the labour itself is productive labour." 
Marx further explains: "This method of determining, from the standpoint of the labour-process alone, what is productive labour, is by no means directly applicable to the case of the capitalist process of production". (Marx, Ibid)
The productive process takes on, under capitalism, specific forms which no longer permit us to content ourselves with this first, almost tautological formulation. There are two reasons for this. The first appears when we look at capital from the viewpoint of labour. In capitalism, labour power is purchased by capital, of which it seems to become an integral part. This labour power can affirm its productive capacity only by restoring to capital a sum of value higher than which it receives from capital (wages). In order to be productive as living labour and not as mere part of capital, it is no longer sufficient for labour to produce any random value in any random quantity: it must create a ‘super value', a surplus value. The second reason becomes obvious if we look at living labour from the viewpoint of capital. The goal of capital is not to satisfy needs, but to produce profit, surplus value. This does not eliminate the first determination of productive labour, to the extent that capitalism produces commodities and therefore use values. But this determination becomes ‘insufficient'. Use value is no longer, as it was in previous systems, the basis of production; it remains only as something which must be tolerated, a necessary support for exchange value, but whose specific content is of no importance to the capitalist.
Thus, it is not sufficient, for the individual capitalist, for the labour he buys to concretise itself in any particular use value: that labour must expand his capital.
"Surplus value, the specific product of the capitalist process of production, is only created as a result of the exchange with productive labour. What constitutes its specific use-value for capital is not its particular useful character, any more than the particular useful qualities of the product in which it is materialised, but its character as the element which creates exchange-value (surplus value)." 
"That labour alone is productive whose process is the same as the productive process of the consumption of labour power by capital or the capitalist." 
In short, the way of determining productive labour under capitalism differs from the general determination which held true for previous systems by virtue of the difference which capitalism introduces at the level of ‘the productive process of the consumption of labour power'.
This difference does not reside in the actual extraction of surplus labour: the feudal lord or the slave owner of antiquity also disposed of the surplus labour of their serfs and slaves. What distinguishes surplus value from other forms of surplus labour is the fact that it is transformed into new capital and not consumed in a non-productive way as in antiquity or feudalism. It is only when this transformation of surplus value is completed that the expansion of capital has been achieved (the accumulation of capital). The problem of the possibility of this transformation is thus fundamental in the definition of productive labour in the capitalist system.
"The distinction between productive labour and unproductive labour is important with respect to accumulation, because the exchange against productive labour alone is one of the conditions for the reconversion of surplus-value into capital." 
But it is at this level that most of the confusion about the problem of productive labour arises. In effect, the conditions for the transformation of surplus value differ depending on whether they involve the individual capitalist or global capital.
A capitalist consumes what his own enterprise produces only in exceptional cases. Each fraction of capital is only an atom within a network whose complexity increases at the same rate as the socialisation of production. Each capitalist has only the most fragmented consciousness of this network and has practically no control over it. Once a commodity has been produced, it plunges into commercial circuits and disappears from the sphere of the capitalist who produced it. What is important for the capitalist is to receive the monetary counterpart of the commodity which will permit him, in turn, to take from the productive circuit the goods necessary for the reconstitution and expansion of his capital.
It follows that for the individual capitalist, or in a more general way, a fraction of global capital, "... the fact that labour is productive has absolutely nothing to do with the determinant content of the labour, its particular utility or the particular use-value in which it is materialised". 
Whatever the use value of his product might be, from the moment that he succeeds in realising their exchange value, his surplus value can be transformed into capital, the determination of productive labour is independent of the content of that labour.
Global capital, on the other hand, itself consumes the essential part of its product since it is constituted by all the capitalists. The use value content of what it produces directly conditions its possibilities of expansion and realisation.
"To accumulate it is necessary to convert a portion of the surplus-product into capital. But we cannot except by a miracle, convert into capital anything but such articles as can be employed in the labour-process (ie. means of production), and such further articles as are suitable for the sustenance of the labourer (ie. means of subsistence). Consequently, a part of surplus-value must have been applied to the production of additional means of production and subsistence, over and above the quantity of those things required to replace the capital advanced. In one word, surplus-value is convertible into capital solely because the surplus-product, whose value it is, already comprises the material elements of new capital." 
From the viewpoint of global capital, the process of the general accumulation of capital - and this is the only point of view which can be taken into consideration when it is a question of judging the capitalist system - productive labour is that which creates surplus value and which is crystallised in use values which themselves can be productively consumed in the process of capital accumulation.
This why we assert that labour (living and dead) which goes into the production of goods such as arms (as well as luxury goods,etc) constitutes unproductive labour.
"The services that the capitalist buys voluntarily or involuntarily (through the state, etc) for his consumption because of their use-value, do not become factors of capital, any more than the commodities bought for his private consumption. Consequently, these do not constitute productive labour and their agents are not productive labourers." 
Most of the confusion amongst marxists on the question of productive labour has two sources.
The first is that Marx mainly dealt with the problem from the viewpoint of the individual capitalist, or from the viewpoint of the worker as someone exploited by capital. Two reasons can explain Marx's insistence on this single aspect of the problem: the polemic he launched against J.B. Say and Bastiat, who located the problem on this terrain, and the relatively secondary importance of this problem in the course of the general development of capitalism at the time that Marx was writing. (Unproductive expenditures of capital took on really significant proportions only with the period begun by World War 1.) 
The second source of confusion perhaps resides in the meaning of the term ‘surplus value'. Let us imagine a capitalist nation which in the course of a year produces only goods and services whose consumption is directly productive, that is to say, commodities entering directly into the productive process. (We therefore assume that, for example, there is no production of armaments or luxury goods.) At the end of the first year, the capitalists find themselves in possession of a mass of products which they dispose of to the exclusion of the rest of society: this is the mass of surplus value, the real surplus value of the global capital of the nation.
Let us further assume that at the end of this first year, they decide not to devote, as they have been doing up until then, all the extracted surplus value in the production of new productive goods. In keeping with their privileged situation, they agree to create luxury industries: in order to defend themselves from possible aggression from other nations, they decide to create a military industry. These two new industries are obviously going to function according to capitalist relations of production: wage labour, etc...
The problem is the following: what is the nature and origin of the profit which the capitalists of these new sectors are going to collect? Do the workers in these industries create surplus value? What is the relationship between the profit of the capitalists in this sector and the real surplus value of the productive sector?
The answer appears clearly as soon as one realises that all the production of these sectors is bought by capitalists, by the capitalist class as a whole (through their state apparatus in the case of military goods). The wages of the workers in these industries, as well as the profits of the corresponding capitalists, are paid for by previously created surplus value. Global capital is not going to be expanded by a single hour of the labour provided by these non-productive industries. It is, on the other hand, going to devote a part of surplus value which could have been devoted to its self-expansion to the maintenance of both the workers and the capitalists of these branches.
From the viewpoint of global capital - or national, in this case - these workers do not create surplus value. On the contrary, they consume it. But from the viewpoint of the capitalist in the sectors in question, these workers give them through their surplus labour, the right to dispose (at a rate determined in relation to the capital they have invested, according to the law of the ‘capitalist environment') of a certain portion of the mass of real surplus value.
Dassault or Chanel provide other capitalists with commodities produced according to the capitalist mode of exploitation: in the value of these commodities is included the unpaid labour furnished by the workers and whose equivalent in value of course belongs, once again by the laws of capitalism, to Dassault and Chanel. The other capitalists buy these commodities at their value and thus realise the surplus value of their colleagues.
The labour power of the unproductive sectors has thus affirmed its productive capacity in relation to the immediate capital which purchased it, by providing that capital with a value higher than the one which it received. Labour power creates surplus value for capital. But from the viewpoint of global capital, it destroyed surplus value.
The worker exploited by an unproductive capitalist industry is just as much a proletarian as the worker in the productive sector. But whereas the surplus value created by the latter is part of the REAL surplus value which expands global capital, the surplus value produced by the workers in the unproductive sector is a source of profit only for the immediate capitalist; it is an unproductive expense for global capital.
Our point of departure was to answer the question: is arms production a solution to capitalist crises? Can it prevent a new crisis? These questions led us to pose the problem of the determination of productive labour.
The answer to this problem in turn now permits us to respond clearly to the first question; ARMS PRODUCTION IS NOT A SOLUTION TO CRISES.
The outlet constituted by military spending also represents a heavy burden for the economy of each nation.. Military spending is an incredible waste for capital and for the development of the productive forces. . It is estimated that in the past few decades the United States has transformed, on average, one third of its annual surplus into military goods. This means that if this spending could have been transformed into productive goods, the growth of the American economy (provided it found the necessary markets) would have been 33 per cent more rapid.
The production of a cannon implies for capital not only the loss of the labour embodied in the cannon itself but even more a sterilisation of this labour, a blockage of the continuing process of the self-expansion of capital. Capital has to bear not only the loss of past labour but also the burden of a paralysis of its productive process.
It is clear that the ‘military stimulant' can in no way assure the eternal expansion of capitalism. As an outlet it can only be an auxiliary factor (in periods of reconstruction, for example) and in any case its effects, like those of reconstruction, are of a limited duration: a nation that transformed its whole annual surplus into arms would find its economic growth totally stopped at the end of a year. A country can resume its economic expansion only by making these armaments profitable, that is through war (and then only if the country is victorious).  The greater the share of surplus transformed into arms, the shorter the duration of the stimulating effects of this outlet, and the sooner the question of their profitability becomes pressing. The longer this solution is put off, the greater becomes the immense weight of this unproductive burden on the national economy: inflation, decline in the competitiveness of the nation's products on the world market (because the cost of these products increasingly comprises military expenses), which in its turn, through the loss of markets which results, aggravates the internal problem of a lack of outlets. The increasing pressure exerted by the American government on the European nations to take care of their own military expenses cannot be explained in any other way.
This said, we must now answer the question implicit from the beginning: if military production is so detrimental to the development of capital, why do all the countries in the world, and first and foremost the great powers, devote such an enormous part of their productive capacity to this type of production?
We have seen that the demand created by military needs possesses certain particular advantages as an outlet: for example it affects nearly all sectors of industry, while bringing special advantages to those sectors where capital is most highly concentrated.
Its role as an economic stimulant has made certain analysts conclude that the development of the armaments industry was the result of a conscious decision on the part of the capitalists to create an artificial outlet invented to fill the needs of an economy constantly threatened by suffocation due to a lack of markets. Nothing could be further from the truth. Capitalism produces commodities. Naturally its primary concern is the exchange value of the commodity, its monetary equivalent. But this does not allow it to ignore the use value of the commodity. A product which has no use value, which does not correspond to any social need whatsoever is anything but a commodity. It has no place in the capitalist world. The capitalist state which buys armaments is, like any other capitalist, a prisoner of the law of value: it can buy only what corresponds to a real need. Outlets which can be invented at will only exist in the dreams of bankrupt capitalists. The development of the armaments industry is linked to the exacerbation of the inter-imperialist antagonisms. In a world entirely split up among the competing powers and where there are too few pieces for all the competitors, the military strength of each nation becomes the indispensable, inevitable tool for its economic survival. The world wars have shown the price a country has to pay for any weakness in this domain.
It is not only armaments which are an unproductive industry for capital. Capitalist decadence since 1914 is characterised by the dizzying growth of a whole series of unproductive economic activities. All these expenses have the same raison d'etre: to moderate the growing difficulties that the economic system encounters in all areas in trying to assure its own reproduction. Armaments are only one of these activities and if we have dealt with them separately it is because of their very important role as an ‘economic stimulant'.
Among the expenditure of this type which has becoming particular important we can cite:
1. Expenditure to maintain 'social peace'. All the costs of maintaining a whole body of employees ranging from the police to the union apparatus, from welfare workers to sociology professors. Also expenses such as unemployment compensation,etc. (Great Britain, for example, has ‘maintained' more than a million unemployed for more than five years!)
2. Expenditure to moderate economic difficulties within each nation - or within each company. The incredible swelling of the general economic administration of the state: planning bodies, regulatory agencies, etc. It is necessary to add to these all the costs of subsidising failing companies and the ever growing aberrations of agriculture under capitalism: payments for letting the land lie fallow, destruction of harvests and surpluses, the cost of stabilising the agricultural markets, etc. There are also the expenditures of private companies constantly confronted with sales problems which competition makes even more acute: marketing expenses, administrative costs, and above all advertising (these costs are generally lower in state capitalist countries but any savings are offset by the enormous waste that bureaucratic irresponsibility entails in the distribution network). The development of the Tertiary sector is largely the result of this type of unproductive expenditure.
3. Expenditure caused by the aberrations resulting from the desperate actions of firms in certain countries (United States and to some extent Europe and Japan) faced with the need to compensate for the lack of markets. The most striking aspect of this phenomenon is in the reduction in the use value of consumer goods (cars, stockings, electric appliances, etc) consciously made with a view towards planned obsolescence. This technique is reinforced by advertising intended to create ‘in' fashions and therefore ‘out' fashions. Amongst these aberrations we can take the example of the automobile: at first a real social need it has gradually become a social disaster. The productive aspect of the car (a means of transportation for labour power) has more and more given way to the unproductive aspect of the commodity as a noxious waste.
4. Expenditure arising from international relations in particular, the extent of military expenditure staggers the imagination. The military expenditure of the United States and the USSR together are greater than the combined national incomes of all of Latin America plus India and Pakistan. The state of the Middle East devote around 25 per cent of their GNP to armaments. In the last ten years world capital has spent more on armaments than in the first fifty years of this century - the two world wars included.
The Disproportionate Growth of the Unproductive Sector: The Characteristic of Decadence
Unproductive expenditure also existed in the ascendant phase of capitalism. Luxury goods for the ruling class, for example, have existed since the beginning of capitalism. The same is true for the police, army and the administrative costs of the state. But the extent of these expenses in the past centuries is in no way comparable to these same types of expenditures in the past sixty years.
The unproductive content of these ‘artificial' costs of production in ascendant capitalism was attenuated because the relations of production were at the time basic to the development of the productive forces. When capitalism was taking over the planet by destroying pre-capitalist relations, when it was imposing its techniques on the entire world, its relations of production were also productive forces. At the time the general costs necessary to the maintenance of the relations of production were offset by the productive nature of these relations. Therefore, when an imperialist army, for instance, imposed capitalist relations in a part of the world, world capitalism was enriched. These armies were on certain occasions true productive forces for world capitalism.
This situation completely changes the moment that the world is divided up amongst the competing powers. The war could only bring about a redivision of the spoils, any new conquest having become impossible. From that time, one capitalist power could militarily win a sphere of influence only at the expense of another power. For world capital war could only represent an internal disruption, a disastrous waste. When capitalist relations of production cease to be the instrument for the development of the productive forces and become fetters, all the ‘artificial' costs that they entail become simple waste. It is important to note that this inflation of artificial costs is an inevitable phenomenon which imposes itself on capitalism with as much violence as its contradictions.
For a half century the history of capitalist nations is filled with ‘austerity programmes', attempts to turn back the clock, struggles against the uncontrollable expansion of government costs and unproductive expenses in general. Every time international competition is aggravated in a critical fashion, the question of these expenses is more sharply raised . All these efforts, however, systematically end in failure. Remember the Nixon austerity plan (August 15,1972) and the anti-inflationary budget that went with it: in spite of all the speeches, military expenditures - the principle source of inflation in the United States - were once again increased. Unproductive expenditure functions for decadent capitalism like some drug, which is at once a medication and a poison, taken for certain illnesses. The worse the disease becomes, the more you have to increase the dosage; the more you increase the dosage, however, the worse the disease becomes. Inflation is the cancer of modern capitalism and unproductive expenses are its main source of nourishment. The more difficulties capitalism faces, the more it must develop its artificial costs. This vicious circle, this gangrene rotting the core of the wage labour system, is only one of the symptoms of the real disease: the decadence of capitalism. The consequences of all this have been known for more than half a century: world war or proletarian revolution: socialism or barbarism.
 Marx, "Materiaux Pour L'Economie", in Oeuvres, Vol.2, La Pleiade, p.387
 Marx, Capital, Vol. 1, Pt. V, Chap. XVI, in Oeuvres, Vol. 2, La Pleiade, p.1001 (We cite this French edition because it renders the passage in question more accurately than the English translation)
 Marx, Capital, Vol. 1, Pt.3, Chap. VII, London 1970, p.180-181
 Marx "Materiaux Pour L'Economie", in Oeuvres, Vol. 2, La Pleiade, p.392
 Ibid, p.388
 Ibid, p.398
 Ibid, p.393
 Marx, Capital, Vol 1, Pt. VII, Chap. XXIV, Section 1, London 1970, p.580-581
 Marx, "Mareriaux Pour L'Economie", in Oeuvres, Vol. 2, La Pleiade, p.390
We can only note the ignorance both of marxism and of capitalist reality shown by ‘marxists' like H. Weber who, taking refuge in "dry and arid theory" see only "ethical criteria" in this definition of productive power.
 The definition of productive labour under capitalism can only be related to the useful content of this labour at the point of a certain development of the system. The first capitalist manufacturers produced mainly ‘unproductive' goods: arms, gun powder, luxury cloth, etc. However, this did not pose a major problem for capitalism's development. The reason was that the capitalist sector remained a mere workshop within a process of social production still largely dominated by pre-capitalist formations (essentially feudal ones). Agricultural and artisan production could still supply the capitalist manufacturers with the basic materials for the means of production and for the consumer goods necessary for production which was functioning with an extremely low technical composition of capital, ie, in which living labour predominated dead labour. (The first workshops were often simply associations of artisans, working with the old methods, but subjected to the regime of wage-labour by an entrepreneur.)
In these conditions, the use-value content of capitalist commodities had little importance for the development of capital. It is only when the capitalist mode of production takes possession of the whole of social production ("the real domination of capital") that the definition of productive labour has to take into account its specific utility. That is why Engels could write: "At the beginning of the fourteenth century, gunpowder came from the Arabs to Western Europe, and as every school child knows, completely revolutionized the methods of warfare. The introduction of gunpowder and fire-arms, however, was not at all an act of force, but a step forward in industry, that is, an economic advance. Industry remains industry, whether it is applied to the production or destruction of things" (Engels, Anti-Duhring, Moscow 1969, p.200)
H. Weber, who had the misfortune to write on this subject in the above mentioned pamphlet, has not failed to deduce that today industry dedicated to the construction of arms, "these modern consumer goods", constitutes an industry as productive as any other. Mr. Weber really doesn't see why anyone should think that anything changes in the productive or unproductive nature of labour in Dassault's factories whether they are producing bombers or civil aviation aeroplanes. In the fifteenth century, all the capitalist enterprises could adapt themselves to producing nothing but arms and still undergo a certain expansion. That Mr. Weber portrays the same thing today and concludes that this represents capitalist expansion - all this falls within the logic of his argument.
 We are presupposing that this nation has not external trade enabling it to deflect the unproductive impact of these industries onto other nations as in the example given above ( the example of Dassault and Peru).
 The living labour and the means of production dedicated to this kind of production are not productive, they are not productive forces - or rather, if one wants to define them from a general abstract point of view, from the point of view of the labour process, they are sterilised, annihilated productive forces. Only the science and the techniques of the production developed in these sectors, to the extent that they remain in existence after production and only to the extent that they can be applied to the creation of productive goods, are productive forces.
 Russia's dismantling of entire factories in Czechoslovakia and Mongolia in order to take them back to its own territory at the end of World War 2 is a spectacular illustration of the ‘profitability' of military production.
 The success in the United States of someone like Ralph Nader, enemy of murderous cars and of products which fall apart, doesn't only derive from the anger of ‘consumers'.