Automation. On Jason E. Smiths “Smart Machines and Service Work”

Jason E. Smith’s book Smart Machines and Service Work can be summed up in one denominator: The theories of post-capitalism and automation turn out to be a hollow form of Californian ideology. In it, the world of digital technologies – be it the smartphone, smart cars or smart cities, smart factories or smart hospitals, be it the implementation of computer technologies that replace and save labour, not only as automation of production processes but also of offices, This would lead to an explosion of productivity, but at the same time to a rapid disappearance of jobs, so that the question would arise as to what people would do without work and whether the new world of joblessness could not lead to social misery of unimagined proportions for the many.

The three decades after World War II are seen as the years of rising labour productivity in industry, which created the material conditions for both rising profits and higher wages, with the trade unions collaborating more and more with the owners and managers of the large industrial companies. This simultaneously led to workers being pushed out of capital-intensive industries and absorbed into service industry sectors such as health and education, state apparatuses and financial enterprises, food industries and other tertiary sector areas in the wake of process innovations. Today, according to the proponents of automation theory, it is no longer primarily industry that is being restructured by job-saving technologies, but the service sector itself, whose jobs that were the result of the first wave of automation in industry are being replaced by a new wave of automation: Automation 2.0. But, Smith immediately interjects, while the first wave of automation took place in the post-war industrial boom, we are now in the midst of a long period of economic stagnation or stagflation, with the core countries of capital in particular characterised by falling rates of GDP growth and labour productivity since as early as the 1970s. Moreover, driven once again by the financial crisis of 2008, workers’ real wages are falling, accompanied by growing unemployment and falling rates of investment in industry and even negative rates in labour productivity. We will examine later whether it is permissible for Smith to reduce the problem of productivity purely to labour productivity and completely ignore the issue of capital productivity.

Smith feels it is imperative to point out that the ubiquity of technological tools such as the smartphone, which integrates telecommunications, shopping, video and sociality in one gadget, in the context of the rise of the big tech companies and the new platform firms, have far-reaching effects on finance, mobility, consumption, etc., but only negligible effects on productivity at work. At this point he quotes the economist Robert Solow, who already in 1987 said that the computer age could be seen everywhere except in the statistics on productivity development. For Smith, on the other hand, there are a number of reasons why the effects of automation and new technologies do not have a significant impact on labour productivity growth rates in developed economies, be it that certain work processes still require intuitive and socially mediated skills and knowledge that cannot yet be replaced even by AI machine learning, be it that the existence of cheap labour prevents companies from investing in expensive machine systems that may not be amortised after all.

The stagnating wages that have been recorded for decades are themselves an effect of technological stagnation, insofar as rising wages are also partly dependent on increases in productivity growth rates, which in turn require higher rates of investment and the implementation of new technologies, although it is precisely in the USA and other developed countries that a decline in private investment has been recorded since the 1980s, which can be traced back to stagnating or falling profit rates also as a result of the expansion of cheap or so-called unproductive labour. Today, for example, the introduction of robots in areas in companies such as Amazon and Walmart does not necessarily lead to the replacement of jobs, but rather to a further demand for cheap labour. Amazon announced in mid-2020 that it would hire 100,000 new workers for operations in its warehouses and for delivery services. In addition, when revenues and profits decline in other industrial sectors, one cannot assume that high investments in fixed capital will take place, while companies with high cash flows often enough use the cash to buy back shares in their own companies. Increases in labour productivity then increasingly occur through increases in labour intensity for a given capital input.

Smith begins his book with a short treatise on the history of automation, starting with the Greek terms automaton and automata in Homer and Aristotle, the use of the terms in Descartes and Liebniz (monads as automata), the description of the organism as a machine, and ending with Norbert Wiener’s definition of cybernetics in the 1940s as the scientific study of the communication and control of animals and machines as a result of feedback processes. Cybernetic feedback is also about whether the disturbances are compensated for by the machine complexes themselves or externally by human input. At least in the input of setpoints, even if it is a one-time programming, the human switching element is still present, the instrumental organon that keeps the automaton from becoming pure perpetual motion. In cybernetic discourse, the self-regulating automaton thus still remains bound to the coupling of machines and humans, whereby human decision-making and control follows a logified structure of either/or, the yes/no decision, and therefore excludes the allowance of infinitesimal difference.

At the same time as Wiener, Ford vice-president Del Harder had simply used the term automation to describe electromechanical, pneumatic and hydraulic parts/machines in industrial production, while a decade later the term was used to describe the further development of feedback technologies and the self-regulation not only of individual machines but of entire machine systems. More than a century before that, however, Andrew Ure, an important reference author for Marx regarding his own theory of machines, was already raving about an automatic factory with clean, well-ventilated and spacious halls, in which the workers were already completely alongside the machines and at best still controlled them, or as Marx put it, functioned as regulators rather than initiators or primary components of production.

Smith mentions the studies of Frederick Pollock, who defines automation as a technique of industrial production in which machines are controlled by other machines, following on from the theory control systems of engineering or the application of the principles of automatic control of dynamic systems, be they biological, mechanical or social systems. We are talking about closed feedback loops in which a given output is monitored and measured by a control system, which then processes the information and passes it on to adjust the input if necessary. Think of thermostats and cameras as well as the first computer systems used in petroleum refineries and chemical distilleries.

At this point, one can note with Hans-Dieter Bahr that cybernetics is not essentially different from mechanical machine discourse as far as the capacities for control and order are concerned, be they internal or external. The construction of a control loop, in which any change in a controlled variable that is considered a deviation is compensated by a variable that counteracts it, is preceded by the Newtonian formulation of the equality of effect and counter-effect. If we speak of cybernetic systems in terms of the fact that, as far as possible, no (disturbing) human element is involved in their processing and the controlling, which here, in contrast to an action, appears only as “behaviour”, is essentially reduced to the switching on or off of the machine systems by the human agent, then the anthropological scheme is by no means turned on its back. For one continues to speak of controlling interventions in a control loop, which are perfect interventions precisely when there is no malfunction or accident, whereby control tends to be reduced to the maintenance of a linear, malfunction-free process that is supposed to correspond to a steady-state equilibrium, which in each of its moments is the respective distance between binary states, up to the limit value of their coincidence.

Digital automation, which uses discrete rather than continuous signals, now functions electro-computationally, constantly generating simulations of feedbacks and other autonomous processes in various social fields. It unfolds in diverse networks made up of electronic and neural connections, in which employees and users alike are integrated as quasi-automatic relay stations in a continuous stream of information from machines that process at different rhythms. In the 1960s, digital implementation was still limited to individual units through numerical control machines, but even in the 21st century, only the most sophisticated factories in certain industrial sectors have automated and centralised control systems for the entire machine complex. In regions where wages are low, there are still many enterprises with high labour-capital ratios and primitive control systems. But, Smith says, ultimately automation is not just about replacing manual tasks, but also about those of supervision and control, and moreover about integrating earlier discrete operations and thus creating an uninterrupted flow in labour processes. This flow production without human intervention was already described by Marx as the maximum continuity of production, and it is not surprising that this type of production first took place in industries with liquid or gas-like products, in oil, electricity and chemical factories, before it was introduced into the automobile industry, for example. Smith sums up that the idea of a fully automated factory with integrated planning, processing and distribution using digital feedback systems has remained largely a vision to this day.

The danger of mass unemployment set in motion by automation, which Pollock also feared, with armies of the unemployed unable to find employment, has not materialised, at least in the highly developed countries, according to Smith, insofar as since the 1960s there has been a massive expansion of the so-called service sector, which has taken over the absorption of the workers made redundant by technological innovation, with the additional integration of millions of women into these sectors. The declining number of workers able to produce higher industrial output in the wake of the use of new technologies (only a quarter of the total labour force in the US in 1970) was matched by an explosion in the absolute number of sercice workers. Smith notes a growth in the number of service workers in the US from 62 million in 1953 to 140 million in 2010. Thus, the rhetoric of job-killing automation increasingly subsided and flared up again in the 1990s with the introduction of barcoding and RFID tracking technologies that enabled important innovations in international supply chains and the organisation of management. It was Jeremy Rifkin who, in his book The End of Work, predicted an almost completely automated service sector by the middle of the 21st century. Since that statement, millions of new workers in developed countries have migrated to the service sector, which remains characterised by low productivity. Overall, the trend of declining labour productivity since the 1970s, interrupted only by a tick in the mid-1990s, continues to be confirmed. This is where Smith’s real investigation begins and where the problems start.

When asked to explain why labour productivity has declined over the last two decades, automation enthusiasts cite difficulties in measuring progress in automation in sectors such as health care and time-lags in implementing or diffusing the techniques into the economy. The breakthrough of the computer began with the Intel 4004 microchip in the 1070s and was followed by exponential growth in digital capacity (Moore’s Law), as well as the use of fibre optic cables, accompanied by a steady fall in the price of microprocessors, leading to the digital landscapes we know today: Split-second financial transactions, streamed video, social media and the big tech companies. It was economist Robert Solow who first spoke of a productivity paradon, insofar as rapid digitisation since the 1970s has not been accompanied by a rise in labour productivity. A new technological regime, as theorist Carlota Perez writes (think here of the use of the electric dynamo, for example), requires the investment of high fixed capital in equipment, machines and factories, the value of which is only gradually amortised over longer periods. (Perez identifies three correlations of innovation thrusts in the “real sketor” with those in the financial industry in the domestic history of capital: Industrial Revolution and local banks in England, Fordist mass production and consumer credit, and finally communication technologies and derivatives). But it also needs an economic and political environment in which technological thrusts can take hold. For example, it was the specific conditions after World War II (destruction of capital, etc.) that were responsible for the establishment of a new technological regime.

The accelerated growth driven by computers and information technologies is called a “singularity” in scientific circles, but a distinction must be made between technical and economic singularity. Economic singularity is measured by the general development of substitutability between information and conventional inputs. The economist William D. Nordhaus has pointed out in a recent study that economic growth depends on the extent to which materiality can be replaced by electronics. In addition, at the macroeconomic level, it must be taken into account that higher productivity leads to lower prices, so that in toto, a rising share of high-productivity sectors can only be proven if their volume growth more than compensates for the decline in prices. Nordhaus also explores the question of the constant substitutability of certain factors of production by information at the level of the organisation. He proves that this was not the case in the past and predicts only a slow development towards economic singularity for the 21st century, although the capital intensity (capital stock compared to labour input) continues to increase and thus also the share of information capital. Nevertheless, Nordhaus assumes that digital information and communication technologies in the new millennium have contributed to establishing a new level of productivity in production and distribution by condensing and accelerating the exchange of goods between companies, enabling rationalisation processes in global value chains and in the supplier industry, the reorganisation of areas such as architecture, urban planning, health care, etc. The software, with which management methods are developed and implemented, is a key factor in this development. The software with which management methods, derivatives and digital logistics are processed can certainly be understood here as a basic innovation that is integrated into a new technical-economic paradigm.

Smith vigorously contradicts this. Since the 1970s, one must assume stagnation and a decline in the growth rates of both labour productivity and industrial investment. At this point, Smith again mentions the objections of some automation theorists such as Paul Mason, who, with recourse to Ernest Mandel, refers to the long Kondratieff waves, according to which the use of the Intel microprocessor in a phase of economic stagnation in the mid-1970s brought about clusters of innovations that were supposed to ensure a new economic upswing after 35 years at the latest, which would then also be reflected in the corresponding economic data (productivity, GDP, investment). Mason attributes the fact that this did not happen at first to the financial crisis of 2008, although he continues to dream of an exponential takeoff in productivity in view of the fact that since the 1970s we have been dealing with a climate of economic stagnation and, according to Smith, a persistent technological inertia in the developed countries. Thus, the circulation of images via networked computer terminals is essentially just an amalgamation of innovations in the field of telecommunications (radio, telephone) and certain image technologies (photography, moving images). And the smartphone is merely a better version of existing devices, it is a toy and not a tool, used primarily in the entertainment and leisure industries and with little effect on productivity and modelling of jobs – ultimately, according to Smith, we are dealing with a tsunami of infantile gadgets for children and adults. For Smith, Facebook and Google represent, in economic terms, nothing more than refined “advertising delivery systems” that generate much of their revenue through the sale of ads. Zuboff, however, goes on to speak of a new logic of accumulation within which the new surveillance capital (data on human behaviour including seemingly useless surpluses of data (errors, syntax, signs, etc.) is extracting, i.e. sucking up data and errors and constantly combining them with other data in order to feed intelligent machines that use algorithmic processes to produce predictions about the future behaviour of users, predictions that are then offered and sold as quasi-derivatives on behavioural futures markets.

The new platform companies like Uber, Lyft and AirBnB are also nothing more than digitally mediated arrangements in which users and providers are brought together in a digital marketplace, with the platform companies collecting fees for each transaction. At best, these are new monopoly business models, but not new innovative technologies related to the application of productive labour.

If a new era of technological innovation and economic growth were really dawning with automation, then we would also have to deal with rising rates of industrial investment and fixed capital. Meanwhile, economist Robert Gordon, in his book The Rise and Fall of American Growth, speaks of a rapid fall in net investment by private US companies since the new millennium. The rate of net investment relative to the capital stock has been falling since the 1970s and dramatically since 2002 (from 1970 to 2002 we are looking at an average of 3.2% net investment in the US, in 2013 it was only 1%). If, in addition, according to the latest statistical measurement methods, financial expenditure on intellectual property also counts as investment, i.e. legal titles for technologies to secure the flows of revenues, then, according to Smith, the investment rates for new, more efficient work processes and new organisational methods should be set even lower. At this point, Smith speaks of the dynamism of industries with high and dense competition, which require constant innovation, versus non-competitive sectors (anti-markets), where market shares are secured through property rights. But as Katharina Pistor’s new book The Code of Capital is about to show, legal coding (property,contract, insolvency law, etc.) can also create and protect capital, modules that assign properties to certain assets, such as priority to claims, permanence that extends temporal claims, universality that extends them spatially, and convertibility that allows capital and asset owners to turn credit claims into state money. For some financial instruments, the code itself is the capital. In the discussion of productive and unproductive labour and the question of finance, we will further clarify whether Smith is not applying a truncated notion of capital and productivity.

In discussing the so-called Big Tech firms, Smith points out that in 2007 the 10 firms with the highest market capitalisation were multinationals in finance, industry, oil, gas and telecoms, while today there are seven Big Tech firms like Apple in the top ten in the ranking of large companies that contribute little innovation to product development and new forms of operational organisation. In 2018, Apple held cash reserves of $300 billion, while as much as $210 billion has been used for share buybacks since 2012, and precisely not put into research and innovation. Smith forgets to mention that Apple is also active as a financial company, but as we will see later, Smith mistakenly understands the financial industry as a purely unproductive sector. Today’s digital technology and internet companies, all of which are oligopolistic or even monopolistic, achieve their success by exploiting network effects; their assets are the users who are networked and whose information, generated by the exchange, is extracted and utilised.

At the same time, according to a study by the OECD, we are increasingly dealing with so-called zombie companies that would no longer be competitive under normal market conditions and would be forced out of the market by more efficient and technologically innovative companies, but which are able to keep their heads above water by borrowing due to ultra-low interest rates driven by the policies of the central banks (QE and cheap money). These companies are themselves struggling to service their debt through their revenues and profits. The companies, which are characterised by low productivity, hang on the drip of cheap credit, which leads to only a suboptimal allocation of economic resources for the total capital. For new firms with higher innovation potential, this can even make market entry more difficult. For Smith, at any rate, the two marketerist business models of the period after the 2008 financial crisis are the platform or big tech companies and the zombie companies. In saying this, of course, he is underestimating the still weighty role of financial companies. Today, there are just 30 companies in the US that rake in about 50% of the profits of listed companies, companies that Bichler/Nitzan call dominant in their concept of capital as power, while in contrast, a large proportion of companies have large loan portfolios that are hedged and traded through financial instruments such as CDLs, which could have a similar dynamic to CDOs, a flashpoint of the great financial crisis of 2008. For example, 85% of the outstanding loan debt of private companies can be traced back to this type of financing.

After the housing crisis, bank demand shifted to a similarly risky asset, collateralised loan obligations (CLOs). A CLO is structured similarly to a CDO, but instead of loans to homebuyers, loans are made to businesses – particularly troubled businesses. CLOs bundle and structure leverage loans, the subprime mortgages of corporations. These are loans to companies that have reached their borrowing limit and are no longer able to sell bonds directly to investors or qualify for a traditional bank loan. There are currently more than $1 trillion worth of leveraged loans outstanding in the US, with the majority held as CLOs.

CDOs and CLOs are remarkably similar because, like a CDO, a CLO has multiple tiers that can be sold separately: The bottom tier is the riskiest, the top tier the safest. If only a few of the loans in a CLO default, only the bottom tier will suffer a loss and the other tiers will remain safe, but if defaults increase, then the bottom tier will lose even more and the other tiers will also suffer losses, but the top tier will remain protected because it will only lose money when the lower tiers are wiped out.

According to many estimates, the CLO market is bigger than the CDO market with its subprime mortgages at their peak. The Bank for International Settlements, which helps central banks ensure financial stability, estimated the total size of the CDO market in 2007 at $640 billion, and it estimated the total size of the CLO market in 2018 at $750 billion already.

In the next chapter, Smith turns to employment and first finds in the US the paradox of falling real wages with a simultaneous rise in employment. Smith looked more closely at US statistics on employment and found that in 2016 only 69% of workers were actively participating in the labour market, meaning that 31% were no longer even looking for jobs. This means, of course, that the unemployment rate is higher than officially proclaimed, standing at 20% in the US in 2017, according to the Financial Times. For Smith, however, the permanent stagnation of wages has its main reason not in the relationship between supply and demand for labour, but on the one hand in the relationship between total economic output and workers’ wages. Insofar as labour income is inverse to the share of output that flows back to the capitalists as profit, any change in the distribution of labour and capital income with constant labour productivity influences both shares, one share positively, the other negatively. This is a direct expression of class power in capitalist economies, which has led to an increased grip of capital on output as a result of the growing decomposition of the political working class since the 1970s. Secondly, falling wages are due to a reduction in labour productivity growth rates. If the division of income between capital and labour remains constant, real wages can only rise if labour productivity rises (higher output per worker or hour), so that the cost per unit of output falls, allowing rising margins to be shared between capital and labour while incomes remain constant, as was the case in the first three decades after World War II due to the use of new technologies in developed countries. And rising labour productivity is possible, among other things, on the basis of rising rates of the use of fixed capital per worker, the investment in machinery. And Smith states that in developed countries wages have fallen not because of an oversupply of labour, but because these economies are in the process of perpetual stagflation. While there is a discourse that with globalisation and outsourcing of business, just-in-time production and computerised logistics, products are circulating around the world cheaper and faster or at shorter intervals than ever before, logistics and distribution clusters in the commercial sales sector are becoming more efficient, and at the same time more and more labour is being sucked into these sectors, Smith says that McKinsey’s figures show that labour productivity growth is at a low level. Smith even states a rate of productivity growth of -0.2% for the interval from 2007 to 2017 in the US.

Smith quotes Anwar Shaikh, who cannot establish any a priori correlation between the movement of wages and productivity rates, the latter being merely a material condition for potential wage increases without determining them exactly. Wage increases require certain social institutions and class struggles. For the period from 1970 to 2008 in the USA, Shaikh speaks of a technological dynanism with productivity increases accompanied by real wage reductions, which led to an increase in capital accumulation until 2008, whereby Smith accuses Shaikh of only taking into account the productivity movement in industry, but not that in all employment sectors, in which 4 out of 5 workers are employed in the USA. And it is precisely in the service industry that there have been no significant increases in labour productivity.

However, Shaikh’s argument is more complex. For Shaikh, a stagnating or falling rate of profit cannot be the sole decisive factor in explaining the course of a boom or a depression, at least as far as the short-term or medium-term view of capital accumulation is concerned. Shaikh’s justification for the prolonged boom in the USA from 1982 to 1997 is based, on the one hand, on the decline in real wages produced by neoliberal strategies and on the sharp fall in interest rates – factors that would both have led to an increase in corporate profits. In addition, however, there is the fall in the organic composition of capital, especially for the period in which the new micro-technologies massively lowered the value of constant capital. In Shaikh’s view, the moderate increase in the rate of profit between 1982 and 1997 is thus by no means only due to the increase in the rate of surplus value, but also to the temporary fall in the organic composition of capital, whereas in his view it was the low interest rates that made the flat boom in net profit rates and a relatively constant rate of accumulation possible. So what really motivated or drove the boom in the 1980s in the US, according to Shaikh, was effectively the dramatic fall in the general rate of interest, which fell from 14% in 1981 to over barely 1% in 2003. And it was precisely this that led to the rise in the net profit rate that substantially accelerated capitalist growth in the two decades after 1982. At the same time, falling interest rates favoured, among other things, the increase in debt-financed spending by private households, which meant that industrial growth ran parallel to the upward movements or bubbles on the property and financial markets. But interest rates also fell in other parts of the world, sometimes even faster than in the US, and this in turn led to an international boom in industrial accumulation and financialisation, especially in the so-called emerging markets. It was thus the fall in interest rates, among other things, that was largely responsible for the rise in private debt without, at least in phases, drastically increasing the debt repayments of wage earners.

In the next chapter, Smith refers in detail to the economist William Baumol, who assumes that the developed industrial countries are divided into two important economic sectors, namely the technologically progressive sector, whose production processes include innovations, high accumulation rates and large-scale processes, and the technologically stagnant sector, whose technological structures tend to prevent a significant increase in labour productivity. It is precisely the dynamism of the first sector that causes the fall in productivity rates in the second sector, insofar as machines are used in the innovative sector over certain periods of time, enabling higher output through the use of fewer and fewer workers (if output remains constant, then the use of new machines definitely means that fewer workers are employed, who then have to find employment in other sectors). At this point, Smith again only considers labour productivity, without examining the issue of more efficient use of machinery, which is reflected in capital productivity. Further, by increasing the efficiency of the use of labour, the products of the innovative sector become cheaper and cheaper, so that even if consumers’ incomes remain constant, a smaller proportion of their income is used to buy these goods. This also leads to more and more labour being used in technologically stagnant sectors.

Thus, the stagnant sector will tend to increase in terms of employment figures, although income ratios are relatively elastic here. And because productivity in this sector remains low, the increase in output here leads to an increase in employment. Ultimately, this development, i.e. the differentiation in the growth rates of productivity, leads to a gap, whereby productivity in the first sector continues to rise, while in the second sector (service) it tends to remain constant, and this in turn leads to the fact that the jobs that are lost in the dynamic sector are absorbed in the stagnant sector. For Baumol, this development leads to a growth rate in the economy that asymptotically approaches zero. Solow’s productivity paradox, which states that the computer is everywhere today except in the statistics on productivity, would now have to be corrected to the effect that the rapid computerisation of the innovative sector of the economy has led to declining growth rates in productivity throughout the economy.

In many labour-intensive sectors, the quality of a product is determined by the mass of labour used, the knowledge and the skills of the workers. Smith cites the example of the teaching profession, which remains largely resistant to time-saving methods and technological innovations. The service sector is an extremely heterogeneous area that contains a wide range of economic activities that differ greatly in terms of labour income, qualification, location and size of the enterprises and finally the labour-capital ratio, so that with Smith we can assume a rather negative definition of this sector, namely that the service sector includes all economic activities except those of industry and agriculture. Here the question already arises whether the financial industry, as Smith suggests, can be added to the rather unproductive service sector, or whether the financial industry, as we suggest, is not sui generis productive for capital. We will come back to this later.

In the US, Smith points out, the statistics refer to firms rather than individual employments in questions of employment figures. For example, Smith notes that in the case of Apple, a company with an enormously high market capitalisation, which is listed as an industrial company even though it has no factories and thus only a small part of its costs can be attributed to production in China and other countries, all workers are subsumed in the US statistics as industrial employees, so that here the number of employees in the statistics increases. These misleading national statistics for companies that outsource much of their productive economic activity to other countries indicate that the classification of employment in this or that sector remains precarious. Moreover, there is a deep dichotomy between two separate service sectors, the business sectors, which are often intermediary and supply products to industry, thus belonging to it, and the consumer sectors, which supply products to individuals and families. For example, areas such as transport, research and design, which in the statistics belong to the service sector, are in fact part of the industrial sector, although many companies actually outsource these areas in the course of the division of labour and globalisation.

On the other hand, restaurants are categorised as food service, although the provision of food to consumers is linked to highly efficient processes reminiscent of industrial processes. At this point, Smith embarks on a historical outline of the term “service”, starting with Adam Smith’s definition, according to which it is an activity whose product dies instantaneously during its performance (e.g. the hairdresser), i.e. production and consumption coincide. Going further, the German economist Georg Quaas, for example, has defined the service as a transformation of a material thing carried out through labour, which is owned by the person who demands the service, whereby the service is provided by non-owners, the so-called service providers. In this context, the service provider (hairdresser, shoeshiner, beautician, etc.) remains the owner of his or her most important work tools, which he or she needs for his or her activity, whereby he or she works on “objects” such as hair, face or shoes, which are to be understood as the property of another from the outset.

This presupposes that, unlike in industrial production processes with a dense division of labour, producers are usually only engaged in the manufacture of a discrete “product”, whereas today a dense division of labour prevails in industrial production, in which only a few workers are directly engaged in the manufacture, design and transport of the products, and some of the workers participate only indirectly in the production process precisely by monitoring production, assisted by computer-controlled equipment. Here, the strict distinction between activities that lead to the production of products and those that have no direct relation to the production of use values becomes fragile. The latter, like even those of the large cleaning firms for the production processes, nevertheless remain essential for production and certainly have an effect on the labour productivity of the workers directly employed in industrial production. One might even conceive of them as productive activities themselves.

For Smith, however, it is still the direct industrial labour processes (automobile, steel, oil, etc.) that are the most mechanised and capital-intensive and, indeed, the most productive, taking place in large factory spaces on a national and international scale, although, as the critics again argue, there do not seem to be any compelling reasons for not extending the scope of the highly-poductive sector to certain service sectors, where processes of rationalisation through the use of digital technologies are constantly taking place and destroying jobs. Smith disagrees with these critics because there are more and more sectors in the economy where there is no significant increase in labour productivity. For example, many jobs in the public health and education sectors are much less exposed to the pressures of competition than private economic activities, insofar as private companies have to make their production processes more efficient in order to offer cheaper products and services on the market. On the other hand, a large number of labour-intensive jobs in the service sector are much less subject to competitive pressure or outsourcing to other countries because they have to be consumed close to where they are produced. These jobs are often poorly paid, so there is often no reason for companies to replace these precarious occupations with the use of new machinery, which also takes years to pay off. Finally, many of these jobs cannot (yet) be replaced by even the most intelligent AI machines because their operational structure, such as nursing, requires intuitive complexity (from haptic to affective), emotional intelligence and dealing with uncertainty. Thus, and this is Smith’s resume again, a polarisation between a highly mechanised, capital-intensive and productive sector and a much larger service sector with low productivity gains has become entrenched in developed countries in recent decades.

But at this point, one would at least have to differentiate further, so that the argumentation would become even more complex. Of course, with Tony Norfield, the question immediately arises as to whether these processes that Smith is referring to are represented in the standard statistics at all.

According to our understanding, the process of “service” is precisely then included in the network of “productive” money capital circuits (which of course make it a process of potential surplus value production), if it is a) a so-called investment use and not just a purely consumptive use of the service, and if one can b) expect that a profit is realised with the price of the product of the service. Thus, here too, it should not be the reference back to materiality or skills, however justified, which ultimately only defines value-creating or productive labour; rather, according to a thoroughly Marxist reading, services as investment and thus potentially value-creating production processes take place precisely when the corresponding services enter into capitalised production and the realisation of its commodities, for example, for example, when a painting company finishes the construction of a factory building with the design of the facades, whereby value creation has taken place, the value of the product “design” has been objectified in the value of the factory building, which is successively transferred to the value of the products manufactured in the factory building through depreciation in the following years. Or one buys all the material of a factory hall in individual parts and commissions a service company to build or assemble the parts, the result being an installed factory hall that will certainly remain standing for a few years for the production of profit-making goods. And with regard to the concept of information work, it can similarly be said that the materiality of the manufactured product alone can hardly be considered a decisive determinant of the production of surplus value; first of all, information always possesses a material carrier, be it, for example, according to Brown, moving gas molecules, sound waves propagating in water, concrete or other material, electromagnetic fields, hard drives or other storage devices. Now, certain types of information can hardly be used profitably, whereas on the other hand information can also be kept scarce or others can be excluded from the use of knowledge/information and thus information can be sold profitably, and this should not be disputed, for example, in the case of Microsoft; at most, it would still be open to debate here whether this type of informationalisation is about the extraction of added value or a skimming off of informational rent. And indeed, social values such as truth, power, justice and beauty should also be regarded as objective wealth, i.e. as bearing exchange value. Roads as parts of the infrastructure function as collective means of production for capital and therefore have to be considered as productive state expenditures, universities or hospitals as productive enterprises if they invent profitable medicines, technologies and apparatuses.

Let us now follow Smith further in the extremely exciting discussion on the question of how productivity or labour productivity in the terms of value, price and materiality can be defined at all. Smith first mentions Baumol’s quite common definition, according to which the rate of productivity involves the division of a given output of an enterprise expressed in money (whereby Smith would then have to distinguish between value and price here) by the quantity of labour employed for this output in money (one usually assumes market prices here). However, productivity can also be measured by replacing monetary output with the output of physical units (of a given product). Smith first mentions this in terms of “universal” methods of measurement in a post-capitalist (planned) economy. In which the social resources of labour, raw materials and machines should be applied and distributed without the intermediary of prices, market and profit.

However, the relationship between these two measures of labour productivity is not so simple; an increase in the material measure of labour productivity does not necessarily lead to an increase in labour productivity measured in money. If prices fall more than productivity rises (as a result of technological innovation), then this leads to a fall in the net product of productivity measured in money. In turn, a firm that can set the price of a product from 100 to 50 euros while increasing the output of that product from 50 000 to 100 000 has the same output, namely 5 million euros. In monetary terms, there would be no growth in productivity even through the use of new machinery leading to a doubling of production. So far, monetary productivity only applies to the numerator, while the denominator involves the measurement of units in labour time, which does not fully express the dynamics of capitalist production, insofar as enterprises are mainly interested in the relation between the price of the output and the cost of the input, often especially that of the input in labour. If one now replaces labour time with labour costs, then further problems arise. One can either increase the price of output by increasing the physical output per hour (while keeping wages the same) (which often leads to a falling price per unit), or one can reduce costs while keeping output the same in physical and monetary terms. One way to do this is to reduce the number of workers while increasing labour intensity, or to reduce wages while keeping the number of workers the same. While in the first case productivity was increased by increasing output qua a given unit of labour measured in time, in the second case the quantity of output was increased relative to the cost of labour (not in physical units). Here, when output and input are expressed in monetary terms, labour productivity can apparently increase without a change in the production processes in the technological structure, simply by reducing labour costs.

When measuring productivity through physical outputs, the question again arises whether different types of output can be compared with each other, which is why monetary terms are used in the statistics when productivity in different sectors is measured and compared with each other. In turn, certain activities that produce an output that is consumed, but without having a market price, such as the reproductive work of women in the family, are not included in the statistics as economic activities. On the other hand, there is again a production of services without use-value production such as security guards. which facilitate the sale and purchase of other goods.

Smith does not consider here methods of production in which, with each increase in productivity, there is precisely no even greater increase in the technical composition of capital, which is accompanied by an increase in the savings rate of the capitalists; on the contrary, here the rate of growth of productivity will actually increase more than the percentage increase in the technical composition of capital, with the composition of labour employed remaining constant or falling. Now it is possible that even when the value composition of capital remains constant or falls, productivity increases, and this is not primarily as the productivity of labour, but rather this is due to the degree of efficiency in the application of constant capital, which is called capital productivity.

As an unproductive sector, Smith cites finance, which essentially reallocates already existing capital. We will formulate some objections to this, especially in the next section. First of all, our own definition of productivity. We can only briefly note here that in Marx’s understanding not every form of wage labour is considered “productive”, but only that which actually produces surplus value. Marx wrote the following in the Theories of Surplus Value: “An actor, for example, even a clown, is a productive worker if he works in the service of a capitalist (the entrepreneur), to whom he returns more labour than he receives from him in the form of a salary, whereas a tailor who comes to the capitalist’s house and mends his trousers, creating for him a mere use-value, is an unproductive worker. The labour of the former exchanges itself for capital, that of the latter for revenue. The former creates a surplus-value; in the latter a revenue is consumed.” (MEW 26.1: 127) Thus the labour of a clown, who works for the private pleasure of the capitalist and is paid out of his revenue, is indeed to be assessed as unproductive, and only when the clown, for example, performs in the circus Sarrasani, is the capitalist paid out of his revenue. In the Sarrasani Circus, for example, when the clown works in front of spectators and the spectators pay an entrance fee that brings the owner a profit, only then does the clown’s work transform into surplus value-generating and thus productive labour, without any concrete or qualitative change in comparison to the work of the clown who works for the private pleasure of the capitalist. Unproductive labour, on the other hand, does not contribute to the production of surplus value, but is paid out of surplus value as so-called revenue and thus reduces the potency of capitalist accumulation. In this context, Marx accused Adam Smith, for example, of constantly mixing the “material determinations of labour” with its social form determinations. (Ibid.: 122/127)

For Lohoff/Trenkle, from the 1980s at the latest, financial capital is the motor for the expansion of global commodity production, which had already been taking place since the 1960s at a high level of productivity and progressive process automation. The authors speak of “induced value production” or of “inverse capital” because value production is not based on the extraction of surplus value through the use of labour power, but is driven by the growing accumulation of fictitious capital, indeed the entire global value production today is induced by fictitious capital. For without fictitious capital, the functioning capital (the capital invested in the “real economy”) would have had to enter into an intensifying cycle of mass devaluation.

Lohoff/Trenkle have also attempted to provide empirical support for their theses. First, they point to the well-known figures. In 2010, the Global Wealth Report put financial assets (excluding derivatives) at 231 trillion dollars, four times the current global GDP. The total volume of derivatives grew from $72 trillion to $673 trillion between the years 1998 and 2008 – reaching twelve times the global GDP. Lohoff/Trenkle find an important indicator pointing to the hegemony of financial capital in the economic relations between the USA and China, commonly referred to as “Chimerica”. The steadily growing trade and current account deficit of the USA since the 1980s (higher import of goods and services than their export), which was financed by the inflow of capital from abroad, corresponds to a surplus in the capital account (increase in capital imports). In the process, the USA has increasingly imported goods, especially from China, while its financial markets have increasingly sold securities and derivatives to foreign capital investors (fictitious/speculative capital). Industrial products from China, but of course also from other export-oriented countries (such as the FRG) were sold to the USA, while at the same time there were purchases of all kinds of securities in the USA. Chinese private investors and Chinese sovereign wealth funds also invested their profits in the US capital markets, especially until the financial crisis of 2008, i.e. they bought securities and derivatives “produced” in the USA, and this especially since the end of the 1990s, at a time when China was once again making a big leap forward in terms of its economic growth rates. China has also driven financialisation in its own markets, with total debt (government, financial, industrial and private sectors) growing from 153 per cent of GDP in 2008 to 282 per cent today, with capital flowing mainly into the construction industry and infrastructure development. In the next section we continue this discussion.

In contrast to Smith, we consider the financial sector to be a productive sector in the wake of Greek economist John Milios. An intermediate position is taken by Tony Norfield in his book The City, in which he shows that sui generis capitalist enterprises need the financial sector to conduct their business, although they play no role in current production for commodities and their profits are therefore only a proportion of the total value production that would take place mainly in the productive industrial sector. The metaphor “central nervous system of capital” used by Tony Norfield in his book The City to characterise the current financial system accurately suggests the development of capitalist economies. If the capital principle is the engine of the breathing monster called total capital, then the financial system is its brain and central nervous system. Randy Martin has pointed out in this context that the financial system is intrinsic to the three volumes of Marx’s Capital in that it performs an important function for the reproduction of capital in the movements of production and circulation and the related need to anticipate risk. The financial system executes to a not inconsiderable extent the competition, coordination and regulation of enterprises (in all sectors), which in turn are presupposed by the a priori of total capital, which is actualised through the real competition of enterprises, which for Marx is always not a ballet but a war. Financial capital constantly modulates and reignites the competition of all enterprises – it is therefore an integral part of the economy of capital, not a cancer that a doctor removes to restore the body of capital to health.

For Norfield, the operations of the financial system are by no means limited to the multiple strategies of banks, investment funds and other financial institutions; rather, they affect the capitalist system and its enterprises as a whole, insofar as industrial and commercial enterprises must also constantly carry out a variety of financial transactions. Thus, internationally operating companies use private banks to obtain the currencies they need to buy imported goods or to exchange the profits derived from their export transactions into domestic currency. Companies borrow short-term from private banks to secure their cash flows, or they take out longer-term loans to finance their investments. They issue bonds or shares in the financial markets to raise money from investors, and they use derivatives to hedge against adverse movements in interest rates that limit their profitability. For example, purchases of raw materials, it systems, buildings, machinery and labour made to produce new goods to be sold at a profit can be reduced by pending interest payments. And the net profits of industrial enterprises are affected by all sorts of financial transactions, from hedging currencies to interest rate risks, especially when the enterprises themselves invest in financial collateral. The financing of capitalist production and circulation is a crucial aspect of the reproduction of capital on an extended ladder.

Let us now quote John Milios on this subject at length (deepl translation): “According to this approach (Ricardo), profitability in capitalism can be derived through two different routes: a productive one (M-C-M′, where M stands for money, C for commodities and M′ = M+ΔΜ) and a parasitic or speculative one (M-M′′, with M′′=ΔΜ′). Ricardo, of course, never came to this conclusion or categorisation. Nonetheless, the latter can be seen as a direct consequence of his reasoning, if this reasoning is extended to cover the developments of contemporary finance.In this respect, it can be suggested that when the absentee owner becomes dominant in the organisation of capitalist life, the ‘productive’ aspects of the latter are displaced and speculative and predatory activities move into a dominant position. This would be the case because the dominant motive of capitalism would amount to the pursuit of profit in the sphere of financial circulation, i.e. the appropriation of profit created by other fractions of (productive) capital or even from the income of (productive) workers. Circulation becomes the main means of absorbing the profit previously generated by production; all this would lead to stagnation and instability in the production of use-values….

In the context of capitalist economic and social relations, the movement of money as capital binds the production process to the circulation process: commodity production becomes a phase or moment (and indeed, for the whole valorisation process, the decisive moment) of the circulation of social capital: M-CMLPp … P … C′ … M′(1.1) The capitalist appears on the market as the possessor of money M and buys commodities C, consisting of means of production Mp and labour power Lp. In the production process (P), these commodities C are productively consumed to produce an output of other commodities, a product C′, the value of which is supposed to exceed that of C. The capitalist then sells this output to a third party. Finally, it sells this output to obtain a sum of money M′ that is higher than M….

Finance “creates costs”. It uses labour and means of production to create and sell certain (suigeneris) goods (exchange values that are also use values for others). In other words, financial intermediation can take different forms and involve different types of institutions, but in any case it is associated with a certain set of financial services, which are in fact capitalist commodities. We will not get into the discussions about the functions of financial intermediaries. But in general, they mediate the investment process under certain conditions that follow the institutional trends of capitalist economies.3 This mediation is itself a service and therefore a productive activity that seeks to maximise profit, like any other sector of the capitalist economy. It is therefore considered a productive activity.” (Dimitris P. Sotiropoulos, John Milios, and Spyros Lapatsioras: A Political Economy of Contemporary Capitalism and its Crisis)

Our argument in Capital and Power in the 21st Century is different again. In contrast to Milios, we point out that the financial enterprises, as opposed to the industrial sector, engage in their own form of profit-making, which is nonetheless necessary for capital as a whole: leveraging, credit, derivatives, etc. Alan Freeman writes here (translation deepl): “We can understand neither the crisis nor Marx’s theory of the rate of profit if we abstract from Marx’s concept of finance. More precisely, money is capital. This may seem obvious when put like that, but it is missing from almost all Marxist writings on the rate of profit.” All this, of course, has implications for the question of how rates of profit can be read out of official statistics, which unfortunately we cannot explore in more detail here. Incidentally, Freeman also assumes a falling rate of profit in the USA since the 1980s.

If the private banks today take on a supporting and at the same time driving role in the allocation of monetary capital for the various industries and sectors of the economy, then it is precisely some important differences between financial capital and industrial and commercial capital in terms of profit mechanisms and economic power that must be taken into account. There is definitely no uniform mechanism that leads to the equalisation of the rates of return of the financial enterprises with the rates of profit of the industrial and commercial enterprises, or, to put it differently, the profits of the private banks do not have a strictly lawful relation with the investments and the profits of the industrial capital. It is not that the profits of the private banks cannot be calculated at all, or that they have no relation at all to the profits of the industrial enterprises, but there is no mechanism of equalisation of the rates of profit between the private banks, as is found in the industrial sector, nor is there any integration of the bank profits in the processes of producing the (industrial) average rates of profit. To prove this, one needs to examine the various forms of investment and speculative activities of private banks a little more closely. The financial capital advance necessary for industrial enterprises to buy technology, machinery, energy, raw materials and buildings and to hire labour does not have the same crucial importance for financial enterprises when it comes to generating their own profits. The returns of private banks are not primarily based on the use of trading screens, buildings, software and labour. Fixed and circulating capital have relatively little to do with the credit creation and capital mechanisms by which private banks create their profits in real terms, because, unlike industrial companies, they can create fiat money themselves and produce fictitious/speculative capital at a profit, and therefore have much higher leverage than industrial companies.

But financial operations does not mean growth in investment. Let us cite in length therefore Jan Toporowski: “Large corporations, which
account for the vast bulk of private sector investment, now have excess
capital and engage more in balance sheet restructuring (buying and
selling financial assets; issuing and repaying liabilities). Such restruc
turing leaves corporations with larger risky financial market exposures,
which therefore require the holding of greater amounts of liquid assets
(short-term deposits, holding of securities). If a company finds itself with
too many liquid assets, profits can be immediately increased by using the
excess liquid assets to repay debt. Indeed, this is a far more certain way
of raising profits than the prolonged and uncertain business of investment
in plant and equipment. Industrial regeneration is a dream of engineers,
from which companies are awoken by their finance directors to face the
irrefutable realities of balance sheet restructuring as the only financially
viable way forward for all companies.” (Jan Toporowski: Credit and Crisis from Marx to Minsky: 129)

After this lengthy discourse, let us return to Smith’s text and in particular to the chapter on “Circulation and Control”, which first deals with investment rates in the productive sector (machinery, computers, factories, research, etc.), which have been persistently stagnant in the USA since the 1980s. Although there is a moderate increase in productivity in the 1990s due to the computerisation of accounting, warehousing and logistics, there is no increase in the new millennium due to e-commerce and smart technologies. Smith attributes the decline in investment rates to the fall in average profit rates across the economy, assuming that the profits generated in the financial sector are part of the profits of the “value-producing” bussines (industry and partly service).

Smith then briefly summarises Marx’s law of the tendential fall of the rate of profit. We refer to our detailed summary of the law in Capitalisation Vol.1. Smith considers work associated with control and supervision of the processes of production to be unproductive. Work in circulation, in turn, includes that of the realisation of commodities, a wide range of transport, distribution, security and insurance, finance, law and accounting, sectors which Smith attributes to the growing service sector, which for him is largely unproductive because it does not produce products or services that can be sold for profits. If this is largely true for government activities, it is far from true for Finance and other service sector enterprises that do indeed realise profits. See the discussion above on unproductive labour.

For Smith, an increase in the rate of surplus value must compensate not only for the reduction of labour relative to capital (qua technological innovation), but also for the increasing expenditure in circulation, with wages generated here drawn from surplus value generated in the productive sector. He subsequently refers to Paul Mattick, who writes that the tendency towards lower productivity increases in circulation exerts a pressure on profit rates, whereby the role of circulation labour in the total circulation of capital would be that, on the one hand, it is necessary for the realisation of commodities, but on the other hand, it does not directly produce value. Here Smith asks why the growing disproportion between rising productivity in the productive sector and stagnating productivity in the service sector would not be reduced by increasing automation of the latter sector, and he answers, following Mattick, that the production processes are concentrated in a few increasingly large enterprises, while the service sector is highly decentralised. The economy of large-scales in industry is contrasted with the labour-intensive activities in circulation.

Smith mentions here the Marxist economist Fred Moseley, who since the 1970s has noted a moderate increase in productivity in the USA with falling real wages, which has atypically led to a fall in the rate of profit. When the rate of surplus value rises more than the organic composition of capital, the rate of profit rises. Smith mentions falling costs of constant capital, falling wages and higher productivity as counter-tendencies against falling rates of profit. Moseley claims that these three factors have been present in the USA since 1980, but that there has nevertheless been a fall in the rate of profit due to a rising rate of unproductive labour compared to productive labour, which Smith qualifies insofar as there is apriroi no reason for the quantity of labour in circulation to increase compared to that in production and, moreover, it must be questioned why productivity in circulation is lower than in production. The reason Smith then gives is that there are inherent difficulties in the growth of automation due to the multiple person-to-person transactions in circulation, such as the highly automated automobile industry, where sales activities are still highly personalised. The substitution of labour in the areas of circulation (security, transport, warehousing, delivery, etc.) has so far been limited, but today it may extend to activities of control and supervision of production processes, which are responsible for increasing the labour intensity in the production processes. Today, this is forced by electronic and biometric methods of measuring and monitoring work. It is more about the automation of high-paid control activities than about the activities of low-paid workers in circulation. In this context, Smith mentions the use of monitoring technologies such as smart sunglasses in production, which would be able to observe and measure the performance of workers on a minute-by-minute basis.
In the next section, Smith discusses service sector jobs in more detail, noting that 6.5 million low-paying jobs have been created in the US since the recession, especially in the food industry, the nursing sector and the private service sector. The health sector is also facing strong expansion. We see an overall increase in low-paid, low-skilled jobs in developed countries, while innovation has tended to take place in administrative and routine office work.

Smith distinguishes in this context between tasks and jobs that combine multiple tasks, either repetitive or intuitive. Thus, automation would replace tasks rather than jobs, that is, it would produce a restructuring of the previous division of labour rather than a factual replacement of activities. While automation in the product area leads to falling costs of the product and probably to an increase in demand, i.e. the product is price-elastic (correlation between price and output) and thus the technological replacement of labour can be compensated by higher demand, there are also saturation effects to be noted here. Smith cites the cheapening of foodstuffs induced by new technologies in the agricultural industry, whereby the absolute quantity of food consumption has not kept pace with innovation, while at the same time consumption in the service sector, including the catering industry, has increased. Service jobs such as care, education, restaurants are income elastic, meaning that as incomes rise, so does the demand for them. If employment is in low-paid and low-productivity service jobs, then competition and oversupply among workers will lead to further wage pressure.

A fully automated economy through all sectors is not possible for Smith in an economy characterised by wage labour, as there is always an oversupply of cheap labour (Marx’s reserve army) that makes full automation unnecessary, rather it drives labour productivity in one sector while stagnating labour productivity in another. And when products and services become cheaper due to productivity growth in one sector, the demand for labour-intensive services in another sector increases.

Yet it remains generally difficult to automate some low-paid work, which requires intuition, understanding of social norms, certain linguistic skills and insight into imponderables. This is different from the automation of repetitive tasks as in Taylorism. At the same time, the concept of skills is quite elastic and thoroughly social. In many cases, activities are not automated precisely because there are enough workers on the market, which lowers labour costs and discourages investors from investing in expensive machines.

For Marx, the value of the commodity labour power is determined by the costs of reproduction, that is, by the socially necessary number of goods and services to reproduce it day after day according to certain social conventions, which precisely includes the costs of education, know-how and qualification (measured in time and money), which in turn secures a certain place in the hierarchy of the wage scale. Nevertheless, a whole range of highly skilled activities that involve high educational costs are unproductive in the Marxian sense, since they do not produce surplus value, and rather than representing pure wage costs, are paid for out of the profits of firms.

In the last section, Smith renews the thesis that a surplus of available labour not only pushes down wages, leading firms not to invest in expensive machinery, but that the excess of labour prevents automation in one sector precisely because of the excess of automation in another. In this context, he mentions Marx’s thesis, which he himself called an absolute law, that for a given physical output, machines reduce the quantity of living labour necessary to produce that output. This is not a special insight unless Marx also assumed that automation in one sector leads to an increase in jobs in another sector. Thus, the automation of one sector of industry leads to the expansion of infrastructure, the cultivation and extraction of more raw materials and the construction of more machinery. In addition, there are sectors such as marketing, law and communication. Thus, high productivity in industry leads to a whole army of workers being driven into sectors with low productivity, in Marx’s sense, into unproductive sectors. Nevertheless, the absolute law of capitalist accumulation provides a limit to the growth of demand for labour. The total output of an industry will gradually grow at a lower rate. In this context, Marx already addressed the existence not only of an industrial reserve army that remains in the labour market, but of a surplus proletariat that is pushed out of the official labour market altogether and has to work in the so-called informal sector with low productivity. Smith does not elaborate on this discussion in his book.

We have raised our objections regarding the neglect of capital productivity, the assessment of the financial sector and unproductive labour. In this context, official statistics would have to be re-read, if Marxist ratios such as the rate of profit can be derived from these statistics at all, which is doubtful. However, we agree with Smith that the theses of the automation theorists do not correspond to reality in large parts.

translated by deepl.

,

Scroll to Top