The election of Donald Trump as the 45th president of the United States, combined with the rise of similar right-wing demagogues in Europe, has prompted a discussion about the cause of the decline in the number of relatively high-wage, “middle-class,” unionized industrial jobs in the imperialist core countries. One view blames globalization and bad trade deals. The European Union, successor to the (West) European Common Market of the 1960s; the North American Free Trade Area; and the now aborted Trans Pacific Partnership have gotten much of the blame for the long-term jobs crisis.
This position gets support not only from President Trump and his right-hand man Steve Bannon and their European counterparts on the far right but also much of the trade-union leadership and the “progressive” and even socialist left. The solution to the problems caused by disappearing high-paid jobs in industry, according to economic nationalists of both right and left, is to retreat from the global market back into the safe cocoon of the nation-state. Economic nationalists insist that to the extent that world trade cannot be entirely abandoned, trade deals must be renegotiated to safeguard the jobs of “our workers.”
Most professional economists have a completely different explanation for the jobs crisis. They argue that changes in technology, especially the rapid growth of artificial intelligence in general (1) and machine-learning in particular, is making human labor increasingly unnecessary in both industrial production and the service sector. Last year—though it now seems like centuries ago—when I was talking with one of this blog’s editors about possible new topics for future blogs, a suggestion was made that I take up a warning by the famous British physicist Stephan Hawking that recent gains in artificial intelligence will create a massive jobs crisis. This is a good place to examine some of the subject matter that might have been in that blog post if Brexit and Donald Trump had been defeated as expected and the first months of the Hillary Clinton administration had turned out to be a slow news period.
It is a fact that over the last 40 years computers and computer-controlled machines—robots—have increasingly ousted workers from factories and mines. The growth of artificial intelligence and machine learning is giving the “workers of the brain” a run for their money as well. This has already happened big time on Wall Street, where specially programmed computers have largely replaced humans on the trading floors of the big Wall Street banks. No human trader can possibly keep up with computers that can run a complex algorithm and execute trades based on the results of the computation in a fraction of a second.
Wall Street traders are not the only workers of the brain whose jobs are endangered by the further development of AI. Among these workers are the computer programmers themselves. According to an article by Matt Reynolds that appeared in the February 22, 2017, edition of the New Scientist, Microsoft and Cambridge University in the UK have developed a program that can write simple computer programs.
He writes: “Created by researchers at Microsoft and the University of Cambridge, the system, called DeepCoder, solved basic challenges of the kind set by programming competitions. This kind of approach could make it much easier for people to build simple programs without knowing how to write code.”
The ability of “DeepCoder” to create programs is very limited today. However, Reynolds indicates that in the future the successors to DeepCoder’s ability to write programs will increase enormously. “Ultimately,” Reynolds quotes Marc BrockSchmidt, one DeepCoder’s creators, “the approach could allow non-coders to simply describe an idea for a program and let the system build it.”
“The potential for automation that this kind of technology offers,” Reynolds quotes MIT computer scientist Armando Solar-Lezama, “could really signify an enormous [reduction] in the amount of effort it takes to develop code.”
In economic terms, an enormous reduction in the amount of effort means that companies like Microsoft will be able to spend less money hiring programmers to do “routine” coding tasks that are today done by human programmers. Everything remaining equal, the development of this AI technology capable of writing computer programs will radically reduce demand for computer programmers, just like factory mechanization and automation have been reducing the demand for factory workers since the industrial revolution.
Machine learning is actually a catch-all term that refers to a variety of computer programming techniques. Instead of having the computer follow a fixed series of instructions like is the case with traditional computer programs, the program modifies the instructions it executes in response to its data. As a result, the computer and the “robots” it controls seem to learn, much like an animal or even a human, through practice and experience.
For example, if a computer making trades for a Wall Street bank makes a “bad trade,” it will “learn” to avoid making the same mistake again. If the computer wins big, it will learn from that as well, much as a human trader would. Machine learning is not confined to Wall Street trading. Anybody who uses the Internet experiences the effects of machine learning every day. The computers that feed advertisements to your computer browser program “learn” from the websites you look at what your interests are and base advertisements they send accordingly.
And now computer-controlled automobiles are learning to navigate crowded highways. It is widely predicted that in the near future truck driver and taxi jobs will be replaced by self-driving vehicles controlled by computers that have “learned” and continue to “learn” to drive better and safer than human drivers can.
As a result of the rapid progress now being made in machine learning, it is argued by some that in the not-too-distant future there will be virtually no task now performed by humans that machines will not do better and faster.
Science fiction scenarios
Some computer scientists predict and indeed fear that within a few decades computer-controlled machines will have become more intelligent than humans in every sense of the word. As a result, humans will be replaced not only in the sphere of work but in every other area of human activity. The best that we flesh-and-blood humans can hope for is that the information encoded in our brains will be translated into huge strings of 1’s and 0’s that will be downloaded into computers. Our minds will migrate from the biological entities in which they developed into machines. In that sense we will “live” on. That’s the best case.
In the worst case, the intelligent machines in the spirit of Frankenstein’s monster will decide on their own to purge the world of the very humans that brought them into existence. It should be pointed out that many computer scientists think these scenarios are pure fantasy, while others are on the fence.
Even if we leave aside the science fiction-sounding scenarios of “hard AI,” where machine intelligence replaces humans entirely, most economists agree that flesh-and-blood workers have little future in industrial production. They argue that even if President Trump and his fellow economic nationalists succeed in vastly increasing the share of global industrial production being carried out within the U.S.—and other imperialist countries—the “middle-class” industrial jobs of the 1950s and 1960s will not come back. Assuming that new factories spring up in the “rust belts” of the U.S. and Europe, these economists argue, they will be filled with intelligent machines with hardly a human worker in sight.
Not a new process
The process by which factory workers are replaced by increasingly powerful machinery is hardly new. It has been going on since the industrial revolution that began in England and Scotland in the last half of the 18th century. By the early 19th century, fears were being expressed that the growth of steam-powered mechanization would soon render most of the working class redundant. In Britain, movements arose among workers called Luddites, led by a mythical King Lud. The Luddites attempted to destroy the new factory machinery taking their jobs and rendering their skills economically useless. While some machines were destroyed, the movement was futile, since there was no stopping the creation of ever more job-destroying machinery, which has continued right down to the present.
The (bourgeois) economists of the time argued, however, that things were not as grim for workers as they appeared. They claimed that while ever more powerful machinery would indeed eliminate some jobs, the machines were actually creating more jobs than they were destroying. The future of the working class under capitalism was bright because the new jobs that were created by the progress of science and technology would be far more pleasant and better paying than the jobs eliminated.
The economists catch up with Ricardo
Among the economists arguing that the growing mechanization was good for workers was David Ricardo. However, upon thinking about this question more carefully the great English classical economist, who was to exercise great influence on Marx, reversed himself. Ricardo came to the conclusion that there was no law that dictated that increasingly powerful machines would create more jobs than they were destroying. The truth-loving Ricardo admitted that the introduction of the process of replacing workers with machines could indeed work against the interests of the working class.
In the last few years, with a lag of only 200 years, our economists in the age of Trump, who so valiantly defend “free trade and globalization,” seem to be finally catching up with Ricardo. These economists insist that industrial jobs of the past are gone for good and that many other types of jobs are doomed to disappear as “intelligent machines” begin to take over from the “workers of the brain.” They seem to have quietly dropped their claim that powerful machines create more and better jobs than they destroy.
What are the limits of automation?
There are really two aspects to the question of how far the replacement of workers of both the hand and brain by machines can go. On one side, the question is a matter of science and engineering. Can “hard AI” ever be achieved, and if the answer is no as many computer scientists believe, what will be the limit beyond which no “intelligent machine” can go? The other side of the question is economic. What are the economic limits on automation that are imposed not by science or technology but by the capitalist system? Though economists may deny it, these are two very different questions.
What is capital?
From the viewpoint of our modern economists, capital is simply another name for the the means of production. The very primitive stone tools made by our pre-human ancestors more than 3 million years ago are an early form of “capital,” they argue. The accumulation of capital, from primitive stone tools right up to the CPUs—central processing units—that form the “brains” of our intelligent machines that are increasingly capable of learning from their own mistakes, are all examples of capital. Even specialized skills—as opposed to the general ability to work possessed by all normal human beings—are also, according to these economists, examples of “capital.”
According to the economists, our pre-human ancestors of 3 million (2) and more years ago—as they learned from the previous generation to make what to the eyes of their remote descendants were crude stone tools—were accumulating human, or in this case perhaps we should say “pre-human,” capital. (For post that discusses “human capital,” see here.) If we go by this definition, the limits of capital are identical to the limits of human skill and its highest manifestations in science, engineering and technology.
In contrast to the bourgeois economists, especially those of the various marginalist schools that have dominated the profession since the late 19th century, Marx defined capital not as a thing or collection of things but as a social relationship of production. This relationship has dominated production only during a specific and relatively short period within what we now know as the more than 3-million-year history of production. In the not too distant future, this relationship is doomed to disappear, because it is becoming a greater and greater hindrance to the further development of production.
A century ago, during the Russian Revolution, Lenin defined socialism as soviet power—the political power of the working class—plus electrification. At the time Lenin coined this slogan, electricity was replacing steam power as the main motive force in industrial production. (3) A century later, we might define socialism as the working class in power plus artificial intelligence and machine learning.
Many visionary bourgeois figures, including the famous U.S. industrial capitalist Elon Musk, foresee a situation arising in the near future where there will be so few jobs that society will be forced to provide an income for all people. Musk writes: “There will be fewer and fewer jobs that a robot cannot do better. I want to be clear. These are not things I wish will happen; these are things I think probably will happen. And if my assessment is correct and they probably will happen, [then] we have to think about what are we going to do about it. I think some kind of universal basic income is going to be necessary.”
If this happens, Musk worries it will no longer be possible for a class to exist that has a monopoly ownership of the means of production while another class exists that is forced to sell its labor power to the owners of the means of production. Instead, the fruits of production will have to be distributed to the population more or less according to their needs. Economic distribution according to human needs will not simply become an economic possibility, it will be an economic necessity.
Whatever such a society will be, it will not be capitalist. Musk in his own way, though we can assume he is not a communist, is admitting, though he really doesn’t wish it, that the end of capitalism is now in sight.
What is capitalist production?
Capitalist production is a process of the accumulation of capital. Capital is a social relationship of production that is quantitatively measured in terms of value—accumulated abstract labor embodied in the means of production. Since value is measured in terms of abstract human labor measured in terms of some unit of time, capital itself is measured in terms of human labor measured in terms of time. However, this cannot be done directly. Instead, value that is embodied in capital, just like the value that is embodied in a simple commodity, must take the form of exchange value—money—which is quantitatively measured in terms of some unit of gold bullion—or other money commodity—which in turn is measured in terms of some unit of weight. In everyday language, this means that capital is measured in terms of dollars and cents, though only a small portion of the total social capital consists of actual money.
The class that owns no means of production besides its ability to work—labor power—sells its labor power to the class that has a monopoly on all other means of production. Once the workers have sold their labor power to the capitalists, the capitalists own all the means of production.
Under the “wages system”—another name for capitalism—the workers sell their labor power for sums of money called wages. During part of the work day, the workers reproduce the value of their wage but then must perform additional unpaid labor—that is, produce surplus value. (4)
Even if we assume that the capitalist pays the workers the full value of their labor power (5)—because the workers are fully protected by their union—the capitalist pays for only a part of the labor the workers perform. If all labor were paid, the capitalist would have no incentive to buy the workers’ labor power. This is the reason that no matter how powerful and well organized the unions are, they cannot end the exploitation of the workers by the capitalists as long as the capitalist system lasts. At best, they can end the super-exploitation of the working class where the capitalists fail to pay the full value of the workers’ labor power.
The three forms capital assumes within circulation
Capitalist production is a unity of production and circulation. If there is no capitalist circulation, there is no—at least not for very long—capitalist production. In the course of its circulation, capital takes on successively three forms. These are (1) money capital, (2) productive capital—the means of production, including the purchased labor power of the workers, and (3) commodity capital, the finished products before they are sold. A common error made by Marxist writers is to confuse commodity capital—inventories in everyday business language—with the means of production the capitalist uses to produce the finished product.
It is true that industrial capitalists have to purchase the means of production—except for labor power, which they must purchase from workers—from other industrial capitalists or from merchant intermediaries. From the viewpoint of the sellers of the means of production, unsold finished products, which in their material use value later on after they have been sold are destined to function as productive capital, are commodity capital. However, from the viewpoint of their industrial capitalist purchasers, these same commodities once they have been purchased and put to use are productive capital.
Industrial capitalists can only complete the cycle of (re)production by selling their finished commodities at profitable prices. If this cannot be done, the whole process of capitalist circulation and production—reproduction—comes to a screeching halt. When industrial capitalists do successfully sell their commodities at their prices of production, they have successfully transformed their commodity capital into money capital. Only then have the capitalists made a profit. The industrial capitalists are then in a position to start another cycle of (re)production on an enlarged scale.
The forms of productive capital
Marx called the part of productive capital that preserves its value—though not its use value—in the process of production constant capital. Constant capital, however, does not create an atom of additional value, and therefore it creates no surplus value. Physical examples of constant capital include raw and auxiliary materials, factory buildings and machinery, farm machinery, seed, animals, growing plants, fertilizers, mines and mine machinery. Variable capital, or sold labor power of workers, does create value. Variable capital both replaces the value of money capital that industrial capitalists use to pay their workers and creates an additional value, a surplus value, which is the value produced by the unpaid labor workers are obliged to perform for the capitalists buying their labor power.
A portion of the surplus value is then converted into new productive capital. According to Marx, this portion of the surplus value converted into new constant capital grows at a faster rate than the portion converted into variable capital. Therefore, in absolute terms on a global basis the number of workers employed in industrial production—broadly defined—grows absolutely but shrinks in relation to the growing scale of production.
This does not preclude that even outside of crises the absolute number of industrial workers might shrink in particular branches of industrial production and even in particular countries. However, an absolute decline in the number of industrial workers worldwide—and here industrial worker is defined broadly as all workers that produce surplus value—must grow if capital as the fund of accumulated abstract human labor is to grow.
The general law of the growth of industrial production, employment and the rate of surplus value
The process by which productive workers are replaced by machines has an upper bound set by the growth of science and technology. Not every task can be mechanized at a given level of technology—at least not yet. It also has a lower bound, set by the profit needs of the capitalists. The industrial capitalists are not interested in fully utilizing existing science and technology to replace as many workers as possible. Rather, they are only interested in realizing the highest possible rate of profit on their capital.
Therefore, at any given level of science and technology industrial capitalists have a “choice of technique” to use, in the bloodless terminology of the economists. However, this choice of technique is not arbitrary, since the industrial capitalists must always choose the technique that yields the highest rate of profit.
The workers find themselves in an increasingly desperate competition with ever more powerful machines. The workers can only win against these machines by agreeing (in effect) to work for the capitalists for free for an ever greater percentage of the working day. The general law of industrial production and employment under the capitalist mode of production, therefore, is that both industrial production—outside of crises—and employment must grow on a global scale, though not necessarily in every branch of industry or even in every country. But industrial production grows faster than industrial employment.
In the competition between machines and workers, the workers must do well enough to keep their number growing absolutely on a global basis. The more capitalism and the productivity of human labor grows, the higher the rate of surplus value can grow, because a given amount of value represents more material use values. The rate of surplus value must rise so that the workers remain sufficiently in the running in their competition against the machines so that the value of the total capital keeps growing.
Therefore, the growing chasm between the profits, not to speak of the accumulated wealth, of the largest capitalists, on one hand, and the rest of the population, on the other, so obvious today, is not an unfortunate accident but the inevitable result of the economic laws that govern the capitalist system. These laws can be modified for periods of time by the class struggle, but they cannot in the long run be negated as long as the class struggle does not lead to the overthrow of the capitalist system.
The tendency of the rate of profit to fall
The general law of the growth of industrial production, employment and exploitation is expressed by the tendency of the rate of profit to fall. In developing this law, Marx was not interested in temporary effects on the rate of profit brought about by problems involving the realization of value and surplus value in money form. These problems are very real and periodically find expression and resolution in the periodic crises of overproduction.
Marx considered the falling tendency of the rate of profit to be the most important law of political economy because it points toward the inevitable end of the rule of capital over production. Capitalist production is production for profit. Capitalists as the necessary agents of capitalist production are—and must be—obsessed with the rate of profit on their investments. However, the very development of capitalism undermines the rate of profit. The law of the falling tendency of the rate of profit implies that capitalism must with the further growth of production, and growth in the productivity of human labor, itself driven ever forward by capital itself, come to an end. These were Marx’s revolutionary conclusions.
Marxist economists are actually divided on Marx’s law of the tendency of the rate of profit to fall. Many Marxists over the decades have disagreed strongly with Marx on this point, but other Marxists have strongly defended Marx. Among the strongest contemporary defenders of Marx’s law of the tendency of the rate of profit to fall is Anwar Shaikh. Perhaps the most important Marxist critic of the law of the tendency of the rate of profit to fall was Paul Sweezy, the founder of the Monthly Review school.
Though Sweezy died in 2004 after a long and productive life, his successors at Monthly Review have actually hardened their position in opposition to the law that Marx considered the most important of all in political economy. In his “Theory of Capitalist Development,” Paul Sweezy rejected Marx’s law of the tendency of the rate of profit to fall on grounds that it was indeterminate whether the rising organic composition of capital—which all else remaining equal lowers the rate of profit—would be offset by the rising rate of surplus value, which all else remaining equal raises the rate of profit.
Other critiques of Marx’s law of the tendency of the rate of profit to fall have attacked the claim that capitalist development actually causes the organic composition of capital to tend to rise. These critiques have claimed that technological innovations are just as likely to be (constant) capital saving as (variable capital) labor saving.
While nobody can deny the growing role of machines compared to human labor—referred to as the technical composition of capital since the latter half of the 18th century—it is also true that the value of machines and raw and auxiliary materials that make up that constant capital are themselves lowered as the productivity of human labor advances. These critics claim, therefore, that the direction of the evolution of organic composition of capital is itself indeterminate and with it the direction of the rate of profit. (6)
Okishio theorem
Another well-known attack on Marx’s theory of the tendency of the rate of profit to fall involves the Okishio theorem, put forward by Japanese economist Nobuo Okishio. Okishio observed that no capitalist would knowingly choose a method of production that would lower the rate of profit. This is, of course, true. Therefore, the only thing that would lower the rate of profit, according to the Okishio theorem, is a rise in real wages. The Okishio theorem has never been very popular, to say the least, among Marxists who support Marx’s law of the tendency of the rate of profit to fall.
From a Marxist point of view, there are many problems with Okishio’s argument. Marx never attempted to show that the rate of profit would fall if real wages remained unchanged or even fell. Instead, in his demonstration of the law in Volume III of “Capital,” Marx held wages in terms of value—not real wages—steady. He showed that if the organic composition of capital rises while the value of wages and the rate of surplus value remains unchanged, the rate of profit falls.
Marx sharply distinguished between the rate of surplus value—the calculation of the mass of profit over the total variable capital—and the rate of profit, which measures the mass of profit over the total capital. Therefore, Marx and Okishio are talking about different things. Indeed, since a rise in the organic composition of capital implies a rise in the productivity of labor, Marx’s demonstration of the falling rate of profit implicitly assumes a rising wage in terms of use values—that is, a rising real wage. So in a certain sense, Okishio and Marx are talking past one another.
This, however, is not the question that interests Shaikh in his “Capitalism,” though he does take note of it. The reason is that Shaikh’s book is about capitalist competition. Marx had intended to write a book that would specifically deal with competition, but as far as we know it was never written. Shaikh’s “Capitalism” therefore, unlike Marx’s “Capital,” (7) is a book about capitalist competition. In it, Shaikh contrasts “perfect competition” of marginalist orthodoxy with what Shaikh calls the “real competition” of the “classical economists”—and, more importantly, the real world. Shaikh attacks the Okishio theorem on grounds that it assumes “perfect competition.” Shaikh demonstrates that once the theory of “perfect competition” is dropped in favor of the “classical” theory of real competition, Oksihio’s so-called theorem falls to the ground.
A note on terminology
Marx used very precise terminology. Unfortunately, as a rule Shaikh in “Capitalism” does not use Marx’s precise terminology—though he does in places. Often, Shaikh uses the vaguer terminology used by his fellow economists that often covers up as much as it reveals.
For example, bourgeois economists talk about capital and labor as “factors of production” as opposed to constant capital that preserves its value and variable capital that alone produces value and surplus value. The economists, as I explained above, define capital to be all non-human means of production plus the skills of workers that are beyond the ability to work possessed by all normal humans. This is doubly wrong because not only does it transform capital into an eternal category of production but it overlooks the fact that once the labor (power) has been sold by the worker to the industrial capitalist, it is very much a form of capital—indeed the most important form of capital because it alone produces surplus value. On this point alone, the vast superiority of Marx’s terminology should be obvious to the reader.
Anybody who is familiar with Shaikh’s work knows that Shaikh is highly fluent in both Marx’s terminology and the terminology of “modern” orthodox economics as well as “Sraffian neo-Ricardian” economics. However, to get his work published by Oxford University Press, Shaikh had to use terminology that would make the work understandable to his fellow economists, who as a rule are uneducated in Marx’s economic writings or indeed the work of the classical economists or even Sraffian neo-Ricardian economics. The price he had to pay for this was to make himself obscure not only to Marxist political activists, who are as a rule not professional economists, but also in terms of clarity more generally.
In order to illustrate Shaikh’s arguments as clearly as possible—assuming I understand them correctly—I will use Marxist terminology not only because it is more familiar to most of our readers but because it is far more precise than the economic vernacular Shaikh was obliged to employ in his “Capitalism.”
Fixed and circulating capital
In order to understand Shaikh’s attack on Okishio, we have to return to the circulation of productive capital. During its circulation, capital must successively assume three forms—money, productive capital and commodity capital. As far as productive capital is concerned, classical political economy, following Adam Smith, ignored the division of productive capital into constant and variable capital. Adam Smith believed that constant capital could in the “final analysis” be reduced to variable capital if you go back far enough. However, the classical economists were very much aware of the difference between fixed and circulating capital. (8)
It is important, therefore, to avoid confusing the distinction between fixed and circulating capital of the classical economists with Marx’s division between constant and variable capital. Another common error is to imagine that circulating capital is a reference to money capital. At least the way Marx used the terms, the distinction between fixed and circulating refers only to productive capital.
The circulation of variable capital
Under capitalist production based on “free” wage labor, all variable capital is circulating capital. During the paid portion of the working day, the workers on a social scale must replace the value of the commodity capital in the form of “wage goods” they must consume to reproduce their labor power. The portion of the commodity capital destined to function as wage goods is commodity capital as long as it is in the hands of the industrial capitalists—and merchant intermediaries—but then becomes means of personal consumption in the hands of the workers and their families.
The act of personal consumption transfers the value embodied in the wage goods to the labor power of the workers. Unlike the capitalists, who destroy not only the use value but the value of the commodities they use as means of personal consumption, the workers preserve the value—though not the use value—of the commodities they personally consume. The value of the workers’ labor power, however, disappears when the industrial capitalist buyer of the labor power productively consumes—puts to work—the labor power of the workers. However, the workers replace the value destroyed when the industrial capitalist consumes their labor during the paid portion of the working day.
The circuit of variable capital is therefore money capital->variable capital-> commodity capital->money capital. Because the full value of the variable capital is replaced with each turnover of the variable capital, variable capital forms part of the circulating capital. Since variable capital alone produces surplus value—though this fact is hidden from both the capitalists and the capitalist economists by equalization of the rate of profit and consequent transformation of direct prices into prices of production—the rate of profit on an annual basis is greatly affected by the number of turnovers of variable capital over a year. In contrast, the rate of profit on constant capital—once prices of production are (re)transformed back to direct prices—is zero. No matter how many times you multiply the number zero, you get zero.
Constant circulating capital is somewhat different than variable circulating capital. Constant circulating capital consists, in use value terms, of both raw and auxiliary materials. An example of auxiliary material is electricity used to power factory machines and factory lighting necessary if industrial production is to be carried out. The electricity, however, does not as a material use value enter physically into the body of the commodities it helps produce. In contrast, raw material—for example, yarn used to produce textiles—enters both physically and in terms of value into the commodity. What raw and auxiliary materials have in common is that they transfer their entire value in a single turnover cycle into the commodity they help produce. Hence, they are both constant capital—they produce no value themselves but pass on their value in a different use value form into the commodity they are being used to produce—and circulating capital.
Fixed and circulating capital and the rise of the space capitalists
In the case of fixed constant capital, the value is transferred to the commodity capital over more than one turnover cycle. To illustrate the difference between fixed and circulating capital, we can look at the first attempts being made to transform rockets used to launch satellites and the resupply of the astronauts (or cosmonauts as the Russians called them) in space stations and exploration of extraterrestrial bodies into a private for-profit business. A new layer of industrial capitalists that for lack of a better term I will call “space capitalists” is beginning to emerge, though this process is still in its earliest stages.
Before the coming of the “space capitalists,” rockets were used only by governments, which were not acting as capitalists. For example, the rockets used in the Apollo moon shots carried out by the U.S. government between 1969 and 1972 were designed to prevent the socialist Soviet Union from being the first nation to land on and return humans from the moon. If the Soviet Union had done that first, it would have been another political mark against capitalism in the wake of Sputnik and the Soviet successes in sending the first man and then the first woman into Earth orbit.
In addition, many scientists were interested in learning more about the moon and its origins, though this was very far from being a chief interest of the U.S. government. None of these motives, whether those of the U.S. government or the scientists and astronauts, involved making a profit on invested capital. The rockets used in space launches, whether the motive was political as in the case with the Apollo launches or military such as is the case with “spy satellites,” were discarded.
If these rockets had been used in a capitalist for-profit business—for example, mining the moon and returning the ore to earth for refining—the rockets since they were used once and then discarded would have functioned as circulating capital. The entire value of the very expensive rockets would be transferred in each shot to the commodity capital produced—in the form of ore returned from the moon. This would have meant that “moon ore” would have a very high individual value. In reality, the individual value of any “moon ore” would have been so high that it would quickly be driven out of the market by far cheaper ores obtained from the Earth’s crust by traditional mining techniques.
This is one of the reasons, much to chagrin of science-fiction writers and other space enthusiasts, that after landing astronauts on the moon the U.S. government suspended the program, and as of this writing almost half a century has passed without any further human landings on the moon.
Today, however, some adventurous industrial capitalists, such as Elon Musk of SpaceX and a few similar companies, are trying for the first time to extend capitalist production for private profit beyond the Earth. However, if this is to succeed, Musk and other would-be “space capitalists” must transform rockets, which are very expensive pieces of equipment, into fixed capital that transfers only a small portion of their value to their payload on each launch. In order to do this, the rocket must be reusable. Only when this can be done with some degree of reliability will the mining of asteroids and other extraterrestrial bodies have any chance of becoming a private profitable capitalist business.
Calculating cost prices and the rate of profit
The rate of profit yielded by a capitalist enterprise is calculated by dividing the profit—interest plus the profit of enterprise—over the enterprise’s total capital. The cost price is the value of the entire circulating capital plus the portion of the fixed capital that actually transfers its value to the commodity during one production cycle. The cost price differs from the price of production in that cost price is the cost that industrial capitalists incur in carrying out production, while the price of production is the cost that society and the individual buyer of the commodity must pay if they are to enjoy the use value of a commodity.
The difference between the price of production—assuming that it coincides with the actual selling price—and the cost price is the profit realized on the commodity once it is sold. This is why the terms cost price and price of production are preferred over the term “cost of production,” which blurs this vital distinction. It must be kept in mind that the cost price includes only the value of fixed capital that is transfered to the commodity produced and not the total capital. In other words, the cost price does not include the portion of the value of the fixed capital that is not used up in production.
The rate of profit is calculated by dividing the profit by the total capital including the portion of the fixed capital that does not transfer its value to the commodities that are produced in a given cycle of production. From the viewpoint of the capitalists, the value locked up in fixed capital is not really that different than capital locked up in a long-term bank account, since the capitalist measures the quantity of capital in terms of money—that is, exchange value.
Keeping this in mind, Shaikh gives a numerical example on page 315 of “Capitalism” showing that if the price is high enough, the method of production with a higher cost price but requiring less capital per commodity produced will yield a higher rate of profit. However, as the price falls towards the cost price, the one that has the lowest cost price becomes the more profitable method.
Shaikh assumes two methods of production, which he calls C and D2, that are used to produce what I will call commodity X. He also assumes that commodity X sells at 100, but let’s call it $100. Keep in mind that I am using Marxist terminology throughout and not necessarily Shaikh’s terminology.
While neo-Ricardians assume the method of production that is always the most profitable regardless of selling price, Shaikh modifies his example making this assumption to show that this is not necessarily true. Shaikh assumes the individual cost price or unit price of our commodity X produced by method C is $78 and the individual cost price of the same commodity produced by method D2 is $75. Should the industrial capitalist producing commodity X use method C or method D2?
Since the cost price with method C is $78 and with method D2 is $75, you would think that the capitalist would choose the method with the lower cost price—method D2. Not so fast! Remember, capitalists should always adopt the method that yields the highest rate of profit on the total capital, which is not necessarily the cheapest method. Is this necessarily the method with the lowest cost price? Shaikh begins by assuming that X has a selling price of $100.
Method C has a higher cost price of $78 per unit of X as opposed to method D2’s cost price of $75 per unit. However, method C has one advantage over D2: It is capital saving, which we can assume is because this method has a lower organic composition of capital than method D2. How much capital do we need to produce a unit of commodity X with method C as opposed to method D2?
It turns out that if we produce a unit of X with method C, we need $137.50 of capital, but with method D2 we need $157.89. Using method C, the industrial capitalist “saves” $20.39 worth of capital for each unit of X produced. This compensates for the higher cost price of method C—$78—versus $75 with method D2.
Using the “capital wasting” method D2, the rate of profit is calculated by dividing the total profit per unit, which is $25, over the total capital set in motion to produce the profit, which is $157.89—yielding a rate of profit of 15.83 percent. But if we used the “capital saving” method C, despite its higher cost price, the total profit is $22 per unit and overall rate of profit is 16 percent.
It turns out that with a selling price of $100, method C yields a higher rate of profit than method D2. Therefore, the industrial capitalist should use method C rather than D2 because its capital saving virtues more than compensate for the cheaper but “capital wasting” method D2.
But our assumption of a $100 selling price is after all purely arbitrary. Suppose that C is the long-established method of producing commodity X and that every industrial capitalist that produces commodity X makes use of it. But then an adventurous industrial capitalist enters the market and starts producing commodity X using method D2. In order to grab market share from the incumbent capitalists that use method C, he decides to sell commodity X at a price of $84. Buyers of X are now offered an identical version at $84 as opposed to $100.00. Capitalists producing commodity X with method C start losing market share to their competitor who produces with method D2. Soon all the capitalists that are producing with the capital-saving method with the higher cost price of method C must lower their selling price to $84 or be driven from the market.
Now which method will be the most profitable? Will the capital-saving but higher cost price method C or the capital-wasting but cheaper in terms of cost price method D2? At the new selling price of $84, it is now method D2 that yields the highest rate of profit (5.70 percent) as opposed to method C’s 5.09 percent. In terms of the rate of profit on total capital, C and D2 have changed places. Now it is method D2 that must be adopted by all the industrial capitalists, though the overall rate of profit has fallen from 16 to 5.70 percent. But 5.70 percent is still better than 5.09 percent.
We should note that the cost price of using method C regardless whether the selling price is $100 or $84 is $78, while the capital-wasting—and thus lower rate of profit, all else remaining equal—method D2 has a cost price of $75. This means that if the price were to fall to the cost price of D2, which is $75, D2 would break even while C would make a loss of $3 per unit. Therefore, at a price of $84 or lower, down to a mathematical limit of $75, cost price method D2 will yield a higher rate of profit than method C. Indeed, at any selling price between $78 and $75, method D2 still makes a profit while method C makes losses.
To generalize, as the selling price of any commodity falls towards the cost price, the method no matter how capital-wasting that has the lowest cost price always wins.
According to Shaikh, who assumes real competition, sooner or later capitalists eager to expand their market share will adopt method D2, while according to the Okishio theorem, which assumes “perfect competition,” capitalists will keep selling at $100 and therefore shun the capital-wasting method D2 in favor of the capital-saving but more profitable method C. Instead of 16 percent on their capital as before, our poor capitalists now have to settle for a mere of 5.70 percent. In a world where everybody after calculating his or her cost price gets to add a 16 percent “normal profit” or “interest,” why would any capitalists spoil everything and introduce a method D2 that lowers everybody’s profit to 5.70 percent?
Before I read Shaikh’s “Capitalism” in criticizing the Monthly Review school theory of monopoly prices, I noted that the Monthly Review school based its theory of monopoly prices—imperfect competition—on the work of the marginalists—perfect competition. While Sweezy and Baran upheld the law of labor value in the abstract, when they got down to their concrete analysis of the economy they made no attempt to connect prices with values.
Instead, they simply reverted to the marginalist methods they learned in their university studies, which Baran and Sweezy assumed is compatible with Marx’s law of (labor) value, which as we know does not directly determine prices. According to the marginalists, the capitalists simply add the prevailing interest rate to their cost price, which will, assuming they adopt the most efficient method of production available, coincide with the prevailing market prices, assuming “perfect competition.”
If competition is “imperfect,” as Baran and Sweezy assume is the case in their “Monopoly Capital,” the capitalists are in a position to add an extra “rent” onto the interest, which will coincide with the selling price in a market where competition is imperfect. I believed—and now having read Shaikh’s “Capitalism” believe even more than before—that a theory of real world concrete competition and prices must be based on the foundations of Marx—labor value—and not the “perfect competition” of the marginalists. In other words, how does the law of labor value assert itself in the real world?
As Shaikh shows in “Capitalism” and his other writings, changes in labor values explain about 80 percent of changes in prices of production using realistic assumptions. It does so precisely through the process of real competition and not the fantastic theory of “perfect competition” of the marginalists.
Unfortunately, in the last and weakest part of Shaikh’s “Capitalism” mistakes he makes on money and price—the form of value—causes Shaikh to fall short in his analysis. But here we see the real strength of Shaikh’s arguments—once we correct for Shaikh’s errors we will see that his case against perfect-imperfect competition in favor of real competition is vastly strengthened.
Foreign trade
Next month, I will deal with Shaikh’s views on foreign trade. Shaikh completely rejects bourgeois economic orthodoxy, which is based on the theory of “comparative advantage.”
The economists use the supposed law of comparative advantage to show that free trade is in the interests of all nations. According to the theory of comparative advantage, to put a nation first—like U.S. President Donald Trump does with his “America First” slogan—makes no economic sense.
Shaikh, however, rejects the theory of comparative advantage, holding that it is absolute not comparative advantage that determines which capitalists emerge victorious in the battle of competition, whether this battle is fought out within the borders of the nation state or on the world market where the competing capitalists belong to different nation states.
If Shaikh is correct, then international trade is a highly antagonistic process. Perhaps President Trump—and other economic nationalists like Bernie Sanders—do have a point if we assume that we are operating in the cutthroat world of capitalist competition as opposed to a socialist world based on international cooperation. I will deal with this crucial subject and its implications in next month’s post.
_______
1 The term “artificial intelligence” was coined by computer scientists back in the 1950s. A rather loose term, artificial intelligence refers to any computer program that gives a computer the appearance of possessing human-like intelligence. In the early years of digital computers, many pioneering computer scientists believed that it would be relatively easy to simulate virtually all the functions of the human mind on a digital computer. For all practical purposes, machines would achieve human and then super-human intelligence. Those computer scientists who believe that virtually all the functions of the human mind can be simulated on digital computers are known as the “hard AI school.”
Since the 1950s, the computing power—measured by the number of computations a computer can perform in a given period of time—has grown exponentially while the physical size of computers and the cost-value or quantity of labor necessary to produce a computer has fallen so far that we now carry around computers many times more powerful than the super-computers of just a few decades ago in our pockets. We call these powerful computers smart phones. However, hard AI has proved stubbornly elusive.
This has not prevented specialized forms of AI from being extremely useful in fields ranging from beating the best human chess players and more recently beating skilled humans in the much more difficult—for a computer—at the board game Go to the use of “expert systems” in making medical diagnoses and the ever-growing role of computer-controlled industrial robots in factory automation. These forms of “soft AI,” however, make no attempt to simulate the human mind in general, though they may appear to be doing so to the lay public when they are used in very specific domains such has board games like chess and Go and beating contestants on the TV game show Jeopardy.
Recently, however, the rise of “machine learning”—for example, teaching a computer-controlled car to drive safely on the highway without any human intervention—has revived speculation about the possibilities of achieving “hard AI” in the coming decades. (back)
2 Recently, what are believed to be extremely primitive stone tools dating back 3.3 million years have been found in Africa. These tools were presumably made by members of the extinct genus Australopithecus—which means southern ape because the first fossils of this genus were found in South Africa. It is believed that one of the species of Australopithecus evolved into the genus Homo—human beings.
It is extremely unlikely that the oldest stone tool will ever be found, so stone tool production extends even further—nobody knows by much—into the past. Tools made of materials such as wood by members of the genus Australopithecus or an even earlier genus that have not been preserved may be considerably older. The history of production has deep roots that extend through human pre-history and into the pre-human animal world that eventually gave rise to humanity. (back)
3 Just as today’s machine learning and computer technology operates on top of “electrification,” electrification operates on top of steam power. Electricity is generated by steam-driven turbines that use electromagnetism to transform energy in the form of steam power into electrical energy. As far as the laws of physics are concerned, electricity is simply a way of using steam power more efficiently. (back)
4 A common error is to define chattel slavery as a situation where the workers—slaves—are not paid, while free wage workers are paid for their labor. In reality, the slaves too must be paid something or else they will perish, which would cost the slave owner the wealth invested in the slave. The real difference between chattel slavery and the “wages system” is that under chattel slavery the workers in their entire persons are the private property of the boss—the slave owner—while under the wages system the workers retain ownership of their persons but are obliged to sell for a period of the time their ability to work to a capitalist buyer. (back)
5 Bourgeois economists and bourgeois trade unions such as the “Gomperite” trade union leadership of the AFL-CIO define exploitation as a situation where the workers are paid less than the value of their labor power. This occurs, for example, in the all-to-common cases where “wage theft” occurs. In the case of wage theft, workers are exploited in the sphere of circulation. They receive less value in the money they are paid than the value of their labor power. Marxists call such cases super-exploitation, and we must of course strongly oppose and fight against all such cases.
However, even assuming the workers receive the full value of their labor power, they are still obliged to perform unpaid labor for the capitalists. This exploitation does not occur in the sphere of circulation—the workers receive the full value of their labor power—but rather in the sphere of production. Trade unions, insomuch as they confine their struggle to improving the position of the workers within the capitalist system, can at best eliminate super-exploitation but cannot end the exploitation of the workers by the capitalists. (back)
6 Marx was well aware that the rise in labor productivity tends to lower the value of the constant capital and thus hold in check the growth of the organic composition, which is a ratio of values and not a technological ratio. Marx assumed that while the value of individual machines of a given power would fall, the total value of machinery and other forms of constant capital would grow causing the rate of profit to fall assuming a constant rate of surplus value. The claim that the rise in labor productivity lowers the value not only of individual machines but of the total constant capital ignores the way technology advances in practice.
For example, when machinery was first introduced in industry the machines themselves continued to be produced by the old handcraft methods of production. Only later was machine production of machines themselves introduced. As a rule, new more powerful types of machines are themselves initially produced by the old, more labor-intensive method of production.
There is also an economic reason that makes unlikely the development of the productive forces such that the value of constant capital is lowered in such a way that the rise of the organic composition of capital is nullified or even reversed. A fall in the value of the total constant capital relative to the total variable capital would lead to extremely severe crises as result of the devaluation of huge amounts of fixed capital. The severity of these crises would drive unemployment to extremely high levels.
The resulting increase of competition for the remaining jobs among the workers would sharply lower wages and increase the rate of surplus value, which would slow once again the rate of growth of the productivity of labor.
The rising rate of surplus value, therefore, along with the rate of scientific and technological change regulates the rate of growth of the productivity of labor. The faster the potential rate of growth in the productivity of labor the higher the rate of surplus value must be to prevent the growth in the productivity of labor from reaching the point where it would destroy the capitalist system.
A rate of growth of labor productivity therefore emerges that is actually much slower than the rate of growth allowed by science and technology alone. This fact shows the increasingly reactionary nature of the capitalist mode of production. Economists and non-economists who predict that revolutionary progress of science is about to eliminate jobs wholesale—and such predictions have been made since the industrial revolution—don’t realize that capitalism is not about the accumulation of the means of production but about the accumulation of capital. And these are, as Marx knew but our modern economists don’t, two quite different things. (back)
7 In Marx’s original vision, “Capital”—three volumes plus a book about the history of theories of surplus value, sometimes considered Volume IV of “Capital,” was to be only one book of a multi-book critique of bourgeois political economy. Other books forming Marx’s total critique of bourgeois classical political economy would deal with such subjects as landed property and rent, wage-labor, and the world-market, competition and crises. Each of these books was also to have the history of the theory on their respective subjects. Shaikh observes that Marx’s ambition far exceeded his ability to complete more than a relatively small part of it. Shaikh’s “Capitalism” is an attempt to fill part of the gap with a book on competition and crises. (back)
8 Marx considered Adam Smith’s claim that constant capital can be reduced to variable capital in the final analysis to have been a major error that held back the progress of classical political economy. However, Shaikh, who is a great admirer of Adam Smith, sees a more positive side to Smith’s analysis on this point, because it foreshadowed the input-output models employed by 20th-century economists. These input-output models have played an especially important role in the work of Sraffa and other neo-Ricardian economists that Shaikh has greatly admired as well as criticized throughout his career. (back)
A Critique of Crisis Theory 