Last week, I began to discuss the question whether Francis Fukuyama’s hypothesis about convergence to liberal democracy and capitalism is at least partially right. While the countries of the world have not been moving towards democracy in the last 25 years, he could still be right that the future belongs to the markets. This week, however, I will argue that this is not the case.


In the works of libertarians like Ayn Rand and David Friedman, one can find a deep-rooted conviction that the outcomes of markets are not only efficient but also morally desirable. This contention is not as preposterous as one might think.

In a market, those are doing well who have to offer something other people desire. Typically, the object of desire may be ordinary capital or (qualified) labor. Let us focus on the latter, because most people make their living by selling their labor. The abilities to write computer programs, conduct orchestras, and to teach mathematics, to name just a few, are demanded by other people, and thus generate income to those who have them. But also solid skills like cooking, carpentry etc. are demanded by others and lead to income.

Markets only yield unsatisfactory results for those who have nothing to offer, for example because they are old or heavily handicapped. In pure capitalism (i.e. without government, as envisioned by David Friedman) only handicapped and old people would have a hard time, as their well-being would depend solely on the altruism of others.

For those who have to offer something in a market, the competition is typically multi-dimensional and therefore not so harsh. Humans can engage in “product differentiation”: as a university instructor, I do not compete with the taxi driver who brings me to my office, as I am not offering taxi services and he is not teaching students. And if there are too many people competing with me in the market for university instructors, I may consider to offer something else, e.g. become a journalist. The taxi driver, on the other hand, may also react to competitive pressure. He can upgrade his qualification and get a driving license also for trucks, he can learn English and become a tourist guide, or he can just offer his manpower in one of the many construction projects going on in the country. So, even though capitalism is based on the principle of competition, everyone is just competing with a relatively small group of people, and the actual competitive pressure is rather cozy.

Even nicer is the fact that one can do a lot to decrease competitive pressure. One possibility is to accumulate human capital – the higher one’s qualification, the less competition one has to face. While I could become a taxi driver (if I would have a driving license), most taxi drivers would have to upgrade their human capital for being capable of teaching students. So, my qualification reduces the competition.

To sum up, the competitiveness of most people is determined by skills, resourcefulness, and diligence, and the market rewards talent and efforts. A just and justifiable system, and the libertarians seem to have got it right!


In contrast, consider a situation in which 90% of the population were severely handicapped or very old. All of these people would have nothing to offer to others, and the market outcome would be unattractive from a moral point of view. The 10% of the population who were competitive would arguably live reasonably well, but the remainder would starve.

In their 2011 book “Race against the Machine”, economists Erik Brynjolfsson and Andrew McAfee argue that we are now at a point where “technological unemployment” becomes a reality. The idea that technical progress could cause net unemployment is very old (at least 250 years) but always turned out to be wrong. Whenever machines made a profession obsolete, new demand was created elsewhere to absorb the supply of human labor. Humans could always deal relatively well with the competition of other humans. In future, however, their main competitors will be machines. Brynjolfsson and McAfee predict that many humans will lose all of their competitiveness.

According to Moore’s Law, computing power available in a given amount of space doubles approximately every 2 years. This exponential growth could be observed since 1965. Brynjolfsson and McAfee argue that with exponential growth of computing power, in future there will be little left that machines cannot do. There are plenty of examples. A few years ago, it was considered impossible that computers would be able to independently drive cars. Now, companies like Google let a car drive without human intervention for thousands of kilometers under normal traffic conditions. Experts assume that in the 2020s driverless vehicles will gradually replace cars driven by humans. Arguably, in 20 years from now taxi driving will be no option anymore to make a living.

What impressed me most about my new smartphone is the speech recognition. I can speak quickly and fluently and the phone transforms my words into written language, almost without mistake. A bit more of that, and we do not need secretaries anymore.

Some months ago, a computer program passed the Touring Test, i.e. it was communicating in a way that was indistinguishable from human communication. A bit more of that, and we do not need shop sellers, consultants, and medical doctors anymore. With enough computing power, there is almost no kind of labor that is safe from being replaced by machines.

Brynjolfsson and McAfee revive the old argument (going back to Ricardo) that there is nothing in a market which prevents wages to fall below subsistence level, i.e. the level needed for a worker to sustain his and his family’s living.

An even stronger point, however, is that machines are not only replacing human labor, but usually offer something which is better. The self-driving cars may soon drive better than human drivers, the electronic secretaries may make less mistakes, and the computer doctor may make the right diagnosis more often than its human counterpart. Today, horses are not only obsolete because their labor value is below what is needed for their subsistence (as argued by Brynjolfsson and McAfee, using this example), but also because steam engines and combustion motors are much more versatilely employable. Horses were used to pull ships, walking on paths next to channels. Today, the strongest ship engines have more than 100,000 horsepower, but it is impossible to let 100,000 living horses pull a ship. Even if the cost of living horses were close to 0, the demand for horses would be much lower than it was 150 years ago.

If now, in an exponential process, more and more humans have nothing to offer anymore – regardless of their talents, resourcefulness, and diligence –, then markets will not produce outcomes anymore that are morally acceptable. The basis of the libertarian view on markets is shaken!


In the 20th century, there was a huge debate between socialist economists and the so-called Austrian School about the question which system (capitalism or socialism) would finally prevail. In the debate evolving around the so-called Calculation Problem, Oskar Lange, a Polish socialist economist, claimed that socialism could never perform worse than capitalism, because a planned economy could simply emulate a market and yield the same results. Hence, a socialist economy would be at least as good as a capitalist one, but maybe better. His Austrian opponents (Mises, von Hayek and others) proved in response to Lange that there was simply not enough computing power to emulate a market economy. What the price system does in a decentralized and heuristic way could not be implemented as a plan.

One of the major problems why socialism failed was the lack of computing power. But the problem of future markets is the abundance of computing power. Isn’t this a striking coincidence?

The economic planning of the Soviet Union was extremely crude. The Gosplan authority set up 5-year plans which were usually outdated from the very beginning. A modern version would be different. It would be adjusted in real time. Whenever somebody would buy a coffee, that would be reported online to the central computer. The production plan would be adjusted instantaneously. Stocks would be very slim (as it is common already in today’s production), and many things would be produced on-demand. If there was an unforeseen event, e.g. an accident in a production plant, the plan would be adjusted immediately.

Even if there would be many problems with such a system, when markets become dysfunctional it might be considered the lesser evil. As democracy does not depend on free markets (see the first part of the article), such a system could possibly be implemented in a democratic way.

My guess is that Marx’s vision of the future will turn out to be more accurate than Fukuyama’s.