The centrally planned economy, Hayek, and the red spot of Jupiter.


After the fall of the Soviet Union, economic marginalists triumphally proclaimed that the market is the only realistic system to manage our complex, global civilization of billions of people.   Although homelessness, vacant houses, starvation and food waste persist, the marginalist will argue that the market is not perfect but that no better alternative exists. Yet, an incoming catastrophe in the form of climate change may pose the question of a democratic and globally planned, socialist economy – after all, it seems self-delusional to think that an economic system made of competing firms and nation-states can respect the planetary boundaries. Global warming was almost entirely caused by the laws of motion of  capitalism, which lead to the wear and tear of bodies and the Earth, given how competition drives firms and nation-states to harvest the cheapest labor and natural resources.  Furthermore, the  fact that the length of the working day has not decreased in almost a century, and imperialism has exacerbated predatory asymmetries between the core and the periphery, should make the question of global, economic planning central for socialists.

The planned economy  has a bad rap, even amongst socialists.   The debate seems to have been settled in the first half of the 20th century with the so called “socialist calculation problem”, where  marginalist economists such as Hayek and  Mises  criticized the inability of centrally planned economies to compute the authentic demand and supply for specific goods. Hayek in particular, gave the most sophisticated attack, with his essay,  The Use of Knowledge in Society where he argued that the market acted as an unconscious, distributed computer where resources are efficiently allocated through the computation of the demand of goods by price signaling between different parts of the system (e.g., individuals and firms).   Many socialists have  criticized Hayek’s arguments throughout the last century, yet many of the socialist retorts are posed from a philosophical and epistemological perspective.  However, I believe the argument for central planning can be contextualized using new mathematical sciences, such as computer science, and nonlinear dynamics – fields that didn’t exist in Hayek’s day. So I will attempt to contextualize Hayek’s argument using a more “mathematical”  ( but not quantitative) method and I will retort from a pro-planned economy, socialist perspective.

The theme of Hayek’s argument is that the central planners have no possibility of knowing all information required to efficiently allocate goods in  a society. In contrast, the market acts as a giant, distributed computing system, where firms and individuals act as “parallel processors”, where each individual processor computes a small, local problem: a shop-keeper computes that a particular brand of cigarettes is very popular in  their neighbourhood, and consumer calculates their individual demand  on cigarettes the moment they look at the available cigarette brands.  The parallel processors, which are embodied in individuals, firms, and institutions, then in turn communicate with each other,  finally collectively computing the price of a particular commodity, which embodies the aggregate demand and supply of a particular good. To conclude,  central planners can never  acquire all the required information to compute efficiently the needs and wants of a particular good, while the market, which acts as a distributed network of processing cores, can efficiently allocate goods because each processor computes a smaller, simpler problem (e.g. the want of an individual for marlboro cigarettes over parliaments, or the observation by a small shop keeper that 5 ft long USB cables sell out abnormally fast in a particular Best Buy), and communicates with other processors through pricing and purchase, leading to the allocation of goods where they are demanded.

However,  the old socialist would counter that there is nothing efficient about the market, given the vacant houses, food that goes to waste, massive poverty, the business cycles,  etc. However the marginalist would retort that the market as a distributed computing system has its problems, but it will still always be better than  central planning,  citing toilet paper shortages, and long bread lines.  Hayek argues that  problem is ultimately about information – the central planner will never have enough or timely information to plan the demand at the granular level – e.g. the demand for a specific brand of cigarettes, or for the right size of a smartphone, or a particular laundry machine.

However, the Hayekian attack on central planning is only valid at the granular level. If the strong form of the Hayekian attack against central planning were true,  then the natural sciences would be invalidated.    The lack of information at the granular level is actually a common problem in the natural sciences – where for example, we can predict the climate (e.g. the average temperature of the Earth averaged throughout ten years) but are unable to forecast the weather (the temperature, precipitation, etc for one hundred kilometre squared at a given day). Similarly, we can predict the average thermodynamic properties of a gas, such as temperature or pressure, but we cannot predict the movement of an individual molecule in a gas.   This can be understood in scientific parlance as random noise at local scales that makes theories more uncertain at smaller scales but still allows for predictions and modelling at larger scales.  The  “random noise” can be thought as unpredictability arising because of lack of information at smaller scales. For example, in the case of weather forecasting,  the lack of information about all the variables affecting the weather, such as precise temperature measurements,  numerical errors arising from the computers solving the hydrodynamic equations that govern the air flow, the  inadequate modelling of the physical geography etc, rapidly leads to  inaccurate results at the local level. However, in the case of making predictions about the global scales of climate, such as the average temperature of the whole earth in the next ten years, the statistical noise at smaller scales becomes irrelevant. What Hayek implies,  is that because  statistical noise  exists in a given economic system, that economic planning is absolutely impossible. He frames his argument as informational, stating that the central planner has not sufficient information on the demand of goods at the local level. Yet, many natural sciences have to deal with extreme statistical noise at small scales, making forecasting only possible at larger scales, so his argument would seem to invalidate the natural sciences such as astronomy, climate science, and ecology as well. Therefore, a physical scientist would reply that his argument only applies to the smaller scales where the noise dominates, and does not say anything about larger scale systems.  Hayek argued  that one cannot compare the economic and natural sciences, because the latter is concerned with objective, natural laws, while the economic sciences are concerned with subjective, human wants; however that argument is irrelevant, because it’s obvious desires can be quantified,  which is precisely what psychologists or firms like Amazon do. Finally,  empirical evidence invalidates his argument against planning, for the institution of private property and the rule of law, which are necessary for the existence of the market, are large-scale, national and sometimes even international systems that require inordinate amount of coordination and planning, given that these institutions have incredible overhead in the form of the police, the paper-pushers, the lawyers and judges. These institutions are formally necessary for the market in general, even if social welfare or food regulations don’t exist.

Another interesting assault against planning comes from Nassim Taleb, who has revived the Hayekian argument in spirit but with the use of modern statistical tools.  His most important point is the existence of “black swans”, rare and unpredictable, extreme events that can  trigger radical changes in a given system.  Some  black swan examples are terrorism,  massive floods of coastal cities, and nuclear melt downs. All of these are rare events with extreme and almost unpredictable consequences. For example. terrorism’s body count is highly variable, from a couple of people dying at a given event, to thousands of people. Terrorism can also   trigger unpredictable  social instabilities in a given polity.   Other black swans come in the form of famous works of arts,  economic crises, the overthrow of governments, and   world-historic events. The problem with economic planning, then, is that by its own nature it’s blind to black swans; thus  planned economies are very fragile to unpredictable shocks, not unlike a very complicated clockwork that can crumble the minute one of the cogs breaks. However, the existence of black swans such as economic shocks is not really an argument against economic planning, given that  human societies throughout history have always been endangered by black swans and shocks, with disease, wars, and technological inventions wiping out whole societies. So whether the  economy is planned or not, dangerous black swans could still appear. A way to deal with black swans is with sensible risk management; although we cannot predict black-swans, we can design systems to be robust to shocks.  For example, buildings in seismically active regions are built to withstand earthquakes, the latter which are black-swans,  given their rarity, extreme nature, and unpredictability. Hypothetically, there is no reason why economies can’t be built to withstand shocks.

My retort to the Hayekian argument is highly abstract and formal given that the original form of  Hayek’s argument is very formalistic. However, it would be interesting to see how a global, central planned economy would look like, and how the granular uncertainty Hayek pointed at would be dealt with.   Economic planning could be made of two processes: a distributed, decentralized planning from below, and a broad-stroke centralized planning from above.  The broad-stroke, central planning would be directed by elected and recallable councils but would deal with planning at the central, global level, dealing with planetary objectives  such as making sure the economy doesn’t surpass ecological constraints (e.g. global warming). Another global, planning objective would be reducing the length of the working day. This latter point is important given that marginalists like Keynes promised a short working day that will triggered by the movement of the market. However, now it is obvious that the working day is entirely a planned and political thing, and cannot be reduced just by the stochastic behavior of the market. In fact, the historical shortening of the working day happened entirely because of legislation triggered by the militant activity  of  the working class. Finally, global central planning would have to deal with global problems capitalism  exacerbated such as global inequality and imperialism.  The  distributed, decentralized planning from below, would be in charge of the micro-economical calculations of supply and demand for particular goods, such as the appropriate way to stock stores for consumers.  Capitalism is  competent at the micro-economic  part, stocking shops  with commodities based on the supply and demand as mediated by price signals – this latter point was at the heart of Hayek’s argument. Yet, there is no reason why efficient, micro-economic calculation couldn’t be made by local, democratic councils with the aid of advanced computers.  Input data on consumer wants and needs, which can be signaled from what individuals pick up at stores, can be quickly processed by machine learning algorithms, not unlike how in capitalism purchases and pricing propagate the information of supply and demand. In fact, this sort of big data processing is already done with intra-firm planning today, with companies such as Amazon and Wal-Mart planning resource allocation based on consumer data that is processed with machine learning algorithms.  The socialist democratic councils that would plan micro-economic movements could act as semi-autonomous, but publicly owned, firms as well, using a similar micro-economic calculation approach to modern capitalist firms, but without having to depend on price signals, and instead using consumer big data and information related to global  constraints such as world resources, global development plans,  ecological risk,  global resource allocation etc (these global constraints  would  be outputted by global planning councils).

Global planning is a very big hypothetical, and would require the existence of a world, socialist council republic.  However, given the hard, planetary constraints that global warming unearthed, it’s urgent to argue for alternatives to the anarchy of the market.  If socialists don’t argue for an alternative, factions of the capitalist class, such as fascists, will certainly come up with their own forms of centralized, authoritarian economies given the social and political threats that global warming would bring to the table.  Global warming is going to trigger humanitarian disasters that will lead to unquantifiable social and political consequences, such as a massive refugee crisis that will embolden reactionaries and nationalists. Therefore, the problem socialists face will not only be ecological, but political, and if we do not bring our own alternative to the table, our enemies will.