Material World: Non-market socialism is feasible (2024)

The Material World Column from the October 2024 issue of the Socialist Standard

All the necessary techno-infrastructure required to enable a post-capitalist society to function effectively already exists today; we don’t need to reinvent the wheel. A self-regulating system of stock control involving ‘calculation-in-kind’, making use of disaggregated physical magnitudes (for instance, the number of cans of baked beans in stock in a store) rather than some single common unit of accounting (such as money) as the basis for calculation, is something that already operates well enough under our very noses within capitalism, alongside monetary accounting. Any supermarket today would, operationally speaking, rapidly grind to a complete halt without recourse to calculation-in-kind to manage and monitor the flow of goods in and out of the store.

At any point in time our supermarket will know more or less exactly how many tins of baked beans it has on its shelves. The computerisation of inventory management has made this task so much simpler. Our supermarket will know, also, the rate at which those tins of baked beans are being removed from the shelves. On the basis of this information it will know when, and how much fresh stock, it will need to order from the suppliers to replenish its existing stock – this simple arithmetical procedure being precisely what is meant by ‘calculation-in-kind’. It is applicable to every conceivable kind of good – from intermediate or producer goods to final or consumer goods.

Calculation-in-kind is the bedrock upon which any kind of advanced and large-scale system of production crucially depends. In capitalism, monetary accounting coexists alongside in-kind accounting but is completely tangential or irrelevant to the latter. It is only because goods – like our tins of baked beans – take the form of commodities that one can be beguiled into thinking that calculation-in-kind somehow depends on monetary calculation. It doesn’t. It firmly stands on its own two feet.

Market libertarians don’t appear to grasp this point at all. For instance, according to Jésus Huerta de Soto:

‘… the problem with proposals to carry out economic calculation in natura or in kind is simply that no calculation, neither addition nor subtraction, can be made using heterogeneous quantities. Indeed, if, in exchange for a certain machine, the governing body decides to hand over 40 pigs, 5 barrels of flour, 1 ton of butter, and 200 eggs, how can it know that it is not handing over more than it should from the standpoint of its own valuations?’ (Socialism, Economic Calculation and Entrepreneurship, 1992, Ch 4, Section 5).

This passage reveals a complete misunderstanding of the nature and significance of calculation-in-kind in a post-capitalist society. Such a society is not based on, or concerned with, economic exchange at all. Consequently, the claim that ‘no calculation, neither addition nor subtraction, can be made using heterogeneous quantities’ is completely irrelevant since such a society is not called upon to perform these kinds of arithmetic operations involving a common unit of account. This is only necessary within an exchange-based economy in which you need to ensure exchanges are objectively equivalent.

On the other hand, even an exchange-based economy, like capitalism, absolutely depends on calculation in kind. As Paul Cockshott rightly notes:

‘Indeed every economic system must calculate in kind. The whole process of capitalist economy would fail if firms like Honda could not draw up detailed bills of materials for the cars they finally produce. Only a small part of the information exchanged between companies relates to prices. The greater part relates to physical quantities and physical specifications of products’ (Reply to Brewster, Paul Cockshott’s Blog, 28 August 2017).

In his Economic Calculation in the Socialist Commonwealth Mises claimed that the application of in-kind calculation would be feasible only on a small scale. However, it is possible to identify extant or past examples of calculation-in-kind being implemented on a fairly – or even very large scale. For instance, Cockshott refers us to the fascinating case of the first Pyramid at Saqqara, built under the supervision of Imhotep, an enormous undertaking by any standard, involving nothing more than calculation-in-kind. Another example was the Inca civilisation, a large-scale and complex civilisation that effectively operated without money.

However, it was really the emergence of linear programming that has effectively delivered the coup de grâce against this particular line of argument peddled by Mises and others. It has removed what Mises considered to be the main objection to calculation in kind – that it could not be applied on a large scale basis.

Linear programming is an algorithmic technique developed by the Soviet mathematician Leonid Kantorovich in 1939 and, around about the same time, the Dutch-American economist, T. C. Koopman. As a technique it is widely and routinely used today to solve a variety of problems – such as the logistics of supply chains, production scheduling, and such technical issues as how to best to organise traffic flows within a highly complex public transportation network with a view to, say, reducing average waiting times.

To begin with, the computational possibilities of this technique were rather limited. This changed with the development of the computer. As Cockshott notes:

‘Since the pioneering work on linear programming in the 30s, computing has been transformed from something done by human ’computers’ to something done by electronic ones. The speed at which calculations can be done has increased many billion-fold. It is now possible to use software packages to solve huge systems of linear equations’ (Paul Cockshott, 2007, Mises, Kantorovich and Economic Computation, Munich Personal RePEc Archive, Paper No. 6063).

Computerised linear programming allows us to solve some very large-scale optimisation problems involving many thousands of variables. It can also help to solve small-scale optimisation problems.

In short, linear programming provides us with a method for optimising the use of resources – either by maximising a given output or by minimising material inputs or both. The problem with any single scalar measure or unit of accounting (such as market price or labour values) is that these are unable to properly handle the complexity of real world constraints on production which, by their very nature, are multi-factorial. Calculation-in-kind in the guise of linear programming provides us with the means of doing precisely this since it is directly concerned with the way in which multiple factors interact with – and constrain – each other.

While a non-market system of production could operate well enough without linear programming, there is little doubt that the availability of such a tool has now put the matter of whether such a system is feasible or not, beyond dispute.

Robin Cox

More FALC: The Walmart hypothesis (2019)

Book Review from the June 2019 issue of the Socialist Standard

Can Walmart Tell Us Anything About Socialism?

Internationally, there is a resurgence in discussing socialism, not just in the American sense of faintly looking at a welfare state, but credibly and seriously taking on the idea of a co-operative democratic abolition of market allocation of goods and resources. Books like Four futures: life after capitalism by Peter Frase have already lead the way, including among the options a world of abundance without money (although reading the text, it seems the author’s preferred or most likely future was some sort of regulated market in a post-ecological-collapse world). That book sprang from Jacobin magazine, and it seems that other contributors to that magazine are beginning to produce other useful examinations of the way towards a market-free world.

The latest example is Philips and Rozworski’s The People’s Republic of Walmart: how late-stage capitalism gives way to early-stage fully automated luxury communism. This book puts forward the simple case that firms like Walmart are in effect massive planned economies, close to the size and scale of the Soviet Union in terms of the number of products and processes they have to employ to run their enterprise. They also note how 3rd party firms that trade with them are effectively locked into their productive ecosystem.

Much of this is achieved algorithmically, with resources being poured into tracking stock through the system, knowing where it is at all times, and using vast storage capacity. The algorithms calculate what resources are available nearest to the point at which they are required, and how to get them there quickly. The authors give the counter example of Sears, which imploded after implementing an internal market and competition within its firm (under libertarian ideological commitment from its directors).

Of course, Walmart and Amazon are planning to exploit, and get the most out of, their workers and advantages over commercial rivals. It is not so much an example to be emulated, but a living demonstration that wide-scale economic allocation via planned structures works. Further evidence is found in a nuanced description of the history of the NHS, noting how it has gone from an ad hoc replication of existing structures, without conscious planning to serve community needs, to being carved up by an internal market. The authors note that it has always existed as a hybrid between the firms of GPs and the hospitals and the pharmaceutical industry, and that attempts to implement society-wide input and control were stymied at Westminster.

Democracy is seen as an inherent good and a necessity for wide-scale planning in an economy, if it is to serve people’s needs. Phillips and Rozworski note that a significant part of the failure of the Soviet Union’s planned economy was because of dictatorial intervention preventing the free flow of information between productive units, and removing their capacity to respond. The other living example they cite is Project Cybersyne in Chile.

British cyberneticist Stafford Beer was called in to help the Allende government implement a planned economy. Given the state of the country at the time, he had to improvise a system of telex and phone lines. His systems were credited with enabling the government to defeat a crippling lorry drivers’ strike, by routing key resources around disruptions and allocating available lorries. The principles behind Beer’s cybernetics were thus not all about enslaving people to the machine, but allowing distributed horizontal decision-making between relatively autonomous units with oversight and regulation to achieve common ends.

If all of this sounds remarkably familiar, it is because this is what we have been saying in our slightly less fashionably way since we were formed. Indeed, our pamphlet Socialism as a Practical Alternative, largely written by the late Pieter Lawrence, based a description of socialism on using the regulated stock control models already developed by the likes of Walmart, coupled with formation clearing houses to allow communities and groups to work to enable production for needs.

It’s clear, though, that while technology has advanced, along with the experience of wide-scale economic organisation, this is not a prerequisite for socialism. Pencil and paper-based systems could have handled the work (and probably still could), but the fact that machines and computers allow us to do it faster, and are increasingly presenting the possibility of co-ordinated democratic production in people’s minds, is making the topic more popular.

Philips and Rozworski cite the work of Scottish academic Paul Cockshott with his demonstration that an economy-wide plan is computable (especially if we exclude millions of null combinations of goods, arguably Walmart does not compute all the possible combinations of goods, but works with what it has and approximates efficient allocation-adjusting over time). Cockshott is not alone in academia, and he and collaborators have made impressive inroads into the possibilities of computational planning through vertical integration of sub-systems. It would be well to think what would be the result of the combined efforts of the best and brightest computational minds being directed towards co-operative economics and the satisfaction of needs, rather than algorithmic stock trading.

We have discussed in these pages before the types of innovations in thought and process that will enable us to make socialism work on a worldwide scale: Leonid Kantarovich’s linear programming; Gale and Shaply’s stable matching algorithm; Brams and Taylor’s envy-free cake-cutting algorithm, etc. The intellectual progress for the machinery of co-ordinating activity is improving, even if the political co-ordination to pick it up and use it is not yet as developed. It is good that these obscure and complex lines of enquiry are beginning to get serious popular attention.

Some commentators have begun to compare capitalism (and/or corporations) to a vast out-of-control Artificial Intelligence, programmed to maximise profit. This is potentially a powerful way of depicting capitalism: but the risk is, in coming up with algorithms and procedures to create a planned economy, that we would end up replacing capitalism with simply a rival AI, rather than returning to a human-centred system.

Demonstrating that we can plan and rationalise resources in the way that capitalism can and does is not a clincher for socialism: it could be argued that in an ends-based economy, using labour inefficiently could well become a goal (as William Morris pictured in his utopian novel News from Nowhere). The most powerful message we can send is that there is not one way to run an economy, and the only real limit is the power of our imaginations.

The tag-line of ‘fully automated luxury communism’ that is gaining some popularity may help raise the profile of consciously planning the economy co-operatively.

Pik Smeet