Action Is Sometimes Clearer Than Talk: Why We Will Always Need Trade
This is a lead essay in the C4SS Mutual Exchange Symposium: "Decentralization and Economic Coordination." The related readings and list of all other articles can be found in the introduction here.

Is it possible for our enemies to discover actual insights? The impulse to deny this is universal. The third reich dismissed special relativity as “Jewish physics” and lost significant advantage. The USSR worried that accepting Darwin’s insights in evolution would open the floodgates to capitalist social darwinism and so they hurt themselves by sticking with Lamarckism.

Most people can admit their enemy invented a useful mousetrap, but it’s much harder when one’s ideological enemies make a claim that has rhetorical power for their position.

The calculation/knowledge problem — a family of critiques discovered by bourgeois economists — has traditionally been used by capitalists to suggest There Is No Alternative to the existing order. If indeed markets are necessary for some complex economic production then that suggests some measure of the horrors and dysfunction of capitalism are tied to it. The instantly obvious retreat is to abandon complex production, or some measure of it. But for those who recognize that technology only possible through complex production offers quite significant expansions to our freedom to act, the problem is a pressing one.

There has been precious little serious work on the topic in the left, and the content of the responses dramatically differ depending on whether one embraces central planning or rejects it, and to what degree. This leads to amusing situations where those in favor of decentralized solutions cite the responses of those in favor of centralized solutions. It’s a bit of a mess.

There are also a number of marxists and anarcho-communists who respond by honestly admitting the problem hasn’t been solved — as G A Cohen put it, “the principal problem that faces the socialist ideal is that we do not know how to design the machinery that would make it run” — but it may still be worth pursuing with open-ended hope.

In contrast the worst response in all of this is arguably the quick dismissal that the problem only applies to state socialism. Simply put, political decentralization is not the same thing as decentralization of information flow. Nor does it provide assurances of accurate information.

A collective meeting may involve equal authority but still have the structure of the meeting functionally centralized. If only one person is speaking at a time in a spokescouncil, that’s a centralized system: at any one time only one person is broadcasting information and everyone else receives. That who is speaking rotates can create an equality of political authority, but it does not change the network structure of how the information flows. This is immediately apparent to those of us who sat through large spokescouncils in the counter-globalization era — once you have dozens or even hundreds of people the meeting form grows catastrophically inefficient at solving anything complex.

Trying to solve the problems by only having small meetings and creating tiers of representation starts to create political authority via representation but it also degrades the fidelity of the information conveyed. We’ve all seen cases where concerns or particulars raised by an individual in an affinity group or breakaway committee are simply not conveyed at all to the larger assembly because of time constraints.

But no matter the structure by which conversation is facilitated, there are still fundamental constraints posed by human language, whether written or spoken. While there are some base needs widely shared by homo sapiens, most of our individual lives are incredibly unique. Our desires, aspirations, and preferences are not easily conveyable to other people. This insight by the Austrians is not a mere rhetorical trick of philosophical skepticism, it finds its ultimate grounding in the rich density of neural information in the human brain and the limited average carrying capacity of language. Consideration of and deliberation between desires can occur far faster and far more productively within the brain of a single individual than it can occur between individuals when forced through the limited channel that is dialogue.

Even the closest lover is usually far less qualified to make a decision than you are yourself. Only you have direct and relatively immediate access to the vast network of contingent desires and preferences in your own head.

When I was homeless as a child I would try to intricately fill out the christmas wishlist forms provided by churches. From such experience I realized the problem: many gifts are of less use without other gifts, some fulfill some needs or desires, but not others. Conveying all those details in a form is nearly impossible. That problem is magnified when dozens, hundreds or ultimately billions of those forms must be processed through and weighed against one another and the various other considerations and costs attendant. Processing is not magic, a supercomputer the size of the sun can’t crack some encryption schemes and so too are there significant and inescapable constraints on processing through economic considerations.

These wider considerations and costs are inherent — there is no such thing as a world entirely free of scarcity. Sure some “needs” can be met rather trivially without the artificial scarcities created by the present capitalist order, but there is no objective line demarking “needs” from “wants.” Anarchists in particular, prone as we are to framing anarchism in terms of pursuing infinite desire, demanding everything, are either forced to take an abrupt u-turn into a kind of Buddhist quietism or we must recognize that our desires and aspirations will always expand and chafe up against material constraints of some kind.

Every choice has tradeoffs, and this is why the most fully agential expression of choice is not a proclamation of desire, but a trade. It’s not enough to say you really really super extra prefer X, such superlatives, no matter how bureaucratized, will never match the frank and immediate clarity of revealed preference. And showing you’re willing to trade a hundred Y for X only means something if you can in fact actually do so. Not with someone else’s Y, but with your “own”. This truth-forcing reality of personal stake is part of why prediction markets are able to — in aggregate across many participants — do so well. 

What is being sketched here takes inspiration from the real world emergence and flourishing of some marketplaces, but is intentionally abstracted from the capitalist hellworld we exist within because as anarchists we should be uninterested in merely solving for “better than” the existing order. This is not a defense of a very specific market form, nor a claim that markets should drive literally every process. This is merely a claim of the usefulness of decentralized networks where individuals trade things of value to them. As I have always argued, markets are a tool, not a god.

The neoclassical textbook just-so-stories of people on a deserted island are not an accurate reflection of the emergence of markets, because in a small commune content with a small number of fairly simple and static needs there is little impetus for much trade. The utility of trade emerges first when you want to coordinate outside of small communities into larger more cosmopolitan societies. But it is reinforced secondly when conditions change rapidly, whether environmental, social, or in terms of individual desire.

It would be one thing if no adaptation was necessary, no matter how complex the considerations, we could in theory eventually solve them to some measure of tolerableness and then just keep that solution in place forever. The problem emerges when an economy needs to adapt, particularly when it needs to handle a high amount of changing particulars at a very fast speed, harder still when inventions spur feedbacking acceleration.

The USSR’s state capitalist dysfunction is now widely acknowledged, but it’s not yet widely understood. Incentive structures were obviously malformed, without personal stake workplaces dealt with uncertainty by over-requesting resources, creating more and more intense requests, hoping to get whatever they could. But more generally prices were indeed a matter of “calculational chaos” with de facto prices or input-output flows set in many cases only thanks to the black market or by stealing prices from the west, with agents literally stealing order catalogs to help them plan:

“Gosplan officials used the prices quoted for goods in the catalogue to obtain relativities between this and that item. They would then try to match the goods of the catalogue to what was available in the Soviet Union and then fix prices according to the relativities prescribed by Sears Roebuck.”


Probably the most cited modern response to the knowledge and calculation problem is that of Cockshott and Cottrell, authors of Towards a New Socialism, and they’ve worked across a number of papers to try to salvage planning and prove that the question of what economic inputs to apply to what outputs can be solved with linear programming. They are statist marxists and are trying to solve for “the planning problem” which is not quite a fully decentralized communism, but they make feints towards how planning can be “decentralized” into regional/etc clusters and endorse “democracy” — which is enough apparently to get them accolades among some anti-market anarchists.

First off it’s worth noting that the toy models of Cockshott and Cottrell do not abolish markets — following Oskar Lange and other socialists in the calculation debate, they intentionally retain a consumer market, explicitly recognizing the problem of individual desires and the utility of revealed preference. You would still go to the marketplace/store equivalent in their system and exchange some form of currency for goods. They’re merely trying to remove market relations in the prior production chain (a rather arbitrary cleaving that requires or assumes an exclusively hierarchical flow from raw inputs to minor parts to larger parts, to consumer goods, to consumption). But this is only doable through a sleight of hand: they dismiss everything related to infrastructure, social organization, and local or environmental particularities into a handwaved matrix assumed to be static. What this means is their whole system falls apart if those things change or themselves become responsive to changes in the flow of inputs and outputs. There is no room for the fluid rise and fall of workplaces or cooperatives from free association, nor is there space for reverse and lateral flows in the production “chain”. Trade networks (ultimately between individuals) provide more fluid capacity to fall apart and restructure, as well as incentives for bottom up adaptation and solution-finding, utilizing the tacit localized knowledge that each individual has, but cannot easily express. To give just another example Cockshott and Cottrell still have to arbitrarily set a number of potential outputs (eg different consumer goods), but there is no efficient mechanism to seek out which potential consumer goods should be considered and which should not. On the market this emerges bottom-up via individuals who have tacit knowledge of niche market demands and their willingness to essentially stake bets on it.

One doesn’t need to work hard to see that Allende’s Cybersyn was not just “before its time” but a doomed dream. The “decentralization” promised in these sorts of schemes is never full or real. It’s more properly thought of as a matter of tiered deputization. Which is to say that relationships and agreements flow upwards in representative scale along set paths, rather than being made in ad hoc and horizontal relationships. In part because iterative-update “planning” in such cases risks becoming a chaotic storm of messages between thousands of collectives without overall price signals to provide quick restorative tendencies.

Another avenue taken by fellow travelers in this branch is to simply point at the accounting, stock, and infrastructure management schemes of modern production behemoths like walmart and extrapolate that we should be able to simply socialize these. It should go without saying that these are not easily adaptable to anarchist aspirations, specifically geared as they are to gargantuan mass production through state-created economies of scale, to say nothing of the centralized, verticalist, and generally authoritarian command premise. But moreover these distribution schemes are in many ways inefficient — COVID-19 has been a great example of how fixed and ossified supply chains give you efficiencies until there’s suddenly a change that requires significant reorganization outside your existing assumptions.

This is not to say that fetishizing competition inside a bloated firm as Eddie Lampert infamously did within Sears is gonna give great results either, since that’s just a top-down imposed market, with arbitrary incentive structures that are hard to determine or even map as a planner.

And this all takes us to a huge expanse of objections to markets that basically boil down to “but capitalism has many planned or non-market aspects and it kinda survives.” This is essentially another variation of the appeal to the USSR. Capitalism is mostly inefficient when measured in terms of the satiation of individual desires, all its efficiencies are optimized around maintaining and strengthening power. It is a project of power, for power.

Break the institutions, liquidate the ecosystems of power, and pretty much anything would be better. But better isn’t the same thing as optimal or even tolerable.

Our target should be many orders of magnitude improvements over capitalism so that we can improve nearly everyone‘s lives, and win a stable long-term consensus in favor of liberation. Any transition to a better world will be rocky, disruptive, and vastly damaging. It’s easier to be less afraid of ruins when we have tools to very rapidly build.

But it is worth noting the bottom-up weeds that emerge on their own in the shell of the old, especially those which capitalism finds need to contain, enslave, and direct. It’s fair to admit that Sears was something of a planned economy, and one would expect its prices to be initially set by merely estimating costs and then adding some desired profit margin. But the reason Sears catalogs were of such value to USSR planners was precisely that Sears faced competition from the scraps of legitimately bottom-up market processes in US capitalism, realities that transmitted — however partially and less efficiently than a pure neoclassical theory — actual pressures of supply and demand.

If the USSR was partially held aloft by the black markets growing up like weeds within it, Sears was likewise held aloft by the markets growing all around it, even though Sears itself would no doubt quite like to enjoy the kind of monopoly power that the USSR state held. Marketplaces exist in virtually every community in the world. While the state has at times in history played a role in the construction of very specific markets with dynamics and incentive structures inclined to reinforce power, other markets have also always emerged from the bottom-up.

Sure, capitalist firms and family households are able to plan internally — to varying levels of positive results for human liberation — but we also know the asymptotic inefficiencies they face with complexity scales. We all know the ways that the capitalist firm is propped up by power against the various diseconomies it faces. The “family” is often likewise propped up by broader patriarchal power systems. Knowledge problems are no less relevant there, we just somehow accept patriarchs, parents, and elders steamrolling over the more complex desires of other family members. 

Anti-market anarchists who don’t retreat from complex technology and desire are stuck in a bind, if you “abolish economy” and allow individuals to pull whatever they desire from the collective pile, there’s little way for even saintly individuals to understand the cost tradeoffs at play. If formal institutions are created — even if these communes or collectives are multiply overlapping in a mesh so as to avoid insular closed social topologies and continue centering individual agency — those institutions face bureaucratization. Even the most certifiably “fair” bureaucratization of Parecon. Not a problem for those extroverts hungering for structured socializing via The Meeting That Never Ends, but a dystopian hell for most people.

Folks love to handwave at cybernetics, and there are some computational aspects that can be facilitated with the surfeit of computational power we now have (although not for long if an anarchist economy flounders), but most of the problems are a result of a gap between the speed of information flow within our brains and that carried on average by our language and culture per phoneme or character. When we have brain-to-brain communications, not to mention spare storage and processing on par with human brains we could begin to offload the drudgery of meetings and logistics or accounting discussions. That day is not today.

Evgeny Morozov inadvertently provides a characteristic description of this kind of optimistic belief that tech can solve all, phrasing it as “ngos, charities, churches—can leverage information technology to do a better job at coordinating disaster relief than centralized government bureaucracies.” But the implication — that some kind of “needs registration” app, as Daniel Saros advocates, can resolve problems better than the market merely because decentralized orgs can coordinate certain problems better than the state — is fallacious and empirically false. Food Banks have adopted internal markets to better coordinate distribution between them because it works better than needs registration with iterative correction.

Further, whatever the system of technological facilitation you might set up… if there’s still revealed preference through testing users with trades where they could legit be stuck with the result… you still have a market. A market with extra steps, maybe, but a market all the same.

The fractal structure of markets in a liberated world, their norms, will probably emerge in complex ways mediating more tensions and considerations than anyone could ever write down. I would hope — via the smashing of power — that this emergence will look emergently bottom-up and as deliberatively considered as possible, but marketplaces are sites of constant contestation that we will have to continuously struggle to assert our values over and within. Same as with any meeting process.

I have grown partial to fully public ledger markets, more akin to the informal markets that emerge prior to state “standardization” and forced anonymization. One of the claims against capitalism is that firm competition drives secrecy, impeding accurate clearing. This is certainly true, and we can argue about the degree to which this norm is able to persist only thanks to the various distortions brought on by state violence, but a market once freed will still reflect an aggregate of our desires and thus our values, we must still work to see our most emphatically held values embodied or normalized. Transparency is a hard won and unending struggle in any context. Removing, marginalizing, or severely impairing anonymous transactions would do wonders for firm transparency, but aggressive reporting and broad social expectations will still be needed. If sometimes actors fail to communicate relevant tacit information to create and exploit asymmetries in markets, well they certainly do the same in collective meetings and every other non-market context ever proposed.

What I increasingly suspect, however, is that just as anarcho-communists and anarcho-collectivists will never be able to fully suppress black markets, we will have to live in a world cut with veins of secrecy, deliberately opaque transactions and relations. The real anarchist economic contest, I believe, will eventually be recognized as over how that secrecy is embraced, contained, and navigated.

I hope we will soon be finally ready to move on to tackle that question, rather than endlessly rehashing whether our enemies could ever have truly stumbled onto a potent realization with the knowledge & calculation problem.

I do not pretend to know with real certainty what sort of market arrangements would likely emerge in a liberated world, or even what would be most optimal. But I do know that markets will have to be part of the mix and that fearing and demonizing them as a whole is shooting ourselves in the foot.

Mutual Exchange is C4SS’s goal in two senses: We favor a society rooted in peaceful, voluntary cooperation, and we seek to foster understanding through ongoing dialogue. Mutual Exchange will provide opportunities for conversation about issues that matter to C4SS’s audience.

Online symposiums will include essays by a diverse range of writers presenting and debating their views on a variety of interrelated and overlapping topics, tied together by the overarching monthly theme. C4SS is extremely interested in feedback from our readers. Suggestions and comments are enthusiastically encouraged. If you’re interested in proposing topics and/or authors for our program to pursue, or if you’re interested in participating yourself, please email or

The Anatomy of Escape
Fighting Fascism
Markets Not Capitalism
Free Markets & Capitalism?
Organization Theory
Conscience of an Anarchist