Why the Left Needs Game Theory
A Framework You Were Taught to Dismiss Is One You Can’t Afford to Ignore
There is a version of this argument that sells out. It goes: game theory is just economics, economics is just capitalism with math, and the left should develop its own frameworks rather than borrowing the tools of the enemy. That argument is wrong, and the left pays a real price for accepting it.
Game theory is not economics. It is a framework for analyzing strategic interaction. It explores situations where what happens to you depends on what other people do. It applies as readily to strike coordination, coalition building, and revolutionary organizing as to market competition. The left has allowed bourgeois economics to colonize a neutral analytical tool, and then rejected the tool because of who was using it.
The cost of that rejection is not abstract. It shows up in organizing failures, in political coalitions that collapse in predictable ways, and in the recurring spectacle of the left building institutions intended to liberate people that end up reproducing the hierarchies they were designed to replace. Without a framework for asking what incentive structures its institutions create, the left navigates like a Roomba in the dark. While sincere in its purpose, energetic in its motion, it seems fundamentally unable to learn from the walls it keeps hitting.
Two of the most serious attempts to build a liberated society illustrate exactly this problem. The Bolsheviks built a structure in 1917 that they believed would end class exploitation, and produced instead a technocratic elite that replaced the aristocrats they had just overthrown. Murray Bookchin spent decades developing the most sophisticated libertarian institutional theory the left has produced, and left it without the tools to prevent the administrative capture he most feared. Neither failure was inevitable. Both were predictable by a framework the left has largely refused to use.
What Structural Analysis Does Well — And Where It Stops
Marxist and structural analysis excels at identifying the large-scale forces that shape political possibility. It explains why capitalism generates crises, why ruling class interests converge across apparent political differences, why reformist parties get absorbed, why empires overextend. It identifies the contradictions that accumulate beneath the surface of stable-seeming systems and explains why ruptures, when they come, take the forms they do.
This is powerful analytical work, and no serious political analysis can do without it.
But it has a persistent blind spot at the level of mechanism. Structural analysis tells you that capitalism creates the conditions for working-class solidarity. It does not explain why, in any specific situation, individual workers defect from solidarity rather than maintain it. It tells you that ruling class unity is structurally advantageous. It does not specify the precise mechanism by which that unity is achieved against the short-term interests of particular capitalists. It predicts crisis without tracing the path.
The gap is not a minor technical omission. It is the difference between knowing what direction history pushes and understanding why actors, caught inside those structural forces, make the specific choices that produce specific outcomes. Game theory is the tool built to close that gap.
When Missing the Mechanism Cost Everything
The Bolsheviks inherited from Marx and Engels a clear theory of class formation: classes are constituted by their relationship to the means of production. Abolish private property in the means of production, and you eliminate the material basis for class formation. The state would wither. Privilege would dissolve. The new society would not reproduce the hierarchies of the old.
What the theory missed is that control over resource allocation is itself a form of property, independent of legal ownership.
The Soviet administrator who decides who gets housing, who gets promoted, who gets a travel permit, who gets scarce consumer goods — that person holds something functionally equivalent to property even if the deed is held by the state. They have power over outcomes. They have information the center doesn’t have. And they have interests that diverge from the interests of the workers they nominally serve.
Game theory has a precise name for this structure: a principal-agent problem. The principal in this case would be the state, the party, or the workers collectively. The principal delegates decisions to agents such as administrators, or collective representatives who have private information the principal cannot verify. Every agent faces the same calculus: use that informational advantage for the principal’s benefit, or use it to extract personal advantage. When the rewards for personal extraction are high, when monitoring is difficult, and when the system provides no price mechanism to reveal whether allocation decisions are efficient, the rational response for agents at every level of the hierarchy is to extract rents.
Multiply this across millions of administrative decisions such as housing allocation, factory management, party appointments or foreign trade and you have not abolished class stratification. Instead you have rebuilt it through the back door, constructed entirely out of information asymmetries rather than formal property titles.
The second mechanism is the coordination game of elite reproduction.
Once a stratum of administrators exists with shared interests in maintaining administrative power, the problem of elite reproduction solves itself, not through conspiracy, and not because administrators are worse people than anyone else, but because the incentive structure makes self-perpetuation the path of least resistance at every decision point.
A parent with preferential access to educational institutions who does not use it for their child is making an active sacrifice against immediate family interest. An administrator who fills appointments on pure merit when patronage costs them nothing and loyalty pays dividends is leaving value on the table by the logic of their position. An incumbent who does not raise the barriers to challengers from below is voluntarily accepting risk that selective opacity would eliminate. None of these behaviors require malice. They require only that people respond to incentives the way people respond to incentives.
This is the point the moralistic analysis of Soviet bureaucracy consistently misses. Attributing the nomenklatura to the wrong people being in charge. They blame the corrupting influence of power on individuals who should have known better. This is not just historically inaccurate, it is strategically catastrophic, because it implies the solution is better people rather than different structures. Put saints into the same incentive architecture and you get the same equilibrium on a longer timeline.
Unanalyzed incentive structures encouraged the Soviet nomenklatura to become a self-reproducing class with its own schools, hospitals, vacation facilities, and supply networks, formally invisible in law but materially decisive in practice. Milovan Djilas, a Yugoslav communist who watched this process from inside, named it in 1957: the party bureaucracy had constituted itself as a new ruling class.
The Bolsheviks did not intend this outcome. But intention is precisely what game theory teaches you to set aside. The question is not what actors want. The question is what structure of incentives they are embedded in, and what equilibrium that structure produces. The equilibrium produced by centralized administrative allocation with unmonitorable agents is rent extraction and elite formation, regardless of the ideological commitments of the administrators. This was predictable as a structural tendency. This is the kind of prediction Marxist analysis makes about capitalism all the time, applied here to the structure the Bolsheviks were building.
Bookchin’s Blind Spot
Murray Bookchin is the most important libertarian institutional theorist the left produced in the twentieth century. His framework is called libertarian municipalism. It is composed of nested confederations of directly democratic assemblies, policy-making kept in the assembly and administration kept strictly subordinate. Libertarian municipalism became the theoretical foundation for the Rojava experiment in northern Syria after Abdullah Öcalan encountered his work while imprisoned and adapted it into democratic confederalism. Bookchin’s ideas are being implemented at scale by millions of people. They deserve serious analytical scrutiny.
Bookchin came out of the anarchist and Frankfurt School traditions, and his intellectual formation was Hegelian Marxism filtered through Kropotkin. He was committed to the idea that human beings have a latent cooperative nature that hierarchical society suppresses, and that the right institutional conditions would allow that nature to express itself. His social ecology framework is fundamentally a theory about the reconciliation of humanity with itself. It is not a theory about strategic interaction under conditions of conflicting interest.
This produced a specific and consequential blind spot. Bookchin’s confederal assemblies assume that participants arrive with roughly cooperative orientations, that deliberation produces genuine consensus, and that administrative structures subordinate to assemblies will remain subordinate. He had a sophisticated theory of institutional form but no theory of institutional capture. He did not systematically ask what happens when participants in an assembly have interests that diverge from the collective good, what payoff structures make defection from cooperative norms individually rational, or what equilibrium an administrative apparatus produces when its agents have private information and incentives misaligned with the principal they serve.
The Soviet problem was administrative elites reconstituting class power through positional control rather than formal ownership. This is exactly the failure mode his confederal design needed to guard against. His framework addresses it primarily through institutional form and political education rather than incentive analysis. Keep the assemblies sovereign. Keep the administrators accountable. Cultivate ecological consciousness. These are necessary conditions. They are not sufficient ones, because they do not specify the mechanisms by which accountability is maintained when agents have every incentive to obscure their behavior from the principals nominally above them.
The Rojava experiment has encountered this problem in practice. The tension between the ideological vanguard structures inherited from the PKK and the genuinely horizontal assembly structures the ideology prescribes is real and persistent, with the former tending to dominate the latter under conditions of military pressure. Bookchin’s framework can identify this as a deviation from the ideal. It cannot fully equip practitioners to prevent it, because it lacks the tools to analyze why the deviation is the individually rational choice for actors inside the structure.
A Bookchin who had internalized game theory would have asked different design questions. Not just “how do we keep assemblies sovereign” but “what are the payoff structures facing administrators, and how do we make transparency and accountability the individually rational choice rather than the heroically virtuous one.” That is a different kind of institutional engineering. It produces designs held together by structural mechanism rather than by the virtue of the people inside them.
The Good People Delusion
The most persistent obstacle to structural thinking on the left is not theoretical. It is psychological. Across a century of organizational failures, the explanation that keeps returning is the same one: the wrong people got into positions of power. The revolution was betrayed by Stalin, not by the incentive architecture Stalin inhabited. The party was captured by careerists, not by the principal-agent dynamics that make careerism rational. The solution, implicitly or explicitly, is always the same: find better people.
This is not analysis. It is a just-world hypothesis wearing analytical clothing.
Carl Jung spent his career mapping the terrain that this explanation refuses to enter. He developed the concept of the Shadow which is the repository of traits, impulses, and capacities that the conscious self disowns and projects outward. This is not a metaphor for individual pathology. It is a structural feature of human psychology.
Every person capable of genuine altruism carries the capacity for self-serving rationalization. Every committed revolutionary carries the capacity for authoritarian imposition. The Shadow does not disappear when someone holds correct political views. It relocates into blind spots, into righteous justifications and into the comfortable certainty that one’s own exercise of power is categorically different from the power one is fighting.
Jung was explicit about what happens when individuals or movements refuse to acknowledge the Shadow: it does not go away. It operates unconsciously, which makes it more dangerous, not less. The revolutionary who has not confronted their own capacity for domination does not thereby lack that capacity. They lack only the self-knowledge to recognize when they are exercising it.
This is not a counsel of despair about human nature. It is a counsel of precision about institutional design. The relevant question is never whether the people running an institution are good or bad — a binary that psychological science has spent decades dismantling. It is what the institution’s incentive structure does with the full range of human motivation that real people actually bring to it.
Philip Zimbardo’s Stanford Prison Experiment demonstrated in controlled conditions what history demonstrates at scale: ordinary people, randomly assigned to roles with asymmetric power, rapidly develop behaviors consistent with those roles rather than their prior personalities. Guards became authoritarian. Prisoners became compliant and broken. The assignment of role was sufficient. Character was not the determining variable. Zimbardo later extended this framework in The Lucifer Effect to argue that Abu Ghraib was the predictable output of a situation that made cruelty the path of least resistance.
That argument has been used to exonerate the soldiers who carried out the abuse which is precisely the wrong conclusion. The structural explanation does not dissolve individual responsibility. It extends it upward, to the commanders who designed the conditions, the lawyers who authorized the techniques, and the officials who created the institutional permission structure that made cruelty normal. Situational causation does not end accountability. It relocates it to where the power actually was.
Hannah Arendt arrived at a compatible conclusion from a different direction. Her concept of the banality of evil was drawn from the Eichmann trial but it is applicable far beyond it. She made the observation that the most systematic harms are not produced by monsters. They are produced by ordinary people following institutional logic, meeting performance metrics, serving organizational interests that have been severed from their human consequences. The evil is in the structure. The person is almost incidental.
Stanley Milgram’s obedience experiments add the final piece. When an authority structure defined delivering electric shocks as the correct behavior, the majority of participants complied. These were ordinary people not clinical sadists, but because the situation defined compliance as normal and resistance as deviance. The proportion of people willing to harm others under institutional pressure is not a small minority of bad actors. It is most people, under the right structural conditions.
The veterans who return from wars marked by atrocity and take their own lives in disproportionate numbers confirm this from the other side of the experience. Some of them enjoyed what they did at the time. The situation didn’t merely extract reluctant compliance, it activated capacities that felt, in the moment, like power and purpose. Jung would recognize this immediately: the Shadow, when the institutional structure finally gives it permission to operate, does not feel like corruption. It feels like release.
What breaks these men later is not that they were forced to act against their nature. It is that they discovered what their nature contains when the situational pressure is removed and the structure that normalized the behavior is gone. The guilt is lethal precisely because the person doing the killing was recognizably themselves. That asymmetry, consisting of situational activation along with individual consequence, is one of the most important and least discussed features of how institutions shape human action.
What this body of evidence points toward is not cynicism about human beings. It is a specific and actionable conclusion: institutional design cannot be premised on the assumption that good intentions do the structural work. Intentions do not hold the structure up. Incentive structures do. The question to ask of any institution is not “will good people run this?” They might, for a while but the important question is “what does this institution do with bad intentions when they arrive, and what does it do with good intentions when those intentions conflict with positional interest?”
The left that builds institutions on the first question will keep being surprised. The left that builds institutions on the second question is doing structural engineering rather than moral aspiration, and its designs will outlast the intentions of the people who build them.
Three Game Theory Concepts Worth Understanding
The Prisoner’s Dilemma
Two workers are deciding whether to join a strike. Each would individually prefer not to strike because it costs wages and risks retaliation. But each knows that if neither strikes, they both lose bargaining power. If one strikes and the other doesn’t, the striker bears all the cost while the scab gets whatever gains follow. If both strike, they share the cost and share the gain.
This structure, where individually rational choices produce a collectively worse outcome, is the prisoner’s dilemma. Neither worker is stupid or cowardly. The problem is not consciousness. The problem is the payoff structure.
This reframes the analysis of scabbing entirely. Scabbing is not primarily a moral failure. It is the individually rational response to a payoff structure that employers deliberately construct. At-will employment, company towns, the destruction of union density and the elimination of strike funds all work by making the defection payoff higher and the cooperation payoff lower.
The left that understands this fights differently. It focuses on changing the payoff structure: strike funds that raise the cost of continuing to work, mutual aid networks that lower the cost of striking, legal frameworks or extra judicial escalation ladders that raise the cost of employer retaliation. The left that doesn’t understand it delivers speeches about betrayal and keeps losing.
Nash Equilibrium
A Nash equilibrium is a state where no actor can improve their outcome by changing strategy, given what everyone else is doing. Nobody chose it or designed it. The incentive structure selects for it.
Capitalism reproduces itself not because capitalists are all-powerful or workers are entirely deceived, but because the existing arrangement is a Nash equilibrium. Any individual deviation from it, absent collective action, makes the deviating actor worse off. This is why individual ethical consumption doesn’t work. Buying fair trade coffee while the rest of the market does not is individually virtuous and structurally irrelevant. The equilibrium itself is intact.
Nash equilibrium also explains ruling class coordination. Capitalists genuinely compete with each other and have real conflicts of interest. But on questions of labor law, union suppression, and state power, they coordinate with remarkable consistency. The reason is that the payoff from collective action against labor is large, and the cost of defecting from that coalition is high. Ruling class unity is the Nash equilibrium of their structural situation. It is not conspiracy. It is incentive structure, and it is far more durable than conspiracy would be.
Coordination Games and Focal Points
Not all strategic problems are prisoner’s dilemmas. Some are coordination games, where the problem is not defection but synchronization. Everyone wants to act together. The question is who moves first, and how others know to follow.
Revolutionary moments are coordination problems of this kind. Workers who would all prefer collective action to continued exploitation nonetheless fail to act because each is uncertain whether others will move. The obstacle is not individual preference. It is uncertainty about collective response.
The game-theoretic solution is a focal point . A focal point is a salient event that makes it common knowledge that others are watching and ready to move. Rosa Parks refusing to give up her seat was not the first such act of resistance on Montgomery’s buses. It became a focal point because its visibility made resistance common knowledge, which made coordination possible.
This reframes the role of the organizer. The organizer is not primarily a persuader trying to change individual preferences one at a time. The organizer is a focal point creator . They are someone who reduces uncertainty about collective response, who makes collective action legible as a possibility, who shifts the coordination equilibrium by making it common knowledge that others are willing to move. Understanding this distinction focuses effort on visibility and the infrastructure of common knowledge rather than the endless conversion of individuals who were already willing.
What Game Theory Cannot Do
Game theory takes the payoff structure as given. It explains what equilibrium a structure produces; it does not explain where the structure came from or why the payoffs are what they are. For that you need historical and structural analysis. The prisoner’s dilemma facing workers in a right-to-work state exists because capital won decades of legislative battles. Game theory models the dilemma. It cannot explain the class power that created it.
Game theory also has limited tools for explaining how preferences change. The worker who becomes a committed socialist has changed what they value. Game theory can model behavior given new preferences. It cannot explain the political education and collective experience that produced them. That remains the domain of ideology, culture, and organizing. These are the irreplaceable work that game theory does not touch.
The frameworks are not alternatives. They are instruments with different domains of competence. Structural analysis explains what contradictions accumulate and what transformations become possible. Game theory explains why, given those structural conditions, specific actors make the specific choices that produce specific outcomes. Neither is complete without the other.
Navigating in the Dark
The left is not lacking in moral clarity, historical knowledge, or genuine commitment to a more egalitarian and peaceful world. What it has consistently lacked is a precise enough map of the terrain between structural conditions and human outcomes. It lacks a way of asking, at the level of individual actors embedded in specific incentive structures, why people do what they do and what institutional designs would produce different behavior.
Without that map, the left navigates like a Roomba in the dark. It moves with energy and purpose. It covers a great deal of ground. And it keeps colliding with the same obstacles. It diagnoses scabbing as betrayal, administrative capture as ideological deviation and coordination failures as insufficient consciousness. It lacks a framework for understanding why these outcomes are the equilibria its structures keep selecting for.
The Bolsheviks hit the wall of principal-agent dynamics in centralized allocation and called the result Stalinism. Bookchin identified the wall of administrative capture and called the solution political education. The left has been running the same pattern across a century of attempts: build institutions based on structural analysis and normative aspiration, watch those institutions produce outcomes the theory didn’t predict, respond with moral condemnation of the actors rather than redesign of the incentive structures.
Game theory does not complete the left’s analytical project. It does not replace the structural analysis of capitalism, the theory of ideology, or the political education that builds genuine solidarity. What it does is turn the lights on for one critical part of the problem: the mechanisms by which actors embedded in structures make choices, and the institutional designs that make cooperation the rational option rather than the heroic one.
The tools exist. The walls are not going anywhere. The question is whether the left will keep navigating by collision, or start drawing the map.
