An authoritarian regime is sustained not only by the dictator but also by thousands of rank-and-file members. Some take part in the repression reluctantly, fearing immediate loss; others can no longer stop because they are ‘tainted’; and still others make a career out of persecution. Behavioural economics and game theory explain why each persists and why opting out is so hard. They also suggest what might break this logic. Written by Maksim Znak.

Complex social systems are like living organisms, and people tend to simplify them by exaggerating the role of those who are in the spotlight. We are accustomed to associating authoritarian regimes with the odious figure of the dictator, but their base is provided by the grassroots level — ordinary functionaries, ‘cogs’: lower and middle-ranking officials, investigators, security forces, lecturers, judges, specialists at state-owned enterprises. Their petty, often invisible operations cumulatively produce a repulsive overall result, and through passivity, inertia, and the forced rationalisation of their own choices, the cogs of the system turn in unison with the state’s repressive mechanism.

The “khodiki”¹ are not homogeneous. Most participate in the repression without enthusiasm, merely to avoid trouble. Some are already so “tainted” that it is dangerous for them to stop. And for others, the persecution has become a social ladder. The behaviour of the first is largely explained by cognitive distortions, while that of the second and third is entirely rational calculation within the framework of the “corrupted mathematics” of incentives that the regime constructs. The previous text focused on the socio-psychological mechanisms of repression; this one focuses on their economics and game theory.

The Reluctant: The Flawed Mathematics of Risk

Daniel Kahneman and Amos Tversky, in prospect theory, showed that decisions under risk systematically deviate from the model of the rational agent. People assess outcomes relative to their current situation, and perceive losses as roughly twice as painful as comparable gains. Furthermore, as Kahneman describes in his book “Thinking, Fast and Slow”, people evaluate each decision individually, “here and now”, and fail to see that a series of small “safe” choices adds up to much greater risk.

For a Belarusian official after 2020, it looks like this. You have to sign a false document, dismiss a dissenter, or approve a repressive initiative. Refusal means a guaranteed and immediate loss: loss of a bonus, contract termination, and possible administrative or even criminal prosecution. Consent promises the preservation of the status quo and only a hypothetical, deferred responsibility — should the authorities ever change, and the matter come to a reckoning.

When choosing between a guaranteed loss and the risk of an even greater loss, people usually choose the risk — this is the so-called reflection effect. Prospect theory also predicts an overestimation of small probabilities, which ought to be frightening in terms of future liability. But when the risk is known from personal experience — “years have passed, and no one has been punished” — rare events, on the contrary, are underestimated. This is compounded by unrealistic optimism: “it could happen to someone else, but not to me”. Thus, by avoiding an immediate inconvenience, a person takes the first step and becomes a hostage to their own actions.

 “Tainted”: Hostages to Their Own Actions

The first unlawful act becomes a lever for subsequent ones: for a prior offence, the disobedient can be punished by the state itself. Everyone in the system understands this, so unlawful acts are concealed: prisoners’ letters are burned without written orders, violations are fabricated, schemes are devised to cover up crimes. The perpetrators know: the wind can change at any moment, and the command “Sic ’em!” can be sounded against an individual or an entire group — as already happened, for example, with the ‘Ezhovites’ in 1938–1939. When the time of impunity ends, dishonest behaviour will result in criminal liability for complicity in crimes, loss of professional reputation, and moral degradation. This fear does not restrain but pushes on: there is nowhere left to retreat.

In game theory, this mechanism is called a “hostage” — a tool of credible commitment. Diego Gambetta has shown that in criminal circles, mutual incrimination guarantees trust: the one who holds compromising material will not betray. Keith Darden described how a “blackmail state” is built on this principle: the authorities deliberately record the misconduct of their officials in order to keep them on a leash. For such a system, the “tarnishedness” of the enforcer is not a side effect, but a tool of control.

There is reason to believe that the number of officials involved in the repression increases over time: these are no longer “special” judges or investigators, but everyone in turn. The system needs not born executioners, but a broad circle of “compromised” people. The noose tightens according to the “foot-in-the-door” principle: the first concession seems trivial — putting a signature, staying silent at a meeting, conducting a “preventive talk” — but after it, a person sees themselves as someone who ‘has already agreed’, and the next one is easier to make.

Next comes the “escalation of commitment”: having started to take part in a dirty business, a person invests more and more in it and justifies past actions with even greater loyalty. The classic escalation is an irrational attachment to sunk costs, but in a repressive apparatus, loyalty can be entirely rational: leaving the game has become genuinely dangerous.

The system rarely issues written criminal orders: it communicates its wishes orally, leaving the executor alone with the legal responsibility. Every episode can be used for blackmail, and supporting the regime becomes a matter of self-preservation. Those ‘tainted’ do not trust guarantees of amnesty from democratic forces for two reasons. The first is psychological: people tend to assume others act like them, and someone who has broken promises expects the same from others. The second is the “commitment problem”: a future government cannot credibly promise today that it will not punish tomorrow.

Careerists: the economics of persecution

For some in the security forces, repression is not a forced necessity, but a profitable business. Lower-level officials, accustomed to rewards for “fighting political enemies”, often become the main driving force behind persecution. Real crime fighting is labour-intensive, dangerous, and not always effective, whereas repression carries minimal risk and offers guaranteed rewards: stars on the epaulettes, early promotions, and bonuses. For example, by mid-August 2020, more than 300 Ministry of Internal Affairs personnel — from commanders to prison van drivers — had received departmental medals.

The incentive grows when a career depends on quantitative metrics: according to “Campbell’s Law”, any metric used for decision-making begins to distort what it is meant to measure — and it becomes advantageous for the security forces to fabricate enemies. Added to this is the competition for loyalty, which Ian Kershaw called “working towards the Führer” and which was discussed in the previous text. References to the law or economic feasibility are perceived as betrayal, which completely devalues the voice of the non-security wing of the nomenklatura.

The Stag Hunt: Why No One Wants to Be First

So why don’t the “khodiki” refuse en masse? If a majority of judges, members of electoral commissions or teachers were to refuse at the same time, the regime would be unable to punish every single one of them — just as it is impossible to jail hundreds of thousands of protesters or to mass-dismiss public sector workers. But no one wants to be among the few who are punished: to lose a bonus, their job or a council flat, or to pay the state a huge fine for “education” and “training”.

This situation is described by a coordination game — the “stag hunt”. The two hunters independently choose between risky cooperation and safe selfishness. Together they can catch a deer. But if one pursues the modest, yet guaranteed, rabbit, he will catch it, and the trusting partner will be left with nothing. Therefore, in the absence of trust, both hunt rabbits. Unlike the prisoner’s dilemma, cooperation here is beneficial for everyone — if they are confident in one another. The problem is not with incentives, but with trust and information.

Therefore, the regime fights not only protests but also independent media, chat groups and any horizontal connections: it destroys not so much motivation as mutual trust. Mark Granovetter showed that collective action depends on the distribution of individual “thresholds” — how many people have to step forward first for the next one to join. Timur Kuran added to this the “falsification of preferences”: under repression, people publicly express views that differ from what they actually think, and no one knows the true level of discontent until the cascade suddenly sets off. Exemplary repressions raise the threshold for everyone else — which is why they are effective.

Another game, the Prisoner’s Dilemma, formulated in 1950, accurately describes another situation within the system: informants. When several perpetrators witness the “extremists” forbidden behaviour in prison, or hear a “seditious” conversation between colleagues, it is in everyone’s interest to be the first to report it: whoever remains silent risks becoming the next target. So, time and again, a denunciation is chosen — even though the atmosphere of denunciation ultimately harms everyone.

Free-riders and the Way Out of the Trap

According to Mancur Olson’s theory of collective action, rational individuals do not act collectively when the group is large, and the outcome benefits everyone, including the ‘free-riders’ who risk nothing. This applies to officials as well: if the system changes, everyone wins, so there is no point in taking personal risks. However, small, cohesive groups with personal (‘selective’) incentives organise easily. The security apparatus is precisely such a group: relatively small, hierarchical, and personally rewarded, whereas society and civil servants are large and diffuse. This is how Olson explains not only the passivity of the majority, but also the advantage of the minority.

Some flee the country, others stand frozen in silent expectation, hoping that the machine of repression will pass by their door. Albert Hirschman described three possible reactions to the decline of an institution — “exit”, “voice” and “loyalty”: when exit is available, the most active leave first, and internal voice weakens. “Exit” can be either dismissal or emigration. Interestingly, however, as Hirschman later demonstrated with the example of the GDR, in 1989 mass emigration, on the contrary, strengthened the “voice”, as it made the discontent visible to all.

The passivity of those who remain recalls Martin Seligman’s learned helplessness: having convinced themselves that outcomes do not depend on their efforts, people stop trying even when circumstances change. This is an individual model, and we can apply it to society only as an analogy. But Seligman’s later work showed that passivity is the default reaction to prolonged threat, and it is a sense of control that is “learned”. Therefore, even a small experience of successful action is enough to break the sense of helplessness.

Until trust is restored, society remains in a “rabbit” equilibrium: everyone avoids personal punishment “here and now”, while the entire country receives a massive prison sentence as the legal default, repressive suffocation and a loss of agency. General knowledge and visible signals help to break out of such a balance: when everyone knows that others are also not supporting the repression, and knows that others know this, the thresholds are lowered. Therefore, every visible refusal — including that of the “khodiki” themselves — changes the calculation for everyone else.

For the “tainted”, however, the main obstacle is a lack of trust in the promises of the future authorities. Monika Nalepa, using the example of Poland, the Czech Republic and Hungary, showed that a promise becomes credible when the person making it has a vested interest in fulfilling it: the opposition there had “skeletons in the cupboard” — agents of the secret services in its ranks — and was in no hurry to carry out lustration itself. For Belarus, this means guarantees must be a mechanism, not a declaration: pre-published, differentiated criteria for accountability; visible protection for those who refused to carry out criminal orders; and international guarantees. For those within the system, a credible prospect becomes crucial — both accountability for those who carry on, and differential treatment for those who stop.

¹‘Khodiki’ (‘The Walkers’) — a term introduced by the author in the preceding text, ‘The Innovation of Evil’, It refers to people who carry out dishonest or harmful acts, sometimes against their will or under a veil of indifference, as instruments of a repressive system.