Case Study 1: The Ethics of Assassination Markets and Where to Draw the Line

Overview

The assassination market thought experiment is the most extreme test case in prediction market ethics. It forces us to confront the limits of market-based information aggregation and to articulate principles for distinguishing acceptable prediction markets from morally impermissible ones. This case study traces the intellectual history of assassination markets, analyzes them through multiple ethical frameworks, examines real-world analogues and near-misses, and develops a practical taxonomy for drawing the line between permissible and impermissible prediction markets.

The Original Concept

Jim Bell's "Assassination Politics"

In 1997, crypto-anarchist Jim Bell published an essay titled "Assassination Politics" that described a mechanism for anonymous, decentralized political assassination. The core concept:

  1. A website collects anonymous, encrypted predictions of the date of death of a public figure.
  2. Contributors who correctly predict the date of death receive a payout funded by earlier contributions.
  3. The mechanism creates an implicit bounty: anyone who plans to kill the target can submit a prediction for the planned date, carry out the act, and collect the accumulated pool.
  4. The anonymity of cryptocurrency prevents tracing the connection between prediction and action.

Bell argued this mechanism would serve as a check on government power, making it too dangerous for officials to act against the public interest. He saw it as a form of direct democratic accountability enforced by market incentives.

Bell was later convicted of tax-related offenses and harassment of government officials. His essay, however, has had a lasting impact on discussions of prediction market ethics, serving as the canonical example of how market mechanisms can create perverse incentives.

The Mechanism in Detail

The assassination market mechanism works through the following steps, which we analyze both technically and ethically:

Step 1: Target Selection. Someone nominates a public figure. In Bell's vision, this was any government official, but the concept generalizes to any individual.

Step 2: Pool Accumulation. Anonymous contributors add funds to the pool. Each contributor may have different motivations: some may want the target dead, others may be hedging against the consequences of the target's death, and still others may simply be speculating.

Step 3: Prediction Submission. Contributors submit encrypted predictions of the date of death. The encryption ensures that predictions cannot be read until after the event occurs (or a verification date passes).

Step 4: Execution (literal). If someone has both the ability and willingness to kill the target, they submit a prediction for their planned date, carry out the act, and the prediction is decrypted and verified.

Step 5: Payout. The contributor(s) with the correct prediction receive the pool. The anonymity of the system protects the assassin from identification.

Technical Feasibility

With the development of anonymous cryptocurrencies (Monero, Zcash), decentralized smart contracts (Ethereum), and zero-knowledge proofs, the technical infrastructure for an assassination market is more feasible today than when Bell wrote his essay. This is not a hypothetical risk; the technical barriers have been substantially reduced.

However, practical barriers remain significant:

  • Pool size problem. For the incentive to be meaningful, the pool must be large. Accumulating a large enough bounty attracts attention from law enforcement.
  • Coordination problem. The market requires someone willing to commit murder. While the financial incentive exists, the vast majority of people will not kill for money, and those who will face enormous personal risk.
  • Trust problem. Contributors must trust that the mechanism will pay out correctly. In a decentralized system, smart contract bugs or oracle manipulation could prevent payout.
  • Detection risk. Law enforcement agencies monitor for exactly this type of activity. The combination of a public prediction market and an actual death would trigger an intensive investigation.

Ethical Analysis

Utilitarian Perspective

Against assassination markets: The utilitarian calculus is overwhelmingly negative. The direct harm (murder) is catastrophic and irreversible. The indirect harms include: - Fear and chilling effects on public service (fewer people willing to serve in government) - Destabilization of political systems - Erosion of the rule of law - Normalization of violence as a political tool - Psychological harm to the target, their family, and society

The consequentialist defense (Bell's argument): Bell argued that the threat of assassination markets would deter government corruption and abuse of power, ultimately producing better governance. This argument fails on several grounds: 1. The deterrent effect is speculative; the direct harm is certain. 2. The mechanism cannot distinguish between corrupt and non-corrupt officials. 3. It replaces democratic accountability with plutocratic violence (whoever funds the largest pool determines who lives or dies). 4. The long-term consequences of normalizing assassination as a governance tool are catastrophic.

Utilitarian verdict: Clearly impermissible. No plausible information value or deterrence benefit outweighs the direct harm of incentivizing murder.

Deontological Perspective

Kant's categorical imperative: The maxim "I will create a market mechanism that incentivizes the murder of individuals I disapprove of" cannot be universalized without contradiction. A world where everyone created such mechanisms would be one of perpetual violence and fear, undermining the conditions necessary for any social cooperation.

Dignity and rights: Assassination markets treat human beings as means to financial ends in the most literal sense: the target's life is commodified, and their death is converted into profit. This violates the fundamental Kantian principle that persons must be treated as ends in themselves.

Deontological verdict: Absolutely impermissible. Assassination markets violate the most fundamental moral prohibitions.

Virtue Ethics Perspective

What kind of society creates assassination markets? Virtue ethics asks not just about actions and outcomes but about character and the kind of community we are building. A society that tolerates assassination markets is one that has abandoned basic commitments to human dignity, non-violence, and the rule of law.

Character of participants: Contributing to an assassination market pool — even without intending to kill anyone — reflects and cultivates callousness, contempt for human life, and a willingness to externalize the costs of one's political frustrations onto innocent individuals.

Virtue ethics verdict: Unequivocally impermissible. Participation in assassination markets corrupts individual and collective character.

Contractarian Perspective

Behind the veil of ignorance: Would rational agents, not knowing whether they would be the market participant or the target, agree to a society with assassination markets? The answer is clearly no. The risk of being targeted (and the impossibility of defending against anonymous, decentralized threats) would make rational agents reject this institution.

Contractarian verdict: Impermissible. No rational agent would consent to an institution that makes them vulnerable to anonymous assassination.

From Extreme to Gray Area: The Severity Spectrum

The assassination market is the extreme case, but the principles it illustrates apply along a spectrum. We can organize prediction markets from clearly impermissible to clearly permissible:

Tier 1: Absolutely Impermissible

  • Markets on the death of specific, named individuals
  • Markets on specific terrorist attacks
  • Markets on the physical harm of identifiable persons
  • Markets designed to incentivize any criminal act

Principle: Markets that create direct, substantial financial incentives for violence against identifiable individuals are absolutely impermissible, regardless of information value.

Tier 2: Strongly Disfavored

  • Markets on the timing of an individual's resignation or removal (where sabotage is feasible)
  • Markets on specific corporate failures (where sabotage is feasible)
  • Markets on specific criminal events in identifiable locations

Principle: Markets where a market participant could plausibly cause the event, even at significant personal cost, require extreme caution. Position limits and surveillance may mitigate but not eliminate the risk.

Tier 3: Permissible with Safeguards

  • Markets on aggregate mortality statistics (pandemic death tolls, actuarial data)
  • Markets on election outcomes (very difficult for a single participant to influence)
  • Markets on corporate earnings (insider trading concerns, but manipulation of the underlying event is difficult)
  • Markets on geopolitical events (regional stability indices, not specific attacks)

Principle: Markets where the event is determined by broad social forces rather than individual action are generally permissible, provided adequate surveillance and position limits are in place.

Tier 4: Generally Permissible

  • Markets on weather events
  • Markets on economic indicators (GDP, unemployment, inflation)
  • Markets on scientific outcomes (replication, discovery)
  • Markets on technology milestones

Principle: Markets on events that no market participant could plausibly cause or significantly influence are generally permissible. The information value is substantial and the moral hazard risk is negligible.

Real-World Analogues and Near-Misses

The DARPA FutureMAP Controversy (2003)

The closest real-world approach to conflict prediction markets was DARPA's Policy Analysis Market (PAM), proposed as part of the FutureMAP program. While PAM would not have allowed direct betting on specific terrorist attacks (a common misconception), the public perception that it was a "terrorism futures market" led to its rapid cancellation.

The FutureMAP controversy illustrates several key principles: 1. Perception matters as much as reality. Even if a market is designed with appropriate safeguards, public perception of "betting on death" can be fatal to the project. 2. Framing is critical. "Geopolitical stability index market" and "terrorism futures market" describe similar instruments but evoke very different reactions. 3. Political dynamics override technical analysis. The merits of the market design were irrelevant once the political opposition mobilized.

Dark Web Prediction Markets

Several dark web platforms have offered prediction-market-like mechanisms on violent events. These platforms have been small, illiquid, and largely non-functional, but they demonstrate that the concept has moved from thought experiment to attempted implementation.

Law enforcement agencies in multiple countries have investigated and shut down such platforms. The combination of anonymity tools (Tor, Monero) and smart contracts makes this an ongoing enforcement challenge.

Life Insurance and Viatical Settlements

The existing life insurance and viatical settlement markets demonstrate that "betting on death" is already normalized in certain financial contexts. Life insurance companies profit when policyholders live longer than expected; viatical settlement buyers profit when policyholders die sooner.

The critical distinction is that these markets involve consenting parties (the policyholder agrees to the insurance contract or settlement) and serve a recognized social function (risk management). Assassination markets lack both consent and social function.

Drawing the Line: A Practical Framework

Based on the analysis above, we propose a five-factor test for evaluating the ethical permissibility of a prediction market:

Factor 1: Causation Potential. Can a market participant plausibly cause the event? - Score 0: Impossible (weather, large-scale economic indicators) - Score 1: Extremely difficult (national elections) - Score 2: Difficult but conceivable (corporate decisions, small elections) - Score 3: Feasible (individual-level events, small-scale events)

Factor 2: Harm Severity. How severe is the harm if the event is caused? - Score 0: No harm (weather, sports) - Score 1: Moderate harm (financial loss, reputational damage) - Score 2: Serious harm (injury, significant disruption) - Score 3: Catastrophic harm (death, mass casualty)

Factor 3: Information Value. How much social value does the market's price signal provide? - Score 0: No information value - Score 1: Modest value (entertainment, curiosity) - Score 2: Significant value (business planning, policy input) - Score 3: Critical value (disaster preparedness, life-saving decisions)

Factor 4: Consent. Have the subjects of the market consented? - Score 0: Explicit consent - Score 1: Implied consent (public figures, public events) - Score 2: No consent, but public event - Score 3: No consent, private matter

Factor 5: Mitigation. Are adequate safeguards in place? - Score 0: Comprehensive (position limits, surveillance, KYC, delayed settlement) - Score 1: Partial (some controls) - Score 2: Minimal (basic controls only) - Score 3: None

Ethical Risk Score:

$$\text{Risk} = (\text{Causation} \times \text{Harm}) + \text{Consent} - \text{Information Value} - (3 - \text{Mitigation})$$

  • Score 0 or below: Generally permissible
  • Score 1-3: Permissible with enhanced safeguards
  • Score 4-6: Strongly disfavored; exceptional justification required
  • Score 7+: Impermissible

For the assassination market: Causation = 3, Harm = 3, Consent = 3, Info Value = 0, Mitigation = 3. Risk = (3 x 3) + 3 - 0 - (3 - 3) = 12. Clearly impermissible.

For a hurricane prediction market: Causation = 0, Harm = 0, Consent = 0, Info Value = 3, Mitigation = 0. Risk = (0 x 0) + 0 - 3 - (3 - 0) = -6. Clearly permissible.

Discussion Questions

  1. The line-drawing problem. Where exactly does the line fall between Tier 2 (strongly disfavored) and Tier 3 (permissible with safeguards)? Is there a bright-line rule, or must every market be evaluated individually?

  2. Decentralized enforcement. If assassination markets are deployed on immutable smart contracts with no identifiable operator, what enforcement mechanisms are available? Is code deployment itself a criminal act?

  3. The information value of the extreme. Some argue that even assassination markets have information value: a rising price signals growing danger to the target, potentially saving their life. Does this argument have any merit?

  4. Cultural relativism. Are the ethical conclusions reached above universal, or do they reflect a particular cultural perspective? How would this analysis differ in a society with different norms about violence and political accountability?

  5. The technology imperative. Given that the technology to build assassination markets exists and cannot be uninvented, should the focus shift from prevention to mitigation? What mitigation strategies are available?

Computational Exercise

The chapter's code files include an implementation of the five-factor ethical risk scoring framework (code/case-study-code.py). Use it to evaluate the following markets:

  1. "Will the CEO of Company X resign within 6 months?"
  2. "Will there be a mass shooting in the US in Q3 2026?"
  3. "Will the global average temperature exceed 1.5C above pre-industrial by 2030?"
  4. "Will Paper Y replicate in a direct replication attempt?"
  5. "Will Politician Z win the upcoming election?"

Compare your scores to those generated by the framework and discuss any disagreements.