Hidden Assumptions: The Ethical Traps of Pandemic Models
What if the mathematical models used to decide whether you can leave your house during a pandemic were hiding a series of disturbing ethical traps? While we often view viral simulations as objective mirrors of reality, a new analysis reveals these tools contain profound philosophical choices.
The Italian Case Study & A Troubling Framework
Writing in a framework calibrated to Italy's "Patient Zero" context, researchers Silvio Vanadia and Charles Shaw explored how social planners use Rank-Discounted Critical-Level Utilitarianism (RDCLU) to navigate impossible crisis choices.
The Model's High Stakes
Their study utilized a social planner’s model with a 5% annual interest rate (). The key finding is immense: when the model accounts for the cost of death rather than just lost productivity, the total welfare costs of a pandemic are roughly three times higher.
Key Findings and Fragile Outcomes
The research, based on a population where 98% were initially susceptible (), uncovered critical insights about lockdown strategies and their underlying logic.
The "Optimal" Lockdown
Under benchmark parameters, an optimal lockdown strategy was found to reduce long-run mortality by approximately 0.80% compared to a scenario with no controls.
The Productivity Collapse Trap
The effectiveness of lockdowns is fragile. In models where testing for recovered individuals is absent (), the simulation forces an abrupt end to restrictions. This occurs because healthy, immune workers are trapped at home, leading to a productivity collapse without further reduction in viral spread.
The Overwhelming Pressure Point
When the fatality rate scales to reach 3% when 40% of the population is infected, the pressure on the "social planner" to enforce stringent lockdowns becomes overwhelming.
The Core Ethical Problem: Hidden Value Judgments
The analysis argues that the difference between a lockdown and a "let it rip" strategy isn't just based on biology—it’s based on a philosophical choice often made by programmers rather than ethicists.
The "Repugnant Conclusion"
The researchers found that the way a model values a person can lead to what philosophers call the "Repugnant Conclusion"—a scenario where a massive population with barely any quality of life is mathematically preferred over a smaller, flourishing one.
The "Ruthless" Utilitarianism Warning
They warn that letting these "workhorse tools" run on autopilot leads to paradoxical or unacceptable policy outcomes. "Ruthless" utilitarianism, baked into the code, makes life-and-death decisions based on "hidden assumptions" about the value of human life.
Limitations and The Path Forward
While the proposed RDCLU framework offers a mathematical escape from these logical traps, the model itself faces significant hurdles that must be addressed.
Critical Model Gaps
- Lack of Age Structure: The current simulation lacks granular age data, creating "unrealistic impacts on mortality risk" given COVID-19's lethality is heavily skewed toward older populations.
- Sensitive Core Metric: The results are deeply sensitive to the "Value of a Statistical Life," which the study calibrated at 20 times annual GDP.
The Final Takeaway: Until ethicists and mathematicians sit at the same table, the hidden assumptions baked into our pandemic responses will continue to make profound decisions for us, long before the first real-world data point is even collected.
Reference: Utilitarianism on the front lines: COVID-19, public ethics, and the "hidden assumption" problem. Silvio Vanadia and Charles Shaw. arXiv:2205.01957v1 [econ.GN] 4 May 2022.