“Communication Failure: The Hidden Face Of The Tragedy Of The Commons” is a forthcoming paper by Philippe Colo and Guillaume Pommey, which was a Global Priorities Institute Working Paper. This post is part of my sequence of GPI Working Paper summaries.

The main takeaway: When self-interested, rational individuals contribute to a public bad, they will often ignore true information from a socially interested expert that is too complex for them to personally verify.

Mini-summary:

Introduction, methodology, and caveats

• Colo argues that the credibility of scientists’ forecasts depends on the public’s confidence in them, and hence their forecasts end up mostly ignored in decisions, even if the scientists are socially interested.
• Colo designs a theoretical model of many contributors to a public bad who decide how much they contribute to the problem after they receive an expert forecast that is too complex for them to verify themselves. The expert aims for social welfare; the contributors aim for their individual good.
• It features only one expert, so the findings only apply when there is a unified information source (e.g., perhaps the IPCC or WHO). It remains to be said whether multiple experts might improve or reduce information transmission.

Findings and key takeaways:

1. Three characteristics of large-scale public goods problems each can cause the contributors to ignore the expert, even if (in some cases) listening to the expert would increase social welfare:
   1. A high number of contributors
   2. Uncertainty about the damage caused by their contributions to the problem (except for the expert’s information)
   3. Contributions that create a low amount of damage on the margin
2. Even with a small number of contributors, as long as the problem’s severity is low enough, they still ignore the expert.
3. The self-interested contributors fail to internalize their actions’ consequences on others, creating a disconnect between their beliefs and the information from the socially-interested expert. This disconnect leads them to distrust and ignore the expert’s information.
4. Colo also considers experts who favor the interests of the individuals most exposed to the problem, finding their information is less likely to be ignored, especially if the contributors have highly heterogeneous (varied) preferences. (Suggesting that experts with Rawlsian preferences, as opposed to utilitarian ones, are more effectively able to communicate with contributors.)
5. Effectively informing the population of the truth is not always what is best for its well-being; there are situations where experts have to choose between informing the population and increasing its welfare.
6. Experts are better able to communicate with small groups (e.g., politicians) when they are more uncertain about the range of possible damage (e.g., climate forecasts) than when the spread of possible consequences is smaller (e.g., COVID-19 infection forecasts).