Jon Murphy: Cascading Expert Failure
This is a guest post by EPERN member Jon Murphy. We’ve invited Jon to follow up on his December 2021 EPERN research presentation Cascading Expert Failure. His paper has been recently accepted for publication in the Journal of Institutional Economics.
Written by Jon Murphy
Experts are generally seen as high-information individuals. We employ experts throughout our lives in order to help us make decisions: mechanical experts help fix our cars, religious experts help prepare us for the Afterlife, medical experts help care for us in this life, etc. In his 2018 book Expert Failure, Syracuse University finance professor Roger Koppl explored how and when experts can fail in their duty as advisors. Different institutional arrangements, such as monopoly expert power and jural authority, can cause experts to fail by ignoring alternative opinions to their own.
Roger’s work places the expert within a complex institutional framework of knowledge and information generation and how that complexity plays on the expert’s behavior. In a forthcoming paper in the Journal of Institutional Economics, I expand on Roger’s framework by placing the expert within a network of other experts beyond their field of expertise. I argue experts interact with, but often misunderstand the importance of, information coming from outside their disciplines. Consequently, experts become more likely to fail in their advice if they do not interact with other disciplines. Additionally, if experts fail in their discipline, given the interconnectedness of knowledge networks, their failure can cascade into other disciplines, and cause multiplying failures throughout the economy.
The theoretical model evolved out of mainstream network analysis. Following the 2008 financial crisis, numerous papers arose exploring the necessity of auto bailouts. Daron Acemoglou and coauthors, in a 2012 Econometrica article developed a theoretical framework of when failures in production networks would average out in the aggregate (as the Law of Large Numbers predicts) and when they would not. David Baqaee, in the pages of the same journal, expanded on that model in 2018 to show how the size and interconnectedness of producers matter in cascades. I take their work and combine it with research on information production, distribution, and cascades (in particular Milgrom and Robert’s famous 1986 paper “Relying on the Information of Interested Parties” and Bikhchandani, Hirshleifer, and Welch’s 1992 article “A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades”) to explore expert failure.
In my article, I look at two case studies, both originating from the COVID-19 pandemic response. In the first, I discuss how the FDA and CDC’s decision in the early days of the pandemic to focus COVID-19 testing only on severe cases led to an overestimation by modelers of the severity and spread of the disease. In turn, their predictions greatly influenced political decisions and individual responses, often leading to responses that were suboptimal. At each and every step, there was vital information from outside the discipline the various experts did not pay attention to, or outright dismissed. In the initial decision to restrict testing, the medical doctors at the CDC and FDA dismissed insights from statisticians and epidemiologists about how random testing would provide better information about the characteristics of this unknown disease. The modelers ignored insights from economics and statistics about how populations would react simply to the news of a new pandemic and would take precautions on their own.
In the second study, I examine how face mask recommendations and policies at the beginning of the pandemic caused increased spread of the disease. Initially, government officials did not recommend face masks for most Americans. However, a few weeks later, they reversed course and required face masks. In the meantime, price gouging legislation was enforced to prevent a rise in personal protective equipment (PPE). Thus, the sudden increase in demand coupled with price controls lead to shortages and hoarding (Dr. Fauci would say in a June 2020 interview that he lied about the necessity of masks in the beginning precisely because he wanted to avoid shortages and hoarding). Additionally, because of the shortages, people had to make repeated trips to different stores trying to find goods, thus contributing to the spread of the disease. Medical experts ignored the insights of economists and psychologists when formulating their recommendations.
Expert advice plays a crucial and indispensable role in our lives. Understanding the institutional structures that generate good advice is necessary for that market to function. Likewise, experts may not be as high-information individuals as they are popularly thought. Experts exist within a highly entangled network, not within clearly delimited silos.
Download Cascading Expert Failure at the Journal of Institutional Economics.
Jon Murphy is currently an Instructor of Economics at Western Carolina University as well as a research fellow at the Institute for an Entrepreneurial Society at Syracuse University. He holds a Ph.D. in Economics from George Mason University and a B.A. from Framingham State University. Jon’s personal website is www.jonmmurphy.com
