Modeling Behavior during a Pandemic: Using HIV as an Historical Analogy
NBER Working Paper No. 28898 Many models from epidemiology are not designed to capture tradeoffs between health and economic well-being and few are equipped to predict how these tradeoffs interact with individuals’ preferences to influence behavior. Policies based on such models may not reflect societal preferences or capture how preferences affect behavior and disease spread. Models with epidemiological processes that do incorporate behavior change or health-wealth tradeoffs (mostly from economics) drastically limit individual-level heterogeneity to remain tractable. This makes it difficult to understand variation in behavioral responses and distributional consequences of the pandemic, which should be central to any evaluation of pandemic-related policy. Without the proper tools, we are unable to fully address even basic policy questions, such as whether the long-run effects of social-distancing justify the costs - and if so, for whom and under what circumstances. After reviewing modeling approaches from epidemiology and economics (among other fields), I propose a framework adapted from earlier work on the HIV/AIDS epidemic. Analysis of individual behavior is based on the idea that health is a form of human capital in which individuals invest through their choices. The model integrates individual level variation that other frameworks omit, but does so at the cost of drastically limiting aggregate economic and epidemiological processes. I conclude that what is needed is a multi-disciplinary effort that helps to bridge gaps across disciplines, which would amount to constructing models that capture key features of different approaches. To be useful in evaluating pandemic-related policy, such a model should have rich heterogeneity and capture how individual choices are linked to the broad contours of disease spread and aggregate economic output. Whether such an effort would be feasible and fruitful remains to be seen. This paper is available as PDF (468 K) or via email
Machine-readable bibliographic record - MARC, RIS, BibTeX Document Object Identifier (DOI): 10.3386/w28898 |
Contact Us
