Optimizing behavior alters predictions of epidemiological models
Dasaratha (2020) and Toxvaerd (2020) assume that individuals can choose to social distance themselves in order to limit their own risk of infection. This alters predictions relative to the SIR model. For example, because individuals reduce contacts when the risk of infection is high, peak prevalance will be lower than in the SIR model. Moreover, Toxvaerd shows that the more infectious a disease is, the higher is peak prevalence according to the SIR model, whereas the opposite can be the case in the economic model. In a similar spirit, Dasaratha shows that polices that make socializing safer can be counterproductive because individuals increase their social contacts. Both papers are related to earlier papers that study individual decision-making in SIR models, for example Fenichel (2013) who also shows that social distancing policies can have paradoxical effects.
Links to papers:
Fenichel (2013): https://www.sciencedirect.com/science/article/pii/S0167629613000039
Dasaratha (2020): https://arxiv.org/abs/2004.14533
Toxvaerd (2020): http://www.econ.cam.ac.uk/research/cwpe-abstracts?cwpe=2021
Stockholm School of Economics