というNBER論文をdigitopolyブログなどで知られるJoshua Gans(トロント大)が上げているSSRN版)。原題は「The Economic Consequences of R̂ = 1: Towards a Workable Behavioural Epidemiological Model of Pandemics」。

This paper reviews the literature on incorporating behavioural elements into epidemiological models of pandemics. While modelling behaviour by forward-looking rational agents can provide some insight into the time paths of pandemics, the non-stationary nature of Susceptible-Infected-Removed (SIR) models of viral spread makes characterisation of resulting equilibria difficult. Here I posit a shortcut that can be deployed to allow for a tractable equilibrium model of pandemics with intuitive comparative statics and also a clear prediction that effective reproduction numbers (that is, R) will tend towards 1 in equilibrium. This motivates taking R̂=1 as an equilibrium starting point for analyses of pandemics with behavioural agents. The implications of this for the analysis of widespread testing, tracing, isolation and mask-use is discussed.