Working Paper: CEPR ID: DP14953
Authors: Jakub Steiner; Andrea Galeotti; Jeffrey Ely
Abstract: We study rotation schemes that govern individuals' activities within an organization during an epidemic. We optimize the frequency of rotation and degree of cross-mixing of the rotating subpopulations. Frequency affects risk over the length of diffusion within the infected subpopulation until the organization detects and/or reacts to the infection. If the reaction time is short, then such risk is undesirable since the growth of the prevalence is initially convex in time. Frequent rotation, which acts as insurance against exposure time risk, is then optimal. Infrequent rotation becomes optimal if the organization reacts slowly. Mixing of the rotating subpopulations is detrimental because it increases the share of interactions between sick and healthy individuals. However, the effect of mixing is small if the terminal prevalence is low in the absence of mixing.
Keywords: rotation; epidemic; mitigation
JEL Codes: I18
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| frequent rotation (J62) | optimal insurance against exposure time risk (G52) |
| organization's reaction time is short (L29) | frequent rotation is optimal (C69) |
| organization's reaction time is slow (L29) | infrequent rotation is optimal (H21) |
| mixing of rotating subpopulations (C69) | increased interactions between sick and healthy individuals (I14) |
| increased interactions between sick and healthy individuals (I14) | detrimental to infection control (I12) |
| terminal prevalence is low (J17) | adverse effect of mixing is minimal (C90) |
| strict rotation without mixing (C69) | expected number of new infections is minimized (C22) |
| optimal rotation schemes (C61) | very frequent or very infrequent (C46) |
| optimality of frequent rotation (C61) | robust to some mixing across groups (C90) |
| overall prevalence remains low (I12) | limited effect of mixing on infection spread (C92) |