Effect of Temperature on the Spread of Contagious Diseases: Evidence from over 2000 Years of Data

The COVID-19 pandemic led to a surge in interest among scholars and public health professionals in identifying the predictors of health shocks and their transmission in the population. With temperature increase becoming a persistent climate stress, our aim is to evaluate how temperature specifically impacts the incidences of contagious disease. Using annual data from 1 AD to 2021 AD on incidence of contagious disease and temperature anomalies, we apply both parametric and non-parametric modelling techniques, and provide estimates on the contemporaneous, and as well as lagged effects, of temperature anomalies on the spread of contagious diseases. A non-homogeneous Hidden Markov Model is then applied to estimate the time-varying transition probabilities between hidden states where the transition probabilities are governed by covariates. For all empirical specifications, we find consistent evidence that temperature anomalies in fact have statistically significant effect of the incidence of the contagious disease in any given year covered in the sample period. The best fit model further indicates that the contemporaneous effect of a temperature anomaly on the response variable is the strongest, and that given temperature anomaly predictions are becoming very accurate, one can prepare effectively with necessary public health response for at least contagious diseases. These findings further have implications for designing cost effective infectious disease control policies for different regions of the world.