Assessing a fossil fuels externality with a new neural networks and image optimisation algorithm: the case of atmospheric pollutants as confounders to COVID-19 lethality.

This paper demonstrates how the combustion of fossil fuels for transport purpose might cause health implications. Based on an original case study [i.e. the Hubei province in China, the epicentre of the coronavirus disease-2019 (COVID-19) pandemic], we collected data on atmospheric pollutants (PM2.5, PM10 and CO2) and economic growth (GDP), along with daily series on COVID-19 indicators (cases, resuscitations and deaths). Then, we adopted an innovative Machine Learning approach, applying a new image Neural Networks model to investigate the causal relationships among economic, atmospheric and COVID-19 indicators. Empirical findings emphasise that any change in economic activity is found to substantially affect the dynamic levels of PM2.5, PM10 and CO2 which, in turn, generates significant variations in the spread of the COVID-19 epidemic and its associated lethality. As a robustness check, the conduction of an optimisation algorithm further corroborates previous results.