Title : A Panel Path Analysis approach to the Determinants of COVID-19 Transmission: Does Testing Matter for Confirmed Cases?

Authors : Gour Gobinda Goswami, ARM Mehrab Ali, Sharose Islam

Abstract : The mysterious character of COVID-19 is mostly unknown to us. Researchers are making attempts to examine the linkage of its transmission through population density or social distancing, government stringency, different types of testing patterns, per capita GDP which works as a proxy for disposable income and act as a measure of willingness and ability of a country to conduct widespread test under the framework of a consumption function hypothesis, prevalence of other deadly diseases like diabetes or cardio vascular diseases etc. Regional and country wise variation based on culture and drive of political authority etc. is also important. Using Our World in Data for 212 countries and areas and 162 time periods on a daily basis since December 31, 2019 to June 09, 2020 on an unbalanced panel framework we have developed a Panel Based Path Analysis Model to explore the interdependence of various actors of COVID-19 cases of transmission across the globe. As an anecdote, we explore the direct effects, indirect effects and total effects of different potential determinants of transmission cases in the world and get an idea about the relative role of each factor in a structural equation framework. First of all, we analyze the global pattern then we extend the analysis to different regions like Africa, Asia, Europe, North America, Oceania, and South America. The robustness is also checked by type of test whether it is based on number of people, number of sample or unknown testing pattern etc. No matter what specification we use we find that high per capita GDP or income leads directly to more tests and more transmission. Countries, which cannot afford more tests, are also the countries where transmission rate is suppressed downward. Using panel path model we find that a 1% change in the standard deviation of test results into 0.69% to 0.70% change in total cases per million after controlling for several variables like per capita GDP, government stringency, people aged above 65 etc.

Journal : Volume : Year : 2020 Issue :
Pages : City : Edition : Editors :
Publisher : ISBN : Book : Chapter :
Proceeding Title : Institution : Issuer : Number :