Weather, Lockdown, and the Pandemic: Evidence from the Philippines

As the landscape of the COVID-19 pandemic continues to evolve, there is a need to better understand the factors that affected COVID-19 health outcomes using a more appropriate dataset and comprehensive variables. This paper constructs a novel daily provincial panel dataset (N = 14,507) during the nascent and important period of the pandemic (April–September 2020) to analyze both the socioeconomic (lockdowns or ECQ, mobility of individuals, health care capacity, and trends in transmission) and environmental factors (rainfall shocks, temperature in Celsius, average relative humidity, and wind speed) that affect COVID-19 health outcomes. A panel dataset is more apt than the other types of datasets since it addresses both spatial and time variations, as well as the time-invariant unobserved heterogeneity that, if ignored, would have resulted in biased estimates and findings. In addition, using a more complete list of explanatory variables could address omitted variable bias, which leads to proper identification and a more reliable set of findings that could aid the government in formulating optimal, multi-faceted, and timely policies for future health crises. Using fixed effects on panel data, our results, which are robust across the different lag structures and time periods used, are consistent with the existing literature with caveats. First, while ECQ is effective in stemming COVID-19 cases, it is ineffective in reducing COVID-19 deaths. Second, exogenous weather variables have heterogenous effects on COVID-19 health outcomes contingent on the period of analysis and the type of health outcome analyzed. Third, public behavior, which is only partially correlated with public policy (ECQ), matters in curtailing viral transmission. We conjecture that individuals voluntarily avoid infection for their own well-being, resulting in positive externalities, or they stay at home due to weather shocks.