Modelling the impact of COVID-19 on unemployment patterns across diverse sectors in the Valencian Community, Spain.

The COVID-19 crisis has brought about a substantial and abrupt disruption to the economy of the EU27, leading to the need for unprecedented policy responses in the European Union (EU) nations. The aim of this research is to analyse the impact of the pandemic on unemployment within the Valencian Community, Spain. Specifically, it aims to examine and compare the effects on the major economic sectors in the region, namely industry, service, agriculture and construction. The study investigates the variation in unemployment rates across municipalities and in those economic sectors. To achieve this objective, we used a dataset comprising economic data collected at the municipal level from 2016 to 2021, including variables such as unemployment rates, the number of companies, government spending, budgetary data and debt levels. To explore these economic phenomena, we used spatio-temporal quantitative models that incorporate integrated nested Laplace approximation (INLA) and stochastic partial differential equation (SPDE) methodologies. These specialized methodologies are employed alongside three distinct approaches with various analytical strategies to effectively analyse and comprehend the dynamics of economic factors, allowing us to gain a deeper understanding of the impact of the pandemic on unemployment within the Valencian community. The study first examines the economic situation in the Valencian community by considering all municipalities, stratifying them by each of the economic sectors considered and separating them into three years: 2019, 2020 and 2021. The second approach categorizes municipalities by population size (small, medium, large) and considers political subsidies based on population and unemployment. The final approach uses temporal models to monitor changes from 2016 to 2021, capturing both political and social shifts. The results reveal a uniform impact of the lockdown on all four sectors. Upon closer examination of the three methodologies, it becomes evident that the initial approach, which did not account for temporal effects, emphasized the significance of spatial effects, and the best models are those that include all the elements, covariates, temporal effects and the spatial distributions of the municipalities.