A Statistical and Machine Learning Framework for Measuring the Economic Impact of Reduced Travel due to COVID-19 in Maryland

The COVID-19 pandemic caused business shutdown worldwide because of which there was a significant reduction in roadway traffic movement and vehicle miles travelled. This, in turn resulted in lower gasoline sales. This paper assesses the economic impact of reduced vehicle miles of travel due to COVID-19 and determines the impact on the revenue generated in the highway trust fund due to reduced gasoline taxes collected in Maryland, USA. An economic loss assessment is performed using gasoline sales and vehicle miles of travel. The effect of declining Annual Average Daily Traffic (AADT) between 2018-2020 is analyzed for the 23 Maryland Counties and Baltimore city, to identity Counties that were significantly impacted due to the declining AADT. Finally, a machine learning model using artificial neural network is performed to predict the significance of various Counties subjected to declining AADT during COVID-19. The results are useful for assessing the economic impact of COVID-19 on roadway improvement projects owing to reduced gasoline sales.