A1127
Title: Scaling up coherent estimation: Evaluating COVID-19s crime impact across U.S. cities using hierarchical time series
Authors: Thomas Fung - Macquarie University (Australia) [presenting]
Joanna Wang - University of Technology Sydney, Australia (Australia)
Abstract: In response to the COVID-19 pandemic, many U.S. cities experienced significant shifts in crime patterns. A prior study examined these impacts using data from 25 large cities, focusing on the early stages of the pandemic. However, traditional approaches to crime analysis often model disaggregated series independently or rely on simple aggregation, which can lead to inconsistencies and reduced accuracy when evaluating intervention effects. Building on a previous work, which utilized grouped time series structures to reconcile estimates across hierarchical levels, this follow-up study expands the analysis to the full set of cities included in the referenced study. A hierarchical time series modelling framework is applied that ensures coherent estimation across both disaggregated and aggregated levels. Using updated data and a more comprehensive city sample, a detailed evaluation of crime pattern changes is offered during the pandemic, and it is demonstrated how estimation reconciliation techniques can improve both predictive performance and the evaluation of policy interventions in the context of criminal justice research.
