Skip to main content
  • Home
  • Happenings
  • Events
  • A study of crime trends during COVID-19 related lockdowns using Bayesian structural time series models
A study of crime trends during  COVID-19 related lockdowns using Bayesian structural time series models

A study of crime trends during COVID-19 related lockdowns using Bayesian structural time series models

Date25th Mar 2021

Time11:00 AM

Venue Webex

PAST EVENT

Details

Purpose: In the proposed research, we inspect changes in crime trends brought about by COVID-19 pandemic in Tamil Nadu, India, pertaining to two-time frames of strict and mild restrictive policy imposed by Government. Specifically, we look at absolute and relative changes in telephonic crime reporting (distress calls) versus actual registration, across different crimes. Such rare pattern changes during a one-off event helps us understand crime motivators, enablers and thereby help practitioners take tangible measures towards its reduction during regular times as well specially to address under-reporting and burking of crime.

Data: Crimes registered as FIRs and Distress calls from Jan 1st 2018 to Jan 1st 2021 pertaining to Property Offences, Offences against Human Body, Crimes against Women, Children and Elders, Cybercrimes, Road Accidents on account of rash and negligence.

Methods: We use Bayesian Structural Time-Series model that estimates the impact of stay-at- home orders on the crime. This effect is measured by analysing differences between expected and observed behaviours. We have the response time series (particular type of crime or distress call) and a set of control time series (average temperature, time of sunrise/sunset), This model tries and predicts the counterfactual, i.e., how the response metric would have evolved after the intervention if the intervention had never occurred.

Results: Mixed crime trends are observed despite an overall depletion in most crimes in the restrictive phases. Property offences has declined during both phases whilst bodily offences sank when acute constraints were warranted, followed by a hike during mild restrictions. Calls for police service spiked in crimes against women, though registered crimes dwindled.

Conclusions: Results have expedited our comprehension of different categories of crime dynamics including emerging crimes during the pandemic. This study, further highlighted the inherent inadequacies such as under-reporting of crimes by victims and non-registration of crime by police, which will suitably assist to enhance preparedness and response to improve public service delivery by reordering priorities in regular times as well

Speakers

Shri. Kandaswamy Paramasivan, MS16D300

Department Of Management Studies