A Generalized Linear Mixed Model for Data Breaches and Its Application in Cyber Insurance
Authors:
Open Access
Abstract
Data breach incidents result in severe financial loss and reputational damage, which raises the importance of using insurance to manage and mitigate cyber related risks. We analyze data breach chronology collected by Privacy Rights Clearinghouse (PRC) since 2001 and propose a Bayesian generalized linear mixed model for data breach incidents. Our model captures the dependency between frequency and severity of cyber losses and the behavior of cyber attacks on entities across time. Risk characteristics such as types of breach, types of organization, entity locations in chronology, as well as time trend effects are taken into consideration when investigating breach frequencies. Estimations of model parameters are presented under Bayesian framework using a combination of Gibbs sampler and Metropolis–Hastings algorithm. Predictions and implications of the proposed model in enterprise risk management and cyber insurance rate filing are discussed and illustrated. We find that it is feasible and effective to use our proposed NB-GLMM for analyzing the number of data breach incidents with uniquely identified risk factors. Our results show that both geological location and business type play significant roles in measuring cyber risks. The outcomes of our predictive analytics can be utilized by insurers to price their cyber insurance products, and by corporate information technology (IT) and data security officers to develop risk mitigation strategies according to company’s characteristics.Keywords
Funding Information
- Natural Science and Engineering Research Council of Canada (RGPIN/05640-2019)
This publication has 52 references indexed in Scilit:
- Heavy-tailed distribution of cyber-risksThe European Physical Journal B, 2010
- Actuarial statistics with generalized linear mixed modelsInsurance: Mathematics and Economics, 2007
- The economics of information security investmentACM Transactions on Information and System Security, 2002
- Data clusteringACM Computing Surveys, 1999
- Markov Chain Monte Carlo Convergence Diagnostics: A Comparative ReviewJournal of the American Statistical Association, 1996
- Markov Chains for Exploring Posterior DistributionsThe Annals of Statistics, 1994
- Estimation in generalized linear models with random effectsBiometrika, 1991
- Model Selection and Akaike's Information Criterion (AIC): The General Theory and its Analytical ExtensionsPsychometrika, 1987
- I. Problems and Designs of Cross-Validation 1Educational and Psychological Measurement, 1951
- The Monte Carlo MethodJournal of the American Statistical Association, 1949