Modeling post-disaster business recovery under partially observed states: A case study of the 2011 great East Japan earthquake

Abstract

Post-disaster business recovery is a complex stochastic process influenced by a variety of factors. Moreover, the recovery rate estimation and prediction are more challenging when limited recovery data are obtained for each firm. This study proposes a novel analytical methodology by incorporating the multi-state Markov model to estimate the conditional recovery rates of firms under partially observed recovery states while considering the impact of sector types, initial damage, and financial resource availability after disasters. The recovery process is an improvement process of multiple production capacity states, and the transition probability among states is estimated by utilizing fixed inspection interval observations under an exponential distribution function assumption. Moreover, the likelihood function is developed and maximized to estimate the parameters in the recovery function. To verify the applicability of the proposed model, 2052 firms' recovery after the 2011 Great East Japan Earthquake was conducted as an empirical case study. Specifically, post-earthquake recovery processes under different socioeconomic factors are compared and discussed, with a special focus on the impact of location and financial conditions. The results indicate that the proposed methodology is practical and effective in estimating the partially observed recovery process, which is more in line with the type of data that actual recovery cases can collect. The estimation is limited to earthquake disasters in this research, however, the concepts and framework in the recovery model are also applicable to other types of hazards.