Time To Liquidation Of Smes: The Predictability Of Survival Models

One of the most crucial information in predicting firm time to liquidation is, given a firm has survived a certain period, its probability of surviving an arbitrary extended period. To provide updated results on the determinants of firm liquidations, we examine the survival of a large number of SMEs in the UK and follow them from 2004 up to 2016. We then compare the baseline hazard model and its updated hazard models on predicting time-to-liquidation using time-dependent performances, and for the latter one, we experiment with special of stratified bootstrap validation. We analyze the risk of going into liquidation of the UK SMEs from 2004 onwards using a baseline survival model and compare its predictive performance with the discrete survival model. A sample of 67,262 UK SMEs is employed in survival analysis with company fixed, demographic characteristics and time-varying financial elements. The results first show the significant effects of firm's demographic characteristics including number of trading addresses, number of directors, number of contacts, and number of subsidiaries. We also further stress on improvement in model accuracy using updated hazard models which utilized on the time-varying nature of firm's financial variables.