Korean Energy Economic Review ISSN:1599-7057(Print) 에너지경제연구

Since energy and environment industries have much publicity,predicting the industries’ corporate bankruptcy in advance is importantin that it can predict and prepare for the possibility of sudden supplyshocks, thereby alleviating the overall economic shock anddeteriorating public welfare. This study analyzes prediction of corporatebankruptcy in energy and environment industries. The logit model isused as basic estimation model, in which 51 financial variablesfrequently used in the previous studies are considered as initialexplanatory variables. And we compare the predictive power betweentwo models, one with and another without the variables multiplied bythe industries' dummy variable as additional variables. On the otherhand, Lasso(Tibshirani, 1996), a model shrinkage method, is applied toalleviate the problem of over-fitting and multicollinearity due to manyexplanatory variables used in the model. The results showed that thelogistic lasso model including the dummy variables of the energyenvironment industry was the best in all predictive measures.