Likelihood of financial distress in Canadian oil and gas market: An optimized hybrid forecasting approach

Abstract

Forecasting models are built on either multivariate parametric or nonparametric methodologies. We attempt to optimize the accuracy of the forecasts combining these approaches to make a robust hybrid forecasting model in predicting the likelihood of financial distress for companies in the Canadian oil and gas market. The proposed approach combined the forecasts out of a multivariate logit model based on the conventional Altman’s Z-score with a nonparametric Artificial Neural Network (ANN) technique. The sample firms are publicly traded and listed on the Toronto Stock Exchange (TSX) and span over a period from first quarter of 1999 to the last quarter of 2014. The results of a proposed three-stage estimation process for the period of 2015-2020 indicated that besides the fact that Canadian energy sector will go through ups and downs regarding the likelihood of financial distress, this industry would face a hard time by late 2020. Results show that the forecasting accuracy out of the proposed three-stage forecasting technique is significantly superior to the outcomes of any individual forecasting techniques, i.e. ANN and logit models.