An agent-based model for bankruptcy avoidance
Millan, Liliana (2009) An agent-based model for bankruptcy avoidance. [Dissertation (University of Nottingham only)] (Unpublished)
The consequences of bankruptcy for the economy, shareholders, creditors, and society in general can be of tremendous impact. That is why diverse models have been developed since the mid thirties for trying to predict bankruptcy with the use of financial ratios in order to identify inflection points in which it is still possible to apply corrective actions that make the firms to avoid bankruptcy. This dissertation focuses on identifying the more significant financial ratios for bankruptcy prediction, explores how financial ratios have been used in formally accepted statistic models of bankruptcy prediction. In addition, it makes use of specific financial ratios to identify the financial health of a firm on a given point in time, which triggers prestablished actions that a firm must follow in order to avoid bankruptcy. All the above integrated in an agent-based model developed for bankruptcy avoidance. This research project was based on a review of relevant literature and the collection and analysis of the data obtained from the development of an agent-based model. The findings from this research provide evidence that the best performance obtained in the proposed model have an efficacy of 86.30% with only twenty firms failing for bankruptcy, that twelve months is the ideal time a firm has to wait for issuing securities so to improve the performance of the firm, that the EBIT parameter must be changed during the simulation, and that the EBIT/TA ratio is the most significant one in the model. The main conclusions drawn from this study are that financial ratios are indispensable in prediction bankruptcy models, that the Z-score model is an easy and reliable model to use for classifying firms, and that the rules and actions elicited from the expert are useful and correct. This dissertation recommends that the EBIT should be changed each time the firm issue securities, or on each time step of the simulation following a normal distribution, and that more research should be done in order to obtain the efficacy of the model when the EBIT is changed during the simulation.
Actions (Archive Staff Only)