Forecasting Corporate Bankruptcy: Applying Feature Selection Techniques to the Pre- and Post-Global Financial Crisis Environments
By Parker Levi
I investigate the use of feature selection techniques to forecast corporate bankruptcy in the years before, during and after the global financial crisis. Feature selection is the process of selecting a subset of relevant features for use in model construction. While other empirical bankruptcy studies apply similar techniques, I focus specifically on the effect of the 2007-2009 global financial crisis. I conclude that the set of bankruptcy predictors shifts from accounting variables before the financial crisis to market variables during and after the financial crisis for one-year-ahead forecasts. These findings provide insight into the development of stricter lending standards in the financial markets that occurred as a result of the crisis. My analysis applies the Least Absolute Shrinkage and Selection Operator (LASSO) method as a variable selection technique and Principal Components Analysis (PCA) as a dimensionality reduction technique. In comparing each of these methods, I conclude that LASSO outperforms PCA in terms of prediction accuracy and offers more interpretable results.
View Data (Email for Access)
Advisors: Professor Andrew Patton, Professor Michelle Connolly | JEL Codes: G1, G01, G33
Market Dynamics and the Forward Premium Anomaly: A Model of Interacting Agents
By Phillip Hogan and Evan Myer
This paper presents a stochastic model of exchange rates, which is used to explain the forward premium anomaly. In the model, agents switch between four trading strategies, and these changes drive the evolution of the exchange rate. This framework is meant to more realistically represent the important market dynamics of exchange rates, as we suspect these to be the cause of the forward premium anomaly. Our simulations of the model indicate two conclusions: (i) many of the statistical regularities observed in currency markets, including the forward premium anomaly, can be thought of as macro-level scaling laws emerging from micro-level interactions of heterogeneous agents, and (ii) the dynamics of estimates of the beta coefficient in tests of UIP are driven by perceived relationships between changes in interest rates and agents’ aggregate views on the value of the exchange rate, which we call the fundamental value. Section I presents an introduction to the topic and section II reviews the relevant literature. Section III provides the theoretical basis of the forward premium anomaly and our model, then the mathematical definition of the model. Section IV presents the results of a typical simulation which section V compares to relevant stylized facts of the currency markets. Sections VI and VII present our results and a conclusion of what we have drawn from the model.
Advisor: Craig Burnside, Michelle Connolly | JEL Codes: G1, G13, G15 | Tagged: Exchange Rates, Forward Premium Anomaly
A Case Study on the Informational Role of Futures Markets: Can Weather Futures Forecast Electricity Consumption?
by Ying Chiat Ho
Abstract
This paper provides a case study on the informational role of futures prices by investigating the ability of Cooling Degree Day (CDD) futures prices to forecast electricity consumption for New York State. I develop a cross-sectional model relating electricity consumption with the cumulative CDDs in a month for New York City and utilize the 30-day ahead settlement prices of the New York CDD futures contracts within the model to forecast electricity consumption. The forecasts derived explain up to 94.68% of the variation in actual electricity consumption, suggesting that the CDD futures prices contain useful forward-looking information about electricity consumption.
Professor Edward Tower, Faculty Advisor
JEL Codes: G13,
