| dc.contributor.author | Roy, Saktinil | |
| dc.date.accessioned | 2009-08-13T15:20:21Z | |
| dc.date.available | 2009-08-13T15:20:21Z | |
| dc.date.issued | 2009-08-13T15:20:21Z | |
| dc.identifier.uri | http://hdl.handle.net/2149/2283 | |
| dc.description | It was a pleasure as well as an opportunity to attend the 15th International Conference on Computing in Economics and Finance. Scholars from around the globe -- many of them from the national central banks -- attended the conference. So, it was also an opportunity to hear how economics and finance researchers in different countries are analyzing the current global crisis. Most papers involved applications of complex time series tools. Since my own research also involves time series and its applications, participation in some of the sessions was very useful for my own advancement.
I presented my paper, “Predicting the Asian Currency Crises with Artificial Neural Networks: What Role of Function Approximation?” at the session on “international transmission of shocks” on July 17th. Other researchers presenting at the session were from Bank of England, Bank of Canada, and Monash University. Many other researchers from several institutions, such as International Monetary Fund, Bank of Canada, Deutsche Bundesbank, participated in the session and made valuable comments. My paper, which recognizes nonlinearity in financial data and suggests a solution with artificial neural networks as an early warning system for financial crisis, was much appreciated by the session participants. The unique property of artificial neural networks as “universal function approximators” and its usefulness in the context of prediction of financial crisis was understood and appreciated.
Currently I am revising this paper so that I can submit it for review at a reputed economics/finance journal. | en |
| dc.description.abstract | Indicators approach and probit and logit models are the two most widely used early warning systems for currency crisis.
Several studies however suggest that despite their strengths these systems have high error rates in prediction. In this paper a recurrent neural network model is proposed as an alternative. Consistent with the notion of model uncertainty the system does not rest on a pre-specified mathematical relationship between the macroeconomic fundamentals and probability of crisis. Under such conditions the neural network model, by virtue of its unique property, serves the purpose of approximating a non-linear function that is assumed to be unknown. The system is implemented for prediction of currency crises in the emerging market economies of Africa, Asia, and Latin America, and the results are compared to those obtained with several probit models by Berg and Pattillo (1999). The recurrent neural network model is found to perform better than the probit models with respect to goodness of fit measures, both within sample and out of sample. In particular, the probit models are outperformed in predicting the1997 Asian crises out of sample. The best of probit models and the neural network (NN) model perform almost equally well in predicting the Asian crises with a probability of 20% -- but the NN model, in addition, predicts a significant percentage of crises with a probability of 50%, whereas the probit models fail completely. We conclude that the NN model, which approximates an unknown non-linear function, has greater chances of performing well out of sample in terms of distinguishing between a ‘tranquil’ month and a ‘pre-crisis’ month. However, both systems fail to predict the crises systematically. | en |
| dc.language.iso | en | en |
| dc.relation.ispartofseries | 92.927.G1154; | |
| dc.subject | Logit Models | en |
| dc.subject | Currency Crisis | en |
| dc.subject | Asian Crisis | en |
| dc.title | Predicting the Asian Currency Crises with Artificial Neural Networks: What Role of Function Approximation? | en |
| dc.type | Presentation | en |
