Custom Search

Most of the interest in using mathematical models in finance and economics involves models based on hypothetical causal relationships described by quantitative mathematical functions. When these models work they may reflect substantial understanding of the system being simulated and provide a wealth of information about the internal relationships of the various factors included in the model. Many of these models have mathematical characteristics which allow them to be solved for optimal results. In the areas of finance and economics, however, the quality of data is often so poor and the complexities of the relationships so difficult to express that such models frequently seem to be unreliable oversimplifications of the real world.

Another scientific tradition, empirical modeling, should then be considered for wider utilization in the areas of finance and economics. Empirical models may be of value when the more formal theoretical models are deemed inadequate because of high uncertainties or an inability to adequately express the complexities. Empirical models are developed by analyzing experimental data (in the case of finance and economics, historical data).

These models provide little insight or understanding of the cause and effect relationships of the process being studied, but they do recognize identifiable events which can then predict other events in the process with various degrees of accuracy. The methodology of theoretical modeling relies on the statistical relevance of quantitative results derived from mathematical functions used to test hypothetical causal relationships. Empirical modeling simply tests the statistical relevance of qualitative correspondence between observed events.

To the extent that the model is accurate, an empirical model may be all that is required by its users. A good analogy may be found in the field of instrumentation. When the EKG was first used in monitoring the heart, the recorder would print results on a moving tape which initially looked like nothing more than oscillating pen strokes. The instrument was recording a great deal of information but no we knew how to read and interpret it. Eventually, after studying thousands of cases, analysts learned how to read and interpret the squiggles. Different patterns reflect different types of heart conditions. The results won't always be identified correctly, but generally speaking, the test has proved to be reliable and useful. The useful knowledge about the technique was developed empirically. That is actually how most theoretical modeling begins. To quote Karl Brunner on methodology in economics: "We begin with empirical regularities and go backward to more and more complicated hypotheses and theories."