May 21st, 2006, 2:09 pm
Noe the title of the book "THEORY of Corporate Finance". Much of the popular theories provide insights on how we should think about things. For example, Miller and Modigliani through their irrelevance propositions taught us what should matter in capital structure when they said that capital structure didn't matter.Models are abstractions of reality and are not meant to be exact forecasting methodologies. If we can build a model that exactly tracks the movement of, say, stocks, then we wouldn't need traders and analysts. No one can make money off anything because we would be certain of what the next price would be. The problem with SOME practitioners is that they just tweak models in order for them to arrive at some workable and practical solution. However, some don't understand why the models they use work and how their "tweaks" make the assumptions of the model invalid. These are often dismissed withtements like "we don't get that much difference in our answers" or other similar statements. In short, inference from a solution that was generated by violating certain model assumptions should not be relied on, but in practice it doesn't seem to hold. A concrete example is the use of econometric methodology to analyze stock price movements, almost all known data problems are present but most still use the simplicity of OLS regressions in presenting results because it is much easier for their audience to accept the results.