Stock volatility worries investors. That is because, if the market is continually deviating from its normal average, it gives the impression that the market as a whole is unstable and uncertain. Investing becomes difficult. Regression modeling can give a scientifically accurate picture of volatility in its proper framework: the variables that cause the volatility and the intensity of that relationship. Regression analysis, therefore, is absolutely necessary to make sense out of volatility.
Lay out your variables. The dependent variable will always be stock volatility. You will be working with a specific stock, a specific sector, or the market as a whole. This must be clarified. Volatility itself must be measured and defined. Your main concern is deviations from the stock price average. The extent to which the average, standard deviation is too much for your comfort is a subjective variable, but it must be laid out in detail.
Define your independent variables. If you are concerned with what exactly causes volatility and to what extent, independent variables should be well defined and many. The basic independent variables are normally macroeconomic variables such as unemployment and economic growth. Also include those variables that are specific to the market, such as investor sentiment, and those specific to the firm, such as employee turnover, managerial abilities and the nature of the competition.
Program your data. Stock volatility issues suggest you take a “pivot point” average. That is, to take the low, medium and high averages for the stock price for each day. Spread this over at least 10 or 15 business days. The more days you use, the more information the model has to develop a useful measure. The more data, the better the model.
Analyze the results. You are looking for several things. First, determine which variables are causing the change. The regression software will show you which variables are significant, and which are not. It will also show you where your variables can be improved. The biggest enemy of a regression model is duplicate coverage. If two variables are covering the same ground, this will distort your results. For example, two macro-variables that often say the same thing are threats of inflation and future interest rate predictions. These might be important for understanding why a stock or the market is volatile, but these are basically the same variables. Eliminate one. A good regression program will let you run diagnostic tests on your variables.
Write up the volatility framework. Begin with your basic theory. For example, you might hold that volatility in this specific stock is being caused by high fuel prices. Then lay out your variables in detail. Report the statistical results from the regression analysis software. Finally, describe which variables were significant and which were not. Once you have done this, your readers will have a detailed and scientific analysis of what is causing market uncertainty and how to take action against it.
- Comstock/Comstock/Getty Images