Why Use Regression Analysis?

There are numerous reasons that you might want to use this technique. In general, these reasons fall into one of two categories:

1. Focus on Drivers. Regression analysis can be used to identify the "levers" (or explanatory variables) that can most effectively influence or control the value of the dependent variable. Looking again at the national merchandise retailer example, regression analysis shows that advertising expenditures are indeed strongly related to total sales. The firm can then view advertising as one "lever" that can be used to influence or control the value of total sales.

2. Predicting the Value of a Variable. Regression analysis can also be used to estimate the future value of a variable. The equation that defines the relationships between the dependent and independent variables can be used in a predictive manner by first replacing explanatory variables with predicted values, then solving the equation to arrive at a predicted value for the dependent variable. To illustrate with the retail example, the analyst may be responsible for creating a five-year strategic plan for the company. This activity requires projecting cash flows for that period of time. If budgetary constraints will reduce advertising expenditures, the analyst can use the regression analysis equation to quantify the future impact on total sales. This function of regression analysis is of great value because of its ability to introduce scientific methods into the forecasting process. In addition, these techniques can be deployed in the economic filter to understand market demand, as well as in the operational filter as future impacts of financial performance drivers on return on ROA. The basic equations for market demand and the DuPont method can serve as the basic regression equation.

We will focus primarily on exploring regression analysis as a predictive tool and discover its utility and application in the context of the Step-Wise Approach to Value (SWAV).