Free Trial

Safari Books Online is a digital library providing on-demand subscription access to thousands of learning resources.

Share this Page URL
Help

4 The Economics (and Econometrics) of Co... > The Econometrics of Cost Modeling: A... - Pg. 104

104 CHAPTER | 4 The Economics (and Econometrics) of Cost Modeling which precludes a finding of cost complementarity and hence subadditivity of the cost function. Thus, it should be data, not the functional form, that determine the findings and support the conclusions reached. 7. Finally, whenever possible, parsimony is key. Include only variables that are theoretically relevant for which accurate data can be obtained. Be certain to review data carefully, looking for outliers and other oddities. THE ECONOMETRICS OF COST MODELING: AN OVERVIEW Ordinary Least-Squares Estimation You may recall that there are certain assumptions that must be met for OLS esti- mators to be the "best" (a.k.a. BLUE--best linear unbiased estimators). Known as classic assumptions, when met, OLS estimators are unbiased, efficient (i.e., minimum variance), consistent, and distributed normally. These assumptions are: 1. The model is linear in coefficients, correctly specified, and the error term ( i ) is additive; that is, Y = X + i : 2. 3. 4. 5. 6. (4:14) The error term has a population mean = 0. Explanatory variables are not correlated with the error term. Error terms are not correlated serially (i.e., no autocorrelation). The error term has a constant variance (i.e., no heteroscedasticity). No explanatory variable is a perfect linear function of another explanatory variable (i.e., no multicollinearity). 7. The error term, i , is distributed normally; that is, i ~ N(0, 2 ), or Eð i Þ = 0 and Varð i Þ = 2 : (4:15) Regression Analysis and Cost Modeling Basically put, regression analysis is a statistical technique that attempts to quantify the effect of a change in an independent variable on a dependent variable. In the case of cost modeling, the dependent variable, cost, is a function of other explanatory variables, such as prices of inputs and, of course, output. That is, at the very least, a cost function is of the general form: C = f ðY, pÞ, where C is cost, Y is output, and p is (a vector of) input prices. (4:16)