EASTERN MEDITERRANEAN UNIVERSITY FACULTY OF BUSINESS AND ECONOMICS DEPARTMENT OF BUSINESS ADMINISTRATION Discussion Questions, Problems and Answers ( Reference chapter 4; D. Salvatore; 2004) Discussion Questions: 1. a) What steps are usually involved in the estimation of a demand equation by regression analysis? b) How does a research determine the demand model to estimate? ANS: (a) The steps that are usually involved in estimation of a demand equation by regression analysis are the following four: (1) the model must be specified (i.e., the variables to include in the demand equation must be determined); (2) the data on each variable or its proxy must be obtained; (3) the researcher must decide on the functional form (e.g. linear or power function formulations) of the demand equation, and (4) the regression results must be evaluated. (b) The researcher determines the demand model to estimate (i.e., the explanatory variables to include in the model) from his or her knowledge of demand theory and the market for the commodity. Demand theory identifies the variables that are generally important in all demand functions. These include the price of the commodity, consumers’ incomes, the number of consumers in the market, the price of related (i.e., substitute and complementary) commodities, and consumers’ tastes. The demand for a commodity also depends on some specific variables that are important for the particular commodity. These might include the level of advertising by the firm and its competitors, the availability of credit incentives, consumers’ expectations about prices of the commodity and business conditions in the future, and so on. Often data are not available on all the variables that are thought to be important. In those cases, the researcher can seek proxies for these variables or collect data on them by consumer surveys. Omitting important variables will bias the results. The number of variables included in a regression analysis must, however, be relatively small (say, no more than 5 or 6) in order to have sufficient degrees of freedom and to avoid problems of multicollinearity. 1
2.How does a research go about obtaining the data to estimate a demand equation by regression analysis?
- Winter '17