PPA207-CaliforniaStateUniversity,Sacramento.docVIP

PPA 207: Quantitative MethodsMeeting 10, Spring 2004 1. Homework a. Studenmund, Chapter 8, Number 9 a. I expect that competitors price (PC), gross domestic product (Y), consumption (C), and number of stores (N) will have a positive influence on sales of durable goods (SQ). I expect that own price (PQ) will have a negative influence on SQ. There are 26 observations, so dof equals 26-5-1=20. These are all one-tailed tests so critical t at 5 percent level is 1.725. The calculated t for PC is 0.80. for PQ it is 1.20, for Y it is 0.51, for C it is 1.49, for N it is 1.94; therefore, can only accept that N has a positive influence on SQ. b. Omitted variables could be measure of differences in advertising activity or product mix across the years. Y and C are really measuring the same thing (the effect of national income on purchases at these stores) and hence they are likely to be highly collinear and one of them is irrelevant. c. The high R squared and the high colinearity between Y and C indicates that multicolinearity is likely to be a problem. The partial correlation coefficient between of 0.813 between PC and PQ is not that large and these two variables measure two separate influences on SQ (the own price effect and the substitute price effect). Therefore I would not consider dropping one of these variables. d. I would recommend removing Y and keeping C as a more direct measure of what influences SQ. Also look to add some explanatory variables to account for marketing and product differences over the years. b. Pollock, Chapter 7, Question 2 Review answers in class c. Go over one student’s regression run and multicolinearity check and correction 2. Studenmund, Chapter 9, Serial Correlation Consider that you have been asked to build a simple model of property tax revenue in Sacramento County in year “t” Property tax revenuet = f (populationt, timet) Either for understanding impact of these causal variables or forecasting Data retrieved from California Institute for Cou