Review a short publication entitled "A Primer on Gasoline Prices" produced by the Energy Information Administration (EIA) of the Department of Energy. A simple Internet Search will locate it quickly. Write a brief summary of the variables that cause fluctuation in gasoline prices. (note: typically economists do an in-depth literature review)

Collect the following data for the January 2001 through August 2004 time period: composite refiner acquisition cost of crude oil(monthly), U.S. refiner motor gasoline prices by grade and sales type(monthly), total gasoline stocks(weekly), and the CPI for all urban consumers base year = 1982-1984 (monthly). If your last name begins with the letter A – H you will be analyzing the price of regular grade gasoline, if your last name begins with I – R you will be analyzing the price of mid-grade gasoline and if your last name begins with S – Z you will be analyzing the price of premium grade gasoline. (Note - the monthly 2001 data from EIA is below) To use the data you will have to do some manipulation in Excel (Note: gasoline units are in cents and oil is in dollars. Converting to the same units may be helpful) The goal is to have a set of real monthly monetary values and monthly stock (inventory) values from 2001 through August 2004. Data sources will be the EIA and BLS.

Run a simple two variable regression where the real price of gasoline is the dependent variable and the real cost of crude oil is the explanatory variable. Have Excel generate a new worksheet of the regression results and format the data to fit on one portrait page. Below the regression results provide the following: 3a) show the regression equation, 3b) interpret each coefficient and determine if the slope coefficient estimate is significant at the 5% level, 3c) how well did the explanatory variable explain gasoline price fluctuations? (be sure to identify the measure used to develop your answer)

Run a multiple regression where the real price of gasoline is the dependent variable and the real cost of crude oil and the stock of gasoline are the explanatory variables. Have Excel generate a new worksheet of the regression results and format the data to fit on one portrait page. Below the regression provide the following: 4a) show the regression equation, 4b) interpret each coefficient and determine if the slope coefficient estimates are significant at the 5% level, 4c) test the validity of the model using an F-test, and 4d) how well did the explanatory variables explain gasoline price fluctuations? (be sure to identify the measure used to develop your answer)

Turn-in the annotated printouts printed portrait format from Steps 3 and 4. Do not print the data. Your work shall have a cover sheet with your name and assignment number with a staple in the upper left corner.

Using the data collected in Computer Assignment #4, add 3 qualitative variables (dummy variables) to account for the four seasons. Since the Primer on Gasoline Prices article discusses changes relative to the winter, make the winter season the omitted season. See page 238 of Statistics with Microsoft Excel for the basic coding logic. (Note: since more than one month is in each season, you will need to modify the IF statement to allow for "nested" if statements. This will be explained in class or can be found in the Excel on-line reference.)

Re-run the regression in step 4 of Computer Assignment 4 except add the three dummy variables. This time select "Residuals" in the regression dialog box before selecting "OK". Have Excel generate a new worksheet of the regression results and format the data to fit on one portrait page. Cut and paste the residuals output into a new worksheet. Below the regression results (with residuals moved) answer the following: 2a) show the regression equation, 2b) interpret each coefficient and determine if it is significant at the 5% level (how do the seasonal dummy variable coefficients compare to the seasonal fluctuation in the Primer on Gasoline Prices article?), 2c) test the validity of the model using an F-test, and 2d) how well did the explanatory variables explain gasoline price fluctuations? Did the added seasonal dummy variables help explain fluctuations? (be sure to identify the measure used to develop your answer)

Step 3 Use the Durbin-Watson formula (d) on page 681 of the Keller text to compute the Durbin-Watson statistic using the residuals from the regression in Step 2 (do not use the Durbin-Watson macro that comes with the Keller text except to check your value). Format the data and computed Durbin-Watson statistic to fit on one portrait page.

Step 4 Now that you have computed the Durbin-Watson statistic, on a separate sheet of paper set-up a hypothesis test and determine if first-order autocorrelation exists. Be sure to explain how you came to your determination.

Turn-in the output from steps 2 through 4 printed in portrait. Your work shall have a cover sheet with your name, and assignment number with a staple in the upper left corner.