Please use http://econ413.wustl.edu/adata/dodgers9-a10.wf1.
About two thirds of the way through the 2008 season, the Los Angeles Dodgers baseball team traded for superstar Manny Ramirez, and the result was a divisional pennant and dramatically increased attendance. Suppose that you've been hired by Manny's agent to help prepare for his upcoming contract negotiations by determining how much money Manny generated for the Dodgers. You decide to build a model of the Dodgers' attendance, and, after learning as much as you can about such modeling, you collect a time series of the 81 home games in 2008 and estimate the following equation:
ATTt= 34857 +4104*MANNYt +2282*PMt +5632*WKNDt +4029*PROMt +8081*TEAMt
( 1021) ( 1121) ( 1096) ( 1068) ( 5819)
4.02 2.04 5.14 3.77 1.39
N 81 R^2=.54 DW 1.30 where:
NOTE the table is an Excel object. In your project ALL tables must be Excel objects.
Note the data set contains 8 equations. Make a table of all 8 equations. Use EQ02R as a place holder to order the variables in the table. Display the table first.
a) Discuss any differences in your output EQ01 and the output above.
b) Without consulting a table, based on the EQ01 DW .299740, explain why you think there is or there is not serial correlation.
c) Discuss whether you accept no serial correlation and why using the BG tests on EQ01.
d) Display a table with only EQ01 and EQ01HAC. Why are the estimates with (EQ01HAC) and without (EQ01) Newey West HAC identical?
e) How do the p-values differ between the two equations? You may use color rather than the actual p-value to indicate the differences.
f) Test the equation EQ01 with Ramsey, 1 to 4 terms. What is your conclusion?
g) A different specification of the effect of MANNY is to regress both the intercept and interactions:
LS att c pm wknd prom team rival manny pm*manny wknd*manny prom*manny team*manny
Test equation EQ02 with Ramsey and BG serial LM. What is your conclusion versus EQ01?
h) Test whether MANNY had an effect on attendance using EQ02 or EQ02HAC depending on your answer in G.
i) Suppose you are more concerned with an omitted variable than with serial correlation, especially because an omitted variable can cause impure serial correlation. Add RIVAL ( a dummy variable equal to 1 if the opponent in the ith game is an in-state rival of the Dodgers, 0 otherwise) to the equation resulting in EQ01R EQ01RHAC EQ02R EQ02RHAC
Do a Ramsey test (1 to 4 terms) and a BG serial LM test, 1 to 4 terms. What is your conclusion about the specification with RIVAL versus without RIVAL.
j) Which equation of the 8 equations EQ01 EQ01HAC EQ02 EQ02HAC EQ01R EQ01RHAC EQ02R EQ02RHAC is best and why? Is MANNY significant in your equation?