Several researchers have investigated the relationship between a mutual fund’s performance and the characteristic of its manager. Assume that you collected the data in “C17-01.xls” file to examine this context. The performance of each fund was measured by its risk-adjusted excess return, which is the difference between the return on investment of the fund and a return that is considered a standard. The standard is based on a variety of variables, including the risk-free rate. Four variables describe the fund manager: age, tenure (how many years the manager has been in charge), whether the manager had an MBA (1=yes, 0=no), and a measure of the quality of the manager’s education (the average SAT score).
Write a report that discusses how the average SAT score of the manager’s alma mater, whether he or she has an MBA, and his or her age and tenure are related to the performance of the fund. The report should be less than 10 pages (excluding figures and tables). Figures and tables should be placed at the end of the document, properly labelled and numbered. Your analysis should include the following discussions:
• estimate a model with all four variables included and interpret the value of each coefficient.
• evaluate whether a more simplified model could be used instead. If so, suggest which variable(s) should be dropped from the analysis and justify why the selected variable(s)
should be removed from the original regression model.
• discuss whether the model results are valid. Some of the validation checks include:
– check normality of residuals by using the STATA command lines: * predict r, resid * kdensity r, normal
– check homoskedasticity by: * plotting the residuals against fitted (predicted) values * running the Breusch-Pagan Test. The null hypothesis of this test is that the variance of the residuals is homogeneous. The command for running the Breusch-Pagan Test is “estat hettest”, which is inserted after you run the regression model.
– check multicollinearity by: * finding the pairwise correlation
* using the “vif” command. If the variable is redundant, it will give a high VIF value. – check linearity and the exogeneity assumption by plotting the scatterplot of the residuals against each predictor variable • discuss whether adding polynomials and/or using logarithms is worthwhile for this analysis.
