Use Auditfees data, sample codes and refer to the course ppt to perform the following tasks: 1. (2 points, show codes only) In the dataset, there is a variable nseg. This variable shows the number of business segments a company has. Notice that data on the number of business segment (nseg) is missing for some firms. Set the missing value of nseg to 1 Hint: use if ... then... command 2. (2 points, show codes only) Create a dummy variable Loss, which is equal to 1 when a firms earnings before extraordinary items (ib) are negative and 0 otherwise. 3. (6 points, report both codes and results). How Accounting firms charge audit fees may differ when firms report loss. The relations between our Xs and Y might be different for these two scenarios. Run the following regression model by whether a firms earnings before extraordinary items is negative or not, respectively. Log(Audit Fees) i,t = 0 + 1 Log(Assets)i, t + 2 Leverage i, t + 3 Profitability t + 4 Rec_Inventory i, t + 5 nseg i, t + 6 Big5 i, t Hint: Remove the outliers of leverage, profitability and rec_inventory first. Briefly comment on the regression results, such as which independent variables are statistically significant and how they are related to audit fees. Are the associations between each independent variables and audit fees same when firms report losses versus when they report profits?

2 4. (5 points, show codes only) Propose a model to examine tax fees. Write down your dependent variable (Y) and independent variables (Xs). Show your codes on how to generate these variables. The independent variables should have some economic meanings and comparable across firms. Briefly explain what firm characteristics each independent variable is used to measure. 5. (3 points, show both codes and results) Run the regression model you proposed for tax fees and report the results. 6. (2 point) Briefly comment on what you find based on results reported in Question 5. Hint: focus on the positive and negative relation between Xs and Y, and whether the coefficients are statistically significant.