Individual Take-Home Assignment for Empirical Methods in Finance Course Code: 6314M0203 Lecturer: Dr. Florian Peters Date Posted: 17 January 2017, 17:00 Deadline: 31 January 2017, 13:00 In this assignment you are asked to replicate selected and slightly modified parts of the article "Bank CEO Incentives and the Credit Crisis" by Fahlenbrach and Stulz, Journal of Financial Economics, 2011 (in the following referred to as "FS"). For this assignment you will need to use four databases, ExecuComp (to compute CEO incentives), CRSP (to construct buy-and-hold returns), the CRSP-Compustat Merged Database (for standard accounting data), and the Compustat Bank Fundamentals Annual Database (for the tier 1 capital ratio). Start from the entire ExecuComp dataset (CEOs only), and apply the sample selection criteria given in Section 2 of the article. Then compute three measures of CEO incentives: First, "Percentage of equity-based pay". This is fair dollar value of option awards plus the fair dollar value of stock awards, all divided by total compensation. Second, "Percentage ownership", defined as the number of shares and options held by the CEO (vested as well as unvested ones), divided by the total number of shares outstanding. (Note that this definition differs from that used in the article.) And third, "Dollar gain from +1%", defined as the natural log of the dollar change in the value of the CEO's portfolio of current and prior grants of shares and options for a 1% change in stock price. As an approximation for this variable, use the natural log of the variable ONEPCT as defined in equation (8) in Bergstresser and Philippon (2006). (The paper is available on the Blackboard page for this assignment.) Next, merge in accounting data from Compustat as well as the Tier 1 capital ratio from the Compustat Fundamentals Annual Database. Finally, construct buy-and hold returns (BHRs) for the banks analyzed. Buy-and-hold returns for bank i and period s, starting from time t, are computed as BHRi,t-->t+s= (1+R i,t+1)*(1+R i,t+2)*…* (1+R i,t+s)-1. Compute these BHRs from CRSP monthly holding period returns for all banks in your sample, and over the period July 2007 to December 2008. (For each bank you should have one BHR for each calendar month in the period July 2007 to December 2008). Normalize BHR to zero in July 2007. Note that, for the plot of question 3, you need to keep one observation for the BHRs for each bank and month while for the regressions, you only need to retain one BHR for each bank, i.e. the BHR from July 2007 to December 2008. You will then need to merge the file containing the buy-and-hold returns with the ExecuCompCompustat–merged file. The merge is somewhat different for question 3 than for questions 4 and 5 due to the different number of BHRs per bank.Questions: 1. Provide a summary statistics table equivalent to FS's Table 1. You only need to report the variables Total assets, Total liabilities, Market capitalization, Net income/total assets, Net income/book equity, Book-tomarket ratio, and Tier 1 capital ratio. Briefly summarize the results and compare them to those of FS. Do you roughly or exactly match the numbers in their Table 1? 2. Provide a summary statistics table equivalent to FS's Table 3. You only need to report the statistics for CEOs, and you only need to report the following variables: Total compensation, Salary, Bonus, Dollar value of annual stock grant, Dollar value of annual option grant, Other compensation, Value of total equity portfolio/total value of annual compensation, Percentage of equity-based pay, Percentage ownership , and Dollar gain from +1% increase in stock price. Note that the fraction of equity-based compensation is not given in Table 1 of the original paper. You will use it as an alternative measure of equity incentives. 3. Split the sample in two halves, high and low incentives according to each of the three measures of incentives, and plot the monthly subsample average of the buy-and-hold return (y-axis) against monthly calendar time (x-axis) separately for banks in the two subsamples. (Plot three graphs (one for each incentive measure) where each graph shows two lines: one for the upper half and one for the lower half. Briefly describe the figures, and comment on whether they are consistent with FS's main hypothesis. (Note that the original article does not contain these figures.) 4. Produce a regression table similar to Table 4 of FS. Report five columns only: the first three show univariate regressions of buy-and-hold returns on the three incentive measures, the fourth column is a multivariate regression including all three incentive measures as independent variables. The fifth regression contains only one incentive measure, the one which is most highly significant in regressions (1) to (3) and further adds the book-to-market ratio and the log of market value. Do you find that highly incentivized CEOs did better or worse during the crisis? 5. Produce the same regression table as for question 4, except that you now replace the dependent variable with the tier 1 capital ratio in fiscal year 2006. Did banks with highly incentivized CEOs have less, similar or more leverage than less incentivized CEOs before the crisis? (Note that the article does not contain this table.) For each of the questions, give your answer in two pages. Provide the respective table or figure(s) including caption on the first page. On the second page, describe your results in words, i.e. comment on the magnitude of the key coefficients or statistics (statistical and economic significance), and relate your results to those in the original paper (are your results similar or different?). Tables and figures should look like those in the top finance journals (Journal of Finance, Journal of Financial Economics, Review of Financial Studies). They must be self-contained, requiring no further information from other sources, including the main text, to broadly understand them. In particular, the caption must define all variables and briefly explain what the table or figure shows. Use descriptive names for the variables, not cryptic ones. E.g. use "Book leverage" instead of "bk_lev". You need to submit your document in pdf format to the Blackboard course page. If you create your document in Microsoft Word or another text editor, you need to convert your file to pdf format before submitting. In addition to the answers to the questions above, you also need to submit the STATA code which you used to prepare and analyze the data. Copy and paste the Stata code into the appendix of the document you submit, so that you upload only one single pdf file to Blackboard (do not upload dta files to Blackboard!). The deadline for the assignment is 31 January, 2017, 13:00. The assignment will be graded on a scale of 1 to 10. It will count for 60% of the final grade. I will use the following scheme for late sub–missions: For each day of delay, 0.5 grade points will be subtracted from the grade (i.e. one day late=-0.5pt, two days late=- 1 pts, etc.). Good luck!