臺大管理論叢 NTU Management Review VOL.29 NO.1

CEO Incentives and Bank Liquidity Management 264 3.1 Data Sources We gathered our sample from the WRDS Compustat Bank file, which mainly consists of US financial institutions with Standard Industry Classification (SIC) codes 602 (commercial banks) and 603 (saving institutions) from period 1992 to 2012. Because our study examined whether bank liquidity policy was associated with CEO incentives of commercial banks, our banks within the sample had to meet the following criteria: 1. Complete information for all the relevant measures had to be available. Therefore, banks with missing information on total deposits, total loans, and liquid funds were excluded from the sample. 2. Banks must not have gone through a merger during the research period. Mergers data from the Federal Reserve National Information Center were analyzed, and any banks that went through a merger during this period were excluded from the sample. 3. To avoid the possibility of outliers having outsized influence on the results, we eliminated all bank-years with 1) asset growth over the last year in excess of 50%, 2) total loan growth exceeding 100%, 3) total loans-to-asset ratio below 10%, and 4) share of credit card loans in the loan portfolio above 50%. These criteria were similar to those used in Loutskina (2011). 4. Data about CEO incentives had to be available either in a database or through manual collection; however, we did not require that each bank have all CEO incentives data available. Data from Compustat Bank database provided us with most of the financial data needed to conduct our empirical analysis except for bank liquidity, transaction deposits, commitments, and the amount of loans in different categories for each bank. 1 These data were available from the Report of Condition and Income (named “Call Report”) that all banks in the U.S. that are regulated by the Federal Reserve System, Federal Deposit Insurance Corporation (FDIC), and the Comptroller of the Currency are required to file. 2 After merging with Compustat data, there are 3,552 observations (450 banks) that including 424 observations of S&P banks (26 banks) and 3,128 observations of non-S&P 1 We have 7,750 Compustat observations (1,035 banks), including 857 observations of S&P 1,500 banks (52 banks) and 6,892 observations of non-S&P banks (983 banks) from years 1992 to 2012. 2 We have total 51,001 observations (2,927 banks) from Call report from years 1992 to 2012.