Obtain the following variables from Compustat North America's Fundamentals Annual data via WRDS: GVKEY (firm identifier), FYEAR, DLTT, DLC, CEQ, PRCC_F, CSHO, SICH, and SIC for all available firm-year records. Compute the following variables for each firm-year record in the data:

BookLeverage = (DLTT + DLC) / (DLTT + DLC + CEQ)

MarketLeverage = (DLTT + DLC) / (DLTT + DLC + PRCC_F * CSHO)

Note that the firm identifier in Compustat is gvkey and the year variable is fyear. For a variety of reasons, these variables are likely to contain outliers. One (fairly unscientific) approach to reducing the influence of outliers is to "winsorize" the data. Winsorizing sets all values of a variable above some upper threshold percentile to the value at that percentile and all values below some lower threshold percentile to the value at that percentile. For example, winsorzing at the 1st and 99th percentiles entails setting all values of a variable above (below) the 99th (1st) percentile of the distribution to the value at the 99th (1st) percentile. Winsorize the two leverage variables at the 1 st and 99th percentiles.

The standard industrial code (SIC) is an industry classification system. The Compustat variable SICH is a company's "historical" SIC - it is SIC as recorded at the time of the record. The Compustat variable SIC is a company's current-in-time SIC. In some cases, a company's SICH will be missing. Replace any missing values of SICH with a firm's current-in-time SIC. Then, calculate and report the mean values of BookLeverage and MarketLeverage and number of observations for each of the following:

All firm-years in Compustat All firm-years for which SICH is in the range 4900-4999 (utilities)

All firm-years for which SICH is in the range 2000-3999 (manufacturing)

All firm-years for which SICH is in the range 6000-6199 (banks and credit unions)

All firm-years for which SICH is 7372 (computer software)