Altman's ZSCORE is one of the most famous bankruptcy production models in Finance. The Altman Z-score...

Transcribed Image Text:

Altman's ZSCORE is one of the most famous bankruptcy production models in Finance. The Altman Z-score is calculated as follows: ZSCORE = (1.2 A) + (1.4 B) + (3.3 C) + (0.6 D) + (1.0 E) == = = where, A = Working Capital/Total Assets, B Retained Earnings/Total Assets, C Earnings Before Interest & Tax/Total Assets, D = Market Value of Equity/Total Liabili- ties, E = Sales/Total Assets. A score below 1.8 means the company is probably headed for bankruptcy, while companies with scores above 3.0 are not are not likely to go bankrupt. Please use the zscore.dta dataset provided in the Brightspace to calculate the Altman's ZS- CORE measure for all manufacturing firms in North America. A manufacturing industry is denoted as SIC code between 2000-3999 in the dataset. Once you calculate your ZSCORE measure, please do the following: 2. Now calculate additional variables to be incorporated as independent/explanatory variables in your bankruptcy-risk prediction model: logarithm of firm-size, defined as: log (at); firms' future growth opportunity, measured by: Tobin's q; book leverage of firms, defined as: dltt/at; firm-level cash-holding measured by: che/at; firms' operating profitability calculated as: ib/sale; firm-level liquidity, captured by the current ratio: act/dlc; and a dummy variable indicating whether the firm pays any dividend or not. Winsorize all these variables at the 1% level on both tails. Give a table of summary statistics for all these explanatory variables. you 3. Now estimate a simple OLS regression model, where ZSCORE is the dependent vari- able and all variables you calculated in the previous step are independent variables. Make have winsorized ZSCORE and all additional variables at the that sure 1% level before you start estimating your regression. Based on your OLS model, which firm characteristics are more informative about predicting default risk? Next, estimate the same OLS using White's (1980) heteroscedasticity-corrected standard errors (the robust option in stata). What happens to your conclusion based on the simple OLS results? Now cluster your standard errors by firm and re-estimate your OLS (the cluster option in stata). What happens to your conclusion based on the simple OLS as well as heteroscedasticity-adjusted results? your 4. So far, you have estimated a pooled regression al model. Now estimate a fixed-effect panel regression model, where each firm has a distinct fixed effect. How does results differ using a fixed-effect panel model as opposed to a simple OLS model earlier? 5. Now create a new dependent variable, which is a dummy variable. If the ZSCORE measure is below 1.8 the new variable gets "1" and if the ZSCORE measure is above 3 the new dummy variable gets "0". Next, use the dummy variable as your dependent variable to estimate a LOGIT and a PROBIT model of bankruptcy risk with all your usual independent variables. Do your results vary compared to the simple cross- sectional OLS and fixed-effect panel models? (15 Marks) Altman's ZSCORE is one of the most famous bankruptcy production models in Finance. The Altman Z-score is calculated as follows: ZSCORE = (1.2 A) + (1.4 B) + (3.3 C) + (0.6 D) + (1.0 E) == = = where, A = Working Capital/Total Assets, B Retained Earnings/Total Assets, C Earnings Before Interest & Tax/Total Assets, D = Market Value of Equity/Total Liabili- ties, E = Sales/Total Assets. A score below 1.8 means the company is probably headed for bankruptcy, while companies with scores above 3.0 are not are not likely to go bankrupt. Please use the zscore.dta dataset provided in the Brightspace to calculate the Altman's ZS- CORE measure for all manufacturing firms in North America. A manufacturing industry is denoted as SIC code between 2000-3999 in the dataset. Once you calculate your ZSCORE measure, please do the following: 2. Now calculate additional variables to be incorporated as independent/explanatory variables in your bankruptcy-risk prediction model: logarithm of firm-size, defined as: log (at); firms' future growth opportunity, measured by: Tobin's q; book leverage of firms, defined as: dltt/at; firm-level cash-holding measured by: che/at; firms' operating profitability calculated as: ib/sale; firm-level liquidity, captured by the current ratio: act/dlc; and a dummy variable indicating whether the firm pays any dividend or not. Winsorize all these variables at the 1% level on both tails. Give a table of summary statistics for all these explanatory variables. you 3. Now estimate a simple OLS regression model, where ZSCORE is the dependent vari- able and all variables you calculated in the previous step are independent variables. Make have winsorized ZSCORE and all additional variables at the that sure 1% level before you start estimating your regression. Based on your OLS model, which firm characteristics are more informative about predicting default risk? Next, estimate the same OLS using White's (1980) heteroscedasticity-corrected standard errors (the robust option in stata). What happens to your conclusion based on the simple OLS results? Now cluster your standard errors by firm and re-estimate your OLS (the cluster option in stata). What happens to your conclusion based on the simple OLS as well as heteroscedasticity-adjusted results? your 4. So far, you have estimated a pooled regression al model. Now estimate a fixed-effect panel regression model, where each firm has a distinct fixed effect. How does results differ using a fixed-effect panel model as opposed to a simple OLS model earlier? 5. Now create a new dependent variable, which is a dummy variable. If the ZSCORE measure is below 1.8 the new variable gets "1" and if the ZSCORE measure is above 3 the new dummy variable gets "0". Next, use the dummy variable as your dependent variable to estimate a LOGIT and a PROBIT model of bankruptcy risk with all your usual independent variables. Do your results vary compared to the simple cross- sectional OLS and fixed-effect panel models? (15 Marks)