Hypothesis The rate of exit is lower when the csp of a firm is higher

Part I. Your submitted do-file should be able to reproduce the results in your answers; otherwise your answer to part I will not be scored. Please check to make sure that the do-file runs smoothly before you submit your answers (50 points).

Use Final_BUS918_2020.dta file uploaded on Blackboard. Variables are defined as in their labels.

These days, more and more attention is paid to corporate social performance (Flammer, 2015; Zhao & Murrell, 2016). Let’s examine how corporate social performance (CSP) affects a firm’s ability to survive in the marketplace, and how this relationship differs by whether the firm is an US firm or a non-US firm. Using a global dataset tracking a firm’s corporate social performance, let’s suppose that you want to test the following hypotheses:

Hypothesis 1. The rate of exit is lower when the CSP of a firm is higher.

Hypothesis 2. The rate of exit is lower if the firm is an US firm.

Hypothesis 3. The relationship between CSP and the rate of exit will weaken if the firm is an US firm.

1) Create “Table 2” after running survival models with appropriate specifications. You expect that the following variables also affect firms’ rate of exit and therefore should be controlled: firm size, profitability, current ratio, leverage, foreign sales to total sales, and industry, with all variables except industry and the indicator for US firm lagged by one year. Table 2 should have three columns: first column with exponential hazard, second column with Gompertz hazard, and the third column with Cox proportional hazard model (10 points).

2) Create “Table 3” after running the same survival models as above. This time, please report hazard ratios instead of coefficients. Interpret the results carefully (10 points).

3) Draw necessary graphs and run necessary tests to figure out which of the three models is the best model in testing the hypotheses (10 points).

4) Based on your answer to 3) above, explain whether each of the hypotheses is supported or not (10 points).

5) Create “Table 1” for the variables used in your analysis, including the interaction variable (5 points). Based on these correlations, how much should you be concerned about the issue raised by Kalnins (SMJ 2018) (5 points)?