Specify and estimate 3 multiple regression models to identify the factors that may explain the CEO salary in 1990. Justify your model specification by referring to the literature on determinants of CEO salary. Conduct post-estimation tests for heteroskedasticity, variance inflation factor (multicollinearity) and omitted variables. Which model is preferable and why? Interpret and discuss your findings from the preferred model, referring to the relevant literature on determinants of CEO salary. (15 marks).

You are given a dataset called Wldrg_CEOsalary1, with 209 observations on chief executive officer (CEO) salary and other variables. Brief description of all variables is given below. 1. salary 1990 salary, thousands $ 2. pcsalary % change salary, 89-90 3. sales 1990 firm sales, millions $ 4. roe return on equity, 88-90 avg 5. pcroe % change roe, 88-90 6. ros return on firm's stock, 88-90 7. indus =1 if industrial firm 8. finance =1 if financial firm 9. consprod =1 if consumer product firm 10. utility =1 if transport. or utilties 11. lsalary natural log of salary 12. lsales natural log of sales