Hello Everyone,   I need your help with the SAS codes. I need to use a Heckman selection model to correct for endogeneity. In the first stage of my Heckman model, the dependent variable is CEO expertise. Since there are three categories of CEO expertise in my context, the dependent variable (CEO_Expert) takes value either 0,1, or 2 (0 if the CEO is an accounting expert; 1 if a finance expert; and 2 if a sales expert). I don’t know how to obtain inverse mills ratio when the dependent variable is ternary. In the past instances when my dependent variable was a 1/0 dummy, I used proc qlim to obtain inverse mills ratio.   PROC QLIM data = First_Stage; model Y1 = X1 X2 / DISCRETE; OUTPUT out=Have mills; RUN;   I have seen only examples of using inverse Mills ratio involving binary probit selection equation. I would really appreciate it if someone shared the code to obtain inverse mills ratio when the dependent variable is not binary.   Thanks, Gokul   ... View more

Hello Everyone,   I need your help with the SAS codes. I need to use a Heckman selection model to correct for endogeneity. In the first stage of my Heckman model, the dependent variable is CEO expertise. Since there are three categories of CEO expertise in my context, the dependent variable (CEO_Expert) takes value either 0,1, or 2 (0 if the CEO is an accounting expert; 1 if a finance expert; and 2 if a sales expert). I don’t know how to obtain inverse mills ratio when the dependent variable is ternary. In the past instances when my dependent variable was a 1/0 dummy, I used proc qlim to obtain inverse mills ratio.   PROC QLIM data = First_Stage; model Y1 = X1 X2 / DISCRETE;  OUTPUT out=Have mills;    RUN; I have seen only examples of using inverse Mills ratio involving binary probit selection equation. I would really appreciate it if someone shared the code to obtain inverse mills ratio when the dependent variable is not binary. Thanks, Gokul     ... View more

Hello All,   I am a newbie and need your help with the coding. I have a huge sample (obs around 10,000) which is divided into two subsamples-- sample firms and control firms. I am trying to match each sample firm with a control firm using the following criteria: (1) sample firm and control firm should belong to the same year and same industry. (2) the control firm should be the closest match with the sample firm based on SIZE and BM. That is, [(Sample firm SIZE – Control Firm Size) + (Sample firm BM – Control Firm BM)] = minimum. (3) All sample firms should be matched with a unique control firm. So if a control firm is the closest match to two sample firms, then I want the second sample firm to be matched with its second-closest control firm.   Sample Firms Firm      Year   Industry      Size          BM           Sample_Firm A          2011       4511           8.0         0.20                       1 A          2012       4511          8.7         0.21                        1 B         2012       4511           8.5         0.19                        1   Control Firms Firm         Year     Industry       Size          BM                   Sample_firm X              2011       4511           7.9         0.19                        0 X              2012      4511            8.5         0.2                          0 X              2013      4511            8.3         0.17                        0 Y              2011      7826            8.7         0.21                        0 Y              2012      7826            8.2         0.19                        0 Y              2013      7826            8.7         0.23                        0 Z               2011      4511            8            0.5                          0 Z               2012     4511             8.5         0.18                        0   Desired Matched firms (the output that I am trying to get) Firm             Year            Industry             Size                             BM                      Sample_firm A                  2011             4511                    8                               0.2                          1 X                  2011             4511                    7.9                            0.19                        0 A                  2012             4511                    8.7                            0.21                        1 X                  2012             4511                    8.5                            0.2                          0 B                  2012             4511                    8.2                            0.19                        1 Z                  2012             4511                    8.5                            0.18                        0   Sample_firm is a dummy variable that takes value 1 for sample firms and 0 for the control firms. So in the matched firms, I want to match A_2011 with X_2011, A_2012 with X_2012, B_2012 with Z_2012.   Can someone please share a code that will help me get the desired matched firms?   Best regards, Gokul ... View more

I need your help with a code. I have three samples (Sample A, B, and C). All three samples have information on SIZE, CAPEX, LEVERAGE, INDUSTRY, and YEAR. I want to create two new samples D and E.     In sample D, I want to match sample A with sample C based on INDUSTRY, YEAR, SIZE, and LEVERAGE. That is, the matched firm from sample C should belong to the same industry and year as the firm in sample A. Also, the difference in SIZE + difference in LEVERAGE in the matched firms should be minimum.   In sample E, I want to match sample B with sample C based on INDUSTRY, YEAR, CAPEX, and LEVERAGE. That is, the matched firm from sample C should belong to the same industry and year as the firm in sample B. Also, the difference in CAPEX + difference in LEVERAGE between the firm in samples B and C should be minimum.   Another condition is that the firms from sample C that are matched with sample A should not be used in the matching procedure of sample B.   In the attached file I have used an example to show the sample that I have and the sample that I want.   I would greatly appreciate it if someone provided me with the code so that I could get the desired output.   Best regards, Gokul ... View more