02-17-2017 08:07 AM - edited 02-17-2017 08:09 AM
I am trying to run a 3-equation 2SLS to correct for multiple sources of endogeneity. I am trying to fine tune my code that I pieced together from various forum discussions, but I am not entirely sure exactly how it is currently operating. For example, which equation does it estimate first? Where are my instruments being included? What is the step by step process this code is running?
Objective: My goal is to estimate how institutional investors affect R&D and ROA during the recession. My question is: do institutional investors influence firms to adopt certain R&D strategies that helps those firms to perform better/worse during the recession? However, there is a bit of a selection bias because institutional investors may prefer firms which have a certain (hi/low) level of R&D or perform a certain way. To make the endogeneity problem even worse, a firm's performance may also influence its level of R&D. I am not entirely sure my simultaneous 3-equation system is the proper route to explore this; I am open to suggestions. But in the meantime, I would like to figure out how this system I currently have coded is operating.
My institutional investor variable = dumii, R&D variable = adjrd, and performance variable = ROA. I have a quarterly panel dataset spanning first quarter 2003 through second quarter 2009.
I am currently running the code:
poc syslin data=data 2sls first;
endogenous dumii roa adjrd;
instruments dumsp500 lagroa lagadjrd;
performance: model roa = dumii adjrd lev lnat rec;
investments: model adjrd = dumii roa liq vol rec;
insitutionalinvestor: dumii = roa adjrd lnat lev;
Thanks in advance for your help!