tag:blogger.com,1999:blog-6288862798546085706.post6169663105805414466..comments2024-02-08T03:39:11.256-05:00Comments on Econometrics By Simulation: Question and Answer: Generating Binary and Discrete Response DataFrancishttp://www.blogger.com/profile/16658586705916884436noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-6288862798546085706.post-89122880150638879232013-08-20T17:23:36.201-04:002013-08-20T17:23:36.201-04:00Hi Syed, I am glad you found the post helpful.
I ...Hi Syed, I am glad you found the post helpful.<br /><br />I am not sure yet if I fully understand what you need. I think this is what your are saying:<br />where R1 and R2 are a random uniform draw <br />theta is the CDF of the normal distribution<br />G is some coefficient on the endogenous variable Y1.<br />B is the coefficient on the exogenous variables X<br />N is the coefficient on the instrumental variables Z<br />M is the coefficient on the exogenous variables X as they predict Y1<br /><br />Y2 = 1[theta(X*B + Y1*G)>R1] <br /><br />Thus: Y1 = = 1[theta(Z*N + X*M)>R2]<br /><br />It is not very hard to generate data looking like this. Is this what you are asking? However, solving this is slightly more complicated. <br /><br />Check out this post related to the Biprobit which I think should answer all of your questions:<br />http://www.econometricsbysimulation.com/search/label/biprobitFrancishttps://www.blogger.com/profile/16658586705916884436noreply@blogger.comtag:blogger.com,1999:blog-6288862798546085706.post-80023047898516923362013-08-20T13:41:46.715-04:002013-08-20T13:41:46.715-04:00Thanks Francis. It is really helpful. I have anoth...Thanks Francis. It is really helpful. I have another question. In case of my research at hand I have a probit model where one independent variable is binary and endogenous. I looked for instruments and due to data limitation they are also binary. I used -ivreg-with instruments. Now I am looking for some help in simulation for the case where I have to generate binary variables. I understand that I cannot use -drawnorm- in this case. <br />Thanks again for your response.Hashttps://www.blogger.com/profile/15631084325984688254noreply@blogger.com