Endogeneity and Instrumental Variables

[adham]

I learned more from your video than my entire econometrics class

[AkbarinSeoul]

Unbeliveble! You can explain complicated concept with a very clear explanation. Best teacher!!!! Thank you very much!

[rainmakr]

all heros dont wear capes. thanks for these awesome videos. i hope u read this. maybe u can make more and keep contributing towards this awesome channel

[MrHadeedian]

Explained the complications of regression in very simple words. Nice and Thankyou.

[muratcaglayan]

i am new for instrumental variables but that is great intuition for solving problems. 

[nikitagauhar]

Please make a video on endogeneity. I have watched several video but there is no good content available. I believe you may explain that very well.

[zeeshan]

You are a super awesome instructor! <3

[loanie]

Thanks for saving my 2 last years of Bachelor…

[witamywholandii]

Next to great videos, also a great voice. You sure you are not the guy from HeadSpace??

[whim]

Omg thanks for the explanation! I finally understand!

[jungikpark]

Thanks a lot. Your videos are really helpful.

[elisanebres]

Hi Ben. I've been watching your channel for a couple of months now. Your videos are very helpful. I have one question though. When is it appropriate to use IV or GMM, in terms of dealing with endogeneity. Which is better.

[winstonchirpsehill]

endogeneity omitted variable bias question
Today, 07:04


I am seeing the effect of derivative usage on firm value, and so regression firm value with my independant variable derivative usage and set of control variables.

I know there is an endogeneity issue in the sense that there a characteristics both unobservable (eg managerial quality) and observable that have a postive effect on firm value and are postively correlated with derivative use.


I understand the enodogenity in this sense mean that these characteristics that are captured by the error term are linked to the the explantoryy variable deriavtive usage.

What my main question is is which variables are ones reffered to as endogenous?

a)is it the firm value that is enodogenous or deriavtive usage that is enodogenous

b) or are firm value and derivative usage both "endogenous variables"

c) or is the observed/unobserved characterisitcs that are referred to as endogenous.
Thanks so much.

[MrJJOBT]

How can this be true if the OLS constructs estimator such that the residuals are independant from X? Endogeneity in OLS cannot exist by definition

[TheBiggestOne]

Guys, I need your advice .. my supervisor misled me with writing my master's thesis. First, he says that I need to run the OLS regression and then to check the Hausman specification test for endogeneity and next do the IV. I followed his advice, run the OLS, then checked the endogeneity, found that it existed, and then did the IV, and then compared the results with OLS regression. I wrote and sent the paper to him to check. However, he commented that it is methodologically wrong approach and I needed to do everything vice verse - first to check the Hausman specification test for endogeneity and run IV, then OLS. I am in a deadlock.. cannot understand what really he wants to see? What is the step by step logic/idea with endogeneity problem? What has to be done first, second, third... Please, advise...

[sadaf]

Basic question. Why do we even have E(e\x) not equal to zero? Isn't Alpha there to pick up any non-zero mean of error?

[zhangliang]

I have a question regarding the Endogeneity. If I knew that that some of the regressor are Endogenous, and if I knew how to find the instrumental variables, Do I still need to use 2SLS? for example I have a linear equation:

y = alpha + beta1*x1 + beta2*x2 + beta3*x3 + e.

I knew that x1 is Endogenous, and I knew that the instrumental variables can be xx1, xx2. Why can I just use xx1, xx2 and x2, x3 to create a new linear model?

y = alpha + beta1*xx1 + beta2*xx2 + beta3*x2 + beta4*x3 + e.

[nganhuynh]

what about the efficiency of an estimate?

[mrsmanssour]

Hi Lambert, Does endogeneity happen in linear regressions only?

[k_o_r_a_y]

Ah, so you call unmeasured confounding as endogeneity.