The Most Simple Explanation of the Endogeneity Bias and 2-Stage Least Squares Regression
HTML-код
- Опубликовано: 16 мар 2016
- A simple, non-mathematical and intuitive explanation of the endogeneity bias and the importance of the 2-stage least squares (2SLS) model.
This is one of the fundamental models of econometrics and is able to untangle causality from spurious correlations. It is crucial for anyone using non-experimental data to understand the 2SLS model.
***
LOOKING FOR AN ONLINE TUTOR?
Are you struggling with statistics or your university studies? I've had several requests for ONLINE TUITION. I highly recommend my colleague Amber who is a gifted online tutor! All her sessions are done via Zoom. I've heard great things from her students, and have seen her teach first hand! Amber is an amazingly talented tutor.
Get in touch with Amber at: studentvip.com.au/tutors/1630...
Good luck!! :D
****
SUBSCRIBE at: ruclips.net/user/subscription_...
Keywords: 2SLS, 2-stage least squares, 2 stage least squares, endogeneity bias, exogeneity, econometrics, statistics, OLS regression, regression analysis, GMM, maximum likelihood, two-stage least squares, instrumental variables, instruments, endogenous, exogenous, TSLS, economics, finance, stata, eviews, endogeniety
Hi my viewers! Are you in need of an online tutor? If so, check out the video description for details 😊
the best explanation I've watched so far
Thank you so much for making this video! The best explanation I've ever found so far
thank you, this explanation was so helpful!
Thanks so much for this video!
Clear, thank you!
Thank you, you explained it very clearly! :)
I really appreciate this vedio .. helpful and useful.
Thank you for the explanation :)
Excellent explanation. Much better than my grad profs who made such concepts impenetrable.
Hi there 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.
excellent explanation!!!!
Thank you so much bro
It's a wonderful explanation. Thanks so much.
thank you do much for this
Thank you!!!
thanks a lot!!! saviour
I am in Indonesia and I need 2LS regression model 😭. Help
Thank you for the video :-)
how to apply it in spss??
Very nice:)
What are the differences between endogeneity and omitted variable bias? Can omitted variable bias be regarded as a type of endogeneity? Can you give me the examples as well?
can you do a video on heterogeneity?
So in theory, if you had a variable measuring competitiveness and it would correlate with both, your main explanatory variable (education) and you dependent variable (wages), then you would simply include it in your regression model? What about correlation with education?
Yes, if you can observe (and measure) competitiveness, you should include it in your model, since it's most likely correlated with both wage and education. Ignoring it would lead to omitted variable bias. If you don't have data for competitiveness, you can use an instrumental variable to avoid omitted variable bias. This instrument has to be both *relevant* (correlated with education) and *exogenous* (uncorrelated with competitiveness which is part of the error term). So in this example the instrument has to affect wage solely through education. 2SLS works as follows:
*1st stage regression:* Regress your endogenous variable (education) on ALL exogenous variables (including your instrument) to isolate the "good" part of education, that is uncorrelated with the error term (which includes the effect of competitiveness) and compute the fitted values of education. These fitted values are now exogenous, since they're just a linear combination of other exogenous variables.
*2nd stage regression:* Regress your dependant variable on the fitted values of education as well as the other exogenous variables (without the instrumental variable). That's basically your original model where you replaced education with the fitted values of education which you calculated in the first stage.
-> Be happy because you got rid of the nasty endogeneity in your model ;-) Your estimator is now unbiased (if there are no other issues).
I am not sure, but I think this illustrates "confounding" or "common causation." Endogeneity selection bias would result from a situation in which wages and education are independent (B1 = 0), but both cause something else, such as happiness. If we then regressed wages against BOTH education and happiness, it would result in a spurious negative relationship between education and wages (B1-hat < 0).
Maybe an issue of discipline (Elwert and Winship 2014. Annu. Rev. Sociol. 2014. 40:31-53
Great video but total clickbait thumbnail. Gotta do what you gotta do I guess
this is such click bait.... didn't even explain 2sls--just a commercial for your course
ruclips.net/video/KDZKjNsdWeE/видео.html
clickbait
How is this the most simple explanation of 2SLS? You didn't explain it at all.
ruclips.net/video/KDZKjNsdWeE/видео.html
@@GarimaGupta Thank you Garima. I already passed my econometrics exam.😊 This guy helped me a lot: ruclips.net/video/3GF3rpkaDfU/видео.html
@@bubbleworld4172 oh! No issues.. That's good to hear! I'm a student and i have just started my own channel.. So was just wondering if you could see my channel and subscribe to it.. Only if u like! 😊
what kinda accent is that?