Today we talked about Instrumental Variables (IV) in Econometrics.
--
[Talking about how it's good to have an overidentified IV regression.]
Professor: What do you do during Thanksgiving?
Student: You eat a lot.
Professor: Okay. Yea. You sit down during Thanksgiving, and you think, "Aw, I really like turkey. Oh! There's turkey!"...
"Aw, I really like gravy... Oh! There's gravy!"
"Aww, I really like mashed potatoes. Oh! There's mashed potatoes!"
Then you say, "Ohhh, I really like pecan pie. I really, really like pecan pie. OH! There's pecan pie too! ... There's also pumpkin pie, and I don't really like that, but that's okay because there's pecan pie."
Like, it's the same thing. "Boy, I really like instruments."
"Oh, why don't you have an instrument! Have a second instrument! Have a third, a fourth!"
"Thank God! Now I can write my paper!"
It's like Thanksgiving. There's 10 of your favorite things.
--
Professor: What's this variance? We don't know.
Student: It could be a zero or not zero.
Professor: That's right. It could be a zero or not because we don't know the true beta. Who knows what the true beta is? If you know the true beta, you know the variance. But nobody knows. Only God knows. God knows.
(Then, he wrote "God knows" on the board.)
--
Professor: Do you know the formula for this?
[nobody answers]
Professor: "No" is a fine answer.. because there's no midterm for this class. "It's midterm week, so I have to study for other important things. I don't have to time to memorize formulas."
Monday, March 7, 2011
Subscribe to:
Post Comments (Atom)
pickles
ReplyDeleteduh man. I posted a comment yesterday and it didn't show up again cause I didn't do your code thing.
ReplyDeleteanyway, I said "I know that picture"