Zoom Logo

Public Lectures on Structural Econometrics - CEMMAP masterclass - Shared screen with speaker view
Handy Tan
03:33
Are there any good reference texts that would capture most of the topics for this course?
Dante Langone
16:38
I can't access the lecture slides of today's lecture, is this an issue for everybody?
Joan Llull
17:02
You mean from the website?
jadesiu
17:05
Yes, me neither :(
Elena Pastorino
17:09
@Dante: they are up here
Elena Pastorino
17:10
http://comlabgames.com/47-812/2%20Estimators/7%20Nonlinear%20Parametric%20Estimators_update.pdf
Joan Llull
17:13
http://comlabgames.com/47-812/2%20Estimators/7%20Nonlinear%20Parametric%20Estimators_update.pdf
jadesiu
17:53
Thank you!
Dante Langone
17:59
Thank you, I was getting a 404 error from the website
Elena Pastorino
18:16
Yes: if you refresh the page all should work now
Irina Luneva
01:26:19
How do we choose theta1 and theta2?
Joan Llull
01:27:03
Note that this is given by the structure of the problema to estimate
Joan Llull
01:27:30
The function f1 only depends on theta_1, whereas the function f2 depends on both
Elena Pastorino
01:27:31
@Irina: it is for sake argument. Bob is saying: suppose you have a model that separates the way he is describing i.e. suppose you have a modular structure
Elena Pastorino
01:28:21
how should you proceed then?
Irina Luneva
01:28:27
I see. Thank you!
Irina Luneva
01:30:20
But we know the distribution of epsilon_n?
Irina Luneva
01:30:40
Otherwise, we wouldn’t be able to simulate
Elena Pastorino
01:31:28
@Irina: the point Bob was making is general. To simulate your model, yes, you need to know the DGP it implies
Irina Luneva
01:31:46
Understand, thanks!
Lisa Botbol
01:39:08
Can someone give me a sense of what the function Psi is? I am still a bit lost as to what it concretely represents
Lisa Botbol
01:39:44
“The behavioural equations of the model”, what does that refer to concretely?
Arie Beresteanu
01:39:53
It's a summary of all the functional and distributional assumptions of the model.
Elena Pastorino
01:40:00
@Lisa: it is like a first order condition, an equilibrium condition, etc
Milena Almagro
01:41:37
@Lisa as concrete example think of labor demand = labor supply in equilibrium => labor demand - labor supply = 0
Milena Almagro
01:41:55
With labor demand and supply given by some functional form
Elena Pastorino
01:45:03
@Lisa&Milena: yes! Think about a budget constraint, a resource constraint, a first-order condition for the choice of a continuous action, and similar. They can all be written in Psi(.)=0 form
Lisa Botbol
01:45:25
thanks everyone! I get it now
Lisa Botbol
01:46:12
Another question: how do you “simulate” x and y? Do you draw them?
Lisa Botbol
01:46:40
Or do you simply simulate epsilon
Joan Llull
01:46:50
That depends. Are they exogenous or endogenous?
XIAOXIA SHI
01:47:05
@Lisa, yes, you will assume a distribution for epsilon, and draw from there. Then get the draw of y and x based on the model and the draw of the epsilon.
Joan Llull
01:47:10
If they are exogenous, then you need to draw them the same way that you draw epsilon
Joan Llull
01:47:40
If they are endogenous, then they are defined by the model
Lisa Botbol
01:47:47
mmh ok I see, thank you
Pengzhan Qian
01:56:21
Could I say that in Simulated Method, we simulated the integral, rather than have the analytic form of the integral (like in previous lectures)?
XIAOXIA SHI
01:57:14
@Pengzhan yes, you use simulated method to compute the integral if there is no analytical solution to the integral.
XIAOXIA SHI
01:58:16
For an example, think multiple choice probit model where you need to compute triple integrals of normal densities .
Pengzhan Qian
01:58:54
@Xiaoxia Thanks. My feeling is that to get the analytical solution we need more assumptions, especially on the error terms.
XIAOXIA SHI
01:59:28
or different assumptions that leads to simpler integrand.
Pengzhan Qian
02:00:24
I mean, it seems that with enough assumptions, we could always get the aggregation expression. Simulated methods is another way to do it.
XIAOXIA SHI
02:01:25
Hmm... I'm not sure I see what you mean. Would you like to ask the professor?
Elena Pastorino
02:02:47
@Pengzhan: in rich structural models you may not be able to derive the analytical form of the integral so you may end up reverting to simulation in this case too
Arie Beresteanu
02:03:44
In many cases due to non-linearities, the dimentionality and the distribution of epsilon it is impossible to compute the integrals analytically. Simulation methods help us get around this.
Pengzhan Qian
02:05:04
@Elena, thanks! It kind of answering my puzzles. My feeling is that there were many paper with anaytic results in Rust times. But nowadays more paper without anaytic solution but SMM.
Laura Hurtado-Moreno
02:05:58
How do we choose A_N in slide 28?
Elena Pastorino
02:06:46
@Penghzan: yes, many interesting models are simply very difficult (or impossible) to solve analytically -- they involve multidimensional functional equations to solve for Psi(.)=0, many dimensions of unobservable to integrate over, etc
Pengzhan Qian
02:07:40
Thank you Elena and Xiaoxia!
Miguel Angel Cabello
02:11:32
Many thanks!
Handy Tan
02:17:51
What is the difference between system and difference gmm? When is it appropriate to use one vs the other?
Milena Almagro
02:18:56
Have to leave now. See you next weekend!
Joan Llull
02:23:36
@Handy, some people use this terminology to refer to different ways of dealing with dynamic panel data methods (Arellano-Bond would be the difference gmm and Arellano-Bover refers to the system gmm). If that is the case, there is a long answer about one versus the other, but the short answer is more assumptions (Arellano-Bover assumes the process is stationary) vs more robustness (Arellano-Bond is not very robust when the autoregressive coefficient is close to 1).
Handy Tan
02:23:58
Ok thank you Joan