Assignment #4 (1 page)
Econometrics II, CERGE-EI, Summer 2004
Daniel Münich
Due: July 7, 2004 (in my mailbox or office)
This assignment is for grade. Submit one output per voluntary formed groupsof two. The length of your output should not exceed 6 one sided pages. The primary purpose of this assignment is your learning-by-doing using pre-cooked data and research itinerrary. The assignment is entirely based on chapter 11 in the Berndt’s book. Scans of the chapter and empirical data file are available from the course web.
Get familiar with the economic and econometric theory as reviewed i) at the lecture (and instructor’s notes) based on Wooldridge and Kennedy books, and ii) in Berndt’s theory chapter 11. The notation and terminology between those two slightly differ. To perform tasks required bellow, you will need to consult examples in Stata 8 manuals to understand individual commands, options and post-estimate tools. Do not execute tasks blindly.Try to understand the essence of what you do. These days, limited depend procedures are very common technique and you should be able to use them even without complete grasp of their technical peculiarities.
Tasks marked as optional bellow will not be graded, although they are relatively easy to implement using existing commands in Stata 8.0. Performing optional tasks would be your own benefit for the future.
1)Get familiar with the dataset MROZ.txt described on p.651 and in README.rtf. Perform Exercise 1 (page 651) but do not provide written answers because data inspection/description is not the primary purpose of this assignment.
2)Perform tasks in Excercise 2, briefly describe what you did and answer questions (briefly, because OLS procedure has serious flaws in this case) .
3)Perform tasks in Excercise 3 (present results and answer questions)
a)Linear probability model (. reg LFP LWW1 KL6 etc.)
b)Logit model is optional
c)Probit model:consult Stata for .probit and .dprobit commands and post-estimate commands like .predict. In addition to tasks required in the book, compute Φ(.), φ(.), and inverse Mill’s ratio, λi, using estimated XB index (.predict XB,xb) and statistical functions in Stata,normprob and normd. Plot XB vs. λi
d)Skip
e)Skip
f)Skip
4)Perform tasks in Excercise 4 (present results and answer questions)
a)Conditional OLS model: (hint: White’s procedure is robust option in Stata)
b)Optional
c)Tobit: Consult Stata manual on the use of command .tobit and its options. Instead of comparisson to 4b),compare resuls to OLS results in Exercise 2.
d)Skip
e)Skip
f)Skip
5)Exercise 5 is optional.
6)Exercise 6 is optional. One can consider own 2 stage procedure or use of command .heckman.