EC 671

Advanced Econometrics

Professor Junsoo Lee

Syllabus

Free counter

DATA SETS and CODES for textbook examples

Wooldridge Data sets via Baum | Sources of the data sets | Datasets from the STATA site

STATA procedures (Wooldridge) | RATS procedures for Wooldridge data sets

Verbeek site

Greene's Official Site (data sets, limdep examples..)  | Greene Data sets via Baum

DATA / PANEL DATA

RFE (Resources for Economists) DATA LINKS

· PSID (Panel study of income dynamics)

· NLS (National Longitudinal Survey of Youth 1997 (NLSY97)

· Start here

· author list

· title list

· source list

· keyword search

STATA RESOURCES

Use of STATA in Textbooks

Stata class notes with movie! (UCLA)

J & J Note

Resources for learning STATA

Econometrics Handbook  (Volume 1 - 4 free pdf files, 5 - 6)

Three tests (undergraduate/MA econometrics): Exam 1, Exam 2, Exam 3

Greene textbook, Typos  | His book link

STATA Help

Help

OUTLINE

-------------------  Review of EC 670 -------------------------------------------

NEW Note 1 Review of Regression Analysis and Extension  | OLD Note 1

NEW Note 2 Review of Instrumental Variables Estimation |  OLD Note 2

NEW Note 3 Review of Additional Single-Equation Topics | OLD Note 3

NEW Note 4 Review of  System of Equations Models | OLD Note 4

-----------------  Main Topics in EC 671 ---------------------------------------

NEW Note 5 Panel Data Models (I)    | OLD Note 5

NEW Note 6 Further Topics in Panel Data Models (II)  | OLD Note 6

NEW Note 7 Note 7: Maximum Likelihood Estimation OLD Part I, Part II, and Part III

NEW Note 8 M-estimator, GMM, and Minimum Distance Estimation | OLD Note 8

NEW Note 9 Binary Discrete Choice Models | OLD Note 9

Added I (delta method)  | Added II (interaction)

NEW Note 10 Extended Choice Models | OLD Note 10

NEW Note 11 Corner Solution Outcomes and Censored Regression Models | OLD Note 11

Note 12  Selection Bias Models and Average Treatment Effects Models | OLD Note 12

Note 13   Count Data Models | OLD Note 13

Note 14  Stochastic Frontier Models | OLD Note 14

Note 15  Duration Models | OLD Note 15

Note 16 Non-linear Panel Data Models (III) | OLD Note 16

Part I.  “Review” of Linear Regression Models *

            (* will go fast)

1.  Regression Analysis and Extension

Read: CT 1.6, 4, 7; Wooldridge Ch 4; Verbeek Ch 2, Ch 3, 4.10; Green Ch 2, 3, 4, 6

NEW Note 1  | old Note 1 Word file (not proofread)

Problem Set 1

Problem Set 2

Self test 1

STATA example and h/w data  asset2.do (code) asset_pricing.zip (data files) fashion_stores.zip (clothing data)

10 Papers that Dr. Cook recommends for finance doctoral students.

Linear Algebra of OLS Estimation

(common ground)

· Assumptions in OLS and Asymptotic Inference Using OLS

· Heteroskedasticity-robust Inference

· HAC Standard errors

· GLS and WLS (CT 4.5)

· Testing Hypothesis and 3 Tests

· Sources of the Endogeneity Problem

· Treating Unobservables

· Nonlinear Models

· Testing for Structural Breaks

· Use of STATA and Empirical Exercises

2.  Instrumental Variables Estimation

Read: CT 2.4, 4.8-4.9; Wooldridge Ch 5, 9; Verbeek Ch 5; Green Ch 5

NEW Note 2 | old Note 2 (pdf) Word file (not proofread)

  Read: Levitt's paper

2SLS stata examples | data | excel | 2sls example (word file) | 2sls.do

Problem Set 3

· Instrumental Variables Estimation and 2SLS

· Testing for Endogeneity

• Endogeneity test and 2SLS: Mroz data

• Stata instruction file
• Data (stata file)
  

• Exercise Questions on IV estimation
  
• Empirical Exercise: Return to school (homework)

• Data (excel file)
• Stata Instruction: testing for endogeneity and 2SLS
  

· Potential Pitfalls with 2SLS

· Issues regarding Weak Instruments

** Murphy_Topell correction | Summary (revised) | STATA code | data(mroz.dta) | output | Jing's comment | limdep code

· OLS with Generated Regressors and IVs

· GIV, GMM estimator (Wooldridge Ch 13, 14)

· GMM in Limdep examples | mroz.dat | mroz.des (description)

· Verbeek zip file including the example for Consumption based Asset Pricing Model: RATS example (verbeek154.prg)

o Eviews example

· GMM Example Papers in Finance

· Huang and Stoll (RFS, 1997)

· Hennessy (JF, 2004)

· GMM and C-CAPM (2003, WP)

3.  Additional Single-Equation Topics

NEW Note 3 | Old  Note 3 (pdf) Word file (not proofread)

Read: CT 4.6; Wooldridge Ch 6; Green Ch 7

References:

Read: Heckmit's standard error |  Murphy and Topel, and Alternative

Read: D-i-D estimator and Trade liberalization | Meyer et al AER 1995

Read: How much do we trust d-i-d?

· Other sampling schemes

Quantile regression (Koenker)

4. System of Equations Models (Moved from Part I)

NEW Note 4 | Old Note 4

Read; CT 6.9-6.10; Wooldridge Ch 7; Green Ch 14

Problem Set 4

· Seemingly Unrelated Regressions

· GLS

· Endogeneity and GMM in system of equations

4.  Specification Tests and Model Selection

Read CT 8

OLD Midterm Exam 1:   2004, 2005A, 2005B

Part II.  Linear Unobserved Effects Panel Data Models

 1.  Panel Data Models (I)

NEW Note 5 | old Note 5

Stata Examples: wagepan.do (code)   wagepan.dta (data)

Manual

hw 1 | hw 2

• panel data hw / data
• panel data hw2 / crime4 stata data / excel

· PSID (Panel study of income dynamics)

· NLS (National Longitudinal Survey of Youth 1997 (NLSY97)

o Start here

• author list

o title list

o source list

o keyword search

Panel data models: example, crime in Brazil

Wooldridge Ch 10; Verbeek Ch 10.1-10.3; Green Ch 13 

· Advantages of Panel data models and Unobserved heterogeneity

 · Three Approaches

o First Difference (FD)

o Fixed Effect (FE) Models

o Random Effect (RE) Models

· Choosing between OLS, FD, FE, BE or RE models

Someone's excellent notes on Panel data models (easy but good discussions and good examples)

 Notes: 1, 2, 3, 4, 5 | 6 (dynamic) | 7 (choice, FE), 8 (choice RE) | 9 (GEE) | 10 (count)

2.  Further Topics in Panel Data Models (II)

NEW Note 6   |    old Note 6

Wooldridge Ch 11; Verbeek Ch 10.4-10.8; Green Ch 13

· Hausman and Taylor method

· Dynamic Linear Models

· Cluster Samples

· Recent Issues in Panel data models

Manuals Backup copy: Read  (1) abond and (2) dpd (new)

Stata files:

· Panel IV estimation (xtivreg.do | xtrivreg.log )

· Hausman & Taylor Method (Stata do file | data set | log file )

· Dynamic panel

o TSP archives for Dynamic Panel Data models

o STATA example (stata example | data)

References:

· Lemieux paper | Hoxby paper | Friedberg paper | Levine's Bad smoking paper

Fama and MacBeth :

· Original Paper

· Example 1 | Example 2 | Example 3 | Example 4 | Example 5 (Fama French) | Example 6 (Fama French) |

 Part III.  “Review” of General Estimation Methods *

      Wooldridge Ch 12, 13, 14; Verbeek Ch 6; Green Ch 17, 18

           (* will go fast)

(a)  Maximum Likeihood Estimation

NEW Note 7 | Old Note 7: Part I, Part II, and Part III

· Properties of MLE

· Hypothesis Testing

· Preliminary and Examples

· Examples 

o Ex1 (Linear)

o Ex2(Gamma)

o Ex3 (weibull data)

o Ex4 (logit)

o Ex5 (probit)

o Ex6 (Poisson) | another

o Ex7 (Linear using own gradient)

o Ex8 (Poisson using own gradient)

o Ex9 (tobit)

o Ex10 (ordered probit)

o Ex11 (negative binomial)

o Ex12 (duration)

Exercises

· Exercise homework

Problem Set 5

(b)  M-estimator, GMM, and Minimum Distance Estimation

NEW Note 8 | Old Note 8

M-estimator, Extreme estimator and Properties of GMM

IVGMM stata note

Part IV.  Advanced Cross-Sectional Econometric Models

1.  Binary Discrete Choice Models

Wooldridge Ch 15; Verbeek 7.1; Green 21.2-21.5

NEW Note 9  | Old Note 9

Added I (delta method)  | Added II (interaction)

· Review on Probit and Logit Models

· Latent Index model? Advantages of LPM?

· Issues for these models

· Endogeneity issue

· Choice based sampling

· On heteroskedasticity and non-normality

· QMLE and sandwich estimator

· Panel Choice Models

· Examples and Review

· (old) choice_binary (output | stata code)

· (revised) choice_binary (stata code file: choice_binary-revised.do  | output | )

· AACSB example (question and stata | excel file data)

· WLS example (Greene Ch 12; copied)

· Panel RE/FE probit and logit models (stata code xtprobit_logit.do | output | smcl output file)

** Murphy_Topell correction | Summary (revised) | STATA code | data(mroz.dta) | output | Jing's comment | limdep code

         Read:   Sandwich estimator for the variance in logit or probit, Why?

2. Extended Choice Models

NEW Note 10  | Old Note 10

      Wooldridge Ch 15; Verbeek 7.2; Green 21.7-21.8

  · Multinomial Logit and Conditional Logit Models

· IIA assumption

· Nested Logit Models

· Ordered Probit Models

· Panel Models

· Examples and Review

Example of extended choice models

stata code extended_choice.do | output word file |

Try this (logit-low1.do).  So, "clogit, group(pair)" and "paired logit" produced the same results. Jing wrote a user-defined code for this issue.

Exercise questions

Data file:  TBL19-2.dta

3. Corner Solution Outcomes and Censored Regression Models

NEW Note 11  | Note 11

Wooldridge Ch 16; Verbeek 7.4-7.5; Green 22.1-22.3

· Inconsistency of OLS

· Estimation and Inference with Censored Tobit

· Censoring and Truncated Regression Models

· Panel Models

· Examples and Review

·  Stata code tobit_cnreg_intreg.do | output log file | word file output

4. Selection Bias Models and Average Treatment Effects Models

Note 12

Wooldridge Ch 17, 18; Verbeek 7.6; Green 22.3-22.4

· Censoring, Truncation and Incidental Truncation

· Two Stage estimation of the Tobit Model

· Models with Self-selectivity

· Corrected Standard Errors of Generated Regressors

· Other extended models

· Panel Models

· Counterfactual Setting and the Self-Selection Problem

· Propensity Score Matching Methods

· Differences-in-Differences Method

· IV Estimation

· Examples and Review

· Stata code heckman_selection.do | output log file | word file output

· PSM example: 

  This needs two files: psmatch.do | psmatch.ado | pamatch.hlp | pdf document for this

5. Count Data Models

Note 13

Wooldridge Ch 19; Verbeek 7.3; Green 21.9

· Count Data

· Poisson regression models

· Negative binomial models

· Hurdle, ZIP and ZAP Models

· Endogeneity Issue

· Panel Count data models

· Examples

· Stata code and output files :  poisson_negbin.do | output log file | word file output

· Limdep code and outputs   :  POISSON.LIM | poisson.out | poisson.lpj 

· Limdep code and outputs   :  count_hw.lim | crime1.dat | fertil1.dat

6. Stochastic Frontier Models

Note 14

(Reference Papers, Green p. 429, 501-505)

· Various Frontier Models

· Efficiency Measures

· Ranking airline companies or banks? 

· Why can’t we use residuals? 

· Technical efficiency, allocation efficiency

· Stochastic Frontier Models

· Panel Data Stochastic Frontier Models

· Examples and Review

· Stata code and output files :  frontier.do | output log file | word file output | frontier.txt

7. Duration Models

Note 15

Wooldridge Ch 20; Verbeek 7.8; Geeen 22.5

 · Duration and Transition Data

· Hazard and survival functions

· Non-parametric Approach

· Plot of Hazard Rates

· Homogeneity test

· Semiparametric and Parametric Duration models

· Proportional Hazard Models

· Parametric Duration Models

· Examples and Review

STATA example:

duration2.do duration2.doc duration2.log

duration.do duration.txt duration.doc duration.log

LIMDEP example: