The Hong Kong University of Science and Technology
Econometrics of Crosssectional and Panel Data
ECON 4284

Fall 2016
Econ 4284
Topic 4: Panel Data
Part C
DifferencesinDifferences
(SW Page 532)
1
Application: differencesindifferences
Econ 4284
This is a special case of panel data: binary regressor with
both entity and time fixed effects for two periods
Suppose we a
The Hong Kong University of Science and Technology
Econometrics of Crosssectional and Panel Data
ECON 4284

Fall 2016
Econ 4284
Topic 6
Maximum Likelihood Estimators (MLE)
1
A Review of Probability DistributionEcon 4284
CDF (cumulative distribution function)
If X is a random variable, then its CDF is defined for
any real number x by
Fx PrX x
So, F 0; F 1, and F takes
The Hong Kong University of Science and Technology
Econometrics of Crosssectional and Panel Data
ECON 4284

Fall 2016
Econ 4284
Topic 7: Binary Dependent Variable
1
Binary Dependent Variables
Econ 4284
We have worked extensively with regression models
in which Y is continuous.
We can predict the effect of smoking on earnings.
We can predict the effect of exogenous cha
The Hong Kong University of Science and Technology
Econometrics of Crosssectional and Panel Data
ECON 4284

Fall 2016
Econ 4284
Topic 7 (part B)
Binary Dependent Variable
Estimation and Inference
1
Nonlinear Least Square Estimation
Econ 4284
When G(z) = z we use ordinary least squares,
which picks the betas to minimize the sum of
squared errors
n
i1 Yi 0 1 X 1i . . . k X
The Hong Kong University of Science and Technology
ECON 4284

Spring 2015
EC 3024
Managerial
Macroeconomics
Introductory Session
We
Instructor: Yang LU ([email protected])
Office Hours: by appointment
Room: 6059 LSK
TA: TBA
Office Hours: TBA
Room: 6066A LSK
2
Relationship with classmates
Your classmates are NOT your competitors
The
The Hong Kong University of Science and Technology
ECON 4284

Spring 2015
Econ 4284
Topic 5: Instrumental Variables (IV)
Part B
1
Outline
Econ 4284
Deriving the IV estimator
Properties of IV estimator
Understanding the IV estimator
(2SLS or TSLS and MOM)
An Example in Stata
2
Instrumental Variable Estimator
Econ 4284
Instr
The Hong Kong University of Science and Technology
ECON 4284

Spring 2015
EC3024
Managerial Macroeconomics
Openness in Goods and Financial
Markets
The Open Economy
1.
2.
3.
Openness has three distinct dimensions:
Openness in goods markets. The ability of consumers and
firms to choose between domestic goods and foreign goods.
Fr
The Hong Kong University of Science and Technology
ECON 4284

Spring 2015
EC3024
Managerial Macroeconomics
Exchange Rate Regimes, Expectations,
and Currency Crisis
Fixed Exchange Rates
Central banks act under implicit and explicit exchangerate
targets and use monetary policy to achieve those targets.
Pegs, Crawling Pegs, Bands
The Hong Kong University of Science and Technology
ECON 4284

Spring 2015
EC3143
Macroeconomics Theory II
The Crisis
From a Housing Problem to a Financial Crisis
The Global Recession.
World GDP Growth,
20072010.
World GDP declined
sharply in both the
last quarter of 2008
and the first quarter
of 2009.
Emerging market countries
The Hong Kong University of Science and Technology
ECON 4284

Spring 2015
EC3024
Managerial Macroeconomics
Goods and Financial Markets:
The ISLM Model and Expectations
The ISLM Model without Expectations
IS r e la tio n : Y C (Y T ) I (Y ,i ) G
M
L M r e la tio n :
Y L (i)
P
Figure 5  6
The ISLM Model
Equilibrium in the
goo
The Hong Kong University of Science and Technology
ECON 4284

Spring 2015
EC3024
Managerial
Macroeconomics
Determination of Interest Rates
and the LM Curve
The Supply of Money
Major players:
1.
2.
3.
4.
Central Bank (e.g. the Federal Reserve)
Banks
Depositors
Borrowers from Banks
Tools of Central Banks
1.
2.
3.
4.
Open Market O
The Hong Kong University of Science and Technology
Econometrics of Crosssectional and Panel Data
ECON 4284

Fall 2016
Problem Set 3, ECON 4284, 2016, HKUST
(Due: December 5, Monday)
Please submit your problem set to the mailbox with the label ECON4284 on the sixth
oor of LSK building, which is close to the general o ce of economics department.
1. (30 points) Consider the
The Hong Kong University of Science and Technology
Econometrics of Crosssectional and Panel Data
ECON 4284

Fall 2016
Econ 4284
Numerical Minimization
1
Numerical Minimization
Econ 4284
We want to minimize the nonlinear square
function, or we want to maximize the loglikelihood function:
For example:
min f0 ,1 ,2
0 ,1 ,2
n
y
i1 i
0 1 x 1i 2 x 2i 2
2
Numerical Minimiza
The Hong Kong University of Science and Technology
Econometrics of Crosssectional and Panel Data
ECON 4284

Fall 2016
ECON4284 Topic 1: LLN, CLT and Sampling Distribution
August 29, 2016
Basic Ideas:
1. Averages approximate expectations. To be precise, averages are consistent estimators
of expectations. Consistency is a statistical term that implies that two random varia
The Hong Kong University of Science and Technology
Econometrics of Crosssectional and Panel Data
ECON 4284

Fall 2016
Econ 4284
Topic 4: Panel Data
(Part A)
1
Regression with Panel Data
Econ 4284
A panel dataset contains observations on multiple entities
(individuals), where each entity is observed at two or more
points in time.
Hypothetical examples:
Data on 420 Califo
The Hong Kong University of Science and Technology
Econometrics of Crosssectional and Panel Data
ECON 4284

Fall 2016
Econ 4284
Welcome to
Econ 4284
Econometrics of
Crosssectional and Panel Data
1
Teaching Team
Econ 4284
Instructor: Xun (Sean) Lu ()
Associate Professor in Economics
Specialty: Econometrics
TA: TBA
2
Textbook
Econ 4284
Required:
Stock, James H., and Ma
The Hong Kong University of Science and Technology
Econometrics of Crosssectional and Panel Data
ECON 4284

Fall 2016
Problem Set 1, ECON 4284, Fall 2016
Due on October 3 (Monday)
1
Question 1 (20%):
This years graduating class at Uniform University contains exactly 600 seniors. Each senior
has a random number of parents (0, 1, or 2) who will attend the graduation ceremo
The Hong Kong University of Science and Technology
Econometrics of Crosssectional and Panel Data
ECON 4284

Fall 2016
Econ 4284
Topic 5: Instrumental Variables (IV)
Part A
1
Randomization
Econ 4284
Consider y = b0 + b1x + e ,
From last lecture, we know that x and e are uncorrelated
if x is randomly assigned.
However, most applied research in economics uses
observationa
The Hong Kong University of Science and Technology
Econometrics of Crosssectional and Panel Data
ECON 4284

Fall 2016
Econ 4284
Topic 2B: Internal and External Validity
1
Internal Validity
Econ 4284
Internal validity: the statistical inferences about
causal effects are valid for the population being
studied.
1. Violation of E(ux)=0, or there is correlation between
u an
The Hong Kong University of Science and Technology
Econometrics of Crosssectional and Panel Data
ECON 4284

Fall 2016
Econ 4284
Topic 1Part B
Sampling Distribution,
Central Limit Theorem,
Asymptotic Distributions of OLS Estimators
1
Sampling
Econ 4284
A sample
Sampling is the process of taking a smaller group of
subjects from a larger population.
2
Sampling
.
. .
. .
.
The Hong Kong University of Science and Technology
Econometrics of Crosssectional and Panel Data
ECON 4284

Fall 2016
Econ 4284
Topic 1Part A
Law of Large Numbers and
Consistency of OLS estimator
1
LLN and CLT
Econ 4284
Two cornerstones of statistics and econometrics:
Law of Large Numbers (LLN)
Central Limit Theorem (CLT)
Basic Ideas
LLN: the sample average becomes cl
The Hong Kong University of Science and Technology
Econometrics of Crosssectional and Panel Data
ECON 4284

Fall 2016
Econ 4284
Topic 4: Panel Data
(Part B)
1
Fixed Effects Regression
Econ 4284
W h at if y o u h av e m o re th an 2 tim e p erio d s ( T > 2 )?
Y it = 0 + 1 X it + 2 Z i + u it , i = 1 , ,n , T = 1 , ,T
W e can rew rite th is in tw o u sefu l w ay s:
1.
F i
The Hong Kong University of Science and Technology
Econometrics of Crosssectional and Panel Data
ECON 4284

Fall 2016
Econ 4284
Topic 2A: Correlation and Causality
1
Econ 4284
Two main uses of regression:
causal inference
forecasting
2
Causal inference
Econ 4284
y = alpha+x*beta + u,
Causal inference:
we are interested in the effects of x on y
x is a potential caus
The Hong Kong University of Science and Technology
ECON 4284

Spring 2015
Econ 4284
Topic 5: Instrumental Variables (IV)
Part C
Instrumental Variables & 2SLS
Multiple Regression
1
IV Estimation
Econ 4284
IV estimation can be extended to the multiple regression
case
Call the model we are interested in estimating the
structural