econometric analysis of panel data panel data analysis – random effects assumptions gls estimator...
TRANSCRIPT
Econometric Analysis of Panel Data
• Panel Data Analysis– Random Effects• Assumptions• GLS Estimator• Panel-Robust Variance-Covariance Matrix• ML Estimator
– Hypothesis Testing• Test for Random Effects• Fixed Effects vs. Random Effects
Panel Data Analysis
• Random Effects Model
– ui is random, independent of eit and xit.
– Define it = ui + eit the error components.
' ( 1,2,..., )
( 1,2,..., )i
it it i it i
i i i T i
y u e t T
u i N
x β
y X β i e
Random Effects Model
• Assumptions– Strict Exogeneity
• X includes a constant term, otherwise E(ui|X)=u.
– Homoschedasticity
– Constant Auto-covariance (within panels)
( | ) 0, ( | ) 0 ( | ) 0it i itE e E u E X X X
2 2 '( | )i i ii e T u T TVar ε X I i i
2 2
2 2 2
( | ) , ( | ) , ( , ) 0
( | )
it e i u i it
it e u
Var e Var u Cov u e
Var
X X
X
Random Effects Model
• Assumptions– Cross Section Independence
2 2 '
1
2
( | )
0 0
0 0( | )
0 0
i i ii i e T u T T
N
Var
Var
ε X I i i
ε X Ω
Random Effects Model
• Extensions– Weak Exogeneity
– Heteroscedasticity and Autocorrelation
– Cross Section Correlation
1 2
1 2
( | , ,..., ) ( | ) 0
( | , ,..., ) 0
( | ) 0
iit i i iT it i
it i i it
it it
E E
E
E
x x x X
x x x
x
2( | )itit iVar X
( , | , ) 0it jt i j ijCorr X X
Model Estimation: GLS
• Model Representation
2 2 '
2 22
2
'
,
( | )
( | )
1
i
i i i
i
i i i
i i i i i T i
i i
i i i e T u T T
e i ue i T i
e
i T T Ti
u
E
Var
TQ Q
where QT
y X β ε ε i e
ε X 0
ε X I i i
I
I i i
Model Estimation: GLS
• GLS
11 1 1 1 1
1 1
11 1 1
1
21
2 2 2
21/2
2 2
ˆ ( )
ˆ( ) ( )
1
1
i
i
N N
GLS i i i i i ii i
N
GLS i i ii
ei i T i
e e i u
ei i T i
e e i u
Var
where Q QT
and Q QT
β XΩ X XΩ y X X X y
β XΩ X X X
I
I
Model Estimation: RE-OLS
• Partial Group Mean Deviations' '
'
2
2 2
' '
'
( )
( )
1
( ) [(1 ) ( )]
it it it it i it
i i i i
ei
e i u
it i i it i i i i it i i
it i it
y u e
y u e
T
y y u e e
y
x β x β
x β
x x β
x β
Model Estimation: RE-OLS
• Model Assumptions
• OLS
'
' 2 2 2 2 2
' 2 2 2 2
2
2 2
( | ) 0
( | ) (1 ) (1 2 / / )
( , | ) (1 ) ( 2 / / ) 0,
: 1
it i
it i i u i i i i e e
it is i i u i i i i e
ei
e i u
E
Var T T
Cov T T t s
NoteT
x
x
x
1' 1 ' '
1 1
12 ' 1 2 '
1
2
ˆ ( )
ˆˆ ˆ ˆ( ) ( )
ˆ ˆ ˆ ˆˆ ' / ( ),
N N
OLS i i i ii i
N
OLS e e i ii
e
Var
NT N K
β XX Xy X X X y
β XX X X
ε ε ε y Xβ
Model Estimation: RE-OLS
• Need a consistent estimator of :
– Estimate the fixed effects model to obtain– Estimate the pooled model to obtain– Based on the estimated large sample variances, it
is safe to obtain
2
2 2
ˆˆ 1ˆ ˆ
ei
e i uT
2ˆe2 2ˆ ˆe u
ˆ0 1
Model Estimation: RE-OLS
• Panel-Robust Variance-Covariance Matrix– Consistent statistical inference for general
heteroscedasticity, time series and cross section correlation.
1 1' ' ' '
1 1 1
1 1' ' '
1 1 1 1 1 1 1
ˆ ˆ ˆˆ ( ) [( )( ) ']
ˆ ˆ
ˆ ˆ
ˆˆ ˆ ,
i i i i
N N N
i i i i i i i ii i i
N T N T T N T
it it it is it is it iti t i t s i t
i i i it
Var E
β β β β β
X X X ε ε X X X
x x x x x x
ε y X β
' ˆit ity x β
Model Estimation: ML
• Log-Likelihood Function
' '
2 2 '
2 2 1
( ) ( 1,2,..., )
( 1, 2,..., )
~ ( , ),
1 1( , , | , ) ln 2 ln
2 2 2
i i i
it it i it it it i
i i i
i i i e T u T T
ii e u i i i i i i
y u e t T
i N
iidn
Tll
x β x β
y X β ε
ε 0 I i i
β y X ε ε
Model Estimation: ML
• ML Estimator
2 2 2 2
1
2 2 1
2 22
2
2 2' 2 '
2 2 21 1
ˆ ˆ ˆ( , , ) argmax ( , , | , )
1 1( , , | , ) ln 2 ln
2 2 2
1ln 2 ln
2 2
1( ) ( )
2i i
N
e u ML i e u i ii
ii e u i i i i i i
i e ue
e
T Tuit it it itt t
e e i u
ll
where
Tll
T T
y yT
β β y X
β y X ε ε
x β x β
Hypothesis Testing
• Test for Var(ui) = 0, that is
– If Ti=T for all i, the Lagrange-multiplier test statistic (Breusch-Pagan, 1980) is:
, , ,( ) ( ) ( )it is i it i is it isCov Cov u e u e Cov e e
22
1 1 2
2
1 1
'
ˆ1 ~ (1)
2 1 ˆ
ˆˆ 1
ˆ
N T
iti t
N T
iti t
it it it
Pooled
eNTLM
T e
where e yu
βx
Hypothesis Testing
– For unbalanced panels, the modified Breusch-Pagan LM test for random effects (Baltagi-Li, 1990) is:
– Alternative one-side test:
22 2
1 1 1 2
2
1 11
ˆ1 ~ (1)
ˆ2 ( 1)
i
i
N N T
i iti i t
N TN
iti i i ti
T eLM
eT T
0~ (0,1)
: Pr ( )n
LM N under H
P Value z LM
Hypothesis Testing
• Fixed Effects vs. Random Effects '
0
'1
: ( , ) 0 ( )
: ( , ) 0 ( )
i it
i it
H Cov u random effects
H Cov u fixed effects
x
x
Estimator Random EffectsE(ui|Xi) = 0
Fixed EffectsE(ui|Xi) =/= 0
GLS or RE-OLS(Random Effects)
Consistent and Efficient
Inconsistent
LSDV or FE-OLS(Fixed Effects)
ConsistentInefficient
ConsistentPossibly Efficient
Hypothesis Testing
• Fixed effects estimator is consistent under H0 and H1; Random effects estimator is efficient under H0, but it is inconsistent under H1.
• Hausman Test Statistic ' 1
2
ˆ ˆ ˆ ˆ ˆ ˆ( ) ( )
ˆ ˆ ˆ~ (# ), # # ( )
RE FE RE FE RE FE
FE FE RE
H Var Var
provided no intercept
β β β β β β
β β β
Hypothesis Testing
• Alternative Hausman Test– Estimate the random effects model
– F Test that = 0
' ' ' '( ) ( ) ( )it i it i it i ity y e x x β x x γ
0 0: 0 : ( , ) 0i itH H Cov u γ x
Example: Investment Demand
• Grunfeld and Griliches [1960]
– i = 10 firms: GM, CH, GE, WE, US, AF, DM, GY, UN, IBM; t = 20 years: 1935-1954
– Iit = Gross investment
– Fit = Market value
– Cit = Value of the stock of plant and equipment
it i it it itI F C