Econometrics I: Professor William Greene Stern School of Business Department of Economics
Econometrics I: Professor William Greene Stern School of Business Department of Economics
Econometrics I: Professor William Greene Stern School of Business Department of Economics
Part 6 – Estimating
the Variance of b
e y - Xb
y X ( X ' X )1 X ' y
1
[I X( X ' X) X ']y
My M( X ) MX M M
e'e (M'(M
'M'M 'MM 'M
(X’X)-1
s2(X’X)-1
b = + (X'X)-1 * i 1 xi i
n
= A sandwich matrix. = A B A
What does the variance of the sum (the meat) look like?
Leading cases.
(1) Heteroscedasticity
(2) Autocorrelation
(3) Grouped (clustered) observations with common effects.
i1 i
n 2
2
e
Use
ˆ
n
nei2 ˆ ˆ )=n
ˆ i = 2 , Ω=diag( ˆ i ) note tr(Ω
ˆ
ˆ X'X X'ΩX
2 1
ˆ X'X
Est.Var[b]
n n n n
ˆ
2 X'ΩX 2 X'ΩX
Does
ˆ n 0?
n
6-23/49 Part 6: Estimating the Variance of b
Groupwise Heteroscedasticity
Countries
are ordered
by the
standard
deviation of
their 19
residuals.
Regression of log of per capita gasoline use on log of per capita income,
gasoline price and number of cars per capita for 18 OECD countries for 19
years. The standard deviation varies by country. The “solution” is
“weighted least squares.”
n
l
wl 1 = "Bartlett weight"
L 1
1 1
1 X'X X'X
Est.Var[b]= [S 0 S1 ]
n n n
= ( X'X ) 1 Ni=1 Ct=1
i
Cs=1
i
eit eis x it x is ( X'X ) 1
e a least squares residual
(If Ci = 1, this is the White estimator.)
Bootstrapped sample
results
Asymptotic Theory
Asymptotic Variance:
1
N
A 1 C A 1
1 11
A = E[f u (0) xx] Estimated by i 1 1| uˆi | B xi xi
N
N B2
Bandwidth B can be Silverman's Rule of Thumb:
1.06 Q(uˆi | .75) Q(uˆi | .25)
Min su ,
.2
N 1.349
(1- )
C = (1- ) E[xx] Estimated by XX
N
2
su XX .
1
For =.5 and normally distributed u, this all simplifies to
2
But, this is an ideal application for bootstrapping.
= .50
= .75