Large-N and Large-T Properties of Panel Data Estimators and the Hausman Test



where G1 = G-,1TA1-1G-,1T .

Notice that

θT GR1G-,1T´ = θTG-T (A1 + θ2Fι)-1 G⅛
×Gx,TF1Gx,TG1Gx,TF1Gx,TG1.

Lemma 2(a) and Assumption 12 imply

G1 =Op(1).

In addition, by Lemma 2(c)

Gx,TF1Gx,T= Jx-,T1 Dx,T F1Dx,T Jx-,T1 =Op(1),           (102)

since Jx-,T1 = O (1) . Thus,

σ2θ2r (g-,1tRiG⅛´ = Op (θT) = op (1).             (103)

Now, consider

θτ √NT θT G-,1τ R2

=  θτT [√Tg-,1t (A-1F1A-1F2) + Tg-,1tRi (A2 + θTF2)]

n'2  -~ 1   - 1z-y -1            -   ʌf/ɑ-1   -1 -1        √~V

= θτ VT


θT Gx,T A1 Gx,T (Gx,TF1Gx,T) Gx,T A1 Gx,T Gx,T NF2

+θT (g-,1tRiG-,tX√NGχ,τA2 + θTNGχ,τF2)

Under the local alternatives (Assumption 11), we may deduce from Lemmas 2
and 3 that

Gι, Gχ,τFιGχ,τ, Gχ,τVNf2 = Op (1) ;

θT GRiG-,T ) , NGχ,τA2 = op (1).

Since θτ T = O (1) , we have

θτVntθTg,1tR2 = O (1) 2τOp (1) + op (1)] = op (1).      (104)

Using the results (103) and (104) , we now can approximate the Hausman
statistic as follows:

HMNT

= θτNT [(-T1DDχ,τFιDχ,τ J-1τ) (GιGχ,τA2) - J-TDχ,τF2 + op (1)] 0
σv

× ( Jχ-τDχ,τFτDχ,τ J-τ + op (1)) -l

×θτNT [ J-τDχ,τFτDχ,τ JχΓ (GτGχ,τA2) - J-τDχ,τF2 + op (1)]
σv

= θτNT (Dχ,τF2)0 (Dχ,τFτDχ,τ)-1 — √NT (Dχ,τF2)

σv                                  σv

+op (1) .

62



More intriguing information

1. What Lessons for Economic Development Can We Draw from the Champagne Fairs?
2. A Computational Model of Children's Semantic Memory
3. The name is absent
4. The name is absent
5. Federal Tax-Transfer Policy and Intergovernmental Pre-Commitment
6. Picture recognition in animals and humans
7. Moi individuel et moi cosmique Dans la pensee de Romain Rolland
8. The name is absent
9. The name is absent
10. The name is absent