Handling the measurement error problem by means of panel data: Moment methods applied on firm data



or, compactly,

yi = (∆Xi)β + ∆i.

The IV matrix (cf. Proposition 1) is the ((2T - 3) × KT(T - 2)) matrix

xi (21)
..
.

∙∙       0

.

..

0

.

.

∙∙      0

..

..

(41)            Zi=

.

0

..

∙∙ xi(T,T-1)

.

0

..

∙∙      0

0
.

∙∙       0

..

xi2

..

∙∙      0

.

.

.

0

..

..

∙∙       0

.

.

0

..

..

∙∙ xi,T-1

We here use different IV’s for the (T - 1)+(T - 2) equations in (40), with β as a
common slope coefficient. Let

y= [(∆y1)0,...,(∆yN)0]0,     ∆= [(∆1)0,...,(∆N)0]0,

X= [(∆X1)0,...,(∆XN)0]0, Z=[Z10,...,ZN0]0.

The overall GMM estimator corresponding to (32), which we now write as
E[Zi0(∆i)] = 0T (T -2)K,1, minimizing [N-1 (∆) 0Z](N-2V )-1[N-1Z0(∆)] for V =
Z 0Z, can be written as

(42)      βbDx = [(∆X) 0Z(Z0Z)-1Z0(∆X)]-1 [(∆X) 0Z(Z0Z)-1Z0(∆y)]

-1

= [Pi(∆Xi)0Zi][PiZi0Zi]-1 [PiZi0(∆Xi)]
× h[Pi(∆Xi)0Zi][PiZi0Zi]-1 [Pi Zi0(∆yi)]i .

This estimator examplifies (23). The consistency of βDx relies on Assumptions
(A), (B1), and (E).
If ∆ has a non-scalar covariance matrix, a more efficient
estimator is obtained for
V = V Z(∆) = E[Z0(∆)(∆)0Z], which gives

-1

βeDx = (∆X)0ZVZ-(1∆)Z0(∆X)    (∆X)0ZVZ-(1∆)Z0(∆y) .

We can estimate VZ(δ,) /N consistently from the residuals obtained from (42),
dei = ∆yi - (∆Xi)βDx, by cZ(∆,)/N = (1 /N) P= Z'(def)(def)0Zi. The re-
sulting asymptotically optimal GMM estimator, which examplifies (24), is

0                               -1

(43)     βeDx = [Pi(∆Xi)0Zi][Pi Zi0didi0Zi]-1[Pi Zi0(∆Xi)]

0

× [Pi(∆Xi) 0Zi][Pi Zi0didi0Zi]-1[Pi Zi0(∆yi)] .

21



More intriguing information

1. The name is absent
2. DISCUSSION: POLICY CONSIDERATIONS OF EMERGING INFORMATION TECHNOLOGIES
3. Fiscal Rules, Fiscal Institutions, and Fiscal Performance
4. The name is absent
5. Why Managers Hold Shares of Their Firms: An Empirical Analysis
6. Temporary Work in Turbulent Times: The Swedish Experience
7. Regionale Wachstumseffekte der GRW-Förderung? Eine räumlich-ökonometrische Analyse auf Basis deutscher Arbeitsmarktregionen
8. Willingness-to-Pay for Energy Conservation and Free-Ridership on Subsidization – Evidence from Germany
9. The name is absent
10. Picture recognition in animals and humans