Testing the Information Matrix Equality with Robust Estimators



C.2 Skewed normal alternative

The skewed-normal log-density is

-2 log(2π) - log σ - u2,               if y ≤ β;

log f (У ; β,σ,γ) = <                                      2

[ - 1 log(2 π ) - log σ + log(1 + γ ) - uγ, if y>β ;

where u = (y - β) and uγ = u(1 + γ). The score function, evaluated at
Y = 0, is

(u∕σ

( u2 - 1)

I(y > β)(u2 - 1)

Now let Fn be skewed normal with β = 0, σ = 1, and γn = e∕√n > 0.
Then,
γоn = 0, and by the results of Appendix B.2,

β0 ,n  = EFn ( Y ) = γnφ (0),

σ о ,n  = 1+^2^+o ( γn ) '

It follows that

I(Y>βо,n)} j


EFn [ s ( Y ; β0 ,n0 ,n, 0)] = -


0
o ( Yn ) .
EFn { [(Y-вП)2 - 1

The third element in parentheses is

∞ ,     Yn). - β0n ʌ 2 dφw - 1 + Φ(βоn)

∙M 0 ,n σ 0 ,n /

= (----—) (ββ0,nφ(β0,n) + 1 - ф(β0,n)-------2 0П ’ φ(β0,n) - 1

σ 0 ,n J                                          σ 0 ,n

+H β0 ,n ) + o ( γη )

- 2Yn (φ(0)) - 1 + Ф(β0,n)


= (1 + γn )( γn (φ (0)) +1 - Ф( β0 ,n ))

+O ( Yn )

= Yn (I - (φ(0))2) + 0(Yn)

= Yn ( π ) + 0 ( Yn ).

34



More intriguing information

1. The migration of unskilled youth: Is there any wage gain?
2. Globalization, Redistribution, and the Composition of Public Education Expenditures
3. Une Gestion des ressources humaines à l'interface des organisations : vers une GRH territoriale ?
4. The name is absent
5. AN ANALYTICAL METHOD TO CALCULATE THE ERGODIC AND DIFFERENCE MATRICES OF THE DISCOUNTED MARKOV DECISION PROCESSES
6. The name is absent
7. Industrial districts, innovation and I-district effect: territory or industrial specialization?
8. The name is absent
9. The name is absent
10. Pupils’ attitudes towards art teaching in primary school: an evaluation tool