Testing the Information Matrix Equality with Robust Estimators



and for which ρc(0) = 0. This estimator is consistent for σ and has influence
function (Huber, 1981)

IF(u; σ∙,Φ) = Pc(^> - bc .

E φ [ pc ( u ) u ]

For ^ to be robust, ρc has to be bounded and / has to be a robust estimator
of location (e.g. the median). The ML estimator of
σ, for example, is an
M-estimator of scale for which
ρc(u) = u2 (unbounded), bc = 1, and / is the
sample average, and hence is non-robust.

of scale


defined by Tukey’s biweight


Consider the robust M-estimator
(TB) function,

{u 2

2

C2

6 ,


u 4
2c2


+ u6

+ 6c4,


if |u| ≤c;


if |u| >c.


The choice of c is governed by the desired breakdown point of ^. The details
of how to compute
T , for any choice of c, are given in Appendix A. Table 1
gives the numerical results for 10%, 25% and 50% breakdown points that
are needed to compute
T using the TB estimator.

Another, simpler, robust M-estimator of scale is the (standardized) Me-
dian Absolute Deviation (MAD),

σ = Ф-1(3/4) medi{lYi - medjYj|}

for which ρc(u) = I(|u| ≥ c) with c = Φ-1(3/4) = 0.6745 and bc = 1 /2. The
breakdown point of the MAD is 50%. Table 1 gives the constants, derived
in Appendix A, that are needed to compute
T using the MAD.

4 Local asymptotic power

Let Fn be a sequence of local alternatives tending to the null distribution,
i.e.
Fn Fθ. Under a sequence of local alternatives,

Hn : Y ~ Fn,

the IM indicator M is, given some regularity conditions, asymptotically
normally distributed,

n(M ( M - mn ) 4 N(0 ,v ),



More intriguing information

1. WP 92 - An overview of women's work and employment in Azerbaijan
2. The name is absent
3. The name is absent
4. sycnoιogιcaι spaces
5. The name is absent
6. The name is absent
7. Washington Irving and the Knickerbocker Group
8. Macro-regional evaluation of the Structural Funds using the HERMIN modelling framework
9. The name is absent
10. The name is absent