15 Implications of Conditional Homoskedasticity

# 15 Implications of Conditional Homoskedasticity - ³ x tk x...

This preview shows pages 1–7. Sign up to view the full content.

Implications of Conditional Homoskedasticity Hayashi 2.6 Goal: behavior of test statistics under conditional ho- moskedasticity Conditional Homoskedasticity (Assumption 2.7) E u 2 t j x t ± = 2 > 0 Unconditional Homoskedasticity E u 2 t ± = 2 > 0 Conditional , Unconditional?

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Relation Between Conditional and Unconditional Conditional ) Unconditional E u 2 t ± = E X h E u 2 t j x t ±i = 2 Unconditional ; Conditional Example u t = x t v t v t independent of x t Ev 2 t = 2 Ex 2 t = ! 2 E u 2 t ± = E x 2 t ± E v 2 t ± = ! 2 2 E u 2 t j x t ± = x 2 t E v 2 t j x t ± = x 2 t 2
Estimation of S Need to estimate S = E u 2 t x t x 0 t ± Conditional homoskedasticity estimator ^ S = ^ 2 X x t x 0 t ^ u t = y t x 0 t ^ ± ^ 2 = 1 n P ^ u 2 t ^ S consistent for S

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Consistency of S S = E x E u 2 t x t x 0 t j x t ± = E h x t x 0 t E u 2 t j x t ±i = E x t x 0 t 2 ± = 2 E x t x 0 t ± Implications 1) A2.4 redundant A 2.5 S nonsingular ) 2 > 0 ) E ² x t x 0 t ³ nonsingular (A2.4)
Assumption Implications A 2.4 redundant and 1 n P t x t x 0 t p ! E x t x 0 t ± (ergodic stationarity A2.2) 1 n P t ^ u 2 t p ! 2 (A2.1-A2.4) no need for A2.6 E ²

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: ³ x tk x tj ´ 2 µ < 1 Asymptotic Behavior Proposition 2.5 Under A 2.1-2.5, 2.7 a) (asymptotic distribution of b ) The OLSE b of & is consistent and asymptotically normal with Avar ( b ) = ± 2 h E & x t x t ±i & 1 b) (consistent estimation of asymptotic variance) The quantity Avar ( b ) is consistently estimated by \ Avar ( b ) = ^ ± 2 ² P x t x t ³ & 1 Asymptotic Behavior continued c) (asymptotic distribution of test statistics) Under H : & = & , t statistic is asymptotically distributed as N (0 ; 1) Under H : R& = r , # r & F is asymptotically distributed as ± 2 (# r )...
View Full Document

{[ snackBarMessage ]}

### Page1 / 7

15 Implications of Conditional Homoskedasticity - ³ x tk x...

This preview shows document pages 1 - 7. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online