slides_Ch2_W[1]

1 slr4 fails if slr1 fails see chp 13 for methods that

Info iconThis preview shows page 1. Sign up to view the full content.

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

Unformatted text preview: xi x ) ˆ = in 1 β 1 = i =1 n 2 ¯ ¯2 ∑i =1 (xi x ) ∑i =1 (xi x ) Melissa Tartari (Yale) Econometrics (2.49) 51 / 93 Unbiasedness of the OLS Estimator III Recall ¯ ¯ ¯ ∑n= yi (xi x ) ∑n (yi y ) (xi x ) ˆ = in 1 β 1 = i =1 n 2 ¯ ¯2 ∑i =1 (xi x ) ∑i =1 (xi x ) (2.49) ˆ We can then write β1 in terms of ( βo , β1 ), the u 0 s , and the x 0 s : ∑n ( β + β xi + ui ) (xi ˆ β 1 = i =1 o n 1 ¯2 ∑i =1 (xi x ) x) ¯ = β1 + ¯ ∑n=1 ui (xi x ) i n ¯2 ∑ (xi x ) | i =1 {z } generally 6=0 Melissa Tartari (Yale) Econometrics 51 / 93 Unbiasedness of the OLS Estimator IV - Theorem Theorem 2.1: under assumptions SLR.1 to SLR.4 ˆ ˆ E β1 jx = β1 and E βo jx = βo . Melissa Tartari (Yale) Econometrics (2.53) 52 / 93 Unbiasedness of the OLS Estimator IV - Theorem Theorem 2.1: under assumptions SLR.1 to SLR.4 ˆ ˆ E β1 jx = β1 and E βo jx = βo . (2.53) ˆ This means that the distribution of β1 is centered about β1 and the ˆ is centered about β . distribution of βo o Melissa Tartari (Yale) Econometrics 52 / 93 Unbiasedness of the OLS Estimator V - Proof Proof of Theorem 2.1: ˆ E β1 jx = E [ β1 jx ] + E " ¯ ∑n=1 ui (xi x ) i jx n ¯2 ∑i =1 (xi x ) # ¯ ∑n=1 E [ui jx ] (xi x ) i 2 n ¯ ∑i =1 (xi x ) n ¯ ∑ = 0 (xi x ) = β1 + in 1 = β1 + 0 = β1 ¯2 ∑i =1 (xi x ) = β1 + and ˆ E βo jx = E y ¯ M elissa Tartari (Yale) ˆ¯ β1 x jx = y ¯ ˆ E β1 jx x = y ¯ ¯ Econometrics β1 x = βo . ¯ 53 / 93 Unbiasedness of the OLS Estimator: Failures I Unbiasedness generally fails if any of SLR.1-SLR.4 fails: if SLR.1 fails =) see Chp 13 for methods that handle non-linear models, Melissa Tartari (Yale) Econometrics 54 / 93 Unbiasedness of the OLS Estimator: Failures I Unbiasedness generally fails if any of SLR.1-SLR.4 fails: if SLR.1 fails =) see Chp 13 for methods that handle non-linear models, if SLR.2 fails (e.g. when samples are not representative of the population because the data is collected by over-sampling some sub-populations)...
View Full Document

This note was uploaded on 02/13/2014 for the course ECON 350 taught by Professor Donaldbrown during the Fall '10 term at Yale.

Ask a homework question - tutors are online