1. Trang chủ
  2. » Thể loại khác

Chapter 11_Seemingly Unrelated Regressions (SUR)

8 46 0

Đang tải... (xem toàn văn)

THÔNG TIN TÀI LIỆU

Thông tin cơ bản

Định dạng
Số trang 8
Dung lượng 117,43 KB

Nội dung

Advanced Econometrics Chapter 11: Seemingly unrelated regressions Chapter 11 SEEMINGLY UNRELATED REGRESSIONS I MODEL Seemingly unrelated regressions (SUR) are often a set of equations with distinct dependent and independent variables, as well as different coefficients, are linked together by some common immeasurable factor Consider the following set of equations: there are = Y1 (T ×1) = Y2 (T ×1) X β + ε1 country X β + ε2 country 1 (T ×k ) ( k ×1) 2 (T ×k ) ( k ×1) 1, 2, … M, such that (T ×1) (T ×1) … = YM X M M + M (T ì1) (T ×k ) ( k ×1) Assume each country M (T ×1) (i = 1, 2, …, M) meets classical assumptions so OLS on each equation separately in fine • Although each of M equations may seem unrelated, the system of equations may be linked through their mean – zero error structure • We use cross-equation error covariance to improve the efficiency of OLS M equations are estimated as a system ' E (ε= σ= σ ii IT iε i ) i IT E (ε iε 'j ) = σ ij IT Where ij: contemporaneous covariance between errors of equations i and j Nam T Hoang University of New England - Australia University of Economics - HCMC - Vietnam Advanced Econometrics Chapter 11: Seemingly unrelated regressions    X1  Y1   (T ×k )   (T ×1)     X2 Y  =  (T ×k )              Y   0  XM   (TM×1)   (T ×k )      (TM ×kM ) ( MT ×1)  β1   ε1   ( k ×1)   (T ×1)  β  ε   +        β  ε   ( k ×N1)   (T ×N1)    ( kM ×1) ( NT ×1) Assumption: there is a such that: (1) ↔= Y Xβ +ε  σ 11 I σ I E (εε ′) =  21   ( MT × MT )  σ M I  σ 11 σ  21 Where: Σ =   σ M II σ 12 I σ 22 I  σ 1M I   σ M I  = Σ⊗ I      σ M I  σ MM I  σ 12  σ 1M  σ 22  σ M       σ M  σ MM  GENERALIZED LEAST SQUARES ESTIMATION OF SUR MODEL (GLS) The equation (1) can be estimated by GLS if E(εε’) is known: βˆSUR = [ X '( E (εε ')) −1 X ]−1[ X '( E (εε ')) −1Y ] βˆSUR = [ X '(Σ ⊗ I ) −1 X ]−1[ΣX⊗'( I ) −1Y ] GLS is the best linear unbiased estimator: ( ) VarCov βˆSUR = [ X '( E (εε ')) −1 X ]−1 Advantages of SUR over single-equation OLS Gain in efficiency: Because βˆSUR will have smaller varriance than βˆOLS Nam T Hoang University of New England - Australia University of Economics - HCMC - Vietnam Advanced Econometrics βˆOLS Chapter 11: Seemingly unrelated regressions  βˆ1( OLS )   ( k ×1)    β =  2( OLS )        β M ( OLS )  k ×1) (  (TM ×1) Note that βˆi (OLS ) is efficient estimator for i, but βˆOLS is not efficient estimator for , and βˆSUR is efficient estimator for Test or impose cross-section restriction (Allowing to test or impose) Usually E(εε’) unknown Feasible GLS estimation Estimate each equation by OLS, save residuals ei , i = 1, 2, …, M (T ×1) Compute sample variances and covariances T σˆ ij = ∑e e it t =1 jt  σˆ11 σˆ12  σˆ σˆ 22 Σ = 21     σˆ M σˆ M E (εε ′) = ( MT × MT ) all ij pairs T −k  σˆ1M   σˆ M  =    T −k   σˆ MM   e1/ e1 e1/ e2  / /  e2 e1 e2 e2     / / eM e1 eM e2 e1/ eM    e2/ eM       eM/ eM   Σˆ ⊗ I ( M ×M ) (T ×T ) βˆFGLS = [ X '(Σˆ ⊗ I ) −1 X ]−1[ΣX⊗'( ˆ I ) −1Y ] → Σˆ is a consistent estimator of It is also possible to interate & until convergence which will produce the maximum likelihood estimator under multivariate normal errors In other words, βˆFGLS and βˆML will have the same limiting distribution such that: asy βˆML , FGLS  N ( β , ϕ ) Nam T Hoang University of New England - Australia University of Economics - HCMC - Vietnam Advanced Econometrics Where Chapter 11: Seemingly unrelated regressions is consistently estimated by = ϕˆ [ X '(Σˆ ⊗ I ) −1 X ]−1 III KRONECKER PRODUCT: Definition: For any two matrices A,B A ⊗ B is defined by the matrix consisting of each element of A time the entire second matrix B Propositions: (1) ( A ⊗ B )( C ⊗ D ) = AC ⊗ BD  a11 B a12 B   c11 D c12 D   a B a B  c D c D  = 22 22  21   21          (2) ( A ⊗ B) −1 ∑ (a ∑ (a c j ) BD   = j c j ) BD   AC ⊗ BD   1j =A−1 ⊗ B −1 if inverses are defined ( A ⊗ B ) ( A−1 ⊗ B −1 ) = Because:  ∑ (a1 j c j1 ) BD   ∑ (a2 j c j1 ) BD    ( AA −1 ) ⊗ BB −1 = I → ( A ⊗ B ) =A−1 ⊗ B −1 −1 (3) IV ( A ⊗ B) / =A/ ⊗ B / (you show) TWO CASE WHEN SUR PROVIDES NO EFFICIENCY GAIN OVER SINGLE OLS: When ij = for all i≠j: the equations are not linked in any fashion and GLS does not provide any efficiency gains → we can show that βˆOLS = βˆSUR ( ) VarCov βˆSUR = [ X '(Σ ⊗ I ) −1 X ]−1 Nam T Hoang University of New England - Australia University of Economics - HCMC - Vietnam Advanced Econometrics Chapter 11: Seemingly unrelated regressions  σ I  11  (Σ ⊗ I ) −1 = Σ −1 ⊗ I = =       (      I  σ 22        I  σ MM   0 ) VarCov βˆSUR =    X 1/   (T × k )   =          X 1/ X   σ 11   =        / (T × k ) X       σ I   11       X M/   (T × k )     X 2/ X σ 22              X M/ X  σ MM  ( X 1/ X ) −1σ 11  / ( X X ) −1σ 22 =     0  (  I      σ 22 σ MM     X1       I   X2               X M     −1 −1         ( X M/ X M ) −1σ MM    ) VarCov βˆiOLS = ( X i/ X i ) −1σ ii ( ) ( ) → VarCov βˆSUR (i ) = VarCov βˆi → no efficiency gains at all Exercise: Show: βˆSUR Nam T Hoang University of New England - Australia  βˆ1OLS     βˆ2OLS  =  in this case     βˆ   MOLS  University of Economics - HCMC - Vietnam Advanced Econometrics Chapter 11: Seemingly unrelated regressions Note: The greater is the correlation of disturbance, the greater is the gain in efficiency in using SUR & GLS The less correlation then is in between the X matrices, the greater is gain in using GLS When X 1= X 2= = X M= X ( ) VarCov βˆSUR = [ X '(Σ −1 ⊗ I ) X ]−1 = [( I ⊗ X ) / (Σ −1 ⊗ I )( I ⊗ X )]−1 = [Σ −1 ⊗ ( X / X )]−1 = Σ ⊗ ( X / X ) −1 σ 11 ( X / X ) −1 σ 12 ( X / X ) −1  σ ( X / X ) −1 σ 22 ( X / X ) −1 =  21      (    → no efficiency gain     σ MM ( X / X ) −1    0  ) VarCov βˆiOLS = σ ii ( X / X ) −1 X 0 X=    0 V X   0       X HYPOTHESIS TESTING: Contemporaneous correlation (spatial correlation): σ ij 0 σ ij / E (ε iε j ) =        0       σ ij  H0: σ ij = for all i≠j HA: H0 false M i −1 LM test statistic: λ = T ∑∑ rij2  χ M2 ( M −1) =i 2=j Where rij is calculated using OLS residuals: Nam T Hoang University of New England - Australia University of Economics - HCMC - Vietnam Advanced Econometrics rij = Chapter 11: Seemingly unrelated regressions ei/ e j (ei/ ei )(e /j e j ) Under H0 → λ  χ M2 ( M −1) If accept H0 → no efficiency gain 2 Restrictions on coefficients: H0: R β = HA: H0 false The general F test can be extended to the SUR system However, since the statistic requires using Σˆ , the test will only be valid asymptotically Where β = ( β1 , β , , β M ) Within SUR framework, it is possible to test coefficient restriction across equations One possible test statistic is: ( R βˆFGLS q ) / [ R VarCov( βˆFGLS ) R /−]−1 ( RβˆFGLS q ) W=  ( m×k ) ( k ×1) ( m×k ) ( k ×m ) ( m×1) k ×k )    (  ( m×1) ((1×m ) ( m×m ) asy W  χ m2 under H0 VI AUTOCORRELATION: Heteroscedasticity and autocorrelation are possibilities within SUR framework I will focus on autocorrelation because SUR systems are often comprised of time series observations for each equation Assume the errors follow: ε i ,t ρiε i ,t −1 + uit = Where uit is white noise The overall error structure will now be: σ 12 Σ12  σ 11Σ11 σ Σ σ 22 Σ 22 E (εε ′) =  21 21     σ M 1Σ M σ M Σ M  ρj     ρj Where: Σij =     T −1 T −2  ρj  ρ j Nam T Hoang University of New England - Australia σ 1M Σ1M   σ M Σ MM      σ MM Σ MM  MT ×MT ρ Tj −1   ρ Tj −1        T ×T University of Economics - HCMC - Vietnam − Advanced Econometrics Chapter 11: Seemingly unrelated regressions  ε i1  ε  / E (ε iε j ) =  i  ε j1 ε j  ε jT       ε iT  Estimation: Run OLS equation by equation by equation Compute consistent estimate of ρi: T ρˆ i = ∑e e it it −1 t =2 T ∑e t =1 it Transform the data, using Cochrane-Orcutt, to remove the autocorrelation Calculate FGLS estimates using the transformed data • Estimate Σ using the transformed data as in GLS • Use Σˆ to calculate FGLS Nam T Hoang University of New England - Australia University of Economics - HCMC - Vietnam ...Advanced Econometrics Chapter 11: Seemingly unrelated regressions    X1  Y1   (T ×k )   (T ×1)     X2 Y  =  (T ×k ) ... Australia University of Economics - HCMC - Vietnam Advanced Econometrics βˆOLS Chapter 11: Seemingly unrelated regressions  βˆ1( OLS )   ( k ×1)    β =  2( OLS )        β M ( OLS... Australia University of Economics - HCMC - Vietnam Advanced Econometrics Where Chapter 11: Seemingly unrelated regressions is consistently estimated by = ϕˆ [ X '(Σˆ ⊗ I ) −1 X ]−1 III KRONECKER

Ngày đăng: 09/12/2017, 08:39

w