You are on page 1of 26

Multi-Equations Econometric

Models
-- Nasrudin --
Sekolah Tinggi Ilmu Statistik, Jakarta - 2015
Apa hubungannya, bagaimana caranya?
Bagaimana pendapatmu, jika:

Yt= 0 + 1. + 2. + 3. +t

Multicollinearity, karena:
I = f(r)

Bagaimana seharusnya?
Multi-equations model:
(1) Yt= 0 + 1. + 2. +1t
recursive

(2) It= 3 + 4. +2t


atau
(1) Yt= 0 + 1. + 2. +1t
simultan

(2) It= 3 + 4. + 5. 2t
EKONOMETRIKA (Pasca-UTS)
Materi Sifat Kegiatan Penugasan Evaluasi Program
UTS
9 Introduction to Multi-Equations Econometrics Model: Model
Specification (Recursive, SUR, Simultaneous Equation, VAR) Core Course Econometric Models

10 Simultaneous Equations: Identification Problems, Test of Individual paper


Simultaneity and Reduce Form Core Course Kuis-4 Eviews
(draft)
11 Simultaneous Equations: Simultaneous Equation Method
(Indirect Least Square, Two Stages Least Squares) and Its Mathematical
Core Course SAS
background
Intepretation
12 Simultaneous Equations: Validation and Simulation Output
Additional Course Kuis-5 SAS
Intepretation
13 Application: 'Skripsi' and 'Thesis' Simultaneous Equations
(Presentation) Core Presentation Group Presentation

14 Application: 'International Journal' of Simultaneous


Equations (Presentation) Core Presentation Group Presentation

15 Simultaneous Equations: Dynamic Simultaneous Equations Individual paper


and Intro to Three Stages Least Squares Additional Course Kuis-6 SAS
(final)
16 Multi-Equation Econometrics Time Series Models: VAR
Analysis Optional Course Eviews

UAS
Multivariate Bias
Bias secara teori ekonomi
Bias karena melanggar properties model
statistik
Variables in the multi-equations
econometric model

Endogenous variable
Exogenous variable
Lagged endogenous variable
Lagged exogenous variable
Pre-determined variables
Explanatory variables
Not use independent variable, because its not exactly independent
Multi-equations econometric model:
1. Recursive model
2. Seemingly Unrelated Model:
3. Simultaneous Equation: contain the
simultaneous problems
4. Time series multiequations: i.e Vector
Autoregressive Model
(1) Recursive Model: straight forward,
triangular, no simultaneity problems

(zero contemporaneous correlation)


FIRST EQUATION:
it contains only the exogenous variables on the right-hand side
Exogenous variables are uncorrelated with the disturbance term u1t,
satisfies the critical assumption of the classical OLS,
Hence, OLS can be applied straightforwardly to this equation.
SECOND EQUATION:
contains the endogenous variable Y1 as an explanatory variable along with
the nonstochastic Xs.
Y1t and u2t are uncorrelated.
Hence, one can proceed with OLS estimation of this equation.
THIRD EQUATION
Carrying this argument a step further, we can also apply OLS to the third
equation because both Y1 and Y2 are uncorrelated with u3.

Thus, in
the recursive system OLS can be applied to each
equation separately. Actually, we do not have a
simultaneous-equation problem in this situation.
Jika u1t berkorelasi dengan u2t, apakah
OLS bisa digunakan?
(2) Seemingly Unrelated (SUR) Model:
seemingly unrelated among equations, but related

Y1t= 01 + 11. + 21. +1t


Y2t= 02 + 12. + 22. +2t

Ynt= 0 + 1. + 2. +nt

Theyre correlated via error term, because of unobservable effect.


When the error terms of two or more equations are correlated, OLS is not
efficient.
In the panel model: SUR (correlated among observation/individual)
In the SUR model: correlated occurred among equation, its mean that there
are correlated among endogenous variables
Inter-correlated is not a simultaneity
Correlation: strong/weak a relation
Simultaneity: co-causality
(3) Simultaneous Equation
Reason:
Base on economic theory
Test of simultaneity: endogenity test and
exogenity test
Simultaneous bias problem
Equation types:
1. Structural or behavioral equation:
portray the structure (of an economic model) of an
economy or thebehavior of an economic agent
Statistically, stochastic
Ex: Consumption (C) = Co + .INCOME(Y), the coef
estimated by statistical method
2. Identity equation:
Aggregation, or other mathematical operation
among variables.
Mathematically, deterministic
Ex: Income (Y) = C + I + G + NX, its determined by
mathematical calculation.
Keynesian Model (Gujarati & Porter, 2008)
Langkah
Kerangka teori: alur diagram untuk
mempermudah
Spesifikasi model
Identifikasi: just/exact, under, over identified
Estimasi: OLS, ILS, 2SLS, 3SLS, FIML
Intepretasi
Validasi (jika untuk peramalan)
Simulasi (jika untuk peramalan)
Next: 10th 15th session, Applied Econometrics
Multi-Equations in Econometric Time Series:
Vector Auto Regressive (VAR)
Example: Gujarati & Porter (2008)
Data

Unit Root Test

Non Stationer
Stationer

Cointegration Not Cointegration


VAR

VECM VAR in the first


Impulse Respond difference
Funstion Transfer
Variance decomposition

Next: 16th session, Applied Econometrics


Fifteen minutes
Buat contoh model yang menggunakan
keempat bentuk model yg telah dipaparkan
Nomor absen 1, 5, 9, 13, . Recursive model
Nomor absen 2, 6, 10, 14, . SUR model
Nomor absen 3, 7, 11, 15, . Model Simultan
Nomor absen 4, 8, 12, 16, . Model Simultan
EVALUASI UTS
Tidak memperhatikan soal, atau hapalan
rumus?
Penulisan estimasi model
Pengujian dan kesimpulan dari pengujian
Cara baca koefisien regresi
Cara baca koefisien fixed/random effect
Lembar jawaban bisa diambil masing-
masing di Ruang Jurusan Statistika, kapan
saja ketika saya ada di tempat.
Rata-rata 81.7
Std Dev 6.3
Coef. Var (%) 7.7
Max 94.0
Min 68.0
The best performance: