Scribd
Upload
220 views

Introduction To Econometrics: Wk14. Simultaneous Equations and IV Technique

This document discusses simultaneous equations and instrumental variable (IV) techniques. It explains that estimating structural equations using ordinary least squares (OLS) can result in biased coefficients due to endogeneity. The document introduces indirect least squares and two-stage least squares as IV methods to consistently estimate structural equations. These methods work by first estimating a reduced form model using only exogenous variables, then calculating the structural coefficients from these estimates. Identification requires that the number of excluded exogenous variables be at least as large as the number of endogenous regressors. A valid IV must be correlated with the endogenous variable but not with the error term.

Uploaded by

Nuur Ahmed
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
220 views

Introduction To Econometrics: Wk14. Simultaneous Equations and IV Technique

This document discusses simultaneous equations and instrumental variable (IV) techniques. It explains that estimating structural equations using ordinary least squares (OLS) can result in biased coefficients due to endogeneity. The document introduces indirect least squares and two-stage least squares as IV methods to consistently estimate structural equations. These methods work by first estimating a reduced form model using only exogenous variables, then calculating the structural coefficients from these estimates. Identification requires that the number of excluded exogenous variables be at least as large as the number of endogenous regressors. A valid IV must be correlated with the endogenous variable but not with the error term.

Uploaded by

Nuur Ahmed
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 10

Introduction to Econometrics

Fall 2009
Wk14. Simultaneous Equations
and IV Technique

2
Structural and reduced-form
equation
Structural form
C =
0
+
1
Y+u
Y = C + I

Reduced form [endogenous variables
as a function of exogenous variables]
C = (
0
/1-
1
)+(
1
/1-
1
)I+(1/1-
1
)u
Y = (
0
/1-
1
)+(1/1-
1
)I+(1/1-
1
)u
3
Simultaneous equation bias
If you apply OLS on the structural
equations,
1
will be biased.
This is because there is a correlation
between the endogenous variable on
the RHS and the error of that equation.
i.e. Cov (Y,u)0
4
Indirect least squares
Estimate the reduced form model and
find the parameters of the structural
equation
Example from agriculture:
Demand: Q =
0
+
1
P +
2
Y + u
Supply: Q =
0
+
1
P +
2
R + v
where: Y is [exogenous] income
R is [exogenous] rainfall
5
Reduced form is:

P = (
0
-
0
)/(
1
-
1
) -
2
/(
1
-
1
)Y +
2
/(
1
-
1
)R + e
1

=
0
+
1
Y+
2
R+e
1

Q = (
0
+
1

0
) + (
1

1
+
2
)Y +
1

2
R + e
2
=
0
+
1
Y +
2
R +e
2

We can now run OLS to get estimates of
and .
6
Hence:
2
2
1

o =
1 1 1 2


o o =
|
|
.
|

\
|
+ =
1
2
1 1

o
o |
0 1 0 0


o o =
)

(

1 1 2 2
| o | =
)

1 1 0 0 0
| o o | + =
All structural parameters can be estimated
unambiguously.
7
Identification (Order condition)
X = number of exogenous variables in
the system of equations that are
excluded from a particular equation
N = number of endogenous variables on
the RHS of a particular equation
Equation not identified if X<N
Equation exactly identified if X=N
Equation overidentified if X>N
8
IV Estimator
Y =
0
+
1
X+u
Variables correlated with the error term
are called endogenous variables
Variables uncorrelated with the error
term are called exogenous variables
Instrumental variables technique is a
general single equation estimation
method useful to tackle endogeneity
issues.
9
Two conditions a valid instrument
What we need is a good instrument, Z.
Instrument relevance: corr(Z,X)0
Instrument exogeneity: corr(Z,u)=0
Then we do the following two-stage
least squares (2SLS)
10
2SLS
Stage 1: Regress X on instrumental variables
Z and included exogenous variables W using
OLS. Computed the predicted values for the
regression .
Stage 2: Regress Y on the predicted values
on the endogenous variables and
included exogenous variables W using OLS.

X

You might also like