CHAPTER TWO

INTRODUCTION TO SIMULTANEOUS
EQUATION MODELS

2.1 Nature of Simultaneous Equation models

2.2 Simultaneity Bias & Inconsistency of OLS estimators
2.3 Identification problem
2.4 Formal Rules (Conditions) for Identification
2.5 Estimation of Simultaneous Equations Models1
 2.1 Nature of Simultaneous Equation models ----

 So far we focused on the cause-and-effect relationship in

single equation models between the dependent and
independent variable, i.e., unidirectional relations.
 That is, the explanatory variables are the cause and the
independent variable is the effect.
 But there are situations where such one-way or
unidirectional causation in the function is not meaningful.

2
 2.1 Nature of Simultaneous Equation models ----

 This occurs when Y (dependent variable) is not only a function of X’s (explanatory
variables) but also all or some of the X’s are, in turn, determined by Y.

 There is, therefore, a two-way flow of influence between Y and (some of) the X’s
which in turn makes the distinction between dependent and independent variables
a little doubtful.
 In the simultaneous model there is more than one equation –one for each of the
mutually, or jointly, dependent or endogenous variables.
 The number of equations in such models is equal to the number of jointly
dependent or endogenous variables involved in the phenomenon under analysis.
 2.1 Nature of Simultaneous Equation models ----

➢ Thus, in simultaneous equation models, it is not usually possible to

estimate a single equation of the model without taking into account
the information provided by other equations of the system.

➢ If one applies OLS to estimate the parameters of each equation

disregarding other equations of the model, the estimates so obtained
are not only biased but also inconsistent, i.e. even if the sample size
increases indefinitely, the estimators do not converge to their true
values.
4
 2.1 Nature of Simultaneous Equation Models ----

 Example: the classic example of simultaneous causality in economics is

supply and demand.
 Both Prices and quantities adjust until supply and demand are in
equilibrium.
 A shock of demand or supply causes both prices and quantities to move.
 As well known, the prices P of a commodity and quantity Q sold are
determined by the intersection of the demand and supply curves for that
commodity.

5
 2.1.1 Model specification

 Thus, assuming for simplicity that the demand and supply

curves are linear and adding the stochastic disturbance term
U1 and U2, we may write the empirical dd and ss function as:
Demand function :
Qt =  0 + 1 Pt +  2Yt + U1t − − − 1  0
d

Supply function :
Qt =  0 + 1 Pt +  2Yt + U 2t − − − 1  0
s

Equilibrium condition: Qt = Qt
s d

 Where Qtd= quantity demand

 Qts=quantity supplied
 t=time;  0 , and 1 are the parameters
 2.1.1 Model specification ----

 B/C of simultaneous dependence between Q and P, then U1t and Pt, and U2t and
pt cannot be independent.
 If U1t in the above equation changes b/c changes in other variables affecting Qtd
such as income, wealth, and tastes), the demand curve will shift upward if U1t is
+ve and downward if U1t is –ve.
 Thus, a shift in the demand curve changes both P and Q.
 Similarly, a change in U2t b/c of weather, import or export restrictions, etc; will
shift the ss curves, again affecting both P and Q.
 B/c of this simultaneous dependence b/n Q and P, U1t and Pt and U2t and pt
cannot be independent.
 Thus, a regression of Q and P as in above equation would violate an important
assumptions of the classical linear regression model; namely the assumption of no
correlation b/n the explanatory variable(s) and the disturbance term. 7
 ,

2.1.1 Definitions of Some Concepts

In simultaneous equation models variables are classified as endogenous

and exogenous.
 Endogenous variables: are variables that are determined by the
economic model (within the system) and
Exogenous variables: are those determined from outside of the system.
Exogenous variables are also called predetermined. Since the exogenous
variables are predetermined, they are supposed to be independent of the
error terms in the model/ non stochastic.
Exogenous variables includes:
contemporaneous variables (, lagged exogenous variables and lagged
endogenous variables. Predetermined variables are non-stochastic and
hence independent of the disturbance terms. X t , X t −1 and Yt −1 are regarded as
predetermined (exogenous) variables.

8
 2.1.1 Definitions of Some Concepts -----

 Structural models: A structural model describes the complete

structure of the relationships among the economic variables.

 Structural equations of the model =endogenous variables+ exogenous

variables + disturbances (random variables).

 The parameters of structural model express the direct effect of each

explanatory variable on the dependent variable.

 The variables not appearing in any function explicitly may have an

indirect effect and is taken into account by the simultaneous solution
of the system.
 2.1.1 Definitions of Some Concepts

 Reduced form of the model: The reduced form of a structural model is

the model in which the endogenous variables are expressed as a
function of the predetermined variables and the error term only.
 Example: The following simple Keynesian model of income
determination can be considered as a structural model.
=--- ---------(1)
----------(2)

 where: C=consumption expenditure; Z=non-consumption expenditure;

Y=national income; C and Y are endogenous variables while Z is an
exogenous variable.
 Find the reduced form of the above structural model. Since C and Y are
endogenous variables and only Z is the exogenous variable, we have to
express C and Y in terms of Z. 10
 2.1.1 Definitions of Some Concepts

To do this substitute Y=C+Z into equation (1).

-------(3
Substituting again (3) into (2) we get;
  1  U
Y= +  Z +
1−  1−   1−  ----------(4)
 Equations (3) and (4) are called the reduced form of the structural
model of the above. We can write this more formally as:
 2.1.1 Definitions of Some Concepts

Parameters of the reduced form measure the total effect (direct and
indirect) of a change in exogenous variables on the endogenous
  
variable. For instance,  1 −   in the above reduced form equation(1),
measures the total effect of a unit change in the non-consumption
expenditure on consumption. This total effect is 1 , the direct effect,
1−  
times, the indirect effect.   12
 2.1.2 Inconsistency and Simultaneity Bias of OLS Estimators

 Biasedness:
▪ The two-way causation in a relationship leads to a violation of the important
assumption of a linear regression model, i.e. one variable can be the
dependent variable in one of the equations but becomes also an explanatory
variable in the other equations of the simultaneous-equation model.
▪ In this case E[XiUi] may be different from zero. To show simultaneity bias,
let’s consider the following simple simultaneous equation model.
Y =  0 + 1 X + U 

X =  0 + 1Y +  2 Z + V 
X = f (Y )
▪ Y = f ( X ) this
shows that the 2-way causation in a relationship leads to
violations of the important assumptions of linear regression model
13
 2.1.2 Inconsistency and Simultaneity Bias of OLS Estimators

 Suppose that the following assumptions hold,

(U ) = 0 , (V ) = 0
(U 2 ) =  u2 , (V 2 ) =  u2
(U iU j ) = 0 , (ViV j ) = 0, also (UiVi) = 0;

where X and Y are endogenous variables and Z is an

exogenous variable.
The reduced form of X of the above model is obtained by
substituting Y in the equation of X.
X =  0 +  1 ( 0 +  1 X + U ) +  2 Z + V
 0 +  0 1  2    1U + V 
X = +
1−   
 Z + 
 1−  

1 −  1 1    
14 1 1 1 1
 2.1.2 Inconsistency and Simultaneity Bias of OLS Estimators

 Applying OLS to the first equation of the above structural model

will result in biased estimator. Now, let’s proof whether this
expression.
cov( X iU i ) = ( X iU j )  0
 1  1 u2
=  (U ) =
2
0
 1 −  
1 1  1 −  
1 1

 That is, covariance between X and U is not zero. As a consequence, if

OLS is applied to each equation of the model separately the coefficients
will turn out to be biased. Now, let’s examine how the non-zero co-
variance of the error term and the explanatory variable will lead to
biasness in OLS estimates of the parameters. 15
 2.1.2 Inconsistency and Simultaneity Bias of OLS Estimators

 To show that the OLS estimator 𝜷 ෡ 𝟏 is biased and

inconsistent estimator of 𝜷𝟏 because of correlation
between Yt and ut in the consumption equation, we
proceed as follows:
ˆ =
 (C − C )(Y − Y )
t t

 (Y −Y )
1 2
t

=
c yt t
 ( +  1 Y t + ut ) yt
ˆ =
0

y
2

 yt
1 2
t

=
C y t t
= 1 +
y u t t

y y
2 2
t t

16
 2.1.2 Inconsistency and Simultaneity Bias of OLS Estimators

 Taking expectation of both sides, we have:

 y u 
E ( ˆ ) =  + E  
t t

 y 
1 1 2
  t

 An estimator is said to be consistent if its probability limit (plim) is

equal its true population value.
  y ut 
p lim (ˆ ) = p lim ( 1) + p lim  
t

  yt 
n →  
1 2
n → n →

  y ut n 
=  1 + p lim  
t

  yt n 
n →  
2

p lim ( y t u t n )
=  1 + n →
p lim  y t n
2

n →
( )
 2 (1 −  1)
= 1 +
17  Y2
 2.2. Identification Problem

 In simultaneous equations model, identification problem is to mean whether

numerical estimates of the parameters of the structural equations can be obtained
from the estimated reduced-form coefficients.
 If this can be done, we say that the particular equation is identified. If this cannot
be done, then we say that the equation under consideration is not identified, or
under identified.
 Specifically,
 An equation is said to be unidentified/not identified iff we cannot get the structural
coefficients from the reduced form estimates.
 An equation is said to be exactly identified iff we can get unique structural/behavioural form
coefficient estimates from the reduced form estimates.
 An equation is said to be over-identified iff more than one set of structural coefficients
could be obtained from the reduced form estimates.
 2.2 Identification Problem

 In the reduced equation, the endogenous variable is expressed as a function of

the exogenous variable.
 OLS method can be applied to estimate the coefficients of the reduced-form
equations.
 From the estimated reduced-form coefficients one may estimate the structural
coefficients.
 This procedure is known as indirect least squares (ILS), and the estimated
structural coefficients are called ILS estimates.
 The identification concept concerns whether the numerical estimates of
structural equations can be obtained from the estimated reduced form
coefficients.
 An identification may be either exactly (fully or just identified) or over-
identified or under-identification.
 2.2. Identification Problem

A. Under identification (SEP>REP)

 It occurs when the parameters of structural equation is higher than
reduced form parameters.
 If the coefficients of the structural equations are greater than the
coefficients of the reduced form, then we can say that the equation is under-
identified.
 Let us consider the demand and supply model with the market clearing
condition: 𝑸𝒅 = 𝜶𝟎 + 𝜶𝟏 𝑷𝒕 + 𝒖𝟏𝒕 , 𝑸𝒔 = 𝜷𝟎 + 𝜷𝟏 𝑷𝒕 + 𝒖𝟐𝒕
 At equilibrium, 𝑸𝒅 = 𝑸𝒔
⇒𝜶𝟎 + 𝜶𝟏 𝑷𝒕 + 𝒖𝟏𝒕 = 𝜷𝟎 + 𝜷𝟏 𝑷𝒕 + 𝒖𝟐𝒕
 2.2 Identification Problem

⇒𝜶𝟎 + 𝜶𝟏 𝑷𝒕 + 𝒖𝟏𝒕 = 𝜷𝟎 + 𝜷𝟏 𝑷𝒕 + 𝒖𝟐𝒕

𝜶𝟏 𝑷𝒕 − 𝜷𝟏 𝑷𝒕 = 𝜷𝟎 − 𝜶𝟎 + +𝒖𝟐𝒕 − 𝒖𝟏𝒕
 𝑷𝒕 𝜶𝟏 − 𝜷𝟏 = 𝜷𝟎 − 𝜶𝟎 + 𝒖𝟐𝒕 − 𝒖𝟏𝒕
𝜷𝟎 −𝜶𝟎 𝒖𝟐𝒕 −𝒖𝟏𝒕
𝑷𝒕 = + --------------------(1)
𝜶𝟏 −𝜷𝟏 𝜶𝟏 −𝜷𝟏

𝑷𝒕 = 𝝅𝟎 + 𝝃𝟎 ---------------------------- (2)
𝜷𝟎 −𝜶𝟎
Where 𝝅𝟎 =
𝜶𝟏 −𝜷𝟏

𝒖𝟐𝒕 − 𝒖𝟏𝒕
𝝃𝟎 =
𝜶𝟏 − 𝜷𝟏
 2.2 Identification Problem

Now, substitute equation (1) in to demand equation above:

𝒅
𝜷𝟎 − 𝜶𝟎 𝒖𝟐𝒕 − 𝒖𝟏𝒕
𝑸 = 𝜶𝟎 + 𝜶𝟏 ( + ) + 𝒖𝟏𝒕
𝜶𝟏 − 𝜷𝟏 𝜶𝟏 − 𝜷𝟏
𝜶𝟏 𝜷𝟎 −𝜶𝟏 𝜶𝟎 𝜶𝟏 𝒖𝟐𝒕 −𝜶𝟏 𝒖𝟏𝒕
⇒ 𝑸𝒅 = 𝜶𝟎 + + + 𝒖𝟏𝒕
𝜶𝟏 −𝜷𝟏 𝜶𝟏 −𝜷𝟏
𝜶𝟎 𝜶𝟏 − 𝜶𝟎 𝜷𝟏 𝜶𝟏 𝜷𝟎 − 𝜶𝟏 𝜶𝟎 𝜶𝟏 𝒖𝟐𝒕 − 𝜶𝟏 𝒖𝟏𝒕 + 𝜶𝟏 𝒖𝟏𝒕 − 𝜷𝟏 𝒖𝟏𝒕
𝑸𝒅 = + +
𝜶𝟏 − 𝜷𝟏 𝜶𝟏 − 𝜷𝟏 𝜶𝟏 − 𝜷𝟏

𝜶𝟏 𝜷𝟎 −𝜶𝟎 𝜷𝟏 𝜶 𝒖 −𝜷 𝒖
𝑸𝒅 = + 𝟏 𝟐𝒕 𝟏 𝟏𝒕 −−−−−− −(𝟑)
𝜶𝟏 −𝜷𝟏 𝜶𝟏 −𝜷𝟏
𝜶𝟏 𝜷𝟎 −𝜶𝟎 𝜷𝟏 𝜶𝟏 𝒖𝟐𝒕 −𝜷𝟏 𝒖𝟏𝒕
Let 𝝅𝟏 = , 𝝃𝟏 =
𝜶𝟏 −𝜷𝟏 𝜶𝟏 −𝜷𝟏
⇒𝑸𝒅 = 𝝅𝟏 + 𝝃𝟏 −−−−−−−−−−−−−−−− −(𝟒)
 Under identification (SEP>REP)
▪ Equations 2 and 4 are the two reduced-form equations derived from the structural
equations.
▪ Now our demand-and-supply model contains four structural coefficients (α0, α1,
β0, and β1) where as the reduced equations have only two coefficients (π0 and π1).
▪ The coefficients of reduced form contain the coefficients of the structural
equations i.e α0, α1, β1, and β2 are found in π0 and π1.
▪ But how we can find the values of α0, α1, β1 and β2 from π0 and π1.
▪ There is no way in which the four structural unknowns can be estimated from only
two reduced-form coefficients.
▪ Since it is not possible to find these values from π0 and π1 or the coefficients of the
structural equations are greater than the coefficients of the reduced form then we
can say that the equation is under-identified and we can not compute four
structured coefficients from two reduced coefficients.
 Why under identification is happened?
▪ The reason to have under identified function in the previous
demand and supply function was that:
▪ The same variables P and Q are appearing in both functions
(only endogenous variables in both equation)
▪ There is no additional information.
 Exact /Just/ Identification (SEP=REP)
▪ It occurs when structural coefficients are equal to reduced form
coefficients.
▪ Now let’s incorporate additional variables in the demand function in order
to solve the above problem.
5
6
7
 Exact /Just/ Identification (SEP=REP)

9
9 5
 Exact /Just/ Identification (SEP=REP)

10
9 10
5
6

2
 Exact /Just/6 Identification (SEP=REP)

10 9

11
8

10

2
 Exact /Just/ Identification (SEP=REP)
▪ But in case of the demand function α0, α1, and α2, is 3 structural
coefficients but in reduced form of equation the coefficients are two.
▪ Since in the demand function the coefficient of the reduced form (10)
is less than the coefficients of the structural equation (5).
▪ We can concluded that the demand function is under identified since
(π2,π3) are less than α0,α1,and α2).
▪ But in case of supply function π2,π3 are equal to β0 , β1 then it is just
identified.
▪ In conclusion, we can say that the supply function is identified but the
demand function is not identified on the basis of this one can say that
the system as a whole is not identified.
 Over identification (SEP<REP)
▪ It occurs when the coefficients (parameters) of structural
equation is less than the coefficients (parameters) of reduced
forms.
▪ Let’s modify the demand function by incorporating wealth (R)
and supply function by incorporating the lagged price we will
have the following equation.
 12
Over identification (SEP<REP) 13

14
 Over identification (SEP<REP)
𝑷𝒕 = 𝝅𝟎 + 𝝅𝟏 𝑰𝒕 + 𝝅𝟐 𝑹𝒕 + 𝝅𝟑 𝑷𝒕−𝟏 + 𝒗𝒕 --------- (14)
𝑸𝒕 = 𝝅𝟒 + 𝝅𝟓 𝑰𝒕 + 𝝅𝟔 𝑹𝒕 + 𝝅𝟕 𝑷𝒕−𝟏 + 𝒘𝒕 -------- (15)
Where
 Over identification (SEP<REP)
▪ From equation number (12) and (13) we have seven structural coefficients
but in equation (14) and (15) we have eight reduced form coefficients.
▪ Since the coefficients of reduced form coefficients are greater than the
structural coefficients we can say that the system as a whole is over
identified.
▪ A function (an equation) belonging to a system of simultaneous equations
is identified if it has a unique statistical form, i.e. if there is no other
equation in the system, or formed by algebraic manipulations of the other
equations of the system, contains the same variables as the
function(equation) in question.
 2.2.1 Formal Rules (Conditions) for Identification

▪ Identification problems do not just arise only in two

equation-models.
▪ Using the above procedure, we can check identification
problems easily if we have two or three equations in a given
simultaneous equation model.
▪ However, for ‘n’ simultaneous equation model, such a
procedure is very cumbersome.
▪ In general, for any number of equations in a given
simultaneous equation, we have two conditions that need to
be satisfied to say that the model is in general identified or
not.
 2.2.1 Formal Rules (Conditions) for Identification
▪ There are two conditions which must be fulfilled for an equation
to be identified. These are:
▪ 1. the order condition for identification
▪ 2. the rank condition for identification
▪ The identification of a system means the identification of each
question.
▪ The parameters identification in any equations means there is
unique value for each parameter in equations.
 A. The order condition for identification
Let, G = total number of equations (= total number of
endogenous variables)
K= number of total variables in the model (endogenous and
predetermined)
M= number of variables, endogenous and exogenous, included in
a particular equation/ in a specific equation.
 A. The order condition for identification
▪ Then the order condition for identification may be symbolically
expressed as:

▪ The guidelines is that:

▪ If (K-M)> (G-1); the equation is identified.
▪ If (K-M)= (G-1); the equation is just/exactly identified.
▪ If (K-M)< (G-1); the equation is under identified.
▪ If (K-M)>(G-1); the equation is over identified.
 A. The order condition for identification
Example 1:
Qd = α + α1 P1 + α2 I + U 1 − − − −( 1 )
Qs = β0 + β1 P1 + U 2 − − − − − − − ( 2)
• Take the dd equation
• G= total number of equations/ total number of endogenous variables=2
• K=total number of exogenous and endogenous variables in the model =3
• M= total number of exogenous and endogenous variables in equation (1), i.e.,
in demand equation=3
• The solution is that: (K-M)________(G-1)
• (3-3)____(2-1)=0<1, we conclude that the demand equation is under
identified.
 A. The order condition for identification
Take the ss equation
▪ Given: G=2; K=3; M=2;
▪ Solution:
▪ K-M-------------G-1
▪ (3-2)--------------(2-1)
▪ 1=1→ from these we can conclude that the supply function is
exactly identified.
 A. The order condition for identification
Example 2: Given the structural model and determine whether the
equation are identified or under identified.
y1 = 3 y 2 − 2 x1 + x2 + u1 − − − −1
y 2 = y3 + x3 + u 2 − − − − − − − −2
y3 = y1 − y 2 − 2 x3 + u3 − − − − − 3
• Take equation (1);
• Given; M (endogenous and exogenous variables) in this specified equation
is 4 (y1, y2, x1 and x2); K=6; G=3;
• (K-M)----------(G-1)
• 6-4-------------(3-1)
• 2=2--→ this equation is identified and it is exactly identified.
 A. The order condition for identification
• Take equation (2);
• Given; M (endogenous and exogenous variables) in this specified
equation is 3 (y2, y3, & x3); K=6; G=3;
• (K-M)----------(G-1)
• 6-3-------------(3-1)
• 3>2--→ this equation is identified and it is over identified.
 A. The order condition for identification
• Take equation (3);
• Given; M (endogenous and exogenous variables) in this specified
equation is 4 (y3, y1, y2 and x3); K=6; G=3;
• (K-M)----------(G-1)
• 6-4-------------(3-1)
• 2=2--→ this equation is identified and it is exactly identified.
▪ Example 3: if a system contains 10 equations with 15 variables,
ten endogenous and five exogenous, an equation containing 11
variables is not identified, while another containing 5 variables is
identified.
 A. The order condition for identification
▪ For 1st equation we have:
G=10; K=15; M=11;
Order condition:
 K-M> G-1
 15-11> 10-1
 4<9→ that is the order condition is not satisfied.
▪ For the 2nd equation we have:
▪ G=10; K=15; M=5
▪ Order condition:
▪ (K-M)> (G-1); 10>9-----the order conditions is satisfied.
 B. The rank condition for identification
▪ The rank condition states that: in a system of G equations any
particular equation is identified if and only if it is possible to
construct at least one nonzero determinant of order (G-1) from
the coefficients of the variables excluded from that particular
equation but contained in the other equations of the model.
▪ The practical steps for tracing the identifiablity of an equation of
a structural model may be outlined as follows.
▪ Firstly, write the parameters of all the equations of the model in a
separate table, noting that the parameter of a variable excluded
from an equation is equal to zero.
 B. The rank condition for identification

▪ Where y’s are the endogenous variables and

 x’s are the exogenous variables

▪ Ignoring the random disturbance the table of the parameters of the

model is as follows:
 B. The rank condition for identification

Secondly, Strike out the row of coefficients of the equation which is

being examined for identification. For example, if we want to
examine the identifiability of the second equation of the model we
strike out the second row of the table of coefficients.
 B. The rank condition for identification
▪ Thirdly, Strike out the columns in which a non-zero coefficient of the equation being examined
appears.
▪ Table of structural parameter
Equatio Y1 Y2 Y3 X1 X2 X3
ons

1st equ. -1 3 0 -2 1 0

2nd equ. 0 1 1 0 0 1

3rd equ. 1 1 1 0 0 2

▪ By deleting the relevant row and columns we are left with the coefficients of variables not
included in the particular equation, but contained in the other equations of the model.
▪ For example, if we are examining for identification the second equation of the system, we will
strike out the second, third and the sixth columns of the above table, thus obtaining the
following tables.
B. The rank condition for identification
 B. The rank condition for identification
▪ Fourthly, form the determinant(s) of order (G-1) and examine their value.
▪ Guide line:
▪ If at least one of these determinants is non-zero, the equation is
identified.
▪ If all the determinants of order (G-1) are zero, the equation is under
identified.
▪ In the above example of exploration of the identifiability of the second
structural equation we have three determinants of order (G-1)=3-1=2.
They are:
B. The rank condition for identification
 B. The rank condition for identification
▪ The identification of a function is achieved by assuming that some
variables of the model have zero coefficient in this equation, that is, we
assume that some variables do not directly affect the dependent variable in
this equation.
▪ This, however, is an assumption which can be tested with the sample data.
▪ We will examine some tests of identifying restrictions in a subsequent
section.
▪ Some examples will illustrate the application of the two formal conditions
for identification.
 2.3 Test of Simultaneity problem/Endogeneity
If there is no simultaneous equation or simultaneity problem, the OLS estimators
produce consistent and efficient estimators.
On the other hand, if there is simultaneity, OLS estimators are not even consistent
(instead we use 2SLS & IVs).
This discussion suggests that we should check for the simultaneity problem before we
discard OLS in favor of the alternatives.
the simultaneity problem arises because some of the regressors are endogenous and
are therefore likely to be correlated with the disturbance, or error, term.
Therefore, a test of simultaneity is essentially a test of whether (an endogenous)
regressor is correlated with the error term.
 If it is, the simultaneity problem exists, in which case alternatives to OLS must be
found; if it is not, we can use OLS.
 2.3 Hausman Specification Test for Simultaneity/Endogeneity
To see how the test work, let us consider the two equations:

Assume that I and R are exogenous. Of course, P and Q are endogenous.

Now consider the supply function. If there is no simultaneity problem (i.e., P and
Q are mutually independent), Pt and u2t should be uncorrelated.
On the other hand, if there is simultaneity, Pt and u2t will be correlated.
To find out which is the case, the Hausman test proceeds as follows:
 2.3 Hausman Specification Test for Simultaneity
First, from the demand and supply equations, we get the
following reduced-form equations:

Estimating reduced form price using OLS, we have:

Therefore,
 2.3 Hausman Specification Test for Simultaneity
Now, if the null hypothesis is that there is no simultaneity, that is, Pt is not an
endogenous variable, the correlation between 𝒗 ෝ𝒕 and 𝒖𝟐𝒕 should be zero,
asymptotically.
Thus, if we run the OLS regression on the last equation and find that the
coefficient of 𝒗 ෝ𝒕 is statistically zero, we can conclude that there is no
simultaneity problem.
If it is, we have the simultaneity problem (𝑷𝒕 is endogenous).
Of course, this conclusion will be reversed if we find this coefficient to be
statistically significant.
In passing, note that Hausman’s simultaneity test is also known as the Hausman
test of endogeneity:
 2.4. Test for Exogeneity
Suppose we have a three-equation model in three endogenous variables,
𝒀𝟏 , 𝒀𝟐 𝒂𝒏𝒅 𝒀𝟑 and suppose there are three exogenous variables,
𝑿𝟏 , 𝑿𝟐 𝒂𝒏𝒅 𝑿𝟑 .
Further, suppose that the first equation of the model is;
𝒀𝟏𝒊 = 𝜷𝟎 + 𝜷𝟏 𝒀𝟐𝒊 + 𝜷𝟐 𝒀𝟑𝒊 + 𝜶𝟏 𝑿𝟏𝒊 + 𝒖𝟏𝒊
If Y2 and Y3 are truly endogenous, we cannot estimate this equation by
OLS.
But how do we find that out? We can proceed as follows.
First, obtain the reduced-form equations for Y2 and Y3.
 2.4 Test for Exogeneity
Second, from these reduced-form equations, we obtain 𝒀 ෡ 𝟐𝒊 𝒂𝒏𝒅 𝒀෡ 𝟑𝒊 .
Then, in the sprit of the Hausman test discussed earlier, estimate using the
OLS:
𝒀𝟏𝒊 = 𝜷𝟎 + 𝜷𝟏 𝒀𝟐𝒊 + 𝜷𝟐 𝒀𝟑𝒊 + 𝜶𝟏 𝑿𝟏𝒊 + 𝜹𝟐 𝒀෡ 𝟐𝒊 + 𝜹𝟑 𝒀
෡ 𝟑𝒊 + 𝒖𝟏𝒊
Using the F test, we test the hypothesis that 𝜹𝟐 = 𝜹𝟑 =0.
If this hypothesis is rejected, Y2 and Y3 can be deemed endogenous, but if
it is not rejected, they can be treated as exogenous.
 2. 5. Estimation of Simultaneous Equations Models

▪ The major applied in the estimation of simultaneous equation

methods are:
 1. Ordinary least squares (OLS)
 2. Indirect least squares (ILS)

 3. Two stage least squares (2SLS)

▪ As we saw in the previous sessions, using OLS to estimate

structural parameters in the presence of simultaneity is a risky
move.
 2.5. Estimation of Simultaneous Equations Models

2. Indirect least squares (ILS)

▪ It is the method of obtaining the estimates of the structural
coefficients from the OLS estimates of the reduced-form
coefficients.
▪ It is applied for exactly identified structural equations.
▪ ILS involves the following three steps:
1. Obtain the reduced-form equations.
2. Apply OLS to the reduced-form equations individually.
3. Obtain estimates of the original structural coefficients from the estimated
reduced-form coefficients obtained in Step 2.
 2.5 Estimation of Simultaneous Equations Models

2. Indirect least squares (ILS) -----

Example: Consider the demand-and-supply model introduced in the previous
session:

 The reduced-form equations corresponding to the preceding structural equations

are:

𝜷𝟎 −𝜶𝟎 −𝜶𝟐 𝜷𝟎 𝜶𝟏 −𝜶𝟎 𝜷𝟏 −𝜶𝟐 𝜷𝟏

Where 𝝅𝟎 = ; 𝝅𝟏 = ; 𝝅𝟐 = ; 𝝅𝟑 = ;
𝜶𝟏 −𝜷𝟏 𝜶𝟏 −𝜷𝟏 𝜶𝟏 −𝜷𝟏 𝜶𝟏 −𝜷𝟏
𝒖𝟐𝒕 −𝒖𝟏𝒕 𝜶𝟏 𝒖𝟐𝒕 −𝜷𝟏 𝒖𝟏𝒕
𝒘𝒕 = ; 𝒗𝒕 =
𝜶𝟏 −𝜷𝟏 𝜶𝟏 −𝜷𝟏
 2.5 Estimation of Simultaneous Equations Models

2. Indirect least squares (ILS)-----

 Notice that each reduced-form equation contains only one endogenous variable,
which is the dependent variable and which is a function solely of the exogenous
variable X (income) and the stochastic disturbances.
 Hence, the parameters of the preceding reduced-form equations may be
estimated by OLS. These estimates are:
 2.5 Estimation of Simultaneous Equations Models

2. Indirect least squares (ILS)----

 The supply function is exactly identified. Therefore, its parameters can be
estimated uniquely from the reduced-form coefficients as follows:

 Hence,the estimates of these parameters can be obtained from the estimates of

the reduced-form coefficients as:
Instrumental Variables
Instrumental Variables
Instrumental Variables
Instrumental Variables
3.
3.
▪ End of Chapter 2

▪ For Numerical Exercises, please read Gujarati Chapter 18-

20.