Chapter 14

Simultaneous Equations Models

14.1 Introduction
Simultaneous equations models differ from those we have considered in previous
chapters because in each model there are two or more dependent variables rather than
just one.
Simultaneous equations models also differ from most of the econometric models we
have considered so far because they consist of a set of equations
The least squares estimation procedure is not appropriate in these models and we must
develop new ways to obtain reliable estimates of economic parameters.

Slide 14.1
Undergraduate Econometrics, 2nd Edition-Chapter 14

14.2 A Supply and Demand Model

Supply and demand jointly determine the market price of a good and the amount of it
that is sold
An econometric model that explains market price and quantity should consist of two
equations, one for supply and one for demand.
Demand: q = 1 p + 2 y + ed

(14.2.1)

Supply: q = 1 p + es

(14.2.2)

In this model the variables p and q are called endogenous variables because their
values are determined within the system we have created.
The income variable y has a value that is given to us, and which is determined outside
this system. It is called an exogenous variable.

Slide 14.2
Undergraduate Econometrics, 2nd Edition-Chapter 14

Random errors are added to the supply and demand equations for the usual reasons,
and we assume that they have the usual properties
E (ed ) = 0, var(ed ) = d2
E (es ) = 0, var(es ) = 2s

(14.2.3)

cov(ed , es ) = 0
The fact that p is random means that on the right-hand side of the supply and demand
equations we have an explanatory variable that is random.
This is contrary to the assumption of fixed explanatory variables that we usually
make in regression model analysis.
Furthermore p and the random errors, ed and es, are correlated, making the least
squares estimator biased and inconsistent.

Slide 14.3
Undergraduate Econometrics, 2nd Edition-Chapter 14

14.3 The Reduced Form Equations

The two structural equations (14.2.1) and (14.2.2) can be solved, to express the
endogenous variables p and q as functions of the exogenous variable y.
This reformulation of the model is called the reduced form of the structural equation
system.
To find the reduced form we solve (14.2.1) and (14.2.2) simultaneously for p and q.
To solve for p, set q in the demand and supply equations to be equal,
1 p + es = 1 p + 2 y + ed

Slide 14.4
Undergraduate Econometrics, 2nd Edition-Chapter 14

 Then solve for p,

p=

2
e e
y+ d s
(1 1 ) (1 1 )

(14.3.1)

= 1 y + v1

Substitute p into equation 14.2.2 and simplify.

2
ed es
q = 1 p + es = 1
y+
+ es

(
)
(
)
1
1
1
1
1 2
e 1es
y+ 1 d
=
(1 1 )
(1 1 )

(14.3.2)

= 2 y + v2

Slide 14.5
Undergraduate Econometrics, 2nd Edition-Chapter 14

 The parameters 1 and 2 in equations 14.3.1 and 14.3.2 are called reduced form
parameters.
The error terms v1 and v2 are called reduced form errors, or disturbance terms.
The reduced form equations can be estimated consistently by least squares.
The least squares estimator is BLUE for the purposes of estimating 1 and 2.
The reduced form equations are important for economic analysis.
These equations relate the equilibrium values of the endogenous variables to the
exogenous variables. Thus, if there is an increase in income y, 1 is the expected
increase in price, after market adjustments lead to a new equilibrium for p and q.
Secondly the estimated reduced form equations can be used to predict values of
equilibrium price and quantity for different levels of income.

Slide 14.6
Undergraduate Econometrics, 2nd Edition-Chapter 14

14.4 The Failure of Least Squares Estimation in Simultaneous Equations Models

Demand: q = 1 p + 2 y + ed

(14.2.1)

Supply: q = 1 p + es

(14.2.2)

In the supply equation, (14.2.2), the random explanatory variable p on the right-hand
side of the equation is correlated with the error term es.

Slide 14.7
Undergraduate Econometrics, 2nd Edition-Chapter 14

14.4.1

An Intuitive Explanation of the Failure of Least Squares

Suppose there is a small change, or blip, in the error term es, say es.
The blip es in the error term of (14.2.2) is directly transmitted to the equilibrium
value of p.
This is clear from the reduced form equation 14.3.1. Every time there is a change in
the supply equation error term, es, it has a direct linear effect upon p.
Since 1 > 0 and 1 < 0 , if es > 0, then p < 0 .
Thus, every time there is a change in es there is an associated change in p in the
opposite direction. Consequently, p and es are negatively correlated.

Slide 14.8
Undergraduate Econometrics, 2nd Edition-Chapter 14

 Ordinary least squares estimation of the relation between q and p gives credit to
price for the effect of changes in the disturbances.
In large samples, the least squares estimator will tend to be negatively biased.
This bias persists even when the sample size is large, and thus the least squares
estimator is inconsistent.
The least squares estimator of parameters in a structural simultaneous equation
is biased and inconsistent because of the correlation between the random error
and the endogenous variables on the right-hand side of the equation.

Slide 14.9
Undergraduate Econometrics, 2nd Edition-Chapter 14

14.4.2

An Algebraic Explanation of the Failure of Least Squares

First, let us obtain the covariance between p and es.

cov ( p, es ) = E p E ( p ) es E (es )
= E ( pes )

= E [1 y + v1 ]es

[since E (es ) = 0]
[substitute for p]

e e
= E d s es
1 1

[since 1 y is exogenous]

[since ed , es assumed uncorrelated]

E (es2 )
1 1

(14.4.1)

2s
=
<0
1 1

Slide 14.10
Undergraduate Econometrics, 2nd Edition-Chapter 14

 The least squares estimator in equation 14.2.2 is

b1 =

pt qt

(14.4.2)

2
t

Substitute for q from equation 14.3.2 and simplify,

b1 =

p ( p + e ) = + p
p
p

t
2
t

st

e = 1 + ht est
2
st
t

(14.4.3)

where

ht =

pt
pt2

Slide 14.11
Undergraduate Econometrics, 2nd Edition-Chapter 14

 The expected value of the least squares estimator is

E (b1 ) = 1 + E ( ht est ) 1

(14.4.4)

The expectation E ( ht est ) 0 because es and p are correlated.

In large samples there is a similar failure.
Multiply through the supply equation by price, p, take expectations and solve.
pq = 1 p 2 + pes

E ( pq ) = 1E ( p 2 ) + E ( pes )
(14.4.5)
1 =

E ( pq )
E ( p2 )

E ( pes )
E ( p2 )

Slide 14.12
Undergraduate Econometrics, 2nd Edition-Chapter 14

 In large samples, as T, sample analogs of expectations, which are averages,

converge to the expectations. That is,

q p / T E ( pq ), p
t

2
t

/ T E ( p2 ) .

Consequently, because the covariance between p and es is negative, from equation

14.4.1,
q p /T
E ( pq )
b =

=
p /T E ( p )
t

2
t

E ( pes )
E( p

= 1

2s (1 1 )
E( p

< 1

(14.4.6)

The least squares estimator of the slope of the supply equation, in large samples,
converges to a value less than 1.

Slide 14.13
Undergraduate Econometrics, 2nd Edition-Chapter 14

14.5 The Identification Problem

In the supply and demand model given by equations 14.2.1 and 14.2.2, the parameters
of the demand equation, 1 and 2, can not be consistently estimated by any estimation
method.
The slope of the supply equation, 1, can be consistently estimated.
The problem lies with the model that we are using. There is no variable in the supply
equation that will shift it relative to the demand curve.
It is the absence of variables from an equation that makes it possible to estimate its
parameters.

A general rule, which is called a condition for identification of an

equation, is this:

Slide 14.14
Undergraduate Econometrics, 2nd Edition-Chapter 14

A Necessary Condition for Identification: In a system of M simultaneous

equations, which jointly determine the values of M endogenous variables, at
least M1 variables must be absent from an equation for estimation of its
parameters to be possible. When estimation of an equations parameters is
possible, then the equation is said to be identified, and its parameters can be
estimated consistently. If less than M1 variables are omitted from an equation,
then it is said to be unidentified and its parameters can not be consistently
estimated.

Slide 14.15
Undergraduate Econometrics, 2nd Edition-Chapter 14

 In our supply and demand model there are M=2 equations and there are a total of three
variables: p, q and y.
In the demand equation none of the variables are omitted; thus it is unidentified and its
parameters can not be estimated consistently.
In the supply equation M1=1 and one variable, income, is omitted; the supply curve
is identified and its parameter can be estimated.
The identification condition must be checked before trying to estimate an equation.

Slide 14.16
Undergraduate Econometrics, 2nd Edition-Chapter 14

Remark: The two-stage least squares estimation procedure is developed in

Chapter 13 and shown to be an instrumental variables estimator. The number
of instrumental variables required for estimation of an equation within a
simultaneous equations model is equal to the number of right-hand-side
endogenous variables. In a typical equation within a simultaneous equations
model several exogenous variables appear on the right-hand-side.

Thus

instruments must come from those exogenous variables omitted from the
equation in question. Consequently, identification requires that the number of
omitted exogenous variables in an equation be at least as large as the number of
right-hand-side endogenous variables. This ensures an adequate number of
instrumental variables.

Slide 14.17
Undergraduate Econometrics, 2nd Edition-Chapter 14

14.6 The Two-Stage Least Squares Estimation Procedure

In this section we briefly describe two-stage least squares (2SLS) estimation
q = 1 p + es

(14.2.2)

The variable p is composed of a systematic part, which is its expected value E(p), and
a random part, which is the reduced form random error v1.
p = E ( p ) + v1 = 1 y + v1

(14.6.1)

In the supply equation (14.2.2) the portion of p that causes problems for the least
squares estimator is v1, the random part.
Suppose we knew the value of 1. Then we could replace p in (14.2.2) by (14.6.1) to
obtain
Slide 14.18
Undergraduate Econometrics, 2nd Edition-Chapter 14

q = 1[ E ( p ) + v1 ] + es
= 1E ( p ) + (1v1 + es )

(14.6.2)

= 1E ( p ) + e*
We could apply least squares to equation 14.6.2 to consistently estimate 1.
We can estimate 1 using 1 from the reduced form equation for p.
A consistent estimator for E(p) is
p = 1 y
Using p as a replacement for E(p) in (14.6.2) we obtain
q = 1 p + e*

(14.6.3)
Slide 14.19

Undergraduate Econometrics, 2nd Edition-Chapter 14

 In large samples, p and the random error e* are uncorrelated, and consequently the
parameter 1 can be consistently estimated by applying least squares to (14.6.3).
Estimating the equation 14.6.3 by least squares generates the so-called two-stage least
squares estimator of 1 , which is consistent and asymptotically normal.

Slide 14.20
Undergraduate Econometrics, 2nd Edition-Chapter 14

13.4 An Example of Two Stage Least Squares Estimation

Consider a supply and demand model for truffles:
demand: qt = 1 + 2 pt + 3 pst + 4 dit + etd
supply: qt = 1 + 2 pt + 3 pft + ets

(13.4.1)

(13.4.2)

In the demand equation q is the quantity of truffles traded in a particular French

market at time t, p is the market price of truffles, ps is the market price of a substitute
for real truffles (another fungus much less highly prized), and di is per capita
disposable income.

Slide 14.21
Undergraduate Econometrics, 2nd Edition-Chapter 14

 The supply equation contains the market price and quantity supplied. Also it includes
pf, the price of a factor of production, which in this case is the hourly rental price of
truffle-pigs used in the search process.
In this model we assume that p and q are endogenous variables.
The exogenous variables are ps, di, pf and the intercept variable.

Slide 14.22
Undergraduate Econometrics, 2nd Edition-Chapter 14

13.4.1 Identification

The rule for identifying an equation is that in a system of M equations at least M1

variables must be omitted from each equation in order for it to be identified.
In the demand equation the variable pf is not included and thus the necessary M1=1
variable is omitted.
In the supply equation both ps and di are absent; more than enough to satisfy the
identification condition.
We conclude that each equation in this system is identified and can thus be estimated
by two-stage least squares.

Slide 14.23
Undergraduate Econometrics, 2nd Edition-Chapter 14

13.4.2 The Reduced Form Equations

The reduced form equations express each endogenous variable, p and q, in terms of the
exogenous variables ps, di, pf and the intercept variable, plus an error term.
qt = 11 + 21 pst + 31dit + 41 pft + vt1
pt = 12 + 22 pst + 32 dit + 42 pft + vt 2
Data for each of the endogenous and exogenous variables are given in Table 13.1.
The price p is measured in $ per ounce, q is measured in ounces, ps is measured in $
per pound, di is in $1000 and pf is the hourly rental rate for a truffle-finding pig.

Slide 14.24
Undergraduate Econometrics, 2nd Edition-Chapter 14

Table 13.1
OBS
1
2
27
28
29
30

Sample Truffle Supply and Demand Data

p
q
ps
di
9.88
19.89
19.97
21.03
13.41
13.04
18.04
20.43
27.90
20.81
28.98
46.32
27.00
14.95
18.52
48.94
29.48
26.27
28.16
51.25
35.15
20.65
28.43
48.36

pf
10.52
19.67
27.80
30.34
24.12
34.01

Table 13.2a Reduced Form Equation for Quantity of Truffles (q)

Variable
Estimate
Std. Error
t-value
p-value
Const
7.895
3.019
2.615
0.0147
PS
0.656
0.133
4.947
0.0001
DI
0.217
0.065
3.323
0.0026
PF
0.113
0.0001
0.507
4.491

Table 13.2b Reduced Form Equation for Price of Truffles (p)

Variable
Estimate
Std. Error
t-value
p-value
Const
2.478
0.0002
10.837
4.374
PS
0.569
0.109
5.229
0.0001
DI
0.253
0.054
4.736
0.0001
PF
0.451
0.093
4.872
0.0001

Slide 14.25
Undergraduate Econometrics, 2nd Edition-Chapter 14

 The reduced form equations are used to obtain p t which will be used in place of pt on
the right-hand side of the supply and demand equations in the second stage of twostage least squares.
p t = 12 + 22 pst + 32 dit + 42 pft
= 10.837 + .569 pst + .253dit + .451 pf t

Slide 14.26
Undergraduate Econometrics, 2nd Edition-Chapter 14

 The 2SLS results are given in Tables 13.3a and 13.3b.

Table 13.3a 2SLS Estimates for Truffle Demand (qd)
Variable

Estimate

Std. Error

t-value

p-value

Const

4.279

5.161

0.829

0.4145

1.123

0.460

2.441

0.0217

PS

1.296

0.331

3.919

0.0006

DI

0.501

0.213

2.359

0.0261

Estimate

Std. Error

t-value

p-value

20.033

1.160

17.264

0.0001

1.014

0.071

14.297

0.0001

PF

1.001

0.078

12.784

0.0001

Table 13.3b 2SLS Estimates for Truffle Supply (qs)

Variable
Const

Slide 14.27
Undergraduate Econometrics, 2nd Edition-Chapter 14

 The estimated demand curve results are in Table 13.3a.

Note that the coefficient of price is negative, indicating that as the market price rises
the quantity demanded of truffles declines, as predicted by the law of demand.
The standard errors that are reported are obtained from 2SLS software.
They and the t-values are valid in large samples.
The p-value indicates that the estimated slope of the demand curve is significantly
different from zero.
Increases in the price of the substitutes for truffles increase the demand for truffles,
which is a characteristic of substitute goods.
Finally the effect of income is positive, indicating that truffles are a normal good.

Slide 14.28
Undergraduate Econometrics, 2nd Edition-Chapter 14

 The supply equation results, appear in Table 13.3b.

As anticipated, increases in the price of truffles increase the quantity supplied,
and increases in the rental rate for truffle-seeking pigs, which is an increase in the
cost of a factor of production, reduces supply.
Both of these variables have statistically significant coefficient estimates.

Slide 14.29
Undergraduate Econometrics, 2nd Edition-Chapter 14