Chapter 21: Simultaneous Equation Models Identification

Chapter 21 Outline

Review
o Demand and Supply Models
o Ordinary Least Squares (OLS) Estimation Procedure
o Reduced Form (RF) Estimation Procedure
Two Stage Least Squares (TSLS): An Instrumental Variable Two Step Approach
Comparison of Reduced Form (RF) and Two Stage Least Squares (TSLS) Estimates
Statistical Software and Two Stage least Squares (TSLS)
Identification of Simultaneous Equation Models
o Underidentification
o Overidentification
Summary of Identification Issues: Reduced Form and Two Stage Least Squares
Estimation Procedures

Chapter 21 Preview Questions

Beef Market Data: Monthly time series data relating to the market for beef from 1977 to 1986.

Q
t
Quantity of beef in month t (millions of pounds)
P
t
Real price of beef in month t (1982-84 cents per pound)
FeedP
t
Real price of cattle feed in month t (1982-84 cents per pounds of corn cobs)
Inc
t
Real disposable income in month t (thousands of chained 2005 dollars)
ChickP
t
Real rice of whole chickens in month t (1982-84 cents per pound)
Year
t
Year

Consider the model for the beef market that we used in the last chapter:

Demand Model: Q
D
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+ e
D
t

Supply Model: Q
S
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S
t
Equilibrium: Q
D
t
= Q
S
t
= Q
t

Endogenous Variables: Q
t
and P
t
Exogenous Variables: FeedP
t
and Inc
t


1. We shall now introduce another estimation procedure for simultaneous equation models, the
two stage least squares (TSLS) estimation procedure:

1
st
Stage: Estimate the variable that is creating the problem, the explanatory endogenous
variable:
Dependent variable: Original endogenous explanatory variable that creates the
bias problem.
Explanatory variables: All exogenous variables.

2
nd
Stage: Estimate the original models using the estimate of the problem
explanatory endogenous variable
Dependent variable: Original dependent variable.
Explanatory variables: 1st stage estimate of the problem explanatory
endogenous variable and any relevant exogenous explanatory variable.

2
Naturally, begin by focusing on the first stage.
1
st
Stage: Estimate the variable that is creating the problem, the explanatory endogenous
variable:
Dependent variable: Original endogenous explanatory variable that creates the
bias problem. In this case, the price of beef, P
t
, is the problem explanatory
variable.
Explanatory variables: All exogenous variables. In this case, the exogenous variables
are FeedP
t
and Inc
t
.

Using the ordinary least squares (OLS) estimation procedure, what equation estimates
the problem explanatory variable, the price of beef?
Click here to access data {EViewsLink}
EstP = ______________________________________________

Generate a new variable, EstP, that estimates the price of beef based on the 1
st
stage.


2. Next, we focus on the 2
nd
stage and consider the demand model:

Demand Model: Q
D
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+ e
D
t


2
nd
Stage: Estimate the original models using the estimate of the problem
explanatory endogenous variable
Dependent variable: Original dependent variable. In this case, the original
explanatory variable is the quantity of beef, Q
t
.
Explanatory variables: 1st stage estimate of the problem explanatory
endogenous variable and any relevant exogenous explanatory variable. In this
case, the estimate of the price of beef and income, EstP
t
and Inc
t
.

Beef Market Demand Model: Dependent Variable: Q
Explanatory Variables: EstPrice and Inc

a. Using the ordinary least squares (OLS) estimation procedure, estimate the EstPrice
coefficent of the demand model.

b. Compare the two stage least squares coefficient estimate for the demand model with
the estimate computed using the reduced form estimation procedure in the previous
chapter.

3
3. Now, consider the supply model:
Supply Model: Q
S
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S
t
and the second stage of the two stage least squares estimation procedure.

2
nd
Stage: Estimate the original models using the estimate of the problem
explanatory endogenous variable
Dependent variable: Original dependent variable. In this case, the original
explanatory variable is the quantity of beef, Q
t
.
Explanatory variables: 1st stage estimate of the problem explanatory
endogenous variable and any relevant exogenous explanatory variable. In this
case, the estimate of the price of beef and income, EstP
t
and PFeed
t
.

Beef Market Supply Model: Dependent Variable: Q
Explanatory Variables: EstPrice and FeedP

a. Using the ordinary least squares (OLS) estimation procedure, estimate the EstPrice
coefficient of the supply model.

b. Compare the two stage least squares coefficient estimate for the supply model with
the estimate computed using the reduced form estimation procedure in the previous
chapter.

4. Reconsider the following simultaneous equation model of the beef market and the reduced
form estimates:

Demand and Supply Models:
Demand Model: Q
D
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+ e
D
t

Supply Model: Q
S
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S
t

Equilibrium: Q
D
t
= Q
S
t
= Q
t

Endogenous Variables: Q
t
and P
t
Exogenous Variables: FeedP
t
and Inc
t


Reduced Form (RF) Estimates
Quantity Reduced Form Estimates: EstQ = a
Q
Const
+ a
Q
FP
FeedP
t
+ a
Q
I
Inc
t

Price Reduced Form Estimates: EstP = a
P
Const
+ a
P
FP
FeedP
t
+ a
P
I
Inc
t

a. Focus on the reduced form estimates for the income coefficients:
1) The reduced form income coefficient estimates, a
Q
I
and a
P
I
, allowed us to
estimate the slope of which curve? ___ Demand ___Supply
2) If the reduced form income coefficient estimates were not available, would
we be able to estimate the slope of this curve? ___

b. Focus on the reduced form estimates for the feed price coefficients:
1) The reduced form feed price coefficient estimates of these coefficients, a
Q
FP

and a
P
FP
, allowed us to estimate the slope of which curve?
___Demand ___ Supply
2) If the reduced form feed price coefficient estimates were not available, would
we be able to estimate the slope of this curve? ___
4
Review: Demand and Supply Models

In simultaneous equation models the value of an explanatory variable is determined within the
model. For the economist, arguably the most important example of a simultaneous equations
model is the demand/supply model:

Demand Model: Q
D
t
=
D
Const
+
D
P
P
t
+ Other Demand Factors + e
D
t

Supply Model: Q
S
t
=
S
Const
+
S
P
P
t
+ Other Supply Factors + e
S
t

Equilibrium: Q
D
t
= Q
S
t
= Q
t

Endogenous Variables: Q
t
and P
t
Exogenous Variables: Other Demand and Supply Factors

Project: Estimate the beef market demand and supply parameters

In a simultaneous equation model, it is important to emphasize the distinction between
endogenous and exogenous variables. Endogenous variables are variables whose values are
determined within the model. In the demand/supply example, both quantity and price are
determined simultaneously within the model; the model is explaining both the equilibrium
quantity and the equilibrium price as depicted by the intersection of the supply and demand
curves. On the other hand, exogenous are determined outside the context of the model; the
values of exogenous variables are taken as given. The model does not attempt to explain how the
values of exogenous variables are determined.
Endogenous variables Variables determined within the model: Quantity and Price.
Exogenous variables Variables determined outside the model.

Unlike single regression models, an endogenous variable can be an
explanatory variable in simultaneous equation models. In the
demand and supply models the price is such a variable. Both the
quantity demanded and the quantity supplied depend on the
price; hence, the price is an explanatory variable. Furthermore, the
price is determined within the model; the price is an endogenous
variable. The price is determined by the intersection of the supply
and demand curves. The traditional demand/supply graph clearly
illustrates that both the quantity, Q
t
, and the price, P
t
, are
endogenous, both are determined within the model.
In our last lecture, we showed why simultaneous equations cause
a problem for the ordinary least squares (OLS) estimation
procedure:

Simultaneous Equations and Bias: Whenever an explanatory variable is also an endogenous
variable, the ordinary least squares (OLS) estimation procedure is biased.

In the demand/supply model, the price is an endogenous explanatory variable. When we used the
ordinary least squares (OLS) estimation procedure to estimate the value of the price coefficient in
the demand and supply models we observed that a problem emerged. In each model, price and the
error term were correlated; unfortunately, such correlation results in bias:


Figure 21.1: Demand/supply Model

S
D
Price
Quantity
P
Q
Q = Equlibrium Quantity
P = Equilibrium Price
5
Demand Model: Supply Model:
Q
D
t
=
D
Const
+
D
P
P
t
+ Other Demand Factors + e
D
t
Q
S
t
=
S
Const
+
S
P
P
t
+ Other Supply Factors + e
S
t

S
D (e
D
down)
D
D (e
D
up)
Price
Quantity
P
P (e
D
down)
P (e
D
up)


Figure 21.2: Effect of Demand Error Term Figure 21.3: Effect of Supply Error Term

e
D
t
up e
D
t
down e
S
t
up e
S
t
down

P
t
up P
t
down P
t
down P
t
up
Explanatory variable and Explanatory variable and
error term and positively correlated error term and negatively correlated

OLS estimation procedure OLS estimation procedure
for coefficient value for coefficient value
biased upward biased downward

So, where did we go from here? We explored the possibility that the ordinary least squares (OLS)
estimation procedure might be consistent. After all, is not half a loaf better than none? We took
advantage of our Econometrics Lab to address this issue. Recall the distinction between an
unbiased and consistent estimation procedure:
Unbiased: The estimation procedure does not systematically underestimate or overestimate
the actual value; that is, after many, many repetitions the average of the estimates equals the
actual value.

Consistent but Biased: As consistent estimation procedure can be biased. But, as the sample
size, as the number of observations, grows:
The magnitude of the bias decreases. That is, the mean of the coefficient estimates
probability distribution approaches the actual value.
The variance of the estimates probability distribution diminishes and approaches 0.
Unfortunately, the Econometrics Lab illustrates the sad fact that the ordinary least squares (OLS)
estimation procedure is neither unbiased nor consistent.

We then considered an alternative estimation procedure: the reduced form (RF) estimation
procedure. Our Econometrics Lab taught us that while the reduced form (RF) estimation
procedure is biased, it is consistent. That is, as the sample size grows, the average of the
coefficient estimates gets closer and closer to the actual value and the variance grew smaller
and smaller after many, many repetitions. Arguably, when choosing between two biased
estimates, it is better to use the one that is consistent. This represents the econometricians
pragmatic, half a loaf is better than none philosophy.

We shall now quickly review the reduced form (RF) estimation procedure.
S
S (e
S
down)
D
S (e
S
up)
Price
Quantity
P
P (e
S
down)
P (e
S
up)
6
Review: Reduced Form (RF) Estimation Procedure One Way to Cope with
Simultaneous Equation Models

We begin with the simultaneous equation model and then constructed the reduced form
equations:
Demand and Supply Models:
Demand Model: Q
D
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+ e
D
t

Supply Model: Q
S
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S
t

Equilibrium: Q
D
t
= Q
S
t
= Q
t

Endogenous Variables: Q
t
and P
t
Exogenous Variables: FeedP
t
and Inc
t

Reduced Form (RF) Estimates
Quantity Reduced Form Estimates: EstQ = a
Q
Const
+ a
Q
FP
FeedP
t
+ a
Q
I
Inc
t

Price Reduced Form Estimates: EstP = a
P
Const
+ a
P
FP
FeedP
t
+ a
P
I
Inc
t
We can use the coefficient interpretation approach to estimate the slopes of the demand and
supply in terms of the reduced form estimates:

Suppose that FeedP increases while Suppose that Inc increases while
Inc remains constant: FeedP remains constant:
Q = a
Q
FP
FeedP Q = a
Q
I
Inc
P = a
P
FP
FeedP P = a
P
I
Inc












Figure 21.4: Reduced Form Summary and Coefficient Interpretation Approach


Q
P
=
a
Q
FP
FeedP
a
P
FP
FeedP
=
a
Q
FP
a
P
FP

Q
P
=
a
Q
I
Inc
a
P
I
Inc
=
a
Q
I
a
P
I


We are moving from one equilibrium to We are moving from one equilibrium to
another on the same demand curve. another on the same supply curve.
This movement represents a change in This movement represents a change in
the quantity of beef demanded, Q
D
: the quantity of beef supplied, Q
S
:
b
D
P
=
Q
D
P
=
a
Q
FP
FeedP
a
P
FP
FeedP
=
a
Q
FP
a
P
FP
b
S
P
=
Q
S
P
=
a
Q
I
Inc
a
P
I
Inc
=
a
Q
I
a
P
I

D
Price
Quantity
FeedP increaes
S
S
P =

F
Q
P
FeedP

F
Q
P
FeedP
Q =
Inc constant
S
Price
Quantity
FeedP constant
D
D
P =

I
Q
Inc

I
Q
Inc
Q =
Inc increases
Initial
equilibrium
7
Intuition: Critical Role of the Exogenous Variable Absent from the Model
Demand Model: Q
D
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+ e
D
t

Changes in the feed price, the exogenous variable absent from the demand model, allow us
to estimate the slope of the demand curve. The supply curve shifts, but the demand curve
remains stationary. Consequently, the equilbria trace out the stationary demand curve.
Supply Model: Q
S
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S
t

Changes in income, the exogenous variable absent from the demand model, allow us to
estimate the slope of the supply curve. The demand curve shifts, but the supply curve
remains stationary. Consequently, the equilibria trace out the stationary supply curve.
Key Point: In each case, changes in the exogenous variable absent in the model allow us to
estimate the parameters of the model.

Calculating the Reduced From Estimates

We use the ordinary least squares (OLS) estimation procedure to estimate the reduced form
parameters and then use the ratio of the reduced form estimates to estimate the slopes of the
demand and supply curves:
Click here to access data {EViewsLink}
Quantity Reduced Form Equation: Dependent Variable: Q
Explanatory Variables: FeedP and Inc
Dependent Variable: Q
Method: Least Squares
Sample: 1977M01 1986M12
Included observations: 120

Coefficient Std. Error t-Statistic Prob.

FEEDP -331.9966 121.6865 -2.728293 0.0073
INC 17.34683 2.132027 8.136309 0.0000
C 138725.5 13186.01 10.52066 0.0000

Table 21.1: EView Regression Results Quantity Reduced Form Equation
Price Reduced Form Equation: Dependent Variable: P
Explanatory Variables: FeedP and Inc
Dependent Variable: P
Method: Least Squares
Sample: 1977M01 1986M12
Included observations: 120

Coefficient Std. Error t-Statistic Prob.

FEEDP 1.056242 0.286474 3.687044 0.0003
INC 0.018825 0.005019 3.750636 0.0003
C 33.02715 31.04243 1.063936 0.2895

Table 21.2: EView Regression Results Price Reduced Form Equation
Estimated Slope Estimated Slope
of the Demand Curve of the Supply Curve

Ratio of Reduced Form Ratio of Reduced Form
Feed Price Income
Coefficient Estimates Coefficient Estimates

Estimate of
D
P
= b
D
P
=
a
Q
FP
a
P
FP
Estimate of
S
P
= b
S
P
=
a
Q
I
a
P
I

=
332.00
1.0562
= 314.3 =
17.347
.018825
= 921.5
8
Two Stage Least Squares (TSLS): An Instrumental Variable Two Step
Approach A Second Way to Cope with Simultaneous Equation Models

Another way to estimate simultaneous equation model is the two stage least squares (TSLS)
estimation procedure. As the name suggests the procedure involves two steps. As we shall see,
two stage least squares (TSLS) uses a strategy that is similar to the instrumental variable (IV)
approach.

1
st
Stage: Estimate the variable that is creating the problem, the explanatory endogenous
variable:
Dependent variable: Original endogenous explanatory variable that creates the bias
problem.
Explanatory variables: All exogenous variables.

2
nd
Stage: Estimate the original models using the estimate of the problem explanatory
endogenous variable
Dependent variable: Original dependent variable.
Explanatory variables: 1
st
stage estimate of the problem explanatory endogenous
variable and any relevant exogenous explanatory variables.

We shall now illustrate the two stage least squares (TSLS) approach by consider the beef market:

Beef Market Data: Monthly time series data relating to the market for beef from 1977 to 1986.

Q
t
Quantity of beef in month t (millions of pounds)
P
t
Real price of beef in month t (1982-84 cents per pound)
FeedP
t
Real price of cattle feed in month t (1982-84 cents per pounds of corn cobs)
Inc
t
Real disposable income in month t (thousands of chained 2005 dollars)
ChickP
t
Real rice of whole chickens in month t (1982-84 cents per pound)
Year
t
Year

Consider the model for the beef market that we used in the last chapter:

Demand Model: Q
D
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+ e
D
t

Supply Model: Q
S
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S
t
Equilibrium: Q
D
t
= Q
S
t
= Q
t

Endogenous Variables: Q
t
and P
t
Exogenous Variables: FeedP
t
and Inc
t




9
The strategy for the first stage is similar to the strategy used by instrumental variable (IV)
approach. We use the ordinary least squares (OLS) regression to estimate an equation in which
the problem endogenous explanatory variable becomes the dependent variable. The
explanatory variables are all the exogenous variables. In our example, price is the problem
explanatory variable; consequently, price becomes the dependent variable in the first stage. The
exogenous variables, income and feed price, are the explanatory variables.

1
st
Stage: Estimate the variable that is creating the problem, the explanatory endogenous variable:
Dependent variable: Original endogenous explanatory variable that creates the bias
problem. In this case, the price of beef, P
t
, is the problem explanatory variable.
Explanatory variables: All exogenous variables. In this case, the exogenous variables are
FeedP
t
and Inc
t
.

Click here to access data {EViewsLink}
1
st
Stage: Dependent variable: P
Explanatory variables: FeedP and Inc

Dependent Variable: P
Method: Least Squares
Sample: 1977M01 1986M12
Included observations: 120

Coefficient Std. Error t-Statistic Prob.

FEEDP 1.056242 0.286474 3.687044 0.0003
INC 0.018825 0.005019 3.750636 0.0003
C 33.02715 31.04243 1.063936 0.2895

Table 21.3: EViews Regression Results TSLS 1
st
Stage

Estimated Equation: EstP = 33.027 + 1.0562FeedP + .018825Inc

Using these regression results we can estimate the price of beef based on the exogenous variables,
income and feed price.

The strategy for the second stage is also similar to the instrumental variable (IV) approach. The
dependent variable is the original dependent variable, quantity. The explanatory variables do not
include the problem endogenous explanatory variable; instead, the estimate of the problem
explanatory variable based on the first stage is used. Instead of using the price as an explanatory
variable, we use Stage 1s estimate of the price.

For the the demand model, the dependent variable is the quantity of beef, Q, and the explanatory
variables EstP and Inc:

2
nd
Stage: Estimate the original models using the estimate of the problem explanatory
endogenous variable
Dependent variable: Original dependent variable. In this case, the original
explanatory variable is the quantity of beef, Q
t
.
Explanatory variables: 1st stage estimate of the problem explanatory endogenous
variable and any relevant exogenous explanatory variable. In this case, the estimated
of the price of beef and income, EstP
t
, and Inc
t
.

10
For the demand model, the dependent variable is the quantity of beef, Q, and the explanatory
variables EstP and Inc:

2
nd
Stage Beef Market Demand Model: Dependent variable: Q
Explanatory Variables: EstP and Inc
Dependent Variable: Q
Method: Least Squares
Sample: 1977M01 1986M12
Included observations: 120

Coefficient Std. Error t-Statistic Prob.

ESTP -314.3312 115.2117 -2.728293 0.0073
INC 23.26411 2.161914 10.76089 0.0000
C 149106.9 16280.07 9.158860 0.0000

Table 21.4: EViews Regression Results TSLS 2
nd
Stage Demand

Estimated Equation: EstQ
D
= 149,107 314.3EstP + 23.26Inc

We estimate the slope of the demand curve to be 314.3.

For the supply model, the dependent variable is the quantity of beef, Q, and the explanatory
variables EstP and FeedP:

2
nd
Stage: Estimate the original models using the estimate of the problem explanatory
endogenous variable
Dependent variable: Original dependent variable. In this case, the original
explanatory variable is the quantity of beef, Q
t
.
Explanatory variables: 1st stage estimate of the problem explanatory endogenous
variable and any relevant exogenous explanatory variable. In this case, the estimated
of the price of beef and income, EstP
t
, and PFeed
t
.

2
nd
Stage Beef Market Supply Model: Dependent variable: Q
Explanatory Variables: EstP and FeedP
Dependent Variable: Q
Method: Least Squares
Sample: 1977M01 1986M12
Included observations: 120

Coefficient Std. Error t-Statistic Prob.

ESTP 921.4783 113.2551 8.136309 0.0000
FEEDP -1305.262 121.2969 -10.76089 0.0000
C 108291.8 16739.33 6.469303 0.0000

Table 21.5: EViews Regression Results TSLS 2
nd
Stage Supply

Estimated Equation: EstQ
S
= 108,292 + 921.5EstP 1,305.2 FeedP

We estimate the slope of the demand curve to be 921.5.

Compare the estimates from the reduced form (RF) approach with the estimates from the two
stage least squares (TSLS) approach:
Estimate of Reduced Form (RF) Two Stage Least Squares (TSLS)

D
P
314.3 314.3

S
P
921.5 921.5
The estimates are identical. In this case, the reduced form (RF) estimation procedure and the two
stage least squares (TSLS) estimation procedure produce identical results.
11
Software and Two Stage Least Squares (TSLS)

Many statistical packages provide an easy way to apply the two state least squares (TSLS)
estimation procedure so that we do not need to generate the estimate of the problem
explanatory variable ourselves.

Getting Started in EViews

EViews makes it very easy for us to use the two stage least squares (TSLS) approach. EViews
does most of the work for us eliminating the need to generate a new variable:
In the Workfile window, highlight all relevant variables: q p feedp income
Double click on one of the highlighted variables and click Open Equation.
In the Equation Estimation window, click Options and then select TSLS Two-Stage Least
Squares (TSNLS and ARIMA).
In the Instrument List box, enter the exogenous variables: feedp income
In the Equation Specification box, enter the dependent variable followed by the
explanatory variables (both exogenous and endogenous) for each model:
o To estimate the demand model enter q p income
o To estimate the supply model enter q p feedp

Click here to access data {EViewsLink}
Beef Market Demand Model: Dependent variable: Q
Explanatory variables: P and Inc
Instrument List: FeedP and Inc
Dependent Variable: Q
Method: Two-Stage Least Squares
Sample: 1977M01 1986M12
Included observations: 120
Instrument list: FEEDP INC

Coefficient Std. Error t-Statistic Prob.

P -314.3188 58.49828 -5.373129 0.0000
INC 23.26395 1.097731 21.19276 0.0000
C 149106.5 8266.413 18.03763 0.0000

Table 21.6: EViews Regression Results TSLS Demand

Beef Market Supply Model: Dependent variable: Q
Explanatory variables P and FeedP
Instrument List: FeedP and Inc
Dependent Variable: Q
Method: Two-Stage Least Squares
Sample: 1977M01 1986M12
Included observations: 120
Instrument list: FEEDP INC

Coefficient Std. Error t-Statistic Prob.

P 921.4678 348.8314 2.641585 0.0094
FEEDP -1305.289 373.6098 -3.493723 0.0007
C 108292.0 51558.51 2.100372 0.0378

Table 21.7: EViews Regression Results TSLS Supply

Note that these are the same estimates that we obtained when we generate the estimate of the
price on our own.

12
Taking Stock

Let us step back for a moment to review our beef market model:
Demand and Supply Models:
Demand Model: Q
D
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+ e
D
t

Supply Model: Q
S
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S
t

Equilibrium: Q
D
t
= Q
S
t
= Q
t

Endogenous Variables: Q
t
and P
t
Exogenous Variables: FeedP
t
and Inc
t

Reduced Form (RF) Equations
Quantity Reduced Form Equation: Q
t
=
Q
Const
+
Q
FP
FeedP
t
+
Q
I
Inc
t
+
Q
t

Price Reduced Form Equation: P
t
=
P
Const
+
P
FP
FeedP
t
+
P
I
Inc
t
+
P
t
The reduced form estimation procedure uses the the reduced form estimates to estimate the
slopes of the demand and supply curves:


In each model there is one exogenous variable absent and one endogenous explanatory variable.
This one to one correspondence allows us to estimate the coefficient of the endogenous
explanatory variable.
Demand Model: The absent exogenous variable, FeedP, plays a critical role. Changes in FeedP shift
the supply curve allowing us to estimate the slope of the demand curve, changes in FeedP allow
us to estimate the coefficient of the demand models endogenous explanatory variable, P.
Figure 21.5: Reduced Form Summary and Coefficient Interpretation Approach

Changes in the feed price shift the Changes in income shift the
supply curve, but not the demand curve demand curve, but not the supply curve

b
D
P
=
Q
D
P
=
a
Q
FP
FeedP
a
P
FP
FeedP
=
a
Q
FP
a
P
FP
b
S
P
=
Q
S
P
=
a
Q
I
Inc
a
P
I
Inc
=
a
Q
I
a
P
I


The exogenous variable absent The exogenous variable absent
in the demand model, FeedP, in the supply model, Inc,
allows us to estimate the coefficient allows us to estimate the coefficient
of the endogenous explanatory of the endogenous explanatory
variable, P, in the demand model. variable, P, in the demand model.
D
Price
Quantity
FeedP increaes
S
S
P =

F
Q
P
FeedP

F
Q
P
FeedP
Q =
Inc constant
S
Price
Quantity
FeedP constant
D
D
P =

I
Q
Inc

I
Q
Inc
Q =
Inc increases
Initial
equilibrium
Estimated slope of demand curve: Estimated slope of supply curve:
b
D
P
=
Q
D
P
b
S
P
=
Q
S
P


13
Supply Model: The absent exogenous variable, Inc, plays a critical role. Changes in Inc shift the
demand curve allowing us to estimate the slope of the supply curve, changes in Inc allow us to
estimate the coefficient of the supply models endogenous explanatory variable, P.

The Order Condition formalizes this relationship:

Number of Less Than Number of
exogenous variables Equal To endogenous explanatory
absent from the model Greater Than variables in the model


Model Model Model
Underidentified Identified Overidentified

No Estimates Unique Estimates Multiple Estimates

Figure 21.6: Order Condition

Underidentification

We shall now illustrate the underidentification problem. Suppose that no income information
was available. Obviously, if we have no income information, we cannot include Inc as an
explanatory variable in either the original demand and supply models or the reduced form
equations:

Demand and Supply Models:
Demand Model: Q
D
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+ e
D
t

Supply Model: Q
S
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S
t

Equilibrium: Q
D
t
= Q
S
t
= Q
t

Endogenous Variables: Q
t
and P
t
Exogenous Variables: FeedP
t
and Inc
t


Reduced Form (RF) Equations
Quantity Reduced Form Equation: Q
t
=
Q
Const
+
Q
FP
FeedP
t
+
Q
I
Inc
t
+
Q
t

Price Reduced Form Equation: P
t
=
P
Const
+
P
FP
FeedP
t
+
P
I
Inc
t
+
P
t


Let us now apply the order condition by counting the number of absent exogenous variables and
endogenous explanatory variables in each model:

Demand Model Supply Model
Exogenous Endogenous Exogenous Endogenous
variables explanatory variables explanatory
absent from variables in absent from variables in
the model the model the model the model

FeedP P None P
1 1 0 1


14
The order condition suggests that we should
still be able to estimate the coefficient of the endogenous explanatory variable, P, in the
demand model.
not be able to estimate the coefficient of the endogenous explanatory variable, P, in the
supply model.

The coefficient interpretation approach explains why. We can still estimate the slope of the
demand curve, however, by calculating the ratio of the reduced form feed price coefficient
estimates, a
Q
FP
and a
P
FP
. We shall use the coefficient estimate approach to explain this phenomenon
to take advantage of the intuition it provides.



There is both good news and bad news when we have feed price information but no income
information:
Good news: Since we still have feed price information, we still have information about
how the supply curve shifts. The shifts in the supply curve cause the equilibrium
quantity and price to move along the demand curve. In other words, shifts in the supply
curve trace out the demand curve; hence, we can still estimate the slope of the
demand curve.
Figure 21.7: Reduced Form Summary and Coefficient Interpretation Approach

Changes in the feed price shift the Changes in income shift the
supply curve, but not the demand curve demand curve, but not the supply curve

b
D
P
=
Q
D
P
=
a
Q
FP
FeedP
a
P
FP
FeedP
=
a
Q
FP
a
P
FP
b
S
P
=
Q
S
P
=
a
Q
I
Inc
a
P
I
Inc
=
a
Q
I
a
P
I


The exogenous variable absent The exogenous variable absent
in the demand model, FeedP, in the supply model, Inc,
allows us to estimate the coefficient allows us to estimate the coefficient
of the endogenous explanatory of the endogenous explanatory
variable, P, in the demand model. variable, P, in the demand model.
Estimated slope of demand curve: Estimated slope of supply curve:
b
D
P
=
Q
D
P
b
S
P
=
Q
S
P


D
Price
Quantity
FeedP increaes
S
S
P =

F
Q
P
FeedP

F
Q
P
FeedP
Q =
Inc constant
S
Price
Quantity
FeedP constant
D
D
P =

I
Q
Inc

I
Q
Inc
Q =
Inc increases
Initial
equilibrium
15
Bad news: On the other hand, since we have no income information, we have no
information about how the demand curve shifts. Without knowing how the demand
curve shifts we have no idea how the equilibrium quantity and price move along the
supply curve. In other words, we cannot trace out the supply curve; hence, we cannot
estimate the slope of the supply curve.

To use the reduced form (RF) approach to estimate the slope of the demand curve, we first use
ordinary least squares (OLS) to estimate the parameters of the reduced form (RF) equations:
Click here to access data {EViewsLink}
Quantity Reduced Form Equation: Dependent Variable: Q
Explanatory Variable: FeedP
Dependent Variable: Q
Method: Least Squares
Sample: 1977M01 1986M12
Included observations: 120

Coefficient Std. Error t-Statistic Prob.

FEEDP -821.8494 131.7644 -6.237266 0.0000
C 239158.3 5777.771 41.39283 0.0000

Table 21.8: EViews Regression Results RF Quantity

Price Reduced Form Equation: Dependent Variable: Q
Explanatory Variable: FeedP
Dependent Variable: P
Method: Least Squares
Sample: 1977M01 1986M12
Included observations: 120

Coefficient Std. Error t-Statistic Prob.

FEEDP 0.524641 0.262377 1.999571 0.0478
C 142.0193 11.50503 12.34411 0.0000

Table 21.9: EViews Regression Results RF Price

Then, we can estimate the slope of the demand curve by calculating the ratio of the feed price
estimates:
Estimated slope of the demand curve = b
S
P
=
a
Q
I
a
P
I
=
821.85
.52464
= 1,566.5

Now, let us use the two stage least squares (TSLS) estimation procedure to estimate the slope of
the demand curve:

Beef Market Demand Model: Dependent variable: Q
Explanatory variable: P
Instrument List: FeedP
Dependent Variable: Q
Method: Two-Stage Least Squares
Sample: 1977M01 1986M12
Included observations: 120
Instrument list: FEEDP

Coefficient Std. Error t-Statistic Prob.

P -1566.499 703.8335 -2.225667 0.0279
C 461631.4 115943.8 3.981510 0.0001

Table 21.10: EViews Regression Results TSLS Demand

In both cases, the estimated slope of the demand curve is 1,566.5.
16
Similarly, an underidentification problem would exist if income information was available, but
feed price information was not.

Demand and Supply Models:
Demand Model: Q
D
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+ e
D
t

Supply Model: Q
S
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S
t

Equilibrium: Q
D
t
= Q
S
t
= Q
t

Endogenous Variables: Q
t
and P
t
Exogenous Variables: FeedP
t
and Inc
t


Reduced Form (RF) Equations
Quantity Reduced Form Equation: Q
t
=
Q
Const
+
Q
FP
FeedP
t
+
Q
I
Inc
t
+
Q
t

Price Reduced Form Equation: P
t
=
P
Const
+
P
FP
FeedP
t
+
P
I
Inc
t
+
P
t

Again, let us now apply the order condition by counting the number of absent exogenous
variables and endogenous explanatory variables in each model:

Number of Less Than Number of
exogenous variables Equal To endogenous explanatory
absent from the model Greater Than variables in the model


Model Model Model
Underidentified Identified Overidentified

No Estimates Unique Estimates Multiple Estimates

Figure 21.8: Order Condition

Demand Model Supply Model
Exogenous Endogenous Exogenous Endogenous
variables explanatory variables explanatory
absent from variables in absent from variables in
the model the model the model the model

None P Inc P
0 1 1 1


The order condition suggests that we should
still be able to estimate the coefficient of the endogenous explanatory variable, P, in the
supply model.
not be able to estimate the coefficient of the endogenous explanatory variable, P, in the
demand model.



17
The coefficient interpretation approach explains why.



Again, there is both good news and bad news when we have income information, but no feed
price information:
Good news: Since we have income information, we still have information about how the
demand curve shifts. The shifts in the demand curve cause the equilibrium quantity and
price to move along the supply curve. In other words, shifts in the demand curve trace
out the supply curve; hence, we can still estimate the slope of the supply curve.
Bad news: On the other hand, since we have no feed price information, we have no
information about how the supply curve shifts. Without knowing how the supply curve
shifts we have no idea how the equilibrium quantity and price move along the demand
curve. In other words, we cannot trace out the demand curve; hence, we cannot
estimate the slope of the demand curve.

Figure 21.9: Reduced Form Summary and Coefficient Interpretation Approach

Changes in the feed price shift the Changes in income shift the
supply curve, but not the demand curve demand curve, but not the supply curve

b
D
P
=
Q
D
P
=
a
Q
FP
FeedP
a
P
FP
FeedP
=
a
Q
FP
a
P
FP
b
S
P
=
Q
S
P
=
a
Q
I
Inc
a
P
I
Inc
=
a
Q
I
a
P
I


The exogenous variable absent The exogenous variable absent
in the demand model, FeedP, in the supply model, Inc,
allows us to estimate the coefficient allows us to estimate the coefficient
of the endogenous explanatory of the endogenous explanatory
variable, P, in the demand model. variable, P, in the demand model.
D
Price
Quantity
FeedP increaes
S
S
P =

F
Q
P
FeedP

F
Q
P
FeedP
Q =
Inc constant
S
Price
Quantity
FeedP constant
D
D
P =

I
Q
Inc

I
Q
Inc
Q =
Inc increases
Initial
equilibrium
Estimated slope of demand curve: Estimated slope of supply curve:
b
D
P
=
Q
D
P
b
S
P
=
Q
S
P


18
To use the reduced form (RF) approach to estimate the slope of the supply curve, we first use
ordinary least squares (OLS) to estimate the parameters of the reduced form (RF) equations:
Click here to access data {EViewsLink}
Quantity Reduced Form Equation: Dependent Variable: Q
Explanatory Variable: Inc

Dependent Variable: Q
Method: Least Squares
Sample: 1977M01 1986M12
Included observations: 120

Coefficient Std. Error t-Statistic Prob.

INC 20.22475 1.902708 10.62946 0.0000
C 111231.3 8733.000 12.73690 0.0000

Table 21.11: EViews Regression Results RF Quantity

Price Reduced Form Equation: Dependent Variable: P
Explanatory Variable: Inc
Dependent Variable: P
Method: Least Squares
Sample: 1977M01 1986M12
Included observations: 120

Coefficient Std. Error t-Statistic Prob.

INC 0.009669 0.004589 2.107161 0.0372
C 120.4994 21.06113 5.721413 0.0000

Table 21.12: EViews Regression Results RF Price

Then, we can estimate the slope of the supply curve by calculating the ratio of the income
estimates:
Estimated slope of the supply curve = b
D
P
=
a
Q
FP
a
P
FP
=
20.225
.009669
= 2,091.7

Once again, two stage least squares (TSLS) provide the same estimate:

Beef Market Supply Model: Dependent variable: Q
Explanatory Variable: P
Instrument List: Inc
Dependent Variable: Q
Method: Two-Stage Least Squares
Sample: 1977M01 1986M12
Included observations: 120
Instrument list: INC

Coefficient Std. Error t-Statistic Prob.

P 2091.679 1169.349 1.788756 0.0762
C -140814.8 192634.8 -0.730994 0.4662

Table 21.13: EViews Regression Results TSLS Supply

Conclusion: When a simultaneous equations model is underidentified, we cannot estimate all its
parameters. For those parameters we can estimate, however, the reduced form estimation procedure
and the two stage least squares (TSLS) estimation procedures are equivalent.
19
Overidentification

While an underidentification problem arises when too little information is available, an
overidentification problem arises when too much information is available. To illustrate this
suppose that in addition to the feed price and income information, the price of chicken is also
available. The simultaneous equation model and the reduced form estimates would become:

Demand and Supply Models:
Demand Model: Q
D
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+
D
CP
ChickP
t
+ e
D
t

Supply Model: Q
S
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S
t

Equilibrium: Q
D
t
= Q
S
t
= Q
t

Endogenous Variables: Q
t
and P
t
Exogenous Variables: FeedP
t
, Inc
t
, and ChickP
t

Reduced Form (RF) Equations
Quantity Reduced Form Equation: Q
t
=
Q
Const
+
Q
FP
FeedP
t
+
Q
I
Inc
t
+
Q
CP
ChickP
t
+
Q
t

Price Reduced Form Equation: P
t
=
P
Const
+
P
FP
FeedP
t
+
P
I
Inc
t
+
P
CP
ChickP
t
+
P
t

Let us now apply the order condition by counting the number of absent exogenous variables and
endogenous explanatory variables in each model:

Number of Less Than Number of
exogenous variables Equal To endogenous explanatory
absent from the model Greater Than variables in the model


Model Model Model
Underidentified Identified Overidentified

No Estimates Unique Estimates Multiple Estimates

Figure 21.10: Order Condition


Demand Model Supply Model
Exogenous Endogenous Exogenous Endogenous
variables explanatory variables explanatory
absent from variables in absent from variables in
the model the model the model the model

FeedP P Inc and ChickP P
1 1 2 1


The order condition suggests that we should
still be able to estimate the coefficient of the endogenous explanatory variable, P, in the
demand model.
should encounter some difficulties when estimating the coefficient of the endogenous
explanatory variable, P, in the supply model.

We shall now explain these difficulties.

20


Now we have two exogenous factors that shift the demand curve: income and the price of
chicken. Consequently, there are two ways to trace out the supply curve. There are now two
different ways to use the reduced form (RF) estimates to estimate the slope of the supply curve:

Ratio of the reduced form Ratio of the reduced form
income coefficients chicken feed coefficients

Estimated slope Estimated slope
of supply curve: of supply curve:

b
S
P
=
a
Q
I
a
P
I
b
S
P
=
a
Q
CP
a
P
CP



Figure 21.11: Reduced Form Summary and Coefficient Interpretation Approach

Changes in income shift the Changes in the chicken price shift the
supply curve, but not the demand curve demand curve, but not the supply curve

b
S
P
=
Q
S
P
=
a
Q
I
Inc
a
P
I
Inc
=
a
Q
I
a
P
I
b
S
P
=
Q
S
P
=
a
Q
I
ChickP
a
P
I
ChickP
=
a
Q
I
a
P
I


The exogenous variable absent The exogenous variable absent
in the demand model, FeedP, in the supply model, Inc,
allows us to estimate the coefficient allows us to estimate the coefficient
of the endogenous explanatory of the endogenous explanatory
variable, P, in the demand model. variable, P, in the demand model.
Estimated slope of demand curve:
b
S
P
=
Q
S
P


S
Price
Quantity
FeedP constant
D
D
P =

I
Q
Inc

I
Q
Inc
Q =
Inc increases
Initial
equilibrium
ChickP constant
S
Price
Quantity
FeedP constant
D
D
P =

F
Q
P
ChickP

C
Q
P
ChickP
Q =
Inc constant
Initial
equilibrium
ChickP increases
21
We shall now go through the mechanics of the reduced form (RF) estimation procedures to
illustrate the overidentification problem. First, we use the ordinary least squares (OLS) to
estimate the reduced form (RF) parameters:
Click here to access data {EViewsLink}
Quantity Reduced Form Equation: Dependent Variable: Q
Explanatory Variables: FeedP, Inc, and ChickP

Dependent Variable: Q
Method: Least Squares
Sample: 1977M01 1986M12
Included observations: 120

Coefficient Std. Error t-Statistic Prob.

FEEDP -349.5411 135.3993 -2.581558 0.0111
INC 16.86458 2.675264 6.303894 0.0000
CHICKP 47.59963 158.4147 0.300475 0.7644
C 138194.2 13355.13 10.34765 0.0000

Table 21.14: EViews Regression Results RF Quantity

Price Reduced Form Equation: Dependent Variable: P
Explanatory Variables: FeedP, Inc, and ChickP
Dependent Variable: P
Method: Least Squares
Sample: 1977M01 1986M12
Included observations: 120

Coefficient Std. Error t-Statistic Prob.

FEEDP 0.955012 0.318135 3.001912 0.0033
INC 0.016043 0.006286 2.552210 0.0120
CHICKP 0.274644 0.372212 0.737870 0.4621
C 29.96187 31.37924 0.954831 0.3416

Table 21.15: EViews Regression Results RF Price

Estimated slope Estimated slope Estimated slope
of demand curve of the supply curve of the supply curve

Ratio of reduced form Ratio of reduced form Ratio of reduced form
feed price income chicken price
coefficient estimates coefficient estimates coefficient estimates

b
D
P
=
a
Q
FP
a
P
FP
=
349.54
.95501
= 366.0 b
S
P
=
a
Q
I
a
P
I
=
16.865
.016043
= 1051.2 b
S
P
=
a
Q
CP
a
P
CP
=
47.600
.27464
= 173.3

The reduced form (RF) estimation procedure produces two different estimates for the slope for
the supply curve. The slope of the supply curve is overidentified.


22
Two-Stage Least Squares (TSLS)

While reduced form (RF) estimation procedure cannot resolve the overidentification problem,
two stage least squares (TSLS) approach can. The two squares least squares estimation procedure
provides a single estimate of the slope of the supply curve. The following regression printout
reveals this:
Click here to access data {EViewsLink}
Beef Market Demand Model: Dependent variable: Q
Explanatory Variables: P, Inc, and ChickP
Instrument List: FeedP, Inc, and ChickP
Dependent Variable: Q
Method: Two-Stage Least Squares
Sample: 1977M01 1986M12
Included observations: 120
Instrument list: FEEDP INC CHICKP

Coefficient Std. Error t-Statistic Prob.

P -366.0071 68.47718 -5.344950 0.0000
INC 22.73632 1.062099 21.40697 0.0000
CHICKP 148.1212 86.30740 1.716205 0.0888
C 149160.5 7899.140 18.88313 0.0000

Table 21.16: EViews Regression Results TSLS Demand

The estimated slope of the demand curve is 366.0. This is the same estimate as
computed by the reduced form (RF) estimation procedure.

Beef Market Supply Model: Dependent variable: Q
Explanatory Variables: P and FeedP
Instrument List: FeedP, Inc, and ChickP
Dependent Variable: Q
Method: Two-Stage Least Squares
Sample: 1977M01 1986M12
Included observations: 120
Instrument list: FEEDP INC CHICKP

Coefficient Std. Error t-Statistic Prob.

P 893.4857 335.0311 2.666874 0.0087
FEEDP -1290.609 364.0891 -3.544761 0.0006
C 112266.0 49592.54 2.263769 0.0254

Table 21.17: EViews Regression Results TSLS Supply

Two stage least squares (TSLS) provides a single estimate for the slope of the supply curve:
b
S
P
= 893.5

23
Overidentification: Comparison of Reduced Form and Two Stage Least Squares Estimates

Table 21.18 compares the estimates that result when using the two different estimation procedures:

Estimated slope Estimated Slope
of demand curve of supply curve
Reduced Form 366.0
Based on Income Coefficients 1051.2
Based on Chicken Price Coefficients 173.3
Two Stage Least Squares 366.0 893.5
Table 21.18: Comparison of RF and TSLS Estimates

Note that the slope of the demand curve is not overidentified; furthermore, both the reduced
form (RF) estimation procedure and the two stage least squares (TSLS) estimation procedure
provide the same estimate. On the other hand, the slope of the supply curve is overidentified.
The reduced form (RF) estimation procedure provides two estimates; the two stage least squares
(TSLS) estimation procedure provides only one.

Summary of Identification Issues: Reduced Form and Two Stage Least Squares
Estimation Procedures

Number of Less Than Number of
exogenous variables Equal To endogenous explanatory
absent from the model Greater Than variables in the model


Model Model Model
Underidentified Identified Overidentified

No Estimates Unique Estimates Multiple Estimates

Figure 21.12: Order Condition

Reduced Form and Two Stage Least Squares Estimation Procedures: A Comparison
Identified: The procedures are equilivalent.
Underidentified: The procedures are equilivalent.
Overidentified: The reduced form estimation procedure produces multiple estimates
while two stage least squares produces a single estimate.

Chapter 21 Review Questions

1. What does it mean for a simultaneous equation model to be underidentified?

2. What does it mean for a simultaneous equation model to be overidentified?

3. Compare the reduced form (RF) estimation procedure and the two stage least squares (TSLS)
estimation procedure:
a. When will the two procedures produce identical results?
b. When will the two procedures produce different results? How do the results differ?

24
Chapter 21 Exercises

The following workfile contains the data we used in class to analyze the beef market:

Beef Market Data: Monthly time series data relating to the market for beef from 1977 to 1986.

Q
t
Quantity of beef in month t (millions of pounds)
P
t
Real price of beef in month t (1982-84 cents per pound)
FeedP
t
Real price of cattle feed in month t (1982-84 cents per pounds of corn cobs)
Inc
t
Real disposable income in month t (thousands of chained 2005 dollars)
ChickP
t
Real rice of whole chickens in month t (1982-84 cents per pound)
Year
t
Year


Consider the following constant elasticity model describing the beef market:
Demand Model: log(Q
D
t
) =
D
Const
+
D
P
log(P
t
)

+
D
I
log(Inc
t
) +
D
CP
log(ChickP
t
) + e
D
t

Supply Model: log(Q
S
t
) =
S
Const
+
S
P
log(P
t
)

+
S
FP
log(FeedP
t
) + e
S
t
Equilibrium: log(Q
D
t
) = log(Q
S
t
) = log(Q
t
)

1. Suppose that there were no data for the price of chicken and income; that is, while you can
include the variable FeedP in your analysis, you cannot use the variables Inc and ChickP.
Click here to access data {EViewsLink}
a. Consider the reduced form (RF) estimation procedure:
1) Can we estimate the own price elasticity of demand,
D
P
? If not, explain why
not. If so, does the reduced form estimation procedure provide a single
estimate? What is (are) the estimate (estimates)?
2) Can we estimate the own price elasticity of supply,
S
P
? If not, explain why
not. If so, does the reduced form estimation procedure provide a single
estimate? What is (are) the estimate (estimates)?

b. Consider the two stage least squares (TSLS) estimation procedure:
1) Can we estimate the own price elasticity of demand,
D
P
? If so, what is (are)
the estimate (estimates)?
2) Can we estimate the own price elasticity of supply,
S
P
? If so, what is (are) the
estimate (estimates)?


25
2. On the other hand, suppose that there were no data for the price of feed; that is, while you
can include the variables Inc and ChickP in your analysis, you cannot use the variable FeedP.
Click here to access data {EViewsLink}
a. Consider the reduced form (RF) estimation procedure:
1) Can we estimate the own price elasticity of demand,
D
P
? If not, explain why
not. If so, does the reduced form estimation procedure provide a single
estimate? What is (are) the estimate (estimates)?
2) Can we estimate the own price elasticity of supply,
S
P
? If not, explain why
not. If so, does the reduced form estimation procedure provide a single
estimate? What is (are) the estimate (estimates)?

b. Consider the two stage least squares (TSLS) estimation procedure:
1) Can we estimate the own price elasticity of demand,
D
P
? If so, what is (are)
the estimate (estimates)?
2) Can we estimate the own price elasticity of supply,
S
P
? If so, what is (are) the
estimate (estimates)?


3. Last, suppose that you can use all the variables in your analysis.
Click here to access data {EViewsLink}
a. Consider the reduced form (RF) estimation procedure:
1) Can we estimate the own price elasticity of demand,
D
P
? If not, explain why
not. If so, does the reduced form estimation procedure provide a single
estimate? What is (are) the estimate (estimates)?
2) Can we estimate the own price elasticity of supply,
S
P
? If not, explain why
not. If so, does the reduced form estimation procedure provide a single
estimate? What is (are) the estimate (estimates)?

b. Consider the two stage least squares (TSLS) estimation procedure:
1) Can we estimate the own price elasticity of demand,
D
P
? If so, what is (are)
the estimate (estimates)?
2) Can we estimate the own price elasticity of supply,
S
P
? If so, what is (are) the
estimate (estimates)?


26
Chicken Market Data: Monthly time series data relating to the market for chicken from 1980 to
1985.

Q
t
Quantity of chicken in month t (millions of pounds)
P
t
Real price of whole chickens in month t (1982-84 cents per pound)
FeedP
t
Real price chicken formula feed in month t (1982-84 cents per pound)
Inc
t
Real disposable income in month t (thousands of chained 2005 dollars)
PorkP
t
Real price of pork in month t (1982-84 cents per pound)
Year
t
Year

Consider the following constant elasticity model describing the beef market:
Demand Model: log(Q
D
t
) =
D
Const
+
D
P
log(P
t
)

+
D
I
log(Inc
t
) +
D
PP
log(PorkP
t
) + e
D
t

Supply Model: log(Q
S
t
) =
S
Const
+
S
P
log(P
t
)

+
S
FP
log(FeedP
t
) + e
S
t
Equilibrium: log(Q
D
t
) = log(Q
S
t
) = log(Q
t
)

4. Suppose that there were no data for the price of pork and income; that is, while you can
include the variable FeedP in your analysis, you cannot use the variables Inc and PorkP.
Click here to access data {EViewsLink}
a. Consider the reduced form (RF) estimation procedure:
1) Can we estimate the own price elasticity of demand,
D
P
? If not, explain why
not. If so, does the reduced form estimation procedure provide a single
estimate? What is (are) the estimate (estimates)?
2) Can we estimate the own price elasticity of supply,
S
P
? If not, explain why
not. If so, does the reduced form estimation procedure provide a single
estimate? What is (are) the estimate (estimates)?

b. Consider the two stage least squares (TSLS) estimation procedure:
1) Can we estimate the own price elasticity of demand,
D
P
? If so, what is (are)
the estimate (estimates)?
2) Can we estimate the own price elasticity of supply,
S
P
? If so, what is (are) the
estimate (estimates)?
27
5. On the other hand, suppose that there were no data for the price of feed; that is, while you
can include the variables Inc and PorkP in your analysis, you cannot use the variable FeedP.
Click here to access data {EViewsLink}
a. Consider the reduced form (RF) estimation procedure:
1) Can we estimate the own price elasticity of demand,
D
P
? If not, explain why
not. If so, does the reduced form estimation procedure provide a single
estimate? What is (are) the estimate (estimates)?
2) Can we estimate the own price elasticity of supply,
S
P
? If not, explain why
not. If so, does the reduced form estimation procedure provide a single
estimate? What is (are) the estimate (estimates)?

b. Consider the two stage least squares (TSLS) estimation procedure:
1) Can we estimate the own price elasticity of demand,
D
P
? If so, what is (are)
the estimate (estimates)?
2) Can we estimate the own price elasticity of supply,
S
P
? If so, what is (are) the
estimate (estimates)?

6. Last, suppose that you use all the variables in your analysis.

Click here to access data {EViewsLink}
a. Consider the reduced form (RF) estimation procedure:
1) Can we estimate the own price elasticity of demand,
D
P
? If not, explain why
not. If so, does the reduced form estimation procedure provide a single
estimate? What is (are) the estimate (estimates)?
2) Can we estimate the own price elasticity of supply,
S
P
? If not, explain why
not. If so, does the reduced form estimation procedure provide a single
estimate? What is (are) the estimate (estimates)?

b. Consider the two stage least squares (TSLS) estimation procedure:
1) Can we estimate the own price elasticity of demand,
D
P
? If so, what is (are)
the estimate (estimates)?
2) Can we estimate the own price elasticity of supply,
S
P
? If so, what is (are) the
estimate (estimates)?

In general, compare the reduced form (RF) estimation procedure and the two stage least squares
(TSLS) estimation procedure.

7. When the reduced form estimation procedure (RF) provides no estimates for a coefficient,
how many estimates does the (TSLS) estimation procedure provide?

8. When the reduced form estimation procedure (RF) provides a single estimate for a
coefficient, how many estimates does the (TSLS) estimation procedure provide? How are the
estimates related?

9. When the reduced form estimation procedure (RF) provides multiple estimates for a
coefficient, how many estimates does the (TSLS) estimation procedure provide?


28
Appendix 21.1 Algebraic Derivation of the Reduced From Equations -
Underidentification

Demand Model: Q
D
t
=
D
Const
+
D
P
P
t
+ e
D
t

Supply Model: Q
S
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S
t
Equilibrium: Q
D
t
= Q
S
t
= Q
t


There are 5 parameters in the demand and supply models:
D
Const
,
D
P
,
S
Const
,
S
P
, and
S
FP
. Ideally,
we would like to estimate them all. Unfortunately, we will not be able to do so. To show this
formally we shall algebraically derive the reduced form equations.


Strategy to derive the reduced form equation for P
t
:
Substitute Q
t
for Q
D
t
and Q
S
t
.
Subtract the equation for the supply model from the equation for the demand model.
Solve for P
t
.

Q
D
t
=
D
Const
+
D
P
P
t
+ e
D
t

Q
S
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S
t
Substitute
Q
t
=
D
Const
+
D
P
P
t
+ e
D
t

Q
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S
t
Subtract
0 =
D
Const

S
Const
+
D
P
P
t

S
P
P
t

S
FP
FeedP
t
+ e
D
t
e
S
t

Solve

S
P
P
t

D
P
P
t
=
D
Const

S
Const

S
FP
FeedP
t
+ e
D
t
e
S
t

(
S
P

D
P
)P
t
=
D
Const

S
Const

S
FP
FeedP
t
+ e
D
t
e
S
t


P
t
=

D
Const

S
Const

S
P

D
P

S
FP

S
P

D
P
FeedP
t
+
e
D
t
e
S
t

S
P

D
P



Strategy to derive the reduced form equation for Q
t
:
Substitute Q
t
for Q
D
t
and Q
S
t
.
Multiply the equation for the demand model by
S
P
and the equation for the supply
model by
D
P
.
Subtract the equation for the supply model from the equation for the demand model.
Solve for Q
t
.

29
Q
D
t
=
D
Const
+
D
P
P
t
+ e
D
t

Q
S
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S

Substitute.
Q
t
=
D
Const
+
D
P
P
t
+ e
D
t

Q
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S

Multiply.

S
P
Q
t
=
S
P

D
Const
+
S
P

D
P
P
t
+
S
P
e
D
t


D
P
Q
t
=
D
P

S
Const
+
D
P

S
P
P
t
+
D
P

S
FP
FeedP
t
+
D
P
e
S

Subtract.

S
P
Q
t

D
P
Q
t
=
S
P

D
Const

D
P

S
Const
+ 0
D
P

S
FP
FeedP
t
+
S
P
e
D
t

D
P
e
S

Solve.
(
S
P

D
P
)Q
t
=
S
P

D
Const

D
P

S
Const

D
P

S
FP
FeedP
t
+
S
P
e
D
t

D
P
e
S


Q
t
=

S
P

D
Const

D
P

S
Const

S
P

D
P

D
P

S
FP

S
P

D
P
FeedP
t
+

S
P
e
D
t

D
P
e
S

S
P

D
P


Compare the reduced form equations for Q
t
and P
t
:

Q
t
=

S
P

D
Const

D
P

S
Const

S
P

D
P

D
P

S
FP

S
P

D
P
FeedP
t
+

S
P
e
D
t

D
P
e
S

S
P

D
P

P
t
=

D
Const

S
Const

S
P

D
P

S
FP

S
P

D
P
FeedP
t
+
e
D
t
e
S
t

S
P

D
P


Next, let the s represent the constants and coefficients of the reduced form (RF) equations:
Q
t
=
Q
Const
+
Q
FP
FeedP
t
+
Q
t


P
t
=
P
Const
+
P
FP
FeedP
t
+
P
t

where the following equations specify the 4 s:

Q
Const
=

S
P

D
Const

D
P

S
Const

S
P

D
P



Q
FP
=

D
P

S
FP

S
P

D
P


P
Const
=

D
Const

S
Const

S
P

D
P



P
FP
=

S
FP

S
P

D
P


There are 5 parameters in the original demand/supply model, 5 unknown s, and only 4
equations specifying the s. We cannot solve for all 5 unknowns with only 4 equations. The
original demand/supply model is underidentified. More specifically, we cannot solve for the
slope of the supply curve,
S
P
; on the other hand, we can solve for the slope of the demand
curve,
D
P
:

Ratio of FeedP
t
coefficients:

Q
FP

P
FP
=

D
P

S
FP

S
P

D
P

S
FP

S
P

D
P

=
D
P
= Slope of the demand curve
30

Appendix 21.2 Algebraic Derivation of the Reduced From Equations -
Underidentification

Demand Model: Q
D
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+ e
D
t

Supply Model: Q
S
t
=
S
Const
+
S
P
P
t
+ e
S
t
Equilibrium: Q
D
t
= Q
S
t
= Q
t


There are 5 parameters in the demand and supply models:
D
Const
,
D
P
,
D
I
,
S
Const
, and
S
P
. Ideally, we
would like to estimate them all. Unfortunately, we will not be able to do so. To show this
formally we shall algebraically derive the reduced form equations.

Strategy to derive the reduced form equation for P
t
:
Substitute Q
t
for Q
D
t
and Q
S
t
.
Subtract the equation for the supply model from the equation for the demand model.
Solve for P
t
.

Q
D
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+ e
D
t

Q
S
t
=
S
Const
+
S
P
P
t
+ e
S
t
Substitute
Q
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+ e
D
t

Q
t
=
S
Const
+
S
P
P
t
+ e
S
t
Subtract
0 =
D
Const

S
Const
+
D
P
P
t

S
P
P
t
+
D
I
Inc
t
+ e
D
t
e
S
t

Solve

S
P
P
t

D
P
P
t
=
D
Const

S
Const


+
D
I
Inc
t
+ e
D
t
e
S
t

(
S
P

D
P
)P
t
=
D
Const

S
Const


+
D
I
Inc
t
+ e
D
t
e
S
t


P
t
=

D
Const

S
Const

S
P

D
P




+

D
I

S
P

D
P
Inc
t
+
e
D
t
e
S
t

S
P

D
P


Strategy to derive the reduced form equation for Q
t
:
Substitute Q
t
for Q
D
t
and Q
S
t
.
Multiply the equation for the demand model by
S
P
and the equation for the supply
model by
D
P
.
Subtract the equation for the supply model from the equation for the demand model.
Solve for Q
t
.

31
Q
D
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+ e
D
t

Q
S
t
=
S
Const
+
S
P
P
t
+ e
S

Substitute.
Q
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+ e
D
t

Q
t
=
S
Const
+
S
P
P
t
+ e
S

Multiply.

S
P
Q
t
=
S
P

D
Const
+
S
P

D
P
P
t
+
S
P

D
I
Inc
t
+
S
P
e
D
t


D
P
Q
t
=
D
P

S
Const
+
D
P

S
P
P
t
+
D
P
e
S

Subtract.

S
P
Q
t

D
P
Q
t
=
S
P

D
Const

D
P

S
Const
+ 0 +
S
P

D
I
Inc
t
+
S
P
e
D
t

D
P
e
S

Solve.
(
S
P

D
P
)Q
t
=
S
P

D
Const

D
P

S
Const


+
S
P

D
I
Inc
t
+
S
P
e
D
t

D
P
e
S


Q
t
=

S
P

D
Const

D
P

S
Const

S
P

D
P




+

S
P

D
I

S
P

D
P
Inc
t
+

S
P
e
D
t

D
P
e
S

S
P

D
P


Compare the reduced form equations for Q
t
and P
t
:

Q
t
=

S
P

D
Const

D
P

S
Const

S
P

D
P




+

S
P

D
I

S
P

D
P
Inc
t
+

S
P
e
D
t

D
P
e
S

S
P

D
P

P
t
=

D
Const

S
Const

S
P

D
P




+

D
I

S
P

D
P
Inc
t
+
e
D
t
e
S
t

S
P

D
P


Next, let the s represent the constants and coefficients of the reduced form (RF) equations:
Q
t
=
Q
Const
+
Q
I
Inc
t
+
Q
t


P
t
=
P
Const
+
P
I
Inc
t
+
P
t

where the following equations specify the 4 s:

Q
Const
=

S
P

D
Const

D
P

S
Const

S
P

D
P



Q
I
=

S
P

D
I

S
P

D
P


P
Const
=

D
Const

S
Const

S
P

D
P



P
I
=

D
I

S
P

D
P


There are 5 parameters in the original demand/supply model, 5 unknown s, and only 4
equations specifying the s. We cannot solve for all 5 unknowns with only 4 equations. The
original demand/supply model is underidentified.
More specifically, we cannot solve for the slope of the supply curve,
D
P
; on the other hand, we
can solve for the slope of the demand curve,
S
P
:

Ratio of Inc
t
coefficients:

Q
I

P
I
=

S
P

D
I

S
P

D
P

D
I

S
P

D
P

=
S
P
= Slope of the supply curve
32

Appendix 21.3 Algebraic Derivation of the Reduced From Equations -
Overidentification

Demand Model: Q
D
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+
D
CP
ChickP
t
+ e
D
t

Supply Model: Q
S
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S
t

Equilibrium: Q
D
t
= Q
S
t
= Q
t


There are 7 parameters in the demand and supply models:
D
Const
,
D
P
,
D
I
,
D
CP
,
S
Const
,
S
P
, and
S
FP
.
These models are overidentified meaning that (at least) one of the parameters can be estimated in
two different ways. To show this formally we shall algebraically derive the reduced form
equations.

Strategy to derive the reduced form equation for P
t
:
Substitute Q
t
for Q
D
t
and Q
S
t
.
Subtract the equation for the supply model from the equation for the demand model.
Solve for P
t
.


Q
D
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+
D
CP
ChickP
t
+ e
D
t

Q
S
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S
t
Substitute
Q
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+
D
CP
ChickP
t
+ e
D
t

Q
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S
t
Subtract
0 =
D
Const

S
Const
+
D
P
P
t

S
P
P
t

S
FP
FeedP
t
+
D
I
Inc
t
+
D
CP
ChickP
t
+ e
D
t
e
S
t

Solve

S
P
P
t

D
P
P
t
=
D
Const

S
Const

S
FP
FeedP
t
+
D
I
Inc
t
+
D
CP
ChickP
t
+ e
D
t
e
S
t

(
S
P

D
P
)P
t
=
D
Const

S
Const

S
FP
FeedP
t
+
D
I
Inc
t
+
D
CP
ChickP
t
+ e
D
t
e
S
t


P
t
=

D
Const

S
Const

S
P

D
P

S
FP

S
P

D
P
FeedP
t
+

D
I

S
P

D
P
Inc
t
+

D
CP

S
P

D
P
ChickP
t
+
e
D
t
e
S
t

S
P

D
P

33
Strategy to derive the reduced form equation for Q
t
:
Substitute Q
t
for Q
D
t
and Q
S
t
.
Multiply the equation for the demand model by
S
P
and the equation for the supply
model by
D
P
.
Subtract the equation for the supply model from the equation for the demand model.
Solve for Q
t
.

Q
D
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+
D
CP
ChickP
t
+ e
D
t

Q
S
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S
t
Substitute.
Q
t
=
D
Const
+
D
P
P
t
+
D
I
Inc
t
+
D
CP
ChickP
t
+ e
D
t

Q
t
=
S
Const
+
S
P
P
t
+
S
FP
FeedP
t
+ e
S
t
Multiply.

S
P
Q
t
=
S
P

D
Const
+
S
P

D
P
P
t
+
S
P

D
I
Inc
t
+
S
P

D
CP
ChickP
t
+
S
P
e
D
t


D
P
Q
t
=
D
P

S
Const
+
D
P

S
P
P
t
+
D
P

S
FP
FeedP
t
+
D
P
e
S
t
Subtract.

S
P
Q
t

D
P
Q
t
=
S
P

D
Const

D
P

S
Const
+ 0
D
P

S
FP
FeedP
t
+
S
P

D
I
Inc
t
+
S
P

D
CP
ChickP
t
+
S
P
e
D
t

D
P
e
Solve.
(
S
P

D
P
)Q
t
=
S
P

D
Const

D
P

S
Const

D
P

S
FP
FeedP
t
+
S
P

D
I
Inc
t
+
S
P

D
CP
ChickP
t
+
S
P
e
D
t

D
P
e
Divide.
Q
t
=

S
P

D
Const

D
P

S
Const

S
P

D
P

D
P

S
FP

S
P

D
P
FeedP
t
+

S
P

D
I

S
P

D
P
Inc
t
+

S
P

D
CP

S
P

D
P
ChickP
t
+

S
P
e
D
t

D
P
e

S
P

D
P

Compare the reduced form equations for Q
t
and P
t
:

Q
t
=

S
P

D
Const

D
P

S
Const

S
P

D
P

D
P

S
FP

S
P

D
P
FeedP
t
+

S
P

D
I

S
P

D
P
Inc
t
+

S
P

D
CP

S
P

D
P
ChickP
t
+

S
P
e
D
t

D
P
e

S
P

D
P
P
t
=

D
Const

S
Const

S
P

D
P

S
FP

S
P

D
P
FeedP
t
+

D
I

S
P

D
P
Inc
t
+

D
CP

S
P

D
P
ChickP
t
+
e
D
t
e
S
t

S
P

D
P


Next, let the s represent the constants and coefficients of the reduced form (RF) equations:

Q
t
=
Q
Const

Q
FP
FeedP
t
+
Q
I
Inc
t
+
Q
CP
ChickP
t
+
Q
t


P
t
=
P
Const

P
FP
FeedP
t
+
P
I
Inc
t
+
P
CP
ChickP
t
+
P
t


where the following equations specify the 8 s:


Q
Const
=

S
P

D
Const

D
P

S
Const

S
P

D
P

Q
FP
=

D
P

S
FP

S
P

D
P

Q
I
=

S
P

D
I

S
P

D
P

Q
CP
=

S
P

D
CP

S
P

D
P


P
Const
=

D
Const

S
Const

S
P

D
P

P
FP
=

S
FP

S
P

D
P

P
I
=

D
I

S
P

D
P

P
CP
=

D
CP

S
P

D
P


34
There are 7 parameters in the original demand/supply model, 7 unknown s, and 8 equations
specifying the s. We have more equations than unknowns. The original demand/supply model
is overidentified. More specifically, we can solve for the slope of the supply curve,
S
P
, in two
ways:

Ratio of Inc
t
coefficients:

Q
I

P
I
=

S
P

D
I

S
P

D
P

D
I

S
P

D
P

=
S
P
= Slope of the supply curve


Ratio of ChickP
t
coefficients:

Q
CP

P
CP
=

S
P

D
CP

S
P

D
P

D
CP

S
P

D
P

=
S
P
= Slope of the supply curve