Econometrics Module

lOMoARcPSD|6162234
Econometrics Module
Introduction to Econometrics (Mekelle

University)
Scan to open on Studocu
Downloaded by Demlie Basha Hailu

Studocu is not sponsored or endorsed by any college or university

Prepared by: Bedru B. and Seid H. ( June,
ECONOMETRICS
A TEACHING MATERIAL FOR DISTANCE STUDENTS
MAJORING IN ECONOMICS
Prepared By:
Module
Bedru I
Babulo Seid
Hasseл
Department of Economics
Faculty of Business and
Economics
Mekelle University
June, 2005
Mekelle
Econometrics, Module I 1

Econometrics
Module I
Introduction to the
The principal objective of the course, “Introduction to Econometrics”, is to provide an
elementary but comprehensive introduction to the art and science of econometrics. It enables
students to see how economic theory, statistical and mathematical methods are combined in the
analysis of economic data, with a purpose of giving empirical content to economic theories and
verify or refute them.
Module I of the course includes the first three chapters. The first chapter introduces students
with the definition and some fundamental conceptualization of econometrics. In chapter two a
fairly detailed treatment of the simple classical linear regression model will be made. In this
chapter students will be introduced with the basic logic, concepts, assumptions, estimation
methods, and interpretations of the simple classical linear regression models and their
applications in economic science. Chapter three, which deals with Multiple Regression Models, is
basically the extension of the simple regression models. But in chapter three attempts will be
made to expand the linear regression model by incorporating more than one explanatory variables
or regressors to the model. In both chapters (chapter one and chapter two), due attention will be
given to the basics of ordinary least square (OLS) method of estimation and investigating the
statistical properties of the parameter estimates which are summarized by the Gauss-Markov's
BLUE (Best, Linear, Unbiased, estimator) properties.
Chapter 1. Introduction
Contents of the Module in
 Definition and scope of econometrics
 Economic models vs. econometric models
 Methodology of econometrics
 Desirable properties of an econometric model
 Goals of econometrics
Chapter 2. The Classical Regression Analysis: The Simple Linear regression Models
 Stochastic and non-stochastic relationships
 The Simple Regression model
 The basic Assumptions the Classical Regression Model
 OLS Method of Estimation
 Properties of OLS Estimators
 Inferences/Predictions
Chapter 3. The Classical Regression Analysis: The Multiple Linear Regression Models
 Assumptions
 Ordinary Least Squares (OLS) estimation
 Matrix Approach to Multiple Regression Model
 Properties of the OLS estimators
 Inferences/Predictions

Chapter One
Introduction
1.1 Definition and scope of econometrics
The economic theories we learn in various economics courses suggest many

relationships among economic variables. For instance, in microeconomics we
learn demand and supply models in which the quantities demanded and supplied
of a good depend on its price. In macroeconomics, we study ‘investment function'
to explain the amount of aggregate investment in the economy as the rate of
interest changes; and ‘consumption function' that relates aggregate consumption
to the level of aggregate disposable income.
Each of such specifications involves a relationship among economic variables. As

economists, we may be interested in questions such as: If one variable changes in
a certain magnitude, by how much will another variable change? Also, given that
we know the value of one variable; can we forecast or predict the corresponding
value of another? The purpose of studying the relationships among economic
variables and attempting to answer questions of the type raised here, is to help us
understood the real economic world we live in.
However, economic theories that postulate the relationships between economic

variables have to be checked against data obtained from the real world. If
empirical data verify the relationship proposed by economic theory, we accept the
theory as valid. If the theory is incompatible with the observed behavior, we either
reject the theory or in the light of the empirical evidence of the data, modify the
theory. To provide a better understanding of economic relationships and a better

guidance for economic policy making we also need to know the quantitative
relationships between the different economic variables. We obtain these
quantitative measurements taken from the real world. The field of knowledge
which helps us to carryout such an evaluation of economic theories in empirical
terms is econometrics.
Distance students! Having said the background statement in our attempt for
defining ‘ECONOMETRICS', we may now formally define what econometrics is.
WHAT IS ECONOMETRICS?
Literally interpreted, econometrics means “economic measurement”, but the scope
of econometrics is much broader as described by leading econometricians. Various
econometricians used different ways of wordings to define econometrics. But if
we distill the fundamental features/concepts of all the definitions, we may obtain
the following definition.
“Econometrics is the science which integrates economic theory, economic

statistics, and mathematical economics to investigate the empirical support of the
general schematic law established by economic theory. It is a special type of
economic analysis and research in which the general economic theories,
formulated in mathematical terms, is combined with empirical measurements of
economic phenomena. Starting from the relationships of economic theory, we
express them in mathematical terms so that they can be measured. We then use
specific methods, called econometric methods in order to obtain numerical
estimates of the coefficients of the economic relationships.”
Measurement is an important aspect of econometrics. However, the scope of

econometrics is much broader than measurement. As D.Intriligator rightly stated

the “metric” part of the word econometrics signifies ‘measurement', and hence
econometrics is basically concerned with measuring of economic relationships.
In short, econometrics may be considered as the integration of economics,

mathematics, and statistics for the purpose of providing numerical values for the
parameters of economic relationships and verifying economic theories.
1.2 Econometrics vs. mathematical economics

Mathematical economics states economic theory in terms of mathematical
symbols. There is no essential difference between mathematical economics and
economic theory. Both state the same relationships, but while economic theory
use verbal exposition, mathematical symbols. Both express economic
relationships in an exact or deterministic form. Neither mathematical economics
nor economic theory allows for random elements which might affect the
relationship and make it stochastic. Furthermore, they do not provide numerical
values for the coefficients of economic relationships.
Econometrics differs from mathematical economics in that, although econometrics

presupposes, the economic relationships to be expressed in mathematical forms, it
does not assume exact or deterministic relationship. Econometrics assumes
random relationships among economic variables. Econometric methods are
designed to take into account random disturbances which relate deviations from
exact behavioral patterns suggested by economic theory and mathematical
economics. Further more, econometric methods provide numerical values of the
coefficients of economic relationships.
1.3 Econometrics vs. statistics

Econometrics differs from both mathematical statistics and economic statistics.
An economic statistician gathers empirical data, records them, tabulates them or

charts them, and attempts to describe the pattern in their development over time
and perhaps detect some relationship between various economic magnitudes.
Economic statistics is mainly a descriptive aspect of economics. It does not
provide explanations of the development of the various variables and it does not
provide measurements the coefficients of economic relationships.
Mathematical (or inferential) statistics deals with the method of measurement

which are developed on the basis of controlled experiments. But statistical
methods of measurement are not appropriate for a number of economic
relationships because for most economic relationships controlled or carefully
planned experiments cannot be designed due to the fact that the nature of
relationships among economic variables are stochastic or random. Yet the
fundamental ideas of inferential statistics are applicable in econometrics, but they
must be adapted to the problem economic life. Econometric methods are adjusted
so that they may become appropriate for the measurement of economic
relationships which are stochastic. The adjustment consists primarily in specifying
the stochastic (random) elements that are supposed to operate in the real world and
enter into the determination of the observed data.
1.4 Economic models vs. econometric models

i) Economic models:
Any economic theory is an observation from the real world. For one reason, the
immense complexity of the real world economy makes it impossible for us to
understand all interrelationships at once. Another reason is that all the
interrelationships are not equally important as such for the understanding of the
economic phenomenon under study. The sensible procedure is therefore, to pick
up the important factors and relationships relevant to our problem and to focus our
attention on these alone. Such a deliberately simplified analytical framework is
called on economic model. It is an organized set of relationships that describes the

functioning of an economic entity under a set of simplifying assumptions. All

economic reasoning is ultimately based on models. Economic models consist of
the following three basic structural elements.
1. A set of variables
2. A list of fundamental relationships and
3. A number of strategic coefficients
ii) Econometric models:

The most important characteristic of economic relationships is that they contain a
random element which is ignored by mathematical economic models which
postulate exact relationships between economic variables.
Example: Economic theory postulates that the demand for a commodity depends
on its price, on the prices of other related commodities, on consumers' income and
on tastes. This is an exact relationship which can be written mathematically as:
Q  b0  b1 P  b2 P0  b3Y  b4t
The above demand equation is exact. How ever, many more factors may affect
demand. In econometrics the influence of these ‘other' factors is taken into
account by the introduction into the economic relationships of random variable. In
our example, the demand function studied with the tools of econometrics would be
of the stochastic form:
Q  b0  b1 P  b2 P0  b3Y  b4 t  u
where u stands for the random factors which affect the quantity demanded.
1.5 Methodology of econometrics

Econometric research is concerned with the measurement of the parameters of
economic relationships and with the predication of the values of economic
variables. The relationships of economic theory which can be measured with

econometric techniques are relationships in which some variables are postulated as

causes of the variation of other variables. Starting with the postulated theoretical
relationships among economic variables, econometric research or inquiry
generally proceeds along the following lines/stages.
1. Specification the model
2. Estimation of the model
3. Evaluation of the estimates
4. Evaluation of he forecasting power of the estimated model
1. Specification of the model

In this step the econometrician has to express the relationships between economic
variables in mathematical form. This step involves the determination of three
important tasks:
i) the dependent and independent (explanatory) variables which
will be included in the model.
ii) the a priori theoretical expectations about the size and sign of the
parameters of the function.
iii) the mathematical form of the model (number of equations,
specific form of the equations, etc.)
Note: The specification of the econometric model will be based on economic

theory and on any available information related to the phenomena under
investigation. Thus, specification of the econometric model presupposes
knowledge of economic theory and familiarity with the particular phenomenon
being studied.
Specification of the model is the most important and the most difficult stage of any
econometric research. It is often the weakest point of most econometric
applications. In this stage there exists enormous degree of likelihood of

committing errors or incorrectly specifying the model. Some of the common

reasons for incorrect specification of the econometric models are:
1. the imperfections, looseness of statements in economic theories.
2. the limitation of our knowledge of the factors which are operative in any
particular case.
3. the formidable obstacles presented by data requirements in the
estimation of large models.
The most common errors of specification are:
a. Omissions of some important variables from the function.
b. The omissions of some equations (for example, in simultaneous
equations model).
c. The mistaken mathematical form of the functions.
2. Estimation of the model

This is purely a technical stage which requires knowledge of the various
econometric methods, their assumptions and the economic implications for the
estimates of the parameters. This stage includes the following activities.
a. Gathering of the data on the variables included in the
model.
b. Examination of the identification conditions of the function
(especially for simultaneous equations models).
c. Examination of the aggregations problems involved in the
variables of the function.
d. Examination of the degree of correlation between the
explanatory variables (i.e. examination of the problem of
multicollinearity).
e. Choice of appropriate economic techniques for estimation,
i.e. to decide a specific econometric method to be applied
in estimation; such as, OLS, MLM, Logit, and Probit.

3. Evaluation of the estimates

This stage consists of deciding whether the estimates of the parameters are
theoretically meaningful and statistically satisfactory. This stage enables the
econometrician to evaluate the results of calculations and determine the reliability
of the results. For this purpose we use various criteria which may be classified
into three groups:
i. Economic a priori criteria: These criteria are determined by economic
theory and refer to the size and sign of the parameters of economic
relationships.
ii. Statistical criteria (first-order tests): These are determined by statistical
theory and aim at the evaluation of the statistical reliability of the
estimates of the parameters of the model. Correlation coefficient test,
standard error test, t-test, F-test, and R2-test are some of the most
commonly used statistical tests.
iii. Econometric criteria (second-order tests): These are set by the theory
of econometrics and aim at the investigation of whether the assumptions
of the econometric method employed are satisfied or not in any
particular case. The econometric criteria serve as a second order test (as
test of the statistical tests) i.e. they determine the reliability of the
statistical criteria; they help us establish whether the estimates have the
desirable properties of unbiasedness, consistency etc. Econometric
criteria aim at the detection of the violation or validity of the
assumptions of the various econometric techniques.
4) Evaluation of the forecasting power of the model:

Forecasting is one of the aims of econometric research. However, before using
an estimated model for forecasting by some way or another the predictive
power of the model. It is possible that the model may be economically

meaningful and statistically and econometrically correct for the sample period
for which the model has been estimated; yet it may not be suitable for
forecasting due to various factors (reasons). Therefore, this stage involves the
investigation of the stability of the estimates and their sensitivity to changes in
the size of the sample. Consequently, we must establish whether the estimated
function performs adequately outside the sample of data. i.e. we must test an
extra sample performance the model.
1.6 Desirable properties of an econometric model

An econometric model is a model whose parameters have been estimated with
some appropriate econometric technique. The ‘goodness' of an econometric
model is judged customarily according to the following desirable properties.
1. Theoretical plausibility. The model should be compatible with the

postulates of economic theory. It must describe adequately the economic
phenomena to which it relates.
2. Explanatory ability. The model should be able to explain the observations

of he actual world. It must be consistent with the observed behaviour of the
economic variables whose relationship it determines.
3. Accuracy of the estimates of the parameters. The estimates of the

coefficients should be accurate in the sense that they should approximate as
best as possible the true parameters of he structural model. The estimates
should if possible possess the desirable properties of unbiasedness,
consistency and efficiency.

4. Forecasting ability. The model should produce satisfactory predictions of

future values of he dependent (endogenous) variables.
5. Simplicity. The model should represent the economic relationships with

maximum simplicity. The fewer the equations and the simpler their
mathematical form, the better the model is considered, ceteris paribus (that
is to say provided that the other desirable properties are not affected by the
simplifications of the model).
1.7 Goals of Econometrics

Three main goals of Econometrics are identified:
i) Analysis i.e. testing economic theory
ii) Policy making i.e. Obtaining numerical estimates of the
coefficients of economic relationships for policy simulations.
iii) Forecasting i.e. using the numerical estimates of the coefficients
in order to forecast the future values of economic magnitudes.
Review questions
 How would you define econometrics?

 How does it differ from mathematical economics
and statistics?
 Describe the main steps involved in any
econometrics research.
 Differentiate between economic and econometric model.
 What are the goals of econometrics?

Chapter Two
THE CLASSICAL REGRESSION ANALYSIS
[The Simple Linear Regression Model]
Economic theories are mainly concerned with the relationships among various
economic variables. These relationships, when phrased in mathematical terms, can
predict the effect of one variable on another. The functional relationships of these
variables define the dependence of one variable upon the other variable (s) in the
specific form. The specific functional forms may be linear, quadratic, logarithmic,
exponential, hyperbolic, or any other form.
In this chapter we shall consider a simple linear regression model, i.e. a

relationship between two variables related in a linear form. We shall first discuss
two important forms of relation: stochastic and non-stochastic, among which we
shall be using the former in econometric analysis.
2.1. Stochastic and Non-stochastic Relationships
A relationship between X and Y, characterized as Y = f(X) is said to be

deterministic or non-stochastic if for each value of the independent variable (X)
there is one and only one corresponding value of dependent variable (Y). On the
other hand, a relationship between X and Y is said to be stochastic if for a
particular value of X there is a whole probabilistic distribution of values of Y. In
such a case, for any given value of X, the dependent variable Y assumes some
specific value only with some probability. Let's illustrate the distinction between
stochastic and non stochastic relationships with the help of a supply function.

Assuming that the supply for a certain commodity depends on its price (other
determinants taken to be constant) and the function being linear, the relationship
can be put as:
Q  f (P)    P                            (2.1)
The above relationship between P and Q is such that for a particular value of P,
there is only one corresponding value of Q. This is, therefore, a deterministic (non-
stochastic) relationship since for each price there is always only one corresponding
quantity supplied. This implies that all the variation in Y is due solely to changes
in X, and that there are no other factors affecting the dependent variable.
If this were true all the points of price-quantity pairs, if plotted on a two-
dimensional plane, would fall on a straight line. However, if we gather
observations on the quantity actually supplied in the market at various prices and
we plot them on a diagram we see that they do not fall on a straight line.
The derivation of the observation from the line may be attributed to several
factors.
a. Omission of variables from the function
b. Random behavior of human beings
c. Imperfect specification of the mathematical form of the model
d. Error of aggregation
e. Error of measurement

In order to take into account the above sources of errors we introduce in

econometric functions a random variable which is usually denoted by the letter ‘u'
or ‘ ' and is called error term or random disturbance or stochastic term of the
function, so called be cause u is supposed to ‘disturb' the exact linear relationship
which is assumed to exist between X and Y. By introducing this random variable
in the function the model is rendered stochastic of the form:
Yi    X  ui............................................................................................................................................................................ (2.2)
Thus a stochastic model is a model in which the dependent variable is not only
determined by the explanatory variable(s) included in the model but also by others
which are not included in the model.
2.2. Simple Linear Regression model.
The above stochastic relationship (2.2) with one explanatory variable is called
simple linear regression model.
The true relationship which connects the variables involved is split into two parts:
a part represented by a line and a part represented by the random term ‘u'.

The scatter of observations represents the true relationship between Y and X. The
line represents the exact part of the relationship and the deviation of the
observation from the line represents the random component of the relationship.
- Were it not for the errors in the model, we would observe all the points on the
line Y ' ,Y ' ,......,Y ' corresponding to X , X ,.. ,
X . However because of the random
1 2 n 1 2 n
disturbance, we observe Y ,Y ,......,Y corresponding to X , X ,...., X . These

1 2 n 1 2 n
points diverge from the regression line by u1 , u2 ,...., un .
⏟
Y
⏟i
   xi u
the dependent
iable
i var 
random var iable
–_–,
the regression line
- The first component in the bracket is the part of Y explained by the changes
in X and the second is the part of Y not explained by X, that is to say the
change in Y is due to the random influence of u i .
2.2.1 Assumptions of the Classical Linear Stochastic Regression Model.
The classicals made important assumption in their analysis of regression .The most
importat of these assumptions are discussed below.
1. The model is linear in parameters.

The classicals assumed that the model should be linear in the parameters
regardless of whether the explanatory and the dependent variables are linear or
not. This is because if the parameters are non-linear it is difficult to estimate them
since their value is not known but you are given with the data of the dependent and
independent variable.
Example 1. Y    x  is linear in both parameters and the variables, so it

u
Satisfies the assumption


2. ln Y     ln x  u is linear only in the parameters. Since the

the classicals worry on the parameters, the model satisfies
the assumption.
Dear distance students! Check yourself whether the following models satisfy the
above assumption and give your answer to your tutor.
a. ln Y 2     ln X 2 U
i
b. Yi   X i U i
2. U i is a random real variable
This means that the value which u may assume in any one period depends on
chance; it may be positive, negative or zero. Every value has a certain probability
of being assumed by u in any particular instance.
2. The mean value of the random variable(U) in any particular period is

zero
This means that for each value of x, the random variable(u) may assume
various values, some greater than zero and some smaller than zero, but if we
considered all the possible and negative values of u, for any given value of X,
they would have on average value equal to zero. In other words the positive
and negative values of u cancel each other.
Mathematically, E(Ui )  0............................................................... (2.3)

3. The variance of the random variable(U) is constant in each period

(The assumption of homoscedasticity)
For all values of X, the u's will show the same dispersion around their mean.
In Fig.2.c this assumption is denoted by the fact that the values that u can
assume lie with in the same limits, irrespective of the value of X. For X 1 , u
can assume any value with in the range AB; for X 2 , u can assume any value
with in the range CD which is equal to AB and so on.

Graphically;
Mathematically;
Var(U )  E[U  E(U )]2  E(U )2   2 (Since E(U )  0 ).This constant variance is
i
i i i
called homoscedasticity assumption and the constant variance itself is called

homoscedastic variance.

4. The random variable (U) has a normal distribution

This means the values of u (for each x) have a bell shaped symmetrical
distribution about their zero mean and constant variance 2 , i.e.
U  N(0, 2 ).......................................................................... 2.4
i
5. The random terms of different observations U ,U are

i
independent. (The assumption of no autocorrelation)
This means the value which the random term assumed in one period does not
depend on the value which it assumed in any other period.
Algebraically,

Cov(u u )   [(u  (u )][u  (u )] 
i j i i j
 E(uiu j )  0.................................................... (2.5)
6. The X are a set of fixed values in the hypothetical process of repeated

i
sampling which underlies the linear regression model.

- This means that, in taking large number of samples on Y and X, the X i
values are the same in all samples, but the
ui values do differ from sample
to sample, and so of course do the values of
yi .
7. The random variable (U) is independent of the explanatory variables.

This means there is no correlation between the random variable and the
explanatory variable. If two variables are unrelated their covariance is
zero.
Hence Cov( X i ,U i )  0........................................................................... (2.6)
Proof:-
cov( XU )  [( X i  ( X i )][U i  (Ui )]

 [( X i  ( X i )(U i
given E(Ui )  0
)]
 ( X iUi )  ( X i )(Ui )
 ( X iUi )
 X i (Ui ) , given that

the xi are fixed
0
8. The explanatory variables are measured without error
- U absorbs the influence of omitted variables and possibly errors of
measurement in the y's. i.e., we will assume that the regressors are error
free, while y values may or may not include errors of measurement.
Dear students! We can now use the above assumptions to derive the following
basic concepts.
A. The dependent variable Yi

is normally distributed.
i.e Yi ~ N(  x ), ………………………………(2.7)
2
Proof:
Mean: (Y )    xi  ui 
   X
i Since (ui )  0
Variance:
Var(Y )  Y  (Y ) 
2
i i i
      X  u  (   X ) 
2
i i
 (u ) 2
i
 2
(since (u ) 2   2 )
i
..................................................................
 var(Y )   2 (2.8)
i
The shape of the distribution of is determined by the shape of the distribution of

Yi
ui which is normal by assumption 4. Since and

 , being constant, they don't

affect the distribution of

yi . Furthermore, the values of the explanatory variable,
xi , are a set of fixed values by assumption 5 and therefore don't affect the shape of
the distribution
of yi .
~ N(  x , 2 )
Yi i
B. successive values of the dependent variable are independent, i.e

Cov(Yi ,Y j )  0
Proof:
Cov(Yi ,Y j )  E{[Yi  E(Yi )][Y j  E(Yj )]}
 E{[  X i  U i  E(  X i  U i )][  X j  U j  E(  X j  U j )}
(Since Yi    X i 
Ui andYj    X j  U j )
= E[(  X i  Ui    X i )(  X j  U j    X j )] ,Since (ui )  0

 E(UiU j )  (from equation (2.5))
0
Therefore, Cov(Y Y )  0 .
i, j
2.2.2 Methods of estimation

Specifying the model and stating its underlying assumptions are the first stage of
any econometric application. The next step is the estimation of the numerical
values of the parameters of economic relationships. The parameters of the simple
linear regression model can be estimated by various methods. Three of the most
commonly used methods are:
1. Ordinary least square method (OLS)
2. Maximum likelihood method (MLM)
3. Method of moments (MM)
But, here we will deal with the OLS and the MLM methods of estimation.


2.2.2.1 The ordinary least square (OLS) method

The model
Yi     X i  is called the true relationship between Y and X
Ui
because Y and X represent their respective population value, and

 and  are
called the true parameters since they are estimated from the population value of Y
and X But it is difficult to obtain the population value of Y and X because of
technical or economic reasons. So we are forced to take the sample value of Y and
X. The parameters estimated from the sample value of Y and X are called the
estimators of the true parameters  and 
and are symbolized as ˆ ˆ .
and
The model Yi  ˆ  ˆX i  ei , is called estimated relationship between Y and X since

ˆ and
are estimated from the sample of Y and X and ei represents the sample
ˆ
counterpart of the population random disturbance U i .
Estimation of  and  by least square method (OLS) or classical least square
(CLS) involves finding values for the estimates ˆ and

which will minimize the
ˆ
sum of square of the squared residuals (  e 2 ).

i
From the estimated relationship Yi  ˆ  ˆX i  ei , we obtain:
ei  Yi  (ˆ  ˆX i )............................................. (2.6)
ˆ e 2
 (Y  ˆ 
) 2 ……………………….(2.7)
X
i i i
To find the values of ˆ and ˆ that minimize this sum, we have to partially
differentiate e 2
with respect to ˆ and ˆ
and set the partial derivatives equal to
i
zero.
e2
1. i
 2 (Yi i ˆ  ˆX )  0.......................................................(2.8)
 ˆ
Rearranging this expression we will get: If you divide (2.9) by
Y i  n 
‘n' and rearrange, we
get ……(2.9)
ˆ  Y  ˆX................................................................................(2.10)

e2
2. i  2 X (Y  ˆ  ˆX )  0..................................................(2.11)
ˆ i
i
Note: at this point that the term in the parenthesis in equation 2.8and 2.11 is the
residual,
e  Yi  ˆ  ˆX i . Hence it is possible to rewrite (2.8) and (2.11) as
 2 ei 
and  2 X i  0 . It follows that;
0
ei
e i  0 and X i ei  0 .......................................(2.12)
If we rearrange equation (2.11) we obtain;
Y Xii  ˆX
i
ˆ X
 2
i ……………………………………….(2.13)
Equation (2.9) and (2.13) are called the Normal Equations. Substituting the
values of ˆ from (2.10) to (2.13), we get:
Y Xi i  X (Y  ˆX )  ˆX 2
i i
 YX i
 ˆXX  ˆX 2
i
i
Y Xi i
 YX i  ˆ(X 2  XX )
i i
XY  nXY
= ˆ ( X 2  nX 2)
i
ˆ   XY n X Y ………………….(2.14)
X 2 nX 2
i
Equation (2.14) can be rewritten in somewhat different way as follows;

( X  X )(Y  Y )  ( XY  XY  XY  XY )
 XY  YX  XY  nXY

 XY  nYX  nXY  nXY
( X  X )(Y  Y )  XY  n X Y              (2.15)
( X  X )2  X 2  nX 2                  (2.16)
Substituting (2.15) and (2.16) in (2.14), we get


( X  X )(Y  Y )
ˆ  ( X  X ) 2
Now, denoting
(Xi X) xi , (Yi  Y ) as yi we get;
as and
ˆ   x i y i ……………………………………… (2.17)
 x i2
The expression in (2.17) to estimate the parameter coefficient is termed is the

formula in deviation form.
2.2.2.2 Estimation of a function with zero intercept

Suppose it is desired to fit the lineYi    X i U i , subject to the restriction
  0. To estimate ˆ , the problem is put in a form of restricted minimization

problem and then Lagrange method is applied.
We
minimize: e2  n(Y  ˆ  ˆX ) 2
i i i
i1
Subject to: ˆ  0
The composite function then becomes
Z
 ˆ  )2  where  is a Lagrange multiplier.
 (Yi  ˆX i ˆ ,
We minimize the function with respect to ˆ, ˆ, and 

Z
   ˆ  )    0        (i)
ˆ 2(Yi  ˆX i

Z   ˆ  ) ( X )  0        (ii)
 2(Yi  ˆX i i
ˆ
z
 2  0                    (iii)

Substituting (iii) in (ii) and rearranging we obtain:
X i (Yi  ˆX i )  0
Yi X i  ˆX 2  0
i

ˆ  X iYi
……………………………………..(2.18)
Xi 2
This formula involves the actual values (observations) of the variables and not
their deviation forms, as in the case of unrestricted value of ˆ .
2.2.2.3. Statistical Properties of Least Square Estimators

There are various econometric methods with which we may obtain the estimates of
the parameters of economic relationships. We would like to an estimate to be as
close as the value of the true population parameters i.e. to vary within only a small
range around the true parameter. How are we to choose among the different
econometric methods, the one that gives ‘good' estimates? We need some criteria
for judging the ‘goodness' of an estimate.
‘Closeness' of the estimate to the population parameter is measured by the mean

and variance or standard deviation of the sampling distribution of the estimates of
the different econometric methods. We assume the usual process of repeated
sampling i.e. we assume that we get a very large number of samples each of size
‘n'; we compute the estimates ˆ 's from each sample, and for each econometric
method and we form their distribution. We next compare the mean (expected
value) and the variances of these distributions and we choose among the
alternative estimates the one whose distribution is concentrated as close as
possible around the population parameter.
PROPERTIES OF OLS ESTIMATORS

The ideal or optimum properties that the OLS estimates possess may be
summarized by well known theorem known as the Gauss-Markov Theorem.
Statement of the theorem: “Given the assumptions of the classical linear regression model, the
OLS estimators, in the class of linear and unbiased estimators, have the minimum variance, i.e. the OLS
estimators are BLUE.

According to the this theorem, under the basic assumptions of the classical linear
regression model, the least squares estimators are linear, unbiased and have
minimum variance (i.e. are best of all linear unbiased estimators). Some times the
theorem referred as the BLUE theorem i.e. Best, Linear, Unbiased Estimator. An
estimator is called BLUE if:
a. Linear: a linear function of the a random variable, such as, the
dependent variable Y.
b. Unbiased: its average or expected value is equal to the true population
parameter.
c. Minimum variance: It has a minimum variance in the class of linear
and unbiased estimators. An unbiased estimator with the least variance
is known as an efficient estimator.
According to the Gauss-Markov theorem, the OLS estimators possess all the
BLUE properties. The detailed proof of these properties are presented below
Dear colleague lets proof these properties one by one.
a. Linearity: (for ˆ )
Proposition:
ˆ & are linear in Y.
ˆ
Proof: From (2.17) of the OLS estimator of ˆ is given by:

x y xi (Y  Y ) xiY  Yxi
ˆ  x i2 i  x 2  ,
i i x 2i
(but
xi   ( X  X )   X   nX  0)
nX nX
x Y
 ˆ  i ; Now, xi
 (i  1,2,.. .n)
let x K
2 i i
x 2
i
 ˆ  KiY                          (2.19)
 ˆ  K1Y1  K 2Y2  K3Y3      KnYn
 ˆ is linear in Y


Check yourself question:

Show that îs linear in Y?Hint: ˆ  1 n  Xkii Y .Derive this relationship
between ˆ and Y.
b. Unbiasedness:
Proposition: ˆ &
are the unbiased estimators of the true parameters  & 
ˆ
From your statistics course, you may recall that if ˆ is an estimator of  then
E(ˆ)    the amount of bias and if ˆ is the unbiased estimator of  then bias =0
i.e. E(ˆ)    0  E(ˆ)  
In our case, ˆ & ˆ are estimators of the true parameters  &  .To show that they
are the unbiased estimators of their respective parameters means to prove that:
(ˆ) 
and (ˆ )  

 Proof (1): Prove that ˆ is unbiased i.e. (ˆ)   .
We know
that ˆ   ki (  X i U i )
kYi
 ki  ki X i  kiui ,
but
ki  and ki X i  1
0
xi (X  X ) X  nX nX  nX
ki   2
 x2  x2 x
i i i
x2
i
  k i  0 ........................................................................................................... (2.20)
ki X i  xi X ( X  X ) Xi
i  x 2
x 2 i
i
X 2  XX X  nX
2 2
 X 2  nX 2 X 2  nX 2 1
  k i X i  1 .................................................................................................... (2.21)
ˆ    kiui  ˆ    kiui                          (2.22)

(ˆ)  E( )  ki E(ui ), Since

are fixed
ki

(ˆ)   , since (ui )  0
Therefore, ˆ is unbiased estimator of  .

 Proof(2): prove that ˆ is unbiased
(ˆ )  
i.e.:
From the proof of linearity property under 2.2.2.3 (a), we know that:
ˆ  1 n  Xk Y
i i
 1 n  Xk   X U  , Since Y    X  U

i i i
i i i
  1
n X i  1
n ui  Xki  Xki X i  Xkiui
  1
n ui  Xkiui ,  ˆ    1
n ui  Xkiui
   1 n  Xk i )u i ........................................................... (2.23)
(ˆ )    1 n (u i )  Xk i (u i )
(ˆ )                                (2.24)
ˆ
is an unbiased estimator of .
c. Minimum variance of ˆ and ˆ

Now, we have to establish that out of the class of linear and unbiased estimators
of and
 , ˆ and ˆ possess the smallest sampling variances. For this, we shall
first obtain variance ofˆ and ˆ and then establish that each has the minimum
variance in comparison of the variances of other linear and unbiased estimators
obtained by any other econometric methods than OLS.
a. Variance of ˆ
...............................................................
var( )  (ˆ  (ˆ))2  (ˆ   )2 (2.25)
Substitute (2.22) in (2.25) and we get
var(ˆ)  E(  k u ) 2
i
 [k u  k u 2  ............  k 2u 2  2k k
2 2 2
uu ........ ku u ]
2k
1 1 2 2 n n 1 2 1 2 n1 n n1 n


 [k 2u 2  k 2u 2  ............  k 2u 2 ]  [2k k

uu ........ 2k k u u ]
1 1 2 2 n n 1 2 1 2 n1 n n1 n

k 2u 2 )  (k k u u i j
( )
i i i j i j
 k 2(u 2 )  2k k (u u )   2k 2
(Since (u u ) =0)
i i i j i j i i j
x x i2 1
k  i , and therefore, k 2  
i
x 2 i
(x2 )2 x 2
i i i
2
 var(ˆ )  22 k
i ……………………………………………..(2.26)
x2i
b. Variance of ˆ
var(ˆ )  (ˆ  (
)
2
 ˆ                             (2.27)
2
Substituting equation (2.23) in (2.27), we get


var(ˆ )   1  Xk
n i
u 2
i
2

  1  Xk  (u
2
)2
 n i
  2( 1  Xk ) 2
n
  2( 1 
Xki  X k )
2 2 2
n
n
i
  ( 2 1  2X
k  X 2k 2 ) , Since  k  0
n i
i i
  2 ( 1  X 2k 2 )
1 n
X2 i
x 2 1
2(  ) , Since k 2  i

 i
2
i
(x2i )2 x i2
n x
Again:
1 X 22 x i2  nX 2  X 22 

n x  nx 2  nx
i i  i 
 X 2
 X 
2
2 1
 var( )   n  2   nxi 2  ....................................................................(2.28)
ˆ   2
 x
 i   i


Dear student! We have computed the variances OLS estimators. Now, it is time to
check whether these variances of OLS estimators do possess minimum variance
property compared to the variances other estimators of the true  and  , other
than ˆ and ˆ .
To establish
that ˆ ˆ possess minimum variance property, we compare their
and
variances with that of the variances of some other alternative linear and unbiased
estimators of
 and  , say  * and  * . Now, we want to prove that any other
linear and unbiased estimator of the true population parameter obtained from any
other econometric method has larger variance that that OLS estimators.
Lets first show minimum variance of ˆ and then that of ˆ .
1. Minimum variance of ˆ
Suppose:  * an alternative linear and unbiased estimator of  and;
Let
* wi Y i................................................................................................................................................................. (2.29)

where
wi  ; but: wi  ki  ci
,
ki
 *  wi (  X i  ui ) Since Yi    X i U i
 wi  wi X i  wiui
(  *)  wi  wi

Xi ,since (ui )  0
Since  * is assumed to be an unbiased estimator, then for  * is to be an unbiased

estimator of  , there must be true that
wi  and wi X  in the above
equation. 0 1
But, wi  ki  ci
wi  (ki  ci )  ki  ci
Therefore, ci 
since ki  wi  0
0
Again
wi X i  (ki  ci )  ki X i  ci X i
Xi
Since
wi X i  and ki X i   ci X i  0 .
1 1
From these values we can drive

ci xi  0, xi  X i  X
where

ci xi   ci ( X i  X ) ci X i  Xci
Since
ci xi  ci   ci xi  0
1 0
Thus, from the above calculations we can summarize the following results.
wi
 wi xi  ci  ci X i  0
0, 1, 0,
To prove whether ˆ has minimum variance or not lets compute var( *) to
compare
with var(ˆ) .
var( *)  var(wiYi )
 w 2 var(Y )
i i
 var( *)   2w2 since Var(Y )   2

i i
But, w  (k  c )  k 2  2k c  c 2

2 2
i i i i i i i
 w2  k 2  c2 Since

k ci xi
c  0
i i i
i i
x 2
i
Therefore, var( *)   2 (k 2  c 2 )   2k 2   2c 2
i i i i
var(  *)  var( ˆ )   c22i
Given that ci is an arbitrary constant,  2c 2 is a positive i.e it is greater than zero.
i
Thus
var( *)  var(ˆ) . This proves that ˆ possesses minimum variance property.
In the similar way we can prove that the least square estimate of the constant
intercept (ˆ ) possesses minimum variance.
2. Minimum Variance of ˆ
We take a new estimator  * , which we assume to be a linear and unbiased
estimator of function of . The least square estimator ˆ is given by:
ˆ  ( 1 n  Xki )Yi
By analogy with that the proof of the minimum variance property of ˆ , let's use
the weights wi = ci + ki Consequently;
 *  ( 1n  Xwi )Yi

Since we want  *
to be on unbiased estimator of the true , that is, (*)   ,
we substitute for Y    xi  ui in  * and find the expected value of  * .
 *  ( 1n  Xwi )(  X i  ui )
 X ui
 (    Xw   XX w
 Xw u )
i i i i i
n n n
*    X  ui / n  Xwi  Xwi X i  Xwiui
For  * to be an unbiased estimator of the true , the following must hold.
 (wi )  (wi X i )  and  (wiui )  0

0, 1
i.e., if
wi  and wi X i  1 . These conditions imply that ci  and ci X i  0 .
0, 0
As in the case of ˆ , we need to compute Var(  * ) to compare with var(ˆ )

var( *)  var( 1 n  Xw )Y 
i i
 (  Xw ) var(Y )
1 2
n i i
  2( 1  Xw ) 2
n
  2( 1  X 2 w 2  2 1 Xw )
n2 i n
  2 ( n  X 2 w 2  2 X 1 w )
n2 i n
var( *)   2
 1 n
 X 2w 2  ,Since wi  0
i
but w  k 2  c 2 2
i i i
 var( *)   2
 1
n
 X (k  c 2  2 2
i i
2
2 1 X  2 2 2
var( *)   n  x 2   X c
 i
 i 
 2
  2 X 2 2c

  nxi2 
2
i
 i
The first term in the bracket it var(ˆ ) , hence
var( *)  var(ˆ )   2 X i2c 2
 var( *)  , Since  2 X 2c 2  0

var(ˆ )

i

Therefore, we have proved that the least square estimators of linear regression
model are best, linear and unbiased (BLU) estimators.
The variance of the random variable (Ui)

Dear student! You may observe that the variances of the OLS estimates involve
 2
, which is the population variance of the random disturbance term. But it is
difficult to obtain the population data of the disturbance term because of technical
and economic reasons. Hence it is difficult to compute 2
; this implies that
variances of OLS estimates are also difficult to compute. But we can compute
these variances if we take the unbiased estimate of  2 which is ˆ computed from
2
the sample value of the disturbance term ei from the expression:
e2
ˆ 2  i.............................................................2.30
u
n2
To use
ˆ in the expressions for the variances of ˆ ˆ , we have to prove
2 and
whether ˆ is
e    22
Ethe unbiased estimator of  , i.e., E(ˆ 2 ) 

2
2 i
n2

To prove this we have to compute e 2
i
from the expressions of Y,
Yˆ , y, yˆ and ei .
Proof:
Yi   ˆ   ˆ X i  e i
Yˆ   ˆ   ˆ x
 Y  Yˆ  ei......................................................................................................................................................................................... (2.31)
 ei
 Yi  Yˆ..................................................................................................................... (2.32)
Summing (2.31) will result the following expression

Yi  yi  ei
Yi 
sin (ei )  0
Yˆ i
ce
Dividing both sides the above by ‘n' will give us

Y
 Y  Yˆ                    (2.33)
Yˆ
 i
n n

Putting (2.31) and (2.33) together and subtract

Y  Yˆ  e
Y  Yˆ
 (Y  Y )  (Yˆ  Yˆ )  e
 yi  yˆ i  e .......................................................................... (2.34)
From (2.34):
ei  yi  yˆ i ....................................................................................................................................... (2.35)
Where the y's are in deviation form.

Now, we have to express
y i and in other expression as derived below.
yˆ i
From:
Yi    X i U i
Y    X U
We get, by subtraction
yi  (Yi  Y )   i ( X i  X )  (Ui U )  xi  (U U )
 yi  x  (U U )........................................................................................(2.36)
Note that we assumed earlier that , (u)  0 , i.e in taking a very large number
samples we expect U to have a mean value of zero, but in any particular single
sample U is not necessarily zero.
Similarly: From;
Yˆ   ˆ   ˆ x
Y   ˆ   ˆx
We get, by subtraction
Yˆ  Yˆ   ˆ ( X  X )
 yˆ  ˆx........................................................................................................ (2.37)
Substituting (2.36) and (2.37) in (2.35) we get

ei   xi  (u i  u )  ˆxi
 (u i  u )  ( î   )xi
The summation over the n sample values of the squares of the residuals over the
‘n' samples yields:
e2   u )  (ˆ   )x ]2
[(ui i
i
 [(u  u ) 2  (ˆ   ) 2 x 2  2(u  u )(ˆ   )x ]
i i i
 (u  u )2  (ˆ   )2 x 2  2[(ˆ   )x (u  u )]
i i i
Taking expected values we have:

(e2 )  [(u  u )2 ]  [(ˆ   )2 x 2 ]  2[(ˆ   )x
(u  u )]......................(2.38)
i i i i i
The right hand side terms of (2.38)may be rearranged as follows

a. [(u  u )2 ]  (u 2  uu )
i i
 2 (u i )2 
 ui  
 n 
1
 (u 2 )  (u) 2
i
n
 n  (u  u  .......  u )2 since (u 2 )   2
2 1
n 1 2 i i u
 n 2  1 (u 2  2u u )
n i i j
 n  1 ((u 2 )  2u u ) i  j
2
n i i j
 n  n  (u u )
2 1 2 2
n u n i j
 n   2 2
(given (u u
)  0)
u u i j
  2 (n  1)............................................................................... (2.39)
u
b. [(ˆ   )2 x 2 ]  x 2 .(ˆ   ) 2
i i
Given that the X's are fixed in all samples and we know that
1
(ˆ   ) 2  var(ˆ)   2
u
x2


1
Hence x 2 .(ˆ   )2  x 2 .  2
i i u
x 2
 x 2i . ( ˆ   ) 22  u ……………………………………………(2.40)
c. -2 [(ˆ   )xi (ui  u )]  2[(ˆ   )(xi ui  uxi )]
= -2 [( ˆ   )(xi u i )] , sin ce  xi  0
But from (2.22) , (ˆ   )  and substitute it in the above expression,

kiui
we will get:

-2 [(ˆ   )xi (ui  u )  2(kiui )(xi ui )]

 x i u i  
x
= -2    (x i u i ) ,since k  i
2
 x  i
x
2
  i 
 i
(xi ui )2 
 2 
2
  x  
i
x u  2x x u u 
2 2
 2 i i i j i j

 x i 2

x 2 (u i 2 )  2(xi x j )(ui u j ) 
 2 i  j
 x i x 2
2

i
x 2(u i 2 )
 2
x 2 ( given (uiu j )  0)
i
 2(ui 2 )  2 …………………………………………………….(2.41)

2
Consequently, Equation (2.38) can be written interms of (2.39), (2.40) and (2.41)
as follows: e2   n  1 2   2  2 2  (n  2) 2......................................... (2.42)
i u u u
From which we get

 e2  ..............................................................................
 i   E(ˆ 2 )   2 (2.43)
u u
n2
e2i
Since ˆ 2 
u
n2

2
Thus, ˆ 2   i is unbiased estimate of the true variance of the error term( 2 ).
n2
Dear student! The conclusion that we can drive from the above proof is that we
2
 2
ˆ , since
can substitute ˆ  i for ( ) in the variance expression of ˆ
2
n  2 and
E(ˆ 2 )   2 . Hence the formula of variance of ˆ and

becomes;
ˆ
Var(ˆ)
 ˆ = e2i ........................................................................................................
(2.44)
2
x 2 (n  2) 2i
i  x
 X 2 
Var(ˆ )  ˆ 2  i 
e  X2 2
i........................................................................................
 n(n i 2) (2.45)
 x
2
nx
 i   i
2
Note: e 2
can be computed as  e 2   y   ˆ x y .
i i i i
Dear Student! Do not worry about the derivation of this expression! we will
perform the derivation of it in our subsequent subtopic.
2.2.2.4. Statistical test of Significance of the OLS Estimators

(First Order tests)
After the estimation of the parameters and the determination of the least square
regression line, we need to know how ‘good' is the fit of this line to the sample
observation of Y and X, that is to say we need to measure the dispersion of
observations around the regression line. This knowledge is essential because the
closer the observation to the line, the better the goodness of fit, i.e. the better is the
explanation of the variations of Y by the changes in the explanatory variables.
We divide the available criteria into three groups: the theoretical a priori criteria,
the statistical criteria, and the econometric criteria. Under this section, our focus
is on statistical criteria (first order tests). The two most commonly used first order
tests in econometric analysis are:


i. The coefficient of determination (the square of the

correlation coefficient i.e. R2). This test is used for judging
the explanatory power of the independent variable(s).
ii. The standard error tests of the estimators. This test is used
for judging the statistical reliability of the estimates of the
regression coefficients.
1. TESTS OF THE ‘GOODNESS OF FIT’ WITH R2

r2 shows the percentage of total variation of the dependent variable that can be
explained by the changes in the explanatory variable(s) included in the model. To
elaborate this let's draw a horizontal line corresponding to the mean value of the
dependent variable Y . (see figure‘d' below). By fitting the line Yˆ  ˆ 0  ˆ X
we
1
try to obtain the explanation of the variation of the dependent variable Y produced
by the changes of the explanatory variable X.
.Y
Y = e  Y  Yˆ
Y  Y= Yˆ Yˆ  ˆ 0  ˆ X
1
= Yˆ  Y
Y.
X
Figure ‘d'. Actual and estimated values of the dependent variable Y.
As can be seen from fig.(d) above, Y  Y represents measures the variation of the
sample observation value of the dependent variable around the mean. However
the variation in Y that can be attributed the influence of X, (i.e. the regression line)
is given by the vertical distance Yˆ  Y . The part of the total variation in Y about


Y that can't be attributed to X is equal to

e  Y  Yˆ which is referred to as the
residual variation.
In summary:
e i  Yi 
= deviation of the observation Yi from the regression line.
Yˆ
yi  Y  Y = deviation of Y from its mean.
yˆ  Yˆ  Y = deviation of the regressed (predicted) value ( Yˆ ) from the mean.
Now, we may write the observed Y as the sum of the predicted value ( Yˆ ) and the
residual term (ei.).
⏟
Yˆ
⏟ ⏟i
Yi   e
Observed Yi predicted Yi Re sidual
From equation (2.34) we can have the above equation but in deviation form
y  yˆ  e . By squaring and summing both sides, we obtain the following
expression:
y 2  ( yˆ 2  e)2
y 2  ( yˆ 2  e2  2 yei)
i
 y  e 2  2yˆ ei
2
i i
But yˆ ei = e(Yˆ  Y )  e(ˆ  ˆxi

Y)
 ˆei  ˆexi  Yˆei
(but  e  0 ,  ex  0 )
i
i
  yˆ e  0 ...........................................................................(2.46)
Therefore;
⏟i ⏟ ⏟i
y 2  yˆ 2 e2 ………………………………...(2.47)
Total Explained Un exp lained var ation
var iation var iation
OR,


Total sum of Explained sum Re sidual sum


square of square of square
–_––, ––_––, ––_––,
TSS ESS RSS
i.e
TSS  ESS  ……………………………………….(2.48)

Mathematically; the explained variation as a percentage of the total variation is
explained as:
ESS  yˆ 2
……………………………………….(2.49)
TSSy 2
From equation (2.37) we

yˆ  ˆx . Squaring and summing both sides give us
have
yˆ 2  ˆ x 2                        (2.50)
2
We can substitute (2.50) in (2.49) and obtain:

ˆ x 2
2
ESS / TSS …………………………………(2.51)

 y
2
2
 xy  x
ˆ x y
2
  , Since   i i
i
 x2  y x 2
2 i
xy xy................................................................
 (2.52)
x 2 y 2
Comparing (2.52) with the formula of the correlation coefficient:

r = Cov (X,Y) /  2 2
= xy / n 2 2
=
xy / ( x2 y 2 )1/2 ………(2.53)
x x x x
Squaring (2.53) will result in: r2 = ( xy )2 / ( x2 y 2 ).
………….(2.54)

Comparing (2.52) and (2.54), we see exactly the expressions. Therefore:

 xy xy
ESS/TSS = r2
x2 y2
From (2.48), RSS=TSS-ESS. Hence R2 becomes;

RSS  1  ei
2
TSS  RSS
R 2
1 ………………………….…………(2.55)
 TSS  TSS y 2

From equation (2.55) we can drive;

RSS  e2  y 2 (1  R 2 )                            (2.56)
i i
The limit of R2: The value of R2 falls between zero and one.
i.e. 0 R2 1.
 Interpretation of R2

Suppose R 2
 0.9 , this means that the regression line gives a good fit to the
observed data since this line explains 90% of the total variation of the Y value
around their mean. The remaining 10% of the total variation in Y is unaccounted
for by the regression line and is attributed to the factors included in the disturbance
variable ui .
Check yourself question:

Exercise:
a. Show that
Suppose rxy 0 is
 the 1.
correlation
R2  coefficient between Y and X and is give by:
b. Show that the square of the coefficient of correlation is equal to ESS/TSS.
xi yi

 x i2  y i2
And let ry 2yˆ  the square of the correlation coefficient between Y and Yˆ , and is
 (yyˆ) 2
given by: r 2y yˆ
y 2yˆ 2
Show that: i) r y2yˆ  R 2 ii) r r

yyyx
2. TESTING THE SIGNIFICANCE OF OLS PARAMETERS

To test the significance of the OLS parameter estimators we need the following:
 Variance of the parameter estimators
 Unbiased estimator of  2

 The assumption of normality of the distribution of error term.

We have already derived that:
 var(ˆ) ˆ 2
 x 2
ˆ 2X 2
 var(ˆ ) nx 2

 ˆ e 2 RSS

2  n2
n2
For the purpose of estimation of the parameters the assumption of normality is not
used, but we use this assumption to test the significance of the parameter
estimators; because the testing methods or procedures are based on the assumption
of the normality assumption of the disturbance term. Hence before we discuss on
the various testing methods it is important to see whether the parameters are
normally distributed or not.
We have already assumed that the error term is normally distributed with mean
zero and variance 2 , i.e. U ~ N(0, 2 ) . Similarly, we also proved
i
thatYi ~ N[(  x), 2 ] . Now, we want to show the following:
ˆ  2 
1.  ~ N  , 
  2
  2X 2 
, ˆ ~ N
2. 2

 nx 
To show whether ˆ and

are normally distributed or not, we need to make use of
ˆ
one property of normal distribution. “........ any linear function of a normally
distributed variable is itself normally distributed.”
ˆ  ki Yi  k1Y1  k2 Y2i . . kn Yn
ˆ  wi Yi  w1Y1  w2 Y2i .  wn Yn
Since ˆ and
are linear in Y, it follows that
ˆ


ˆ  2
  2X 2 
 ~ N  , 2 ˆ 
,;  ~ N
nx 2 
   
The OLS estimates ˆ and
are obtained from a sample of observations on Y and
ˆ
X. Since sampling errors are inevitable in all estimates, it is necessary to apply
test of significance in order to measure the size of the error and determine the
degree of confidence in order to measure the validity of these estimates. This can
be done by using various tests. The most common ones are:
i) Standard error test ii) Student’s t-test iii) Confidence interval
All of these testing procedures reach on the same conclusion. Let us now see these
testing methods one by one.
i) Standard error test
This test helps us decide whether the estimates ˆ and
are significantly different
ˆ
from zero, i.e. whether the sample from which they have been estimated might
have come from a population whose true parameters are zero.
  0 and / or   0 .
Formally we test the null hypothesis
H 0 : i 
against the alternative hypothesis H1 :  i  0
0
The standard error test may be outlined as follows.
First: Compute standard error of the parameters.
SE(ˆ)  var(ˆ)
SE(ˆ ) var(ˆ )
Second: compare the standard errors with the numerical values of ˆ

and ˆ .
Decision rule:
 If
SE( î )  1
2 î , accept the null hypothesis and reject the alternative
hypothesis. We conclude that

î is statistically insignificant.

 If
SE( î )  1
2 î , reject the null hypothesis and accept the alternative
hypothesis. We conclude that

î is statistically significant.
The acceptance or rejection of the null hypothesis has definite economic meaning.
Namely, the acceptance of the null hypothesis
 (the slope parameter is zero)
0
implies that the explanatory variable to which this estimate relates does not in fact
influence the dependent variable Y and should not be included in the function,
since the conducted test provided evidence that changes in X leave Y unaffected.
In other words acceptance of H 0 implies that the relation ship between Y and X is
in fact Y    (0)x   , i.e. there is no relationship between X and Y.
Numerical example: Suppose that from a sample of size n=30, we estimate the
following supply function.
Q  120 0.6  ei
 p
SE : (1.7) (0.025)
Test the significance of the slope parameter at 5% level of significance using the
standard error test.
SE )  0.025
( ˆ
)  0.6
(ˆ
1
2 ˆ  0.3
This implies that

SE( î )  1
2 î . The implication is ˆ is statistically significant at
5% level of significance.
Note: The standard error test is an approximated test (which is approximated from
the z-test and t-test) and implies a two tail test conducted at 5% level of
significance.
ii) Student’s t-test
Like the standard error test, this test is also important to test the significance of the
parameters. From your statistics, any variable X can be transformed into t using
the general formula:

X 
t , with n-1 degree of freedom.
sx
Where
i value of the population mean

sample estimate of the population standard deviation
sx 
( X  X ) 2
sx  n 1
n  sample size
We can derive the t-value of the OLS estimates
ˆ i  
tˆ 
 

SE(ˆ  with n-k degree of freedom.
) 
tˆ 
ˆ 


SE(ˆ
Where: )
SE = is standard error
k = number of parameters in the model.
Since we have two parameters in simple linear regression with intercept different
from zero, our degree of freedom is n-2. Like the standard error test we formally
test the H 0 : i  against the H1 :  i  for the slope
hypothesis: 0 alternative 0
parameter; and
H0 :  against the alternative H1 :   for the intercept.
0 0
To undertake the above test we follow the following steps.

Step 1: Compute t*, which is called the computed value of t, by taking the value of
 in the null hypothesis. In our case   0 , then t* becomes:
ˆ  0 ˆ
t*  
SE(ˆ) SE(ˆ)
Step 2: Choose level of significance. Level of significance is the probability of

making ‘wrong' decision, i.e. the probability of rejecting the hypothesis when it is

actually true or the probability of committing a type I error. It is customary in

econometric research to choose the 5% or the 1% level of significance. This

means that in making our decision we allow (tolerate) five times out of a hundred
to be ‘wrong' i.e. reject the hypothesis when it is actually true.
Step 3: Check whether there is one tail test or two tail test. If the inequality sign
in the alternative hypothesis is  , then it implies a two tail test and divide the
chosen level of significance by two; decide the critical rejoin or critical value of t
called tc. But if the inequality sign is either > or < then it indicates one tail test and
there is no need to divide the chosen level of significance by two to obtain the
critical value of to from the t-table.
Example:
If we
H0 :i 0
have
against:
H1 :  i  0
Then this is a two tail test. If the level of significance is 5%, divide it by two to
obtain critical value of t from the t-table.
Step 4: Obtain critical value of t, called tc at

2 and n-2 degree of freedom for two

tail test.
Step 5: Compare t* (the computed value of t) and tc (critical value of t)
 If t*> tc , reject H0 and accept H1. The conclusion is ˆ is statistically
significant.
 If t*< tc , accept H0 and reject H1. The conclusion is ˆ is statistically
insignificant.
Numerical Example:
Suppose that from a sample size n=20 we estimate the following consumption
function:
C  100 0.70  e
 (0.21)
(75.5)


The values in the brackets are standard errors. We want to test the null hypothesis:
H 0 : i 
against the alternative H1 :  i  using the t-test at 5% level of
0
0
significance.
a. the t-value for the test statistic is:
t* 
ˆ  0 ˆ 0.70
 
0.21
SE(ˆ) = 3.3
SE(ˆ)
b. Since the alternative hypothesis (H1) is stated by inequality sign (  ) ,it is a

two tail test, hence we divide   0.05 2 
2 to obtain the critical value of
0.025
‘t' at 
2 =0.025 and 18 degree of freedom (df) i.e. (n-2=20-2). From the t-
table ‘tc' at 0.025 level of significance and 18 df is 2.10.
c. Since t*=3.3 and tc=2.1, t*>tc. It implies that ˆ is statistically significant.
iii) Confidence interval

Rejection of the null hypothesis doesn't mean that our estimate
ˆ and is the
ˆ
correct estimate of the true population parameter
 and  . It simply means that
our estimate comes from a sample drawn from a population whose parameter  is
different from zero.
In order to define how close the estimate to the true parameter, we must construct
confidence interval for the true parameter, in other words we must establish
limiting values around the estimate with in which the true parameter is expected to
lie within a certain “degree of confidence”. In this respect we say that with a
given probability the population parameter will be with in the defined confidence
interval (confidence limits).

We choose a probability in advance and refer to it as confidence level (interval

coefficient). It is customarily in econometrics to choose the 95% confidence level.
This means that in repeated sampling the confidence limits, computed from the

sample, would include the true population parameter in 95% of the cases. In the
other 5% of the cases the population parameter will fall outside the confidence
interval.
In a two-tail test at  level of significance, the probability of obtaining the specific
t-value either –tc or tc is 
2 at n-2 degree of freedom. The probability of obtaining
any value of t which is equal to ˆ  

at n-2 degree of freedom is
SE(ˆ
1   2   2  i.e. 1   . )
i.e. Pr t  t*  t   1 
c c …………………………………………(2.57)

but
ˆ  
t* …………………………………………………….(2.58)
 SE(ˆ
)
Substitute (2.58) in (2.57) we obtain the following expression.


Pr tc ˆ   
   t   1   ………………………………………..(2.59)
SE(ˆ) 
 
Pr  SE( ˆ )t  ˆ    SE( ˆ )t  1        by multiplying SE( ˆ )
c c
Pr ˆ  SE( ˆ )t      ˆ  SE( ˆ )t   1        by subtracting ˆ

c c
Pr ˆ  SE( ˆ )    ˆ  SE( ˆ )t   1        by multiplying by  1

c
Prˆ  SE( ˆ )t    ˆ  SE( ˆ )t   1        int erchanging

c c
The limit within which the true  lies at (1   )% degree of confidence is:
[ ˆ  SE( ˆ )t c , ˆ  SE( ˆ )t c ] ; where

is the critical value of t at 
2 confidence
tc
interval and n-2 degree of freedom.

The test procedure is outlined as follows.
H0 : 0
H1 :   0
Decision rule: If the hypothesized value of  in the null hypothesis is within the
confidence interval, accept H0 and reject H1. The implication is that ˆ is


statistically insignificant; while if the hypothesized value of  in the null
hypothesis is outside the limit, reject H 0 and accept H1. This indicates ˆ is
statistically significant.
Numerical Example:
Suppose we have estimated the following regression line from a sample of 20
observations.
Y  128.5  2.88X  e
(38.2) (0.85)
The values in the bracket are standard errors.

a. Construct 95% confidence interval for the slope of parameter
b. Test the significance of the slope parameter using constructed confidence
interval.
Solution:
a. The limit within which the true  lies at 95% confidence interval is:
ˆ  SE( ˆ )t c
ˆ  2.88
SE (ˆ)  0.85
tc at 0.025 level of significance and 18 degree of freedom is 2.10.
 ˆ  SE( ˆ )t c  2.88  2.10(0.85)  2.88  1.79.
The confidence interval is:

(1.09, 4.67)
b. The value of  in the null hypothesis is zero which implies it is out side the
confidence interval. Hence  is statistically significant.
2.2.3 Reporting the Results of Regression Analysis

The results of the regression analysis derived are reported in conventional formats.
It is not sufficient merely to report the estimates of  's. In practice we report

regression coefficients together with their standard errors and the value of R2. It
has become customary to present the estimated equations with standard errors
placed in parenthesis below the estimated parameter values. Sometimes, the
estimated coefficients, the corresponding standard errors, the p-values, and some
other indicators are presented in tabular form.
These results are supplemented by R2 on ( to the right side of the regression
equation).
Example: Y  128 . 5  2 . 88
X , R2 = 0.93. The numbers in the
parenthesis below the parameter estimates are the standard errors. Some
econometricians report the t-values of the estimated coefficients in place of the
standard errors.
Review Questions
Review Questions
1. Econometrics deals with the measurement of economic relationships which are stochastic
or random. The simplest form of economic relationships between two variables X and Y
can be represented by:
Yi   0  1 X i  ; where  0 and 1  are regression parameters and

Ui
U i  the stochastic disturbance term
What are the reasons for the insertion of U-term in the model?
2. The following data refers to the demand for money (M) and the rate of interest (R) in for
eight different economics:
M (In billions) 56 50 46 30 20 35 37 61
R% 6.3 4.6 5.1 7.3 8.9 5.3 6.7 3.5
a. Assuming a relationship M    R U i , obtain the OLS estimators of
 and 
b. Calculate the coefficient of determination for the data and interpret its value
c. If in a 9th economy the rate of interest is R=8.1, predict the demand for money(M) in
this economy.


3. The following data refers to the price of a good ‘P' and the quantity of the good supplied,
‘S'.
P 2 7 5 1 4 8 2 8
S 15 41 32 9 28 43 17 40
a. Estimate the linear regression line (S)    P
b. Estimate the standard errors of ˆ and ˆ

c. Test the hypothesis that price influences supply
d. Obtain a 95% confidence interval for
4. The following results have been obtained from a simple of 11 observations on the values
of sales (Y) of a firm and the corresponding prices (X).
X  519.18
Y  217.82
 X  3,134,543
2
i
 X Y  1,296,836
i i
Y  539,512
2
i
i) Estimate the regression line of sale on price and interpret the results
ii) What is the part of the variation in sales which is not explained by the
regression line?
iii) Estimate the price elasticity of sales.
5. The following table includes the GNP(X) and the demand for food (Y) for a country over
ten years period.
year 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989
Y 6 7 8 10 8 9 10 9 11 10
X 50 52 55 59 57 58 62 65 68 70
a. Estimate the food function
b. Compute the coefficient of determination and find the explained and unexplained
variation in the food expenditure.
c. Compute the standard error of the regression coefficients and conduct test of
significance at the 5% level of significance.
6. A sample of 20 observation corresponding to the regression model Yi    X i U i

gave the following data.

Y
 Y  Y 
2
i
 21.9  86.9
 i  X   215.4
2
i i
X 186.2
X
 X i  X Yi  Y   106.4
a. Estimate  and 
b. Calculate the variance of our estimates
c. Estimate the conditional mean of Y corresponding to a value of X fixed at X=10.
7. Suppose that a researcher estimates a consumptions function and obtains the following
results:
C  15  0.81Yd n  19
(3.1) (18.7) R 2  0.99
where C=Consumption, Yd=disposable income, and numbers in the parenthesis are the ‘t-ratios'
a. Test the significant of Yd statistically using t-ratios
b. Determine the estimated standard deviations of the parameter estimates
8. State and prove Guass-Markov theorem
9. Given the model:
Yi   0   1 X i 
with usual OLS assumptions. Derive the expression for the error
Ui
variance.


Chapter Three
THE CLASSICAL REGRESSION ANALYSIS
[The Multiple Linear Regression Model]
3.1 Introduction
In simple regression we study the relationship between a dependent variable and a
single explanatory (independent variable). But it is rarely the case that economic
relationships involve just two variables. Rather a dependent variable Y can
depend on a whole series of explanatory variables or regressors. For instance, in
demand studies we study the relationship between quantity demanded of a good
and price of the good, price of substitute goods and the consumer's income. The
model we assume is:
Yi   0  1 P1   2 P2   3 X i  u i --------------------------------(3.1)
Where Yi  quantity demanded, P1 is price of the good, P2 is price of substitute

goods, Xi is consumer's income, and
 ' s are unknown parameters and is the
disturbance. ui
Equation (3.1) is a multiple regression with three explanatory variables. In general
for K-explanatory variable we can write the model as follows:
Yi  0  1 X 1i   2 X 2i  3 X 3i  .........   k X ki  ui----------(3.2)
Where
X ki   1,2,3,......., K ) are explanatory variables, Yi is the dependent
(i
variable and
 j ( j  0,1,2,....(k  1)) are unknown parameters and ui is the
disturbance term. The disturbance term is of similar nature to that in simple
regression, reflecting:
- the basic random nature of human responses
- errors of aggregation
- errors of measurement
- errors in specification of the mathematical form of the model

and any other (minor) factors, other

xi that might influence Y.
than
In this chapter we will first start our discussion with the assumptions of the
multiple regressions and we will proceed our analysis with the case of two
explanatory variables and then we will generalize the multiple regression model in
the case of k-explanatory variables using matrix algebra.
3.2 Assumptions of Multiple Regression Model

In order to specify our multiple linear regression model and proceed our analysis
with regard to this model, some assumptions are compulsory. But these
assumptions are the same as in the single explanatory variable model developed
earlier except the assumption of no perfect multicollinearity. These assumptions
are:
1. Randomness of the error term: The variable u is a real random variable.
2. Zero mean of the error term:
E(ui )  0
3. Hemoscedasticity: The variance of each ui

is the same for all xi values.
the
i.e. E(u 2 )   (constant)
2
4. Normality of u: The values of each are normally distributed.

ui
i.e. U i ~ N(0, 2)
5. No auto or serial correlation: The values of ui (corresponding to Xi ) are
independent from the values of any

other ui (corresponding to Xj ) for ij.
i.e.
E(ui u j )  for xi  j
0
6. Independence of ui and Xi : Every disturbance term u is independent of

i
the explanatory variables. i.e.

E(ui X 1i )  E(ui X 2i )  0

This condition is automatically fulfilled if we assume that the values of the

X's are a set of fixed numbers in all (hypothetical) samples.
7. No perfect multicollinearity: The explanatory variables are not perfectly
linearly correlated.

We can't exclusively list all the assumptions but the above assumptions are some
of the basic assumptions that enable us to proceed our analysis.
3.3 A Model With Two Explanatory Variables

In order to understand the nature of multiple regression model easily, we start our
analysis with the case of two explanatory variables, then extend this to the case of
k-explanatory variables.
3.3.1 Estimation of parameters of two-explanatory variables model
The Y   0  1 X 1   2 X 2 ……………………………………(3.3)
model: U i
is multiple regression with two explanatory variables. The expected value of the
above model is called population regression equation i.e.
E(Y )   0  1 X 1   2 X 2 , Since E(Ui )  0................................................. (3.4)
where
i is the population parameters.  0 is referred to as the intercept and  1
and
 2 are also some times known as regression slopes of the regression. Note
that,
 2 for example measures the effect E(Y ) of a unit change X 2 when
on in
X 1 is held constant.
Since the population regression equation is unknown to any investigator, it has to

be estimated from sample data. Let us suppose that the sample data has been used
to estimate the population regression equation. We leave the method of estimation
unspecified for the present and merely assume that equation (3.4) has been
estimated by sample regression equation, which we write as:
Yˆ  ˆ0  ˆ  ˆ 2 X ……………………………………………….(3.5)
1
X1 2
Where
ˆ j e estimates of the
ar  j and Yˆ is known as the predicted value of Y.

Now it is time to state how (3.3) is estimated. Given sample observation on

Y, X 1 & X 2 , we estimate (3.3) using the method of least square (OLS).
Yˆ   ˆ 0   ˆ 
X  ei.................................................................................................................. (3.6)
X 1i
ˆ2 2i
1
is sample relation between Y, X 1 & X 2 .
ei  Yi  Yˆ  Yi  ˆ 0  ˆ X 1  ˆ 2 X 2 ........................................................................................... (3.7)
1
To obtain expressions for the least square estimators, we partially differentiate
e 2
with respect to ˆ ,
1 and 
ˆ and set the partial derivatives equal to zero.
i
0 
ˆ 2
  e    
2
i   X  X   0 ………………………. (3.8)
 ˆ
 0 ˆ 1i ˆ2 2i
ˆ 1
0 2 Yi
  e  2
   X  X   0 ……………………. (3.9)
i  2 X
 ˆ 1i Yi  0 ˆ 11i ˆ 11i
ˆ
1
  e  2
   X  ˆ X   0 ………… ………..(3.10)
i  2 X
ˆ 2 2i Yi ˆ0 ˆ 11i 2i
2
Summing from 1 to n, the multiple regression equation produces three Normal

Equations:
 Y  nˆ  ˆ X  ˆ X 0 1 1i 2 2i ............................................................................................... (3.11)
 X Y  ˆ  ˆ X  ˆ X 1 1
2
1 2 1 1 …………………………(3.12)
2 i
X X
0
X Y
 ˆ  ˆ X  ˆ X 2
.........................................
(3.13)
2 i
X X
0 2 1 1 2 2 2
From (3.11) we
obtain ˆ 0
ˆ 0  Y  ˆ 
X2 (3.14)
X1 1 ˆ
2
Substituting (3.14) in (3.12) , we get:

X 1i
Yi  (Y  ˆ X 1
 ˆ 2
1 1 2 2
X 2 )X 1i ˆ X 1i 2 X
  ˆ
 X 1i Yi  YˆX  ˆ (X 2  X X )  (X X  X 2 X 2i )

1i 1i ˆ 2 1i 2i
X
Yi  nY X  (X )
2
1i
2
 nX (X 1i X  nX X 2 )----------(3.15)
1i ˆ2 1i ˆ 2 1
1 2
We know that

X Y 2  (X  nX Y )  x y
Y
i i i i i i i i
 X  2
X  (X 2
 nX )  x
2 2
i i i i
Substituting the above equations in equation (3.14), the normal equation (3.12) can
be written in deviation form as follows:
 x y  ˆ x
1 1
2
1
 ˆ 2x 1 2…………………………………………(3.16)
x
Using the above procedure if we substitute (3.14) in (3.13), we get
x y  ˆ x
 ˆ
2
………………………………………..(3.17)
x 2 1 1 2
x
2 2
Let's bring (2.17) and (2.18) together
 x y  ˆ x 2
 ˆ x
……………………………………….(3.18)
x 1 1 1 2 1 2
x y  ˆ x
 ˆ
2
……………………………………….(3.19)
x 2 1 1 2
x
2 2
ˆ and
1 ˆ can easily be solved using matrix
2
We can rewrite the above two equations in matrix form as follows.
x 2
1
x 1
ˆ
= x 2 y ………….(3.20)
x2
1
xx 1 2
2
ˆ 2 x 3y
x 2
If we use Cramer's rule to solve the above matrix we obtain

x1 y . x2 2  x1 x2 . x 2y
ˆ  …………………………..…………….. (3.21)
1
x . x  (x x )
2 2 2
1 2 1 2
x2 y . x1 2  x1 x2 . x 1y

ˆ  ………………….……………………… (3.22)
2
x . x  (x x )
2 2 2
1 2 1 2
We can also express ˆ and ˆ in terms of covariance and variances of

2
1
Y , X 1 and ˆ

X2 Cov( X 1 ,Y ) . Var( X 1 )  Cov( X 1 , X 2 ) . Cov( X 2 ,Y )

1
          (3.23)
Var( X ).Var( 2 )  [cov(X 1 , )]2
1 2
X X
Cov( X 2 ,Y ) . Var( X 1 )  Cov( X 1 , X 2 ) . Cov( X 1 ,Y )

ˆ2           (3.24)
Var( X 1).Var( X )])2 2 [Cov( X1 , X 2

3.3.2 The coefficient of determination ( R2):two explanatory variables case

In the simple regression model, we introduced R2 as a measure of the proportion
of variation in the dependent variable that is explained by variation in the
explanatory variable. In multiple regression model the same measure is relevant,
and the same formulas are valid but now we talk of the proportion of variation in
the dependent variable explained by all explanatory variables included in the
model. The coefficient of determination is:
R 2  ESS RSS ei-------------------------------------------

2
1  1  (3.25)
TSS  TSS y 2
i
In the present model of two explanatory variables:

e2  ( y  ˆ x  ˆ x ) 2
i i 1 1i 2 2i
 ei ( y i  ˆ x  ˆ 2 x 2i )
1 1i
 ei y  ˆ x e  ˆ e x
1 1 i 2 i 2
 ei yi
since ei x1i  ei x2i  0
 y i ( y i  ˆ x  ˆ 2 x 2i )
i.e e2  y 2  ˆ x y  ˆ x y
1 1
i 1 1i i 2 2i i
 y
 ˆ y  ˆ x y
2
----------------- (3.26)
⏟
e 2
⏟
x 
Total sum of
– – – _–2 ––
1 1 i i i
Re sidual sum of squares
square (Total
var iation)
2i
,i (un exp lained var iation)
Explained sum of
square ( Explained
var iation)
ESS ˆ x y11i i22i
ˆ i x y
 R2   ----------------------------------(3.27)
TSS y 2
As in simple regression, R2 is also viewed as a measure of the prediction ability of

the model over the sample period, or as a measure of how well the estimated
regression fits the data. The value of R2 is also equal to the squared sample
correlation coefficient between Yˆ & Yt . Since the sample correlation coefficient
measures the linear association between two variables, if R2 is high, that means
there is a close association between the values of Yt and the values of predicted by


the model,
Yˆ . In this case, the model is said to “fit” the data well. If R2 is low,
t
there is no association between the values of Yt and the values predicted by the
model, Yˆ and the model does not fit the data well.
t
3.3.3 Adjusted Coefficient of Determination ( R 2 )

One difficulty with R 2 is that it can be made large by adding more and more
variables, even if the variables added have no economic justification.
Algebraically, it is the fact that as the variables are added the sum of squared
errors (RSS) goes down (it can remain unchanged, but this is rare) and thus
R 2 goes up. If the model contains n-1 variables then R 2 =1. The manipulation of
model just to obtain a high R 2 is not wise. An alternative measure of goodness of
fit, called the adjusted
R 2 and often symbolized as R 2 , is usually reported by
regression programs. It is computed as:
e2 / n  k
n1
R2 1 i  1  (1  R 2 ) -----------------------------------(3.28)
y / n  1
2
nk
This measure does not always goes up when a variable is added because of the
degree of freedom term n-k is the numerator. As the number of variables k
increases, RSS goes down, but so does n-k. The effect on R 2 depends on the
amount by which R 2 falls. While solving one problem, this corrected measure of
goodness of fit unfortunately introduces another one. It losses its interpretation;
R 2 is no longer the percent of variation explained. This modified R 2 is sometimes
used and misused as a device for selecting the appropriate set of explanatory
variables.
3.4. General Linear Regression Model and Matrix Approach
So far we have discussed the regression models containing one or two explanatory
variables. Let us now generalize the model assuming that it contains k variables.
It will be of the form:
Y   0  1 X 1   2 X 2  ......   k X k  U


There are k parameters to be estimated. The system of normal equations consist of

k+1 equations, in which the unknowns are the parameters
 0 , 1,  2....... and the
k
known terms will be the sums of squares and the sums of products of all variables
in the structural equations.
Least square estimators of the unknown parameters are obtained by minimizing
the sum of the squared residuals.
e2  (   ˆ  ˆ 2 X  ....... X )2
y i i
 ˆ0 X 1 2 ˆ k
1 k
With respect
 j ( j  0,1,2,.. (k  1))
to
The partial derivations are equated to zero to obtain normal equations.

e2
ˆ    X  ˆ X  ....... Xk) 0
1 2
0 2(Yi  0 ˆ 2
ˆ ˆ k
1
e 2
X k )(xi )  0
 ˆ    X 1  ˆ 2 X  .......
1
2(Yi  ˆ0 ˆ 2 ˆ k
1
……………………………………………………..
e2
ˆ    X  ˆ X  ....... X k )(xki )  0
1 2
k 2(Yi  0 ˆ 12
ˆ ˆ k
The general form of the above equations (except first ) may be written as:
e2
 ˆ    X       ˆ X )  0 ; where ( j  1,2,..k)
k
j 2(Yi  0 ˆ 11i
ˆ ki
The normal equations of the general linear regression model are

Yi  nˆ0  ˆ X 1i  ˆ 2 X 2i  ......................... ˆ k X ki
1
Yi X1i
 ˆ 0  ˆ 1 2
 .................................  ˆ kX 1X ki
X 1i X1i
Yi X 2i
 ˆ 0 21  ˆ X  ˆ 2 2
 ..........  ˆ X X ki
X i X1i 2i X
1 2 k 2
: : : : :
: : : : :
Yi X ki 2
 ˆ 0 X  ˆ 1X X ki   X X ..................  ˆ k X ki
1i 2i ki
ki

Solving the above normal equations will result in algebraic complexity. But we
can solve this easily using matrix. Hence in the next section we will discuss the
matrix approach to linear regression model.
3.4.1 Matrix Approach to Linear Regression Model

The general linear regression model with k explanatory variables is written in the
form:
Yi   0  1 X 1i   2 X 2i ............  k X +Ui
ki
where
(i  1,2,3,......n) and  0  the 1 to  k = partial slope coefficients
intercept,
U= stochastic disturbance term and i=ith observation, ‘n' being the size of the
observation. Since i represents the ith observation, we shall have ‘n' number of
equations with ‘n' number of observations on each variable.
Y1   0  1 X 11   2 X 21  3 X 31...................  k X k1 U 1
Y2   0  1 X 12   2 X 22  3 X 32...................  k X k 2  U 2
Y3   0  1 X 13   2 X 23  3 X 33...................  k X k 3  U 3
…………………………………………………...
Yn   0  1 X 1n   2 X 2n   3 X 3n .............   k X kn U n
These equations are put in matrix form as:
Y1  X X k1 
1 X .......  0 U1 
11 21 
Y  1 X ....... X    U 
X
 2   1  2
X k3
12 22 k2
Y3   X 13 X .......  2  U 3 
1 23 

       
. ....... .   .   . 
  . . .
Y n   X X 2n ....... X kn   n  U n 
Y 1 1n
X   U
 . 

In short.......................................................................................................(3.29)
Y  X
The order of matrix and vectors involved are:
Y  (n
X  (n  (k    (k  1) and U  (n 1)
1),
1), 1

To derive the OLS estimators of  , under the usual (classical) assumptions
mentioned earlier, we define two vectors ˆ and ‘e' as:

 ˆ 0  e1 
 ˆ e 
 1   2
ˆ   .  and e.
 .   
  .
en
ˆ
 k 
Thus we can
write: Y  X ˆ  and e  Y  Xˆ
e
We have to minimize:
e n 2
i
 e2  e2  e2 ......... e 2
1 2 3 n
i1
e1 
 
 e2 
 [e1, e2.....en ] . e'e

 
. 
 e n 
  e 2  e' e
i
e'e  (Y  Xˆ)'(Y  Xˆ)
 YY 'ˆ' X 'Y  Y ' Xˆ  ˆ' X ' Xˆ..................................(3.30)
Since ˆ' X 'Y is scalar (1x1), it is equal to its transpose;

'
ˆ' X 'Y  Y ' Xˆ
e'e  Y 'Y  2ˆ' X 'Y  ˆ' X ' Xˆ-------------------------------------------(3.31)
Minimizing e'e with respect to theelements in ˆ

e2 (e'e)
ˆ i    ˆ
 ˆ 2X 2X'
Since ( X ' AX )  2 and also too 2X'A

AX
ˆ
Equating the expression to null vector 0, we obtain:


 2 X 'Y  2 X ' Xˆ  0  X ' Xˆ  X 'Y
ˆ  ( X ' X )1 X ………………………………. ………. (3.32)
Hence ˆ is the vector of required least square estimators, ˆ0 , ˆ , ˆ 2 ,. ˆ k .

1
3.4.2. Statistical Properties of the Parameters (Matrix) Approach

We have seen, in simple linear regression that the OLS estimators
(ˆ & ˆ) satisfy
the small sample property of an estimator i.e. BLUE property. In multiple

regression, the OLS estimators also satisfy the BLUE property. Now we proceed
to examine the desired properties of the estimators in matrix notations:
1. Linearity
We know
that: ˆ  ( X ' X )1 X 'Y
Let C= ( X X ) 1 X 
 ˆ  CY................................................................................... (3.33)
Since C is a matrix of fixed variables, equation (3.33) indicates us ˆ is linear in Y.

2. Unbiased ness
ˆ  ( X ' X )1 X 'Y
ˆ  ( X ' X )1 X '( X  U )
ˆ    ( X ' X )1 X 'U..................................................................(3.34)

since ( X ' X ) 1 X ' X  I 
(ˆ)    ( X ' X )1 X 'U

 ( )   ( X ' X )1 X 'U 
   ( X ' X ) 1 X '(U )
  , since (U )  0
Thus, least square estimators are unbiased.
3. Minimum variance
Before showing all the OLS estimators are best(possess the minimum variance
property), it is important to derive their variance.

We know that, var(ˆ)   (ˆ   )2   (ˆ   )(ˆ   )'   

 (ˆ   )(ˆ   )'  
(ˆ  

) 2 
 (ˆ   )(ˆ   ....... 
 ( ˆ   )  )  
1 1
) ( ˆ
1 1 k k
1 1 2 2 2

 ˆ ˆ
) ( 2   2 )(  11
 ˆ
)  22
( ....... ( ˆ ˆ

2   2 )(  k   k ) 

 : : : 
 

ˆ ( (ˆ )
:  )    ( (ˆ ):
ˆ
 )  ........ ˆ(  : ) 2 
 k 1 1 k k 2 2 k k 
 var(ˆ ) cov(ˆ , ˆ ) ....... cov(ˆ , ˆ ) 

1 2 1 k
 
ˆ 1 ) var(ˆ 2 ) ...............cov(ˆ 2 , ˆ k )
1
cov(ˆ 2 , 
  : : : 
 : : : 
cov(ˆ , ˆ ) cov(ˆ , ˆ ) ....... var(ˆ ) 
 k 1 k 2 k 
The above matrix is a symmetric matrix containing variances along its main
diagonal and covariance of the estimators every where else. This matrix is,
therefore, called the Variance-covariance matrix of least squares estimators of the
regression slopes. Thus,

var(ˆ)   (ˆ   )(ˆ   )' ...........................................(3.35)
From (3.15) ˆ    ( X ' X )1 X 'U
 ˆ    ( X X ) 1 X U.....................................................................................(3.36)
Substituting (3.17) in (3.16)


var(ˆ)   ( X ' X )1 X 'U ( X ' X )1 X 'U ' 
var(ˆ)  ( X ' X ) 1
X 'UU ' X ( X ' X ) 1 
 ( X ' X )1 X ' (UU ') X ( X ' X ) 1
 ( X ' X ) 1 X '  2 I X ( X ' X ) 1

u n
  ( X ' X ) X ' X ( X ' X )1

2 1
u
var( ˆ )   2 ( X ' X )1.........................................................................................(3.37)

u


Note: ( 2 being a scalar can be moved in front or behind of a matrix while

u
identity matrix I can be suppressed).

n
Thus we
obtain, var(ˆ)   2 ( X ' X )1
u
 n X ....... X kn 
1n
X X ....... X X 
21n
Where,  1n 1n kn 
(X'X) 1
 : : : 
 : : : 
 
X 1n X ....... X 2 kn
  kn 
kn
We can, therefore, obtain the variance of any estimator say ˆ by taking the ith term
1
from the principal diagonal of ( X ' X )1 and then multiplying it by  2 .
u
Where the X's are in their absolute form. When the x's are in deviation form we
can write the multiple regression in matrix form as ;
ˆ  (xx) 1 xy
ˆ
2 x
1 1   x1 x2 ....... x1 xk 
ˆ ˆ ....... x2 xk
x2 x1 x2
2
2  
where  = : and (x x)   : : : 
: : : :
 
ˆ k x x x x ....... x 2

 n 1 n 2 k 
The above column matrix ˆ doesn't include the constant term ˆ .Under such
0
conditions the variances of slope parameters in deviation form can be written
as: var(ˆ)   2 (x' x)1........................................................................................... (2.38)

u
(the proof is the same as (3.37) above). In general we can illustrate the variance of
the parameters by taking two explanatory variables.
The multiple regression when written in deviation form that has two explanatory
variables is
y1  ˆ x  ˆ x
1 1 2 2

var(ˆ)   (ˆ   )(ˆ   )' 

In this
model; ((ˆˆ  )   )
1 1
 
  ( ˆ 2   2 )  
(ˆ   )' 
 1 ) 2) 
( ˆ
1
( ˆ 2
 (ˆ   )(ˆ   )' 

(ˆ  )
 ( ˆ  )  ) 
 (ˆ
1
 1 1 2 2
  (  ˆ 2   2 )  1
ˆ (ˆ   ) 2 (ˆ   )(ˆ   ) 

ˆ
and  (   )(   )'  
  1 )(ˆ   ) 1
(ˆ1 2 ) 2 
(ˆ 1 1 2 2 2 2 
 var(ˆ
cov(ˆ , ˆ 2)
) 1 1
ˆ ˆ var(ˆ 2 ) 
cov( 1,  2 )
In case of two explanatory variables, x in the deviation form shall be:
x11
x12 x21 
x22  x11 x1 ....... x1n 
2
x  and x'  x ....... x 
: :

  x  2n 
x x 
12 22
 1n 2n 
 x 2
x1x 2 1
 2 (x' x) 1   2  1 
x x x 
u u 2
1 2 2
2  x
 2 2
 x1 x2 

 x x1 2 x 1 
u 2
Or  2 (x' x)1 
u x1 2 x1 x2
x1 x2 x22
i.e.,  u2x 22
var(ˆ ) ........................................................................................................
(3.39)
1
x 2x 2  (x x )2
1 2 1 2
ˆ  u2x 12 .................................................................................
and, var( 2 )  (3.40)
x x  (x x )
2 2 2
1 2 1 2

() 2ux 1x 2
cov(ˆ1, ˆ 2 )  ..........................................................................................................
(3.41)
x x  (x x )
2 2 2
1 2 1 2

The only unknown part in variances and covariance of the estimators is  2 .

u
 e2 
As we have seen in simple regression modelˆ 2   i  . For k-parameters
n  2 
 e2 
(including the constant parameter)ˆ 2   i  .
n  k 
In the above model we have three parameters including the constant term and
 e2 
ˆ 2   i 
 n  3
e 2
 y2 
x y x y......... x y....................................... (3.42)
i i 1

1 2 2 K K
this is for k explanatory variables. For two explanatory variables
e 2
 y2 
x y x y.................................................................... (3.43)
i i 1
1 2 2
This is all about the variance covariance of the parameters. Now it is time to see
the minimum variance property.
Minimum variance of ˆ
To show that all the

i ' s in the ˆ vector are Best Estimators, we have also to
prove that the variances obtained in (3.37) are the smallest amongst all other
possible linear unbiased estimators. We follow the same procedure as followed in
case of single explanatory variable model where, we first assumed an alternative
linear unbiased estimator and then it was established that its variance is greater
than the estimator of the regression model.
Assume that ˆ is an alternative unbiased and linear estimator of  . Suppose that
 
ˆ  ( X ' X )1 X 'B Y
Where B is (k x n) matrix of known constants.
 
 ˆ  ( X ' X )1 X 'B X U 
ˆ  ( X ' X )1 X '( X U )  B( X U )



(ˆ)   ( X ' X )1 X '( X U )  B( X U ) 
  ( X ' X ) 1
X ' X  ( X ' X ) 1 X 'U  BX  BU 
   BX , [since E(U) = 0]...................................................(3.44)
Since our assumption regarding an alternative ˆ is that it is to be an unbiased
estimator of  , therefore,
(ˆ) should be equal to  ; in other (XB) should
be a null matrix. words
Thus we say, BX should be =0 if ( )  ( X ' X ) X 'BY is to be an unbiased

ˆ 1
estimator. Let us now find variance of this alternative estimator.

var(ˆ)  (ˆ   )(ˆ   )'
 
( X ' X ) X 'BY   ( X ' X ) X 'BY   '
1 1
 ( X ' X ) X 'B( X U )   ( X ' X ) X 'B( X U )   '

1 1
 [( X ' X ) X ' X  ( X ' X ) X 'U  BX  BU   

1 1
( X ' X ) X ' X  ( X ' X ) X 'U  BX  BU   '}

1 1
 [( X ' X ) X 'U  BU ( X ' X ) X 'U  BU '}

1 1
( BX
 0)

  ( X ' X )1 X 'U  BU U ' X ( X ' X ) 1 U ' B' 
 ( X ' X ) BUU 'X ( X ' X ) U ' B'
1 1
 ( X ' X ) X ' B(UU ' )X ( X ' X )  B'

1 1
2I
( X ' X ) 1

X 'B X ( X ' X ) 1  B' 
u n
 2
( X ' X ) 1
X ' X ( X ' X ) 1  BX ( X ' X )1  ( X ' X ) 1 
( X ' X ) 
u
 2 1
X ' X ( X ' X ) 1  BX ( X ' X )1  ( X ' X )1 X ' B'BB'
u
 2
( X ' X ) 1

 BB' (  0)
BX u
var(ˆ )  u2 ( X ' X12) u BB'……………………………………….(3.45)

Or, in other words,

var(ˆ) is greater than var(ˆ) by an expression  2 BB' and it
u
proves that ˆ is the best estimator.
3.4.3. Coefficient of Determination in Matrix Form

The coefficient of determination (R 2 ) can be derived in matrix form as follows.
We know
that e2 i  e'e  Y 'Y  2ˆ' X 'Y  ˆ' X ' since ( X ' X )ˆ  X 'Y and
X ˆ
Y 2
i
 Y Y
 e'e  Y 'Y  2ˆ' X 'Y  ˆ' X 'Y
e'e  Y 'Y  ˆ' X 'Y...................................................................................(3.46)
ˆ' X 'Y e' e  Y 'Y...................................................................................... (3.47)
We know, y  Y  Y
i i
1
 y 2  Y 2  (Y ) 2
i i i
n
In matrix notation
1 ......................................................................................
y 2  Y 'Y  (Y ) 2 (3.48)
i i
n
Equation (3.48) gives the total sum of squares variations in the model.
Explained sum of squares  y 2  e2
i i
1
 Y 'Y  (y)2  e' e
n
1 ............................................
 ˆ' X 'Y  (Y ) 2 (3.49)
i
n
Explained sum of squares
Since R 2  Total sum of squares
1
ˆ ' X 'Y (Y )i 2
n  ˆ ' X 'Y  nY
R2  ……………………(3.50)
1 Y 'Y  nY 2
Y 'Y (Y )i 2
n


Dear Students! We hope that from the discussion made so far on multiple
regression model, in general, you may make the following summary of results.
(i) Model: Estimators: Y  X U
(ii) Statistical properties: ˆ  ( X ' X )1 X 'Y
(iii) Variance-covariance: BLUE
(iv)
var(ˆ )  u2 ( X ' X1 )
(v) Estimation of (e'e): e'e  Y 'Y  ˆ' X 'Y
1
ˆ ' X 'Y (Y )i 2
n  ˆ ' X 'Y  nY
(vi)Coeff. of determination: R2 
1 Y 'Y  nY
Y 'Y ( Yi )2
n
3.5. Hypothesis Testing in Multiple Regression Model
In multiple regression models we will undertake two tests of significance. One is

significance of individual parameters of the model. This test of significance is the
same as the tests discussed in simple regression model. The second test is overall
significance of the model.
3.5.1. Tests of individual significance
If we invoke the assumption that U ~. N(0, 2 ) , then we can use either the t-test or
i
standard error test to test a hypothesis about any individual partial regression
coefficient. To illustrate consider the following example.
Let Y  ˆ0  ˆ
X1  ˆ 2 X  ei................................................................................................... (3.51)
1
2
A. H 0 : 1  0
H1 : 1  0
B. H 0 :  2  0
H1 :  2  0
The null hypothesis (A) states that, holding X2 constant X1 has no (linear)
influence on Y. Similarly hypothesis (B) states that holding X1 constant, X2 has no
influence on the dependent variable Yi.To test these null hypothesis we will use
the following tests:
i- Standard error test: under this and the following testing methods we
test only for ˆ
 .The test ˆ will be done in the same way.
1
ˆ for 2
var(ˆ ) e2
1
SE(1 )  ; where ˆ 2  i
̂ 22 x2i n3

 xx(x x)
22
1i 2i 1 2
2
 If SE(ˆ )  1 , we accept the null hypothesis that is, we can

2
ˆ
1 1
conclude that the estimate i is not statistically significant.

 If SE(  1 ˆ , we reject the null hypothesis that is, we can
2
ˆ
1 1
conclude that the estimate
i is statistically significant.
Note: The smaller the standard errors, the stronger the evidence that the estimates
are statistically reliable.
ii. The student’s t-test: We compute the t-ratio for each ˆ
i
ˆ i  
t*  ~ t n- , where n is number of observation and k is number of
SE( ˆ k
i
)
parameters. If we have 3 parameters, the degree of freedom will be n-3. So;
ˆ 2  
t* 
2
; with n-3 degree of freedom
SE( ˆ 2 )
In our null
 2  0, the t* becomes:
hypothesis
ˆ 2
t* 
SE( ˆ 2 )
 If t*<t (tabulated), we accept the null hypothesis, i.e. we can conclude
that
ˆ 2 is not significant and hence the regressor does not appear to
contribute to the explanation of the variations in Y.

 If t*>t (tabulated), we reject the null hypothesis and we accept the

alternative one; ˆ
 2 is statistically significant. Thus, the greater the value
of t* the stronger the evidence
i is statistically significant.
that
3.5.2 Test of Overall Significance

Through out the previous section we were concerned with testing the significance
of the estimated partial regression coefficients individually, i.e. under the separate
hypothesis that each of the true population partial regression coefficient was zero.
In this section we extend this idea to joint test of the relevance of all the included
explanatory variables. Now consider the following:
Y   0  1 X 1   2 X 2  .........   k X k  U i
H 0 : 1   2   3  ............   k  0
H1 : at least one of the k is non-zero
This null hypothesis is a joint hypothesis that 1,  2 ,........ k are jointly or
simultaneously equal to zero. A test of such a hypothesis is called a test of overall

significance of the observed or estimated regression line, that is, whether Y is
linearly related to X 1 , X 2 ,. X k .
Can the joint hypothesis be tested by testing the significance of individual
significance of î 's as the above? The answer is no, and the reasoning is
as follows.
In testing the individual significance of an observed partial regression coefficient,

we assumed implicitly that each test of significance was based on different (i.e.
independent) sample. Thus, in testing the significance of ˆ under the hypothesis
2
that
 2  0 , it was assumed tacitly that the testing was based on different sample


from the one used in testing the significance of

ˆ3 under the null hypothesis that
3  0 . But to test the joint hypothesis of the above, we shall be violating the
assumption underlying the test procedure.
“…..testing a series of single (individual) hypothesis is not equivalent to testing

those same hypothesis. The institutive reason for this is that in a joint test of
several hypotheses any single hypothesis is affected by the information in the
other hypothesis.”1
The test procedure for any set of hypothesis can be based on a comparison of the
sum of squared errors from the original, the unrestricted multiple regression
model to the sum of squared errors from a regression model in which the null
hypothesis is assumed to be true. When a null hypothesis is assumed to be true,
we in effect place conditions or constraints, on the values that the parameters can
take, and the sum of squared errors increases. The idea of the test is that if these
sum of squared errors are substantially different, then the assumption that the joint
null hypothesis is true has significantly reduced the ability of the model to fit the
data, and the data do not support the null hypothesis.
If the null hypothesis is true, we expect that the data are compliable with the
conditions placed on the parameters. Thus, there would be little change in the sum
of squared errors when the null hypothesis is assumed to be true.
Let the Restricted Residual Sum of Square (RRSS) be the sum of squared errors
in the model obtained by assuming that the null hypothesis is true and URSS be
the sum of the squared error of the original unrestricted model i.e. unrestricted
residual sum of square (URSS). It is always true that RRSS - URSS  0.
1
Gujurati, 3rd ed.pp

Consider Yˆ  ˆ 0  ˆ
X1
 ˆ 2 X  ......... ˆ k X  ei .
1
2 k
This model is called unrestricted. The test of joint hypothesis is that:

H 0 : 1   2   3  ............   k  0
H1 : at least one of the k is different from zero.
We know
Yˆ  ˆ 0  ˆ  X  ......... X ki
that: ˆ2 2i ˆk
X 1i
1
Yi  Yˆ  e
e i  Yi  Yˆ
i
e  (Y  Yˆ )2
2
i i i
This sum of squared error is called unrestricted residual sum of square (URSS).
This is the case when the null hypothesis is not true. If the null hypothesis is
assumed to be true, i.e. when all the slope coefficients are zero.
Y  ˆ 0  e i
ˆ 0  Y (applying OLS).............................................(3.52)

Y i
n
eY
but ˆ0  Y
 ˆ0
eY Y
e2  (Y  Yˆ ) 2  y 2  TSS
i i i
The sum of squared error when the null hypothesis is assumed to be true is called
Restricted Residual Sum of Square (RRSS) and this is equal to the total sum of
square (TSS).
The ratio: RRSS  URSS / K  1
~ F (k 1,nk ) ……………………… (3.53);
URSS / n  K
(has an F-ditribution with k-1 and n-k degrees of freedom for the numerator and denominator respectively)
RRSS  TSS
URSS  e2  y 2  ˆ yx  ˆ yx  ..........ˆ
yx  RSS
i 1 1 2 2 k k

(TSS  RSS) / k  1
F RSS / n  k
ESS / k  1....................................................................................
F  RSS / n  k (3.54)
If we divide the above numerator and denominator by y 2 

TSS then:
ESS
/k1
TSS
F  RSS
/kn
TSS
R2 /k 1
F
1R2/n …………………………………………..(3.55)
k
This implies the computed value of F can be calculated either as a ratio of ESS &
TSS or R2 & 1-R2. If the null hypothesis is not true, then the difference between
RRSS and URSS (TSS & RSS) becomes large, implying that the constraints
placed on the model by the null hypothesis have large effect on the ability of the
model to fit the data, and the value of F tends to be large. Thus, we reject the null
hypothesis if the F test static becomes too large. This value is compared with the
critical value of F which leaves the probability of in the upper tail of the F-
distribution with k-1 and n-k degree of freedom.
If the computed value of F is greater than the critical value of F (k-1, n-k), then the
parameters of the model are jointly significant or the dependent variable Y is
linearly related to the independent variables included in the model.
Application of Multiple Regression.

In order to help you understand the working of matrix algebra in the estimation of
the regression coefficient, variance of the coefficients and testing of the
parameters and the model, consider the following numerical example.
Example 1. Consider the data given in Table 2.1 below to fit a linear function:
Y    1 X 1   2 X 2   3 X 3 U


Prepared by: Bedru B. and Seid H. ( June, 2005)
Table: 2.1. Numerical example for the computation of the OLS estimators.
nY X1 X2 X 3yi x1 x2 x3 yi 2 x1 x2 x2 x3x1 x3 x1 2 x22 x32 x1 yi x2 yix3 yi
1 49 35 53 200 -3 -7 -9 0 9 63 0 0 49 81 0 21 27 0
2 40 35 53 212 -12 -7 -9 12 144 63 -108 -84 49 81 144 84 108 -144
3 41 38 50 211 -11 -4 -12 11 121 48 -132 -44 16 144 121 44 132 -121
4 46 40 64 212 -6 -2 2 12 36 -4 24 -24 4 4 144 12 -12 -72
5 52 40 70 203 0 -2 8 3 0 -16 24 -6 4 64 9 0 0 0
6 59 42 68 194 7 0 6 -6 49 0 -36 0 0 36 36 0 42 -42
7 53 44 59 194 1 2 -3 -6 1 -6 18 -12 4 9 36 2 -3 -06
8 61 46 73 188 9 4 11 -12 81 44 -132 -48 16 121 144 36 99 -108

9 55 50 59 196 3 8 -3 -4 9 -24 12 -32 64 9 16 24 -9 -12
1 64 50 71 190 12 8 9 -10 144 72 -90 -80 64 81 100 96 108 -120
0
520 420 620 2000
=0
Σ =0
Σ =0
Σ =0
94
Σx1x2 40
20
Σx1x3= 30
70
30
50
Σx3yi 19
Σx2yi 92
25
Σ
Σx3yi=
Σx2x3=
Σyi
Σx1
Σx2
Σx3
2
2
From the table, the means of the variables are computed and given below:
Y =52 ; X 1 =42 ; X 2 =62; X 3 =200
Based on the above table and model answer the following question.
i. Estimate the parameter estimators using the matrix approach
ii. Compute the variance of the parameters.
iii. Compute the coefficient of determination (R2)
iv. Report the regression result.
Solution:
In the matrix notation: ˆ  (x' x) 1 x' y ; (when we use the data in deviation form),
 
 1ˆ x x x 
ˆ ˆ 11 21 31
 x22 x32  ; so that
Where,    2 , x 
 x12
ˆ   : : : 
 3 x x x
 1n 2n 3n 
 x 2
x x x x  x y 
(x' x) 1
1 2 1
3 x y1
x1 x2 x 2
x2 x3  and x' y 
  
x x
1

x x x 2  x
 y2
 1 3 2 3 3   3 
(i) Substituting the relevant quantities from table 2.1 we have;
 270
(x' x)   240  330
 420 and
 319 
x' y   492 
240 630
   
 330  420 750  625

Note: the calculations may be made easier by taking 30 as common factor from
all the elements of matrix (x'x). This will not affect the final results.
 270 240  330
| x' x |  240 630  420  4716000
 
 330  420 750 
 0.0085
 0.0012 0.0031
(x' x)   0.0012
1
 0.0027 0.0009

 0.0031 0.0009 0.0032


 1ˆ 
ˆ   0.0085  0.0012 0.0031  319   0.2063 
ˆ 1      
   2  (x' x' y   0.0012 0.0027 0.0009  492 
  0.3309 
 x)
 ˆ   0.0031 0.0009 0.0032  625  0.5572
 3     
And
  Y  ˆ
 ˆ 2 X  ˆ 3 X 3
X1 1
2
 52  (0.2063)(42)  (0.3309)(62)  (0.5572)(200)

 52  8.6633  20.5139  111.4562  134.2789
(ii) The elements in the principal diagonal of (x' x)1 when multiplied
 2 give the variances of the regression parameters, i.e.,
u
var(ˆ )   2 (0.0085) 
1 u
 e2 17.11
var( )   (0.0027)ˆ 2  i
2
 2.851
ˆ 
ˆ2 u
 u nk 6
var( 3 )  2 (0.0032) 
u
var(ˆ  SE(ˆ
) ) 
1
ˆ
0.0243, var( 2 ) 0.1560 SE( ˆ )
1
2
ˆ
 0.0077, var( 3  0.0877 SE( ˆ
3
)  0.0093, )  0.0962
1
ˆ' X 'Y  (Y ) 2
(iii) ˆ
R2 n
i
ˆ x y  ˆ y   x 
575.98
 0.97
  1 1 y
x 2 2 3 3
1 y 2
594
Y 'Y  (Yi )
n 2 i
(iv) The estimated relation may be put in the following form:

Yˆ  134.28  0.2063X1  0.3309X 2  0.5572X 3
SE( î ) (0.1560) (0.0877) (0.0962) R 2  0.97

t* (1.3221) (3.7719) (5.7949)
The X 1 , X and X 3 explain 97 percent of total variations.
variables 2
We can test the significance of individual parameters using the student's t-test.
The computed value of ‘t' is given above as t * .these values indicates us only
1
insignificant. ˆ is

Example 2. The following matrix gives the variances and covariance of the of the
three variables:
y x1 x2
y 7.59 3.12 26.99 

x  
1
29.16 30.80 
x2    133.00
The first raw and the first column of the above matrix shows y and the first
2
raw and the second column

shows  and so on.
yx1i
Consider the following model
Y  AY 1 Y  2 e vi
1 2 2
Where; Y1 is food consumption per capita

Y2 is food price
Y3 is disposable income per capita
And
Y  ln , X 1  ln and X 2  ln Y3
Y1 Y2
y  Y  Y , x1  X  X , and x2  X  X
Using the values in above matrix answer the following questions.

a. Estimate
1 and  2
b. Compute variance
ˆ and ˆ
of 2
1
c. Compute coefficient of determination
d. Report the regression result.
Solution: It is difficult to estimate the above model as it is, to estimate the above
model easily let's take the natural log of the above model;
ln Y1  ln A  1 ln Y2   2 ln Y3 V i
And let:
 0  ln A Y  ln , X 1  ln and X 2  ln the above model
, Y1 Y2 Y3
becomes:
Y   0  X 1  X 2 V i
The above matrix is based on the transformed model. Using values in the matrix
we can now estimate the parameters of the original model.

We know
ˆ  (x' x) 1 x' y
that
In the present question:
x11 x12 
ˆ  x x 
ˆ   1 and , x   21 22 

  ˆ 2   : : 
 x1n x nn 
 x 2
1 x x  x1 y 
 x' x   1 2
and x' y   
2 
x1 x2 x  x2 y
2
Substituting the relevant quantities from the given variance-covariance matrix, we

obtain:
30.80   3.12 
133.00 and
x' x 
29.16 30.80 x' y  26.99 
   
| x' x | 30.80  2929.64

29.16 133.00
30.80
1  30.80  0.0105
(x' x)1 133.00  0.0454

2929.64    29.16    0.0099 
  

 ˆ 
(a) ˆ  (x' x) 1 x' y   1 
  ˆ 2  
 0.0454  0.01053.12   0.1421 
 0.0105 0.0099 26.99  0.2358
    
(b). The element in the principal diagonal of (x' x)1 when multiplied by  2 give
u
the variances of the ˆ and ˆ .
(c). 2 ˆx 2 y  ˆx3 y

R  y 2
i
 (0.1421)(3.12)  (0.2358)(26.99)
 7.59
 R 2  0.78; e2  (1  R 2 )(y 2 )  1.6680
i i
1.6680
 ˆ 2   0.0981

u
17

var(ˆ )  (0.0981)(0.0454)  0.0045, (ˆ )SE  0.0667
var(ˆ)  (0.0981)(0.0099)  0.0009, (ˆ)SE  0.0312
(d). The results may be put in the following form:

Y1ˆ  AY
(0.1421) (0.2358)
1 Y3
SE (0.0667)(0.0312) R 2  0.78
t* (2.13) (7.55)
The (constant) food price elasticity is negative but income elasticity is positive.
Also income elasticity if highly significant. About 78 percent of the variations in
the consumption of food are explained by its price and income of the consumer.
Example 3:
Consider the model:
Y    1 X 1i   2 X 2i U i
On the basis of the information given below answer the following question
X 2 
32001 X 1 X  X 2  400
2 4300
X 2  X 1Y  X 2Y  13500
73002 8400
Y  800 X 1  n  25
Y 2 250
i
28,000
a. Find the OLS estimate of the slope coefficient

b. Compute variance of ˆ
2
c. Test the significant
of  slope parameter at 5% level of significant
2
d. Compute R 2 R 2 and interpret the result

and
e. Test the overall significance of the
model Solution:
a. Since the above model is a two explanatory variable model, we can estimate
ˆ and
1 ˆ using the formula in equation (3.21) and (3.22) i.e.
2

x yx 2  x yx x
ˆ1  x 2x 2  (x x )2
21 22 21 2 1

x yx 2  x yx x
ˆ2  x2 2x 21  (x 1x ) 2 1
1 2 1 2
Since the x's and y's in the above formula are in deviation form we have to find
the corresponding deviation forms of the above given values.
We know that:
x1 x2  X 1 X 2  nX 1 X 2
 4300  (25)(10)(16)
 300
x1 y  X 1Y  nX 1Y
 8400  25(10)(32)
 400
x2 y  X 2Y  nX 2Y
 13500  25(16)(32)
 700
x  X 2  nX 2
2
1 1 1
 3200  25(10) 2
 700
x  X 2  nX 2
2
2 2 2
 7300  25(16)2
 900
Now we can compute the parameters.
x yx 2  x yx x
ˆ1  x2 2x 22  (x 2x ) 2 1
1 2 1 2
(400)(900)  (700)(300)
 (900)(700)  (300)2
 0.278
x yx 2  x yx x
ˆ2  x2 2x 21  (x 1x ) 2 1
1 2 1 2

(700)(700)  (400)(300)
 (900)(700)  (300)2
 0.685
The intercept parameter can be computed using the following formula.

 ˆ  Y  ˆ X 1  ˆ 2 X 2
1
 32  (0.278)(10)  (0.685)(16)
 18.26
ˆ 2x 2
b. var( )  2
ˆ
1
x 2x 2  (x x ) 2
1 2 1 2
e2
 ˆ 2  i Where k is the number of parameter
nk
In our case k=3
e2
 ˆ 2  i
n3
e2  y 2  ˆ x y  ˆ x y
1 1 1 2 2
 2400  0.278(400)  (0.685)(700)

 1809.3
2

ˆ  i 2
n3
1809.3
 25  3
 82.24
(82.24)(900)
 var(ˆ )   0.137
1
540,000
SE(ˆ) var(ˆ ) 0.137

1  0.370

ˆ 2x 2
ˆ
var( 2 )  1
x 2x2  (x x )2
1 1 1 2
(82.24)(700)
 540,000  0.1067
SE(ˆ) var(ˆ 2 )  0.1067  0.327




c. ˆ can be tested using students t-test

1
This is done by comparing the computed value of t and critical value of t which is
obtained from the table at 
2 level of significance and n-k degree of freedom.
Hence; t* 0.278
  0.751
SE(ˆ1 0.370
)
The critical value of t from the t-table at 
2  0.05
2  0.025 level of significance
22 degree of freedom is 2.074. and
tc  2.074
t*  0.755
 t*  tc
The decision rule if

t*  tc is to reject the alternative hypothesis that says  is
different from zero and to accept the null hypothesis that says  is equal to zero.
The conclusion is
ˆ is statistically insignificant or the sample we use to estimate
1
ˆ is drawn from the population of Y & X1in which there is no relationship
1
between Y and X1(i.e.
1  0 ).
d. R 2 can be easily using the following equation
2
ESS
R RSS
 1- TSS
TSS
We know that RSS  e2
and TSS  y 2 and ESS i yˆ 2  ˆ x y  ˆ x y  ......  ˆ x y
1 1 2 2 k k
For two explanatory variable model:
R 2  1-
RSS 10809.3
 1  2400
TSS
 0.24
 24% of the total variation in Y is explained by the regression line
Yˆ  18.26  0.278X 1  0.685X 2 ) or by the explanatory variables (X1 and X2).


e2i / n  k (1  R 2 )(n  1)
Adjusted R 2
 1   1 
y 2 / n  1 nk
(1  0.24)(24)
1 22
 0.178
e. Let's set first the joint hypothesis as
H 0 : 1   2  0
against
H1 : at least one of the slope parameter is different from zero.
The joint test hypothesis is testing using the F-test given below.
F *(k 1),(nk ) 
ESS / k 1
RSS / n  k
R2 /k
1 1  R 2 / n 
k
From (d)
R2  and k  3
0.24
F *(2,22) 
3.4736 this is the computed value of F. Let's compare this
with the critical value F at 5% level of significance and (3,.23) numerator and
denominator respectively. F (2,22) at 5%level of significance = 3.44.
F*(2,22) = 3.47
Fc(2,22)=3.44
 F*>Fc, the decision rule is to reject H 0 and accept H1. We can say that
the model is significant i.e. the dependent variable is, at least, linearly
related to one of the explanatory variables.


Sample exam questions :

Mid-exam (May, 2004/2005)
 Instructions:
Read the following instructions carefully.
 Make sure that your exam paper contains 4 pages
 The exam has four parts. Attempt
 All questions of part one
 Only two questions from part two
 One question from part three
 And the question in part four.
 Maximum weight of the exam is
40% Part One: Attempt all of the following questions
(15pts).
1. Discuss briefly the goals of econometrics.
2. Researcher is using data for a sample of 10 observations to estimate the relation
between consumption expenditure and income. Preliminary analysis of the sample data
produces the following data.
 xy  700 ,  x 2
 1000 , X  100
Y  200
Where
x Xi X and y  Yi  Y
i
a. Use the above information to compute OLS estimates of the intercept and slope
coefficients and interpret the result
b. Calculate the variance of the slope parameter
c. Compute the value R2 (coefficient of determination) and interpret the result
d. Compute 95% confidence interval for the slope parameter
e. Test the significance of the slope parameter at 5% level of confidence using t-test
3. If the model Yi=+1X1i +2X2i +Ui is to be estimated from a sample of 20 observation using
the semi- processed data given in matrix in deviation form.
(xx)1  0.5  0.08

0.08 0.6
xy 
100
X1 X 2 =25 and Y  30
250 =10,

Obtain the OLS estimate of the above parameters.

4. Linearity is one assumption of classicalist in simple regression analysis. Identify which of the
following satisfies this assumption. Discuss why?
a. LnY2=+1X1i +2X2i +Ui
b. Y=+(1/1) X1i +2X2i +Ui
c. Y=+1X21i +2X2i +Ui
d. LnY=+1LnX1i +2LnX2i +Ui
e. Yi=+2Xi + Ui

Part Two: Attempt any two of the following questions.(10pts). 1.

Consider the model
Yi=+Xi + Ui
Show that the OLS estimate of is unbiased.
2. Suppose  2 is the population variance of the error term and ˆ 2 is estimator of 2 .
Show that ˆ 2 of maximum likely hood is the biased estimator of the true 2 for the model
Yi=+Xi + Ui.
3. In the model
Yi=+Xi + Ui
Show that ˆ =Y  ˆX possesses minimum variance.
4. using the assumptions of simple regression model show that
a. YN (+Xi, 2)
b. Cov ( ˆ , ˆ ) = - X Var( ˆ )
For the model Yi=+Xi + Ui

Part Three: Attempt any one the Following(10 pts.)
1. The model Yi=+1X1i +2X2i +3X3i +Ui is to be estimated from a sample of 20
observations. Using the information below obtain the OLS estimate of the parameters of the
above model.
10,000
0.1  0.12  0.03
(x / x)1   0.12 0.04 0.02 20,300
X Y 10,100
 0.03 0.02 0.08
30,200
X 1  400 ,  X 2  200 , and X 3  600
Where
x Xi X and y  Yi  Y
i
2. In a study of 100 firms, the total cost(C) was assumed to be dependent on the rate of out put
(X1) and the rate of absenteeism (X2). The means were: C  6 X 1  3 X 2  4 . The matrix
, and
showing sums of squares and cross products adjusted for means is
c x1 x2
c 100 50 40
x1 50 50 -70 where, xi  X i  X i and c  Ci  C
x2 40 -70 900
Estimate the linear relation ship between C and the other two variables. (10points)
3. Consider the linear regression model

Yi=+1Xi +Ui
Suppose that there are no data on Xi but we have data on Zi=ao+a1X1, whrer ao and a1 are
arbitrary known constants. Using data of variable Zi we can estimate
Yi=co+c1Zi +Ui
Show how from the estimates of cô and cˆ1 you can obtain the estimates of the original
model.


Econometrics Module

Uploaded by

Copyright:

Available Formats

Econometrics Module

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Econometrics Module

Uploaded by

Copyright:

Available Formats

lOMoARcPSD|6162234

Introduction to Econometrics (Mekelle

Scan to open on Studocu

Downloaded by Demlie Basha Hailu

Downloaded by Demlie Basha Hailu

Downloaded by Demlie Basha Hailu

Downloaded by Demlie Basha Hailu

1.1 Definition and scope of econometrics

The economic theories we learn in various economics courses suggest many

Each of such specifications involves a relationship among economic variables. As

However, economic theories that postulate the relationships between economic

Downloaded by Demlie Basha Hailu

“Econometrics is the science which integrates economic theory, economic

Measurement is an important aspect of econometrics. However, the scope of

Downloaded by Demlie Basha Hailu

In short, econometrics may be considered as the integration of economics,

1.2 Econometrics vs. mathematical economics

Econometrics differs from mathematical economics in that, although econometrics

1.3 Econometrics vs. statistics

Downloaded by Demlie Basha Hailu

Mathematical (or inferential) statistics deals with the method of measurement

1.4 Economic models vs. econometric models

Downloaded by Demlie Basha Hailu

functioning of an economic entity under a set of simplifying assumptions. All

ii) Econometric models:

1.5 Methodology of econometrics

Downloaded by Demlie Basha Hailu

econometric techniques are relationships in which some variables are postulated as

1. Specification of the model

Note: The specification of the econometric model will be based on economic

Downloaded by Demlie Basha Hailu

committing errors or incorrectly specifying the model. Some of the common

2. Estimation of the model

Downloaded by Demlie Basha Hailu

3. Evaluation of the estimates

4) Evaluation of the forecasting power of the model:

Downloaded by Demlie Basha Hailu

1.6 Desirable properties of an econometric model

1. Theoretical plausibility. The model should be compatible with the

2. Explanatory ability. The model should be able to explain the observations

3. Accuracy of the estimates of the parameters. The estimates of the

Downloaded by Demlie Basha Hailu

4. Forecasting ability. The model should produce satisfactory predictions of

5. Simplicity. The model should represent the economic relationships with

1.7 Goals of Econometrics

 How would you define econometrics?

Downloaded by Demlie Basha Hailu

In this chapter we shall consider a simple linear regression model, i.e. a

2.1. Stochastic and Non-stochastic Relationships

A relationship between X and Y, characterized as Y = f(X) is said to be

Downloaded by Demlie Basha Hailu

Downloaded by Demlie Basha Hailu

In order to take into account the above sources of errors we introduce in

Downloaded by Demlie Basha Hailu

disturbance, we observe Y ,Y ,......,Y corresponding to X , X ,...., X . These

points diverge from the regression line by u1 , u2 ,...., un .

2.2.1 Assumptions of the Classical Linear Stochastic Regression Model.

1. The model is linear in parameters.

Example 1. Y    x  is linear in both parameters and the variables, so it