Introduction and subject matter of Econometrics

Lesson: Introduction and subject matter of

Econometrics

Lesson Developer: Neha Goel

Shyamlal College, University of Delhi

1
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

Table of Contents

 1: Learning Outcomes
 2: Introduction
 3.1: Definition
 3.2: Scope
 3: Methodology or steps in an Econometric study
 3.1: Creation of hypothesis or statement of theory
 3.2: Data collection
 3.3: Specification of the mathematical economic model
 3.4: Specification of the statistical/econometric model
 3.5: Estimation of the parameters of the chosen econometric
model
 3.6: Testing of model specification
 3.7: Testing and analyzing the hypothesis derived from the
model

2
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

 3.8: Forecasting or predicting values using the model

 4: Summary
 5: Exercises
 6: References
 7: Quiz
 8: MCQ

1. Learning Outcomes

After reading this chapter, you will be able to:

 Explain the meaning of Econometrics

 Understand the methodology in an Econometric study and promote
development of an understanding of the statistical properties of such
methodology
 Understand why we study Econometrics
 Understand the importance and strong intuitive knowledge of Econometrics
study

3
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

 Provide practical skills required to fir models for a given data

 Improve your understanding of how an economy works at both
microeconomic and macroeconomic level.

2. Introduction

“Econometrics may be stated as the statistical methods used by economists”.

Definition:

According to Arthur S. Goldberger, Econometrics is the social science which applies the
tools of economic theory, mathematics and statistical inference to analyze the economic
phenomena.

According to P. A. Samuelson, T.C. Koopmans, and J.R.N. Stone, Econometrics consists

of the application of mathematical statistics to economic data to lend empirical support to
the models constructed by mathematical economics and to obtain numerical results,
resulting out of a certain outlook on the role of economics.

Econometrics is derived from mathematics and statistics (mainly regression and trending
techniques).

Scope:

From the above definitions we can say that econometrics applies mathematical statistics,
economic statistics, economic theory and mathematical economics. Main aim of
econometrics is to test a hypothesized relationship between an independent (predictor) and
a dependent (predicted) variable. Following are the reasons for why study econometrics:

Mathematical Statistics provides data that cannot be controlled directly. Most economic
data are unique in nature requiring special methods to be developed by the econometrician.
For example, some non-experimental data collected by the public and private agencies, like
data on income, consumption, prices, savings, investments etc. The Econometrician
assumes such data as given. Mathematical statistics normally fail to deal with such special
problems as these data generally consists of measurement errors like omission of relevant

4
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

variables or commission of irrelevant variables. Thus, such measurement errors may be

rectified if econometricians develop special methods of analysis for its rectification.

Economic statistics is used to collect, process, and present economic data in the form of
tables, charts and diagrams. A statistician uses this statistics to collect data on employment,
unemployment, GDP, prices etc. for econometric work.

Economic theory creates hypotheses or statements which are generally qualitative in

nature. But there is no provision for any quantitative measure of the strength of the relation
between variables. Thus, such numerical estimates are provided by an econometrician.
Econometrics is useful for the provision of empirical study to most economic theory.

Mathematical economics expresses economic theory in equations or mathematical forms

without paying heed to empirical verification or measurability of the theory. However, this
empirical verification of the economic theory is provided by econometrics.

Econometrics would be beneficial for students majoring in business and economics as they
can be asked to estimate demand and supply functions or to forecast money supply, sales,
interest rates or to estimate price elasticity for products in their job. Thus, econometrics is
an integral part of their training.

5
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

Types of Econometrics:

1. Theoretical
2. Applied

1. Theoretical:

Theoretical econometrics is used to develop relevant methods to measure the economic

relationships explained by the econometrics models. It depends on such methods to
measure the economic relationship, single and simultaneous equation techniques,
mathematical statistics etc.

Simple equation: Y = a + bx

Simultaneous equation: Y = a+ b +b

2. Applied:

It is used to obtain practical values of our economic research. It uses the econometric
methods from theoretical econometrics in the application to different fields of economic
study like demand and supply, consumption function etc. It is due to applied econometrics
that researchers have got numerical results to such fields of important economic studies.

3. Methodology or steps in an Econometric study

Following are the steps to do an econometric study:

3.1 Creation of hypothesis or statement of theory:

Development of a model generally depends on economic theory. We need to find the

economic theory of the subject we want to study. Take, for example, the demand
relationship. Suppose we want to see the relationship between price of a commodity and
quantity demanded. We know that as the price of the commodity increases, its demand
decreases. Thus, there is a negative relationship between price and demand of a commodity.
However, we know that demand of a commodity also depends upon price of other

6
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

commodities, relative income, taste and preferences of consumers etc. We assume these
factors to be constant or ceteris paribus. But how do we find out this relationship? This
raises empirical question.

3.2 Collection of data:

Quantitative information is required on the two variables for empirical purposes. For this
purpose, three types of data are generally available:

1. Time Series data: Collection of such data is done over a period of time. It
has separate observation for each time period. For example, data on money
supply, employment and unemployment, GDP, government deficit etc. is
collected at some intervals regularly; such as share prices are recorded daily,
money supply is measured weekly, unemployment rate is calculated monthly,
GDP is calculated quarterly and government budget is announced annually.
The data collected may be qualitative (like gender, race, marital status,
employment status etc.) or quantitative (like money supply, income, prices
etc.) in nature. Trends and seasonality is important in such type of data.
2. Cross-sectional data: Collection of such data on one or more than one
variables is done at one point in time. For example, government of India
collects census of population every 10 years.

It is usually a random sample from some population.

3. Pooled data: It has the characteristics of both time series and cross-
sectional data. For instance, data collected on unemployment rate for 20
economies for a period of 10 years, will constitute an example of pooled data
– time series data is the data on the unemployment rate for each economy
for the 10-year period, while cross-sectional data is the data on the
unemployment rate for the 20 countries. Thus, there are 200 observations in
this pooled data.

Panel Data: it is a special kind of pooled data and is also known as micro-
panel or longitudinal data. In this type of data, same cross-sectional unit,
such as household or firm, is studied over a period of time. For instance, the

7
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

government of India collects census of population of the most populous area

in India. At each periodic survey, the same area is studied and interviewed to
check out if there has been any change in the population.

3.3 Specification of the mathematical economic model:

Development of the model is sometimes based on results of previous empirical studies. A

mathematical economic model must specify-

a) All the behavioral and statistical relations, identities and equilibrium

conditions
b) The functional forms of each relationship
c) All the variables to be included in the model
d) Limitations of the model.

Suppose we take the demand relationship and define it as-

q = a + bp

To see how quantity demanded (q) behaves in relation to price (p) of the commodity, we
take an example of demand for sports clubs and its price and then draw a scatter diagram:

Demand and price data for sports club:

8
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

The scatter diagram above shows that price and quantity demanded are inversely related i.e.
there is a downward trend. Using approximation, a straight line has been drawn through the
scatter points and a relationship between price and quantity demanded is written by the
using simple mathematical model as follows:

q=α+βp

The above equation states that quantity (q) holds linear relationship with price (p), α and β
are called the parameters of the linear function. Α is also called the intercept and it gives
the value of q when p is 0. Whereas, β is called the slope and it shows the rate of change in
q per unit change in p. The slope coefficient can be positive or negative, depending on the
relationship between the two variables. In the above example, it is negative.

3.4 Specification of the statistical/econometric model:

It is assumed by the mathematical model that the relationships between two variables
under consideration are exact or deterministic, but actually they are not neat or exact as
the relationships are statistical in nature. Let us continue with the price and quantity
demanded example:
q=α+βp

9
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

We know that quantity demanded also depends upon many random factors like prices of
other commodities, relative income, changing tastes and preferences etc. However, all of
these factors cannot be captured in the model above, and many of them cannot even be
measured. From the scatter diagram drawn above we can say that the relationship between
price and quantity is not exact or linear as all the data points in the diagram do not lie
exactly on the straight line. Thus, we can conclude that the value of q on the basis of
information of price alone will be greater or less than real world value and will be, only by
chance, equal to it. Hence, we can re-write the above equation as:

q=α+βp+u

This is the linear regression model (econometric model) where ‘u’ is a random variable
known as the disturbance or the error term. ‘u’ represents the effect of factors disturbing
the mathematical relationship. This error term represents all those forces which affect ‘q’
but are not explicitly included in the model. Regression analysis is a technique that
models and investigates the relationship between variables. It is the study of
dependence of one dependent variable on one or more explanatory variables.

Simple linear regression model is used to describe linear relationship of one variable on
other. It predicts values of one variable from given values of other variables. Through a
scatter-plot of diagram, it determines the best-fit line, minimizing the sum of squared
residuals and the error variance. Simple linear regression model is represented by:

= + +

10
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

Linear regression states that there is just one line (among many) which best explains the
trend and the relationship between the two variables (with minimum error). Linear
regression model is the prime subject of econometrics. Variable on the L.H.S. i.e. ‘q’ is
called the dependent variable while that on the R.H.S i.e. ‘p’ is called the independent or
explanatory variable as it explains the variation in ‘q’. In regression analysis, we study the
effect of one variable (independent) on the other variable (dependent). However, we should
keep in mind that the relationship between both the variables is not causal. Thus, we can
call the relationship as a predictive relationship, but can we predict ‘q’, for a given ‘p’?

3.5 Estimation of the parameters of the chosen econometric model:

For computation and estimation, we develop the method of Ordinary least squares (OLS).
OLS is an algorithm that defines the values of q, p and u such that the distance between the
actual and the predicted values of q, p and u are at its minimum. Taking the data from the
above demand and price example and using the gretl software, we get the following results:

Model 1: OLS
q

11
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

Coefficient Std. Error t-statistic p-value

Constant 211.315 14.8139 14.2647 <0.00001 ***

(α)

p −0.354631 0.0364517 -9.7288 <0.00001 ***

Mean dependent var 69.16667 S.D. dependent var 26.10062

Sum squared resid 716.0733 S.E. of regression 8.462111

R-squared 0.904443 Adjusted R-squared 0.894887

F(1, 10) 94.64944 P-value(F) 2.04e-06

The estimated regression equation may be written as:

= 211.315 – 0.354631

By applying OLS, we see that the above equation is found to be the best-fitting regression
line as it is linear, unbiased and minimizes the distance between the line and the data points.

12
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

Where and are estimates of q and p respectively. The estimated value of α is 211.315
and that of β is 0.354631. The symbol hat, i.e. ‘^’, is used to denote the estimate. The
interpretation of this equation is: if the price increases by one unit, ceteris paribus, the
quantity demanded is expected to decrease on an average by about 0.35 units. We use the
words ‘on an average’ since the relationship between the dependent and the independent
variable becomes somewhat imprecise due to the presence of the error term u. Thus, the
estimated regression line gives the relationship between average dependent variable and
independent variable i.e., how ‘q’ reacts to a per unit change in ‘p’. ‘α’ value 211.315 gives
the average value of quantity demanded, if the price were zero, i.e. 211.315 units.

β = 0.354631 = if price increases by one unit, quantity decreases by 0.35 units.

α = 211.315 = if price is zero, demand increases by 211 units.

R-squared = 0.904443 shows that model is a good it. If is greater than 0.75, we
normally assume model to be a good fit. We will read it in details in further chapters.

3.6 Testing of model specification:

Taking the demand and price relationship, we need to find out how adequate is our model.
We know that the demand for a particular commodity depends on price of that commodity,
income of the individual, prices of related commodities etc. Thus the relationship can be
represented by:

=f( , , , …., , Y)

Where is the quantity demanded of commodity i, is the price of commodity i, is the

price of other commodity j, and Y is income of the individual. The mathematical relation
thus becomes:

= + + + Y+ u

Where summarizes the effect of all the factors which are not supposed to change during
the period under study. The above equation is called a multiple regression model, in
contrast to the initial simple linear regression or two-variable/bi-variate model. Previously in
two-variable regression model, there was one explanatory variable, whereas, now there are
several explanatory variables. Here we again added the disturbance term ‘u’ because even if
we keep on adding multiple explanatory variables, we cannot clearly explain the effect of

13
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

the dependent variable in the model. The number of explanatory variables to be included in
the model is the decision of the researcher. The model chosen should clearly reflect the
economic phenomena. However, the underlying economic theory should explain what these
variables might be. We can estimate the above multiple regression in the same way as we
did for simple regression model, for a given data.

3.7 Testing and analyzing the hypothesis derived from the model:

After choosing an appropriate model, we may want to test the hypothesis derived from by
economic phenomena and / or some priori experience. For instance, there is a negative
relationship between price of a commodity and quantity demanded of a commodity. Is this
hypothesis borne out by our results? As per our estimation in the example above, the
estimated coefficient of price was negative, thus our statistical results goes hand in hand
with the hypothesis.

3.8 Forecasting or predicting values using the model:

We estimated the regression model for prediction or forecasting. Suppose we have 2015
data on the demand and price of the sports club. Suppose the price of the sports club was
500 units in 2015. If we substitute this value in the estimated regression line:

= 211.315 – 0.354631

We get 33.9995 units as the predicted value of quantity demanded for sports club in the
year 2015 i.e. if price of the sports club in 2010 was 500 units, demand for sports club
would be approximately 34 units. When the real data on price and quantity of sports club
will be available for 2015, we can compare the predicted value with the real value. Any
difference between the two values will give the predicted error. The statisticians try to keep
this predicted error as small as possible.

Like we examined econometric methodology using price and demand relationship, we can
use similar procedure to predict the quantitative relationships between variables in any
theory of economics, politics, sociology, psychology, meteorology, international relations etc.

14
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

4. Summary:

 Econometrics may be stated as the statistical methods used by

economists.
 Econometrics uses mathematical statistics, economic statistics, economic
theory and mathematical economics
 Development of an econometric model generally depends on economic
theory. For example, relationship between price and quantity demanded
of a product.
 Quantitative information is required on the variables for empirical
purposes. For this purpose, time-series, cross-sectional and pooled data
are generally available.

15
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

 Development of the model is sometimes based on results of previous

empirical studies. A mathematical economic model for price and demand
relationship may be given by: q = a + bp.
 Econometric model for the price and quantity demanded relationship may
be given by: q = α + β p
 q = α + β p + u is the linear regression model (econometric model)
where ‘u’ is the disturbance or the error term. ‘u’ represents the effect of
factors disturbing the mathematical relationship.
 The estimated regression equation may be written as: = 211.315 –
0.354631 , for a given data.
 To check the adequacy of the price and demand model, we introduce
other factors that affects demand like income, price of other commodities
etc. The mathematical relation thus becomes: = + + +
Y+ u. Where summarizes the effect of all the factors which are not
supposed to change during the period under study. The above equation is
called a multiple regression model
 After choosing an appropriate model, we may want to test the hypothesis
explained by the economic phenomena and / or a prior experience. We
can do this by testing whether < 0 or > 0. This testing will be taught
in further chapters.
 We estimated the regression model for prediction or forecasting the
unknown or future values.

5. Exercise:

1. Define Econometrics.
2. What is the scope of Econometric study?
3. What are the steps involved in the methodology of Econometric study?

16
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

4. Distinguish between simple and multiple regression line.

5. Define the parameters α, β and ε in the regression equation: Y = α + βx + ε.
6. What are the steps in the traditional methodology of econometrics? Illustrate
using an example.
7. Explain the difference between a statistical and a deterministic relationship?
8. Explain the difference between cross-sectional data and time series data.
Give examples of each.
9. What are pooled data? What are panel, longitudinal, or micropanel data?
10.Explain the difference between experimental and non-experimental data.
11.What is the population regression line? Illustrate in a diagram.
12.What is the population regression function?
13.Explain the difference between linear in the variables and linear in the
parameters. Give examples of each. Explain the difference between linear in
the variables and linear in the parameters. Give examples of each.
14.What is the “stochastic specification of the population regression function”?
Explain.
15.Give four reasons why we introduce a stochastic disturbance term.
16.What is the difference between a parameter and an estimate of a regression
function?
17.What does it mean when we say that an estimator is an unbiased estimator?
18.What does it mean when we say that the least squares estimator is the best
linear unbiased estimator?
19.What does it mean when we say that an estimator is a consistent estimator?
20.Break down the variation of Yi into two components. Illustrate.

17
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

6. References:

D.N. Gujrati, Essentials of Econometrics, 3rd Edition, McGraw Hill International

Edition

7. Quiz:

1. In the regression specification y = α+βx +ε

a) y is called the dependent variable or the regressand, and x is called the

regressor
b) y is called the dependent variable or the regressor, and x is called the
regressand
c) y is called the independent variable or the regressand, and x is called the
regressor
d) y is called the independent variable or the regressor, and x is called the
regressand
2. In the regression specification y = α+βx +ε

a) α is called the intercept, β is called the slope, and ε is called the residual
b) α is called the slope, β is called the intercept, and εis called the residual
c) α is called the intercept, β is called the slope, and ε is called the error
d) α is called the slope, β is called the intercept, and ε is called the error

3. In the regression specification y = α + βx + ε. which of the following is not

a justification for epsilon (ε)?

a) it captures the influence of a million omitted explanatory variables

b) it incorporates measurement error in x
c) it reflects human random behavior
d) it accounts for nonlinearities in the functional form

4. In the regression specification y = α + βx + ε, if the expected value of epsilon

is affixed number but not zero

a) the regression cannot be run

b) the regression is without a reasonable interpretation
c) this non-zero value is accommodated by the βx term
d) this non-zero value is incorporated into α

5. In the regression specification y = α + βx + ε, the conditional expectation of y

is

18
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

a) the average of the sample y values

b) the average of the sample y values corresponding to a specific x
value
c) α + βx
d) α + βx + ε

6. In the regression specification y = α + βx + ε, the expected value of y

conditional on x=1 is

a) the average of the sample y values corresponding to x=1

b) α+β+ε
c) β
d) α

7. In the regression specification y = α + βx + δz + ε, the parameter β is

interpreted as the amount by which y changes when x increases by one and

a) z does not change

b) z changes by one
c) z changes by the amount it usually changes whenever x increases by
one
d) none of the above

8. In the regression specification y = α + βx + δz + ε, the parameter α is called

a) the slope coefficient

b) the intercept
c) the constant term
d) both b) and c) above

9. The terminology ceteris paribus means

a) all else equal

b) changing everything else by the amount by which they usually
change
c) changing everything else by equal amounts
d) none of the above

10.In the regression specification y = α + βx + δz + ε, the parameter βis called

a) the slope coefficient

b) the intercept
c) the constant term
d) both b) and c) above

19
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

11.Which of the following are alternative names for the dependent variable
(usually denoted by y) in linear regression analysis?

i) The regressand
ii) The regressor
iii) The explained variable
iv) The explanatory variable

a) (ii) and (iv) only

b) (i) and (iii) only
c) (i), (ii) and (iii) only
d) All (i), (ii), (iii) and (iv)

12.The residual from a standard regression model is defined as

a) The difference between the actual value, y, and the mean, y-bar
b) The difference between the fitted value, y-hat, and the mean, y-bar
c) The difference between the actual value, y, and the fitted value, y-hat
d) The square of the difference between the fitted value, y-hat, and the
mean, y-bar

Answers:

1. a
2. a
3. a
4. a
5. b
6. a
7. a
8. b
9. a
10. a
11. b
12. c

8. MCQ:

1. The regression model includes a random error or disturbance term for a variety of

20
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

reasons. Which of the following is NOT one of them?

a. measurement errors in the observed variables

b. omitted influences on Y (other than X)
c. linear functional form is only an approximation
d. the observable variables do not exactly correspond with their
theoretical counterparts
e. there may be approximation errors in the calculation of the least
squares estimates

2. Which of the following assumptions about the error term is not part of the so called

"classical assumptions"?

a. it has a mean of zero

b. it has a constant variance
c. its value for any observation is independent of its value for any other
observation
d. it is independent of the value of X
e. it has a normal distribution

3. Which of the following is NOT true?

a. the point Xbar, Ybar always lies on the regression line
b. the sum of the residuals is always zero
c. the mean of the fitted values of Y is the same as the observed values
of Y
d. there are always as many points above the fitted line as there are
below it
e. the regression line minimises the sum of the squared residuals
4. In a simple linear regression model the slope coefficient measures
a. the elasticity of Y with respect to X

21
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

b. the change in Y which the model predicts for a unit change in X

c. the change in X which the model predicts for a unit change in Y
d. the ratio Y/X
e. the value of Y for any given value of X
5. Changing the units of measurement of the Y variable will affect all but which
one of the following?
a. the estimated intercept parameter
b. the estimated slope parameter
c. the Total Sum of Squares for the regression
d. R squared for the regression
e. the estimated standard errors
6. A fitted regression equation is given by Yhat = 20 + 0.75X. What is the value
of the residual at the point X=100, Y=90?
a. 5
b. -5
c. 0
d. 15
e. -15
7. What is the number of degrees of freedom for a simple bivariate linear
regression with 20 observations?
a. 20
b. 22
c. 18
d. 2
8. One tailed tests are sometimes used to test hypotheses about regression
coefficients. In which of the following circumstances?
a. when the estimated coefficient has the sign predicted by theory
b. when you wish to use a larger significance level than 5%
c. when the sample size is large enough to use the normal approximation

22
Institute of Lifelong Learning, University of Delhi
 Introduction and subject matter of Econometrics

to the t distribution
d. when the etsimated coefficient has the opposite sign to that predicted
by theory
e. when you are testing a hypothesis other than that the parameter
equals zero

ANSWERS:

1. e
2. e
3. d
4. b
5. d
6. b
7. c
8. a

23
Institute of Lifelong Learning, University of Delhi