Econometrics I

Nicolás Corona Juárez, Ph.D

24.08.2020
Econometrics I

Nicolás Corona Juárez, Ph.D

Universidad de las Américas Puebla

nicolas.corona@udlap.mx

Office hours: Please send me an E-mail. Office NE-242

2
General Information

• Lecture Notes and Problem Sets will be sent to your institutional email account.
• The Problem Sets are not graded but you are encouraged to work in groups and
be able to solve them on your own.

• Datasets will be provided by myself.

• Introduction to Stata: I will be introducing you to Stata as we progress in the

course.

3
General Information

• Evaluation of the course:

• -1st Exam: September 18th 2020 (30%)
• - 2nd Exam: October 16th 2020 (25%)
• - 3rd Exam: November 6th 2020 (25%)
• - Final Exam (accumulative): The examination office will notify us about the date
before the semester ends (20%)

4
General Information

• J.M. Wooldridge
Introductory Econometrics- A modern Approach, 6th edition

• Gujarati
Basic Econometrics, 4th, 5th edition

Ocassionally I will distribute some research papers.

5
Econometrics I

• Econometrics makes use of Economic Theory, Statistics and Math to draw

conclusions about the interrelation among several economic, financial and
socioeconomic variables.

• Need of data

• Those data are obtained from several sources: Government, Research Institutions
etc.

6
Econometrics I:

• Econometrics is concerned with the empirical determination of economic laws.

• The word “empirical” indicates that the data used for this determination have
been obtained from observation, which may be either controlled
experimentation designed by the econometrician interested, or “passive”
observation.

7
Two definitions of probability for Econometrics

• Random Variable: An event whose outcome is uncertain.

• If 𝑥 is a random variable it can take on different values.

• We know the possible values that the variable takes on….but…. we do not know
the exact outcome before the event.
Examples:
i. Flipping a coin (outcomes are head or tail).
ii. Election of a political candidate (win or lose).
iii. Exchange rate movement (upward or downward movement).

8
Two definitions of probability for Econometrics

• Discrete Random Variable: it represents an outcome that can only take on a fixed
number of values.
Examples:
I. The number of dots on a die.
II. The number of crimes happening in a place.

• Continuous Random Variable: it represents an outcome that cannot be

technically counted.
Examples:
I. The height of an individual.
II. Exchange rate fluctuation.
9
Two definitions of probability for Econometrics

• Bayesian: probability expresses the degree of belief a person has about an event
by a number between zero and one.

• Classical: the relative number of times that an event will occur as the number of
experiments becomes very large.

𝑟𝑜
lim 𝑃 𝑂 =
𝑁→∞ 𝑁
Examples?
Bayesian relies on previous knowledge about the event.
Classical relies on the relative number of attempts.

10
Econometrics

1. Underlying theory and hypothesis set up

2. From theory to the mathematical model
3. Econometric model specification
4. Getting the data
5. Econometric model estimation
6. Hypothesis test
7. Prediction or Forecast
8. Use of the results in policy advice

11
1. Underlying hypothesis

• Take the case of the Millenium Development Goals

• (More info in: http://www.un.org/es/millenniumgoals/)

• Our expectation is: Aid for Health reduces the child mortality in developing
countries.

12
2. From theory to the mathematical model

• Previously we mentioned a relationship between two variables

𝑌 = Child mortality in developing countries

𝑋 =Aid for health

• We do not know the functional form among these variables but we asume they
follow a linear behavior.

13
2. From theory to the mathematical model.

Dependent variable Independent variable

𝑌 = 𝛽0 − 𝛽1 𝑋

Intercept Slope coefficient

14
2. From theory to the mathematical model.

• We have:

𝑌 = 𝛽0 − 𝛽1 𝑋

With: 0 < 𝛽1 < 1

• Why do we have a negative sign in 𝛽1 ?

15
2. From theory to the mathematical model.

• We could have written:

𝑌 = 𝛽0 + 𝛽1 𝑋

With: 0 < 𝛽1 < 1

• The sign of the 𝛽1 coefficient will depend upon the theory

This model assumes a deterministic relationship

16
3. Econometric model specification.

• Theoretically: More Aid for Health will reduce Infant Mortality in the developing
world
• However, there are several factors which might undermine the effectiveness of
Aid for Health on Infant Mortality reduction. For instance, corruption.

• Can you think about a channel of how corruption might exert an impact on the
effectiveness of Aid for Health?

• This means, we do not have any deterministic relationship in the model.

17
3. Econometric model specification .

• A model which considers stochastic variables will look as:

This model is an
example of a linear
regression This is a stochastic
𝑌 = 𝛽0 − 𝛽1 𝑋 + 𝑢 variable (random
variable)

It is also known as
econometric model
Error term

18
4. Getting the data

• Once the model has been specified, we need data in order to estimate the
coefficients 𝛽0 and 𝛽1 .

• We can get data from government offices, international organizations, research

institutions and the like.

• This will depend on the research question

19
4. Getting the data: examples

20
4. Getting the data: examples

21
4. Getting the data: examples

22
5. Econometric Model Estimation

• After getting the data, we proceed to the estimation of the coefficients 𝛽0 y 𝛽1 .

• The most basic method used for the estimation of these coefficients is called
Ordinary Least Squares (OLS).

• There are several econometric methods and their use depends on the data to be
used and the problems we face in our research.

23
5. Econometric Model Estimation

• Suppose that after you estimate the model, you get:

𝑌෠𝑡 = 9.59 − 0.02𝑋𝑡

• Note that the dependent variable is written now in a different form. The hat ෡
indicates this is an estimated variable.

• Thus, 𝑌𝑡 and 𝑌෠𝑡 are not the same.

24
5. Econometric Model Estimation

• Suppose our data look like:

1) Time series
2) Cross Section
3) Pooled Cross Section
4) Panel Data

25
5. Econometric Model Estimation

• Theory indicates: Aid for Health reduces Infant Mortality.

• Suppose you get data for Paraguay on Infant Mortality and Aid for Health
received by this country.

• Observations are available for the 1970 – 2015 period.

26
5. Econometric Model Estimation

If we plot the data:

𝑌

27
5. Econometric Model Estimation

𝑌
What we observe

𝒖
What we estimate

28
6. Hypothesis test

• After the estimation we have to gauge how far is our estimate from reality.
• Important questions:

• How likely is it that we accept something that is indeed false?

• How likely is is that we reject something that is indeed true?

• Need statistical inference to give answer to this questions

29
7. Forecast

• If the estimated model does not reject the null hypothesis, then it can be used to
predict future values of the dependent variable given the expected value of the
independent variable.
• Suppose you want to predict the average infant mortality rate for the year 2010.

• Aid for health was 10 dollars in 2010.

30
7. Forecast

It follows that:

𝑌෠2010 = 9.59 − 0.02 10

𝑌෠2010 = 9.3
• With 10 dollars, the infant mortality rate for the year 2010 is 9 children for each
1000 births.
• Note: Infant mortality rate is defined as the number of dead infants for each 1000
infants born alive.

31
7. Forecast

• The model can over- or underestimate the infant mortality rate.

The model shows: 𝑌෠2010 = 9.3

• With 10 dollars, the infant mortality rate in 2010 is 9 infants per each 1000 births.

• If the real infant mortality rate for the year 2010 would have been 4 children per
each 1000 births; then the model shows an overestimated infant mortality rate.

32
8. Policy

• How much Aid for Health is needed if we want to reach an infant mortality equal
to 500 keeping 𝛽1 = 0.02 ?

• 𝛽1 = 0.02, this is the impact that Aid for Health has on the Infant Mortality Rate.

500 = 9.59 − 0.02 𝑋

Solve for 𝑋

33
References

• J.M. Wooldridge
Introductory Econometrics- A modern Approach, 6th edition
Chapter 1

• Gujarati
Basic Econometrics, 4th, 5th edition
Chapter 1

34