Scribd
Upload
4 views32 pages

Lecture 1 2025

The document outlines the curriculum for an Econometrics I course, including required textbooks, exam structure, and the importance of the term paper. It emphasizes the need for precise research questions and understanding population samples in econometric analysis, particularly in the context of education and economic returns. Additionally, it discusses correlation and regression analysis, highlighting the complexities of interpreting relationships between variables.

Uploaded by

Flower Fry
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
4 views32 pages

Lecture 1 2025

The document outlines the curriculum for an Econometrics I course, including required textbooks, exam structure, and the importance of the term paper. It emphasizes the need for precise research questions and understanding population samples in econometric analysis, particularly in the context of education and economic returns. Additionally, it discusses correlation and regression analysis, highlighting the complexities of interpreting relationships between variables.

Uploaded by

Flower Fry
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 32

Econometrics I:

Text Books
• William Greene “Econometric Analysis”
• Jeffrey Wooldridge
• “Introductory Econometrics: A Modern Approach”
•“Econometrics Analysis of Cross Section and Panel
Data” (Selected Sections)
•Angrist and Pischke
•“Mostly Harmless Econometrics”
•Cameron and Trivedi
•“Microeconometrics using STATA”
Final Exam: Two Components: Written Exam +
Term Paper (60%), Mid Term (20%); Every two
Weeks Class Tests-best 5- (20%)
Term Paper
• CRUX OF THIS COURSE
• In addition to the project course
– FOR JOBS
– FOR APPLICATION ABROAD
– TO LEARN!
ASSIGNMENTS

• Computer assignments
– I expect student to learn STATA (since I don’t work
with R).
Regressions

THE BOTTOM LINE UPFRONT!


Population and Sample
• Are there economic returns from studying
science in high school among high school
educated males in urban India?

• Are there economic returns to males from


studying science in high school in urban India?
Population and Sample
• Are there economic returns from studying
science in high school among high school
educated males in urban India?
– High School educated Males in Urban India (past
and present)
• Are there economic returns to males from
studying science in high school in urban India?
– All Males
Population and Sample
• In the latter question we are asking what would
happen if all males in the population had
studied science in high school

• All Males = those who never studied in high school +


those who did
Population and Sample
• In the latter question we are asking what would
happen if all males in the population had
studied science in high school

• All Males = those who never studied in high school +


those who did

Here one would have to separate out


the impact of high school and impact
of science in high school
Population and Sample
• The two questions are different as they are
statements about different populations

– To understand who to sample, you have to define a


population
– To define a population, you have to define your
question
Example

• What are the average labour market returns to


being literate for women in India?

• Ans?
Example

• What are the average labour market returns to


being literate for women in India?

• Average wages of Literate Women


– Average wages of Illiterate Women
Research Question
• What is our research question?
– ELEPHANT IN THE ROOM:
• Wages are observed when women work!

– So we are comparing conditional on working


Research Question
• What is our research question?

– Depends on what is the ”experiment” in our head

• Typically we may be interested in understanding what would be the


labour market “impact” of women having more education
– So we are asking how much more would Basanti earn if she “were” literate
Research Question
• What is our research question?

– Depends on what is the ”experiment” in our head

• Typically we may be interested in understanding what would be the


labour market “impact” of women having more education
– So we are asking how much more would Basanti earn if she “were” literate
» Problem is Basanti may just not work if she were more literate.
» Are those who are working a good proxy for Basanti?
Research Question
• One can ask whether educated working women
earn more than uneducated working women?
– Population: working women

• But where is Basanti then?


Research Question
• A better question would be what is the impact
of education on wages that women can
command
– Population : All Women (say above 25)
Research Question
• Suppose we decide that all women are the
population
– Sample from this population

• Suppose this is rural India


– Most Univ. educated women don’t work. Illiterate
women do (and say earn 500 INR per day)
– Is the return to University then negative?
Research Question
• We need more clarity:
– What we want to typically estimate is how much
incremental income would a illiterate woman get if she were
university educated
– Here it is not visible to us for most univ. educated
women
– So the income from her education is not zero. It is
just that we don’t observe it since she does not work
Research Question
• We need more clarity:

– So the income from her education is not zero. It is


just that we don’t observe it since she does not work (it could,
for example, be very high but her reservation wage to work
could be even higher)
Bottom Line
• Define a precise “economically-meaningful”
research question
• Define the population
– The population model is a statement about that
population and we seek to estimate parameters of
that model
• Draw a sample from that population
– Deal with issues that comes from that sample
Correlation
• Relation between two variables.
• Rough Logic: How much do the variables co-
vary
• Equivalent: How does one variable change when
the other variable changes.
• Example: Household income and Average
Health of the Household.
– Lets say we find a positive correlation!
Interpretations

Common Interpretations:
1. Households that have higher income also have
higher average health
2. Households that have higher health also have
higher income
3. Better Income leads to better health
4. Better health leads to better income.
Problems
1. Households that have higher income also have
higher average health
2. Households that have higher health also have
higher income

• THESE ARE TRUE STATEMENTS BUT…


Problems

3. Better Income leads to better health


4. Better health leads to better income.

THESE STATEMENTS IMPLY


1. Direction of Effect: Health to Income OR
Income to Health
One Solution: Avoid Contemporaneous
Variables!
Problems

2. If I agree with, say “Better Income Leads to


Better Health”

Logical Conclusion: If I transfer some money to


households, I should see better health!

YOU MAY BE COMPLETELY WRONG!


Problems (Contd)
• For example, Higher Income households may
also be households with better educated adult
members and it may be education that drives
health and not income!

Logical Question: Does Income have an effect on


Health after you take into account education?
Partial Correlation Coefficient
• Is a measure that takes into account the effect
of the “other factor(s)”

• It measures how much does income and health


co-vary after you subtract out the effect of
education from each of them.

• “Partial Out” the effect of education.


Education
Health H-Ed

Income Education Inc- Ed

LOOK AT HOW

AND Inc- Ed COVARY!


H-Ed
Regression
• Is an equivalent way of answering the same
question

• It asks what is the effect of a change in one unit


of income on health, education kept constant

• “Education kept constant” IS ACTUALLY “the


effect of education partial-ed out!”
Lab Experiment
• Scientific Experiment: If households were
guinea pigs,
– Take two households exactly the same in everything
– Let us give one unit income higher in one household
(and not to the other)
– Measure Impact
(sounds familiar???)

You might also like