Scribd
Upload
57 views

APEC 5031 APEC 5031: Elton Mykerezi Applied Economics Applied Economics

• What is economics? • Empirical analysis • Correlation vs. causality • Economic data – Cross-sectional data – Time series data – Pooled cross-sectional data – Panel data

Uploaded by

resperado
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
57 views

APEC 5031 APEC 5031: Elton Mykerezi Applied Economics Applied Economics

• What is economics? • Empirical analysis • Correlation vs. causality • Economic data – Cross-sectional data – Time series data – Pooled cross-sectional data – Panel data

Uploaded by

resperado
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/ 14

APEC 5031 APEC 5031

Elton Mykerezi
Applied Economics Applied Economics
Outline Outline
What is economics? What is economics?
Empirical analysis
C l ti lit Correlation vs. causality
Economic data
Cross-sectional data
Time series data
Pooled cross-sectional data
Panel data
What Is Economics What Is Economics
From Webster: From Webster:
A social science concerned with description and
analysis of production, distribution and consumption
of goods and services.
Econometrics: The application of statistical methods
to the study of economic data and problems to the study of economic data and problems.
I prefer:
Economics: Study of human behavior and interaction Economics: Study of human behavior and interaction
under scarcity
Econometrics: Statistics for non experimental data
Empirical Analysis Empirical Analysis
An empirical analysis uses data to test a An empirical analysis uses data to test a
theory.
Example Example
Hypothesis: Language immersion instruction
leads to better academic results leads to better academic results.
Demand for food items is downward sloping
Th h d t il Theory here doesnt necessarily mean a
formal model
Example 1: Language Immersion
Economic Model: Economic Model:
Set up an economic model the equations Set up an economic model the equations
that describe relationships.
Example Example
y = f (immersion, school funding, family SES)
Specify an econometric model based on Specify an econometric model based on
this economic model.
The econometric model specifies functional e eco o et c ode spec es u ct o a
form, defines variables, accounts for
unobservable factors.
Econometric model for language immersion
i t impact
AcadPerf
i
=
0
+
1
Imm
i
+
2
CperT
i
+
3
Disadv
i
AcadPerf
i

0

1
Imm
i

2
CperT
i

3
Disadv
i
+ u
i
i = school indicator
AcadPerf
i
= academic performance, an average of
reading and math proficiency scores for the school
I 1 if l i i h l 0 th i Imm
i
= 1 if language immersion school, 0 otherwise
CperT
i
= children per teacher ratio in school
Disadv = % economically disadvantaged Disadv
i
= % economically disadvantaged
ui = unobservable factors
Hypothesis:
1
> 0 Hypothesis:
1
> 0
Correlation V.S. Causalityy
Two factors may be correlated: y
y =
0
+
1
x + u
But this is not the same as saying one factor But this is not the same as saying one factor
causes the other.
We need to hold all other relevant factors We need to hold all other relevant factors
fixed to be able to say that x causes y.
Thi i th i i l f t i ib This is the principal of ceteris paribus
all other things being equal.
Examples: p
Intuitive
College Educated Individuals make more
moneyy
Not Intuitive
Cities with more police have higher violent Cities with more police have higher violent
crime rates
Ambiguous: g
People on Food Stamps have higher BMI
than income-eligible non participants g p p
Economic Data: Cross-sectional Economic Data: Cross sectional
Cross-sectional data are a sample of units
( fi h h ld ) t k t ti (consumers,firms, households) taken at one time.
Random sampling is often assumed to: p g
aid in treatment of unobservable factors.
provide correct interpretation of findings over appropriate
populations
Random sampling is typically not an appropriate
assumption for observational data.
Example
A phone survey might undercount workaholics and
carpooling parents, while over counting retirees.
Economic Data: Time Series Economic Data: Time Series
Time series data are an observation on a set of
variables over time.
Example: GDP, inflation and unemployment rates
sampled from 1943 1980 (Same agent; US) sampled from 1943-1980. (Same agent; US)
Chronological order matters
When data frequency is less than annual When data frequency is less than annual,
seasonality may matter.
Example: Home sale dip dramatically during the
holiday season.
Electronic Sales increase during winter holiday
season
Economic Data: Pooled cross-
i l d sectional data
Pooled cross-sectional data are cross- Pooled cross sectional data are cross
sectional samples taken at two or more points in
time.
Units sampled vary between points in time.
Examples p
High school graduation rates sampled before and
after the "No Child Left Behind" legislation.
Standard of Living Measurement Survey
Consumer Expenditure Survey (Rolling Panel)
Economic Data: Panel
Panel data are a time series for each member of
a cross-sectional sample a cross-sectional sample.
Example
The Medical Expenditure Panel Survey (MEPS) The Medical Expenditure Panel Survey (MEPS)
is a periodic collection of data on households,
their insurance status, their medical spending, t e su a ce status, t e ed ca spe d g,
and other relevant factors.
Distinguished from pooled cross-sectional data g p
because it is the same units that are sampled
over time.
Other surveys: PSID, NLSY, NELS, ECLS, SIPP
Economic Data: Panel II Economic Data: Panel II
Advantages: g
Make some analysis possible/cleaner that cross-
sections cant.
Does childhood stress predict depression in Does childhood stress predict depression in
teenagers
Help establish causality
C S Change in FSP status correlated to change in BMI
Change in student grades correlated with
implementation of a policy
Concern:
Attrition could be a problem
Attrition may be non random Attrition may be non-random

You might also like