Ch. 2 Tools of Positive Economics
Theoretical Tools of Public
Finance
•theoretical tools The set of
•tools designed to understand
•the mechanics behind economic
decision making.
•empirical tools The set of
tools designed to analyze data and
answer questions raised by
theoretical analysis.
The Role of Theory
• Economic models
– virtue of simplicity
– judging a model
– limitations of models
• Empirical analysis
•There are many examples
where causation and
correlation can get confused.
•In statistics, this is called
the identification problem:
given that two series are
correlated, how do you
identify whether one series
is causing another?
Causation vs. Correlation
• Statistical analysis
– Correlation
– Control group
– Treatment group
• Conditions required for government action X to
cause societal effect Y
– X must precede Y
– X and Y must be correlated
– Other explanations for any observed correlation must
be eliminated
Experimental Studies
• Biased estimates
• Counterfactual
• Experimental (or randomized) study
Conducting an Experimental
Study
• Random assignment to control and
treatment groups
Pitfalls of Experimental Studies
•
•
•
•
•
Ethical issues
Technical problems
Response bias
Impact of limited duration of experiment
Generalization of results to other
populations, settings, and related
treatments
• Black box aspect of experiments
Observational Studies
• Observational study – empirical study relying on
observed data not obtained from experimental
study
• Sources of
observational data
(American Wind Energy Association, 2007)
– Surveys
– Administrative records
– Governmental data
• Econometrics
– Regression analysis
Conducting an Observational
Study
• L = α0 + α1wn + α2X1 + … + αnXn + ε
–
–
–
–
Dependent variable
Independent variables
Parameters
Stochastic error term
L
• Regression analysis
– Regression line
– Standard error
Intercept
is α0
Slope
is α1
α0
wn
Types of Data
• Cross-sectional data
• Time-series data
• Panel data
3.3
•Estimating Causation with Data We
Actually Get: Observational Data
•Time Series Analysis
3.3
•Estimating Causation with Data We
Actually Get: Observational Data
•Time Series Analysis
•When Is Time Series Analysis Useful?
3.3
•Estimating Causation with Data We
Actually Get: Observational Data
•Cross-Sectional Regression Analysis
•Example with Real-World Data
Pitfalls of Observational Studies
• Data collected in non-experimental setting
• Specification issues
Quasi-Experimental Studies
• Quasi-experimental study (= natural
experiment) – observational study relying
on circumstances outside researcher’s
control to mimic random assignment
3.3
•Estimating Causation with Data We
Actually Get: Observational Data
•Quasi-Experiments
Conducting a Quasi-Experimental
Study
• Difference-in-difference quasi-experiments
• Instrumental Variables quasi-experiments
• Regression-Discontinuity QuasiExperiments
Pitfalls of Quasi-Experimental
Studies
• Assignment to control and treatment
groups may not be random
• Not applicable to all research questions
• Generalization of results to other settings
and treatments