Applied Microeconometric Techniques
Utrecht University – ECB3AMT
Written by Lisanne Louwerse
Summary
,Table of content
WEEK 1 – CAUSALITY ........................................................................................................................................ 3
INTRODUCTION ....................................................................................................................................................... 3
CORRELATION VS. CAUSALITY ..................................................................................................................................... 3
THE RUBIN CAUSAL MODEL ....................................................................................................................................... 6
WEEK 2 – REGRESSION AND BIAS ..................................................................................................................... 8
ENDOGENEITY IN THE REGRESSION MODEL .................................................................................................................... 8
WEEK 3 – RANDOMIZED CONTROLLED TRIAL (RCT) ........................................................................................ 10
RCTS .................................................................................................................................................................. 10
TREATMENT EFFECTS .............................................................................................................................................. 10
INTERNAL AND EXTERNAL VALIDITY ............................................................................................................................ 12
LIMITATIONS AND ALTERNATIVES .............................................................................................................................. 13
WEEK 4 + 5 – INSTRUMENTAL VARIABLES (IV) ................................................................................................ 15
THE IV METHOD.................................................................................................................................................... 15
INTERNAL VALIDITY ................................................................................................................................................ 16
EXTERNAL VALIDITY ................................................................................................................................................ 17
2-STAGE LEAST SQUARES (2SLS) ............................................................................................................................. 18
WEEK 6 – REGRESSION DISCONTINUITY DESIGN (RDD) .................................................................................. 20
MAIN IDEA OF RDD ............................................................................................................................................... 20
INTERNAL VALIDITY ................................................................................................................................................ 21
EXTERNAL VALIDITY ................................................................................................................................................ 22
FUZZY RDD.......................................................................................................................................................... 23
WEEK 7 + 8 – DIFFERENCE-IN-DIFFERENCES (DID) ........................................................................................... 24
MAIN IDEA OF DID ................................................................................................................................................ 24
COMMON TREND ASSUMPTION ................................................................................................................................ 25
DD REGRESSION.................................................................................................................................................... 26
POOLED OLS ........................................................................................................................................................ 27
FIXED EFFECTS ESTIMATOR ...................................................................................................................................... 28
WEIGHTING ......................................................................................................................................................... 29
INTERNAL VALIDITY ................................................................................................................................................ 30
EXTERNAL VALIDITY ................................................................................................................................................ 32
KEY TAKEAWAYS ............................................................................................................................................ 34
WEEK 1 - CAUSALITY ............................................................................................................................................. 34
WEEK 2 – REGRESSION AND BIAS ............................................................................................................................. 35
WEEK 3 - RCT ...................................................................................................................................................... 36
WEEK 4 + 5 - IV ................................................................................................................................................... 38
WEEK 6 - RDD ..................................................................................................................................................... 40
WEEK 7 + 8 - DID ................................................................................................................................................. 41
USEFUL LINKS ................................................................................................................................................. 45
2
,Week 1 – Causality
Key Words
▪ Randomized experiments, natural experiments and quasi-experiments.
▪ Correlation vs. causality
▪ Reverse causality, simultaneity, omitted variable bias and selection bias
▪ Policy evaluation
▪ Fundamental problem of causal inference (counterfactual situation)
▪ Rubin causal model
▪ SUTVA
Introduction
The gold standard to research a cause-effect-relationship is to use randomized experiments
like lab or field experiments. But this is often infeasible in economics and business.
This course: how can we use natural experiments in order to study relationships?
Natural experiments: real-life situations that economists study and analyse to determine cause
and effect relationships.
Difference between natural experiment and quasi-experiment:
▪ In a natural experiment the assignment occurs ‘naturally’, without the researcher's
intervention.
▪ In a quasi-experiment the criterion for assignment is selected by the researcher.
Correlation vs. causality
▪ Correlation does not imply causality. Just because two things correlate, it does not mean
that one of them is causing the other.
▪ Lack of correlation does not imply lack of causality. Just because two things do not
correlate, it does not mean that there is no relationship between them.
Why does correlation not imply causation?
▪ Reverse causality: X and Y are associated, but not in the way you would expect. Instead
of X causing a change in Y, it is really the other way around: Y is causing changes in X.
causality
reverse causality
▪ Simultaneity: the explanatory variable is jointly determined with the dependent variable.
In other words, X causes Y but Y also causes X.
3
, ▪ Omitted variable bias: occurs when a variable O affects both X and Y but is not
(adequately) taken into account.
▪ Selection bias: occurs when the subjects who select or who are selected into treatment
differ from the subjects who don‘t.
Why should we care about causality?
Causal effect is very important in policy evaluation.
Policy evaluation: a systematic assessment of the change in an outcome variable (y) that can
be ascribed to a specific policy measure/intervention (x).
Why does this matter?
▪ Actual effect of measures/interventions → evidence-based policy
→ We want to know whether the policies (measures/interventions) we implement
are actually effective.
▪ Allocation of (financial) resources.
→ As a firm you have limited resources so you want to know which measures give
you the most benefits.
▪ Accountability of decision makers
→ Show what the actual effect of a measure is, instead of having to rely on the
opinions of decision makers (like politicians).
How do you evaluate a policy?
Key element of policy evaluation: construction of an adequate counterfactual situation.
Counterfactual situation: What would have happened in the absence of the intervention?
What would have happened to the treated unit had it not been treated?
Key aim of policy evaluation: identifying the causal effect of an intervention.
The counterfactual situation is unobservable. You can identify the causal effect by comparing
the actual situation to the hypothetical counterfactual situation.
Steps of policy evaluation:
Step 1. Defining the unit of observation.
→ At which level do you measure the outcome variable?
→ E.g. individuals, regions, firms, households, schools, etc.
4
Utrecht University – ECB3AMT
Written by Lisanne Louwerse
Summary
,Table of content
WEEK 1 – CAUSALITY ........................................................................................................................................ 3
INTRODUCTION ....................................................................................................................................................... 3
CORRELATION VS. CAUSALITY ..................................................................................................................................... 3
THE RUBIN CAUSAL MODEL ....................................................................................................................................... 6
WEEK 2 – REGRESSION AND BIAS ..................................................................................................................... 8
ENDOGENEITY IN THE REGRESSION MODEL .................................................................................................................... 8
WEEK 3 – RANDOMIZED CONTROLLED TRIAL (RCT) ........................................................................................ 10
RCTS .................................................................................................................................................................. 10
TREATMENT EFFECTS .............................................................................................................................................. 10
INTERNAL AND EXTERNAL VALIDITY ............................................................................................................................ 12
LIMITATIONS AND ALTERNATIVES .............................................................................................................................. 13
WEEK 4 + 5 – INSTRUMENTAL VARIABLES (IV) ................................................................................................ 15
THE IV METHOD.................................................................................................................................................... 15
INTERNAL VALIDITY ................................................................................................................................................ 16
EXTERNAL VALIDITY ................................................................................................................................................ 17
2-STAGE LEAST SQUARES (2SLS) ............................................................................................................................. 18
WEEK 6 – REGRESSION DISCONTINUITY DESIGN (RDD) .................................................................................. 20
MAIN IDEA OF RDD ............................................................................................................................................... 20
INTERNAL VALIDITY ................................................................................................................................................ 21
EXTERNAL VALIDITY ................................................................................................................................................ 22
FUZZY RDD.......................................................................................................................................................... 23
WEEK 7 + 8 – DIFFERENCE-IN-DIFFERENCES (DID) ........................................................................................... 24
MAIN IDEA OF DID ................................................................................................................................................ 24
COMMON TREND ASSUMPTION ................................................................................................................................ 25
DD REGRESSION.................................................................................................................................................... 26
POOLED OLS ........................................................................................................................................................ 27
FIXED EFFECTS ESTIMATOR ...................................................................................................................................... 28
WEIGHTING ......................................................................................................................................................... 29
INTERNAL VALIDITY ................................................................................................................................................ 30
EXTERNAL VALIDITY ................................................................................................................................................ 32
KEY TAKEAWAYS ............................................................................................................................................ 34
WEEK 1 - CAUSALITY ............................................................................................................................................. 34
WEEK 2 – REGRESSION AND BIAS ............................................................................................................................. 35
WEEK 3 - RCT ...................................................................................................................................................... 36
WEEK 4 + 5 - IV ................................................................................................................................................... 38
WEEK 6 - RDD ..................................................................................................................................................... 40
WEEK 7 + 8 - DID ................................................................................................................................................. 41
USEFUL LINKS ................................................................................................................................................. 45
2
,Week 1 – Causality
Key Words
▪ Randomized experiments, natural experiments and quasi-experiments.
▪ Correlation vs. causality
▪ Reverse causality, simultaneity, omitted variable bias and selection bias
▪ Policy evaluation
▪ Fundamental problem of causal inference (counterfactual situation)
▪ Rubin causal model
▪ SUTVA
Introduction
The gold standard to research a cause-effect-relationship is to use randomized experiments
like lab or field experiments. But this is often infeasible in economics and business.
This course: how can we use natural experiments in order to study relationships?
Natural experiments: real-life situations that economists study and analyse to determine cause
and effect relationships.
Difference between natural experiment and quasi-experiment:
▪ In a natural experiment the assignment occurs ‘naturally’, without the researcher's
intervention.
▪ In a quasi-experiment the criterion for assignment is selected by the researcher.
Correlation vs. causality
▪ Correlation does not imply causality. Just because two things correlate, it does not mean
that one of them is causing the other.
▪ Lack of correlation does not imply lack of causality. Just because two things do not
correlate, it does not mean that there is no relationship between them.
Why does correlation not imply causation?
▪ Reverse causality: X and Y are associated, but not in the way you would expect. Instead
of X causing a change in Y, it is really the other way around: Y is causing changes in X.
causality
reverse causality
▪ Simultaneity: the explanatory variable is jointly determined with the dependent variable.
In other words, X causes Y but Y also causes X.
3
, ▪ Omitted variable bias: occurs when a variable O affects both X and Y but is not
(adequately) taken into account.
▪ Selection bias: occurs when the subjects who select or who are selected into treatment
differ from the subjects who don‘t.
Why should we care about causality?
Causal effect is very important in policy evaluation.
Policy evaluation: a systematic assessment of the change in an outcome variable (y) that can
be ascribed to a specific policy measure/intervention (x).
Why does this matter?
▪ Actual effect of measures/interventions → evidence-based policy
→ We want to know whether the policies (measures/interventions) we implement
are actually effective.
▪ Allocation of (financial) resources.
→ As a firm you have limited resources so you want to know which measures give
you the most benefits.
▪ Accountability of decision makers
→ Show what the actual effect of a measure is, instead of having to rely on the
opinions of decision makers (like politicians).
How do you evaluate a policy?
Key element of policy evaluation: construction of an adequate counterfactual situation.
Counterfactual situation: What would have happened in the absence of the intervention?
What would have happened to the treated unit had it not been treated?
Key aim of policy evaluation: identifying the causal effect of an intervention.
The counterfactual situation is unobservable. You can identify the causal effect by comparing
the actual situation to the hypothetical counterfactual situation.
Steps of policy evaluation:
Step 1. Defining the unit of observation.
→ At which level do you measure the outcome variable?
→ E.g. individuals, regions, firms, households, schools, etc.
4
