L3: Empirical Tools of Public Finance Flashcards
correlation
two economic variables are correlated if they move togethers
causal
two economic variables are causally related if the movement of one causes the movement of the other
identification problem
if two series are correlated, how do you identify whether one causes the other?
explanations when we see a correlation between A and B
A causes B
B causes A
some third factor causes both
randomised trials
ideal type of experiment designed to test causality, whereby a group of individuals is randomly divided into a treatment group, which receives the treatment of interest, and a control group, which does not
solves the identification problem
- rules out reverse causation
- treatment and control differ only by treatment (no third factor causes)
- any difference is due to treatment
problems with randomised trials
external validity
- results are only valid for the sample and not necessarily the population as a whole
attrition
- individuals may leave the experiment over time, which if not random leads to biased estimates
observational data
data generated by individual behaviour observed in the real world, not in the context of deliberately designed experiments
three types of observational analyses
time series analysis: analysis of co-movements of two or more series over time
cross-sectional analysis: analysis of the relationship between two or more variables across a population of individuals at a given time
panel data analysis: combined approach looking at a population of individuals over time
regression line
line that measures the best linear approximation between two variables
control variables
additional variables that are included in cross-sectional regression models to account for differences between treatment and control groups that can lead to bias
natural experiment / quasi-experiment
changes in the economic environment that create nearly identical treatment and control groups for studying the effect of that change
difference-in-differences estimator
difference between change in outcomes for the treatment group that experiences an intervention and the control group that doesn’t
problems with difference-in-differences estimation
requires common trends assumption - rates of change would have been the same without the treatment
regression discontinuity estimator
estimate the effect of a policy based on a discontinuous change in the policy based on some exogenous factor
structural vs reduced form estimates
structural estimates: estimates of features that drive individual decisions
reduced form estimate: measures of total impact of an independent variable on a dependent variable, without decomposing the source of that behaviour response in terms of underlying utility functions
