Question

Suppose the legislature passes a law providing unemployment benefits to people who make less than $20000 a year. I want to see how these benefits affect labor force participation.

a) Why does this setup work well for regression discontinuity?

b) What would the treatment and control groups be if I did a discontinuity regression?

c) In general, what are the advantages of doing a DID or regression discontinuity regression? What are the downsides?

Answer 1

Answer (a): The process of performing regression discontinuity allows is to confidentially determine which factors matter most, which factors can be ignored, and how these factors influence each other.

Using the non-parametric methods in regression discontinuity is very beneficial, and the major benefit is that it provides estimates based on data closer to the cut-off, which is intuitively appealing. It reduces some bias that can result from using data farther away from the cut-off to estimate the discontinuity at the cut-off.

Answer (b): The intution behind the regression discontinuity design is well illustrated using the evaluation of merit-based scholarships. If high performing students are more likely to be awarded the merit scholarship and continue performing well at the same time, comparing the outcomes of the awardees and non recipients would lead to an upwards bias of the estimates. Even if the scholarship does not improve grades at all, awardees would have performed better than non recipients, simply because scholarships were given to students who were performing well.

Comparing the outcome of the treatment group to the counterfactual outcome of the control group will hence deliver the local treatment effect.

Answer (c): * Advantages of doing a regression discontinuity are as follows:

• Well executed regression discontinuity can generate treatment effect estimates similar to estimates from randomized studies.

• When properly implemented and analized, the regression discontinuity yields an unbiased estimates of the local treatment effect. The regression discontinuity design can be almost as good as randomised experiments in measuring a treatment effects.

• Regression discontinuity design, as a quasi experiment, does not require ex ante randomization and circumvents ethical issues of random assignments.

* Disadvantages of doing a regression discontinuity design are as follows:

• The estimated effects are only unbiased if the functional form of the relationship between the treatment and outcome is correctly modelled. The most popular caveats are non linear relationships that are mistaken as a discontinuity.

• Contamination by other treatments. If another treatment occurs at the same cutoff value of the same assignment variable, then the measured discontinuity in the outcome variable may be partially attributed to this other treatment.