n

Imagine you want to find the causal effect of participating in a job search program on employment. Participation in the program is open for everyone who is unemployed and voluntary. a) Using potential outcomes notation, define the causal effect of participating in the program. Explain your notation with words and explain why it is unlikely that you will be able to find the causal effect. (5p) b) Imagine that you could create an experiment where you randomise unemployed people into either participating in the program or not participating in the program. Using potential outcomes notation, explain why such as experiment would solve the issues you had in a). Explain your notation with words as well. What treatment effect would you be able to estimate? (5p) c) Assume now that the job search program is not open for everyone and you cannot randomize. Instead, you find out that everyone who got unemployed after April 1st 2020 gets forced to participate in the program and everyone who got unemployed before April 1st 2020 does not get to participate. This means you have a sharp threshold that determines treatment, and you could estimate a model using one of the quasi-experimental methods we have discussed. Which one? Set up the equation you would like to estimate. Define each thing you include in the equation. Which parameter gives you the causal effect of interest? (5p)