I want to test the causality of part-time workers on unemployment rate.

  1. Should i use part-time workers as a share of total employment or total labour force?

  2. If I run a regression of part-time workers on unemployment rate, there could be reverse causality. So I would have to use 2 stage least square. What are the possible instruments I could use?

  • 1
    $\begingroup$ No regression will tell you about causality. It's the wrong tool for the job. $\endgroup$
    – 410 gone
    Jun 23, 2015 at 14:01

1 Answer 1


This is mostly related to 2):

Not only is there reversed causality, but also potential for omitted variable bias. Identifying causality instead of correlation is something that no econometric method can do for you, you will have to work for that yourself.

For example, you can

  • Find some proxy that is going to move part-time employment, but not any other variable that could be somehow linked to the unemployment rate
  • Find exogenous (random?) variation in part-time employment. The most common tool here are (quasi-) natural experiments.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.