I'm planning on using a regression discontinuity design (RDD) to estimate the impact of a sector-specific law on the firms in that sector. I need to create a control group and I have read somewhere in the literature that you should use "a similar sector". When it comes down the selecting the similar sector, I don't know where to start. Does anyone have any suggestion (readings and such)? The sectors which would be (separately) analysed are: telecommunication, education, construction, banking, finance.
Ideally, but not always feasible, the first option would be to select a similar region where the law is not in place and compare them (e.g. different country / State / city)
If it is not possible to compare with another region, then chose a similar industry. I would try to justify it well though, and comparability would be harder to prove. For example, if you want to study subsidised education, you might want to compare it with subsidized healthcare - the argument could be they're both services, and provide basic needs at some point, with standardised outcomes expected.