If I had infinite (ok, just twice more) time I would like to investigate the following:
What is the effect of limiting the number of foreign players in soccer leagues? A cross-country study.
Ideally, this should be a panel with TSCS variation on the legal limits variable.
DepVar: domestic players competitiveness (measured by performance of the country-i team on FIFA tournaments through time)
IndepVars: Number of foreign players allowed or average foreigners enrolled, initial competitiveness, GDP per capita, number of teams in top tier league or number of large stadiums per country… etc.
Logic: if foreign players crowd out domestic talent, local teams should underperform when more foreigners are allowed. But if foreign players raise competitive standards and complement local talent (think of the self-selection and screening effects on the pool of domestic players), then national teams may become more competitive in the presence of foreigners.
Comment: I am not enough of a sports fan to care a lot about this–but whatever the finding, it would give a good economics lesson to all those sports-pundits out there.
Narco-campaigning? Drug money and local democracy in México (or Colombia for that matter)
This also has to be a panel where some localities become “treated” by drug dealers.
DepVar: Turnout or closeness in local elections.
IndepVars: Some sort of index of drug dealing activity at the local level (maybe proxied by drug-related crime?), typical political controls within state and municipality, two-way fixed effects, etc.
Logic: Ok, we know/suspect that drugdealers may want to buyout local authorities. They can do this at the electoral or post-electoral stage. If they give out money to candidates, and we assume some common campaigning technology, you could expect (i) that the extra money leads to higher turnout (compared to state and municipal averages), or (ii) maybe more lopsided elections returns. If the effect is significant it would be indirect evidence of “narco-campaigning”. If the effect is not significant, we would still face many possibilities: either narco campaigning is not affecting election outcomes because drug dealers don’t care much about the electoral stage (vis a vis post-electoral maneuvering)–or maybe drug dealers giving is indistinguishable from other donors elsewhere, or maybe candidates optimize their spending to just about what they need to win, and pocket the rest, etc. (yes, there are many explanations for any nonsignificant outcome!).
Comment: Getting a good proxy for drug-activity may be hard (but who knows, maybe the AFI got some data at hand). But the real downside is worse: if the narc variable turns out to be positive and significant, your life is in danger.