I can't find that paper using this site's search feature, and I'd be very surprised if it didn't have a questionable definition of "reliable" or considered only a limited set of possible reforms.

There may be some clever game-theoretic argument that could be made about spherical voters in a vacuum, but I think that "political science" results are likely to depend too much on data with too small a sample size (for example, the set of countries that use Approval Voting for national elections).