Following up on a comment he made at my earlier post on this topic, Declan has some interesting thoughts on the question. I will simply add to his points–with which I almost entirely agree–that I accept the idea of some conditions on referendum majorities.
For instance, in federal systems, concurrent majorities make sense. An example of a concurrent-majority requirement would be saying a constitutional amendment must be approved by a national majority and also majorities in a majority of states/provinces, or something of that sort.
I also accept the idea of turnout requirements. Stipulating a minimum quorum of voters participating in the question for it to pass makes sense to prevent, say, 50% of a 40% turnout from deciding policy or constitutional questions, as may happen in California on November 8. There is a problem with quorum rules, however, in that it gives opponents who can’t get a 50% ‘no’ vote an incentive to simply boycott, to defeat the measure by denying it a quorum. This has happened in several Italian referenda, and in a recent series of referenda in Colombia, for example. (A fix for this would be to stipulate a majority of a given quorum or a super-majority of whoever does turn out.)
But while I accept these qualifications on referendum approval, I really do not see much point in having referenda to guage the “will of the people” and then mandating that it will be valid only if 60% or two thirds share the same “will.” And, as I have said before, the rule by which the Iraqi constitution was just approved is truly bizarre.
(A concluding note, which answers a specific question asked by Declan: A few years ago in California a statewide initiative passed to lower the requirement to pass local school bonds to a 55% majority of those voting instead of the former two-thirds requirement. The supporters feared going all the way and proposing simple majority, but at least they cut the minority veto potential from one third to 45%. There continue to be 2/3 requirements on some statewide measures.)