The time-inconsistency of a referendum on independence

I have done work (with Paul Levine, Jon Stern, Jordi Gual and M.A. Montoya, in different papers) on time-inconsistency in regulation as compared for example with time-inconsistency in monetary policy. A decision is time-inconsistent when the decision maker would have taken another decision if she had been able to take and maintain that decision ex ante, before other players had made long lasting decisions. The following situations would be described as time inconsistent behaviour by a regulator:  a regulatory decision is made that could not have been predicted, is the opposite of what had previously been decided (and communicated), and sunk investments had taken place based on the previous decision. For example, a national government expropriating a foreign-owned electricity firm without sufficient compensation would be time-inconsistent.  Or a regulator fixing motorway tolls lower than those that had been committed to at the time of contracting with third party construction/operators, would be time-inconsistent. I would argue that the idea of time-inconsistency can be applied much more generally, and that it is a risk inherent in majoritarian decision-making. Unanimity or super-majorities are required in most democracies to avoid reneging on important commitments. For example, as I argue here, a secession referendum of a relatively rich region divided across ethnolingüistic lines is not necessarily a good idea, despite its popularity. It creates the risks of reneging on the investments by potential losers on human capital, and other assets whose value depend on the stability of institutions. Given a flat-out choice between “yes” and “no” to independence, Catalans for example would be forced to choose between extremes. This would unfairly eradicate the significant middle ground that exists between the two black and white options. Under a federalized system, Spain and Catalonia could continue to enjoy the benefits of union, and Catalans could operate with the enhanced autonomy they desire, without breaching any explicit or implicit contract. By forcing voters to choose between two extreme views, an independence referendum favors extremist thinkers and movements, rather than those seeking to compromise. This will empower those who are intolerant of the opposing side, raising tensions within Catalonia, further creating uncertainty, as well as those between Catalonia and the Spanish state. In addition, by being forced to choose sides, anti-secessionist leftists are pushed into an uncomfortable alliance with rightists who partially share their views on the issue. A referendum on independence is a democratic tool. A deliberative process that ends in a broad compromise, followed by a referendum, is as democratic as that, and scores much higher in terms of commitment and stability. You might be tempted to think that I am like those that when they have a hammer, they think everything is a nail, but I learn these arguments from the readings of great economists and social scientists. Weingast argues here that "unfettered democracy fails to provide the conditions for democratic stability." And Silvestre vindicates the virtues of unanimity or at least qualified majorities as a principle of social justice as defended long ago by Wicksell.