Rick Hills has written a post offering conservatives advice on how to talk civilly about pension reform. I'm a libertarian, not a conservative, and some of my detractors would probably just actually sa that I'm a crank. But some of this probably applies to me anyway, so I'm willing to take a bite.
Still, a concern about the proposed solution of referenda:there seems to be a classic public-choice-ish problem here where the diffuse interests of a large group (the taxpayers) diverge considerably from the interests of a small group (public employees who are self-interested in receiving generous pensions.) The public employees have every incentive to form powerful interest groups that would campaign aggressively for generous pensions. The diffuse mass of taxpayers -- the group that should care the most about promoting fiscal responsibility for self-interested reasons -- have far less incentive to care. Given these incentives, I imagine that many non-public-employees would sit these elections out altogether, and/or that others would follow the public employee interest groups' lead. The result might well be liabilities as large or perhaps larger than we're seeing now.
Voter ignorance is a real concern here; as Ilya has written about before, many voters are unaware of even much simpler facts such as which party controls the House of Representatives or the names of their own senators. Fifty-seven percent of voters didn't know who Newt Gingrich was immediately following the controversial 1994 elections. As Ilya also notes, better informed voters often choose to follow political news as a form of infotainment. Newt Gingrich is a colorful figure who's fascinating to read about; pension reform can seem dry even to professional policy wonks. Getting the electorate to participate in these referenda and to make well-informed decisions will be difficult.
I freely admit that the current systems for setting public employee pay have their own perverse incentives and problems. I'm willing to entertain the possibility that the system Hills proposes would be marginally or incrementally better than what we have now.