The other day we had a discussion of whether referendums can defuse political polarization. The context was the recent referendum in Kansas which allowed the majority to express their opposition to removing abortion rights in a way that didn’t get tangled in partisan politics. That is, some large chunk of Kansas voters were free to be Republicans while still supporting abortion rights.
This made me think that referendums could be a way to escape our current era of extreme partisan polarization.
The general reasoning goes like this. Politicians have policy goals, and they want to be reelected. (They have other goals too: living in a nice place, making money, not spending all their time on the phone asking for donations, etc etc., but let’s keep it simple for now.) In a polarized environment, you sometimes have to take extreme positions to satisfy primary election voters. And when it comes to policy goals, if elections are mostly determined by party ID, the economy, and timing, then you might as well swing for the fences on policy and get what you can when you can get it. By decoupling some issues from partisanship, referenda can get policy to a more moderate position.
I send the above-linked post to some colleagues who shared some thoughts.
Justin Phillips:
This is interesting and I do not really disagree. Simply put, referendums are a way to get liberal policy in a red state and conservative policy in a blue state. And where they exist, voters tend to like them and elected officials in the majority party tend to hate them. I am not sure if they are a “fix” for polarization but in my opinion they can be useful and I would like to see them in more states (though I prefer initiatives to referendums). Of course, they are awful for unpopular minorities….just ask the LGBTQ community which lost something like 35 state-wide ballot measures in a row (until public opinion changed).
The public has never supported unfettered access to abortion nor have the supported total bans. Both parties seem stuck with the policy preferences of their most extreme members.
Dan Hopkins:
The ability of initiatives/referenda to defuse some of the extremist policy outcomes depends on how they are used, and the underlying dimensionality of political conflict. In the Republic of Ireland, holding referenda on divisive social issues like divorce or abortion has enabled the party system to continue to divide the population on more distributive issues. (That’s the story of Sean McGraw’s book.) So if we could use initiatives/referenda to channel certain issues, we could potentially shift the issues on which the party system is defined. That said, in an era of nationalized party coalitions, I am doubtful that the limited use of state-level initiatives/referenda to meaningfully shift the party system.
Greg Wawro:
My sense is that elites are fomenting polarization, presumably because they see political gains from doing so, and they will have a big role in determining how referenda questions are framed and how the discourse around them will be shaped. Thus, if their interests are aligned with a polarized polity, they will do what they can in shaping referenda to maintain that state.
California has long had referenda-centered politics. Can we say that it is less polarized than other states?
I replied to Greg:
I think maybe California is less polarized than some other states, but it would be hard to measure, as I think polarization as we are experiencing it now is inherently a national phenomenon. That is, California’s left-wing politics are part of national polarization but are not polarized in themselves.
Interesting point about (some) elites fomenting polarization. Maybe that’s one reason that a referendum can push back, because it’s an aspect of politics that’s less controllable by political elites.
Mo Fiorina:
It all depends on who gets to design the referendum question. I don’t think I’ve ever said this in print but in a lot of talks I’ve said that if I were allowed to design national referenda on abortion, guns, etc., I could get 60:40 majorities or better on these contentious issues. And yes, the key to why the alternatives in the current system are so poor is party polarization. The Kansas recall was unusual (I think) in that the status quo was reasonably moderate and the referendum was radical. If the referendum had been extreme in the other direction (eg. abortion on demand in the third trimester), it would have failed as well.
I responded to Mo:
That’s a good point about the wording—indeed, the polarizing aspect of Brexit can in large part be explained by how close the split in public opinion was.
But, regarding the wording of the Kansas referendum: that wording was not exogenous, right? It’s more a product of political polarization, that the anti-abortion activists within the Kansas Republican party got greedy, in the same way that the Republicans in the Oklahoma legislature got greedy. Political polarization pushed them in that direction and kept the moderates within the party quiet.
To which Mo responded:
Yes, that’s what I’ve argued in the past three books.
Bob Erikson:
There is a literature (Matsusaka, not saying he’s right) that claims referenda produce more conservative policies.
It’s “complicated.” I will note that just like there are examples of liberal referendum successes like Kansas on abortion, there are lots going the other direction—NIMBYism for example. And lots of stupid anti-tax referenda have passed. Referenda allow for correcting legislature abuse but also provide a device for demagoguery.
Yes, we should not assume that reducing political polarization will lead to policies that we prefer (whoever “we” happen to be). The hope would just be to avoid some of the worst of the tangles of U.S. politics and political extremism.
>”In a polarized environment, you sometimes have to take extreme positions to satisfy primary election voters.”
If the locus of the problem is primary voters, then eliminating primaries seems the more straightforward approach. Ranked ballots (either instant runoff which seems to have more mindshare or Condorcet which has cleaner aggregate behavior) or approval voting both handle large fields of candidates without requiring a two-stage election.
Two thoughts, from a Swiss citizen:
1) Do referenda defuse political polarization? Yes! The best example of this are the national referenda Switzerland held on federal covid-19 regulations, which were challenged by a coalition of anti-vax, anti-government and anti-science actors. For months, they led a Trump-inspired vitriolic campaign with flag-waving marches, online and offline harassment of opponents, conspiratorial claims to speak for the “silent majority” rising up against the “dictatorial state” — trying to stoke the sort of polarization that is undesirable in a democratic society. When they finally lost two referendum votes in a row, it became clear that their views were not shared by a majority, the coalition began to disintegrate, and the political space again became pacified.
Why can’t elections achieve the same thing? First, because politics are multidimensional and “noisy”, elections will not typically reflect popular preferences on any single issue. But isolating the single, divisive issue of covid-19 regulations, and settling it in a referendum, was required to delegitimise conspiratorial narratives of a “silent majority” that are so pernicious to democratic debate. Relatedly, the referendum does not (to a first approximation) entail delegation of policy-making to representatives, and thus kills any narrative opposing the “people” against the “elites”.
Note: What I mean by saying that the referendum defused polarization is not that views on any particular issue converged. They probably did not. What I mean is that the referendum result made it hard to sustain conspiratorial, populist narratives, which are responsible for much of the undesirable toxicity in contemporary democracies.
2) Justin Phillips writes that “[referenda] are awful for unpopular minorities…”. I would want to know: Which system of governance isn’t?
> the polarizing aspect of Brexit can in large part be explained by how close the split in public opinion was.
I think there are a lot of reasons for the Brexit referendum being so divisive, and a lot of the perception of that is because of what happened after the referendum rather than during it. I felt the 2017-19 parliament was *much* worse than the referendum itself and there are many reasons why things ended in the way they did, which were specific to British politics at that time.
I think the three things that led to the referendum itself being such a problem are:
* The question wasn’t remotely clear – leave can cover anything from no deal to EEA and remain can cover reluctantly keeping the status quo to joining the Euro and pushing other countries for further integration. I think this uncertainty is what led to such negative campaigning, because nobody really knew what a victory for either side meant. Putting pre-written legislation in is a much better option for referendums, something that the US needs to be careful of if there is a referendum on the 2nd amendment.
* Brexit is a big issue, which covers economic policy, foreign policy, and national identity, and big changes tend to need big swings in the electorate (which happened in 2019).
* The result was too narrow for the size of the issue. If this result was on cannabis legislation, it would have been passed easily within a few months because very few people have a strong opinion.
However, there is now a political settlement that we are out of the EU and while some people may not like it, nobody who seriously wants to be Prime Minister for the next thirty years will dare propose rejoining so the pain settled what was a contentious issue bubbling under the surface. Sometimes it can be good to bring disagreements into the open.
> Putting pre-written legislation in is a much better option for referendums, something that the US needs to be careful of if there is a referendum on the 2nd amendment.
Do you think, with the benefit of hindsight, that this is something that could have been fixed in the case of Brexit? Of course the question can always be clearer, but short of pre-negotiating the terms of exit (quite a difficult negotiation to do) it seems there was always some inherent ambiguity in how things would be resolved.
In the case of the US amendments, they are by precedent relatively quite short, and since they’re so tough to change, it’s hard to imagine we would actually have much elaboration on policy specifics for gun control in any changes to the amendment.
Moritz wrote:
“…conspiratorial claims to speak for the “silent majority” rising up against the ‘dictatorial state’”
+1
This does indeed seem to be precisely where referenda can help. The referendum in Kansas fits nicely here as an example.
“The referendum in Kansas fits nicely here as an example.”
That’s a pretty tenuous claim, considering the Kansas vote was polling slightly pro-life, and of course here we worship the sanctity of the pollster. I guess you’re saying the silent pro-choice majority was actually pro-choice? They hid from the pollsters? :)
Andrew claims that in Oklahoma the 51% pro-choice majority in polls represents “public opinion”. However, in Kansas pro-choice was in the minority in polls (47-43 / 10 undecided, with the vote landing at 41-59). So does the Kansas vote imply that polls aren’t accurate? Or does it imply that the voting reflects a bias in turnout?
So I ask again: which tally is the actual “public opinion”? Which brings us True Democracy? The poll or the vote?
Chipmunk –
> of course here we worship the sanctity of the pollster.
Assuming that you’re not just trolling, could you provide some evidence for that characterization?
> However, in Kansas pro-choice was in the minority in polls (47-43 / 10 undecided,…
The referendum was on banning all abortions. Just considering a dichotomous variable like “pro-choice” (versus “pro-life”) leaves out a lot of important context. I doubt thhat in very many places, a “pro-choice” item on a survey would equate precixekybwirh an “against banning all abortions” item.
This is another example of why referendums could potentially lead to less polarization than rhetorical framing that is based on facile labeling.
… would equate precisely with…
Just consider that pro-choice people showed up to vote and pro-life people didn’t. Then all your analysis about dichotomous variables is meaningless puff.
Chipmunk:
Joshua is accurately pointing out that lots of people fall in between absolute pro- and anti-abortion positions. For example, a quick google reveals this from Pew Research, “About six-in-ten Americans (62%) say abortion should be legal in all (29%) or most (33%) cases. Around a third of the public (36%) says abortion should be illegal in all (8%) or most (28%) cases.” On this four-point scale, 41% of respondents take an intermediate view.
So for you to say to Joshua, “all your analysis about dichotomous variables is meaningless puff,” contributes negatively to this discussion.
Joshua:
Yeah, I hadn’t caught that “sanctity of the pollster” line. That’s really obnoxious, especially given that so much of my work is on survey adjustments and nonsamplng errors.
Andrew –
It’s much easier to convince yourself that your arguments are better than the arguments of those you disagree with, if you reduce what they say into stupid caricatures.
Chipmunk:
What Joshua said. Public opinion is complicated, and political polarization is a simplifying factor than in the opinion of many (including myself) can lead to bad outcomes, unnecessary distrust, cynicism, etc.
To answer your questions:
1. Yes, the polls “aren’t accurate”—at least, not exactly so. My colleagues and I published an article on the topic.
2. Yes, voting “reflects a bias” in turnout, in the sense that turnout is not a constant, and as I mentioned in my earlier post there has been a concern that low-salience referenda can be manipulated by big spending on one side or another.
3. No tally is “the actual public opinion.” Public opinion is complicated. Most Kansans prefer the Republican party to the Democratic party. In addition, most Kansans oppose the strict restrictions on abortion that are favored by the Republican party. The referendum on abortion gave an important subset of voters an opportunity to support abortion rights while not supporting the entire Democratic party package.
4. There is no “true democracy,” capitalized or otherwise. Again, public opinion is complicated. We have a federated representative democracy in the U.S., and there’s no reason to think that is perfect, any more than one would think that a system based purely on referenda would be perfect, or that an elected dictatorship would be perfect, etc.
“Public opinion is complicated”
Thank you. That’s my point.
“political polarization is a simplifying factor”
I agree you’re missing the boat on “polarization.” But I think it is pretty simple.
It’s just the latest gimmick Democrats are employing to convince themselves that their issue positions are just what everyone wants, if it weren’t for (Russian Election interference, X2, X3, X4, Polarization).
Our paper publishes about polarization frequently. It always goes something like this: “If we could just get over this polarization, we could work together to pass the Liberal Agenda”. ROFL! I wish I could see the light: even more homeless people, a massive crime wave, no pesky police, transportation gridlock, screaming property and now energy prices; higher taxes; reparations for the people my relatives died to set free; it all sounds so wonderful, if I could only accept Leftism into my heart!!
The Devil’s in the details. California’s referendum process has produced some truly awful results like Prop 13 and the California Coastal Commission/Prop 20. What is subject to referendum, what happens when a referendum passes, who can fund them, how does repeal work, etc?
Both this comment and Greg Warren’s in the original post are wrong. California has both initiative and referendum processes that allow citizens to take control. A referendum deals only with existing legislation while an initiative allows for new legislation to be proposed. The Prop 13 capping property taxes was an initiative spearheaded by anti-taxers led by Howard Jarvis. It’s important to distinguish between the two: https://ballotpedia.org/Initiative_and_referendum
Alan:
Thanks for clarifying. In my post I’m using common language and considering initiative to be a subset of referendum.
It is interesting though to consider how the two mechanisms may differ in their potential to reduce political polarization. In the terminology as used in California, a referendum becomes a check on unpopular legislation that nevertheless passes, perhaps because it appeals to an important sub-constituency in a polarized setting. A direct initiative, on the other hand, can be a way to settle issues without the party having to alienate constituents or subfactions.
However, that depends on whether the legislature is allowed to modify or repeal initiatives, and how large of a majority is needed to do so. If the legislature can easily modify an initiative, then an initiative hardly settles a polarizing issue. But if they can’t, then the issue, no matter how contentious and energizing to a subfaction of a party, becomes a moot question (at least from the perspective of the legislature). In that respect, California is a bit more extreme compared to most states: the legislature cannot directly modify passed initiatives at all.