The recent referendum in Kansas (in which 59% of voters “decided against removing the right to abortion from the State Constitution”) made me think about the general idea of referendums as a way to defuse political polarization.
This came up a few months ago around the time of the Supreme Court decision, when we discussed abortion attitudes in Oklahoma. According to Pew Research, 51% of adults in Oklahoma say abortion should be legal in all or most cases, and 45% say illegal in all or most cases—but when the state legislature considered a bill that “prohibits abortion from the moment of fertilization,” they voted 73 to 16 in favor.
The extreme vote in the legislature, so starkly different from the balance of public opinion in the state, is not a shock. As I wrote at the time, it does not defy political gravity for a legislature to vote in a way different from public opinion: issues are bundled, the whole thing is tangled up with national politics, also there’s some sort of pent-up demand from activists who can push anti-abortion legislation in a way that they could not do for fifty years. So, lots going on.
The point is that voters don’t have many options. If you want to vote Republican, you pretty much have to choose anti-abortion. Polarization in action.
A referendum, though, opens up more possibilities.
The role of referendums in a representative democracy is not always clear. Sometimes political scientists have opposed referendums on the grounds that they sidestep the political process. A referendum just gives you one choice, but on a complicated issue, legislators have staffs and can evaluate, deliberate, and find a compromise solution.
Given this, I can think of three reasons to have referendums on major policies:
1. Political polarization. As in the Oklahoma abortion example, sometimes a legislature can’t or won’t “evaluate, deliberate, and find a compromise solution.” That’s fine—it’s not the legislature’s job to compromise on an issue that they favor by a 73-16 margin—but it’s a failure from the standpoint of representing the popular will.
2. Principal-agent problems. A few decades ago, California had successful referendums on taxes and term limits. These are two issues where legislators of both parties are, to some extent, interested parties: tax cuts reduce the government’s power, and term limits threaten to remove people from the legislature entirely. So an extra-legislative solution can make sense.
3. Finally, it’s a safety valve: the threat of referendum can motivate a legislature to action.
I’m not saying that referendums are absolutely necessary: in a functioning democratic system, officeholders can ultimately be removed from office by the voters. But, given the general view that politics in the United States is too polarized, it might be worth considering the value of referendums as a force for moderation.
Having said this, I guess we should consider the opposite position, which is that referendums can increase polarization, with the example of the Brexit vote in the U.K. And, hey! here’s a research article, “Divided by the Vote: Affective Polarization in the Wake of the Brexit Referendum,” by Sara Hobolt, Thomas Leeper, and James Tilley, who write of “affective polarization, not by partisanship, but instead by identification with opinion-based groups.” I guess the difference is that British politics is not polarized by party in the same way as politics in the U.S. Yes, voters have strong party allegiances in Britain, as they have in the U.S. for a long time—but strong party ID is not the same as political polarization. Consider, for example, the U.S. in the 1950s.
Anyway, I think the topic is worth further research.
British politics is polarised by party on many issues – but the Brexit issue happened to cut across left and right (essentially, it was the moderates vs. the extremes on each side). You might see the same thing, albeit with different numbers, if you asked Americans about their support for Ukraine.
Referendum results are only as good as the referendum law defining who/what/when/how can do with them; there is much creative use toward bandwagon effects, mobilization, whatnot. I am fairly curious if the US is new to the practice, as that one referendum seems to be making this much news.
Av:
The rules on referendums and the frequency of referendums vary a lot from state to state. There are no national referendums in the United States. Perhaps the most famous U.S. referendum was Proposition 13 in California in 1978, which cut property taxes.
Wow, we must have been writing about Prop 13 at the same time!
Prop 2-1/2 in Massachusetts, in 1980, was a pretty big deal also.
Hi Andrew:
A thought-provoking blog post as always. I’m glad you brought up the down-side like Brexit. For me, when I got to 2. all I could think about was Prop 13
https://en.wikipedia.org/wiki/1978_California_Proposition_13
As a Californian, I question whether either Prop. 13 or the term limits referendum, reduced polarization. The same is true for a lot of the referendums we vote on, such as a recent one that allowed more gig workers to be classified as independent contractors instead of employees. Or, the one that abolished the state income tax, which was steeply progressive. A big part of the problem is that, under the ‘money is speech’ doctrine, the people can be paid to gather signatures to qualify referendums for the ballot.
> As a Californian, I question whether either Prop. 13 or the term limits referendum, reduced polarization.
It’s an interesting question. How could it be measured?
I’m not sure where, in my own priority list of reasons for referendums, where “reducing polarization” would rank relative to other advantages.
> A referendum just gives you one choice, but on a complicated issue, legislators have staffs and can evaluate, deliberate, and find a compromise solution.
Assuming there is only one choice in the referendum (i.e., nothing like rank choice referendums) or only one referendum regarding a particular issue on the ballot). And of course, that’s only looking at one referendum without any subsequent follow on referendums at a later time. In practice, sometimes referendums overturn recently passed legislation or previous referendums.
What about referendums with rank choice voting?
From a friend:
The Newsom recall example is technically multiple choice: Q1: do you support recalling the current Governor ? Q2: if so who should replace him (multiple questions with multiple choices for question #2.
The general thinking is the more complex the question the harder the campaign is in terms of communication. Ex. Do you support a 1% increase in the state sales tax to pay for xyz versus Do support an increase in the state sales tax and how much should it be (.25, .5, or 1%?). Remember we also have to leave room to explain in some detail what we’ll spend it on. There is also a monitor on the number of words on the ballot (we’re trying to change that) as well as a single subject rule for measures.
From me – I know that an issue with rank choice voting is the complexity and also that’s an issue with referendums. So combining the two….
But, maybe we should reconsider the race to the bottom in our electoral process.
Referendums mostly have two options. In the US, they are likely to increase polarization.
> . In the US, they are likely to increase polarization.
Is that an argue by assertion thing or would you care to explain your reasoning?
Seems to me a big part of why there’s so much polarization in this country is the influence of party identity and identity-aggressive/identity-defensive reasoning (a flavor of motivated reasoning). And we mostly have a “two options” party system
Referendums might largely circumnavigate that influence.
Maybe if in contrast we had a multi-party representative government (like in Norway) where parties need to build alliances, then the “two choices” downside of binary choice referendums would be more applicable.
Which one of these is “polarization”? I can’t tell if it is the public or legislature that is more polarized.
Besides how to measure and define polarization, I also wonder about the assumption that polarization is something that should be reduced. After all, at the other extreme lies a monoculture of groupthink. Lack of diversity can be even worse than multiple opposing agendas.
While I am not religious at all, in fact probably the least religious person who posts in this thread (I won’t go into a church even for weddings or funerals), I do see that the competing authority figures of church and state work to temper each others excesses. Many people rely almost solely on authority/consensus heuristics so it benefits them to have multiple authorities and be part of multiple groups that differ in opinion.
> I also wonder about the assumption that polarization is something that should be reduced
It’s not polarization per se that is the problem, imo. Not all polarization is the same.
The party identification-based polarization we have in this country, which basically splits the public into two enemy camps, effectively immobilizes our political processes. If you believe that the founders designed a reasonably good political structure, then it seems pretty clear (imo) why party-based polarization makes that design relatively disfunctional compared to times when polarization wasn’t necessarily less, but party-based polarization wasn’t as prevalent.
Of course, there are those who think that disfunctional government is the best type of government, but I can’t recall seeing anyone make an argument that comes close (imo) to demonstrating non-governmental societies that out-perform government-based societies along any meaningfully significant metrics. In my experience, that viewpoint tends to be associated with rather facile binary thinking about “unintended consequences.”
I don’t think so. Instead it is ~50% of the population that is in neither camp. Then ~25% each of republicans and democrats. And out of each of those 25% only about half are really devoted to the “team”.
So really the polarization I see is something like 75% vs the 25% who have heavily bought into democrat vs republican and are so politically active they keep forcing the others to ignore it or side with one of the two main teams.
> I don’t think so. Instead it is ~50% of the population that is in neither camp. Then ~25% each of republicans and democrats. And out of each of those 25% only about half are really devoted to the “team”.
Well, sure, I was being too broad there. The actual numbers are important to quantify, and quantifying the polarization around political identification is complicated, and quantifying the impact of partisan identification (among the VOTING public) on the functionality of our government is the more specific question I was thinking about – in which case the primary focus would be on voters not just the general public, and perhaps even more so frequent voters and the types of voters who would likely vote in off-year elections.
Not sure what the actual numbers are there, but I’m not sure how applicable your quantification is.
But like I said, it’s worth some. Googling when I get a chance.
Just a very quick Google netted this as the first hit.
Yikes:
https://fivethirtyeight.com/features/just-how-many-swing-voters-are-there/amp/
Regarding swing voters, see here.
I wasn’t thinking of swing voters, although they should be related. I was thinking of people who don’t consider the other team “enemies”. Politics is a relatively minor issue in their lives and they don’t personally feel this polarization between left/right being discussed.
Essentially I am saying there are vocal/visible minorities of democrats/republicans that get way too much attention. They do not represent the population so you will get a very wrong picture by only looking at them.
Andrew said:
“Regarding swing voters, see here.”
Here’s a different way to think about swing voters: put all voters along a line of political views with “far right” on one end and “far left” on the other end. Lets specify that each end is numbered “1.0” – one unit from the center. The center is 0.0. Presumably, swing voters would sit in the middle quarter or third of that line, centered around 0.0.
Now extend the left end of the line to 1.3 units from the center and move the center mark accordingly (0.15 units to the left). You’ve just shifted some proportion of the swing voters out of the “swing” category and into the “right” category. From the voting standpoint the country appears to be more conservative, but what’s really happening is the left is moving away from the political center, so voters that were centrists are now in the right camp.
That’s what happened with Clinton / Trump. Obama’s final term and Clinton’s agenda shifted collectively shifted the spectrum left – away from swing voters – while the voters stayed in the same place.
It doesn’t make sense to talk about “swing voters” without talking about the relative width of the political spectrum from one election to the next. IMO from BClinton1 to Obama 1, the spectrum shifted right: BClinton enacted welfare reform, three-strikes laws and other legislation to the right of Dem’s normal political positions. That was followed by Bush’s non-extreme Republican presidency; and Obama was elected as a relatively conservative Democrat. Obama shifted dramatically left in his second term, HClinton followed suit and lost the election.
Andrew if you want an interesting political science project, create a “left-right” rating scheme, apply it to all candidates, *then* track the swing relative to the *political* center rather than party vote.
Chipmunk –
>… but what’s really happening is the left is moving away from the political center, so voters that were centrists are now in the right camp.
Assuming you’re not just trolling, I hope you realize how subjective that statement is.
Joshua said:
“I hope you realize how subjective that statement is.”
A while back a famous political cartoonist depicted women lining up in Idaho to come to WA for abortions. I’m like, where’s the fetuses lining up in Washington to get into Idaho? Dude is so deep in his beliefs he’s completely oblivious to half of reality. I’m pro-choice. But I can still see past the end of my nose.
Got mis-nested.
Chipmunk –
Not sure what that comment means
But if you’d like to offer some evidence to support your opinion (stated as fact) feel free.
“it does not defy political gravity for a legislature to vote in a way different from public opinion”
In your Oklahoma article you didn’t present evidence that legislators voted against the public opinion in their particular districts.
But what really constitutes “public opinion”? The voting booth or the poll? In Kansas the polls were much closer than the vote result. Which *really* counts? Voting turnout can be influenced by a wide variety of factors – weather, current economic conditions, other candidates and issues on the ballot, etc – whereas polling is subject to numerous possible errors.
What’s the effect of advertising dollars on voting turnout? I don’t know for issues or initiatives. But from what I’ve read candidates with the most money usually have that money because they’re the most popular candidates. I don’t follow the issue but I haven’t heard of any research that shows money causes elections to be decided against public opinion. Our local newspaper is now making a big deal about industry campaign contributions to an anti-GMO foods initiative which lost, insinuating but not claiming outright that hidden contributions shifted the election. However, in addition to being a hotbed of anti-GMO left-wingers, we’re also a hotbed of hi-powered science and biotech and at least a warm-bed of agriculture, which are all strongly pro GMO.
The other apparently unsolvable conundrum is the difference in action from vote to poll: you can say anything you want in a poll and nothing happens. There is no result. OTOH, when you punch the box you’re making a commitment. I poll strongly for Party Y, but Party Y’s wacko candidates sometimes force me to vote for Party Z or not at all.
So really its a pretty complex game. Its not just as simple as having an initiative.
Oh, and with respect to advertising dollars, especially on initiatives, shouldn’t we really look at the presentation of the advertising to see if it presents the issues fairly? IMO it’s pretty hard to present a legitimate claim that GMO’s should be banned. There’s just no scientific evidence supporting that claim. Zero.
Cipmunk,
I mostly agree regarding GMO, which is a process better called genetic engineering. Whether a GMO product is good or bad depends on what the product is, not how it got created. On the other hand, having run for office (successfully, twice), I can assure you that money matters. As Jesse Unrah, a Democratic Party boss of decades past said, ‘Money is the mother’s milk of politics.’
With regard to money, I’m sure in a close race money matters. But I’ve also read that candidates who have lost previous elections usually don’t fare much better when given sizable additional funds.
I doubt Rand Paul could win a presidential election, no matter his cash pile.
The point of a Democratic Republic is to outsource political decisions to representatives, with some form of democracy to decide who the representatives are. Referendums bypass that structure.
I can picture all sorts of reasons why a Democratic Republic might not work well, but it is not at all obvious that bypassing that structure in this particular way cures any particular problem. And it might work well for some issues but not others (would you have a referendum on steel tariff levels, for example?) And I’m not sure a large disconnect between opinion polling on some issue and legislative votes on the same issue *ought* to be resolved by asking the people what they think. Majority rule by geographic district, the linchpin of a Democratic Republic, is obviously not the same thing as majority rule over all. And everyone knew that when they set up the system in the first place. Feature, not bug.
Jonathan:
I disagree with your statement that “everyone knew that when they set up the system in the first place.” Rules depend on context. The U.S. constitution was set up at a time when political parties are much different than they are today, and when geography had a much different meaning than today. Political polarization as we know it has changed a lot just in the past 20 years.
But even then the system was set up so that the more numerous North couldn’t simply get what they wanted against the less numerous South. And lots of things were set up to require substantial supermajorities (treaties, impeachment, veto overrides, amendments) and others were set up to be immune to voter preference at all (Bill of Rights.) So while I agree with your point about dysfunctional polarization, the notion that “majority numbers should be the ruling numbers” had significant caveats from the beginning.
Jonathan:
I’m not saying that “majority numbers should be ruling numbers,” nor am I even saying that referendums in general would lead to non-polarized politics. Rather, I’m saying that in the current setting of very high partisan polarization, referendums could be a way to reduce this polarization.
Maybe. But to the extent that polarization stems from what Tyler Cowen calls “mood affiliation,” it’s not clear that you’d get much more than the polarization spit right back at you. There may be some discrete issues (abortion might be one, gun regulations might be another, immigration policy might be a third) in which the mood affiliation could be severed for the better. Or not.
Jonathan:
But polarization isn’t just a set of attitudes; it’s also behavior. If politicians are afraid of losing to extremists in primary elections, that pushes the system toward more polarization. But if politicians are afraid of pushing policies that are so extreme that they’ll get overturned in referendums, that could push things toward less polarization. Voters’ polarized attitudes about political parties are not just “mood affiliation.” The parties really do push some extreme positions (as can be seen for example in Donald Trump’s alliance with Alex Jones, to take a particularly extreme case of extremism).
Jonathan –
> it’s not clear that you’d get much more than the polarization spit right back at you.
I’d never heard of “mood affiliation” and a quick Google didn’t turn up much outside pretty much a pet theory of one guy. Perhaps you have some links?
On the other hand, identity-based motivated reasoning (or things like “reasoning motivated by political identity”) seems to have a pretty extensive literature.
I could see where issue-based referendums could help circumnavigate political identity-based polarization. One related example might be where “Obamacare” is so deeply polarized whereas polling on the underlying issues shows less polarization.
Andrew: I hear you, and it’s not like I feel strongly about this. But being “afraid of being overturned in a referendum” doesn’t sound like something a politician, particularly one who is only extremist to ward off more extremist political views in primaries, would be particularly afraid of. It might be something he welcomes, allowing him to become even more extreme knowing that the referendum process will correct his extremism and he can point to the external result to say: “I tried.” But like I said, maybe.
Joshua: I don’t think I’ve got a great definition, but it’s that political positions come in bundles that are sometimes difficult to mutually justify through some coherent worldview but which make sense as adherence to a party view. That sounds like what you’re talking about, although I’m not sure motivated reasoning is a necessary adjunct.
Some issues are supposed to “energize” the base, and parties (supposedly) deliberately try to “bundle” that issue to drive support for other less popular parts of their platform.
But are there cases of the reverse? That is, situations where party leaders recognize some stance on a particular issue is popular but impossible to bundle with the rest of their platform without driving away important constituencies. If so, you might imagine party leaders try to get such issues resolved as referendums in order to deny support to an opposing party or avoid an insurgent faction within their own party.
Is the Brexit referendum an example of this that perhaps backfired? (I’m afraid I don’t know British politics at all.)
What about marijuana legalization referendums in many US states? I’ve always had in mind that it was an example of activists doing an end-run around legislatures that refused to embrace a position that the majority advocated for. But maybe in some instances it was also deliberately posed as a referendum issue to avoid having to attach it to a party platform / cause dissent within parties.
“That is, situations where party leaders recognize some stance on a particular issue is popular but impossible to bundle with the rest of their platform without driving away important constituencies.”
I think that, in general, current progressive economic proposals, the kind of things that Sanders and Warren advocate, fit this bill. They poll well in opinion surveys, but adopting them would poison the donor well.
But I haven’t heard anybody propose that these be put to a referendum. (Of course, these policies would have to be implemented nationally to be effective, and we don’t have a mechanism for national referendums.
A referendum is a way of answering a question. So then the discussion becomes one of who is proposing the question. If the questions decided in a referendum are only ever going to be asked by the polarised parties then the referenda become a reflection of that polarisation. Switzerland has about 4 referenda per year, each with 3-4 topics for voting. There are a couple of ways in which subjects can be raised for these referenda but one of them is that any citizen can propose a change to the constitution if they can collect enough signatures to support their proposal. This seems to work, but Switzerland is a consensus based direct democracy.
In Switzerland regular referenda are part of the standard political process. I believe that in such a system citizens are encouraged to educate themselves better about political issues given that they regularly and directly take part in the legislation, and votings will not be determined by party commitment. I’m not that optimistic about this if referenda are one-off freak events.
Quoting from the OP, “As I wrote at the time, it does not defy political gravity for a legislature to vote in a way different from public opinion: issues are bundled, the whole thing is tangled up with national politics, also there’s some sort of pent-up demand from activists who can push anti-abortion legislation in a way that they could not do for fifty years. So, lots going on.”
Well, certainly there’s a lot going on, but the quote seems to miss the elephant in the room: gerrymandered districts. With all the splitting and packing of voters to enable supermajorities in the state houses and tip the scales in the House; it’s no wonder that many voters seek to use referendums to bypass the current form of representation.
Chipmunk –
Not sure what that comment means
But if you’d like to offer some evidence to support your opinion (stated as fact) feel free.