This is part of my constitutional design series
In the US, there is a 60 vote filibuster supermajority requirement in the Senate for all legislation, meaning that all legislation has to pass on a bipartisan basis - except for bills using this reconciliation loophole that is limited to one per fiscal year and can only deal with fiscal issues1. Additionally, because of midterm backlash, a majority of the time there is divided government: no party controls all three of the Senate, House, and Presidency. Because of these factors, and because of rising partisanship, very few non-fiscal bills addressing partisan issues have been passed in decades. We haven’t had immigration reform since 1986, we got a tiny bit of gun control legislation under Biden but haven’t really done anything substantial since 1994, all abortion action has taken place outside of Congress, etc.
But supermajority requirements shouldn’t prevent partisan action like this. Just because legislation is passed on a bipartisan basis doesn’t mean it can’t contain partisan provisions that favor a certain party. If Democrats like certain gun control legislation more than they hate certain immigration restrictions, and Republicans like the immigration restrictions more than they hate the gun control legislation, the two should be able to do a trade - pass a bill that contains gun control and immigration restrictions.
Such trading is rare. It’s rarely ever seen for partisan issues, and when it is it’s almost always on a single-issue basis - the aforementioned 1986 immigration reform trading amnesty for tougher enforcement is a good example, as is the Clinton legislation trading increased welfare for work requirements. There’s quite a lot of this type of trading for less partisan issues taking place in the yearly budget and omnibus bills every year, and even that is less than it should be.
This post will explore the reasons for this.
Why is bipartisan legislation hard?
There is incentive to sabotage the party with the presidency
This is the most substantive reason I can point to and is something I touch on heavily in the first chapter of my series. Public opinion tends to revolve around the leader of the country, regardless of how much power that leader actually has. No matter what happens politically in a country, the leader is rewarded or punished. So if the public sees illegal immigration as a big issue and blames the leader for it, opposition legislators are not going to want to mitigate this by agreeing to an amnesty-border security trade. We know this is what sank the legislators’ attempts at immigration reform in 2024, and that one didn’t even have any amnesty in it.
Balancing priorities is hard
People are generally bad at prioritizing. If you asked a random Democratic legislator whether they thought a given gun control or voting rights provision was more important, most probably would have to think about it. Even harder, I think, is prioritizing among bad options; what would be worse, Republicans’ abortion legislation or another Iraq war? Since trading requires you to have a sense for what you would be willing to trade for what, this makes it difficult.
Priorities differ within a party
Not only is it hard for most elected officials to know their priorities, priorities among different politicians in a party can differ substantially. It’s hard to know this stuff in the US given the limited info we have about legislators’ positions, but I think it’s a safe bet that people on a committee for a certain issue care more about legislation in that area, for example.
Status Quo Bias
The last and perhaps most important reason I can think of for the how difficult it is to pass bipartisan legislation is status quo bias, the tendency of people to prefer the status quo over proposed changes, even if they might like the changes in a vacuum. Status quo bias is the biggest thing missing from the rational decision theory of politics that I wrote about as part of the foundation of my constitutional design series. It’s a really important effect that seems to slip under the radar until you actually think about it, and then you see it everywhere.
As with all alleged bias, we must ask:
Is status quo bias maybe rational?
I can think of two arguments for why status quo bias might be rational.
One is from stability. Public policy should be relatively stable so that people, businesses, other countries, etc. can work around it. There are downsides to stability - it allows people to discover and abuse loopholes, it can lead to outdated laws, and it means less data informing policymakers on what good policy looks like - but there’s clearly some legitimate value in it.
The other is deference to past legislators. If we live in a world where legislation gets thoroughly scrutinized before passage, that legislation should on average be better than any random proposed legislation, so you should thoroughly scrutinize new legislation before passing it as well.
Still, I think it’s pretty clear that we have too little lawmaking at the moment.
The two levels
Status quo bias functionally occurs at two levels. The first is in the public - people (both the electorate and elites/interest groups whose signals influence the electorate) punish politicians more for supporting bad stuff than they reward them for supporting good stuff. There’s political science research suggesting that “no” generally outperforms on ballot measures. Backlash to radical change - particularly midterm backlash - seems to be a consistent feature in US politics.
The second is among politicians. Politicians, being human as well, are repulsed more by bad changes than they are attracted to good changes. This is harder to find direct evidence of, but I think I’ve noticed it a bit paying attention to congressional negotiations, and just on priors I think we’d expect politicians to be influenced by status quo bias too.
Consequences For Constitutional Design
High Supermajority Requirements
The upshot of all of this is that when it takes a large supermajority (say 2/3) to change something, you can expect it to be in place for a long time.
This is first and foremost a problem in undermining accountability democracy. By kneecapping majorities with higher-level (ie constitutional) law that takes a supermajority to change, you do not allow them to pursue actions that could make them more popular, or that could indicate to the public what they believe. This is the source of many of the problems in the US.
Accountability democracy and consensus democracy are always tradeoffs, so this might be something you’re willing to accept if you want to require greater consensus for something before passing it. The bigger issues have to do with how law ages.
A provision can be too concrete and then become outdated. The provision that only licensed gun sellers need sales to go through background checks, for example, might have made sense when it was written but after the advent of the internet meant that virtually anyone can buy a gun online. Another good example is the Constitutional provision that requires jury trials for all crimes with punishment over $20, a number which (I am not kidding) is not adjusted for inflation.
Sometimes a system does not become outdated per se but was just subpar to begin with, and under supermajority requirements is then locked in. Presidential democracy in the US is a good example.
The other common failure mode is when a provision is too abstract, and thus gives a lot of leeway to the judiciary. This is especially problematic because the world that a law was written in can be very different from the world you currently live in. What the hell does “equal protection of the laws” mean? Does it cover puberty blockers for trans youth? What about gay marriage?
There are different judicial approaches to answer these questions, and I don’t want to downplay them too much; there is method to some of it. I just think you have to acknowledge that ultimately there is a lot of place for the individual opinions and biases of the judges to creep in. And once rulings just become judges exercising their discretion, you lose the advantages of the judiciary in the first place and just get an inferior version of the legislature: the court is likely not representative of the population, they are bounded by precedent, they can only make law in discrete rulings, etc.
For these reasons I’d be very careful about what laws to put in a constitution that take a high supermajority, eg 2/3, to change.
What about the stuff I talk about in this series, “democracy-relevant” law - does it belong in a constitution with supermajority amendment requirements? I think there is a good case for yes, that if you don’t add supermajority requirements politicians are going to change the rules for their benefit. With other areas of law, if the politicians in power change it you can just vote them out and revert; if you change democracy-relevant law you might not be able to vote them out.
On the other hand, ultimately politics is just norms all the way down - just as there is a strong norm to follow the law, there can be a strong norm not to change democracy-relevant law for unprincipled political gain. If you add supermajority requirements, that can lock in bad systems like the Senate that are near impossible to change because half the people in power benefit from it.
The optimal policy of course is to add supermajority requirements if you have a good constitution, but keep majority rule if you have a bad constitution. And of course I personally am the ultimate judge of good constitutional design.
Low Supermajority Requirements
All of the points above applies if the supermajority requirements are really high, like 2/3, which will require buy-in from both the left and the right. Low supermajority requirements - say, 53% - function differently. They’re basically a way of rewarding parties that get a larger majority.
These sound cool at first - make the lowest level of law (level 1) take 50%+1 to change, make level 2 take 53% (and override level 1 law), level 3 56% (and override levels 1 and 2), etc. But I’d really caution against using low supermajority requirements for higher level law like this.
One reason is that, as above, they hurt a lot with accountability, as the majority bloc will not be able to pass laws if they have a small majority. Again though, you might be ok with this accountability-consensus tradeoff.
More fundamental are, here too, practical issue with how law works. What it “really” means for a law to contradict another law can be quite ambiguous, so if a higher level of law is relied upon frequently, courts will have to constantly deal with cases on whether the low-level law contradicts the high-level law. This gives courts a lot of legislative power, which as stated above I think is generally bad.
There’s also the issue of executive power. If you have a supermajority requirements for minister or prime minister appointments, you are likely to much more often get no governing coalition. Probably the more sensible way is to keep simple majority requirements for minister / prime minister appointments even with low supermajority requirements for legislation. But then you run into a similar issue as the previous paragraph: a leader with a narrow coalition will always be looking to work around higher level law by testing the limits of lower level law, which will give the court a lot of power in policing this.
There are better ways of rewarding parties for larger majorities: party candidate selection or multi-party systems. If you force a coalition to represent the views of its members proportionally, there will be significant disagreement, leading to bigger majorities being very valuable. The issue in the US is that geography-based first-past-the-post primaries and general elections do not allow for such proportionality, leading to extremely homogenous parties.
Enduring Majority Requirements
There’s a different type of requirement that I’ll call an “enduring majority” requirement: require that bills be passed in multiple consecutive terms before becoming law. (Or you can go even fancier and require that bills be passed in, say, 3 out of 4 consecutive terms). I think these generally function as a worse version of low supermajority requirements.
If a party wins a few consecutive terms in a majority, often in at least one of them they’ll win a big majority, and thus be able to pass laws under supermajority requirements. In the cases that they don’t, I don’t think they should have some advantage just from holding a majority several times. If the public votes 51/49 in favor of left-wing parties multiple times, is that really legitimate enough to allow them to modify higher-level law?
Enduring majority requirements also have the awkward consequence that laws can become outdated between the time they are passed the first time and the second time.
I thought I’d at least mention these as an option though.
And except for a year after 2008 where Democrats had 60 senators, although the 60th vote on legislation was very moderate so they couldn’t get radical partisan bills passed.