Regionalism

This is the latest post in my constitutional design series. Up until now in this series I’ve been concerned with constitutional design regarding a single government of a unified country. This post is about what happens when you have regional differences that you want reflected in government somehow.

Breadth Requirements

There are countries with strong regional differences where you want buy-in from many different regions before enacting certain laws, which I’ll call a breadth requirement. You don’t want a more populous region to be able to enact laws over the objection of smaller regions.

The US may have been one of them at the founding (although I’m skeptical¹). It is not anymore. The places I’m thinking of are where there is a big ethnic (or religious, racial, etc.; I’ll just use the term ethnic for simplicity) divide that falls along regional lines. Many countries in Africa are like this, for instance, because country lines are drawn on colonial boundaries rather than pre-colonial ethnic ones.

I think you should consider splitting a region into multiple countries if the ethnic divides are vicious enough - I think the empirical track record of having single countries bitterly divided with breadth requirements isn’t amazing (see Austria-Hungary, Czechoslovakia). If you don’t, I think you should consider just having regular democracy with no explicit mechanisms preventing regional domination. Democracy is about every person having equal political power, not every region. As I wrote in my post on voting laws: Everyone has the same capacity for forming good political beliefs, and liberalism suggests that everyone ought to have equal political power. Everyone’s welfare matters just as much as everyone else, so people should have equal say so that their welfare isn’t sacrificed in favor of those with more power.

But if you’re dead set on having a single country with a breadth requirement, which I think can make sense in some deeply divided places, here’s how you’d do it.

First, how you shouldn’t do it. For some reason constitutional designers are obsessed with this idea of having a second legislative chamber (the Senate) where every region has equal say and which can block legislation. The idea is supposed to be that since you require both a majority of the population (represented equally in the other chamber) and a majority of the regions to agree on something, you should get the best of both worlds, right? Wrong, this is a terrible idea. The problem is that the Senate does not stop the passage of legislation that targets specific smaller regions, it just requires a majority of the regions to be in agreement. As long as a majority of the regions and a majority of the population agree on something, they can pass it.

This leads to bizarre and bad results. In the US for example, the Senate is insanely biased toward Republicans, which means going forward there’s a good chance they get a veto on all legislation unless Democrats miraculously pull out a Senate majority. This means that some of the time Republicans will be able to pass things with no Democratic votes, and the rest of the time Republicans and Democrats will both have to sign on. That’s tyranny of the minority, which is even worse than tyranny of the majority, what breadth requirements are supposed to protect against.

The better way of implementing breadth requirements is to require that legislation gets the support of, say, 30% of the legislators from each region². You could make this threshold 50%, but I think if 30% of the region agrees with something, that’s a sign that the issue is not actually that bitterly divided on regional lines.

I also think you should make this only apply to specific categories of laws laid out in the constitution, rather than all laws, (with amendments to these categories also having a breadth requirement). The default should be that laws need the support of a majority of the country to take effect. Only on the issues that different ethnic groups are really sensitive about should you require broad support. The one downside of having only specific categories with breadth requirements is that you need a single court responsible for determining what falls in the category and what doesn’t. You should probably add the same breadth requirement to nominations to this court, but even then you’re going to have trust people to make impartial decisions, which can be hard when tensions are high.

You can extend this logic beyond regional lines to explicitly ethnic lines. In that case you’d likely have a separately elected body representing certain ethnicities in which who you’d need 30% approval to pass certain laws. The problem is that you have to officially classify people by ethnicity to determine if they’re allowed to vote for this body. In addition to being kind of icky, it’s hard to do this accurately (people have political incentive to register as the opposite ethnicity), and there are lots of people who are mixed-ethnicity that you have to figure out what to do with. Still, I think this can be an ok thing to do, particularly if you have a very narrow class of laws they have to agree to. You can also just have protections for a minority ethnicity and not for the majority, which might be a better option for very small ethnicities.

Federalism

Federalism is when regional governments have some category of law set out that is either exclusively their domain or in which their laws override federal laws.

Countries in green have federalism. Though even countries in blue often have some form of de facto federalism even if they don’t call it that

If you want to implement federalism through a formal mechanism, you should use breadth requirements. The national government can add a law to a specific set of laws that override regional laws, but only if it meets a threshold in each region. I’m more ok having this threshold be 50% in this case, since other regions can always adopt their own policy on the issue. As with regular breadth requirements, you should make this apply to only certain categories of laws, a list of categories that can only be amended meeting breadth requirements.

I also think it’s often ok to just implement federalism with norms. This Founding Fathers framing of “the national government will always seek more power” is just untrue, as evidenced by the many countries without federalism or without constitutions to begin with where local issues like local road maintenance are left up to municipalities (see: UK, Israel, New Zealand).

You can also go further if you really hate local governance and disallow local governments from making laws at all, perhaps in certain areas or perhaps for all areas besides a specified set. (The latter is what they do for regions in France, which are pretty much only responsible for high school education policy).

The Pros

Why should you leave policy up to regional governments?

There are three main reasons you might give.

The biggest one is that you want public policy to more closely match people’s preferences. The justifications are then the same as the justification for democracy to begin with:

• People know the needs of their region best. This encompasses both:

  • Political views: people have the best knowledge of their own region that they can use to set local policy.

  • Personal interests: people are directly affected by the policy set by their own region and so are best in charge of setting it.

• People will develop strongly anti-government attitudes when they feel they have no voice. This can lead to worse national government policy (like a turn to more populist candidates) or, in cases where there are deep regional divisions, civil war

The second reason for regional power is that, even if the national government could match people’s preferences in every place, they’re simply too busy with other stuff to take care of it. Maybe if roads were left to the national government they would go into disrepair, not because national politicians don’t care that people want them fixed but because they’re focused on more national issues.

The third reason is that of “laboratories of democracy”, as Justice Brandeis put it. A region can make a law, and then people can study how well it worked and adopt it in other regions or nationally.

If you know anything about social science, you’ll know that what’s important is randomness, that the adoption of certain policies in places was done independently of all other effects that could cause the outcome you’re studying from that policy. Funnily, this means that ideally what you’d want is regional policy that does not represent the people well. You want policy whose adoption is independent of confounding factors that would make people in these regions want to adopt such a policy (factors which could also impact the outcome you care about). In a perfect world, what you’d want is for the national government to pass policy that only goes into effect in some regions chosen at random.

I’d really love for this type of thing to be done. Short of that however, regional governments can be an ok substitute. You can try to suss out the effect of policy by comparing things like trendlines, and good policy can sometimes have a big enough effect that you don’t even need to be a social scientist to see its effect. Regional governments also have the advantage of actively seeking out policy that helps their region.

The Cons

One thing that makes federalism a bad idea is that regional governments can often actually be less democratic than national governments. For whatever reason, people tend to pay more attention to national politics than to local politics. This isn’t a law of nature, you can definitely find times where this wasn’t true, but it is generally true. Now granted, just because people aren’t paying attention to regional politics, it doesn’t necessarily mean their regional politicians don’t represent them well, particularly on an issue that is sharply split on regional lines. But it does suggest it.

Another con of federalism is that one region’s policy very oftem affects other regions. People often underestimate how much this is true; blocking the construction of housing, for example, might seem to only affect people who live in that region but in fact raises housing prices for all nearby regions. Almost anything that affects the economy of one place can affect neighboring places due to trade. The Supreme Court agrees with this logic by the way, because they have interpreted the Constitution’s “interstate commerce” and “necessary and proper” clauses to give the national government authority over basically all fiscal policy and economic regulations.

You have to accept some degree of spillover. But if you have too much of it, you start to directly undermine democracy, as a region is affected by policy it doesn’t control. You can often run into prisoner’s dilemma type issues where a region adopts policy because it reaps the benefits while most of the costs are borne by other regions.

The last con of federalism is that, well, sometimes regions have bad beliefs. Much of segregation largely did not affect people in other states, was done with the support of the white populations in those states, and was nonetheless horrific. Sometimes you just want to give the national government power.

Guidelines

There are two obvious guidelines on when to use federalism.

First, don’t use it it when a region’s policy has a big effect on neighboring regions. This means no federalism for gun laws (you can bring guns across state lines) or marriage law (you can just get married in a different region; you should probably force other regions to recognize it or you face legal hell).

Second, use it when the conditions of different regions differ in policy-relevant ways, and not just their political views. If a region has a lot of farmland, or a lot of oil, or a lot of manufacturing, it makes sense to leave form, oil, or manufacturing regulation up to that region. It makes less sense to leave regulatory policy up to a left-wing college town just so they can use it to adopt more stringent regulations.

Besides these guidelines, I don’t think it’s super clear where to have federalism. A lot of it depends on how much you value laboratories of democracy, how much people pay attention to regional politics, etc.

International Agreements

Most agreements are purely voluntary and are not enforced by a court. Different countries agree to all do something that would be mutually beneficial, and if one country stops doing it the other countries stop too. In that case there’s not much constitutional design to talk about.

For more substantive agreements like the EU, where there are rules all countries have to follow and an enforcement mechanism, the considerations are very similar to those of federalism. The main difference is that it should be the default that national laws override international laws, and then have specific areas where the opposite is true. This is because people pay much more attention to national-level politics so the draws of federalism are much stronger there. Also, this doesn’t apply to the EU but in other agreements country borders do a lot to reduce spillover effects.

1. The immediate proliferation of national political parties in the US suggests that the country was divided if anything along regional lines rather than state lines. And even on regional lines I’m skeptical the divide was anything close to the ethnic divides we see in any other country

2. If you were doing list proportional representation, you would not have one national list but instead have a list in each region (or split the region up further into district). If you do this you should probably set any threshold on a regional level rather than nationally. The same thing applies if you’re using a multi-member system like single transferable vote (STV) or proportional multi-vote representatives (PMR).

If you’re using mixed systems (ie MMP or DMP), you have to set proportionality on the regional level rather than nationally. This can lead to a lot more list votes or worse proportionality.