For a Right of Unilateral Secession, Part Three: Design

In the third and final part of my series summarizing my working paper, “Designing a Constitutional Right of Secession” (here are parts one and two), I examine the legitimate objections we can raise to a right to secede. Some of these other scholars have previously mooted, while others are apparently original. Regardless, I will argue that all of these arguments merely establish reasons to qualify the right to secede, rather than to abolish it altogether.

Strategic Demarcation of Territory

When secessionists can determine the territory subject to a referendum, they can include as much territory as they feel confident getting away with, while still winning a referendum. They may include territory dominated by opponents of secession, so long as the rest of the people living in the seceding unit can outvote them.

One way of dealing with this problem is to allow recursive secession: territories that oppose secession can themselves secede from the seceding state. Indeed, that may be a good idea, although it poses some risks: the central government could use recursive secession as an excuse to attack a secessionist state, and there will always be controversies about where boundaries should lie. What if secessionists don’t permit recursive secession? Simply banning secession altogether wouldn’t make sense as a response, since that would violate the rights of the many to safeguard the rights of the few.

A better solution is to limit the right of secession to top-tier geographical subunits of the state. Now, there are some problems with this solution as well. First, administrative boundaries may be morally arbitrary, just like
interstate boundaries. Second, relying on administrative boundaries to set the limits of regions that enjoy the right to secede gives central governments an incentive to manipulate administrative boundaries to dilute potential secessionist challenges. Third, allowing regions to secede along existing administrative boundaries may trap significant minorities within the new state. These are all very real problems, as Yugoslavia’s attempted recursive secessions demonstrate. The secession of Croatia trapped Serbs, and the secession of Bosnia trapped both Croats and Serbs. The Badinter Commission denied these groups a recursive right of secession, and therefore they saw their only option as war combined with ethnic cleansing to alter ethnic balances.

So the administrative-boundaries solution works best if there’s a good procedure for letting people decide to which region they will belong. There should be an easy way for people to secede by referendum and set up new administrative regions. It should also be easy to hold referendums in small areas on moving from one region to another. Had Yugoslavia followed these principles, Serb-majority areas of Croatia and Bosnia would not have been part of Croatia and Bosnia, respectively, at all, and there would have been no reason for war.

Even when there are no good procedures for making sure administrative boundaries line up with what people on the ground want, it is better to make existing regions the subjects of a legal right to secede than to allow secessionists or the central government to redefine the scope of the territory subject to secession without the agreement of the other side.

Irredentism

Irredentism refers to movements seeking to take territory from one state and give it to another. Irredentism needs to be regarded more skeptically than secessionism, because irredentism has often been a cause of war among states, including both world wars. Regional irredentists often instigate violence to try to draw in their “parent” state.

Legalizing irredentism by plebiscite would encourage states to meddle in each other’s domestic politics in hopes of boundary revisions. Instead, irredentism can usually be alleviated with generous autonomy, as has occurred in Aaland and South Tyrol. Where irredentism cannot be satised with autonomy, such as where the irredentist group is a minority in the disputed region (e.g., Northern Ireland), then a solution will be more difficult, and the states involved may not wish to rule out completely a future transfer of sovereignty. In most cases where a secessionist movement has irredentist potential (e.g., Nagorno-Karabakh), it is reasonable to impose a conditions on a right to independence that the new state will never be allowed to join its ethnic-kin state, that the new state will be demilitarized, and so forth.

Vague Referendum Questions

While the problem of “vanity secessions” is overstated, secession without a settled majority in favor of the move can happen when the referendum question is vague enough to get support from pivotal voters without a decided view in favor of independence. For instance, Quebec’s 1995 referendum asked voters whether they believed Quebec should “become sovereign,” because the alternative phrase “become a sovereign country” polled lower.

The terms of a secession referendum should be negotiated between the central government and secessionist or set forth in law. The recent Scottish experience is a model in this regard. The Scottish and British governments have agreed on the conditions for the 2014 referendum, including a clear question: “Should Scotland be an independent country?” The threshold for success is a simple majority. In the Union of Serbia and Montenegro, the threshold was 55 percent. There is nothing magical about any particular threshold, so long as it is above 50 percent. The key concern is that the procedures for secession be legally defined in advance of any secession attempt, in such a manner as to command the assent of both the central government and potential secessionists.

Transaction Costs and Bargaining Failure

“No-fault secession” might be analogous to no-fault divorce. Economists have found that no-fault divorce has increased divorce rates in the U.S., suggesting, per the Coase Theorem, that transaction costs prevent efficient bargains in marriages.

For example, if a husband is guilty of “fault,” then under laws requiring a demonstration of fault for divorce, the husband does not have a right to terminate the marriage, but the wife does. If the wife does not consent to the divorce, the husband will have to “buy” her consent, making both parties better off. If the laws allow unilateral termination of marriage, then the husband does have the right to terminate the marriage, and the wife must “buy” his
consent to remain in the marriage. No matter which laws are in place, if there are no barriers to buying spousal consent, marriages that both spouses want will survive, and ones that at least one spouse does not want will break up.

So why did no-fault divorce increase divorce rates? One possibility is that spouses derive most of their value in a marriage from the children, who are a kind of nontransferable public good (Zelder 1993). If spouses derived most of their value from private, transferable goods, then spouses could still transact efficiently, but if they lack the resources to compensate each other within marriage, then even marriages that could make both spouses better off will not survive.

The most plausible way that transaction costs could permit inefficient secessions is through the time-inconsistency problem. For example, a central government may wish to make a large, one-time investment in a region, in exchange for which the region agrees not to secede (politics dictate that the terms of the bargain will usually not be so explicit, of course). Once the central government has made the investment, the region has no incentive to abide by the bargain if it may secede unilaterally and keep the investment. As a result, the central government will not invest, and the region will definitely secede.

To solve this problem, allow regions to bind themselves to the union for a period of years. For instance, regional governments could be required to put up bonds if they want to hold referendums on secession. If the referendum fails and the regional government decides to hold another referendum within a specified period of years, then the bond would be forfeited. A simpler way to do it would be just to ban the holding of secession referendums more frequently than, say, every 20 years. The Scottish National Party has agreed to just such a requirement for Scotland.

Distributive Justice

Richer regions tend to be more secessionist, at least in democracies. The reason for this correlation is apparently that voters in relatively rich regions see a potential benefit in independence or full scale autonomy: cutting off redistributive flows to poorer parts of the country. Some theories of distributive justice generate strong duties of redistribution within advanced capitalist societies. Secession for the purpose of reducing redistribution
could be morally problematic on these views.

Nevertheless, this isn’t a good enough reason to ban secession. Were it otherwise, egalitarian theorists would have to believe that maintaining one’s independence against conquest by a poorer country is also illegitimate. Is there an enforceable duty to unify with poorer countries? Intuition suggests no. Unless we think mere status quo bias is a good moral reason, then if we accept there’s no duty not to resist conquest by poorer countries, there’s also no good reason to think there’s a duty not to secede from poorer countries.

At most, strongly egalitarian views of distributive justice justify a “thumb on the scale” in favor of tougher standards for secession, not a ban altogether.

Conclusion

To sum up, here are the elements of a well-designed legal right of unilateral secession:

1. The territorial subunits are the subjects of the right to secede, and ideally, these subunits would themselves be subject to popular consent.
2. The plebiscitary threshold for secession should be established with the consent of the territorial subunits.
3. Since a mutually agreeable bargain can’t be guaranteed, there should be an established, accepted procedure for arbitration of disputes over the secession process.
4. Secession plebiscites should follow international standards for clarity of question by including generally accepted legal language.
5. Irredentist disputes should be settled by the states involved. It is reasonable to impose higher thresholds for plebiscites on joining another country, or to allow a right of the rump state to intervene militarily if the seceding state ever attempts to join a third country.
6. Reasonable limits on the frequency of secessionist plebiscites are possible.