When should international boundary lines change?  This is a difficult question where a number of factors come into play:

1. Does the affected national government give consent?

2. Does the change affect de jure boundaries?  How about de facto boundaries?

3.  Do local residents favor a change?

4.  Are there severe human rights abuses against local residents?

5.  Would a boundary change involve a great deal of violence?

Let’s illustrate a few of these points with examples:

There are separatist movements in both Catalonia and Taiwan.  In both cases, the relevant central government is opposed.  In both cases, the regions are de jure part of a larger country (Spain and China).  In both cases, many local residents favor independence?  So why does the Chinese government’s case seem in some sense “worse” than the Spanish case?

First, unlike with Catalonia, Taiwan has de facto independence.  And second, enforcing China’s claim to Taiwan would probably require a lot of violence, whereas Spain seems able to hold onto Catalonia without much violence.

Or take Kosovo and the Crimea.  Why was the West more sympathetic to a border change in Kosovo than in Crimea?  Probably mostly because the Serbs had committed human rights abuses in majority Muslim areas of former Yugoslavia, whereas Ukraine had not committed major human rights abuses in Crimea.

Kuwait (1990) and the Falkland Islands (1982) were two examples where boundary changes were successfully resisted and where there was opposition to change from local residents.  In contrast, Crimean residents did not exhibit strong opposition to the boundary change, and Ukraine did not receive outside military support.

In my view, the first item on my list is the most important.  That might seem odd, given that these decisions seem so arbitrary.  The Czechs allowed the Slovaks to secede.  Quebec residents and Scots were offered referenda on independence.  In contrast, Spain refuses to give that option to Catalonia.  So why is this principle so important, given that the enforcement is so arbitrary?

Central government consent is important because of point #5, the danger of violence.  The international community has decided that unless there is some sort of extenuating circumstances (i.e. Kosovo), boundaries should not be changed without the consent of the central government.  While this is an arbitrary rule, it minimizes the threat of violence.  Throughout the world, there are dozens of places where international boundaries don’t make sense.  If you allowed boundary changes without central government consent, it would open a can of worms.  There’d be many attempts for one region to break off and join another.  Often, the dispute would become violent.

You might wonder why unconstrained boundary changes would necessarily end up in violence.  If the local people agree, why can’t the change occur peacefully?  There are two problems.  First, the local people don’t all agree–it’s complicated.  The majority of Northern Irish don’t want to be part of Ireland, but a large minority of Northern Irish don’t want to be part of the UK.  In former Yugoslavia, the ethnic map is like a jigsaw puzzle.  Whenever you move the line, you hurt one group.

Economic considerations are also important.  If the oil-rich southeast of Nigeria declares independence, the rest of Nigeria would feel deprived of oil revenues.

Even with all of these considerations, there are tricky cases that are hard to resolve.  Thus how much weight should be put on de facto boundaries that are not internationally recognized?  The US considers Taiwan to be part of China (de jure) but treats it like an independent country (de facto.)  Somaliland is another place that is de jure part of another country (Somalia) but is de facto independent.  How long does de facto independence have to last before it becomes internationally recognized?  In my view, the best policy toward Taiwan is the one that minimizes the risk of war.  But what policy is that?  The same question applies to Ukraine.  What weight should be put on deterring future aggression?

America stole land from Mexico in the 1800s.  At what point do we no longer have the moral obligation to give it back?  Russia stole land from China in the 1800s, from Germany and Japan in 1945, and from Ukraine in 2014.  At what point does that theft become accepted by the international community?  At what point does the Golan Heights become part of Israel in a de jure sense?  Today, economic sanctions over the Crimea probably make more sense than economic sanctions over the Soviet theft of Königsberg or the Kuril Islands, but I’d be hard-pressed to provide a rigorous proof of that claim.  In other words, there are lots of “borderline cases”.  (Pun intended.)

Europeans have wisely decided (outside of Yugoslavia) to accept international borders as of 1945, even though (for instance) lots of Hungarians live in Romania. They recall that the German claim on the Sudetenland was based on there being a large population of ethnic Germans in that part of Czechoslovakia.  It also makes sense to accept mutually agreed border changes, as with Czechoslovakia’s “velvet divorce”.  This principle is mostly accepted in Latin America, although that may be mostly because it’s the weaker countries like Paraguay and Bolivia that lost territory in the past.  In Africa, central governments are weaker, and the use of force to change boundaries is still an ongoing problem.  But even in Africa, the principle of central government consent is viewed as being very important.

So what should we do about Ukraine?  Russian apologists insist that the Russians view NATO as a security threat.  That’s nonsense. Putin knows that NATO has no interest in invading Russia.  The real problem is that Russia has seller’s remorse over its decision to allow the Soviet Union to break-up.  It would like to grab back parts of the old Soviet Union that are ethnically Russian.  And it (correctly) fears that if those areas join NATO, it will be unable to do so.  Thus it’s too late to grab back the Baltic states.  NATO is not a threat to Russia, it’s a threat to Russia’s intention to grab territory from its neighbors.

On the other hand, I’d oppose admitting any country into NATO that does not have clearly established boundaries.  That’s a recipe for war.  One compromise might be to assure Russia that NATO would not accept Ukraine until or unless the Russian government accepted Ukraine’s international borders.  The same assurance could be made for Georgia and Belarus.  That would give Russia an effective veto, while holding out the prospect of an eventual expansion of NATO if a future Russian government is more reasonable.