How do states act after they get nuclear weapons?

by James Fearon on January 29, 2012 · 50 comments

in Foreign Policy,International Relations,International Security,War

All this talk about a possible US and/or Israeli preventive war against Iran got me wondering about the historical record concerning the conflict behavior of states after they acquired nuclear weapons.  Does the rate at which states are involved in serious international disputes tend to go up, down, or see no change after they get the bomb?

Advocates of preventive war on Iran like Matthew Kroenig expect Iran to become much more aggressive if it gets nukes.  “Proliferation optimists” like Ken Waltz, by contrast, argue that we have repeatedly expected terrible things from nuclear-armed adversaries but repeatedly found, if anything, the opposite to be the case.  For example, both the Soviets and the US contemplated preventive strikes to prevent Mao from getting the bomb, as they considered him, not without some evidence, to be aggressive, dangerous, and fanatical.  But the Chinese bomb was arguably followed by a more status quo oriented Chinese foreign policy.

Especially in the last 5 years, a number of IR scholars have started to look for patterns concerning nuclear weapons and international disputes in the available quantitative data (see in particular work by Erik Gartzke and Dong-Joon Jo, Robert Rauchhaus, and Michael Horowitz).  Various interesting findings, but looking around I didn’t see exactly what I wanted:  Just a simple country-level examination of dispute involvement before and after nuclear acquisition.

So I put it together myself, using Zeev Maoz’s version of the Correlates of War’s militarized interstate dispute data.  The following graph shows, for each of the nine states that acquired nuclear capability at some time between 1945 and 2001, their yearly rate of militarized disputes in years when they didn’t have nukes, and the rate for years when they did.  Note that for the US we have no data on dispute rate without nukes in this period since we got them in 1945; the rate for non-nuclear years for Russia/USSR is only for 1945-1948; the rate for South Africa (SAF) is for 1982-90; and the dispute data only goes to 2001.

China, France, India, Israel, Pakistan, and the UK all saw declines in their total militarized dispute involvement in the years after they got nuclear weapons.  A number of these are big declines.  USSR/Russia and South Africa have higher rates in their nuclear versus non-nuclear periods, though it should be kept in mind that for the USSR we only have four years in the sample with no nukes, just as the Cold War is starting.

Now it could be that getting nukes means that other states become more likely to initiate a dispute with you, rather than you becoming more aggressive.  So maybe we should restrict attention to disputes initiated by the state in question.  This is shown in the next graph (using the coding in the Maoz data called “primary initiator”).


Same pattern, maybe a little less pronounced, but you still see all but USSR and South Africa decreasing their dispute initiation after they get nukes.

What happens if you control for other stuff, like aggregate GDP (a proxy for total military capability) or secular change over time for all states, in a statistical model?  I’ve done some of this, with a panel data approach using country and year fixed effects and clustering the errors by country.  I get that states see on average about one half fewer disputes per year when they have nuclear weapons, an amount that is close to “statistically significant” at p = .10.  For various reasons I wouldn’t put a lot of weight on this but it does suggest that the patterns seen above don’t go away, and in fact might be somewhat strengthened, when you control for aggregate capabilities and time trends.

Obviously the fact that the other members of the nuclear club generally didn’t get much more aggressive in their foreign policy behavior after they tested doesn’t mean that Iran won’t.  But I think it’s astonishing how weak a case for this we are hearing from the preventive war advocates like Kroenig, or politicians contemplating it like Ehud Barak as reported in the Times article.  Barak says he is worried that Iran would threaten nuclear attack if Israel were to strike at Hezbollah or Hamas in response to rocket or other attacks.  That is not a credible threat.

If Iran had some long and deeply held territorial claim on one of its neighbors I might be worried that its temptation to invade would go up (had Saddam had nukes in 1990, getting him out of Kuwait would have been much harder, or impossible).  But I don’t think it does, and that wouldn’t be a major concern for Israel anyway.  Iran’s Sunni-dominated neighbors are worried that a nuclear Iran would be in a better position to foment Shiite discontent in their countries.  But the fear that the Saudis and others would seek to acquire weapons in response to Iranian nukes will likely weigh on the Iranian leadership (whoever it is).  In fact this could be a consideration leading them to want to stay short of weaponization.  And even if they go to weapons, the risks of nuclear balancing could favor a more status-quo oriented Iranian foreign policy in this area.  That is what has happened in a couple of other cases of nuclear acquisition.

To be clear, I’d strongly prefer that the Iranian regime not get the bomb, mainly because of the risks of further proliferation in the region and attendant risks of preventive war and loss of control of weapons.  But attacking Iran seems likely to guarantee pursuit till acquisition, to more effectively license future attacks on Israel, and to greatly increase popular support for the current Iranian regime and a course of nuclear self-defense.  (Netanyahu is reported in this NYT article to believe that an attack might actually “be welcomed by Iranian citizens.”  If that’s his true view and not purely strategic talk, then sheesh, it looks delusional in light of the historical record on that one.)  On the other side of the ledger are vague, weak, or barely developed arguments and claims about terrible things Iran would do if it got nukes.

We’ve heard these same concerns before, regarding Stalin’s USSR, Mao’s China, Kim Jong-il’s North Korea, and about the mortal mutual enemies of India and Pakistan.  All these cases have been very scary, and it’s understandable that the prospect of a nuclear Iran is incredibly scary for Israelis.  But so far, in none of these prior cases do the more extreme fears look historically justified.


Comments on this entry are closed.

Previous post:

Next post: