Tuesday, November 30, 2010

Commentary on the Commentary on Korea

I've been meaning to post about the use and abuse of the word "rational" for some time.  But I've actually said most of what I want to say here, and Scott Wolford beat me to the punch on most of what else I might have said here.  So I'm not going to talk about how it's less than useful that so many people, pundits and scholars, keep repeating the meme that Kim Jong-Il is irrational.

I'm instead going to take issue with three other frequent claims: 1) Kim Jong-Il has deliberately fostered his reputation of being crazy because there is a strategic incentive for him to do so, 2) failure by the US or South Korea to take a hardline here undermines their reputation, inviting future aggression, and 3) recent actions by North Korea constitute an "act of war".

Monday, November 29, 2010

What Bargaining Looks Like

One of the areas where I think the study of international conflict has in fact made significant strides is the application of bargaining theory.  There are a few reasons I see promise here, which I'll discuss in a future post.

For now, I want to talk about a simple question that I think advocates of bargaining theory have not answered very well so far -- why should anyone believe that issues are actually divisible?

Saturday, November 27, 2010

If Our Theories Were "Right", Would We Know It?

I've mentioned a paper with Kyle Joyce a few times on here.  We've finished the current round of revisions.  You can find it here, and some thoughts on it below the fold.

Wednesday, November 17, 2010

What Your Control Variables Don't Do for You

Following up on this post, I want to say a little more about the problems associated with throwing every variable you can measure into the model and assuming that absolves you of any concern about omitted variable bias.

After this post, I'm apparently due to post pictures of puppies.  (I respond to my reader's whims, and the results of the poll so far have been clear.)

We're All Modelers

As I said yesterday, one of my big pet peeves is that there's no obvious consensus on what we mean by a lot of the words we use in this field.  My favorite example is the word "rational" (which, for that very reason, I try never to use in any of my research, even though most of what I do these days is game theoretic), but I'm again holding that aside for another day.  Today's post is concerned with "models" and "modelers".

Monday, November 15, 2010

Why Would Anyone Try to Test a Theory?

One of my biggest pet peeves about our discipline is that we refuse to each of us use words to mean the same thing.  For example, rationality means something very different to critics of "rational choice theory" than it does to just about every game theorist I've ever met.  I'll have more to say about this at some point.

Similarly, I think there are at least 30,000 different definitions of "theory".  So before elaborating on the provocative title of this post, I first want to make clear what I mean by theory.  I will then explain why I think it's meaningless to test a theory.

Friday, November 12, 2010

US Troops to Remain in Afghanistan Until 2014

Story here.  Who would have guessed the US wasn't really going to pull out in 2011, but would wait until a year or so into Obama's second term...

Thursday, November 11, 2010

My View on Making Quantitative IR Credible

In these two posts (here and here), I made some strong criticisms of quantitative IR (as it is typically practiced).  The idea that it's hard to draw causal inferences from observational data is not new, and the more methodologically sophisticated amongst us have been castigating standard practice for a long time (see here for an excellent recent discussion).

But some will understandably be unpersuaded by the fact that a sufficiently motivated crank can come up with a stylized example where the results we observe would be identical under each of two different stories that have fundamentally different implications.  My example, and the continued remonstrations from methodologists, only amount to a claim that you can't know for certain that your analysis tells you what you want it to tell you.  This is not the same as proving that the results you get from your favorite statistical package are certain to be wrong.  Fair enough.

Let's take this line of thinking further, below the fold.

What's Wrong with the Strongest Finding in IR

Following up on this post, I'm going to elaborate further on why I'm skeptical even of our allegedly strongest finding with respect to international conflict: the democratic peace.   And at the moment, I'm actually talking about the empirical association, my concerns about the theory held aside for the time being.

Tomorrow I'll post some thoughts on what I think we can do to increase the credibility of empirical work.  I doubt my views there will be terribly popular, even (especially?) amongst those who have been receptive to the critiques I've been making here.  So I want to be clear to disentangle the two points.  But for now, let's stay nice and negative.

We've been told over and again that the democratic peace is a very robust finding.

For a simple illustration of why you should not think this robustness is all it's cracked up to be, look below the fold.

Wednesday, November 10, 2010

The Shrinking British Military

A group of former admirals is ringing the alarm bell over the Conservative-Lib-Dem coalition government's radical cuts to defense spending (and every other area).  One of their key arguments is that the UK is inviting Argentina to retake the Falklands (Malvinas) Islands, though naturally we can't assume anyone will take us seriously these days if we don't include at least a passing reference to Hitler, so they do that as well. I, for one, am far more persuaded than I would have been without that.

Tuesday, November 9, 2010

Domestic Politics of an Attack on Iran

Tierney has a great piece on the likely domestic political reaction, both short term and long term, that we could expect if Obama followed the advice many are offering to attack Iran.

I would add a few points though.

Unanswered Questions About the Iraq War

With Bush back in the news, pushing his book and defending his record, I thought I'd bring up some of what, in my mind, are unanswered questions about the invasion of Iraq.

Depending on your partisanship, you probably think the answers to many of these are obvious.  I'd like to know how we know what we know, rather than simply asserting that we know it because we just do, but I recognize not everyone has my hangups.

The list is below the fold.

Saturday, November 6, 2010

Make Your Efforts Count

In a previous post, I mentioned that when it comes to voting, it is not true that every little bit counts, though it is true when it comes to (some) charitable giving.

That was not a throwaway line.  When people argue that their vote counts, sooner or later they slip into magical thinking and say something about how their actions, along with those of others, will do this or that wonderful thing.  When you give to charity, it is entirely possible that your individual actions, holding the actions of everyone else constant, make a difference.

It may be a small difference.  Saving one person's life by giving blood, for example, may not sound as lofty as "helping" (read: your action was purely symbolic) to elect the party that, in your biased assessment, is more likely to enact policies that will benefit everyone (read: everyone to whom you are politically sympathetic, not actually everyone -- after all, virtually all of politics is about choosing how to distribute benefits across society, which creates both winners and losers).  But at least you can say you made a difference  without any funny math or bad logic.

But, if you haven't caught all the qualifiers so far, I would argue that not all charitable giving is the same.  So, for the sake of argument if nothing else, let's assume you care about logic and evidence, not just warm and fuzzy feelings, and you want to be certain that your giving helped someone.

Then I would strongly encourage you to give well by giving to one of the charities recommended by GiveWell.  I do, and I'm a non-voting social misanthrope.

Reflections on the Field

I just finished grading comprehensive exams for the first time.  Almost as much fun as reviewing journal articles.  This also made me wonder exactly what we as a field have really learned about international conflict.  I don't just mean what gets cited a lot.  I mean what do we know.

I'll even more or less leave aside, for the sake of argument, areas where I am skeptical of ideas that are widely accepted by the field (such as the importance of costly signaling or the near-magical properties of democracy).  After all, I make no attempt to deny that I'm a quite a crank.

How many people in our field claim that we have real compelling answers to any of the following questions?

1) Are threats to retaliate with nuclear weapons after suffering a first strike credible?

2) If so, why would the probability of conventional conflict in jointly nuclear dyads be any different than in dyads where neither state possesses nuclear weapons?  Or is it?  What then prevented the United States and the Soviet Union from fighting a conventional war directly against one another rather than relying on proxy wars?

3) How does the systemic distribution of capabilities influence the probability of major power war?  Is the answer to this question different than at the dyadic level?  Why?

4) How did we reach a moment in history where Germany is encouraging deeper military ties between the UK and France?

5) How much of the effects we attribute to political institutions, economic factors, culture etc are genuinely due to those factors themselves, given that so much variation along most of these dimensions can be predicted quite well based on geography?

Finding correlations in observational data, especially when you have hundreds of thousands of dyad-years to analyze, is easy.  Establishing that the relationships are causal is a whole different story.

Don't get me wrong.  I think there's a great deal of value in exposing flaws in putative explanations.  Even if the greatest accomplishment of our field is to whittle down the number of claims worth taking seriously, I think we should be proud of that.  But I honestly wonder if there's a real case to be made that we've in fact done more than that.

US and India

Great NYTimes story on competing pressures within the Obama administration for how to engage with India.

In a nutshell, there are two competing concerns.  On the one hand, both short and long term economic pressures encourage closer ties, as does the desire to balance against China (see earlier post).

Thursday, November 4, 2010

The (Ir)Relevance of the U.S. Midterms for the Israeli-Palestinian Conflict

The Economist, usually a cut above what passes for journalism these days, has this strange story that isn't a story about how the midterm elections in the U.S. will impact the peace process.  Most of the story is a summary of where things stood, which hasn't changed this week, with only a brief, vague attempt on connecting the two.  They try, nonetheless, to strike an upbeat note at the same time that they tell us that the prospects for peace just got worse.  Slow news week?
Negotiations between two unitary actors are tough enough.  They may break down due to the incentive each side has to claim they won't settle for terms that they probably would settle for.  They may break down when one side expects to be more powerful tomorrow than they are today, and the declining state would rather use war as a means of preventing that shift in power. 

As Wolford and Rider demonstrate, things get even worse when one side can't control its own ability to uphold the terms of an agreement. 

An agreement with Abbas and the PLO does not guarantee the Israelis peace, not least because of Hamas.  Promises from the Israeli government can't guarantee the Palestinians that Israeli settlers won't keep taking over land that nominally was given to the Palestinians.  The broad contours of a settlement that each side's leadership would be glad to accept have been known for a long time, with some details to be worked about.  But agreeing to those terms would be pointless, since non-state actors on both sides would defect from the agreement.

Yes, it looks like the Obama administration played some role in persuading the Israelis to place a temporary freeze on settlements.  But it's quite unclear that Netanyahu could credibly commit Israel to abiding by an agreement that the PLO would accept, to say nothing of Hamas. That was true before Tuesday, and it's true today.

If the midterm elections had any implications for U.S. foreign policy, it's with respect to trade agreements (but don't expect much on that front), nuclear agreements with Russia, and climate change.  But most of what the international community might have hoped for would have been held up in the Senate even if the Dems had held on to the House, so even there, I don't see any real story.

I don't want to give the impression that the midterms were of no consequence at all.  For U.S. domestic politics, the range of policies that could be enacted definitely changed.  But one of the greatest impacts of Congress on foreign policy is through rhetoric as much as it is roll calls, so it's not obvious to me just how much it matters for foreign policy if the Republicans hold the majority in one chamber.

Wednesday, November 3, 2010

Negotiated Settlement with the Taliban?

Last week, the Economist ran a great story on the recent push for negotiations between Karzai's regime and the Tablian.

There are a few really good quotes.

Tuesday, November 2, 2010

Obligatory Election Day Post

I get irritated when people argue about whether it's "rational" to vote.  A rational choice is one that follows logically from a preference ordering, where we require that ordering to be complete, reflexive, and acyclical.  Depending upon your preferences, whose substantive content tells me nothing about whether you are rational, provided they exist and do not cycle, either voting or not voting can easily be rational.

Personally, it seems to me that our best explanation for which individuals do vote and which do not has more than a little to do with the fact that some people are likely to be given greater guilt trips than others if they don't vote.  Social pressure appears to be a huge factor.  That's not entirely dissimilar from the "civic duty" argument, but it's more useful in my mind because we can objectively observe how much pressure people face to vote.  But if you prefer the positive normative spin that implicitly suggests voters are responsible people, good for you.

Sure, I might wonder why it is that the desire to honor people who died for their rights hasn't ever motivated *anyone* I've ever met to attend religious services for faiths in which they do not believe regularly every two to four years so that they can also honor the sacrifices of those who left their homelands to come to this country in pursuit of freedom of religion, many of whom died before they arrived, nor to attend political rallies expressing views they do not share, simply to express their gratitude for the freedom of assembly.  I also find it curious that the logic of "what if everyone else thought that way",  does not seem, for most people who invoke such claims with respect to voting, to have any bearing in their minds on the many self interested behaviors in which they engage in everyday life.  I might even point out that while "every little bit makes a difference" is literally true when it comes to giving to some charities, given the abhorrently large number of children dying each year from *preventable* diseases, a fact to which most voters who shun non-voters are blissfully ignorant though their individual actions could in fact save a life, it is not literally true with an election.  Winning is predicated only upon having more votes than the other candidate here in the U.S., not the margin by which a candidate's vote total exceeds that of her/his competitor.

But I certainly won't refer to the decision to vote as "irrational".

Monday, November 1, 2010

Bogus science

Great story on how little of what we think we know about medicine can be backed up.  Drawing causal inferences is always hard when the behavior of human beings is involved, whether it's in the political realm or not.

China's Rise and Balancing

Realists the world over are presumably dancing in joy over this news.  Or they would be, if they weren't all busy pretending a theory that assumes states are unitary actors has important things to say about counterinsurgency, the only topic scholars of international conflict seem to be interested in anymore.

Perhaps that's for the best.  It's not obvious that the core assumptions of realism in fact predict balancing behavior, as Powell (1999) noted.  Nor is it obvious that balanced power promotes peace, according to many recent studies.

The Logic of Suicide Terrorism

In an influential study, Robert Pape argued that suicide terrorism only occurs when democratic states occupy foreign territories populated by adherents of a different religion.