Now, of course, that has to be qualified.
Before I read any more "realists" writing about a Chinese-American cold (or hot) war in Asia, I want them to come up with an explanation of why the War of 1812 was the last Anglo-American war. Why not 54'40" or fight? The claim that "wars are almost never worth the costs, and yet states keep on fighting them" doesn't seem to apply to relationships between some states since 1850 or so.Generally, realists respond simply with "democracies don't fight one another, and we don't know why, but neo-liberals don't know why either". Most of the liberal arguments in favour of their position (that war is irrational, that people generally are rational, and will restrain their leaders from war if given a mechanism) doesn't make sense if you consider that democracies are perfectly willing to war against non-democracies, and that leaders do tend to reap substantial popularity and elites substantial wealth from the enterprise.
The "realist" school has absolutely no clue as to why there has been no Franco-British or Anglo-American war in 190 years, no Mexican-American war in 85, and no Franco-German war in 60. The language of administration in Strasbourg is simply not something that people are willing to kill or die for these days. Which kinds of states study war (against each other at least) no more? And why not? Those are, I think, the most interesting questions in the academic study of international affairs.
(DeLong contends that warfare used to be less costly but now is more so. He cites 1850 as the date when it changed. That's erroneous- Japan profited tremendously from the Sino-Japanese war, the United States seemed to profit quite well from the first Iraq war, and there were certainly enough "dirty little wars" during the Cold War that somebody was profiting from them. Systemic warfare is useless because of nuclear weapons, but that's only one kind.)
It also runs into a significant and substantial problem, which is defining what, precisely, a "democracy" is. Define it too narrowly and you have too small a sample set to mean anything; define it too broadly and the rule is violated. More important is the passage of time: today's democracies may become tomorrow's authoritarian regimes, and the unpredictability of these changes forces democracies to arm themselves against possibilities.
There is a possible answer, though. Why don't democracies fight democracies? They have an army, and there are benefits to reap from warfare. I believe the answer is identity- citizens of democracies derive a measure of superiority from being such, and that shared identity helps mitigate conflicts between them, as long as there are non-democracies around. A democracy need not engage in war against another democracy if the benefits of war can be gained from attacking an authoritarian regime, which can be used to reaffirm the superiority of the citizenry through the simple rhetoric of "liberation". This is the core critique of the newest Gulf War- that it was used to make America feel better about itself; and solves the question. Democracies fight non-democracies.
The remaining problem, though, is this: what happens when there are no non-democracies left?