There is more out there on the subject of war. Economics blogger and libertarian Tyler Cowen seems to think that war (or at least the threat of it) is good for growth. There are two problems with this. First of all, across the post-WWII era, much of the evidence seems to support the opposite hypothesis: countries that are less prone to conflict grow faster. Second, growth is likely not the only variable worth exploring.
On the first issue, one might argue that countries that have been prone to conflict over the last 70 years have mostly been those countries that are also prone to slow growth, for institutional reasons, and that would be fair. So, let's not call in the jury yet on that, although the likes of Dani Rodrik would likely share my skepticism.
On the second issue, even if we found there to be spurts of growth following conflict, we would be hard pressed to show that those spurts would really leave us better off than the alternative scenario in which countries like those in the OECD had done everything else the same, except that they had fought more. Many countries are bound to have growth spurts after conflict simply because of the destruction of capital wrought by conflict. In other words, sometimes growth speeds up precisely because things have exogenously been made worse. This is merely an application of neoclassical growth theory.
Digging deeper into Dr. Cowen's argument, there is a hint that the threat of war might, historically, provided an incentive for investment in public goods, including infrastructure and technology. The mechanism through which this operates in Cowen's model, it seems, is by increasing competition between the incumbent regime and external and internal contestants. To his credit, this is indeed a hypothesis of that is well-established in the literature on institutions, for example by Mancur Olson (1993), but also North, Wallis, and Weingast (2006).
The problem is that in these institutional models, the threat of war as a constraint on power is mostly used in primitive nation-states that rely on fairly closed and autocratic forms of government that are more apt to be extractive. Also, in these models, the main priority of the dictator is to provide security so he can remain in power. This is done in three ways: (1) by buying off former warlords and elites; (2) by eliminating elites those who can't or won't be bought; and (3) by providing public goods.
More open, inclusive, and democratic forms of government can usually achieve the constraints on power necessary to provide public goods effectively through the electoral process, to the extent that the process is not too tainted by asymmetric information and special interest politics. Note that this does not mean that taxes will be lower in a democracy, nor that redistribution will be less. It just means that policies will better reflect the preferences of society at large, and in particular those of the average (median) voter. In fact, a democratic government may have a greater capacity to tax precisely because it uses public funds to provide public goods like infrastructure and technology.
So, it seems myopic at best to think that conflict (or the threat of conflict) should be considered a "good thing." At worst, it is dead wrong and a destructive way of thinking.