Democracy and War February 1, 1999 Eric Rasmusen ABSTRACT Modern democracies do not start wars as often as other kinds of regimes. Regression analysis is sometimes used to make this point. This may be misleading, tho. For one thing, the same methodology would show that Communist regimes do not start wars. For another, it may still be true that modern democracies are disproportionately a cause of wars, by presenting tempting targets for other regimes. Professor of Business Economics and Public Policy and Sanjay Subhedar Faculty Fellow, Indiana University, Kelley School of Business, BU 456, 1309 E 10th Street, Bloomington, Indiana, 47405-1701. Office: (812) 855-9219. Fax: 812-855-3354. Email: Erasmuse@indiana.edu; Erasmuse@Juno.com; Erasmusen@Yahoo.com (for attachments). Web: Php.indiana.edu/~erasmuse. I am sympathetic to the idea that modern republics are less likely to begin wars than other regimes, including, however, less likely than ancient republics. And this is a subject worthy of careful study. I do, however, think that it is inappropriate for regression analysis, due especially to the problem of our inability to control for important variables, or, to put the same problem a bit differently, lack of independence of the error terms. Thus, I read with interest the comments of R. J. Rummel and Ted Carpenter about Professor Rummel's book Power Kills: Democracy as a Method of Nonviolence (Transaction, 1997) in the Summer 1998 Independent Review. In this note, I'd like to make two points to illustrate the importance of careful case-by-case historical study. The first point is empirical, an example to illustrate the hazards of regression. The second is theoretical, a question to illustrate the need for carefully posing the question. The usual regression in this kind of study sees if republics or free countries begin fewer wars, and finds that they do. The variables need precise definition, but while this is difficult, it is not an insurmountable problem. Suppose we take the past 2,000 years as our sample period. We exclude minor border conflicts that do not result in full-fledged wars. We exclude civil wars, defined as conflicts between entities that were part of the same nation at the time or fewer than twenty years before, thus excluding the U.S. Civil War. We say that a country drawn into an already-existing war is not an aggressor, even if it is the one to declare war first, thus excluding the United States as an aggressor in World War I. We also say that simply aiding rebels in a civil war does not make a country an aggressor, thus excluding U.S. aid to the Contras in Nicaragua in the 1980's. I do not list these caveats sarcastically; I think they are all reasonable. But now let us change the dependent variable. Instead of testing whether free countries are warlike, let us test whether Communist countries are warlike. We would find that they are much less warlike even than free countries. First, for most of history, there were no Communist countries. So we will have lots of wars in our sample, making statistical significance easy, but very few started by Communists. Second, even in the 20th Century, by the reasonable criteria above, no war that I can think of at the moment was started by a Communist country. World War II? Russia waited some weeks before joining Germany in declaring war on Poland, which, in any case, was part of Russia till roughly 1921. The Winter War in Finland? Again, Finland and Russia were both the same country till 1921. The same goes for the Baltic Republics, which, in any case, gave up immediately in 1940 rather than be beaten in a war. The Korean and Vietnam Wars? More civil wars-- and China did not join the Korean War till midway thru. The Cuban Revolution? A civil war, tho with foreign aid from Russia. The same for Ethiopia, Angola, Cambodia, Laos, Greece, Spain, and so forth. The Chinese conquest of Tibet? A civil war, according to international law, which did not recognize Tibetan independence. The Sino-Indian War? The restraint of the Chinese made this just a border conflict, even tho they could have advanced deep into India had they wished to. The only exception I can think of is Afghanistan, which is a weak exception, since the Russians came in to restore a regime that was friendly to them and this is more like a foreign-sponsored coup followed by foreign aid in a civil war than like a real war. Thus, our regression would show that Communist regimes are the most peaceful in world history, despite their extremely heavy expenditures on weapons, their successful expansion by conquest, and their explicit desire to impose their system on the entire world by force. My second point is a theoretical one: we must be careful about the question asked. This is one reason the subject is worthy of continued scholarship. The usual regression, the one I just discussed, asks whether free countries are the aggressors in wars. That is a good question, but it is different from the question of whether free countries cause wars. If there were no free countries, would there be fewer wars? I think not. Modern republics have shown themselves to be involved in wars much more than modern dictatorships, and I suspect this is not accidental. Modern republics do not start wars by their aggression, but they start wars by being such temptingly non-aggressive targets. To take the biggest example: would Hitler have gotten anywhere had he been surrounded entirely by dictatorships? It seems not; Stalin's Russia tried desperately to organize his containment until, after repeated rebuffs from France, Poland, and Britain, he compromised with Hitler instead. If we asked what percentage of republics have been involved in wars compared to what percentage of other regimes, would we not find that republics are more often at war? Most of those republics are in Europe, which has been embroiled in two world wars, while most other regimes have been in the Third World, which has had very few wars. Even the few free regimes that do exist in the Third World, such as Israel, Lebanon, and India are not noted for their exemption from warfare. I would guess that free countries tend to create the conditions for war for three reasons. First, with less unified and longlasting political leadership, they are less adept at foreign affairs. Second, the agency problem of giving the leaders incentive to pursue the interest of the country as a whole is worse for foreign affairs, tho better for domestic ones. Domestically, elected officials are less able than dictators to steal from the public. In foreign affairs, however, the interests of a dictator in avoiding conquest is closely aligned with that of the country as a whole, and his freedom from monitoring is useful. An elected official, however, must deal with an electorate which does not understand foreign affairs and is dubious about military spending and military casualties, especially if they seems to be incurred on behalf of foreigners. The U.S. electorate is perenially hostile to foreign aid spending, which continues only because the elected officials realize that it has indirect benefits. Third, free countries may be irritants to unfree ones simply because they are different. Israel allows Arabs to vote, while Syria does not, which must be unpleasant for Syria's rulers to admit. South Korea and South Vietnam were examples to which North Koreans and North Vietnamese might aspire.