In this paper, we review the central claim of a growing literature: that is, that democratic
states rarely, if ever, wage war against and are very unlikely to engage in militarized disputes
with other democratic states. We first examine the analytic foundations of this claim. We
conclude that they are tenuous. Next, we examine the evidence.
We find that no statistically significant relationship exists between regime type and the
probability of war before World War I. We also find that the probability of disputes short of war
is significantly higher for democratic-democratic pairs than for other pairs of states in the pre-
1914 period. In both cases, our analysis shows that the hypothesized relationship prevails only
after World War H.
Because of the Cold War that ensued after 1945, our results suggest that the relationship
we observe between democracy and conflict is the product of common interests rather than of
common polities. An analysis of the relationship between regime type and the probability of
alliance formation lends support to this interpretation.

It is generally recognized that possible gains from cooperation are
not always realized because negotiating parties sometimes fail to reach
agreement. Such negotiation failures are ex post inefficient and have long
defied economist's attempts to explain them. In this paper we discuss
breakdowns in collective bargaining when disputes are resolved by
compulsory arbitration. We argue that the study of arbitration offers
insights into the nature of negotiation failures which may be of general
interest.