Wars happen for essentially only one reason: The legitimacy of a particular regime at a particular time is in doubt.
People do not fight wars because taxes are too high, or they disagree with the way money is allocated in the national budget, or for any other of the multitude of public policy sins that a government can commit. Public policy issues are relevant to which contestants in a war receive support of various kinds while it is waged, and to crafting solutions that will bring legitimacy to a new regime after a war, but are for the most part unimportant in cause wars to arise.
This fact alone illuminates a variety of important empirical facts regarding war.
1) Civil wars are far more common than international wars. This is because a simply succession dispute can create a civil war, while an international war requires something that causes people to re-evaluate the nature of the political community a regime governs. International wars generally involve either a border dispute, or a claim that nationality or some other principle overrides the status quo boundaries of nations.
2) Some of the worst regimes in the world face no serious risk of civil war. North Korea and Saudi Arabia, for example, are almost entirely undemocratic and in their respecive ways, treat their people horribly, and yet, neither is at any serious risk of civil war so long as there are orderly and nearly universally accepted handovers of power within the respective regimes from one individual leaders to the next.
3) We have a civil war in Iraq now because the U.S. led invasion disrupted the continuity of the Baathist regime irrevocably, and because a boycott of the elections for the constitution and the distrust of the military force that implemented the process has prevented the democratic regime put in place by Coalition forces from being viewed as legitimate. Sunni Muslims correctly saw the constitution as a way to deprive them of power permanently, by aligning the other 80% of the nation, against their 20% of the nation, so they had no reason to buy into the arrangement. Yet, preventing a civil war requires much more than 80% of the population to buy into the notion that a particular regime is the legitimate government of a country.
Usually, legitimacy arises from an orderly succession of leaders within a particular regime. For legitimacy purposes, it really doesn't matter what the method of succession happens to be, so long as there is near universal agreement regarding who the successor will be when the process is completed. Heredity, a constitutional succession of leaders in the absence of an election or between elections, or an election with a clear winner, are among the possible ways for this to happen.
Even the establishment of a new regime, like many of the successions of regimes in France and Germany, will not create a civil war, if the transition is cloaked in the appearance of an orderly and lawful succession from one regime to another.
The succession theory of legitimacy is, of course, incomplete. It fails to discuss either how a legitimate regime is established from scratch (suffice it to say for the moment that it is exceedingly difficult, and is basically a form of nation building), and it also fails to discuss how legitimacy can be undermined apart from confusion in the succession process.
Nation building is a very slow thing, unless it happens with the consent of all predecessor regimes. For example, 142 years later, it isn't uncommon to see only grudging acceptance of the legitimacy of the federal government in the former Confederate States of America. There is enough acceptance of that legitimacy that civil war isn't a viable possibility any longer there, but that wasn't true for, at least, a decade or two after the Civil War.
Similarly, the "Indian Wars" in the United States, which involved sovereignty disputes between Native American regimes and the United States of America lasted almost a century, despite the fact that the United States vastly overwhelmed the Native American peoples in numbers and resources, because civil wars persist until there is essentially no organized group of people who deny the regime's de facto legitimacy. Even then, the Indian Wars ended only when the Native American population had reached a historic low, and all the viable Indian tribes received recognition and some sovereignty over some territory. Even a small percentage of the total population can prevent a regime from gaining legitimacy is that percentage of the population is discrete and has some manner of organization.
Ironically, however, an organized insurgency is often essential to ending a civil war, as it gives the proto-regime attempting to assert suppremacy someone to come to terms with who will be accepted by the insurgents as their legitimate representative in such matters. The near absence of an insurgency in post-World War II Japan, for example, owe a great deal to the fact that the Emperor of Japan was retained and had sufficient legitimacy to agree to a change of Japanese regime on behalf of the Japanese people. Similarly, the legitimacy that the monarch of Thailand lent to a political deal to end a civil war there was crucial to the deal's success. New research of the incredible expansion of the Islamic empire in its first couple hundred years, shows that a strategy of co-opting, rather than outright fighting by brute force, local elites, was key to its success. A perceived treat that an acquiring regime could obliterate a nation entirely if it does not accept a conquering regime, is more important than the actual execution of that threat. Of course, once a legitimate succession of leadership is established, the original reasons that legitimacy was achieved hardly matter.
This also explains why it is not uncommon for heinous crimes committed during a war, and grave disloyalty, for which insurgent leaders face death sentences, are so often commuted in exchange for public loyalty oaths to the new regime, which seem like mere pieces of paper. But, loyalty oaths from insurgent leaders to a new regime dramatically and symbolically undermine any shred of legitimacy than an insurgent regime may have had, and thereby prevents further war. Hence, those seemingly empty and symbolic acts can actually save countless lives by preventing a war from continuing longer with predictable casualties. That benefit is often worth more than justice for the individual insurgent, who usually wouldn't have behaved in a deviant or criminal manner in the absence of a legitimacy dispute.
Religion has figured prominently in the history of war, largely because, prior to the widespread adoption of democratic norms, national leaders attributed their legitimacy to a divine right of kings to rule, and religous disputes impacted the legitimacy of the religious stamp of approval that validated their divine right to rule. Religion, like nationalism, is also a basis for casting aside existing arrangements regarding national definitions.
The way legitimacy is undermined is also a difficult part of the theory. A mere percenption that a leader is incompetent, or pursues bad policies is not enough, at least until that leader's legitimacy has been seriously called into question. But, orderly succession, while sufficient to excuse a multitude of sins, is not the exclusive reason that a regime can lose its legitimacy. Suffice it to say that in a constitutional system of government, disrepect for another institutions perogatives in that system undermines the very constitutional system upon which you rely for you own authority to act.