Civil wars are now the most common form of major armed conflict and in this paper we examine the economic and human cost of civil war during 1960-99. The adverse consequences of the war are suffered not by the combatants but by ordinary civilians who have typically no say in either whether the conflict is initiated or whether it is settled. Using a global data set we show that a civil war of five years reduces the average annual growth rate by more than two percent. After the fighting stops a peace dividend is by no means automatic, the economic recovery very much depends on whether the country is able to implement considerable policy reform. Our survey of the human costs of conflict shows that even long after the war stops people are killed or maimed, mainly due to the destruction of public health infrastructure and population displacements. The post-war number of fatalities and casualties occurring is about as high as the numbers incurred during the war. We also consider whether these terrible costs could be seen as a high but necessary price to pay for future improvements. Many rebel movements want to change their countries' political systems for the better. However, using data on economic policy, democracy and political freedom we find that civil wars change countries for the worse. On average countries have lower policy scores than prior to the war. Thus, although it may be possible to find some modern civil war that can be seen as paving the way for social progress, it is likely to be the exception. On average, civil wars during the past forty years have not brought about positive social change but left a terrible legacy of high economic and social costs. They have been development in reverse.