>Except for the stock market crashes of 1929 and 1987, the dot-com bubble, the housing bouble followed a crash, the tulip mania, ... Markets are not always rational.

Crashes are expected in the market theory. Crashes are actually healthy outcomes that get rid of mal-investment. There could have been much less severe crashes in the situations you described if there has been less government involvement (there is ample evidence of the government agencies and the FED being involved in the dot-com bubble and the housing bubble).