Refiners do not keep a large permanent maintenance staff. The bulk of their maintenance crews are hired ad hoc and as needed in the spring and autumn maintenance seasons. Due to the volatility of employment in the ''oil patch'', large numbers of temp maintenance workers have left the industry.
Refiners therefore are in the position of fighting each other for crews. If a wildcard such as Katrina comes along, it makes this problem exponentially worse, because everybody wants the workers all at the same time.
Additionally, there is, as in practically every industry, a high degree of variance from company to company in the quality of their maintenance of capital equipment. The recent AK pipe leak? BPAmoco couldn't be bothered to run the pigs. Texas City still mostly down 18 months and counting since the big explosion? BPAmoco, again (although, to be fair, they've had problems getting crews). The Whiting IN outage? You got it, BPAmoco yet again. See a pattern yet? BP, even well prior to the Amoco purchase/merger, has for decades had indifferent-to-rotten maintenance. They just got away with it for a long time, but no more.
I won't even bother elabourating on the other principal cause of price spikes, to wit, how both market and government uncertainty increases price.
I will second that. I have a friend who is an engineer at the BP Texas City refinery and he has complained for years about the lackadaisical management attitude towards maintenance and equipment reliability. They prefer to skimp on maintenance and then place blame on the engineering staff.
On the other hand take Exxon-Mobil. Before the merger, both had top-notch staffs at all refineries, especially Exxon. They never skimped on maintenance and bought top quality equipment. Several Exxon engineers were marketed by the company to teach equipment reliability to the rest of the industry and even other capital intensive industries, like electric utilities. They are the best are it shows as theri refineries rarely have big problems.