Wednesday, June 9, 2010

Bad Math Reaches Beyond Financial Debacles



In prior blog posts, specifically on September 8, 2009, September 11, 2009 and in sporadic comments throughout, I highlighted the atrocious risk management based on poor mathematical modeling that precipitated the financial meltdown in 2008. Going back ten years earlier, Long Term Capital Management made similar miscalculations despite the fact that two Nobel Prize winners in economics, Myron Scholes and Robert Merton worked there.

Now it seems that British Petroleum in particular and perhaps the oil industry in general are also using ludicrous algorithms for predicting risk. For example, Tony Hayward, the beleaguered CEO has made pronouncements ranging from stupid/insensitive all the way to downright pernicious.

"I would like to get my life back" is a really dumb and selfish thing to have said...hey Tony I've got news for you, 300+ million people in America would like their lives back too. "We are 60-70% confident that the top kill procedure will be effective." In hindsight, it doesn't seem like that prediction was all too accurate...top kill efforts failed repeatedly such that there is no possible way that they ever had anywhere near a 60% chance to succeed. Before cutting the leaking pipe, BP told us that this action might increase the outpouring oil volume by 20%. Now, some scientists are worried that after the cut has been made, the exiting oil could be many times that 20% estimate.

What really frightens me though is something that Hayward let slip this past Sunday and I am a little dismayed that I have not seen anyone in the media call him on it. In a press release Hayward was quoted as noting that the probability against this accident happening was "100,000 to 1 or even 1,000,000 to 1." At face value, this is an utterly absurd statement. It would be like handicapping a horse as being somewhere between a 10-1 and a 100-1 shot!? Probabilities which range a full order of magnitude are inane and useless.

Moving to the next level, even if the odds against the spill happening were indeed one million to one, the risk-adjusted view, which ostensibly should be used for making operational decisions, is far greater (perhaps a thousand times more likely) than that. One million to one is a better chance than someone has to actually win the lottery and yet, lotteries are won every week...get the point?

Finally, how in the hell did anyone manage to come up with this probability appraisal? There have been in aggregate, fewer than 6,000 deep water wells drilled on the planet. Throughout deep water drilling history there have certainly been numerous accidents; the Gulf spill is cataclysmic, but there have been other serious mishaps. One would think that prudent risk management would use prior drilling history to rate the potential danger before knocking open a hole in the bottom of an ocean.

For once I agree with President Obama...Tony Hayward should be terminated as BP's CEO but I have reasons different than Obama's. The President thinks that BP caused the oil spill but I rather believe that the oil industry in cahoots with the government was responsible for it. Shell, Exxon et al are I'm sure on their knees with their Rosary beads thanking the Lord that something like this didn't happen to them...this time (remember the Valdez?) Rather, I think Hayward should be canned because he makes stupid, inaccurate and dangerous public statements. He also allowed his company to behave in a grossly negligent manner, without properly considering the dire consequences that ultimately arose from this carelessness.

It may sound like I abhor oil companies, deep water drilling or even Wall Street, but that assessment is entirely incorrect. I am not even arguing for risk aversion...all progress requires taking a chance on something. But I am advocating rigorous risk awareness/abatement and the mathematical tools to perform this analysis already exist, albeit, outside the realm of the conventional statistics that are most often used. For example, if Exxon, the industry and the government had properly come to grips with the fact that tankers can run into something that punctures them, then maybe double-hulled vessels would have been more prominent at the time the Valdez incident occurred...this wouldn't have averted the spill but it would have mitigated the damage. If BP had realized that its drilling strategy and cost-cutting mentality (see the NY Times link below) were not only bet-the-company but also, bet-the-Gulf of Mexico propositions, then perhaps it would have bored relief wells (assuming that this is the ultimate solution...we'll see if that's true) simultaneously when poking the production hole. I'm not suggesting that I know the specific risk-reduction remedies, but I AM saying that anyone who is in a position to create catastrophes had better completely recognize their predicament. While there are always possibilities for error, in situations like this, trivial considerations such as short-term profits, cost-savings and the like must be dwarfed by overarching, preplanned, redundant remedial resources which are deployed long before any accident can possibly happen...in other words, before drilling begins.

Modifying corporate and governmental behavior patterns with respect to potentially gigantic human, environmental and financial liabilities like the Gulf spill will require not only new, more appropriate risk measurement methodologies but also, a refreshed risk-reward mindset. If projects can't sustain the costs associated with extensive, appropriate prophylactic elements then they should be abandoned. I have no idea how much oil is in that particular reservoir in the Gulf, but whatever it is, it isn't worth the trouble and tragedy that this spill has caused.

For a more detailed description about BP and their missteps, please see the NY Times article below:

http://www.nytimes.com/2010/06/06/magazine/06fob-wwln-t.html?emc=eta1

To read about oil projected to climb up the east coast and beyond, you can check out this piece from the WSJ:

http://online.wsj.com/article/SB10001424052748703340904575284950638479786.html?mod=WSJ_WSJ_US_News_

No comments:

Post a Comment