Imagine you are running a $100B company. Your personal salary is 2% of the excess earnings (over and above the safe treasury rate) of your company. You secretly bet your company's fortune so you get an extra 1% of return on investment with 98% probability, but your company can lose everything with a 2% probability.
Any bookie can see these are terrible odds for the company since the "expectation" is (0.98 X 1) - (0.02 X 100) = -$1.02B. In other words your actions will cause your firm to lose $1.02B a year on average over a long period of time.
But if you are mainly concerned about about your own earnings during your 5 year tenure at the top, then making this bet makes perfect sense. There is an over 90% chance that your company gets that 1% for all five years, netting you $20M every year. If they are unaware of the chances you've taken, then your investors will attribute your "success" to your superior managerial capability. And if that calamity does occur wiping out your investors, you personally get to walk away paying nothing. You even keep your past earnings, go yachting and getting your life back, as BP's CEO Tony Hayward would say.
That in essence is why people can have strong incentives to take on tail risks defined as very unlikely but catastrophic events. Instances of such tail risk taking include:
a) The aforementioned BP spill, where cutting corners and ignoring safety imperatives can save oil companies hundreds of millions of dollars a year. While the other oil chiefs solemnly swear to the complete safety of their practices, they know the chance of any such false claims being exposed on their watch is very low. Just as it was for Tony Hayward who was unlucky enough to have lost the reverse lottery. But the the risk of something terrible happening is very high, when aggregated over all the operating companies and the tens of thousands of wells operating under loose regulations.
b) The financial meltdown led by the collapse of the sub prime mortgage loans market. The easy money architects like Alan Greenspan thought the risk was very low. Many lenders, traders and money managers (backed by their rocket scientist quant analysts) knew that a drop in real estate prices could be catastrophic to the derivatives market. They just figured that the bubble wouldn't burst in their short term trading horizon, and someone else down the line would take the fall. Or if they were too big to fail, that they could collect on the upside while a lot of the downside would be borne by taxpayers. They were right on many counts. Even Goldman Sachs which famously dodged the bullet would have done badly if the housing price collapse had started a year earlier, before they unwound their positions.
c) Hurricane Katrina and the damage to New Orleans. Generations of politicians and lawmakers avoided raising and strengthening New Orleans' barriers. These would have guarded against the very unlikely possibility (in their watch) of a Category 5 hurricane directly hitting the city. They instead could divert such resources for popular "pandering" projects that would win them accololades and political support, with no one the wiser about the risk that did not materialize. But a city's life should be measured in centuries (think of the Netherlands' dikes) and over that horizon the risk was very high.
d) Other as yet unrealized disasters like nuclear accidents (assuming Three Mile Island wasn't bad enough and a while back) or earthquakes where safety codes are not strong or enforced enough.
The common factor in all these instances is that the people taking the risks on average derive a huge benefit from doing so, even if this is severely detrimental to the affected populace. That's why leaving the regulation and policing to the private industry can be so harmful. These special interests can lobby fiercely, or use a portion of their expected benefits to bribe or buy support and intimidate opposition that wants tighter controls.
At "my" University of Chicago the majority academic view leaned heavily towards private enterprise and self regulating markets. It went way beyond the concept of "efficient markets" relating to stock, etc. prices that makes intuitive sense. Many of the arguments and reasoning I heard in support of this more extreme "private and unregulated is generally the best" view was not convincing to me. It generally cited historical correlations between free enterprise and economic prosperity. Many of these no longer hold as even Andy Grove pointed out on July 1 in BusinessWeek in a different context of job creation and the rise of China and other controlled economies.
Still, to its credit the University of Chicago does tolerate dissent and fosters diversity of opinion. Paul Krugman in his April 9, 2009 NYT column, pointed to Raghuram Rajan of this school presciently warning back in 2005 of the risk of a financial meltdown, absent adequate controls. Now we just need to have lawmakers and politicians step up to the plate and have the right government safeguards to watch our collective back - and tail.