The BP oil well is plugged.  Maybe we should consider ourselves lucky that more didn’t spill, that much of what did spill was dissipated by chemical dispersants, and that the oil didn’t, as some had feared, make its way up the Atlantic coast, devastating that coast and those coastal waters the way it devastated the Gulf coast.  Or maybe real luck would have been no oil-well blowout at all.

And maybe we should stop depending on luck to spare us from the consequences of making bets we can’t afford to lose.

Some years ago, I wrote and produced for the Smithsonian Institution an hour-long documentary on first ladies, based on the Smithsonian’s collection of first ladies’ gowns.  After considerable effort, I got an afternoon interview with Nancy Reagan, then the immediate past first lady, in Los Angeles.  To take advantage of the time difference and keep my travel budget down, I booked a morning flight from Washington to LA for myself, the associate producer, and the shooter—the cinematographer, a real talent.  The shooter lived near the associate producer, so he volunteered to pick her up.  On arriving at Dulles Airport, he dropped her at the terminal, set out for the economy parking lot—and missed the plane.

That was the day I learned what Airfones* were good for.  Between us, the associate producer and I secured the services of an LA shooter, and the interview came off without a hitch.

Back in Washington, the shooter and I discussed what had happened.  “It used to be,” he said, “that if I needed something from the drugstore and found a meter in front, I wouldn’t put a coin in.  I felt lucky, and I figured I’d be out before the parking police came.  For years it worked,” he said.  “But lately,” he said, “my luck seems to have run out and I’ve started getting tickets.”

That’s the kind of luck the U.S. has been having.  For years we’ve taken risks and gotten away with them.  Good luck, combined with our desire for the good things that can be purchased with the proceeds of successful wagers, has emboldened us to make and tolerate enormous bets whose success or failure hinge on events we cannot control or, often, even understand, and to risk consequences whose magnitude we shrink from calculating.

The Gulf Coast oil spill is a case in point.  The ecological viability of the Gulf of Mexico and, potentially, the Atlantic coast, and the economies of the Gulf and Atlantic seaboard states, were made contingent on the integrity of an intricately articulated network which depended on a single pipe connecting the well to an oil tanker.  When the connection failed, the entire network failed, with catastrophic consequences that would have been obvious if we could have brought ourselves to look at them.

So, too, with the financial crisis.  Large sectors of the economy, and the health of the entire economy, were placed at risk by an intricate web of financial relationships that connected home-buyers, mortgage lenders, institutional and individual investors, pension funds and, through them, millions of employers and employees.  And all of it, all the millions of financial synapses, depended on the integrity of a single node: home values.  Money was loaned, to sub-prime and prime borrowers alike, on the assumption that whatever might befall the home-buyer, the lender could always repossess the house, sell it, and recover the amount that had been lent and probably—given that home values had increased for decades—more.  On the basis of that assumption, lenders extended credit to home buyers, then sold the mortgages to investors who bought them with borrowed money and then sold them to someone else who paid with borrowed money.

And it all worked, earning investors and speculators millions and billions—it worked until it stopped working, until the one factor on which the entire network depended, home values, failed.

For the first time in decades, home values started dropping.  The sub-prime home buyers, the ones who had bought at artificially low rates, figuring that if worst came to worst, they could always sell the house to pay off the loan, couldn’t sell for enough to repay their loans and defaulted.  The mortgage-holder got the house but he, too, couldn’t sell it for enough to pay off the loan he had taken to buy the mortgage, and defaulted on his loan.  A tidal wave of defaults rippled through the financial sector and then through the economy, destroying jobs and displacing millions of families from their homes.

As in the oil spill, we bet the ranch on the strength of a network of unimaginable complexity, but dependent on a single factor.  There hadn’t been a major oil spill for decades—until there was.  Home prices hadn’t dropped in decades—until they did.

What’s our problem?  Are we unable to look a gift horse—any gift horse—in the mouth?  Are we, as a nation, too afraid of what we might see?  Or are we eternal optimists who continue to believe in the possibility of perpetual-motion machines?

Whatever the reason, we continue to bet the farm, even after losing numerous farms in previous wagers, another opportunity to risk disaster is looming on the horizon—or, more precisely, in the clouds.

Have you heard about cloud computing?  It’s the coming thing.  Instead of working off software on your laptop, desktop or server, you can work off programs on the internet—in the cloud.  And instead of keeping your documents and files near at hand on your hard drive or servers, you can keep them in the cloud, too.  Why use up storage space when space is unlimited in the cloud?  And if the cloud loses or corrupts the programs and files it is keeping for you, or if your connection to the cloud goes on the fritz, and you can’t run your programs or work with your documents, if this unavailability goes on for hours or days—well, how likely are any of those things?

Pretty damn likely, actually.  So likely that it happened just about a year ago.  The Washington Post headlined the October, 2009, story, “Sidekick Users See Their Data Vanish Into a Cloud”:

A server meltdown over the weekend wiped out the master copies of personal data — including address books, calendars, to-do lists and photos — accumulated by users of T-Mobile’s formerly popular Sidekick smartphone.  This computing calamity allows Sidekick owners only a faint hope of backing up the information currently on their devices, and none of recovering anything they’d trusted to online storage.

Microsoft, which stored T-Mobile’s data, was later able to recover it.  Microsoft said no data had actually been lost, that the incident was better described as an “outage.”  Nor was the Sidekick “outage” the only cloud to cast a shadow over cloud computing.  A Network World article listed seven other service outages in the space of a little more than a year.  The companies affected were big: Amazon, Gmail, Twitter, PayPal, Hotmail and Windows Azure, Microsoft’s cloud-computing network.

The outages were blamed on equipment failures, human error, and hackers, and most were restored within hours.  But equipment failures and human error aren’t going away. Hackers will get better.  If Google, Microsoft and Amazon can’t parry their thrusts, why should we assume that more recent, less experienced and more tightly-budgeted cloud-computing service providers will enjoy greater success?  And as cloud-computing spreads, it doesn’t take much imagination to foresee outages lasting days and weeks, with out-of-control servers gushing data like BP deep-sea oil wells.

The problem is not technology that’s not ready for prime-time.  Cloud computing is new, but oil drilling and trading in financial derivatives are mature businesses.  The problem is that all technology, all systems, will occasionally —and unpredictably—fail if they are operated long enough.  Toasters fail.  Lawnmowers fail.

Technological and other systems taken to ever higher levels of development will fail sooner and more often.  And technology and systems pushed to their highest capacity will fail more often still.

So what’s the answer?  Stop taking technology to higher levels?  Stop creating novel financial instruments? Stop taking risks?  Of course not.

But let’s follow the examples of successful risk takers.  Let’s compartmentalize risk: ocean liners do it, building in multiple water-tight compartments so that one or two or three leaks don’t sink a ship.  So do the intelligence services, restricting secrets to those who need to know, so that one compromised agent, one leak, doesn’t swamp the agency.

It’s the principle that a Depression-chastened federal government enacted in the Glass-Steagall Act, which prohibited federally-insured banks from becoming investment bankers, using federal guarantees as ballast against excessive risk.  But it’s a concept that a benighted Republican Congress persuaded a bewitched President Clinton that we no longer needed in the ‘90s, and that a lobbying-addled Congress watered down when it passed financial reforms earlier this year.

Let’s also make sure that those whose actions place the security of others at risk, others who have not agreed to the wager, do so under the adult supervision of those who may not always be right but who act in the name and interest of the larger community, not those who seek private profit.  We have a word for such adult supervision: we call it the government.

And let’s charge those who act for the community with understanding and enforcing something that every successful gambler knows but that some business leaders seem not to have learned: Faced with stakes you can’t afford to lose, choose a safer bet.


  1. Jay Hubelbank
    September 25th, 2010 | 10:21 am

    Cloud computing was formerly known as ASP (Application Service Provider). I am not a believer and we sell our software for ‘in-house’ installation. I am, however, a believer (and victim) of ‘Murphy’s Law’ and the ‘Law of Unintended Consequences’, both of which are prime forces for Cloud Computing to contend with.

    As we approach the 100th birthday of the Titanic we must remember the lessons of arrogance, whether it be Cloud Computing, oil drilling or economics.

    Jay’s Law, “any system that can fail, WILL”.


  2. Mary
    September 25th, 2010 | 1:36 pm

    Ah but there’s the rub. From whence cometh the “adult supervision” you mention? Our technologically savvy younger citizens seem to put little to no stock in anything anyone over 40 has to say. And it seems it was ever thus.

    As always, excellent blog.

  3. Bill Stott
    September 25th, 2010 | 2:57 pm

    Excellent commentary. How is it that the children and grandchildren of the Great Depression have lost so many of its lessons?

  4. Ira H. Klugerman
    September 26th, 2010 | 4:28 pm

    I’m presently working on a story for a Video program on Banking. The story begins with Glass-Steagall and ends with the banking crisis of 2009/10. Trouble is that my economic advisors tell me the problem was not with the repeal of Glass-Steagall, but with the failure of the regulators to regulate. Yet it is clear that with the Glass-Steagall repeal banking became so complex and banks sought so many ways to circumvent the existing regulation that the regulators found it difficult to watch over the banks. This does not excuse the regulators. I’m sure that the word from on high came down that the free market would regulate itself.

    Remember Murphy’s Law. If anything can go wrong it will.

  5. Mitch Hubelbank
    September 29th, 2010 | 7:38 am

    The roots of this concept go back to the early 80’s and a concept and an mebryonic industry called “disaster recovery”. In its infancy, the government gave it a shot in the arm, by passing legistlation through the Fed that required Financial Institutions to have disaster recovery plans. Note: (current) profit was subrogated to effecting a close to 100% availability. These systems were expensive,. The Fed didn’t care. As the system matured, and networks could transport more, it went from physically transporting data to send it over the “net”. 20+ years past, those risk adverse executives were replaced by a new generation of managers and compensation plans, focussed on profits and leverage. Somebody got the brainstorm to stop keeping the data in two places, ergo “data in the cloud”. “Save the cost, make more money, if it goes down, it’s their responsibility, not mine”.
    Note: disaster recovery is like insurance; you pay for it and never want to use it.
    Today’s business environment is not about discipline, long term view and viability. Compernsation drives behavior and as long as the pay is driven by the wrong behaviors, we will continue to see the “risk/reward ratio” be completely out of wack.
    The real estate melt down is indicative of our pervasive this Risk/Reward ratio is completely out of balance.
    And as a parent, I am incredulous to how limited risk assessment skills of upcoming generations.

Leave a reply