Would it have been a good idea to buy a subprime consumer lender in 2006? Not if you believe the received wisdom — that subprime caused the credit bubble and the ensuing collapse and crisis, from which none escaped unscathed.

If that's true, how come Santander managed to buy a subprime lender in 2006 — and then sell a fraction of it in 2011 at a valuation of more than six times its investment, booking a $1 billion profit?

Such an excellent achievement might not have required brilliant operating management, or even good luck. That's because the investment was in subprime auto lending, not mortgages.

Subprime auto is different. Most likely that's because there haven’t been any GSEs dominating it and promoting, for example, a social mandate from Heaven to make automobiles affordable for all — essential as cars may be. Hence subprime auto, like subprime and retail credit cards, came through the crisis just fine. Yes, GMAC required government support and then changed its name to Ally. But its auto lending was not subprime and its problems stemmed from an indefensible detour into 115% LTV and other futuristic mortgages, via its Res Cap subsidiary. (Diversification!)

Successful subprime card lenders such as GE Retail and HSBC balanced their subprime portfolios with substantial volumes of lower-risk card accounts. Deposits are a preferred source of funding for such portfolios—while other banks complain about excess deposits! Some tell me: "We'll do subprime. We just don't want to be seen doing subprime. The market hates it!"

The recent crisis was not even about subprime mortgages. It was about a plethora of creative mortgages, many from people with high credit scores — hence not subprime. What these mortgages had in common was the underwriting assumption that home prices would continue to rise — so that increased payments would never have to be made.

Meanwhile, we spawned a first generation of "enterprise risk executives" (greatest thing since the Y2K Problem, or maybe the Business Continuity Plan) who evidently couldn't imagine home prices going down. Somehow the EVPs of decision management, with their 500-variable models, missed this one.

It's not 500 recognized uncertainties that kill you. It's one, or a few, variables that you didn’t think were variable. To survive, you have to identify them soon enough. Subprime was not such a variable. Neither was mortgage underwriting or credit operations — disgustingly inadequate though they both became. Home prices did the trick.

For a few small banks, valuation of GSE preferred stock was the dynamite variable. Another example: Basel II's assignment of 0 risk to all sovereign debt denominated in a bank's own currency has been the killer for many foreign banks. So much for regulatory optimization, value at risk, formulaic complexity, and risk-weighted capital.

What have we learned from all this? How can you cope with what are imaginatively and misleadingly referred to as "black swan" events? Most of these events, when properly understood, should never have been assumed to be unthinkable. The future does not replay the past!

Write down all the assumptions that limit your perceived uncertainty about asset values. Challenge each vigorously. Make a "loss tree," analyzing possible causation chains of value destruction, top down. No numbers are needed at first — just insights, and the words to describe them. You will probably find more than one wobbly assumption or variable that can mean big trouble. Impossible to know in advance which one of them will strike the spark.

Next, reflect on whether Robert Rubin was right when he told Citi it needed to take more risk during the bubble. Then see what you can do to mitigate exposure to big losses — preferably not through the purchase of credit default swaps, which don't seem ever to have kept many people out of trouble. What will the ISDA mete out to those who failed to read the fine print, and so bought CDSs on Greek bonds? Sincere regrets?

GMAC wasn't the only one to discover that diversification doesn't infallibly reduce overall risk. Gold has lately been moving with the Euro, rather than against it. Assuming negative correlation magnifies exposures and encourages bets on inappropriate assets.

When I recently told a top risk officer of a major bank how he could reduce strategic loss exposure, he replied: "No, we can't stop doing that. We're making too much money at it." I rest my case.

Andrew Kahr is a principal in Credit Builders LLC, a financial product development company, and was the founding chief executive of First Deposit, later known as Providian. He can be reached at akahr@creditbuilders.us.com.