An influential paper released yesterday by the Bank of England argues that the Basel III capital requirements aren’t nearly tough enough. It recommends rules that double the base-case Basel III capital requirements – bringing risk-based capital to about fourteen percent and the leverage percentage up to at least six. The extensive analytics in the paper persuasively argue that these ratios are the minimum needed to prevent another systemic banking crisis. Maybe so, but capital would come to be the be-all and – more likely – the end-all of banking because banks would be so swaddled in capital cushions that they couldn’t make a cost-effective loan, let alone a reasonable profit. Worse, banks – if there still are any – would be at even more risk than before.

The Bank of England analysts aren’t the only ones who want higher capital. One of the heads of Berlin’s prestigious Max Planck Institute recommends leverage at about thirty percent, and a recent Stanford University study goes for fifty. However, capital isn’t the belt, suspenders and safety net rolled into one these analysts hope. Relying on capital as the catch-all, single component of prudential regulation is like relying on a car’s brakes – they are critical to a well-functioning car and had better work under stress. But, safe driving requires lots of other active and passive safety systems in cars, rules of the road, driver-license requirements, and highway-safety standards along with four good brakes.

Why? Because engineers know that redundant systems are essential to failure protection. Further, engineers demand that complex mechanism have different backstops for different risks that work together to create a functioning, safe machine. Build any single component to fail-safe – often an idle dream even when attempted – and you not only don’t do much for safety, but you also undermine functional integrity.

So, back to the Basel critique. Would doubling capital as recommended actually prevent systemic failure? First, it wouldn’t because somewhere, somehow there will be a big bank that – gasp – won’t meet the rules. It might look like it does – think of all the “well-capitalized” banks that failed over the past couple of years – but the capital rules can’t be so perfectly engineered that all opportunities for arbitrage are ended. As with teenagers behind the wheel, bankers will find a way to take new risks for the thrill of it all – in their case, an extra couple of basis points or two on ROE.

But, assume we’re wrong and the capital rules are invulnerable. Does that make banking systems more robust if they are the single, high criterion imposed by regulators? Sadly, no. To switch our metaphors, this is like expecting to live forever with a perfect cardiovascular system and then getting hit by a bus. Banks of course fail for reasons having nothing to do with capital even when they really are well capitalized in fact, not just according to the rulebook.

What are some of these reasons? Fraud is a surprisingly big one – think Lehman 105s for example and, even if these weren’t fraudulent under the law they were, at the least damn misleading to investors. Liquidity risk? For sure a cause of failure and not just for individual banks, but also global financial systems. Concentration risk – same here and look at Fannie and Freddie as sad proof. Externalities to the banking system of course also count – credit rating agencies that done us wrong – and catastrophic risk such as 9/11.

But, discount all this. Assume capital really is the sole salvation of global finance. Here, one hits another analytical road-block: non-banks. Regulators can try to sweep as many companies as they can into the regulatory-capital regime, but the tougher these rules, the more firms will just say no. “Shadow” banking – and then some – of course will result with an even more virulent threat of systemic failures that threaten macroeconomic growth.

Where does this leave one? Not, we hasten to say, with a lenient capital regime that relies on bank models attuned to the most optimistic scenarios under the least possible supervision. That would be like permitting cars to go on the road with rubber-bands instead of brakes. It might have been nice if the Financial Crisis Inquiry Commission had given us a better sense of what rules are really necessary to keep the U.S. financial system safe in the future, looking – as Congress requested – at finance in general, not just banks. But, that is not to be.

So, we’ll stick with our take-away from the financial crisis: the new regulatory regime needs to be a carefully-calibrated array of standards engineered with the goal of creating functioning banking organizations that meet market needs and investor expectations in a robust, prudent and efficient way over a reasonably foreseeable set of circumstances. Banks will still crash – that’s just the way it will have to be – so policy must also be in place to pick up the pieces, preferably without rewarding any evil-doers. That’s why we have ambulances and doctors as well as car-safety systems and that’s why cross-border resolution protocols are an urgent priority virtually unaddressed by commentators on global banking.

Is this perfect? Of course not. But, is there any perfect answer to safe and sound banking? Regulators may think they’ve found one in capital, but engineers know better. Systems are only as safe as their weakest component, and all the capital a bank holds doesn’t mean it isn’t at risk from something else somewhere else.