Analyses of the origins of the financial crisis all too often start from an unquestioned premise: Basel II is to blame for a lot of what befell the global financial system. This point is critical, as it is driving much at both the bank regulators and on Capitol Hill. It’s wrong, though. Basel II is far from perfect, but it didn’t come into effect in the E.U. until 2007 and it was only effective to a limited degree for a small number of big U.S. firms before the crash.

What’s most wrong with Basel II is that Basel I lasted far too long.

Basel II was set in motion in the late 1990s because it was clear that financial markets had developed well beyond the simple structure reflected in the initial capital accord. Most important, banks figured out from Day One in 1988 — when Basel I began — that the treatment of off-balance-sheet assets was a wide-open invitation to regulatory arbitrage. The rules then — and amazingly still now in the U.S. — have no capital charge for any off-balance-sheet commitment with a maturity of less than one year. Shazaam — the world of 364-day commitments was born. From this came structured investment vehicles and so much else that brought banking to its knees, but nothing about this crisis was in fact unforeseen. It was just unaddressed.

Another big problem with Basel I, especially after revisions to it in 2001, was undue reliance on the credit rating agencies. Basel I turned to the rating agencies because it needed someone to rate credit risk for the risk-based regime to work. Rating agencies have become manifestly flawed and conflicted since 1988, but it’s critical to remember that, absent some capital recognition, banks have strong perverse incentives to take lots of risk. A simple leverage ratio — 5% in the U.S. — is way too low for high-risk assets. It also makes holding low-risk ones prohibitive from an economic-capital point of view. Relying on rating agencies at the start of the Basel regime was thus a least-worst decision — it provided some insight into risk on what were then thought to be objective grounds in a reasonably simple way that cured a serious underlying financial system risk.

Basel II in fact sought to reckon with the rating agencies problems by relying more on internal bank models and less on rating agencies. It did this badly, of course, continuing rating agency reliance even as complex, untested models were brought to bear on risks no one — least of all the banks — understood. Worse still, it started in the E.U. at the height of a boom so that capital calculations were based on the best of all possible assumptions in the best of all possible worlds. Reliance on rating agencies is problematic, but nowhere near as much as using short-term-risk scenarios that allowed banks under Basel II to hold capital on grounds that nothing could ever go wrong.

Example: the risk-based charge for residential mortgages. The Basel II rules permit an extremely low risk-based charge for mortgages because the stress scenarios and risk correlations on which they were based were boom-time ones. Even then, they weren’t good enough for the industry — a read through comment letters filed in the U.S. at the time shows one after another arguing that there was simply no risk at all in mortgages because a home is always there to collateralize the loan and house prices always go up. The fault here wasn’t the effort to base capital on risk; rather, it was the signal failure to know how much risk there really was.

Harsh judgments about Basel II fail not only to note what was right about it at the start, but also what’s being done now to fix what was wrong. The Basel Committee has just released a set of final revisions to the accord. These will force banks — very much against their will — to use their own credit-risk analytics and no longer simply to rely on the credit rating agencies. Credit rating agencies’ determinations remain in the rule as a benchmark for risk-based capital, but just that. In essence, credit rating agencies’ determinations now are a floor against which banks must make their own credit-risk judgments and hike capital or show why this doesn’t need to be done. Is this perfect? Of course not. Was Basel I better and thus the standard to which regulators should return? Again, of course not.

The real poster children for Basel II’s critics are the bulge-bracket U.S. investment banks. They went on Basel II in 2004 when the SEC rolled over and gave them the consolidated supervised entity charter. When it did so, the investment banks got the gift of a lifetime — an immediate huge reduction in broker-dealer net capital combined with complex capital rules none of them or their regulator understood without the discipline of a leverage standard or — more important still — any SEC prudential or supervisory standards that would ensure Goldman Sachs, Lehman et al. knew what they were doing and in fact did it.

Is Basel II the cause of the market debacle for the bulge-bracket firms? One might say so, but this too would be way wrong. Basel II on its own — without checks, balances or controls — is in fact considerably worse than a simpleminded capital regime like the net-capital rules that preceded them for the investment firms. But that isn’t Basel II as it was or, still more important, as it will become now that regulators have sobered up.