We hate to do it, but it’s time for math. Or, at least fractions. We have to — U.S. regulators told us yesterday that they’ve finally figured out that the Basel capital rules are a ratio. For the Basel rules to make sense, regulators now acknowledge, one needs to do more than just hike the numerator – the capital requirement. A big number over zero is still zero. So, even as the numerator is set to go up, attention is turning to the Basel denominator – assets against which capital is charged. A bit late, in our view; we learned some time ago that ratios are supposed to represent relationships and, if you don’t get both parts right, you get it wrong.

One might have thought that recognizing that the numerator in Basel III rests atop a fragile denominator would deter regulators from piling still more into the Basel III standards through a systemic surcharge. But, undeterred by arithmetic, that is not the case. While regulators now are talking up efforts to make the Basel denominator better reflect real risk, they are pressing forward on stiff new SIFI surcharges for the numerator that could be announced as early as next week.

The numerator in the Basel rules is the absolute amount of capital a bank must hold. In the wake of the crisis, consensus has determined that the old rules were wrong because both the quantity and quality of capital were unduly lax. But, the new, improved Basel III numerator still rests atop Basel II’s denominator – assets times a weighting factor (expressed as RWA). Assets are given RWAs based on probability of default. Assumed to be zero for some assets like sovereign debt (oops), this translates into a zero RWA that, in the Basel ratio, then results in a zero risk-based capital requirement. The higher the probability of default in the regulator’s minds and/or the bank’s models, the higher the RWA and, then, the capital charge.

When the Basel III process started in 2009, regulators initially said they would work on both the composition of capital and RWAs. Chastened by the crisis, they planned to see how much latitude banks had under Basel II’s RWAs. A lot, as it turned out. In 2009, the U.K.’s regulator gave thirteen banks the same corporate loan and asked them to assess probability of default. For the same loan, banks came up with RWAs ranging from thirty to 189 – and that was for a simple position, a single corporate loan. On more complex structured instruments, RWAs ran wild.

Based on this, regulators thought they would do both parts of the Basel III equation, but they quickly grew faint after hard bargaining over the numerator. Knowing, though, that the denominator did matter, they went to a fall-back. Unable to deal with Basel II’s RWAs, the Basel regulators decided on a floor under them: the leverage requirement. This might control for RWA problems, but only if leverage in fact applies. So far, it’s far from clear that it will. The Basel III rules include a three-percent leverage requirement for both on- and off-balance sheet assets, but only as a supervisory – not minimum – requirement. And, of course, Basel III’s leverage rule only counts if it’s imposed which, at least in the EU, is far from certain.

Which makes RWAs still matter and explains why, as noted, U.S. regulators now are working on them. The goal is to get some sort of super-regulator to review each nation’s approach to RWAs to be sure they are stringent and truly reflect risk. How this will be done is at best unclear. In his testimony on Thursday, FRB Governor Tarullo outlined several RWA regimes that, we expect, will prove hard sells to his recalcitrant Basel buddies.

So, for now, Basel III is a ratio, but a ratio that still doesn’t work. When the Basel Committee piles the SIFI surcharge into the numerator next week, the numerator will grow even as the denominator remains, at best, squishy. It would, we think, make a lot more sense to go back to the 2009 plan: making the numerator and denominator work better in tandem so the total ratio makes sense going forward. But, apparently, that is not to be.