How to Stop Banks from Gaming New Capital Rules

Treasury Secretary Jack Lew made a bold vow recently when he said that the bulk of financial reform would be complete by yearend.

While his pledge got a lot of ink, Lew said something else that was overlooked, offering a bone of sorts to bankers who worry that layers of new rules will bury their companies.

"[T]here is a growing understanding that once the rules are in place, there will be time to see what is working and what is not, and then we can make appropriate adjustments."

I can't remember any policymaker involved in bringing Dodd-Frank to life, much less a senior Obama administration official, saying we might dial some of the rules back. But Lew clearly seems to get that the proverbial pendulum could swing too far.

"[W]hen we finish this work, our regulatory system will strike a balance between incentives for innovation and protections from excessive risk-taking," Lew said. "This is about being effective so our financial system is competitive and efficient, while protecting our economy and taxpayers from excessive risk-taking."

It's good to know someone in power still worries about being competitive and efficient.

If Lew wants a place to start, it should be with simplifying capital requirements.

Like everyone else who loves their mother and feels patriotic eating apple pie, I support strong capital levels, especially at our largest financial institutions.

But we don't need thousands of pages of regulation (the latest Basel III rule alone is 971 pages) to ensure our banks are well capitalized.

When the rule-writing is done we will have a half-dozen separate ratios to measure capital strength, and we'll be calculating the numerator and the denominator differently for each one. Layered over those ratios will be a "surcharge" for being big and two "buffers" — a "conservation" buffer and a "countercyclical" buffer.

U.S. regulators just adopted a new leverage rule that will require holding companies to hold $5 in equity for every $100 of assets and banks to hold $6 for every $100 of assets.

But no one seems able to explain how the new leverage ratios relate to the old one, which is 3% for holding companies and banks alike.

Do the same types of capital count toward the numerator in both? Is the denominator determined the same way? Will the U.S. leverage ratio differ from the one laid out by the Basel Committee on Banking Supervision and adopted by rivals in other countries? Why did regulators decide to go with a lower figure for holding companies? Are they trying to push activities out of banks?

It's a fact that complexity leads to loopholes, and if our large banks have proven an expertise in anything, it's exploiting loopholes.

"The more complicated a rule is, the more unintended consequences it yields and the more opportunities to game it there are," one banker admitted to me. "And when there is an opportunity to game, the banks are always going to beat the regulators. Always."

So doesn't it make sense to simplify? I think we need both leverage and risk-based ratios, but ones that are much less complicated than we are currently building.

The leverage ratio should be an equity capital to total assets ratio that captures both on- and off-balance-sheet assets. I don't know what the "right" number is, but the leverage ratio ought to remain the backstop that it's always been — not the binding constraint.

A U.S. Treasury security should not require the same amount of capital that a corporate loan does.

"To run the financial system on the assumption that all assets have the same risk is nuts," an executive at another large bank told me. (Note to everyone who is rolling their eyes right now: just because a large-bank executive says something doesn't make it wrong.) If the leverage ratio is the binding constraint, guess what will happen? Banks will pile into the riskiest assets to try and earn a market return on this higher capital requirement.

Sources say this is among the Federal Reserve's biggest fears, and may be why it argued for a lower ratio for holding companies. At 5%, firms may shift more of their activities to holding company subsidiaries — under the Fed's supervision.

A reasonable leverage ratio should be bolstered by a fundamental risk-based capital rule.

The current rules are too complex and have been discredited by the fact that regulators and bankers have frequently misjudged the risk posed by various asset categories.

In the late 1990s regulators started letting the large banks use their own models to determine how much risk an asset poses, and therefore how much capital is requires. And these models have grown ever more complicated so it's impossible for outsiders to figure out if a bank's models are effective. Even worse, these models have become negotiating points as bankers try to convince their examiners that a model's results are accurate.

"Figuring out how they work and making sure the banks … aren't gaming the system … sucks up a lot of resources," says one former senior regulator. "It's a waste of resources. We should have something simpler that's transparent but still risk-based.

"I think we should get rid of the models-based stuff."

Reverting to something like the original risk-weighting system in Basel I would mean slotting assets into broad buckets and assigning a capital charge to each bucket.

For instance, because a U.S. Treasury security is unlikely to default, it received a zero weighting under risk-based capital, which meant a bank did not have to hold any risk-based capital against that asset. (These assets are still subject to the leverage ratio.) Likewise, a mortgage secured by a home required half as much risk-based capital as an unsecured commercial loan.

Yes, the buckets are crude and imperfect, but at least they take risk into account and they are better than relying on the banks' own opaque and easy-to-manipulate models. We could refine and update Basel I's buckets to better reflect the risk posed by various asset classes.

And don't forget that capital rules are buttressed by all sorts of other regulation including on-site examiners and rigorous stress-testing and capital planning at the largest banks. And it's also worth remembering that regulators have a lot of power over how much capital a bank holds.

Karen Shaw Petrou of Federal Financial Analytics illustrated that point in an April report full of her usual straight-talk style.

"What's the real risk remedy? I go back to something a lot simpler than leverage: truth and punishment. If regulators come out from behind the curtain — say, by disclosing their Camels ratings in concert with stress-test results — we'd know more not just about banks, but also their ability to spot risk and, then, stop it."

There are lots of ways to ensure banks are well capitalized that don't cause us all to go blind reading Federal Register notices.