Firms Will Need To Prove They Can Fail

As Derek pointed out earlier, Treasury Secretary Timothy Geithner testified today before the House Financial Services Committee. I was pleasantly surprised to see that half of his prepared testimony involved the too big to fail problem. I was even more delightfully shocked to see that he offered a legitimate, tangible solution that could help with the problem. He suggests that firms will need to explain how they could be resolved in the event of their failure.

A Good, But Ultimately Inadequate Proposal

Before getting to his more novel idea, he first suggests a more well-known one:

Second, we will impose tough rules on our largest, most leveraged, and most interconnected firms. We will require these firms to hold more capital to protect the system in the event of the firm's failure. And we will make the financial markets more resilient.

We will require bigger buffers in the financial system to make it strong enough to withstand the failure of individual firms, and will reduce the threat of contagion caused by interconnections among major firms. This will include raising capital and liquidity requirements for all banking firms, and raising capital charges on exposures between financial firms. It will include comprehensively regulating over-the-counter (OTC) derivative markets, including by substantially increasing the use of well-regulated central clearing platforms. And it will include strengthening supervision and regulation of critical payment, clearing, and settlement systems

I support higher capital requirements for financial firms and securities, because it just makes sense. But those requirements alone cannot prevent the too big to fail problem. Let me explain why through an analogy.

Imagine you create a security pegged to some asset. Depending on how that asset's value changes, the security can turn a profit or loss on your initial investment. So you create a scenario where you assume that if things got as bad as they ever had, say like in the Great Depression, you figure out what would happen to that asset's value. You then structure the security in such a way, say with a large cash reserve, to withstand that shock to the underlying asset and at least break even. That's the story of mortgage-backed securities and rating agencies. We all know how well that turned out.

The point is that it's impossible to know how bad things could get, so no one can ever accurately predict the right amount of capital that firms would need to set aside in a catastrophic scenario. That capital may help and provide plenty of cushion most of the time. But it's not most of the time we're talking about -- we care about times that are unprecedented.

A Better Solution

So a better solution must be sought. Luckily, Geithner offers one. And it's a good one:

We will require our major financial firms to prepare and regularly update a credible plan for their rapid resolution in the event of severe financial distress. We will require supervisors to carefully evaluate the plan on an ongoing basis. This requirement will create incentives for a firm to better monitor and simplify its organizational structure and would better prepare the government - as well as the firm's investors, creditors, and counterparties - in the event the firm collapsed.

I love this idea. In theory, it's great on many levels. First, it forces firms who pose a systemic risk to do the lion's share of the work in determining how they could be resolved effectively. Second, it gives regulators a tangible plan to work with for each firm, rather than leaving them to generalize what might happen if certain types of businesses failed.

A while back, I complained that while a resolution authority for non-bank financial institutions sounds like a great idea in theory, I wasn't sure how they'd accomplish the goal of resolving some of these financial giants. This speaks to exactly that problem. It puts that burden on the shoulders of the institutions themselves to show how the resolution authority would do it.

I have also noted in the past that antitrust might be used to prevent firms from becoming too big to fail. Such resolution plans could be used in conjunction with that goal as well. What if, as part of regulators approving a merger, the firms also had to submit a plan showing how resolution of the resulting mega-firm would work? If such a plan can be created, then you've satisfied the worry of the new firm's interconnectedness. If no such convincing plan can be presented, then the merger is forbidden.

Of course, for these plans to matter, regulators would have to be savvy enough to challenge the plans, and imagine ways in which they might break down. For example, if AIG were to have presented such a plan, a regulator should have seen through it and asked, "Well, what if many of those credit-default swaps you've issued are cashed in by big banks simultaneously -- where would you come up with the capital needed?" A shrug of the shoulders wouldn't do. Firms will have to be able to respond to regulators' scenarios, including if other dominos fell due to their failure. But regulators must understand the economy extremely well and know what chain reactions to worry about.

But therein lies a worry, and it's similar to the inadequacy of capital requirements I complained about. Can regulators really predict scenarios that might throw the plan out of whack? It's hard to tell how badly an industry might be ailing or which ones simultaneously. How stressful will the scenarios considered be in evaluating these plans -- and how will regulators know how stressful they should be?

I also have another important question: what if a firm can't come up with such a plan. What if regulators successfully poke holes in every plan a firm presents, and it cannot come up with a working plan that would allow for its resolution without causing a major systemic disturbance? Does that mean the government breaks up the firm? Or does that mean that the firm is simply considered "too big to fail" and must put up additional capital? For the reason I explained in the prior section, I certainly hope it's the former. After all, these plans are worthless if they aren't used in the event of failure.