Metlife, Living Wills, and the Necessity of Technocratic Regulators

Here’s an important thing to remember from these two events: Regulators will be necessary to tackle the problem these two case studies represent, no matter how we structure regulations.

Since the beginning of financial reform, there have been calls for “structural” reforms over “technocratic” ones. This is fed by reasonable concerns over the effectiveness of regulators in the long run and worries about regulatory capture. Structural often means regulations that are more binding, such as price regulations, limitations on size or activities, and ones that are more prophylactic than reactive. These are critical. But structural has also been taken to mean regulations that don’t have any ambiguity at all, ones where we can get rid of regulators and the rules will take care of themselves. The courts appear to be pushing on this more aggressively going forward, and some are saying that the they have a point on Metlife given the way Dodd-Frank works.

But there are no simple “big dumb rules” to follow to tackle these problems. The question is how to make regulators accountable rather than make them unnecessary.

Problem One: Planning for Bankruptcy

Take the living wills. Any responsible way of tackling Too Big to Fail would force the riskiest firms to plan for their bankruptcy in advance. As Matt Levine notes, the living wills force banks to think about their own death in a productive way. This isn’t something that would naturally happen as a matter of economic incentives, especially as the costs of a catastrophic failure are borne by the public as a whole. To internalize these costs you need preventative plans, and as a result you need regulators to see if these plans make any sense or not.

What would be the alternative? On the free market side, Rep. Jeb Hensarling (R-TX) recently said, as he led a vote to repeal parts of Dodd-Frank, ”When it comes to the resolution of these large, complex financial institutions, should we have bailouts or should we have bankruptcy? […] the bankruptcy process is far superior.” But the living wills show that the bankruptcy process would be a disaster as the firms are currently structured. As mentioned above, there’s no incentive for them to change.

On the structural side, can’t we just break them up and be done with it? One problem here is that two of the five banks that failed their living wills, State Street ($224bn in liabilities) and Bank of New York Mellon ($355bn), are already below or near the generally proposed size cap of $350–$500 billion dollars. (Wells Fargo, which also failed its living will, may straddle or be below the line depending on how you measure deposit liabilities in the cap.) Size is important, and should be brought down in case the FDIC needs to do an emergency resolution, but size is clearly not sufficient to ensure a safe failure. Efforts to simply force the banks to carry a lot more of a capital cushion are necessary, but by definition they won’t help in the case of a failure, when those cushions are exhausted.

On the clever market mechanism side, some have suggested removing liability for bank shareholders as a market alternative to dealing with these externality costs, but even then you need regulators to come up with the costs, which they’d determine from living wills tests like these. You can’t simply look at a “too big to fail” subsidy, because that simply is a measure of bailout probability rather than a measure of external damages to the economy.

You need people looking at the plans, trying to figure out if the cases are reasonable, if they make sense, and determine how to proceed. Even further, as Peter Eavis cleverly notes, a little bit of human uncertainty can help deal with the “Minsky moment” where banks game a box-checking operation, making firms think harder and making the financial system safer.

Problem Two: Who Should Have to Follow Structural Banking Rules?

This is also true for Metlife. Here’s an important lesson the housing bubble and financial crisis taught us: There’s not a clear line to be drawn between the activities of commercial banking, with maturity transformation, the creation of money and systemic bank runs, and all other industries, be they investment banking (Bear, Lehman), insurance (AIG), or the finance wing of industrial companies (GE Capital) who take up such activities. As such, we need to both regulate activities broadly and regulate the riskiest institutions engaging in banking-like activities with tough structural rules.

Dodd-Frank attempts to do both. There are rules on everything from consumer lending to credit derivatives to money market mutual funds on the activities side. On the institutions side we have a system of graduated prudential regulations for the largest bank holding companies. As they get bigger they have stricter capital requirements and stress testing.

But then the question becomes: Which firms that aren’t banks should have to follow these tougher rules? We need regulators to have a clear process to determine how it will regulate the riskiest firms as if they had to follow the banking rules. But, given the nature of complex financial structures, we need to make sure the rules aren’t simply manipulated. Not doing this is an invitation to regulatory arbitrage, where activities simply migrate to the institutions with the laxest regulations.

How do you do this? The text of Dodd-Frank calls for an initial cut on firms that receive 85 percent of their revenues (or 85 percent of assets) are financial. The text of Dodd-Frank then gives a 10-point checklist, including leverage, concentration, and off-balance-sheet exposures, for regulators to consider. Regulators in turn created a clear list of what could qualify a firm to be considered for SIFI ($50bn in size plus either $30bn in CDS, 15-to-1 leverage rate, 10 percent short-term debt, etc.), and then a six-point test measuring the remaining risks from the checklist. When it comes to measuring and balancing those risks, regulators are needed to ensure that the case and the objections to the case are sufficiently met.

How else could you do it? It doesn’t make sense to impose the toughest banking regulations for all firms everywhere, or across all industries. Should we force large firms to follow financial regulations even if they don’t do risky financial activities? You could say, “Congress should legislate that no firms that aren’t formal banks should be engaged in maturity transformation.” But how do you define maturity transformation, especially as banks evolve new instruments that they will say aren’t maturity transformation? And again, if it is a simple yes/no test, why won’t financial firms simply innovate around it? The courts clearly overstepped their boundaries, but they also blocked a necessary and important tool towards a safer financial system.

A General Problem

This is one area where my thinking has evolved over the years. All “structural” rules need some technocracy to implement them. It’s difficult to measure the size of a bank, so you need people to do it. Glass-Steagall is often held up as a clear structural regulation because it drew a straight line between commercial and investment banking. Yet that line became hard to draw as time went on, and it eventually dissolved before it was repealed. It’s not trivial for the Volcker Rule to draw a line between hedge fund-like trading and trading for clients. And that’s okay. All lines will have ambiguity around them; the goal is to create the best rules to try and mitigate them.

Crucially, conservatives, when they rail against the administrative state, are simply relying on different people to do the administrative work. Whenever they call for judges and private authority with public power delegated to it, they simply obscure that this power is taken away from public authority.

There’s simply no way out of this half of the equation. You need people to make the call. There are ways to make laws, regulations, and rule-writing better. Senator Warren has some ideas to help us get there, and we at Roosevelt are trying as well. It’s an essential part of the process. But it requires us to tackle the problem of the administration state straight on rather than hoping to abstract away from it.