Economics of Contempt

I think this, from Mike Konczal, is the exact wrong way to think about regulatory discretion in the context of prompt corrective action (PCA):

So this risk council [in the Dodd bill] will come up with rules, and the Federal Reserve will carry them out. Can these rules change on the fly? I’m not sure, but I assume from the bill that they can. Can these rules change in a crisis? I don’t see why not – there’s complete discretion over how they are created by the regulators.

But this is the exact opposite spirit of the 1991 FDICIA bill that created prompt corrective action. When the time comes to resolve a firm, the “ok, you’ve done the best with your discretion, but it’s time to call it” is written right into the bill itself: “The level specified under subparagraph (A)(i) shall require tangible equity in an amount—(i) not less than 2 percent of total assets.”

This is crucial.

The biggest criticism of the current PCA regime for commercial banks (which was established in FDICIA) is that the triggers are outdated, and generally don't come into play until it's too late. The reason they can't be easily updated to accurately reflect the different stages of a bank's deterioration is precisely because the relevant metric (tangible equity to total assets) is set at the statutory level. Why would we want to make that mistake again? If we've learned anything from 18+ years of PCA for commercial banks, it's that we need realistic triggers that can evolve along with the financial sector.

Mike's concern about the systemic risk council changing the PCA triggers on the fly during a crisis is entirely misplaced. Any changes to the PCA rules would undoubtedly have to go through the normal APA rulemaking process, which requires a notice-and-public-comment period (lasting at least 30 days, and likely much longer) before changes to the rules can be made. The choice is absolutely not between setting the PCA triggers at the statutory level, and giving regulators unfettered discretion to change the PCA rules on a moment's notice.

Setting PCA triggers is simply not something that Congress can do. But if you somehow disagree, and think that the PCA triggers can and should be hardwired into the Dodd bill, then surely you have some well-defined set of PCA triggers in mind, right? By all means, I'd love to see them.

This is serious business, because we've made this mistake before, and I don't think anyone who knows this area is keen to make the same mistake again. PCA is not an area where you can set hard-and-fast rules at the statutory level, like the 15:1 leverage cap*. PCA triggers are exceedingly complicated, and they need to be formulated by regulators who have sufficient expertise — and who have sufficient flexibility to update them as the financial sector's funding profile changes — if we want the new PCA regime to stand the test of time.

-----------------
* On a side note, I'm completely fine with a hard leverage cap at the statutory level. I'm not sure 15:1 is necessarily the right number, but I'd be comfortable with anything ≤ 20:1. I'm not sure how a leverage cap would fit in with Basel III though, which could pose problems.

Instead of providing a plan for the financial institution's own resolution, we should instead use living wills as a way to force financial institutions to provide regulators with all of the information they'd need to actually carry out a rapid and orderly resolution. This may not seem like much, but it would actually be critically important.

The most important piece of information that financial institutions should be required to periodically submit would be a detailed description of cross-border and intra-company funding arrangements. This would allow regulators to see the direct cross-border consequences of seizing the holding company and placing it in receivership. For example, had Lehman been required to submit this kind of detailed description, regulators would have seen that Lehman centralized its funding/liquidity in the holdco, and would have known that the moment the holdco filed for bankruptcy, Lehman's European broker-dealer subsidiary (LBIE) would have been insolvent too. At the very least, this would have focused regulators' attention on the direct consequences of LBIE's insolvency. It's possible that had the Bank of England known in advance that LBIE's insolvency would have immediately frozen tens of billions in prime brokerage assets, it would have provided LIBIE with the day-to-day financing necessary to stay out of bankruptcy until the Nomura deal closed (like the Fed did with Lehman's US broker-dealer, which was sold to Barclays).

Living wills should also include an updated list of the financial institution's counterparties, including exposures (on both a gross and net basis). If a financial institution can't provide that kind of comprehensive and up-to-date information, then, well, they're just going to have to invest in the necessarily operational infrastructure — that'll simpy be part of the cost of being a large, complex financial institution.

Another requirement, which is closely related to the description of the institution's funding arrangements, would be a list/map of where all of the institution's assets are held, and a description of the insolvency regime in each jurisdiction. For regulators to effectively coordinate with their international counterparts, they first need to know what issues they need to coordinate, and with whom. Knowing where all a financial institution's assets are held, and how they'll be treated in an involvency proceeding, is a necessary precondition to an orderly resolution of the financial institution.

The financial institution should also be required to submit a list of IT/compliance/general back office personnel that the regulators would need to retain to help them navigate the institution's data management systems. These employees wouldn't know how to make it rain even if they tried, so retaining them wouldn't entail paying ludicrously large bonuses that would tend to stir public anger. But their retention would be necessary, because it would allow regulators to focus on the task of actually carrying out the resolution, instead of fumbling around trying to learn the institution's data management systems.

Ideally, living wills would be folded into the prompt corrective action (PCA) regime. This way, if an institution's capital/liquidity started to deteriorate, regulators could force it to start augmenting its cross-border and intra-company funding arrangements in a way that would make it easier to resolve the financial institution. Or, just as importantly, regulators could start to restrict the financial institution's ability to engage in certain intra-company transfers that would make an orderly resolution more difficult. But for regulators to be able to use PCA to start softening the ground for an orderly resolution, they need to first have all the necessary information in hand. That's something that living wills can realistically do.

This is obviously not an exhaustive list, but I think it's illustrative. I realize that financial institutions would scream and yell about how much it would cost to gather all the information required in a living will on a regular basis ("that's not how our internal systems work!"). And for some financial institutions, I'm sure it would be quite costly to develop the operational capacity to comply with the ongoing requirements of living wills. But you know what? That's just the cost of doing business as a large, complex financial institution.

I suppose I should say something about "living wills," since the idea doesn't seem to be going away. It's hard to critique the idea, since no one seems to be willing to define living wills. The Dodd bill simply requires financial institutions to periodically submit:

[A plan] for rapid and orderly resolution in the event of material financial distress or failure.

Clear as mud. Thanks guys. In any event, there are really two arguments I want to make regarding living wills, so I'm going to split this into two posts: one on what living wills can't do, and one on what living wills can do.

What Living Wills Can't Do

I think it's a bad idea to rely on the financial institution to submit a plan for their own resolution. Resolving failed financial institutions is something that regulators do; it's not something that financial institutions do.

The problem is that the form of the resolution, the cost to the resolution fund, and the damage to the financial system all depend critically on the potential buyers. Who the potential buyers will be, and how much they'll be willing to pay, is inherently unknowable in advance. Think about how much less damage Lehman's failure would have caused had there been a buyer — that is, had the FSA allowed Barclays to buy Lehman, with the Street-financed consortium eating the losses on $60bn of Lehman's worst assets. Without Barclays though, there was no buyer, and the overall cost of resolving Lehman skyrocketed. But who could have known in advance that there would be no buyer when the time came? Put another way, had Lehman submitted a living will saying that if it ran into trouble, it would sell itself to another major bank, with a Street-financed consortium sweetening the deal by eating the losses on Lehman's worst assets (i.e., an LTCM-like solution), would any regulator have rejected that living will? Of course not — because in truth, it wouldn't have been an unreasonable plan. But when it came down to it, Lehman's living will would've been useless. And now we're right back where we started.

Moreover, you can guarantee that any resolution plan that a financial institution submits will be excessively rosy/optimistic. The problem, again, is that everything depends on valuations. For example, one component of any resolution plan will necessarily be a plan to liquidate a portion of the institution's inventory of liquid assets. But what assets will be liquid in a period of financial stress, and how much will those assets fetch? Your guess is as good as mine. But what regulator could possibly have enough market knowledge/experience to challenge the financial institution on its valuations?

For the sake of argument, let's say a financial institution plans as part of its living will to liquidate half of its Agency MBS position to raise cash, and estimates that it can get ~90 cents on the dollar for half its Agency MBS position in a period of financial stress (these numbers are purely illustrative). When the financial institution meets with the Fed and FDIC officials to review its resolution plan, do you think there's any way that any Fed or FDIC official could seriously argue valuations with the financial institution's head MBS trader? Of course not. The trader undoubtedly knows infinitely more about his market than any of the regulators, whose jobs, while important, are most assuredly not to be experts on MBS valuations. So the regulators would ultimately have to accept the financial institution's valuations.

The same argument holds for divestitures as well — regulators wouldn't have been able to seriously challenge AIG's valuation for its ILFC subsidiary, because at the time of AIG's failure, everyone legitimately thought ILFC would fetch top dollar. Again, it all depends on who the buyers would be in a period of financial stress (Warren Buffett? Some sovereign wealth fund? Metlife?), and how much they would be willing to pony up.

I have to roll my eyes every time someone points out that the proposed OTC derivatives reforms still preempt state anti-gambling laws. This is generally presented as some sort of obviously-ridiculous exemption that proves that Dodd, or Geithner, or whoever, is in the pocket of Wall Street. First, I suspect that, if asked, the majority of people complaining about state anti-gambling laws being preempted for OTC derivatives would say that they think anti-gambling laws are anachronistic holdovers from a more paternalistic era, and should be repealed. Second, state anti-gambling laws are famously (over)broad — for instance, New York Law § 5-401 (Gen. Oblig.) provides:

All wagers, bets or stakes, made to depend upon any race, or upon any gaming by lot or chance, or upon any lot, chance, casualty, or unknown or contingent event whatever, shall be unlawful.

Obviously, this could ensnare a whole host of contracts that we consider perfectly legitimate (e.g., insurance). So for this law to stay on the books, we need all sorts of exemptions and carve-outs for legitimate contracts which could nevertheless fall within § 5-401. Once we've decided at the federal level that OTC derivatives are legitimate financial instruments — and like or not, we have decided that — then preempting ridiculous state anti-gambling laws is perfectly appropriate.

The fact that the proposed OTC derivatives reforms preempt state anti-gambling laws is evidence of nothing other than that state anti-gambling laws are silly, overbroad relics.

Now that the financial reform debate is heating up again, I want to caution people against viewing a wider clearing requirement for OTC derivatives as necessarily better. Clearinghouses are not a panacea by any means, and pushing instruments that aren't mature enough or sufficiently liquid for clearing onto clearinghouses would be a very bad idea.

In order to properly manage counterparty risk, a clearinghouse needs to be able to accurately price the instruments it clears. Clearinghouses manage counterparty exposure by requiring counterparties to post initial margin at the beginning of the trade, and — crucially — daily variation margin. As explained by Robert Bliss and Robert Steigerwald in a Chicago Fed paper:

Margining systems are designed to ensure that in the event that a clearing member fails to meet a margin call, sufficient funds remain readily available to close out the member’s positions without loss to the CCP in most market conditions. As a complementary risk-management mechanism, the gains and losses from open positions are posted to a clearing member’s margin account on a regular (usually daily) basis and result in calls for variation settlement (or variation margin). The variation settlement reflects periodic mark-to-market fluctuations and is an important mechanism for assuring the collateral held by the CCP is likely to be sufficient to meet the needs of the CCP in the event of a default.

Given the extent to which clearinghouses rely on market prices to mitigate counterparty risk (which is, after all, the main reason we're pushing central clearing in the OTC derivatives space), it's absolutely crucial that the mark-to-market prices be reasonably accurate. Reliable mark-to-market pricing, of course, requires a liquid market.

It seems that a lot of people are pushing for the government to "err on the safe side" by enacting a very wide clearing requirement for OTC derivatives, with exceptions only for truly bespoke deals. The more OTC derivatives that trade through clearinghouses, the safer the system, right? Well, not really — it's not at all clear that this approach would be "erring on the safe side."

In fact, we have a very good example of the potential dangers of this approach: AIGFP. Yes, I'm fully aware that AIG's CDSs weren't centrally cleared. But they were fully collateralized, requiring AIG to post collateral daily based on market prices — the equivalent of a clearinghouse collecting daily variation margin from counterparties to cleared trades.

Most people agree that it was the decision to collateralize their CDSs that ultimately doomed AIG. Since the contracts required AIG to post collateral based on the mark-to-market prices of the underlying CDOs, the decision to collateralize the trades meant that AIG was exposed to short-term fluctuations in the CDO market. AIG was fine with that, because they foolishly believed that the CDO market was deep enough and liquid enough that mark-to-market prices would remain stable.

As everyone knows by now, liquidity in the CDO market vanished, and the market prices of CDOs plunged. Because they had agreed to collateralize their CDSs, AIG was suddenly forced to post billions of dollars in collateral. The effect would've been the same had AIG been required to clear its CDSs through a clearinghouse. The clearinghouse would've required AIG to post daily variation margin based on market prices — this, again, is the primary way that clearinghouses mitigate counterparty risk — which means that AIG would've been forced to post the same amount of collateral when liquidity dried up in the CDO market.

The point is that forcing OTC derivatives that aren't sufficiently liquid onto clearinghouses is not necessarily "playing it safe" from a regulatory perspective. It greatly increases the chances that a clearinghouse will misprice its counterparty risk, and end up not having collected enough variation margin to cover the losses from a default. And I, for one, would rather not see what a clearinghouse failure looks like in my lifetime.

I hope to get time to write a full post on Sen. Dodd's financial reform bill in the next couple of days, but I do want to highlight one significant provision in the proposed resolution authority (which very much surprised me). In normal commercial bank resolutions, the derivatives safe harbor doesn't apply for the first 24 hours — that is, counterparties to a failed bank's derivatives contracts (and repos, etc.) can't terminate, liquidate, or net the derivatives until 5:00 p.m. on the business day following the FDIC's seizure of the bank. The purpose of this 1-day breather is to give the FDIC time to transfer a failed bank's derivatives positions, usually en masse, to healthy institutions.

Section 210(c)(10)(B) of Sen. Dodd's bill, however, provides that the derivatives safe harbor doesn't apply for five (5) full business days after a covered nonbank financial institution is seized (pp. 231–232). This essentially gives the FDIC five days instead of one to figure out what to do with a failed financial institution's derivatives positions. That's a major difference, especially considering that the House bill stuck with the traditional 1-day breather. Frankly, I think the House bill was wrong, and that the FDIC would almost certainly need more than one day to sort out where to transfer a failed financial institution's derivatives positions. But at first blush, five days strikes me as way too long. (The five days immediately following Lehman's failure felt like an eternity.) I think three days strikes the right balance, but even that's probably optimistic.

In any event, that provision immediately caught my eye.

Simon Johnson, a year ago, on a resolution authority:

Imagine what happens when these powers are passed. The U.S. Treasury and FDIC would immediately have the tools [they] need to walk into America’s largest financial institutions, such as Citibank or Bank of America, and liquidate them, or rewrite their contracts and capital structures. Such powers are clearly useful: if the banks are undercapitalized, and private money is not available, then the government could force creditors to swap claims into equity, thus instantly recapitalizing the banks while avoiding use of taxpayer funds. With such steps, the problem of moral hazard, where creditors to banks are bailed out by taxpayers, would at once be forgotten. Shareholders in banks would lose through dilution, some (unsecured?) creditors would lose with debt-equity swaps, while the nation would be better off having a well-capitalized banking system. The banks would remain private but now be controlled by (ex)creditors.

Simon Johnson, a week ago, on a resolution authority:

But this notion of a resolution authority that can handle massive banks is a complete unicorn – a mythical beast with magical powers that does not really exist. A US resolution authority does nothing to help handle the failure of international banks – there is no cross-border resolution authority, nor will there be one anytime soon. If a Citi or a JP Morgan or a Goldman were to fail, our government would be in exactly the same awkward position as it was in during September-October 2008.

The point, yet again, is that Simon Johnson has absolutely no clue what he's talking about. He is the Ben Stein of financial reform.

Some progressives, Paul Krugman chief among them, want to draw a line in the sand on the CFPA, accepting only a fully independent agency. This is understandable after the year-long health care debate, in which progressives were, at the end of the day, forced to accept a severely diluted and distasteful compromise bill. Nevertheless, I think refusing to compromise on the CFPA is a huge mistake. It's worth examining the severe flaws in the arguments for why progressives should refuse to compromise on the CFPA. First, here's Krugman:

There are times when even a highly imperfect reform is much better than nothing; this is very much the case for health care. But financial reform is different. An imperfect health care bill can be revised in the light of experience, and if Democrats pass the current plan there will be steady pressure to make it better. A weak financial reform, by contrast, wouldn’t be tested until the next big crisis. All it would do is create a false sense of security and a fig leaf for politicians opposed to any serious action — then fail in the clinch.

Better, then, to take a stand, and put the enemies of reform on the spot.
...
In summary, then, it’s time to draw a line in the sand. No reform, coupled with a campaign to name and shame the people responsible, is better than a cosmetic reform that just covers up failure to act.

First of all, Krugman is simply wrong to suggest that a weak CFPA "wouldn't be tested until the next big crisis." This argument applies to crisis response measures, like the resolution authority or Section 13(3) actions, but it does not apply to the CFPA. The CFPA is prophylactic — it's not part of the crisis response measures.

Next, here's James Kwak:

If we want change, someone has to be willing to stand up for it. If you want to win a negotiation, you have to be willing to walk away. If you can’t do that, you will get rolled on every issue.

The argument that "you have to be willing to walk away" doesn't apply to all negotiations. Specifically, it only applies if the Republicans would consider a deal on the CFPA preferable to no CFPA at all. That is most assuredly not the case here. Walking away isn't something we can threaten the Republicans with, because no CFPA is exactly what they want! Our best chance is to find a Republican who, despite having a preference for no CFPA, is nevertheless willing to deal. That Republican, it appears, is Bob Corker. So now we have to find the strongest version of the CFPA that can get 60 votes, and go with it. Look, I wish there were 80 Democrats in the Senate, and we could pass whatever we wanted. But there aren't 80 Democrats in the Senate, and there's no sense pretending that there are.

Finally, both Krugman and Kwak are seriously overestimating the effect of "a campaign to name and shame the people responsible" for killing an independent CFPA. It's always tempting to think that we can force legislators to record "no" votes on a particular issue and then use that against them in the next election, but this is an extremely high-risk strategy, for two reasons. First, it's simply impossible to predict what the important issues in the next election will be. Remember when the defining issue of the 2008 election was going to be immigration? Or in the primaries, when everyone thought the key issue in the general election was going to be how the candidates voted on the Iraq War? Unfortunately, voters have very short attention spans, and the key issue in a given election — to the extent there is one at all — usually ends up being whatever issue the media happens to be hyperventalating about in October.

The other reason a "campaign to name and shame" CFPA opponents is a mistake is that even if it does help to elect more CFPA-friendly legislators, who's to say that an independent CFPA will come up for a vote again during that Congress? The leadership has to choose what issues they want to spend their political capital on in a given Congress, and they're usually not keen on spending political capital on a fight that they just fought in the previous Congress. That's why, for example, you don't see immigration, Social Security, or comprehensive education reform on the agenda for this Congress. It's also why health care took 15 years to come up again after Clinton failed to pass a health care bill in '94. For better or worse, the CFPA's moment is now. If we don't pass some version of the CFPA now, it'll be 10 years before we get another bite at the apple.

The virtues of passing a non-independent CFPA now are that (1) even a non-independent CFPA will provide some protection to consumers (whereas no CFPA will help consumers not a whit); and (2) it's much easier to strengthen an already-existing agency than it is to create a whole new agency. Measures to strengthen a CFPA can be inserted into larger legislative packages in a way that measures to create a whole new independent CFPA cannot. Also, strengthening an existing agency is a convenient and politically acceptable response when scandals inevitably crop up. If there's some scandal involving, say, a mortgage broker that turned out to be little more than a Ponzi scheme, then it's much easier for politicians to say, "Well, we're going to make sure the CFPA has the authority to address that," than it is for politicians to use an isolated scandal to create an independent CFPA with powers that go well beyong the scandal they're supposedly responding to.

So yes, we should be willing to compromise on the CFPA. And if a Fed-housed, self-funded CFPA is the compromise Corker and Dodd are offering, I think we should take it in a heartbeat.

Thanks Citi, I needed a good laugh:

No argument here.

Politico says Dodd and Corker are close to a deal that would house the CFPA inside the Fed. The article also says:

Under this latest broad-brush proposal, the new consumer body would have a presidentially approved director, an independent budget and autonomous rule-writing authority, sources said.

Well, if that's true, then Democrats should say "done" as fast as possible. But I tend to doubt that the CFPA's rulemaking authority will be completely autonomous — Dodd's weekend proposal included a process whereby a prudential regulator could appeal a proposed CFPA rule to a Systemic Risk Council made up of all the different financial regulators. (Hilariously, this sentence from the Politico article was literally the last sentence of the article. "Oh yeah, by the way, here's the substance." Classic Politico.)

But here's why I think housing the CFPA in the Fed could be a very good deal — although I stress "could." (I originally posted this last night as a comment to Yves Smith's post, but I wanted to lay out the argument for housing the CFPA in the Fed here as well, because it wasn't obvious to me at first either.)

Ideally, the CFPA would be a self-funded, independent agency. The great benefit of that arrangement is that it would create an agency (1) whose sole purpose for existing is consumer protection (which is huge, because that means it’ll always find more consumer protection to do), and (2) which is relatively free of political interference. Housing the CFPA inside Treasury is sub-optimal because the Treasury Secretary is a political appointee, and the CFPA would likely not vigorously enforce consumer protection regulations under Treasury Secretary Phil Gramm. (Plus the CFPA could be de-fanged through the budget process.)

Housing the CFPA inside the FDIC would make the CFPA both self-funding and relatively free from political interference, but there’s a huge jurisdictional problem with that plan. The FDIC only has jurisdiction over commercial banks and thrifts, and one of the main purposes of the CFPA is to regulate all the non-bank lenders out there. (That’s the reason Shelby proposed putting it in the FDIC.) So an FDIC-housed CFPA is a non-starter.

The Fed, on the other hand, does regulate non-banks (through its consolidated supervision of bank holding companies). It’s also self-funding, and is the regulator most insulated from political interference, due to the 14-year terms of Fed governors, among other things. So if we set up a CFPA inside the Fed, we will have achieved the goal of having a self-funded regulatory body that’s relatively free from political interference. Now, I know what you’re going to say: “The Fed is institutionally biased in favor of the banking industry, and anyway, Alan Greenspan never would have allowed a strong consumer regulator inside the Fed!” And that’s true. But now we just have to insulate the Fed-housed CFPA, in its rulemaking and enforcement activities, from the Board of Governors. That can certainly be done, and in a much more subtle and less controversial way than just waging an “Independent CFPA or Bust!” campaign. We don’t need total insulation from the Board of Governors either, because if consumer protection is made part of its charter, then the Board won’t be as hostile to consumer regulation as it has been in the past.

Essentially, the idea would be to have the CFPA leech off of the Fed’s independence. If structured properly, it could be a really shrewd move. Seriously, you shouldn't write this plan off before we see it.