« The economics of ticket scalping | Main | Politicized Lending »
The Systemic Risk Council
I received this in an email today:
The best approach to preventing future financial crises is through the creation of a Systemic Risk Council composed of at least the Fed, Treasury, FDIC and SEC, according to William Isaac, former chairman of the Federal Deposit Insurance Corporation, and now chairman of Global Financial Services for consulting firm LECG.
To be sure, systemic risk is an important topic for regulators. But the idea of a "systemic risk council" is much better than the idea of a Systemic Risk Council, all caps. The more formal the institution, the greater its power to spook markets and become a source of systemic risk itself. Let's say that an SRC were summoned into being and one day the Council warned that systemic risk was unacceptably high. Economic activity would plummet and freeze up immediately and furthermore a liability issue could arise for any manager who did not shut down lots of plans. The practical reality is that the Council would have to be very, very cautious in its statements and actions. It's a bit like how the security alert these days is always on "Orange." High enough to indicate some kind of warning and to protect the regulators from a charge of complacency, but not so high as to terrify everyone or indeed inform anyone.
I've heard some people argue that once the Council sees that systemic risk is too high, it will proceed silently and secretly, in Ninja-like fashion (my words, not theirs), to stamp it out by alerting other regulators about the problem. I'm not convinced.
I am not arguing that we should simply stay put when it comes to financial regulation. It would be good to limit forum-shopping, for instance, and also abolish the Office of Thrift Supervision and transfer its powers to the FDIC. They are the group that dropped the ball on AIG, if you recall.
The broader point is this. Better regulation comes through many years of experience and gradual process improvements, built upon some reasonable methods for imposing regulatory accountability. That's how the FDIC got to be good at much of what it does. Better regulation does not come from sitting down, waving a wand, and hoping that a new name or box will address the problem you are concerned about. Keep that in mind next time you hear that "now is the unique moment," etc.
Posted by Tyler Cowen on June 13, 2009 at 07:51 AM in Economics | Permalink
Comments
Great post Tyler. I haven't looked at any formal measures, but the stock market's volatility is surely much much higher than it was before Bear Stearns. I think that is more due to the government's daily announcements, than to the underlying bad assets.
Posted by: Bob Murphy at Jun 13, 2009 8:39:52 AM
How about if the council, instead of having vaguely defined goals, is mandated to take specific actions in response to specific identified risks?
I do support the idea of a regulator tasked with directly measuring risk and responding to it. But I would have it use (at least as one of its tools) techniques from behavioural experiments to measure risk appetite or aversion, and also to control or influence attitudes to risk in how financial products are offered. I am not sure whether it's better to deliberately leave the mandate vague (to avoid people gaming the system) or to have clear rules not subject to interpretation (to reduce speculation about the motives and responses of the regulator).
I recently posted an article at VoxEU about this and would be interested in readers' thoughts on it:
http://www.knowingandmaking.com/2009/05/new-article-at-voxeu.html
Posted by: Leigh Caldwell at Jun 13, 2009 9:23:55 AM
The best goal for systemic risk regulators would be keeping risks uncorrelated. Given the tendency for markets to synchronize expectations (and hence risks), this goal might be accomplished by limiting the percentage of business that financial institutions can do in any one market and requiring that every institutions has a slightly different bundle of percentages. It is far more difficult to synchronize expectations when everybody has a different revenue stream.
The robustness of a partially desynchronized financial industry is a value to be considered alongside rcost-efficiency.
Posted by: Michael F. Martin at Jun 13, 2009 10:47:51 AM
I have no idea what 'systemic risk' is.
If you're saying 'banks shouldn't all make money the same way', then you are going to have to shut a lot of things down just when people are making money with it. Is that always a good idea?
It seems like if you started with the following principles:
-- higher reserves across the board, and even higher reserves or other limits for products/contracts that haven't been tested in a crisis yet or for products based on loans that could be dodgy if macro gets bad
-- improving transparency in consumer lending; ban or penalize consumer loans with negative amortization, high optionality or wrong-way risk (interest rates that are lower because they can spike when my credit gets bad for example)
-- tie mortgage lending to comparable rents as well as comparable purchases
-- centralized settlement and of most derivatives
-- some attempt to solve some agency / short term call option problems with compensation -- not an easy problem but banks can't have enjoyed being looted by their own traders these last few years
@michael martin: it is very difficult to know how correlated financial risks are going to be in a crisis. not all crises are going to be the same in this respect.
Posted by: babar at Jun 13, 2009 2:35:13 PM
Tyler --
Except that, at some point, you need to evaluate whether the materials you have to work with are so corrupted by politics, burdened with institutional inertia, and incapable of rising to the challenge of "gradual process improvements" and increased "regulatory accountability" that it would be better to start over.
As someone who has been under the regulatory supervision of the SEC for 20 years, I believe that institution just cannot be made functional with incremental measures. Its staff is either demoralized or incompetent, its bureaucracy is sclerotic, and the regulations it is charged with enforcing are either outmoded, ineffective, or counterproductive in many instances. I say shut it down, draw up a new blueprint, and start over.
But then again, I believe it would be better to throw away the entire federal tax code and start from scratch there, too. Go ahead, call me a dreamer. I'm used to it.
http://epicureandealmaker.blogspot.com/2009/06/these-dark-satanic-mills.html
Posted by: The Epicurean Dealmaker at Jun 13, 2009 3:51:57 PM
Great plan. The complete regulatory capture of these entities by the likes of JP Morgan, Goldman Sachs and so on has damaged their credibility, so one option for quieting the public is to gin up a new super-duper, no-really-this-time-for-sure regulatory agency. Still staffed by the same people who couldn't figure out that housing prices sometimes go down, still managed by the good old boys from GS to make sure that the bottom lines of the people that matter aren't threatened.
There is a better regulator in the form of the market. The Paulson bailout and the subsequent bailouts should not have taken place, and the rotten banking structures should have been allowed to fall.
Posted by: bbartlog at Jun 13, 2009 10:01:45 PM
@babar
You don't need to know how risks are correlated to know that they are correlated or to prevent them from becoming too correlated.
What we know with certainty is that leverage amplifies risk and that similar bundles of securities result in correlated risk. One way to reduced systemic risk is to limit leverage, but that raises costs without any offsetting benefits. Another way is to require comparably sized institutions to carry different bundles of risk. This also raises costs of but also has the benefit of making the system as a whole more resistan to shocks.
What we built in the '90s was a financial service industry zoo of the fattest juiciest pigs -- which made sense because everybody wants the tenderest cheapest pork. Until swine flu came along. Whoops. Maybe we should have raised some chickens and goats and cows too.
Posted by: Michael F. Martin at Jun 14, 2009 12:06:45 AM
@mm: i'm not sure what you are saying. it almost sounds like you are saying 'risk in the real economy should be diversified' rather than 'risks on bank balance sheets should be uncorrelated'. in the first case i don't see a way to deal with this through regulation (if it is even desirable) and in the second case i don't see how you determine this. for instance, who would have guessed that a banking crisis in the US would lead to the dollar drastically strengthening? i'm sure some people would have, but it's not what i would have thought.
Posted by: babar at Jun 14, 2009 8:08:54 AM
What I'm saying is that one can see with statistics whether buying and selling is getting correlated by looking at autocorrelations in price series without knowing in detail what's causing the correlation at the level of microstructure. Gaussian statistics = uncorrelated or perfectly synchronized; Levy statistics = in between. That much I think is pretty intuitive.
What I'm saying that I admit is more arguable is that we should probably try to keep things in between by building asymmetries into the risk profiles of our financial institutions with regulations. I take your point that this is difficult to do and likely to have unintended consequences. But if I have to choose between trying that and an alternative like simply raising capital requirements across the board, I know what I'd choose.
Posted by: Michael F. Martin at Jun 14, 2009 10:25:04 PM
Food for thought: Why not put the financial services industry into a chimera state to reduce correlations?
http://arxiv.org/PS_cache/nlin/pdf/0407/0407045v1.pdf
Posted by: Michael F. Martin at Jun 15, 2009 11:11:59 AM
There is no "systemic risk." There, however, bona fide, complete incompetents who have hoodwinked the taxpayer with this nonsense and "too-big-to-fail." Can the systemic risk regulation. Let'em fail.
Posted by: dwinds at Jun 15, 2009 5:12:53 PM
Regulatory "reform," as it's conceived by statists such as those currently in charge, suffers from a fundamental psychological disjuncture between the skills needed to get the regulatory job in the first place, and the skills needed to perform that job in the way all of us might like to see.
To get the job, you need to think inside the box---be a respecter of rules, procedures, regulations.
To do the job well, you have to think outside the box---be able to anticipate the next war rather than fight the last one, muster the imagination, creativity, and individuality of thought and judgment to swim against the 'go along, get along' tide that soon consume the lives of most of the folks in your office.
That is one reason, though far from the only one, that the regulatory panacea is so often disappointing. As Veronique de Rugy has pointed out, Bush actually expanded the regulatory state. Not much reason to think Obama's Fix will end up any differently.
The problem is not regulation (except in some fairly narrow and manageable areas such as capitalization ratios), but incentives, and as David Brooks and others have noted, this is an administration singularly lacking in people who understand the importance of fostering proper incentives.
Posted by: Voltaire in '08 at Jun 17, 2009 8:10:15 AM
Posted by: 童裝批發 at Aug 17, 2009 1:10:16 AM
