The following is an edited compilation of comments on Brad Setser's blog in December of 2006.  The two original threads are here and here.

There are always purveyors of doom predicting economic collapse.  Occaisionally, they may be right, but that does not mean they are worth listening too (broken watches, etc).

What is unique about this thread it that the commenters found the exact financial instruments that posed the danger, and guess the exact political choices that the Fed will find itself faced with, and even make pretty good guesses about how the Fed will actually react.

All of these commenters actually have their own blog which are still active.  I highly recommend adding these blogs to your daily news sources, as they've proven they are more on top of things than the NYTimes/Economist.

The commenters are:

Cassandra -  a hedge fund manager in Japan.

Mencius Moldbug - a software engineer on sabbatical after a dot com IPO, he spends most of his time reading and writing.

Brad Setser - an academic economist who has worked for the Federal Reserve and the Council of Foreign Relations

Steve Waldman - another amateur economist.

-------------------------------------------------------------

Steve Waldman

(Last month I couldn’t help but be amused that I was writing about CPDOs while writing c3p0. I like the whole Kate-Moss-thin credit spread thing. I’m quite convinced — probably by reading this blog too much! — that much of the structured credit revolution is about absorbing all of what, er, bulemic central banks seem determined to keep spewing.)

------------------------------

Mencius

If anyone has a better explanation of CPDOs than Steve, I would like to see the link! But frankly, it is not clear to me that the thing can be done. Thank you, Mr. Waldman.

The big picture in CPDOs is that they are a classic example of government failure. Nothing like a CPDO could possibly exist in an unregulated economy. Of course, that doesn’t mean the best choice for this far-from-unregulated economy isn’t to find some way of regulating the damn things out of existence. But fortunately this is not my job.

CPDOs exist because Nationally Recognized Statistical Rating Organizations exist. Rating agencies (Moody’s, S&P) are not private actors. They are granted enormous official authority. Imagining the market price of NRSRO status, sold with no questions asked - for example, to a company which could shake down its clients for ratings - is one way to conceive of the scale of this delegation of power.

In a transparent, professional system of government such as ours, authority of this kind must never be personal and arbitrary. Old John Moody, perhaps, could tell his customers that Joe’s Railroad was a shoddy outfit run by notorious shysters, whose bonds shouldn’t be touched with a ten-foot pole no matter how many payments they’ve made. If his successors rate JRX, they have to justify their result with some serious math. Nothing else would be compatible with “nationally recognized” status.

As Mises pointed out in the ’20s, excessive dependency on calculation is the general flaw in central planning. The planners are constantly being forced to calculate things that cannot possibly be calculated. Credit default probabilities, especially aggregate spread projections, are a classic example.

So Moody’s can’t issue a report saying that these CPDO things just don’t smell right. It can’t call Stratfor, get a ballpark number on the chance of an Israeli-Iranian war, and factor that into the probability of a generalized credit panic. It has to do what it does - run the things through its models. Which predict, as usual, future results from past performance.

And it’s not just that Moody’s has to do this. It’s that it can do this. Because it is effectively a government agency, it has transferred all of its risk for this behavior to the state. It is Uncle Sam that will take the hit if the models fail, and rightly so. Moody’s only existential risk is failure to comply with its own properly approved policies and procedures.

Fortunately, Uncle Sam is perfectly capable of insuring the risk of the models. He can, after all, print more dollars. Which will then be used to buy bonds - keeping those spreads svelte. Moreover, with this same mechanism, he can stimulate the economy, keeping the people who actually have to make their payments flush.

This is a perfect example of an expansionary ratchet. It is a political mechanism that causes immediate pain if the presses are stopped. Hyperinflation happens because the political cost of a liquidating recession exceeds the political cost of continuing around the spiral. Systems that increase sensitivity to default, like the system that the CPDO is gaming, make the spiral harder to escape.

In other words, the more CPDOs are outstanding, the more stress the financial system will suffer in the case of a sharp credit spread widening that overpowers the “stabilizing” reaction when the CPDOs automatically react by selling protection. The CPDO machine sets up a critical point, below which it is a “stabilizing” feedback loop that causes spreads to converge (as CPDOs gear up), and above which it is a thoroughly destabilizing one, that causes them to diverge (as CPDOs max out and fail).

It is, in other words, major “bubble skin.” The lovely old metaphor of a bubble, which really just means “disequilibrium,” can easily be extended to other materials than the usual soapy water. If your bubble is made out of latex, for example, it can get much bigger and sustain a much higher internal pressure. It is harder to pop, but it makes more noise when it does.

With CPDOs, and ultimately with the power of the printing press, the bubble is the size of the Hindenburg, its interior could easily be mistaken for the atmosphere of Jupiter, and its walls are Kevlar and nanotubes. As Steve says, it is very hard to break.

And it is very important to note that it is not just the personal whim of “bulimic CBs” that supports it - it itself enforces exactly that bulimia. The bubble skin is not really the critical point of the CPDOs. It is the fact that CBs cannot allow spreads to reach that critical point.

For all the hawkish talk, they will accept any level of consumer price inflation first. It is much easier to tweak the index again (maybe it could just be the GCPI, the Game Console Price Index, measured in triangles per second per dollar) and suffer the occasional human-interest story in the Times or Post about how the man on the street thinks prices are too high, despite the fact that there is no inflation.

So fasten your seatbelts, everyone. If I am even close to right, there are no brakes on this thing, and we are headed north in a hurry. It may be a happy 2007 indeed.

And if you care to read a retrospective description of “Kate Moss credit spreads,” rendered in lovely old 1930s prose on crumbling yellow paper, you could do a lot worse than an original edition of A Bubble That Broke The World, by Garet Garrett. I am too lazy to type it in, Dave Chiang style, but there is a bit about Brazilian railway bonds that is as resonant, poignant, and hilarious as you’ll ever get - it makes our own dear Cassandra look like Greg Ip. Financial journalism, like a lot of things, isn’t what it once was.

-----------------------------

Cassandra

Mencius, once again, the tapestry you’ve woven dazzles. And not that I want to argue on behalf of Darth Vader (or Bear Stearns) but one must ask the question - truthfully and unjudgementally - precisely what IS it that we really have to be frightened about in respect of financial innovation?? Is is Herstatt risk? Is it moral hazard? Is it implied leverage? Is it complexity itself? For skeptical (ok, even cynical) as I am, it seems to me that the more things issued backed by more things (as opposed to reliable but, at the end of the day, empty promises), even cleaving off the highly volatile, combustible or merely unsavoury bits for consumption by the overtly bold, certifiably insane, merely gullible, or Agent of OPM (other people’s money) Shooting-the-Moon is bona fide progress. The 151 IS in the the gritty bar, or perhaps some Trader Vic concotion, that at worst makes some unsuspecting girls panties in danger of being removed, but at least its NOT being sold in school vending machines or the corner shop. The innovation may create a logjam of cross-default risk, or back-up claims, but I as I look out at the landscape, I see risk transference that will, in the event of a large Shia trebuchet lobbing this or that over Israeli or Saudi borders, make the mess easier to clean up, the logjam easier to break, markets finding clearing prices quicker, and the certifiable idiots meeting their Darwinian (or Amaranthian) fate that much quicker. A Panglossian view perhaps, but one that The Most Sensible Man in The Fed, Tim Geithner, would probably agree with. In a cataclysmic event/contraction/revulsion/crash, most of the financial innovation will sting the reckless, the feckless and the marauders hardest and where it hurts most, with a large chunk broadly socialized - reasonably equitably - amongst pension funds, insurance companies, and banks or all persuasion. And call me an optimist if you must, but this is not an excessive price to pay, and economy-wide, will WILL be an improvement over the mayhem and treacle-like consequences to lending and economic activity in comparison to when it was concentrated all within - and had to be worked out by - but a few lending institutions.

--------------------------------------------------------------

 Mencius

Cassandra,

My worry is moral hazard.

It seems to me that what these instruments are doing is taking ordinary risk and moving it down the tail into Herstatt (systemic) risk. The Fed then uses its power of the press to convert Herstatt risk into dollar risk. The foreign CBs use same to convert dollar risk into euro risk, yen risk, even RMB risk. At the end of the day everything has the same risk, which means there is no risk at all - and everyone makes money.

Or, as it were, mints it.

What all these structured financial techniques are doing is essentially legalized counterfeiting. It is capturing the printing power of the Fed to produce notes that are, in practice, as good as dollars. It is fractional-reserve banking on steroids. The lesson of fractional reserve is that if you can underprice risk, you can turn (n) dollars into (n + m) dollars. This is now being done in a much fancier way, but the principle is the same.

Our financial system does not actually allow you to cut a CPDO into dollar-sized units and buy coffee with them. But suppose it did. Suppose that besides the usual green bills, which are dollars due now and paying no interest, you had blue bills, which were slices of T-bill, red bills, which were slices of AAA bond, and orange bills, which were slices of CPDO. All prorated by their current dollar value, so that a $2.45 espresso was a $2.45 espresso, whatever the color of your bills.

In a credit catastrophe, then, all the orange bills and maybe some of the red ones disappear, or at least change their face value. (Perhaps with some kind of electronic ink.) As a result, people suddenly have less money in their wallets, cash registers, etc, than before.

Obviously, the ideal situation is that the bills are evenly mixed across the economy. Everyone has all different colors of bills. This allows the destruction to be broadly socialized.

At maximum homogeneity, this starts to resemble an event in which all dollar bills with serial numbers divisible (for example) by 3 are destroyed. It is the reverse of Hume’s Archangel Gabriel. Perhaps the Archdemon Beelzebub.

You could make this a completely neutral (absent psychological factors) redenomination by also redenominating debts. Then it would be no different from, for example, the replacement of 1 million old Turkish lira with 1 new TL. A proper redenomination requires the rewriting of contracts, so that if you borrowed 1 billion old TL you don’t now owe 1 billion new TL.

But note how far we have gone from the real world of CPDOs. If money is widely destroyed, but the effect is distributed nonhomogeneously (if, indeed, broadly), and there is no formal way to calculate the transformation from pre-crash money to post-crash money, there is no formal way to, for example, impose partial debt forgiveness. If there is no debt forgiveness, you are privileging lenders over borrowers, which is politically improbable. But the lack of a Schelling point which enables sensible agreement over the proper level of debt forgiveness ensures a long period of confusion and political turmoil. Not good.

The 151, to use your metaphor, is in the school vending machines. It has been diluted considerably. The little tykes are getting Zima, not Bacardi. But it has still spread out across the global economy and stimulated one heck of a lot of demand. An instant mass hangover is certainly not good politics, and I don’t even think it’s good policy.

If there is anything Bernanke is an expert on, it is how dangerous and unpleasant a debt deflation scenario is. At least from an abstract intellectual perspective, it is unfair to blame him for his helicopter metaphor - what he was saying when he said that made perfect sense. But say it he did, helicopters he has, and if the alternative is mass liquidation, to the air they will take.

My feeling is that the feckless will not suffer. The bubble will not pop. The 1929 crash will never happen. The CPDOs will not default. The political cost is too great.

Bernanke is right in a sense. The round of liquidation that started in 1929 happened because the Mellons who encouraged it did not realize what actual liquidation meant after a decade of modern, efficient Fed credit expansion. They were used to liquidating the smaller and more private booms of the 19th century. Politically, liquidation could not have worked in the ’30s, and it probably shouldn’t even have been tried.

And we have gone way past the roaring ’20s here. The more widespread these instruments that push risk into the CB-protected tail become, the better their protection is. Whether it wants to or not (surely not), the Fed has become the FHFIC - the Federal Hedgefund Insurance Corporation.

This is simply lawlessness. You cannot go on letting Wall Street print money while denying that privilege to ordinary folks with their little mimeograph machines. It is incompatible with every ideal held by our political system. It is ultimately incompatible, I think, with the continued existence of that system.

My personal view about how to return from lawlessness to law, which certainly differs from that of most libertarians, is that the only way to do it is to find a formal way to ratify past lawlessness and eliminate future lawlessness. All the land on Earth, for example, has been stolen many times over. You cannot give Massachusetts back to the Algonquins or Londinium back to the Welsh.

This is why I feel that, rather than a deflating crash which tries to sting the “reckless, the feckless and the marauders,” the best thing for all concerned is a “flag day” solution in which everything they have managed to steal is confirmed as their formal property, and no further stealing is permitted. Which means simultaneously printing a whole heck of a lot of money, and transitioning to a post-paper financial system in which no more money can be printed.

Think of it as the financial equivalent of “truth and reconciliation.” It is a very counterintuitive approach. Just as it goes against all our senses of justice to let the murderers of Steve Biko off the hook, it goes against all our senses to let the sharks of Wall Street keep money they have created through borrowing the Fed’s printing press. But if your priority is results rather than revenge, I think it becomes a more attractive solution.

------------------------------------------------------------

 Cassandra responds:

Moldbug,
At last we nail it down. And I am with you on that page (though all the teasers are plausible possibilities for concern).

The game is certainly unfair. It’s NOT unfair from the point of view that some clever dicks have stumbled upon the more optimal solution before most. That is simply goood business. What’s “unfair” distributionally is that - at present - the umpires and arbiters are insisting its a “fair game”. That there remains high credibility to the authorities’ fiscal and monetary policies and so one is deemed potty to believe otherwise. This is why I despire Kudlow and bubblevision. This is why I condemn the Chinese and Japanese, NOT for sheer cheap manufacturing advantage or clever engineering & production efficiency (which must be disheartening for a US manufacturer in the best of times) but for not letting the interest rate market signal to particpants - particularly the masses - that the game is not true. The lie (not to mention the housing bubble) would be much more difficult to perpetuate with rates at 7% or more. The market actually would work if it were allowed and rates (both FX & IR) not so bent. Of course Congress is rather culpable here for we needn’t have run massive fiscal deficits, not coincidental to uber-loose moentary policy. In the absence of that, the lie is perpetuated, so rather than tell the people: “folks, money illusion (dilution inflation etc.) of X% p.a. is essential to continued good times, and so - in the public interest - we just thought you should know in case, like you errr ummm thinking of putting your savings in money markets, buying a US Savings Bond or Treasury Bond, or, in fact being a creditor for anything less a few days unless recompensed for the almost certain inflation that we are bound to see at some point in the future, absence the discovery of cold nuclear fusion. So folks do what you ratonally must to compensate”. After which the people can establish rational expectations based upon sound information, and not Radio Pyongyang blathher and drivel and apology after apology. This caveat is not dissimilar to the placing of photos of gruesome scarred black lung tissue upon cigarrette boxes. But at least “the people” would be on equal fotting, adjustments made, and the lie would be exposed.

---------------------------------------------------------------

Moldbug, like charming cassandra, I am captivated by the web ye weave, and very deeply in agreement with most of what you write. It is a great delight to read plainly written that Wall Street is making its money quite simply because its players have been permitted to print it, and that the distribution of wealth has become largely a game of cleverness or connectedness in accessing central bank printing presses, rather than a matter of conributing to the production of goods and services in the world.

That said, i think there’s a bit of a contradiction in your flag day proposal.  So long as the dudes with the printing press control the terms of mainstream debate (as they do now), there will be no flag day. They’ll just keep printing themselves money. There has to be a problem for there to be a solution. What should that problem be? A great inflation, as more people with a higher propensity to consume, get hold of a mimeograph?  Or a deflation, in which some of that mimeographed money just disappears, and whoever was holding it loses? Something else?

I don’t think we need to ratify anything, as all that printed money, whatever its color, is already effectively legal tender. I don’t think we need to string up anyone for their inflationary synthetic portfolios either. but i think we’ve enetered a game where 1) the accumulation of relative wealth is deeply decoupled from the production of absolute wealth, 2) the game won’t change until there is some kind of a crisis, and 3) the longer the crisis is put off, the worse the putatively ratifiable status quo will be when the crisis happens, and the greater the likelihood it ends up looking like the French Revolution rather than that 80s S&L thing.

One way or another, we will have a great inflation or a great deflation, or both (monetary hypervolatility). I’d like to see at least some deflation, so that some of the right people get burned, but as you say, justice may be too much to hope for in this life and vengeance tends to burn indiscriminately. Whatever form the crisis takes, I just wish we’d get it over with. Only then will there be any hope of a “flag day”, or of any reasonable financial architecture that prevents this kind of thing from happening again.

So, here’s to a crisis, early and soon. And Happy New Year!

----------------------------------------------------------

bsetser :

Cassandra — very interesting set of observations.

Personally, I am worried by:

a)  The implied leverage … and the need to hedge in bad states of the world in ways that can augment moves.
This is something that I think Geithner (my former boss) worries about as well — see his comments on falling margin in his last big speech on systemic issues. Less vol means bigger positions to get the same returns (or put differently, smaller spreads require taking on more leverage to generate outside returns), which is fine so long as the fall in vol is permanent and not temporary …

b)  The complexity itself … I don’t understand 1/2 the products out there, let alone how various folks hedge the nuclear bits of their portfolios, and I would bet I am reasonably sophisticated for someone doesn’t have skin in the game.  From a systemic point of view, complexity introduces more model risk (something that Waldman/ interfluidity nicely highlights)

c) Leveraged, correlated positions — think of all the folks who were long local EM markets (debt and equity) hedged by holding CDS (default protection) on EM $ bonds in May/ June. Those selling the CDS protection were in a position to honor their obligations (they were big boys looking for a yield pickup) but that didn’t change the fact that the leveraged community was all basically long the same stuff (local em assets) and deleveraging meant selling. The correlation between EM local currency sell offs and EM $ debt spread widening held generally speaking, but I do wonder if this is a case where correlations that look good in a backward sense will eventually cause truoble, especially in a world were real EM $ bonds are disappearing …

d) Some concentrated positions in some key market segments. See the FT article on correlation and the CDO market. To generate the tranches that real money investors want, the I-banks of the world ended up being structurally massively long the equity tranches … which they hedged in various correlation trades (as I understand it). In the past I have called this Rajan risk … smart folks take on nuclear stuff to make money selling the less nuclear stuff thinking that they can hedge it … but that generates model risk/ complexity.

e) Is there any risk that some of the prime brokers could get into trouble in a bad state of the world? I.e. one of their clients cannot hedge/ liquidate so the prime broker has to take over the position — and say it isn’t just one client but several who all have the same trade on and cannot get out in time. then prime broker then wants to hedge, and it has touble without moving the market and so on … I am thinking things up but this strikes me as at least possible. I am not 100% that all risk has been dispersed rather than concentrated — there are only so many people who understand certain kinds of complexity and some complex positions may be very concentrated (ok that isn’t necessarily a prime broker point but it is a concern).

------------------------------------------------------

Mencius

The reason recessions and depressions happen, or used to happen, anyway, is that lenders became more reluctant to lend. It was traditionally thought that lending to people so that they can pay off their debts was not sound banking. Of course, we have entered a new era, so this may no longer be true.

So debt deflation requires either a suppression of monetary creation, or at least a redirection of the new money away from existing borrowers.

In the era of classical bank lending, for example, political authorities could trigger debt deflation by raising reserve requirements, or otherwise making lenders reluctant to lend. Or lenders could exhibit such reluctance despite the desires of said authorities - as happened in Japan. They could take the cheap money and lend it to New York, for example.

But most of today’s credit expansion is driven by derivatives. This is a very different phenomenon.

The hallmark of a credit expansion is that the value of dollar claims which are technically mature at any time T exceeds the number of formal dollars in existence at T. In a free market, this is only sustainable when the monopoly supplier of dollars has a clear willingness to print as many dollars as claimants care to redeem.

Obviously, if dollars are a closed system (no printing allowed), and you have a claim for a dollar on an entity E which has issued identical claims for yD dollars, where D is the total number of dollars in the world, and y is greater than 1, your claim cannot possibly be worth a dollar. Whether E is George Bailey, Citibank, or the entire US financial system makes no difference. You are best advised to redeem your dollar at once and secrete it about your person.

But when there is a lender of last resort L that can provide an arbitrary amount of financing to E, there is no need - or almost no need - to redeem the claim. The only thing that can cause a default on the part of E is a breach between L and E. Since E has every incentive to avoid such a breach, your claim is secured by its competence and compliance, which in the case of most financial institutions is very, very high. You have no need to look at E’s own financial position, because what matters is not E’s position, but L’s assessment of it.

The result of this implicit insurance is that claims on entities which fit the description of E are priced higher, relative to dollars, than they would otherwise be. Another way to say this is that if the claims are not mature, they will be priced higher, relative to dollars which are not mature, than otherwise. In other words, these notes will exhibit “unnaturally” low spreads over T-bills.

And in still more words, this process is exactly equivalent to the printing of money. It allows the total value of credit to exceed the total amount of money. These implicitly insured claims to money become, in Misesian language, “money substitutes.” If the insurance is less than perfect, they may be discounted by some fraction, but this reduces net money creation only by that fraction.

The same can be said for the frequently-cited defense that credit expansion also produces liabilities, which must be paid back. Indeed. But if the issuance of new credit exceeds the retirement of old liabilities, the net money flow is still positive. And this is generally the case.

A system with constant monetary creation becomes dependent on the flow of new money. Unless you are a Third World country, the process is unlikely to be as simple as taking out new loans to finance old ones. The new money sloshes around and creates a general feeling of well-being and prosperity. Individuals make different decisions than those they would have made absent this flow.

The challenge in keeping this kind of unstable monetary system alive is to avoid the Scylla of self-reinforcing shutdown and potentially lethal hangover (debt deflation), and the Charybdis of self-reinforcing dependency and potentially lethal overdose (hyperinflation). As a historical rule, the result is always the latter, because it can be selected at any time, and it feels a lot nicer.

Note that consumer prices have very little to do with either of these phenomena. Both debt deflation and hyperinflation, if they get out of control, will inevitably affect consumer prices. But there is no guarantee that they will do so before they get out of control.

Looked at from first principles, this bank-oriented model of credit expansion is a very complex and temperamental system. But it is extremely well-understood, and it affords quite a number of practical levers for administrative management.

In the age of the derivative, however, administrative constraints on liquidity are much less straightforward.

Take, for example, the notorious CPDO. How do you fix a system that assigns an instrument which achieves guaranteed return by exponentially “doubling down” an AAA rating? (Props to Steve Randy Waldman for his excellent explanation of these diabolical devices.) Efficient financial markets have efficiently discovered and exploited the inherent and essential absurdity of stochastic risk modeling.

For example, Citigroup has the incredible and unmitigated gall to describe CPDOs as a “stabilizing” instrument. This is because when the yield spread of the lower-rated instrument protected by the CPDO (eg, some index of A-rated credit) widens, the CPDOs that protect it will gear up and sell more protection. Reducing, therefore, the market-assigned risk of the instrument. So the more CPDOs are sold, the less likely spreads are to widen - making each CPDO inherently safer! This is financial alchemy at its best.

In other words, it reduces the Fisherine concept of “stability” to its natural absurdity. If you define stability as the absence of volatility it is precisely correct. Since this is the way our financial system does define stability, it is indeed correct. And it certainly can’t be perturbed by any force short of actual reality.

So the $64T question is: when this system, which in normal English usage would be defined as the exact opposite of stable, reaches the point of size and fragility where it comprises the overwhelming majority of bank assets, and it will explode if a dog farts in Burma - what happens when the dog farts?

In other words, when CPDOs and similar explosives start going off like firecrackers in the ammo dump, and a lot of people realize that the value of their asset portfolios may be much lower than they thought it was, what does the Fed do to offset this?

Does it try to reopen the floodgates of classical bank lending, relaxing regulations so that profoundly unsound institutions can still lend? Does it try to invent some new infernal financial instrument to take up the slack? Does it buy assets directly with money it prints itself - possibly even the CPDOs themselves, on the grounds that it itself encouraged this gross financial hazard? Does it lend its own money directly? Or does it just let the bankruptcies pile up?