Now for the really big one

by John Quiggin on September 19, 2008

While reviewing this post from 2002, foreshadowing a derivatives crisis like the current one, I found the following:

“At the end of 2002’s first quarter, the notional value of derivatives contracts involving U.S. commercial banks and trust companies was $45.9 trillion, according to the Office of the Comptroller of the Currency’s bank derivatives report. ”

The bulk of the exposure is in interest rate swaps, which are fairly well understood and seem to pose only modest risks in themselves. But there’s still around $1 trillion in more recent derivatives involving securitisation of various kinds of debts. This securitisation is sound only if the credit rating agencies have got their risk assessments right, which in turn requires that the accounts on which those assessments are based should be valid. A few years ago, when the market in debt derivatives was starting up, this assumption seemed safe enough, but now it looks a lot more dubious. The big danger is that defaults in the debt derivatives market could spread to the much larger interest rate derivatives markets.

As an update, the $1 trillion in credit derivatives has exploded to around $50 trillion. While less dramatic in proportional terms, the growth in interest rate swaps is actually more alarming, having reached around $300 trillion in notional values.[1]

It now seems pretty well certain that, as the quote above suggests, the chaos in debt derivatives will shortly spread to interest rate swaps.

There are two reasons for this. First, swaps are essentially bets on interest rate spreads and these have gone wild in the last week, with interest rates on Treasury notes dropping to zero while commercial paper is just about unsaleable at any price. Imperfectly hedged players in the market must be sitting on losses of several percentage points. Depending on how much of this there is, the implied losses could be anywhere from tens of billions to trillions. Crowdsourcing plea: anyone who has a better estimate is welcome to offer it.

Second, hedging only works if you can collect from your counterparties. This Economist story indicates that Lehmans was a big player, but no-one really knows who is owed money by them. And it seems certain that there will be large-scale failures among hedge funds in coming months.

It’s hard to see this crisis being resolved by normal commercial or regulatory means. The hundreds of billions tipped into the market by central banks yesterday is just a drop in the bucket compared to the sums at risk here.

Unless that is, all normal calculations are rendered irrelevant by a US government asset purchase on a scale that will make all past nationalizations look puny. How that will play out I have no idea. For example, will US-based ratings agencies take the step (automatic if it were anyone else) of downgrading US government debt?

fn1. Under normal conditions, the exposure associated with a swap is of the order of 1 per cent of the notional value. But (a) 1 per cent of 300 trillion is 3 trillion (b) conditions are not exactly normal right now.

{ 54 comments }

1

Kenny Easwaran 09.19.08 at 6:47 am

That last link is broken – apparently it’s got “ttp” where it should have “http”.

2

John Quiggin 09.19.08 at 7:01 am

Fixed now, I hope. Thanks for the alert.

3

Andrew Brown 09.19.08 at 7:13 am

Off topic, but I have worked out why your footnotes don’t work: you are doing Textile2 markup, and processing it with textile1

4

Lex 09.19.08 at 7:30 am

If everyone agrees that this market is stuffed, why is there not the option just to void it? Chuck the paper in the shredder and pretend like it never happened? None of these alleged hundreds of trillions of dollars actually exists, after all…

5

Zamfir 09.19.08 at 8:52 am

Lex, I am not sure, but I suppose that a lot of these swaps have been used as hedges for other positions. If some of those other positions are not swaps, companies will be left exposed to to much larger risks than planned.

I guess a total ‘void’ would have the same effect as an unraveling of the market.

6

abb1 09.19.08 at 9:36 am

Also some of the important states, like, say, Japan, China, or Saudi Arabia, might kinda unravel too. That can’t be good. Although, I understand, it already happened once to Japan in the late 80s, so they should’ve known better.

7

Lex 09.19.08 at 10:11 am

I think Saudi Arabia can probably cope – after all, how many diamond-encrusted Mercedes can one man eat?

8

Zamfir 09.19.08 at 10:52 am

Lex, you micht be surprised… I remeber from a few years ago that the saudi royal family went on a surprise holiday to Switzerland. There were not enough S-classes for rent in the country, so more had to be brought in from Germany.

The point is, I guess, that a single person can only need a few dozens Benzes. But a 30.000-member family can need a lot more than that…

9

stuart 09.19.08 at 11:07 am

I seem to remember an article somewhere several years ago talking about how across the board counterparty risk was systematically underestimated, or even completely ignored. One suggestion was that it the fairly unequivocal implicit guarantees of government that any major financial player wouldn’t be allowed to fold – as a worst case the company would be merged into another big player in some way so the intermeshing complex of risk transference that locks all the banking system together can’t be dragged down by a single bad choice.

This reminds me of a gambler using the Martingale system, it looks like a way of guaranteeing winning in the long run, but it will always fail in the end in reality (considering no one has infinite wealth playing against a casino with infinite bet limits).

10

John Quiggin 09.19.08 at 11:20 am

The martingale analogy has occurred to me also, Stuart.

11

Stuart 09.19.08 at 12:22 pm

One other thing I was thinking about as it was being discussed last night on TWIP – the ban on shorting strikes me as being high risk. My understanding is that in recent months a lot of hedge funds have been losing money for their investors, or having trouble finding a way to profit, and that one tool that is being used a lot is shorting stocks. If this means hedge funds lose the confidence of their investors, it seems possible there could be a run on some of these funds as they can’t unwind their positions fast enough, possibly causing a domino effect similar to the one hitting confidence in the banks. Clearly if something like this happens it could have a big impact across a wide range of asset markets which could pile more problems on the banks, insurance companies, etc.

12

Markup 09.19.08 at 1:48 pm

”it seems possible there could be a run on some of these funds as they can’t unwind their positions fast enough, possibly causing a domino effect similar to the one hitting confidence in the banks.”

That just may be a good thing in that it would allow for all the phantoms to either be exposed to harsh light or swept under some corner carpet never to be seen again.

13

Bloix 09.19.08 at 4:42 pm

As an insurance professional, my understanding of what is happening is that people who had no understanding of insurance got involved in insuring risk. Insurance is a seductive business – people give you large amounts of cash now in exchange for a contingent promise to pay in the future – and it’s always very tempting to charge premiums that are too low. That’s why insurance is regulated – to make sure that premiums are sufficiently high to pay claims. But Congress prohibited regulation of credit default swap transactions and what happened is what you’d expect: the insurers cut premiums, and justified their actions by systematically underestimating risk. Since there was no history of losses for these instruments, the risk analysis was entirely speculative anyway, and could easily be manipulated by those whose self-interest consisted in entirely in maximizing cash flow.

14

HH 09.19.08 at 5:12 pm

We seem to be alternating between the administration of morphine and laetrile as cures for societal cancer. Systemic trust collapse resulting from legitimization of lying as a career skill cannot be remediated quickly. It takes decades to drive it out of a corrupt culture.

The reason there is no academic field of “trust studies” is that the culture has normalized falsehood as a useful tool. Until this fundamental dysfunction is corrected, all we will see is variations on the resultant follies.

15

Kieran 09.19.08 at 5:16 pm

The reason there is no academic field of “trust studies”

There is, actually. A substantial literature, anyway.

16

dsquared 09.19.08 at 5:45 pm

I am far too emeshed in conflict of interest to comment on this, but would commend Bloix’s analysis in #13.

17

lemuel pitkin 09.19.08 at 6:12 pm

will US-based ratings agencies take the step (automatic if it were anyone else) of downgrading US government debt?

Why on earth should they? There is literally no conceivable situation in which the US government would be unable to cover its liabilities, given that they’re denominated in dollars.

18

Henry 09.19.08 at 6:21 pm

As Kieran says there is lots of interdisciplinary work on trust there (I work on this myself) – see e.g. the “Russell Sage Foundation program”:http://www.russellsage.org/programs/other/trust/.

19

virgil xenophon 09.19.08 at 6:56 pm

Having a background in both the brokerage AND the insurance industries in part of my former life, let me second Bloix in his analysis.
The mind-set of those who populate the insurance industry and those
who populate the brokerage and financial services conglomerates are not just worlds apart, but many parsecs worth of Galaxies apart. One set is akin to a town full of nothing but gunslingers, the other a town of
nothing but trappist monks (I exaggerate, but not by much.)

20

HH 09.19.08 at 7:08 pm

“The Trust initiative was brought to an official close in November 2005, and RSF does not expect to make any new awards for investigator-initiated research on trust. However, RSF may occasionally consider additional trust-related book proposals of exceptional merit.”

BFD. Where is “Introduction to Trust Synthesis?” or “Principles of Trust Engineering.” Why are there no actionable results of this feeble trickle of scholarship?

The answer is that contemporary society so totally corrupted by the worship of the lie that systematic engineering of right conduct is a profound taboo. Where liars rule, the student of trust plays the fool.

21

Henry 09.19.08 at 7:55 pm

As someone who has a piece coming out in the forthcoming capstone volume of the Trust project, I can testify that it is still alive.

As to the lack of work on the doubtless essential topic of Trust Engineering, I look forward to reading your own insights when you get around to publishing them (unless, natch, the system is _so_ totally corrupted by the Worship of the Lie that it fails to recognize the blinding truths that you have to offer). In the meantime, I’m off to mind me kids …

22

Stuart 09.19.08 at 7:57 pm

I can’t see the US government being in any critical trouble over this either to be honest – the worst case scenario I can see is the Presidents toys are taken away from him and the military is scaled back, and taxes have to rise. Or as Lemuel says, they can take the Zimbabwe route if that is too painful to contemplate.

23

abb1 09.19.08 at 8:15 pm

But the dollar now is 10% higher against the Euro than it was a couple of months ago. What’s the explanation of that?

24

lemuel pitkin 09.19.08 at 8:16 pm

We’re talking about liabilities equal to maybe 10 percent of GDP being absorbed by the feds. Some of that is balanced out by the assets also acquired in the bailout — or maybe all of it: the gov’t could very well end up making a profiit here, by essentially taking over a big chunk of intermediation from the private financial system while borrowing on much more favorable terms.

But assume a lot of that 10 percent has to be paid off from general revenues. Spread out over 20 or 30 years, the additional taxes (or inflation — boo! Zimbabwe!) required would be very modest. Really, talk about downgrading the US credit rating here is just silly.

25

John Quiggin 09.19.08 at 8:31 pm

Lemuel, I think it’s safe to say that under normal circumstances, a government ((with significant outstanding debt and a large structural deficit) that proposed to take on an open-ended liability of the order of 10 per cent of GDP would not retain its AAA rating.

As you say, a likely way out is inflation, and that is not good for bondholders, even if the required rate looks modest. If sovereign debt ratings ignored inflation risk, Zimbabwe would indeed be AAA. But I’m pretty sure Moodys and S&P will agree with you and not with me.

26

abb1 09.19.08 at 8:56 pm

If indeed higher rate of inflation is likely, why’s my TIPS’ market price gone down in the last couple of days? I don’t I understand this thing at all: the government is throwing money away, yet the dollar is steady and interest rates fall.

27

lemuel pitkin 09.19.08 at 8:57 pm

If sovereign debt ratings ignored inflation risk, Zimbabwe would indeed be AAA.

But presumably Zimbabe’s debt is denominated in dollars, or some other international currency?

Are you sure that the credit rating of US debt takes inflation risks into account? Why should it?

28

lemuel pitkin 09.19.08 at 8:58 pm

abb1-

Well, it could be that this was just an liquidity crisis after all, not an insolvency crisis, and the feds have just acquired a bunch of good assets at fire-sale prices. Or it could be, just wait.

29

abb1 09.19.08 at 9:01 pm

I mean, this means that demand for dollar-denominated stuff has not fallen. So, what exactly is it that foreigners are buying these days?

30

lemuel pitkin 09.19.08 at 9:06 pm

Put it another way, if the end result of the bailout is higher inflation, which it may be, the result is a lower value not just for US governement debt but for all dollar-denominated debt. So I still don’t see why it makes any sense to talk about a downgrading of US debt specifically.

(I also thought that credit ratings were specifically intended to measure the probability of default. But maybe I’m wrong about this?)

31

mpowell 09.19.08 at 9:16 pm

30: I am also curious about this. But if you consider a government bond denominated in the currency backed by that government, the likelihood of default could always be presumed to be zero. But in actuality, the market rate for these bonds will depend on some kind of expectation in the future value of that currency (ie inflation), right?

32

abb1 09.19.08 at 9:25 pm

I think government bond denominated in its own currency is always guaranteed. It’s the currency itself that might lose value relative to other currencies.

33

john c. halasz 09.19.08 at 9:37 pm

abb1:

The U.S. $ has appreciated a bit lately not only because foreign economies have tipped into recession faster than the U.S. economy, (at least, according to official GDP statistics), but because wealthy U.S. investors long since got out of the dollar and invested in booming foreign equity markets, but now are withdrawing funds from them to cover their distressed liabilities at home. (Have you looked at the Brazil index lately?)

34

dsquared 09.19.08 at 9:39 pm

No, plenty of countries have local-currency debt ratings worse than AAA. The idea is that as the debt burden grows too big relative to GDP, the implicit inflation cost of the monetary finance needed would be less painful than just defaulting. Lots of countries do in fact default on local currency debt.

35

dsquared 09.19.08 at 9:41 pm

(and in the case of the USA, this is not necessarily theoretical. There are plenty of people in the USA who believe that, under the guise of “Social Security restructuring”, the USA ought to default on billions of dollars of “non-marketable” Treasury bills, and I have occasionally, although not successfully, raised the question that if a state is prepared to bilk their own grandmothers, why on earth would I assume as an investor that they wouldn’t do the same to me?)

36

abb1 09.19.08 at 9:44 pm

the implicit inflation cost of the monetary finance needed would be less painful than just defaulting

more painful?

37

lemuel pitkin 09.19.08 at 9:48 pm

But if you consider a government bond denominated in the currency backed by that government, the likelihood of default could always be presumed to be zero. But in actuality, the market rate for these bonds will depend on some kind of expectation in the future value of that currency (ie inflation), right?

First of all, historically sovereign debt ratings have mostly been for bonds denominated in other currencies. Second, there could still be a default for bonds denominated in the government’s own currency if it simply chose not to honor them, e.g. after a revolution. Third, the credit rating of a bond and its market value can move independently, precisely because there are other risks beside credit risks.

Someone reading this must be able to provide a definitive answer whether credit rating is simply a measure of credit (i.e. default) risk, as I believe, or if inflation risk is independently taken into account. It is certainly true that inflation is taken into account as a predictor of default risk, but that’s different. Note however that the linked article (from the NY Fed) defines credit ratings as “assessments of the relative likelihood that a borrower will default on its obligations.” (Which, if true, means that Moody’s, S&P et al. agree with me and not John Q.)

Another question would be, how much inflation would be required to monetize all the new debt acquired as a result of the bailout? A back of the envelope calculation is that since current debt is at ~40 percent of GDP, an extra 2.5 points of inflation over ten years would get the real value of the debt back to the pre-bailout level. Sustained inflation rates much higher than that have been experienced by the US (and lots of other countries) without sovereign debt downgrades or other serious adverse consequences.

38

HH 09.19.08 at 9:49 pm

The near-term devaluation of the dollar will be nothing compared to what the hideous new taxpayer-funded cronyocracy does to our competitiveness. Note how eager Congress is to pour money into GM and the other auto makers without a whisper about firing the management that has brought them to the brink of bankruptcy. Electric SUVs anyone?

We have now decided to give the cast of corporate clowns and liars who ruined our economy full access to the treasury of the United States. National Socialism has triumphed.

39

soto 09.19.08 at 9:55 pm

To carry bloix and stuart’s examples even further into the absurd, I personally know an enormous Insurance company (NOT American) whose investment dept. was buying synthetic CDO’s! these are ce essentially structures that by Treasuries and then write a credit default swap(s) – In other words, an insurance company!!!!!

40

John Quiggin 09.19.08 at 10:40 pm

“A back of the envelope calculation is that since current debt is at ~40 percent of GDP, an extra 2.5 points of inflation over ten years would get the real value of the debt back to the pre-bailout level.”

You’re forgetting here that fully anticipated inflation will be built into the interest rates the US pays when it rolls over its debts. So, you need to keep on surprising the markets by 2.5 per cent. Three surprises would take you into double-digit range.

And of course, you can’t just cancel the inflation when you’ve got rid of the debt. Starting at 10 per cent it takes at least a decade to wind it back down.

41

lemuel pitkin 09.19.08 at 10:58 pm

Well, you’re right … unless nominal rates hit 0, which is possible. And depending on the term structure of existing debt — even fully anticipated inflation will reduce the value of existing bonds until they have to be rolled over. With ~50% of the debt in notes and another 10-15% in bonds, even anticipated inflation still gets several years to reduce the outstanding debt. And of course inflation isn’t fully anticipated.

The point is just that inflationary episodes comparable to ones the US has expeirenced in the relatively recent past would entirely wipe out the additional liabilities the US is taking on in this bailout. Assuming, again, that there are net liabilities at all — http://delong.typepad.com/sdj/2008/09/thoughts-on-the.htmlBrad DeLong, e.g., doesn’t think there will be.

42

John Quiggin 09.19.08 at 11:36 pm

Lemuel, the AAA rating is supposed to indicate that the chance of default in a given year is well below 0.1 per cent. Faced with the alternative of another decade or two of stagflation, do you really think it’s that improbable that the US might choose instead to stiff some class of creditors?

43

Charlie 09.20.08 at 6:40 pm

It strikes me that these are exciting times to be an economist.

44

Ginger Yellow 09.21.08 at 12:09 am

A few points.

1) I don’t mean to downplay the chaos that has and will be caused by the Lehman bankruptcy, but as far as I know the recent moves in bond rates aren’t that huge a deal in themselves. Most interest rate swaps in my world (structured finance) are between a fixed rate and a floating rate (eg Libor, Euribor, Eonia, Bank of England base rate). Therefore a move in Treasuries will only have an impact on existing swaps in so far as it affects the floating rate. There have been pretty huge fluctuations in Libor and similar rates over the last year and they haven’t caused a swap crisis.

2) Lehman’s collapse is causing big problems in both interest rate and currency swaps, in that lots of people who thought their positions were hedged now find they aren’t. That may not be too big a problem for operating companies, because they could just novate them to another counterparty, probably at a somewhat higher cost. But it’s a big problem for securitisations which don’t have a whole lot of spare cash lying around by design, and which will find their excess spread eaten up even if they can find a replacement quickly.

3) Credit ratings don’t take into account inflation, except in so far as it is an economic stress. The exact definition depends on the agency, but is either the probability of default or the expected loss by final maturity, according to the terms of the deal. It doesn’t matter what the money is worth, as long as it’s paid.

4) <blockquote?I personally know an enormous Insurance company (NOT American) whose investment dept. was buying synthetic CDO’s! these are ce essentially structures that by Treasuries and then write a credit default swap(s) – In other words, an insurance company!!

There’s a lot more to it than that. Most synthetic CDOs are effectively bets on correlation in the underlying pool of reference entities (typically investment grade companies, sovereigns or junk bonds). The value of the different tranches depends on whether people think systemic or idiosyncratic credit risk is more likely. The purchase of Treasuries (or other eligible investments) is just to park the investors’ cash safely and prevent negative carry. If there are defaults in the underlying pool, a proportional amount of the cash is used to pay the buyer of protection under the CDS (typically the bank which arranged it), while the same amount is written off from the bottom of the CDO’s capital structure. That’s a simplified picture, of course.

45

Dr Zen 09.21.08 at 8:38 am

America won’t be downgraded because there’s no way it will default. Your credit rating is a rating of how likely you are to pay. The US is nowhere near defaulting on any of its debts. John, it’s pretty close to unthinkable that the US would “stiff” any creditors this side of Armageddon.

Ginger Yellow’s analysis of the interest-rate swaps seems spot on to me but I think John is pretty much right all the same: the chaos is caused in the main by lack of information about who has lost and how much.

46

abb1 09.21.08 at 11:01 am

Reading this thread, everybody here is surprisingly cool about this thing, I must say.

Yeah, LP, I remember about 10 years ago I lent a few grand to my ex-wife, and she promised to pay it back soon; still waiting. Would you like to take this ‘chunk of intermediation’ off my hands and, perhaps, make a nice profit?

47

Ginger Yellow 09.21.08 at 12:48 pm

“Reading this thread, everybody here is surprisingly cool about this thing, I must say.”

Well, part of that is that there’s not much to be done about it until everyone figures out precisely what their exposure is and how much will be realised from the remains of Lehman. In many cases, counterparties will actually owe Lehman money. It’s not like Lehman were taking huge (relative to their balance sheet) one way bets on interest rates and currency movements. They were offsetting trades against other trades.

One other reason to be somewhat optimistic is that the fallout makes a central clearing house for derivatives almost a certainty, which should make the next counterparty collapse (after its creation) much more manageable.

48

J Thomas 09.21.08 at 1:02 pm

One other reason to be somewhat optimistic is that the fallout makes a central clearing house for derivatives almost a certainty, which should make the next counterparty collapse (after its creation) much more manageable.

By all means, we ought to start planning for the next collapse right away.

Let’s look for ways to make the next collapse run a lot smoother.

I’m not ragging on you, what you say makes a lot of sense. Just….

49

J Thomas 09.21.08 at 1:05 pm

John, it’s pretty close to unthinkable that the US would “stiff” any creditors this side of Armageddon.

Except maybe for people who have Social Security.

It looks to me like we have a bunch of people who say that SS will fail, who repeat that over and over, so that voters will be resigned to it and won’t get too upset when they divert money to something else and SS fails.

50

Martin James 09.21.08 at 4:35 pm

The key to remember is that the net sum of all financial assets is zero.
A bad debt write-off for a bank is a decrease in the liability and increase in the net-worth of a consumer and vice versa. All those derivative bets net out to zero also.

51

abb1 09.21.08 at 5:03 pm

The key to remember is that the net sum of all financial assets is zero.

It’s not net zero over time. Some say (if I understand correctly what LP was saying) that what is considered “bad debt” now may turn out to be “good debt” after all – later, when the dust settles. And then the US government will make a bunch of money and give us some.

I find this scenario… hmm… let’s say ‘highly unrealistic’ . Not because I know anything about economics, but just the way the world normally works.

52

Zamfir 09.21.08 at 6:27 pm

Martin James, I guess that would be relevant if the US was planning to buy literally ALL financial assets. Which, I guess, would make the US officially more communist than Lenin.

53

J Thomas 09.21.08 at 9:55 pm

The key to remember is that the net sum of all financial assets is zero.
A bad debt write-off for a bank is a decrease in the liability and increase in the net-worth of a consumer and vice versa. All those derivative bets net out to zero also.

Sure. But … suppose an insurance company collects premiums against the promise that it will pay for catastrophic health costs. But then we get an epidemic new disease that has a treatment which is very costly. Nobody planned that. Suddenly there are lots of liabilities that the insurance company can’t pay. If the insured victims write off the money and die, that’s a decrease in liability and an increase in the net-worth of the insurance company. A good thing for all the other kinds of victims who need the money. But….

There’s something screwy about this whole idea that a big company will accept risk in return for a steady income stream, and it will depend on the risks to average out. But far too often the risks do not average out.

The idea is to accept a whole lot of liabilities, far more than you can cover, on the assumption that you won’t have to cover very many of them.

Is there anything about this that bothers you?

54

Ray Davis 09.22.08 at 8:20 pm

“Reading this thread, everybody here is surprisingly cool about this thing, I must say.”

My startle reaction went numb about four years ago.

Comments on this entry are closed.