Derivatives datapoint of the day

By Felix Salmon
September 28, 2009
OCC -- page 9 of this pdf. Using second-quarter numbers for each year, I looked at the total nominal derivatives exposure of end users -- the people for whom derivatives are meant to exist -- and for dealers.

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

Here’s a little table I put together with numbers from the OCC — page 9 of this pdf. Using second-quarter numbers for each year, I looked at the total nominal derivatives exposure of end users — the people for whom derivatives are meant to exist — and for dealers.

The results are pretty startling: while end-users have pared their derivatives exposure to a seven-year low, dealers have increased theirs to yet another all-time high. And as the OCC notes, when we say “dealers”, we really mean four banks in particular: JP Morgan Chase, Goldman Sachs, Bank of America, and Citibank.

Oh, and did I mention? The amounts here are in trillions.

Year End Users Dealers Ratio
2003 2.6 62.4 24.0
2004 2.5 76.9 30.8
2005 2.5 96.2 38.5
2006 2.6 110.1 42.3
2007 2.6 138.1 53.1
2008 2.8 163.9 58.5
2009 2.4 187.6 78.2

What has happened in recent years that derivatives dealers now need $78 in nominal derivatives exposure for every $1 in end-user exposure? When Adair Turner talks about “profitable activities so unlikely to have a social benefit, direct or indirect, that [banks] should voluntarily walk away from them”, this is surely a prime example of what he has in mind.

When the OCC tells us that total derivative notionals are now above $200 trillion, we can’t really help but go blank: the number is so many orders of magnitude divorced from any conceivable reality that it’s almost impossible to work out what it could possibly mean. But clearly that kind of exposure wasn’t necessary a few years ago. So why is it now?

More From Felix Salmon
Post Felix
The Piketty pessimist
The most expensive lottery ticket in the world
The problems of HFT, Joe Stiglitz edition
Private equity math, Nuveen edition
Five explanations for Greece’s bond yield
23 comments so far

and why is it not getting smaller?

Posted by winstongator | Report as abusive

I would imagine Warren Buffet has already given us the answer. If two parties can enter into a contract, both book a profit, and both be paid bonuses based on that contract… it will happen.

Posted by Basho | Report as abusive

Do we have any way of knowing in the breakdown of such figures if the ‘easily offsetables’ within those numbers have increased during that timeframe? For example, perhaps a higher amount of the 78:1 is short term forex swaps that the debit and credit for 1 billion could be offset for a couple hundred grand?

How much of these derivatives are real claims?

Posted by Kevin | Report as abusive

notional is not exposure. the big 5 run a balanced book, see table 6

I’m certain the markets have properly discounted this fact.

Posted by csodak | Report as abusive

Thank you, jck. (Although not sure it will drown on the peripatetic cries of the uninformed posters.)

jck and quantacide:

Are you both personally willing to guarantee all the counterparties?

“Balanced” is a complete lie, because payout is only remotely possible with perfectly solvent counterparties.

Repeat amongst yourselves: AIG is solvent.. AIG is solvent..

Posted by Unsympathetic | Report as abusive


You use the combination “OCC” and “derivatives” in a manner that makes sense.

But there is still confusion. To me, the OCC is the Options Clearing Corporation, and ‘derivatives’ refers to equity ad index options.

Is there a good standard we can establish to differentiate?

Re jck and quanticide,

How do you explain the fact that bilaterally netted current credit exposure of the top five banks has also increased three fold over this period (Table 4) — during which time there has been no increase in derivative use by end users? Or Table 12 which shows that the top 5 banks have bought $400 billion more in CDS than they sold? This implies that contracts are offsetting each other less than “running a balanced book” would imply. (It’s pretty obvious that aggregate numbers can’t be used to demonstrate the existence of a balanced book.)

Posted by anonymous | Report as abusive

Give a Statistician/MBA numbers, and he/she can tell any story they want.

I’m (and I believe jck) not arguring that these banks are perfectly flat, but that the numbers are not as incendiary as the commenters purport. 400MM is still a large number, but it is not the 187,600MM SCREAMER.

The myopia that Unsympathetic espouses w/ his/her AIG comment leads me to believe most people read Felix’s post, and erroneously think because notional derivative amounts have increased at a faster rate for banks than for non-trading parties, there is a huge conspiracy here that is stealing government cheese from starving children.

Given that:
1. Bear Stearns, Lehman Brothers, and to some extent ML are no longer Credit Derivs dealers.
2. Capital requirements are now onerous for counterparties (anyone starting up has very prohibitive costs of financing on Credit Derivative trades).

I’m not too shocked to see end users down and dealers up.

400MM = 2,400MM

it is a feature of networked market that the outstanding notional is very large vs net notional held by end-users). the fact that end-users positions have not increased doesn’t mean that they haven’t traded and the trading alone is enough to generate large increases in bilaterally netted positions held by dealers.

Um, quanticide, the evidence of unbalanced books is in the hundreds of billions, not hundreds of millions. Your “statisticians can demonstrate anything” argument would be more convincing if you bothered to get the orders of magnitude in the data right.

Posted by anonymous | Report as abusive

anonymous, mea culpa. I can always count on the internets for an ad hominem attack.

Thank you for bringing this to light, Felix.

Well we are seeing some of the effects of our unwillingness to send folks at places like AIG to prison for fraud for writing what were essentially insurance contracts without reserves and without there being an insurable interest. There is no ability of counterparties to pay when disaster really strikes, and speculators can massively distort the market. We have been unwilling to regulate these products as the insurance products that they plainly are, which says to me that the 2008 blowup was simply too small to stir folks to action.

Thankfully credit default swaps values are slowly receding, but interest rate swaps are at unfathomable levels and growing fast. This at a time when interest rate uncertainty is very high due to the uncertain effects of massive Q.E. and deficits. We have no idea how these would behave if interest rates jump by a few hundred BP or even 1000 BP in the case of an inflationary flare-up.

Does anyone think Wall Street is properly reserved for any *major* interest rate event, even as extraordinary fiscal activities by governments and central banks make such an event likely? Guys like John Paulson are expecting an inflationary flare-up.

Posted by Dan | Report as abusive


My apologies that I sounded “ad hominem” to you. Irritation at inaccurate data references tends to bring out my snarky side — a definite danger of participating in comment threads.

Posted by anonymous | Report as abusive

Is there any doubt as to why derivatives are traded?

Banks make money selling the contracts – and don’t need to bother with considerations of the risk involved. Specifically, in calendar Q209, banks made $5.2B selling this useless paper amongst themselves. cle/ALeqM5ju66OABjQrBVZVWFtIUfgxL4s2ygD9 AUF68O0

Fleece the taxpayer, book a profit!

Posted by Unsympathetic | Report as abusive

The most useful bit is the relationship between end users and among the dealer commmunity.

If anyone wonders why the pace and enthusiasm for reform is so tepid, they merely need to look at your chart. Derivatives are too massive to ignore, and too terrifying to politicians to reregulate quickly. Regulating core banking is easy, but most of the banks activites now are hedge fund(ish) and much more difficult to regulate properly. Since the banks profitablity is so inextricably linked to derivatives markets no one dares try to cut that Gordian knot at this time, lest we have another AIG counterparty calamity.
(My simple solution would be to declare all the trading units at banks as hedge funds, strip banks of the right to run hedge funds, spin them off, and let them compete in the hedge fund arena without government guarantees)

A good chunk of the derivative notionals are associated with the underlying cash instruments (i.e hedges)

The naked spec positons via derivatives is also a huge chunk of the notionals. (think commodities books – Phibro for example).

The banks claims that they’re flat is probably not too far off the mark. The major risk then is counterparty exposure, and since the major players are TBTF, at the moment the banks have little incentive to reduce that risk.

Politicians and regulators are praying that higher capital charges will be an effectively blunt instrument to use while the scalpel needed to amputate trading/ core banking functions (Glass/Steegal II/,derivatives exchanges, size limitations) are developed

One quibble about the OCC data. Bear in mind this only reflects the top US dealers. It doesn’t include DB, CS, non US HSBC, Barclays, Soc Gen, for example.

Posted by michaelc | Report as abusive

…so, jck & quantacide, if one of the larger players goes down (Wells Fargo or BAC will do), it’s all going to work out okay? That is, American taxpayers won’t have to plug in another few hundred billions of dollars a la AIG?

These derivatives are truly a Ponzi scheme that allows the dealers–personally & institutionally–to get rich on the runup, and the public to pay the tab when (surprise) the CDS’ are found to be worth nothing.

I’m with Dan on this one.

Posted by Lilguy | Report as abusive

Traders, Guns and Money. -unknowns-derivatives/dp/0273704745

Derivatives help firms avoid taxes & regulation (leverage and synthetic exposure to a forbidden asset class.) They also add a layer of opacity that ensures the clients have no idea of how they are getting raked.

That in a nutshell is why derivatives will continue to explode in volumes.

Posted by Charles Swann | Report as abusive

Notional amounts in derivatives provide very misleading figures. Let’s say two firms trade a call spread with each other. Bank 1 buys a 100-105 spread for $2 and Bank 2 sells. The actual market risk here is small (range of outcomes [-2, 3] for Bank 1; [-3,2] for Bank 2], but notional amounts usually report absolute values of strike prices $205 or $410 (if double counted). So $3 of risk gets reported as $205 or $410 of derivatives.
I agree that counterparty risk is what you really have to worry about on a system level. What if one side can’t pay? The problem is notional derivative exposure isn’t a good way to assess this risk.

Posted by JackL | Report as abusive

Correlation Mr. Peabody?

According to a study reported in the WSJ, the incomes of the wealthy have become far more variable than incomes for the rest of Americans.

They find that both growth and declines for the top 1% from 1982 to 2006 were more than twice as volatile as the comparable numbers for all taxpayers. The wealthiest of the wealthy had even more volatile incomes, with the top one-10th of 1% experiencing volatility of more than four times the average.

The word ‘nominal’ is used for a reason. It does not — repeat not — represent the amount anybody owes anybody else. What is owed is a small fraction of that amount.

Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see