According to the latest data from the Bank for International Settlements, the central bankers’ central bank, the total amount of outstanding derivative contracts has declined from a 2012 peak of $700tn to about $550tn. To put this into perspective, the figure has fallen from just under three times the value of all the assets in the world to a little over twice the value.
The largest element is interest rate swaps followed by foreign exchange derivatives. Credit default swaps, the instrument at the heart of the 2008 global financial crisis, are now relatively small — if you can accustom yourself to a world in which $15tn is a small number. It is only slightly less than the US gross domestic product (a little more than $18tn in the final quarter of 2015).
Two banks, JPMorgan and Deutsche Bank, account for about 20 per cent of total global derivatives exposure. Each has more than $50tn potentially at risk. The current market capitalisation of JP Morgan is about $200bn (roughly its book value); that of Deutsche, $23bn (about one third of book value). From one perspective, Deutsche Bank is leveraged 2,000 times. Imagine promising to buy a house for $2,000 with assets of $1.
Before you head for the hills, or the bunker, understand that there is no possibility that these banks could actually lose $50tn. The risks associated with these exposures are largely netted out — that is, they offset each other. As you promised to buy the house in question, you also promised to sell it: though not necessarily at just the same time or price or to the same person. That mismatch is the source of potential profit.
How effectively are these positions netted? Your guess is as good as mine, and probably not much worse than those in charge of these institutions. We are reliant on their risk modelling but these models break down in precisely the extreme situations they are designed to protect us against.
You will not find these figures for derivative exposures in the balance sheets of banks nor do such exposures enter directly into capital adequacy calculations. The apparent lack of impact on balance sheet totals is the product of the combination of fair value accounting and the tradition of judging the security of a bank by the size of its credit exposure (counterparty risk) rather than its economic exposure (loss from market fluctuation).
The fair value today of an agreement that has an equal chance of you paying me £100 or me paying you £100 is zero. Since your promise to pay or receive £100 is marked to market at nil there is no credit risk: you cannot default on a liability to pay nothing.
Under generally accepted accounting principles in the US, you are allowed even to net out exposures to the same counter party in declaring your derivative position. This is not permitted under international financial reporting standards, which is why the balance sheets of American banks appear (misleadingly) to be smaller than those of similar European institutions.
The fundamental problem is accounting at “fair value” when that fair value is the average of a wide range of possible outcomes. The mean of a distribution may itself be an impossible occurrence — there are no families with 1.8 children, for example. And netting offsetting positions may also mislead. There is a large difference between being a dollar millionaire and having assets of $100m and liabilities of $99m.
Accounting practices provide an appearance of precision that may be a poor guide to a world characterised by multiple risks and radical uncertainty. The superficial information we have from balance sheets and capital adequacy calculations understates the scale of complexity and interdependence in the global financial system. Market participants are right to be sceptical, and nervous, about banks.
Get alerts on Central banks when a new story is published