Lessons from the Nasdaq Clearing Default

Last month, the loss caused by the default of a single trader in a Nordic power spread contract cleared by Nasdaq Clearing consumed the entire €7 million contribution of Nasdaq to the default waterfall and then wiped out more than two thirds of the €168 million default fund of the Commodities Market segment of Nasdaq (the diagram on page 7 of this document shows the entire default waterfall for this episode).

Nasdaq explained its margin methodology as follows:

The margin model is set to cover stressed market conditions, covering at least 99.2% of all 2-day market movements over the recent 12 month period. In the final step of the margin curve estimation a pro-cyclicality buffer of 25% is applied.

The MPOR (Margin Period of Risk) for the relevant products is two days.

It also provided the following historical data:

• Prior to the default, the largest daily change in spread since 2011 was €1.6.
• 99% of all observations during this period were below €1.
• Nasdaq Clearing’s margin model for this particular spread was calibrated to cover changes in margin spread up to approximately €4
• The change in the spread on the default date was €5.56.

There has been a lot of excellent commentary on this episode:

• Craig Pirrong on his Streetwise Professor blog: here and here
• Jo Burnham at OpenGamma
• Amir Khwaja of Clarus

The episode highlights a number of important lessons about risk management that we knew even before this default happened:

• Electricity is a nasty asset both for traders and for risk managers. Equities are of course the nicest asset class in terms of return distributions, but electricity is really messy compared to even quite ill behaved commodities like natural gas.
• Spreads are a lot worse than the underlying assets from a risk modelling point of view. Over the last few decades, finance has figured out a lot about volatility, but we still do not know much about modelling correlation and dependence. (Ability to utter the magic word “copula” does not amount to knowledge).
• It is very hard to get much beyond 99.9% risk coverage using margins alone – not even in equities. A decade ago, I computed that margins set at eight standard deviations would cover only 99.95% risk (in terms of expected shortfall) in the Indian index futures market.
• To get to a AAA equivalent 99.99x% risk coverage would need other elements of the default waterfall like member and clearing house capital. Member contributions to CCP default funds are therefore at significant risk of loss.
• There is always a risk that all this waterfall would not be enough. If one trader can blow away two-thirds of the default fund of one CCP, some day somebody is bound to blow away the entire default fund of some CCP somewhere. With the increasing clearing of toxic asset classes like credit and power, I think it is only a matter of time before that happens.
• It is important to isolate different market segments with separate waterfalls and segregated default funds. In the Nordic power episode, even if there had been a delay or failure to replenish the default funds, it would have affected only the commodities segment, and Nasdaq would have continued to clear all other products. In India, I have always argued that this is extremely important since cash equities and derivatives trade on the same exchange: even if the derivatives clearing segment goes bust, cash equities must continue to trade.
• Fire sale liquidation of defaulted positions appears to be a growing problem at the clearing houses. I remember when Nick Leeson almost brought down Symex, they chose to wind down the position over several days to reduce losses. The US exchanges seem to prefer a quick auction. When Lehman failed, CME did an immediate auction of the entire Lehman portfolio of positions at that exchange, while LCH used its well oiled machinery for trading down defaulted positions. (Lehman’s bankruptcy examiner complained bitterly about the losses suffered by the Lehman estate due to this auction, but ultimately concluded that the exchange action is protected by law and could not be challenged). In the Nordic episode also, Nasdaq has displayed the American penchant for quick auctions.The Streetwise Professor discusses another power spread related default in the United States where the clearing house abandoned the auction when it “resulted in liquidation prices that were more than six times higher than the actual portfolio losses”. I have long argued that Indian exchanges should build up the operational and technical capabilities to trade down large defaulted positions, but I fear that these capabilities are quite deficient. Regulators also seem to be insufficiently concerned about this problem.

Posted at 6:07 pm IST on Mon, 8 Oct 2018         permanent link

Comments