Central Counterparty Risk Management: Margin, Default Waterfalls, and the 2018 Nasdaq Clearing Default

Central counterparty (CCP) risk management is the set of tools a clearing house uses to keep paying the winners on every contract even when a clearing member fails: entry standards for members, margin that covers most price moves, a default fund that members pre-pay, the CCP’s own capital, and rules for closing out a defaulter’s positions. The international standard for all of it is the Principles for Financial Market Infrastructures (PFMI), published by the committees now known as CPMI and IOSCO on April 16, 2012.

In short: a CCP does not make risk disappear. It concentrates it, measures it every day, and allocates any loss in a pre-agreed order called the default waterfall: first the defaulter’s own margin and default fund contribution, then a slice of the CCP’s own capital (skin in the game), then the default fund contributions of the members that did nothing wrong, then extra calls on those members. In September 2018, a single Norwegian power trader who was his own clearing member at Nasdaq Clearing lost more than his collateral, and about two-thirds of the commodities default fund, €107 million, was used to cover the gap. That case is now the standard teaching example of how the waterfall works, and where it can be thin.

This article goes deeper than how clearing houses work, which explains novation and the basic idea of a guarantee. It focuses on how professionals at CCPs, clearing members and regulators size and test the defences.

A common misconception: “the clearing house guarantees everything, so there is no risk”

A CCP guarantees performance to its members. It does so with resources that mostly come from those same members. When losses exceed what the defaulter posted, surviving members pay through the default fund and, beyond that, through assessments. A clearing member therefore carries two exposures to the CCP: the obvious one (its own positions) and a contingent one (other members’ defaults). Its customers, through their futures commission merchant, are one step further away but not insulated from the system as a whole (see how futures commission merchants work).

The useful question is not “is there risk?” but “how much loss can this CCP absorb, from whom, in what order, and how fast?”

The standard: what the PFMI require of a CCP

The PFMI contain 24 principles. The ones that shape day-to-day CCP risk management are:

Principle	Requirement for CCPs
4: Credit risk	Cover current and potential future exposure to each participant fully with a high degree of confidence. Hold additional resources for extreme but plausible stress: the default of the two participants (and affiliates) creating the largest exposure for a CCP that is systemically important in multiple jurisdictions or has a complex risk profile; the one largest for all others
5: Collateral	Accept collateral with low credit, liquidity and market risk, with conservative haircuts and concentration limits
6: Margin	A risk-based margin system; initial margin should meet a single-tailed confidence level of at least 99% of the estimated distribution of future exposure
7: Liquidity risk	Enough liquid resources, in all relevant currencies, to settle on time if the participant with the largest liquidity obligation defaults (the more systemic CCPs should consider covering two)
13: Participant-default rules	Clear rules and procedures that let the CCP act quickly to contain losses and keep meeting its obligations
18: Access and participation	Objective, risk-based membership criteria, monitored on an ongoing basis

These requirements are known in the industry as “cover 1” and “cover 2.” National rules implement them. In the US, the CFTC’s Part 39 rules for derivatives clearing organizations require margin at a 99% confidence level, with a liquidation period of at least one day for futures, and require systemically important clearing organizations to meet a cover 2 standard. In the European Union, the EMIR regulation requires every CCP’s default fund to withstand the default of the member with the largest exposure, or of the second and third largest if that sum is larger, and requires a CCP to commit a slice of its own capital (at least 25% of its minimum capital, under the supporting technical standards) ahead of the non-defaulting members’ contributions.

Layer 1: who gets in

Membership criteria are the first line of defence. Typical requirements include minimum capital (CME Rule 970 sets $5 million for its clearing members, or $5 billion of Tier 1 capital for a bank), operational capacity to meet intraday margin calls at short notice, and participation in default management, including bidding in auctions. The 2018 Nasdaq default turned on this layer: the defaulter was an individual.

Layer 2: margin

Initial margin is collateral sized to cover the loss on a member’s portfolio between its last good payment and the moment the CCP has closed it out (the margin period of risk). Variation margin settles each day’s gains and losses in cash. The mechanics for an individual account are in how futures margin works; scanning-range models used by CME are in SPAN and futures margin models.

Worked example: a 99% initial margin and its limits

A CCP holds a member’s $100 million futures position whose daily volatility is estimated at 2%.

One-day 99% one-tailed margin, assuming normally distributed returns: 2.326 x 2% x $100 million = $4.65 million.
With a two-day margin period of risk, scaled by the square root of time: $4.65 million x 1.414 = $6.58 million.

Two things follow. First, by construction a 99% model will be exceeded on roughly one day in a hundred. That is why margin is backed by further layers. Second, the model is only as good as its volatility estimate and its assumptions about correlation and liquidity. Concentrated positions in thin markets take longer to close out than the margin period assumes, which is why CCPs add concentration and liquidity add-ons.

Margin is procyclical

Margin models that react to volatility raise requirements in a crisis, which is when members can least afford it. The Basel Committee, CPMI and IOSCO’s Review of margining practices (final report, September 2022) found that initial margin at CCPs rose by about $300 billion over March 2020 and that variation margin calls peaked at $140 billion on March 9, 2020. Managing that trade-off, between protecting the CCP and draining members’ liquidity, is one of the central design choices in CCP risk management. The same dynamic appears in the Treasury market in the Treasury basis trade.

Layers 3 to 6: the default waterfall

When a member defaults, the CCP takes over its positions and works through its resources in order. CME Clearing’s base financial safeguards package, which covers futures and options, illustrates the layers. From CME Group’s annual report on Form 10-K for 2025, as of December 31, 2025:

Order	Resource	CME base package
1	Defaulter’s performance bond (margin) and guaranty fund deposit	Varies by member
2	CME’s designated corporate contribution (skin in the game)	$100.0 million
3	Non-defaulting members’ guaranty fund contributions	$8,306.1 million
4	Assessments on non-defaulting members	$22,841.7 million

The 10-K explains the assessment limit: CME can assess non-defaulting clearing firms 275% of their existing guaranty fund requirements for a single default and up to 550% for multiple defaults. The assessment figure in the table is exactly 2.75 times the guaranty fund, and the amount actually collectible would be lower because defaulted members cannot pay their own assessments. Aggregate performance bond deposits for both CME safeguard packages were $347.3 billion at the same date. CME resizes the guaranty fund regularly, so its financial safeguards page will show different figures on other dates.

Worked example: allocating a default loss

A hypothetical CCP has an $8,000 million default fund. Two members default together. Their margin has already been used, and they contributed $900 million and $600 million to the default fund. The CCP’s skin in the game is $100 million. The surviving members’ contributions total $6,500 million. Consider a member that contributed 2% of the survivors’ fund, or $130 million.

Remaining loss after defaulters’ margin	Defaulters’ fund (1,500)	CCP capital (100)	Survivors’ fund (6,500)	Assessments	Our member’s loss
$1,200m	$1,200m used	0	0	0	$0
$4,000m	$1,500m (all)	$100m (all)	$2,400m (36.9%)	0	$48m
$12,000m	$1,500m (all)	$100m (all)	$6,500m (all)	$3,900m	$130m + $78m assessed = $208m

All figures computed in Python. In the third row, the member loses its entire $130 million contribution and is then assessed $78 million. With two defaulters, a CME-style cap would be 550% of its contribution ($715 million); even the single-default cap of 275% would be $357.5 million. Two features stand out. The CCP’s own $100 million is small next to members’ $6.5 billion; and the member’s loss depends on other firms’ risk-taking, which is why clearing members scrutinize CCP margin models and each other.

Layer 7: default management

A CCP’s job after a default is to return to a matched book quickly. The usual sequence:

Declare the default under its rules and take control of the positions.
Hedge the largest risks so the portfolio stops losing money.
Port client positions and collateral to healthy clearing members where possible.
Auction the remaining house portfolio, usually to clearing members, sometimes split into lots.

The 2008 Lehman Brothers default is the case where this worked. The BIS Quarterly Review (December 2018, “Two defaults at CCPs, 10 years apart”) notes that Lehman’s portfolios at CCPs worldwide, with one exception, were auctioned, liquidated or transferred within weeks without exhausting Lehman’s collateral. Its London-cleared interest rate swap portfolio, about 66,390 trades with roughly $9 trillion of notional, used about a third of the margin held.

Auction design matters. If few members bid, prices are poor and losses larger. CCPs therefore require members to participate, and some subordinate (juniorize) the default fund contributions of members that bid badly or not at all, so they bear losses first.

Case study: the Einar Aas default at Nasdaq Clearing, September 2018

The position

Einar Aas was a Norwegian power trader who held futures on Nasdaq Oslo, cleared by Nasdaq Clearing AB in Sweden. Unusually, he was a direct clearing member in his own name: an individual, not a firm. According to the BIS, he had bet that Nordic and German power prices would converge. Weather forecasts and a change in German carbon policy pushed them apart.

The move

On September 10, 2018, the spread moved by €5.56, against a largest prior daily change since 2011 of €1.60, according to FIA’s account (“Deconstructing a Default,” November 2018). That is about 3.5 times the largest move in seven years. Aas failed an intraday margin call that day, and default procedures began early on September 11.

Why margin was not enough

The BIS described Nasdaq’s margin model for this market: it required 99.2% of the largest two-day market move over the previous year, plus 25% of the largest two-day move that year, and then granted a 50% correlation offset, on the assumption that German and Nordic prices would keep moving together. Aas was not asked for extra margin even though his position was large relative to the Nordic power market, whose volumes had been shrinking. In FIA’s account, ICE Clear U.S.’s Hester Serafini said ICE “would have actually charged a significant add-on because of the size of the position.” A spread position margined as if its legs offset each other is exactly the kind of position that fails when correlation breaks.

The waterfall in action

Nasdaq invited four members to bid in a closed auction for the portfolio. The BIS reports that the winning bid left a loss of €114 million beyond Aas’s collateral. Nasdaq’s Form 10-K for 2018 gives the allocation in dollars: a $133 million loss, of which the first $8 million came from Nasdaq Clearing’s junior capital and the rest from commodities clearing members’ default fund contributions, pro rata.

Computed in Python from the euro figures:

Nasdaq’s junior capital (skin in the game): €7 million.
From the €166 million commodities default fund: €114 million - €7 million = €107 million, or 64.5% of the fund.
A member that had contributed 3% of the fund (€4.98 million) lost €3.21 million, for a default it had no part in.

Members replenished the fund by September 17, according to FIA. Nasdaq’s 2018 Form 10-K adds that Nasdaq Clearing replenished the $8 million of junior capital it had used, increased its commodities junior capital by a further $14 million, and that in November 2018 Aas entered a consensual arrangement with creditors under which any recoveries go first to the default fund participants. Nasdaq also commissioned Oliver Wyman to review its margin, risk management, governance and membership requirements.

What regulators found

Finansinspektionen (FI), the Swedish regulator, issued Nasdaq Clearing a warning and an administrative fine of SEK 300 million, announced January 27, 2021. FI found that “the requirements the firm placed on the clearing members’ financial and operational capacity have been insufficient,” that “an incorrect calculation has resulted in the firm issuing insufficient margin calls,” and that Nasdaq Clearing breached EMIR’s investment rules by keeping its own funds in derivatives “much too long after the default event.” Two days later, FI clarified that it had not evaluated or established a causal link between the default losses and the margin calculation error.
ESMA, the EU securities regulator, published CCPs’ Membership Criteria and Due Diligence (ESMA70-151-2869, January 7, 2020). It surveyed all 16 authorized EU CCPs. None then had an individual as a clearing member; twelve prohibited it, four did not rule it out. ESMA concluded that EMIR’s definition of a clearing member does not exclude individuals and chose not to seek a ban, instead tightening expectations on admission criteria through a new Q&A and reviewing how CCPs monitor members over time. Its later peer review on due diligence of clearing members (published December 2023) noted that the Swedish authority still had to close out all the recommendations it had made after the default.

The lessons professionals drew

Membership standards are a risk tool, not a formality. An individual with no third-party oversight and a concentrated book is a different credit from a regulated bank.
Concentration and liquidity add-ons matter more than the headline confidence level. The position was large relative to a shrinking market.
Correlation offsets fail in exactly the scenario they are meant to cover. Spread margin should be stressed for the spread breaking, not only for both legs moving.
Skin in the game was small. The CCP’s €7 million absorbed about 6% of the loss; members absorbed the rest. FIA’s account of the post-default debate quotes industry participants arguing both sides: Eurex Clearing’s Matthias Graulich said a CCP first-loss layer of 10% of the default fund would cost his firm about two years of profits, while clearing members argued that a loss that falls mainly on survivors is “not defaulter pays.”

The skin-in-the-game debate

Clearing members want CCPs to hold more of their own capital ahead of the default fund, so that CCPs have stronger incentives to set conservative margin. CCPs argue that their capital is small next to member resources by design, that members control the risk they bring, and that large CCP capital could create moral hazard among members. The numbers frame the debate:

CCP layer	Size	Next layer
Nasdaq Clearing commodities junior capital, September 2018	€7m	€166m default fund
LME Clear skin in the game, end of 2021, and default fund from public disclosures (per OFR Working Paper 24-09)	$22.75m	about $1.1bn default fund
CME base package corporate contribution, December 31, 2025	$100m	$8,306.1m guaranty fund

In each case the CCP’s own layer is roughly 1% to 4% of the member-funded layer behind it. The OFR paper on the LME nickel episode, by John Heilbron (December 2024), makes a related point: when a CCP avoids losses by cancelling trades rather than running its waterfall, its skin in the game is never at risk, which may weaken its incentive to monitor members. That episode is covered in the LME nickel crisis of 2022.

How professionals manage CCP exposure

Clearing members and large users treat CCPs as counterparties to be underwritten:

Replicate the margin model. Members ask CCPs for enough model detail to reproduce margin calls and forecast them under stress.
Read the quantitative disclosures. CPMI-IOSCO public quantitative disclosure standards require CCPs to publish figures on margin, default resources, stress test results and backtesting breaches every quarter. Breach counts and the size of the largest breach relative to the default fund are early warnings.
Stress the contingent exposure. A member’s potential loss is its default fund contribution plus the assessment cap, for example 275% to 550% of the contribution at CME. That number belongs in the member’s own capital and liquidity planning.
Prepare for auctions. Members staff and rehearse default management, since they are expected to bid on a defaulter’s portfolio at short notice.
Watch procyclicality. Liquidity plans should assume initial margin can jump in a stress, as it did by about $300 billion across CCPs in March 2020.

What can go wrong

Model failure. A move far outside history, or a correlation that breaks, can exceed margin by a wide margin, as in 2018.
Concentration. A position large relative to market liquidity takes longer to close out than the margin period assumes.
Wrong-way risk. A member whose own creditworthiness falls as its positions lose (for example, a producer hedging its own output) is riskier than its margin suggests.
Liquidity shortfalls. A CCP can be solvent and still lack same-day cash if a large member fails to pay variation margin and collateral cannot be converted quickly.
Multiple defaults. Cover 2 is a planning standard, not a ceiling on reality. Several members can fail together in a systemic crisis.
Non-default losses. Investment losses, custody failures, cyber incidents or operational outages can hit a CCP’s capital without any member default. FI’s finding that Nasdaq Clearing kept its own funds in derivatives after the default is an example of investment risk at a CCP.
Tools outside the waterfall. Trade cancellation, as at the LME in 2022, can protect the CCP and some members while imposing losses on others.

The short version

CCPs manage default risk in layers: membership standards, margin calibrated to at least 99% confidence under the CPMI-IOSCO principles, default funds sized for the default of the largest one or two members, a small slice of CCP capital, assessments on survivors, and default management that hedges and auctions a defaulter’s book. CME’s base package at December 31, 2025 combined a $100 million CME contribution, $8.3 billion of member guaranty funds and $22.8 billion of assessment powers. The 2018 Einar Aas default at Nasdaq Clearing showed each layer at work and where they can be thin: a 50% correlation offset and no concentration add-on on a spread position, an individual as a direct member, €7 million of CCP capital, and €107 million (64.5% of the commodities default fund) taken from members who had done nothing wrong. Sweden’s regulator later fined Nasdaq Clearing SEK 300 million, and ESMA tightened expectations on membership criteria.