×ãÇòtips

Regulators and financial institutions are still grappling with the counterparty risk measurement issues revealed by the 2021 collapse of Archegos Capital Management, the latest breakfast seminar heard.

Archegos was a long equity fund whose investments were concentrated on technology stocks. It defaulted on margin calls from creditors, most notably Credit Suisse, on March 26 2021 after its portfolio value fell around 30 per cent in four days.

Matthias Arnsdorf, Global Head of Traded Risk Modelling at J.P. Morgan, explained that the Archegos story highlighted three linked but distinct drivers of counterparty risk for banks: concentration risk, liquidity risk and wrong-way risk.

These elements, which are often conflated or misunderstood, he said, formed a perfect storm in the Archegos collapse, which ultimately cost Credit Suisse more than $5.5 billion.

The fund managers used total return swaps and derivative contracts to leverage the fund’s exposure with lending banks, leaving the firms vulnerable to severe losses when the trades turned against them.

By revealing the limitations of existing risk frameworks, particularly around counterparty exposure, the collapse prompted significant regulatory scrutiny and continues to haunt the financial services sector.

Countering the counterparty risk

Introducing his presentation, Dr Ansdorf acknowledged the tendency ‘to always fight the last war’ but said the case is still dominating how the sector and regulators think about counterparty risk. He pointed to the Basel document first published around 12 months ago as a valuable tool in the process.

“It is clear that exposure metrics should consider factors such as concentration, liquidity and wrong way risk. How should we think about these risks? How do we think about measuring them or even modelling them? Should they be modelled? If so, where and how? The Archegos default provides a very useful case study for addressing those questions.”

He outlined how Archegos’s portfolio exacerbated the three risks.

Concentration risk: Archegos’s portfolio lacked diversification, making it vulnerable to idiosyncratic market movements. When the market turned against several high-profile tech firms, the concentrated positions amplified the losses.

Liquidity risk: The fund’s portfolio lost around 60 per cent of its value, with nearly half of that loss occurring after the March 26 default, as traders attempted to sell the tech stocks. The trading volume in those shares was 5 to 30 times their typical levels.

Wrong-way risk: This occurs when a counterparty is more likely to default precisely when the bank’s exposure to them is greatest. In this case, Archegos’s deteriorating portfolio triggered margin calls that the fund could not meet, leading directly to its default. The resulting feedback loop turned a manageable loss into a systemic failure.

The bank’s stress tests and potential future exposure (PFE) models underestimated the eventual losses by billions.

“The interesting part in the risk metrics is that Credit Suisse’s margin was just 9 per cent of the net market value – around $1 billion. The potential future exposure was calculated at the 95th percentile over five days, equating to around 16 per cent of the market value before margin.

“Then there was also a stress calculation that was the severe stress was at 30% of the market value to around two to three billion net of margin. The actual loss on that portfolio was $5.5 billion, significantly exceeding any of the predicted risk metrics.

“Models must better account for the potential severity of market moves in illiquid portfolios with limited diversification,” he concluded.

Wrong-way risk, Dr Arnsdorf suggested, is possibly the most controversial dimension of the Archegos story. Such risk is difficult to assess accurately because most models focus on unconditional expectations. However, wrong-way risk increases the probability of large losses conditional on default.

Applying `conditional models’ to Archegos showed that expected exposures could rise to three or four times those estimated by unconditional methods. However, regulators and market participants face a dilemma: better-rated counterparties can actually suggest higher wrong-way risk, because their default implies an extreme market shock. Banks risk alienating seemingly ‘safe’ clients by imposing higher margins tied to conditional losses. Yet they risk being blindsided if those clients default in volatile markets.

He said: “The key lesson from Archegos is not just that risk models failed, but that institutions did not ask the right questions. A dashboard approach – incorporating multiple risk metrics, including both conditional and unconditional exposures, concentration levels, and liquidity impacts – is essential.”

Risk management, he warned, cannot remain static.

The presentation was part of Bayes Business School’s monthly ‘breakfast seminar’ series which brings together senior practitioners, policymakers and academics to discuss a key topic facing business or society.

, Professor of Mathematical Finance and Director of the Quants MSc cluster at Bayes, briefly outlined research into the issue of suitable dynamic models conducted by the Quantitative group at Bayes Business School.

She said: “The Archegos collapse underscores the critical need for risk models that go beyond marginal distributions and simplistic dependence structures. In the presence of wrong-way risk, it is essential to adopt models that can realistically capture tail dependence between exposure and counterparty default. Standard approaches often fail to reflect the amplified losses that arise precisely when market stress and counterparty vulnerability coincide. A robust framework for counterparty credit risk must incorporate structural features and joint tail dynamics to ensure a more accurate and resilient risk assessment.”