With the Fed and the PRA imposing a combined fine of $440m[i] on Credit Suisse / UBS regarding the risk management failures associated with Archegos’s March 2021 blowup, it is worth revisiting the more inexcusable ones.
This article is based on the 172-page special committee report into Archegos[ii], rating agency reviews, news, and the two supervisory fine notices. I provide examples and more colour on the five themes listed by the Fed:
Background
Archegos was the family office run by Bill Hwang, a New York-based hedge fund manager whose blowup caused at least $10.3bn losses to a handful of prime-broker (PB) banks in early 2021.
Importantly, it was a family office because Mr Hwang was barred by the SEC from managing client money after settling an insider trading claim. Archegos was established with $100-200m capital in early 2012 and, by mid-March 2021, had about $10-20bn under management. Thanks to the leverage obtained from its PB banks, its exposure was $100bn in a concentrated basket of Chinese tech firms, and US media conglomerates – most notably Viacom CBS – right before its blowup[iii].
Naturally, losses were not evenly distributed[iv]. Between 25 and 29 March 2021, Credit Suisse was slow to unwind the unhedged positions and took the brunt of the losses at $5.5bn. Arguably, the losses from this single event[v] irreversibly locked Credit Suisse’s path to the UBS absorption in June 2023.
Failures revisited
1. Failure to resolve internal limit breaches
The Credit Suisse Risk department set a $20m potential exposure (PE)[vi] limit as well as a $250m stress scenario limit for Archegos; they also added Archegos in 2018 to the “Watch List” to have closer monitoring. Despite this, since early 2020, the PE and the stressed exposure had breached the limits almost every week and gone unresolved until its blowup in March 2021.
During this period of breaches, Risk and the business discussed mitigation tactics. None were implemented. On the contrary, the business requested that Risk dig its head in the sand by forgoing the “Severe Equity Crash” scenario and using the more lenient “Bad Week” scenario[vii] to stress positions. Business argued that Archegos's position was on liquid and large-cap companies. Risk acquiesced.
Notably, Credit Suisse updated its PE methodology in January 2020. This led both 1 and 2 lines of defence (LoD) to discount the metric’s reliability and meaningfulness[viii].
At the same time, Internal Audit found that limit excesses were not remediated quickly enough. On average, limit excesses had been opened for 47 days for active excesses and 100 days for passive excesses. The audit team concluded that the lack of an established timeline for excess remediation contributed to the delays.
On top of that, Archegos’s Q1-2020 performance had been so poor that its NAV dropped precipitously from $3.5bn to $2.0bn in April 2020. This triggered a termination event, yet the Business opted not to terminate the portfolio, while Risk satisfied itself with the Business assurance of its comfort with Archegos.
The curious case of downgrading Archegos while increasing their limit
In the January 2021 annual review, Risk downgraded Archegos’s credit rating from BB- to B+ while curiously more than doubling its PE and Scenario limits from $20m and $250m to $50m and $500m, respectively.
Risk practised some contortionism to support a $50m limit when the maximum guidance for B+ rated entities was $10m. They argued that an exception to the rule was warranted because the framework indicated “that PE limits should not exceed 10% of NAV”.
The juniorisation of the Prime Services
The Fed and PRA fines highlighted the “lack of experienced staff with sufficient stature” within the Prime Services division and the “injudicious cost-cutting” pursued by Credit Suisse. It is not farfetched that the failure to escalate and remediate the Archegos excesses related to ineffective personnel incapable of understanding the magnitude of the risks they were running.
In the years leading to the Archegos implosion, 40% of managing directors in Risk were separated and replaced by directors, significantly reducing experience and, ostensibly, “stature”. In the specific case of Prime Services, the headcount reductions ranged from -23% to -54% between 2015 and 2020, as documented by the special committee report.
Similarly, the in-business risk function was stripped down of personnel, experience, and stature. Between 2017 and 2021, it lost personnel while at the same time experiencing title inflation: the junior employees were promoted without the experience that a few years back had been expected in each role.
2. Failure to obtain adequate margin from Archegos
The directional add-on that never was
In 2017, changes in Archegos’s portfolio triggered a 10% directional add-on that required them to post additional margin. Archegos requested to be excused, arguing that the add-on would not be necessary if Credit Suisse netted its short and long portfolio.
Credit Suisse forwent the additional margin and removed the bias add-on entirely[ix] on the condition that the combined portfolio bias did not exceed 75%, either long or short. However, breaches became a recurrent affair in 2020 and 2021, and Credit Suisse gave “grace periods” of up to 5 months to reduce the bias.
Halve your margins, secure the client, and double the losses.
In 2019, Archegos asked Credit Suisse to lower its swap margins. At this point, Credit Suisse requested, on average, a 15-25% initial margin on its swap positions in Prime Financing and 15-18% in the dynamically margined Prime Brokerage portfolio. Archegos argued that other PBs offered lower margin rates and allowed them to cross-margin their swaps and cash equities.
Credit Suisse agreed to reduce Archegos’s standard swap margin rate to 7.5%. The supporting analysis from Risk calculated that the new margin would be 3x of its stressed exposure (under the May 2019 portfolio). They also introduced the contractual right to terminate the swap on a one-day notice basis and to change the IM amount at its discretion[x].
These “rights” were never used despite exceeding predefined thresholds, and one was eventually replaced (with a three-day notice). More surprisingly, now that the margin was to be halved to 7.5%, Credit Suisse had to release a substantial “then-excessive” margin to Archegos.
By 2020, the margin didn’t cover the stressed scenarios, nor was it short/long balanced. This meant that the rationale for removing the directional bias no longer held, which should have re-imposed the add-on or re-increased the margin, but nobody did anything.
Prioritising the non-relevant and the non-urgent.
In 2020, Credit Suisse developed a mechanism to introduce dynamic margining to their Prime Financing book and to allow cross-margining[xi]. The priority list of clients was drafted, and the first implementations occurred in September 2020.
Archegos was not on the list despite being among the top exposures and having ongoing breaches. It was only added to the list on 12 Feb 2021 and identified as a high priority on 26 Feb – precisely one month before its default.
To further illustrate the failures of risk management, in Oct 2020, the Risk analyst covering Archegos learned on a due diligence call that Credit Suisse was the only PB “not dynamically margining the swaps or margining the combined swaps portfolio”.
Returning €2.4bn margin weeks before the blowup.
Between 11 and 19 of March, as Credit Suisse discussed how to margin the Archegos position better, the Business rejected the Risk request to post a $1bn additional margin as that would “alienate the client”. Unknown to Risk, Business had received an estimation (from the metrics team) that the appropriate additional margin should be $3bn and requested a forgiving bespoke calculation. The forgiving amount was $1.3bn, higher than Risk deemed necessary.
While Credit Suisse’s calculation of the correct dynamic margin was over a billion dollars, based on the previous static margin rules, Archegos requested $2.4bn back through seven separate requests on the 11th, 12th, 15th, 16th, 17th, 18th and 19th of March, and again, Credit Suisse complied. Each time, in-business risk and Risk approved the requests.
The complex and large structure of the bank also played a role. Risk incorrectly thought that because Archegos was Watch-Listed, any margin payments from Credit Suisse would require in-business risk and Risk approval. This was only the base for Prime Brokerage but didn’t cover the variation margin in the Prime Financing swaps portfolio.
3. Failure to have clear roles, responsibilities and accountability
The cat safeguarding the cheese
The person handling the business relationship with Archegos was also in charge of the in-business risk management. This unnatural and highly unconventional situation came about one year before the implosion when the then-head of in-business risk management passed and was replaced by a 13-year sales and marketing veteran[xii].
As this person served as the sales coverage manager for Archegos, he had a preexisting sales relationship with Archegos when he transitioned to in-business risk. In Sep 2020, 6 months before its implosion– the London Risk team wrote in an email[xiii]:
“The team is run by a salesperson learning the role from people”
“In-business risk is not the best 1 LoD anymore”
“All influential in-business risk people are gone”
The oversight of The Counterparty Oversight Committee (CPOC)
A different hedge fund imploded one year before the March 2021 Archegos case. Hedge fund Malachite Capital imploded from the Covid-19 market volatility, generating $214m in losses.
The Board requested Internal Audit to conduct a review and subsequently embarked on “Project Copper” to enhance the controls and scalation framework during periods of stress. In contrast, Internal Audit concluded that there was “no other profile like that of Malachite”, but Project Copper yielded a new committee: The Counterparty Oversight Committee (CPOC).
Archegos would be one of the inaugural cases to be discussed in the first CPOC.
The September 2020 inaugural CPOC minutes observed that Archegos i) used substantial leverage relative to peers, ii) has generated significant scenario exposures, and iii) has highly concentrated single-name stock, albeit in liquid names. However, there was no action/decision taken or timeline specified to remediate the excesses, the implementation of dynamic margin, or when to revisit the case.
The post-Archegos blowup found that the committee members' engagement depended on whether the client was under their remit. They also note that discussions were collegial with no sense of urgency. They concluded that this approach led to a siloing of expertise and reduced the discussion's richness. The CPOC filed to impose governance or oversight on the more troubling hedge fund relationships.
Two Double-hatted co-headed head in-label-only confusion
What? Exactly. The Archegos blowup revealed that nobody felt responsible for the whole of the Prime Services business. Since 2020, the Business had been led in-label-only by two co-heads who had previously been leading the “Prime Brokerage” and “Prime Financing” businesses. To make matters easier, one was based in London and the other in New York.
Although they were promoted to co-head the Prime Service business, they were also tasked with other functions. None felt responsible for supervising the relationship with Prime Financing clients in the US and claimed no familiarity with the Archegos – despite it being among the top 10 clients and the third most significant gross exposure before its default.
4. Failure to have an adequate reputational risk review process
In a classic case of form over substance, the Business was “the only party that could” initiate a reputational risk review process (RRRP). As such, no committee discussed the news, regulatory notices and associated litigations listed below until 2015.
In 2015, the Anti-money Laundering team in EMEA urged and chased the Business to initiate the RRRP for months. Another RRRP was held in 2018 after the UK supervisor, the Financial Conduct Authority, reviewed the Archegos KYC file and elevated concerns about its completeness.
It is important to note that this was not the case of a “bad employee” who was subsequently fired. The offender was the key person in the hedge fund/family office. The minutes of the RRRP highlight profits, the fact that the fine prevents Archegos from using criminal money and that other PBs continued relationships with Archegos.
Perfunctory unconditional approvals ensued. Archegos continued to expand its business footprint with Credit Suisse.
5. Failure to have adequate data quality management for risk metrics
The Risk committees relied on information that was four to six weeks old, and, as explained in the excesses section, the risk managers discounted large potential exposure numbers because they didn’t think they were valid.
Conclusions
Archegos’ blowup illustrates a fundamental failure of management and controls. Nobody at Credit Suisse appreciated the severe risks that Archegos’s portfolio represented. That was not because the risks were hidden – they were conspicuous, in plain sight.
Between 2020 and 2021, Archegos was “discussed” in 13 UK Investment Bank Credit Risk Committees, 13 Investment Bank Credit Risk Committees, 16 Global Markets Risk Management Committees, and 3 Risk Management Committees. However, each time, the minutes reflected that “Archegos’s risk exposure was being managed”.
The (mis)management of Archegos reflects a careless attitude towards risk, which prioritised a form-over-substance. A first and second line of defence that was accommodative to a degree of incompetence. Incapable of reining in, and even worse, enabled Archegos’s voracious risk-taking.
The deficient risk discipline was abetted by Credit Suisse’s lack of accountability for Risk failures, a cultural unwillingness to engage in challenging discussions and the unwillingness to escalate matters of grave economic and reputational risk.
Archegos is a Greek word for the one who leads the way. It is fitting then that the $10.3bn losses and the insights and research from the Credit Suisse special report will lead the way for Risk functions to improve their governance, their mandates and their value proposition to help create more resilient financial institutions.
[i] The Federal Reserve Board of Governors assessed a civil money penalty of USD 268,494,109.20. At the same time, the Bank of England’s Prudential Regulation Authority imposed a financial penalty of GBP 124,403,000.00 (equivalent to USD 170,235,000). The latter qualified for a 30% discount following the early settlement agreement and was hence reduced to GBP 87,082,000.00 (USD 119,165,000.00). Both notices were dated 21 Jul 2023.
[ii] The Credit Suisse inquiry into the Archegos blowup is fully documented on link.
[iii] Archegos employed a “fundamental research-driven long/short equity strategy focused on long-term (18 months to 3 years) value investing” strategy. The involved stocks and their price drops (in parenthesis) that precipitated the blowup were: Viacom (-53%), Discovery (-46%), RLX Technology (-46%), GSX Tech (-63%), Baidu (-21%), Tencent (-33%), Farfetch (-20%), and VIPs (-35%). As reported by Ryan Eyes and Bespoke Investment Group (1 and 2).
[iv] The blowup impact on PB banks was at least $10.3bn in trading losses, and significant losses were disclosed by Credit Suisse ($5.5bn) and Nomura ($2.9bn).
[v] In the same quarter, Q1-2021, Credit Suisse’s losses from the exposure to Greensill Capital, a supply chain finance specialist, were estimated to be $1.7bn.
[vi] PE is a calculation to assess the maximum potential exposure of the bank (at the 95% confidence level, ignoring the most unlikely tail risk) in the event of a counterparty default. Accordingly, PE considers and provides one measure of the sufficiency of the margins posted by a counterparty.
[vii] These belong to 10 scenarios used to assess a counterparty’s performance. The “Severe Equity Crash” scenario simulated a 30% decrease in equity prices in developed markets and a 45% shock in emerging markets and Japan. The “Bad Week Equity Crash” scenario simulated a 20% decrease in developed markets and a 30% shock in emerging markets and Japan.
[viii] Even more so, in March 2020, the market turmoil related to Covid 19 sent all figures haywire.
[ix] This only applied to Archegos Prime Brokerage's dynamic margining rules.
[x] Within the contract, a framework was laid out that explicitly envisioned that the base margin would increase if a liquidity threshold were reached. Based on Archegos holdings relative to daily trading volumes (DTV), the margin add-on would be 5% to the incremental portion of a position above two days DTV and a further 8.5% add-on to any incremental portion above five days DTV.
[xi] Cross Margining allowed clients to lower their initial margin requirements in circ*mstances where the client’s Prime Brokerage and Prime Financing positions were potentially off-setting. It could also lead to higher initial margin requirements where the client’s swaps portfolio was not offsetting but, instead, additive with respect to directional bias or concentration such that add-ons might be triggered.
[xii] The person chosen to lead the in-business risk function was a managing director who had worked in the trading platform for 13 years in sales and marketing. He had never worked in a risk role at Credit Suisse or elsewhere.
[xiii] Re-drafted quotes to medium.
