Credit Suisse Archegos
What happened to 'risk management'?
I expect that most of you are at least generally familiar with the Credit Suisse Archegos affair, in which CS suffered massive losses resulting from its prime broking relationship with a single hedge fund client, Archegos Capital, managed by ‘Tiger Cub’ Bill Hwang (pictured above). If you don’t already know the basic facts, you can catch up here. This story is especially noteworthy for those of you interested in careers in investment banking, risk management or asset management and of course for any of you planning to work at a hedge fund or in the prime broking unit of a bank.
Credit Suisse reported a loss of $5.5bn from this one hedge fund relationship and the bank—which is one of the world’s largest—was forced to cut its dividend, terminate its share buyback program and raise additional equity capital from its shareholders. The Credit Suisse share price fell 25% when the loss was reported in March 2021 and has not recovered since. Credit Suisse now has a new chairman, new heads of Investment Banking and Group Risk, and a new head of Prime Broking. A number of other senior bankers have been fired and the bank is seeking to claw back $70mm in past compensation from these former employees. Credit Suisse has dramatically scaled back the activities of its prime broking business and has undertaken a firm-wide risk review, which could have significant implications for the CS strategy and culture. In short, this was a BIG DEAL for Credit Suisse (not to mention for Archegos) and we should all attempt to learn something from it.
Credit Suisse was not the only bank caught out in the Archegos fiasco, but it suffered the biggest losses. Collectively a handful of banks lost $10bn or so on this same hedge fund client. CS reported the biggest loss at $5.5bn. Other banks suffering big losses include Nomura (over $2bn), Morgan Stanley ($1bn) and UBS ($800mmn or so). Goldman and Deutsche were also involved with Archegos but suffered much smaller (non-disclosed) losses.
Credit Suisse recently released to the public a 150-page report prepared for its board by the Paul Weiss law firm, which documents in great detail the many risk management failures at Credit Suisse. I have reviewed the entire report and recommend that those of you pursuing related careers read at least the Executive Summary (30 pages or so). You can find the CS press release and an embedded link to the Paul Weiss report here.
Prime broking is the business of servicing hedge fund clients and it is supposed to be a relatively low-risk business, or at least that is what the sponsoring banks tell their board and their shareholders. But a handful of big banks collectively lost $10bn from their prime broking activities with a single hedge fund (family office) client, so it is not unreasonable to think that perhaps prime broking may not in fact be so low-risk after all.
Not every failure of risk management leads to losses, of course, and not all losses are the result of risk management failures. But in this case, it appears that CS lost money because it breached many of the basic tenets of sound risk management, in practice if not in policy. Even when CS had sound risk management policies in place, the Prime Broking team repeatedly ignored internal limits and failed to take prompt action to mitigate the known and increasing risks associated with the Archegos account. And the Prime Broking team was never held accountable by the Group Risk function or by other senior executives of the bank, which was perhaps the biggest risk management failure of all.
So what were some of these risk management failures? Here is my take from reading the Paul Weiss report:
Know Your Client. “Know your client” (KYC) is often regarded as the number one principle of risk management at a financial services company, so let’s start there. Bill Hwang was of course well known to the prime broking team at CS, who were aware of his legal difficulties (charged with insider trading by the SEC and banned from trading in Hong Kong) and his very aggressive investment strategy (further details below). But various senior executives of Credit Suisse, including the head of Investment Banking, the Group CRO (Chief Risk Officer) and the bank’s CEO claim never to have had heard of Bill Hwang or Archegos until they got word of the massive losses incurred by CS in late March 2021. Nor was the Risk Committee of the Board of Directors ever briefed on the matter until after the Archegos default. These are all rather damning facts, particularly given the bank’s gross notional exposure to Archegos of over $20bn, the repeated and substantial breaches by Archegos of various internal CS risk limits, and the bank’s eventual reported losses of $5.5bn. Keep in mind that Archegos was an unregulated single client (family office) hedge fund known for its aggressive investment strategy, ranked by CS at the bottom of its peer group in terms of credit quality, whose principal had been banned from the securities industry and who was managing a concentrated, volatile, illiquid and highly leveraged portfolio across a number of prime broking banks, whose position at CS alone grew to $20bn, about four times larger than any other hedge fund client of the bank. Keep in mind also that Credit Suisse’s $20bn gross exposure to Archegos represented over 40% of the bank’s capital (book shareholder’s equity). If these facts don’t trigger a KYC obligation at the most senior levels of the bank, I don’t know what does.
Understand your real risk exposure: It seems that Credit Suisse did not have a good handle on its actual (and rapidly increasing) risk exposure to Archegos, and I think this is putting it charitably. The bank’s internal risk analytics and IT systems were poor and CS repeatedly chose not to implement more modern systems which would have more accurately portrayed the bank’s large and increasing risk exposure to Archegos. It is hard for an outsider, even an experienced risk professional (which I am not), to evaluate independently some of what is in the report, but it appears that the standard internal CS risk metrics used with hedge fund clients consistently and significantly underestimated the actual risk exposure of CS to Archegos. Internal risk analytics during the six months or so prior to default showed the bank’s “Potential Exposure” peaking at around $500mm and its “Scenario Exposure” peaking at $800mm. These were large numbers well above internal risk limits, but the CS Prime Broking team largely ignored them, allegedly because of the widespread lack of confidence in the integrity of the CS risk metrics (which the Prime Broking team thought overstated the actual risk posed by Archegos). Of course as things turned out, the CS risk metrics were indeed flawed, but not in the way that everyone thought. The PE and SE risk metrics in fact greatly understated the actual risk being taken on by Credit Suisse, whose gross exposure to Archegos reached $20bn with incurred losses in excess of $5bn.
Compliance with internal limits is not optional. In reading the Paul Weiss report, I was particularly struck by the extent to which the CS Prime Broking team and their immediate supervisors largely ignored the bank’s established internal risk limits. As a former banker, I did not know this was even possible. But at CS the Prime Broking team continued to execute increasingly large and highly leveraged trades with Archegos in violation of clearly established risk limits, even in the days immediately preceding the Archegos default. We have just discussed the fact that CS’s own internal risk analytics appear to have greatly understated the amount of actual risk exposure taken on by CS, but the CS Prime Broking team did not abide by the internal limits in any case. In fairness, the Prime Broking team did from time to time get senior executives to sign off on the compliance breaches, but the Paul Weiss report portrays these approvals as largely rubber-stamp affairs, given without much in the way of serious risk analysis.
Margin matters. As a prime broker, Credit Suisse made margin loans to finance the positions held by Archegos at CS. This is what prime brokers do and the risk professionals generally pay great attention to the adequacy of margin for obvious reasons. This may seem obvious, but CS lost money from its prime broking relationship with Archegos in large part because it did not hold enough margin (cash and non-cash collateral) relative to the size and volatility of the Archegos positions being financed. The Paul Weiss report goes into great detail about the many margin failures at CS, which I won’t repeat here, but a few things do stand out. First, it appears that Credit Suisse was hampered by serious weaknesses in its internal risk analytics and IT systems, discussed above. It is hard to get margins right if you really don’t understand the riskiness of the portfolio being financed. Second, it is not adequate or sufficient just to establish the right initial level of margin. Margin needs to be adjusted dynamically over time with changes in the size, composition and risk of the underlying portfolio. Third, even if you get both of these things right, which CS apparently did not, you have to actually enforce your contractual rights to demand additional margin or liquidate the client’s portfolio. The report says that that Archegos always put up more margin when requested, but apparently CS didn’t make such margin calls very often, perhaps for fear of upsetting the client or losing the business to competing brokers. And finally, CS allowed the client to arbitrage differences in the margin requirements between it and other prime brokers as well as within CS itself. Over time CS got more and more of the Archegos business, apparently because it offered more relaxed margin terms than did the competing banks. And even within CS, Archegos was able to get much better margin terms when it executed trades with CS in the form of total return swaps (margin < 7.5%) than when it executed them as cash trades (15% margin). (I have been unable to determine why CS would have allowed this to happen, as the risk of the two forms of execution should have been similar if not identical, but cash equity and TRS trading was done by different units in the bank who perhaps had different internal risk appetites or compliance tolerances.)
Of course Credit Suisse knew that margin matters—in fact, CS deliberately chose to use its “flexible” approach to margins as a competitive advantage to win more business from Archegos. But it wasn’t until a few days prior to default, when belatedly evaluating an internal proposal to move Archegos from “static” to “dynamic” margining, that CS realized that its current margin position was short by $3bn or more. (This incremental $3bn represented about 15% of the gross value of the Archegos portfolio and over half of the losses eventually suffered by CS,) And even then CS chose not to demand full margin from Archegos, instead reducing the request for additional margin by more than 50%, which it thought would be more palatable to the client but which CS was unable to collect in any event. In a classic case of “shut the barn door after the cows have escaped”, Credit Suisse has reportedly now migrated all its hedge fund clients to dynamic margining. Had CS done this sooner with Archegos, it might have saved itself and its shareholders a whole lot of hurt.
See the whole field. The size of Archegos’ portfolio at Credit Suisse grew rapidly over time, reaching $20bn of Gross Market Value (GMV). The composition of the Archegos portfolio also changed, generally in ways that magnified the risk to CS. By the time of default, the Archegos portfolio had a 75% long bias, with half the portfolio invested in just 4-5 individual stocks, which were volatile, highly correlated with one another and practically illiquid due to the size of the positions owned by Archegos (relative to outstanding float and trading volume). We have discussed above the difficulty that CS had assessing accurately the risk characteristics of the Archegos portfolio held at CS. But in reality, the situation was much worse than this, due to Archegos’ undisclosed trading activity at other banks. Archegos was under no obligation to provides its prime brokers with full transparency on its overall positions, unless the brokers required it to do so as a condition of doing business (which apparently they did not, at least not at CS). Credit Suisse was of course aware that Archegos was doing a lot of business with other banks, and at some point it became aware that the positions which Archegos held at other firms were very similar (if not identical) to those it held at CS. This was a very important element of the overall Archegos risk picture which CS appears not to have fully digested until quite late in the game. It turns out that the total size of the Archegos positions at all firms combined grew to $120bn GMV, or 6x the position held by Credit Suisse. As you can imagine, not knowing this made it very difficult for Credit Suisse (and the other banks) to adequately manage their own risks.
Warren Buffett likes to say “if you have been playing poker for 30 minutes and you don’t know who the patsy is, it’s you.” And in this case the biggest patsy (but perhaps not the only one) appears to have been Credit Suisse.
What goes up can come down, and it often will. You are all finance people, so you know this, but in the capital markets it is generally true that what goes up can come down, and often does. (And it usually does so just when we all least expect it.) This is true with share prices of course, and the phenomenon is compounded with a highly leveraged portfolio. It appears that Archegos built a $120bn portfolio with gross leverage in the neighborhood of 90% (more at CS). At its peak, Archegos’ NAV appears to have been in excess of $10bn, up from $500mm when it began the firm. (NAV is Net Asset Value, ie the gross market value of the assets in a portfolio less the amount of debt incurred to fund the investments.) At one point, Archegos held “unencumbered cash” of $6bn. But when the price of just a few stocks in the Archegos portfolio tanked (Tencent fell 20% in one day), Archegos quickly went bust. NAV plummeted and Archegos turned over all its unencumbered cash to satisfy margin calls. Those firms who hesitated to demand additional margin got nothing, and this included Credit Suisse.
Take your losses and move on. I said above that the fundamental problem facing Credit Suisse was one of inadequate margin, attributable to a litany of related risk management mistakes. And I think this is true (almost by definition), but it is also the case that CS compounded the margin problem with the way in which it decided to liquidate the securities after default, in a staged process over time coordinated with several other banks. It is hard for me to say that this was a bad decision a priori—even good decisions can lead to bad outcomes, and vice versa—but in hindsight it appears not to have worked out well for CS. More aggressive banks like Goldman and Deutsche who decided to go it alone and dump their positions quickly seem to have mitigated their own losses, but of course this aggressive selling activity drove down the market price of the Archegos portfolio even further and increased the pain for CS and the others.
Compensation and Skewed Incentives. So many things went wrong with the way Credit Suisse handled the Archegos relationship that we have to ask ourselves why. As is often the case, the answer has a lot to do with compensation and skewed economic incentives. Investment bankers are generally very bright, talented and aggressive people and they respond quickly and strongly to economic incentives, most notably compensation. (At many investment banks, employee compensation can amount to 40-50% of total net revenues.) When bankers are paid based on short-term revenues generated, they will do all they can to generate short-term revenues. And this is exactly what the bank wants them to do. This approach generally works out well for the bank—if it didn’t, then banks wouldn’t do it—but only when the bank establishes and enforces adequate risk and other controls. It is perhaps fair to criticize the CS Prime Broking team for their cavalier approach to risk and their blatant disregard of internal controls, but it would appear that they did do a good job of “winning” the Archegos business and generating revenues for the bank. (The Archegos account grew from annual revenues of $6mm or so to an anticipated $40mm in the year Archegos defaulted.) Yes, the Prime Broking team was meant to be the “first line of defense” from a risk management perspective, but we really have to ask ourselves how good a defense we should expect when the bankers are paid based on maximizing revenues rather than managing risk.
If we want to know where Credit Suisse thinks the blame really lies in the Archegos affair, take note of the fact that it is attempting to claw back $70mm in past and deferred compensation from some number of terminated executives. No doubt senior members of the Prime Broking team got paid a lot (before they were fired), but my guess is not anywhere near this much. I suspect that most of this $70mm will come not from the Prime Broking team but rather from the more senior (and very highly paid) group executives. In my mind, these are the folks who really failed at their jobs, with severe adverse consequences for the bank. It is right that they were terminated and it is right that they cough up some of their ill gotten gains.