Nakedshorts has just published an excellent post on the FSA's response to the recent pricing problems uncovered at a variety of investment banks based in London.
According to a Bloomberg article referenced by Nakedshorts, a "large number" of UK banks have mismarked securities. Ouch, to say the least.
The blog also includes a link to the FSA's "dear CEO" letter regarding valuation. Some highlights:
- "Firms should proactively review their processes and controls to ensure that they are commensurate with the challenges and issues posed by the increased market illiquidity and valuation uncertainty."
- "In the current market environment, characterised by illiquidity, constrained profitability and management stretch, firms are more susceptible to mis-marking frauds perpetuated by trading staff. In the current environment, we are more concerned that weaknesses in processes and controls are more likely to be exploited and exposed."
- "Product Control Staff (i.e. the back office) were unable to adequately challenge the front-office staff, through lack of skills or seniority. Against this background Product Control seemed to be acting too much as a business facilitation function and not enough as a control function."
Once more, we make the same comment as in prior posts. If this is the environment in an investment bank, what is it like in a hedge fund, free of these "bureaucratic controls"?
When thinking of hedge funds, investors must be ever conscious of the clear conflicts of interest and potential for valuation manipulation when portfolio managers mark their own books. Just because the investment banks do it, doesn't mean that it's a good idea.
Some time ago, we commented on Credit Suisse's London office, where a group of CDO traders managed to "deliberately misprice" their portfolios to the tune of nearly $3 billion.
The regulatory process has made some progress: according to the FT, the UK's FSA is close to agreeing a GBP5 million settlement (we wonder how that compares to the bonuses the traders in question hoped to earn based on their fictional prices.)
The FT describes the mispricing event as follows:
"Credit Suisse’s losses occurred in trading books dealing in residential mortgage-backed securities and collateralised debt obligations. They emerged within days of the bank hailing a resilient set of annual results which it argued demonstrated the strengths both of its risk management and risk culture. The mismarkings came to light in an ad hoc review, and there had been no specific indicators that there was any cause for concern."
As always, we remind investors that if this could happen in an investment bank, it could certainly happen in a hedge fund. How many hedge funds, for example, have a process of "ad hoc" reviews of their pricing?
According to Pensions and Investment, the Boston office of the US Department of Labor (the “DOL”) recently issued a letter to an (unidentified) US Pension Plan subject to ERISA (the Employee Retirement Income Security Act) stating that the plan was in violation of ERISA regulations. The DOL is responsible for monitoring - and sanctioning - ERISA plans and, in their letter, threatened legal action if the plan in question did not remedy the noted violations.
When valuing hedge funds and other alternative assets for purposes of the Plan’s annual filing, the pension investor had apparently relied upon valuations provided by the underlying funds’ general partners and, in some cases, on audited financial statements for those funds.
This is, of course, standard practice for many hedge fund investors. It appears, however, that this approach could create a major roadblock for ERISA plans.
According to the DOL, “it is incumbent on the Plan Administrator to establish a process to evaluate the fair market value of any hard to value assets held by the Plan. Such a process would include a complete understanding of the underlying investments and the fund’s investment strategy. In addition, the Plan Administrator must have a thorough knowledge of the general partner’s valuation methodology to ensure that it comports with the fund’s written valuation provisions and reflects fair market value. A process which merely uses the general partner’s established value for all funds without additional analysis may not insure that the alternative investments are valued at fair market value.”
In other words, the entity which has to value all assets – and especially hard to value assets – is the pension investor subject to ERISA. There is no way of dodging this poison chalice - the ERISA investor cannot simply rely on the hedge fund’s own valuation.
This is an enormously challenging obligation, particularly in the context of the severe fiduciary standards set by ERISA. Indeed, the DOL position raises a broad question - is it even possible for ERISA plans (or indeed any hedge fund investor) to meet this duty of care?
We have three observations.
Firstly, very few hedge funds provide position level transparency. However, it is stating the obvious to say that, without position level transparency, it is impossible for an ERISA investor (or any other investor for that matter) to have a “complete” understanding of the underlying investments and the fund’s investment strategy. Moreover, even if managers do provide position information, how can investors ensure that it is timely and accurate? The best solution to the transparency issue is a managed account – as such, would one outcome of the DOL’s position, if enforced, be for ERISA plans to only invest through managed account structures?
Secondly, the DOL states that ERISA plans must have a “thorough knowledge of the general partner’s valuation methodology”. However, in practice, most hedge fund offering documents have deliberately vague and unspecific clauses as to valuation and calculation of the net asset value, especially in relation to hard to value instruments. To add salt to the wound, every prospectus we have ever read includes a final caveat along the lines of “notwithstanding the above policies, the general partner (or the Board of directors in “consultation” with the investment manager for an offshore fund) may elect any “alternative method” of fair valuation. “ There is hence very limited specificity as to valuation procedures in virtually all hedge fund offering materials, and certainly insufficient information to provide a “thorough knowledge” of the valuation methodology which will be applied.
If the prospectus gives an inadequate description of the valuation process, investors need to turn to supplementary information from the hedge fund manager. At this point, however, things get worse - many hedge fund managers have not developed any internal, written valuation policy at all. For those funds which do have a valuation document, there is no standardization, and many valuation policies remain uncomfortably vague and unspecific (although, in fairness, we congratulate the minority of managers who have some stepped up and do furnish investors with comprehensive valuation information.)
The worst case is when a manager does have a valuation document, but will not provide it to the investor. Ironically, the worst culprits in this situation are some of the industry’s largest and most well known hedge fund managers. The issue is liability: hedge fund lawyers now appear to advise managers that the more information provided to investors, the more the potential liability. (As an aside, we recently spoke with the CFO of a large hedge fund: he noted that the sight of the Bear Stearns hedge fund managers being led away in 'cuffs had resulted in urgent calls from the firm's lawyers, advising the manager to reduce the amount of information it provided to investors.)
The third area of concern is the ongoing assumption by many investors, including many ERISA plans, that third party administrators assume responsibility for valuing hedge fund portfolios. As such, the administrator, it is perceived, can provide the necessary independence in the valuation process.
Not so fast. As we have noted before, much of today’s administration industry is now emphatic that they perform only the services of a “calculation agent” not a “valuation agent”. This is a relatively mute point when dealing with exchange traded securities, but it is an enormous issue when looking at a hedge fund which trades hard to value instruments (it goes without saying that we need help to value exotic CDOs, not IBM stock).
As a “calculation agent”, many administrators have amended their legal contracts to retain the right to “consult with” the manager and, indeed, accept prices from the hedge fund manager without further verification. Again, we hate to make an “emperor has no clothes” comment, but this is obviously nonsense: taking prices from the manager is like a police officer issuing speeding tickets on the basis of asking drivers how fast they were going.
These issues, in our mind, share a common theme. In recent years, with an ever-accelerating pace, we have watched the legal pendulum which defines how investors and hedge fund managers transact drift ever further in favor of the manager at the expense of the investor. It is trite, but uncomfortably accurate, to say that, in today’s hedge fund industry, no-one wants to be responsible for anything. Everyone is instead seeking to be indemnified to the point of invulnerability.
And this is the disconnect between the hedge fund industry and DOL. ERISA establishes onerous standards of fiduciary responsibility, deliberately designed to make those responsible for ERISA plans accountable, responsible and liable for their actions. Today’s hedge funds, however, are increasingly structured to ensure the lowest possible degree of accountability and liability on the part of pretty much everyone involved.
Against this background, we will watch with great interest ongoing developments as the DOL monitors ERISA plans with material hedge fund portfolios. The question, of course, is whether investing in opaque, uncommunicative hedge funds (even when they are some of the largest in the world) is too close to pushing a square peg in a round hole for investors who do operate within a strict fiduciary framework.
We've commented more than a couple of times (here, here and here, for example) on the folly of putting portfolio managers and traders in charge of pricing their own books - and in recent months Credit Suisse, Morgan Stanley and Merrill Lynch have all learnt this lesson the expensive way.
Add Canada's Toronto Dominion Bank to the list.
According to a Bloomberg article, one of TD's London traders has just left the bank (i.e. been fired) as a result of "incorrectly priced credit derivatives, costing the bank around C$96 million ($94.3 million) in pre-tax earnings." The culprit is described as a "senior male trader".
Thinking about this, a number of comments come to mind.
Firstly, all the recent mispricing episodes seem to have occurred in London. This is despite the view that the FSA's gentlemanly guidance provides a more tightly controlled regulatory environment than SEC rule making. Whatever that means about the quality of the FSA's guidance, more generally we are of the view that if these problems can happen in London, they can certainly happen in NYC.
Secondly, these episodes - so far, at least - are not the work of disgruntled traders at the bottom of the totem pole, envious of their high flying peers (that's the Paris syndrome, of course.) These are senior traders, who have tenure, high compensation - and the knowledge and authority to find ways around control systems, at least for some period of time.
Thirdly, and very disappointingly for TD, this is not the first time their London trading team has landed the bank in hot water. Barely six months ago, in November 2007, the FSA fined TD GBP490,000 for "systems and control failings."
"Mr. [Simon] Brignall was employed as a senior fixed income trader at the firm. On 9 March , he resigned and revealed to TD Bank that he had been attributing false values to his trading positions for a period of almost two years. He had done this in order to hide significant losses on his trading book. He had also entered into a number of fictitious trades during the two weeks leading up to his resignation.
TD Bank did not identify, through its own systems and controls, either the extent of the mispricing of the trades or the fictitious trades. The FSA identified three main system and control failings in relation to Mr. Brignall's trading book.
- the absence of a system of independent price verification;
- a lack of effective trading supervision; and
- a failure to implement effective trade break escalation procedures.
The most important of these was the failure to have in place a system of independent price verification in relation to Mr. Brignall's trading book. This meant that there was no independent third party check on the valuations that Mr. Brignall had attributed to his own trading positions. The FSA regards this as a fundamental control."
The only good news for TD at the time was that, as a result of the FSA's "executive settlement procedures", they qualified for a 30% discount on their fine (we're not kidding.)
It's clearly a major ouch for TD that they have just let exactly the same thing happen again. More generally, this is (yet another!) example of why traders, compensated based on performance, should not be able to price their own books.
Given the frequency of these mispricing episodes, however, it also raises another question. Obviously, if a multitude of investment banks can lose many millions of dollars when they let their traders mark their own positions, the same could happen in a hedge fund which also puts the PM's in charge of marking their own book. So far, though, we have pretty few examples of deliberate mismarking (Lipper, Beacon Hill etc.) and not many of recent vintage, despite the valuation challenges of recent months.
Why is this? Is this because hedge funds have better controls, or because when the portfolio managers own their own hedge fund, there is no-one to blow the whistle on mispricing events?
Bloomberg reports today that Morgan Stanley London has suspended a credit trader and made a $120 million "negative adjustment"related to "erroneous values of his positions."
This comes, of course, after Merrill London identified a $20 million loss due to deliberate mismarking by a trader on the equity derivatives desk (see our earlier post.)
We shouldn't pick on Merill and Morgan, mind you: their misfortunes fall into insignificance compared to Credit Suisse's stunning $2.65 billion write down related to a trading group that mispriced CDO's (also, see our earlier post.) Lehman has also had some difficulties, let alone the Bank of Montreal which had a near $500 million loss due to two natural gas traders providing bogus prices to their head office.
We continue to believe that pay structures dominated by incentive compensation (be it a hedge fund's "20%" in 2 and 20, or the "eat what you kill" model of prop desk compensation) create an unavoidable conflict of interest (incidentally, hedge funds which have 2 and 20 and then pay their staff on individual P&L is probably the most toxic combination of all).
As such, best pratice must always be for the security pricing process to be owned by the back office, independent of the traders. Yes, we may give up some market color and "accuracy" if we put the back office in charge, but we gain independence. We have yet to find the investor who is not prepared to accept a few basis points of difference between front and back office marks across the aggregate book in return for removing the risk of a 5,000 basis point loss in the event of a hedge fund pricing fraud.
Clearly, if a global, regulated institution like CS, Morgan, Merrill and Lehman can have a pricing problem, then so can a hedge fund. So how can investors respond?
The recent best practice papers from the President's Working Group, the hedge fund working group, AIMA etc. have all emphasized the need for segregation of duties, although with varying degrees of "weasel words" which let the PM's get involved on the hard to price positions. It's time to get rid of these exceptions and make sure that the hedge fund industry adopts a clear and consistent back office pricing regime.
In other words, it's time to walk the walk, not just talk the talk.
We recently commented at length on the issues which arise when traders and portfolio managers are put in charge of pricing their own portfolios. In particular, we highlighted that the the institutional world has already given us plenty of examples of what could happen in a hedge fund if the PM's are in charge - notably Credit Suisse's multibillion dollar loss related to CDO valuations.
As an addendum to our comments, we were unsurprised to see the same thing happening again: this time it's Merrill Lynch. According to Bloomberg, the bank is "probing one of its trading desks in London and has suspended a trader after discovering he may have overstated the value of some of the bank's equity derivatives." The damage? Just about 10 million pounds, or $20 million.
Interested readers may also want to check Greg Newton's post on the same topic at NakedShorts. As Greg says, how many times does this has to happen before someone...
Credit Suisse announced last week that the bank will take a $2.65 billion - yes, that’s billion - write down related to debt securities which were, in Bloomberg’s words, “deliberately mispriced by traders”.
Bloomberg’s coverage continues: “An internal review found that the pricing errors, first announced last month, were made intentionally ‘by a small number’ of traders, who have since been fired or suspended....The Swiss bank hasn't disclosed the names of the traders responsible for the incorrect pricing of residential mortgage- backed bonds and collateralized debt obligations. Credit Suisse said it reassigned trading responsibility for the CDO business and took measures to improve controls to prevent and detect misconduct, which were ``not effective'' previously.”
Sound familiar? It should. A few days earlier, another Bloomberg article noted that Lehman had suspended two London-based equity traders after internal controls identified ``issues'' on share valuations.”
CS and Lehman are not the only institution with valuation problems - earlier in February, AIG announced multi billion dollar write downs on a portfolio of credit default swaps: Bloomberg stated “AIG's difficulty in valuing its derivatives portfolio earned it a rebuke from its auditor, which earlier this month cited "material weakness" in the company's internal controls.” While it’s not clear who was marking these instruments, we’d hazard a guess that front office investment professionals / risk takers who had a hand in developing the models underlying AIG’s initial marks.
These events merit close attention, because they highlight that in today’s investment banks, it is the front office traders who mark securities, subject to checks by mid / back office accountants and product controllers. Many hedge funds take the same approach, with marks for hard to value positions generated by the portfolio managers subject to checking by the fund’s back office and perhaps some form of verification by a fund administrator. Indeed, hedge funds often point to the “best practice” example of investment banks as justification for their own front office pricing approach.
The argument behind putting the front office in charge is simple: some instruments are so complicated, that only the traders who bought (or sold) them in the first place stand any change of valuing them correctly. Perhaps it is only the traders that have the contacts with other brokers and market markets who can give quotes and levels, or perhaps some derivative and structured products are so complex that the back office is (considered) incapable of understanding how valuations change.
Valuation is not a black and white story, and there is certainly merit to these arguments. Faced with three broker quotes, it probably is the case that the portfolio manager could know that one broker is short of inventory and will bid up, while another counterparty would laugh if the fund tried to trade at the level given on their price quote. (The complete lack of accountability Wall Street enjoys when providing these prices, notwithstanding the fact that the banks know that their hedge fund clients need those quotes to cut their monthly NAV’s is, of course, a topic for another day.)
These arguments roll together to form the view that the only way to price complicated hedge fund portfolios with “accuracy” is to put the front office in charge.
Sorry - we disagree. For two reasons.
Firstly, how much is any security actually worth? As we have said before, a security is worth what someone else will pay for it, at the time you want to sell it. The rest of the time, pricing is just an estimate. Let’s emphasize that point again: when pricing a hard to value security, we are not identifying an actual trade, but rather trying to figure out who can come up with the best estimate of what the position would be worth if it was sold. This is not a mathematical science where one price is right, and another wrong.
If a price is not tested in an actual transaction, there is absolutely no way to prove that the PM’s mark is more “accurate” than a price generated by the back office, be that the back office of a hedge fund or an institution. It is a flimsy argument at best for a PM to assume that his or her price of a distressed bond (based, say, on a single quote from Goldman with perhaps a subjective adjustment to make it more "conservative") is more accurate than a price generated by the back office, using, for example, the average of three broker quotes received from different dealers.
Secondly, putting the front office in charge ignores what is the bigger problem - the human dynamics of power and authority.
Risk takers are compensated - handsomely - based on performance. It is pretty unrealistic to assume that the back office has an equal seat at the table when challenging marks (after all, the trader usually has a pretty good story as to why his or her mark is right.) Let’s take an example: what happens if a back office professional making, say, $100k challenges the marks of a front office trader who stands to make a bonus of $10 million based on his or her trading P&L? We don’t think you need to be an advanced student of workplace psychology to understand the dynamics of that conversation.
Front office professionals are compensated based on their performance, and putting the front office in charge of valuation puts, to a significant degree, the front office in charge of their own compensation. Given this conflict, investors will consistently prefer to trade some front office “accuracy” for more back office “independence”, and a reduced risk of deliberate price manipulation.
The Credit Suisse and Lehman examples illustrate exactly this problem. The lesson for hedge fund investors? Just because institutions do something, it doesn’t mean that it’s a good idea.
Bloomberg today carries an interesting story on Merrill's efforts, back in June 2007, to seize collateral from Bear Stearns' High Grade Structured Credit hedge funds when they failed to make margin payments.
On June 15, 2007, the article notes that "Merrill Lynch seized $850 million of the CDOs from Cioffi's funds -- as lenders are allowed to do when their margin calls aren't met -- and tried to sell them in the market. When the firm received bids of 20 cents on the dollar, it abandoned the effort."
The point of the article is to highlight that Merrill was economically irrational: by forcing the Bear Stearns issue and creating some price discovery in the market - which, unfortunately, discovered terribly low prices - Merrill was then faced with the issue of how to mark its own $26 billion CDO book. The ramifications of opening this can of worms are, of course, well known.
William Fitzpatrick, a financial services analyst at Racine, Wisconsin-based Optique Capital Management, states in the article: `It was in Merrill's interest to wait it out and allow the Bear Stearns funds to recapitalize, so they wouldn't have to re-price their (own) assets.''
In other words, don't ask, don't tell.
See no evil, hear no evil, speak no evil.
Er, bury your head in the sand etc. etc.
The story here is familiar - we commented some time ago on an excellent front page Wall Street Journal article about the fall out from the sub prime pricing scandal. One of the vignettes in the piece was to comment on the interplay between the (shortly thereafter closed) Dillon Read hedge fund unit and the fund's parent, UBS. Per the WSJ:
"The hazards of this new age of uncertainty became clear at Dillon Read in March, when rising defaults by homeowners were hammering the value of mortgage securities. John Niblo, a hedge-fund manager at the firm, acted fast. He twice slashed his fund's valuation of securities tied to "subprime" mortgages, knocking them down by about 20%, or nearly $100 million, say traders familiar with the matter.
But managers at UBS AG, Dillon Read's parent company, were irate. The Swiss banking giant was carrying similar securities on its books at a far higher price, the traders say. In conference calls, the UBS managers grilled Mr. Niblo on his move. "I'm marking to where I could reasonably sell them," Mr. Niblo responded during one call, according to the traders familiar with the conversations."
When thinking about hedge fund operational risk, the lesson from these events can be illustrated by what we like to call the "iceberg theory". Sounds funny, but our point is this:
CDOs were a wonderful example of an instrument where a tiny amount of paper (the tip of the iceberg sticking out of the water) could be traded between the investment banks, hedge fund etc. at par. Transaction activity, broker quotes, margin calls etc. could all report with confidence that sales were taking place at strong valuations.
But....underneath the "water", there was a huge mass of paper which was never traded. Any attempt to sell even a small proportion of these "below the waterline" assets would, as Merrill found, quickly reveal that there was no real market at all.
Despite all the sophistication of CDOs and their ever more inventive offspring, Wall Street hence ended up in something very similar to a penny stock market. It's obviously easy for a hedge fund holding an OTC bulletin board stock to buy a few shares to keep the bid / ask high (last day of the month is always a good time) but, of course, any attempt to unload the full position would be impossible without a collapse in the price. The difference with CDOs was that the amounts involved were in the billions rather than the millions.
As always, we return to our favorite valuation maxim: any security is worth what someone will pay for it, at the time you want to sell.
According to Bloomberg today, Goldman Sachs held some $70 billion of "Level 3" assets as of the end of Q3 2007. This amounted to 6.9% of the firm's assets, compared to 5.7% at Citi and 2.5% at Merrill.
FAS 157, the new US accounting standard, defines three "Levels" of pricing sources. Level 1 inputs are quoted prices - obviously the best. Level 2 inputs are "inputs other than quoted prices included within Level 1 that are observable for the asset or liability, either directly or indirectly." This means things like yield curves, fx forward curves etc. that can price instruments like fx forwards, interest rate swaps etc.
Level 3 is the rogue's gallery of everything else: per FAS 157, "Level 3 inputs are unobservable inputs for the asset or liability....allowing for situations in which there is little, if any, market activity for the asset or liability at the measurement date....Unobservable inputs should reflect the reporting entity's own assumptions about the assumptions that market participants would use in pricing the asset or liability (including assumptions about risk). Unobservable inputs shall be developed based upon the best available information in the circumstances, which might include the reporting entity's own data."
Goldman's CFO, however, is quoted in the Bloomberg article as saying: "just because they're in Level 3 doesn't mean we're not
pricing them correctly,'' Goldman Chief Accounting Officer Sarah
Smith said in a Nov. 9 interview. ``We mark our positions to the
point where we could exit at that moment.''
This is $70 billion of assets with "unobservable inputs", guys....
We wrote a post about valuation (no doubt the first of many) a few days ago. Since then, we have seen three notable articles related to pricing - each focuses more on valuation issues within institutions, but each has lessons for hedge fund investors.
A Hall of Mirrors
The first was the front page article in the Wall Street Journal on Friday, October 12 (good job the calendar was not running a day later). According to the WSJ, "U.S. investors face an age of murky pricing - valuation of securities tougher to tie down."
Nakedshorts has already published comments on the article, but to add our two pennies of thought:
According to the WSJ "These days, after a decade of frantic growth in mortgage backed securities and other complex investments traded off exchanges that clarity [the clarity of the ticker tape and its electronic pricing feed descendents] is gone. Large parts of American financial markets have become a hall of mirrors".
The article touches on many topics, including Warren Buffet's suggestion that funds should sell 5% of their hard to value portfolios before setting a price, although a comment from the consultant Janet Tavaloki highlights that even this process could be exploited: "even selling a portion of a position can be gamed by getting four or five investors involved to buy each others' positions."
This is an excellent - although profoundly pessimistic - article, and is mandatory reading for anyone thinking of hedge fund valuations.
Bear Stearns' Bad Bet
The next article was published in this week's Business Week magazine, which provides an in depth analysis of the failure of the Bear High Grade Structured Credit and Structured Credit Enhanced Leverage funds. The article has some interesting factoids: apparently Ralph Cioffi had served as executive producer of the 2006 indie film Just Like My Son,
starring Rosie Perez. Following Ken Lipper's movie exploits, it sounds as if any involvement in the film industry should be a immediate red flag for investors.
More fundamentally, Business Week makes the following comment: "the Bear funds weren't stand-alone portfolios like the ones that blew up on
Amaranth Advisors and Sowood Capital Management in recent years - they carried the
imprimatur of one of the Street's oldest and most storied firms. The funds
marketed themselves with the implicit backing of Bear Stearns and played up the
fact that they were run by its experts in mortgage-backed securities. Now
investors are left with a troubling question: If they can't count on big,
well-established firms to operate hedge funds properly, whom can they count on?"
Goldman - is that a realised gain or just paper profits?
Finally, the UK's storied Daily Telegraph reported yesterday that Goldman Sachs was in a row over "paper profits". According to the Telegraph, "analysts cited by Fortune Magazine claim that almost a third of the bank's revenue came from "short" positions on the lowest tier or mortgage derivatives and other forms of toxic debt. These assets are extremely hard to price, and were not in fact realised. More that $2.62 billion of declared profits were made entirely on house estimates at the underlying value."
Responding to these claims, the Telegraph notes that "A Goldman Sachs spokesman said suggestions of inflated paper gains were preposterous.
do this for a living. It is impossible to manage your risk if you don't
know the value of your assets," he told The Daily Telegraph."
Well, in our humble opinion at least, that's certainly true. But - it might well be impossible to "know the value" of some securities held within today's hedge funds and, of course, in today's investment banks. Then, of course, it's also impossible to manage the risk. Indeed, we're prompted to offer a twist on a well known saying:
You can fool everyone as to the value of hard to value securities some of the time. You can fool some people as to the value of hard to price securities all of the time. But you can't fool everyone all of the time....