Case Studies and Risk Management in Commodity Derivatives Trading

Case Studies and Risk Management
in Commodity Derivatives Trading
January 2011

Hilary Till

Research Associate, EDHEC-Risk Institute
and Principal, Premia Capital Management, LLC

Author’s Note:
This is the pre-peer-reviewed version of the following article:
Till, H. (2008), “Case Studies and Risk Management Lessons in Commodity Derivatives Trading,” a
chapter in Risk Management in Commodity Markets: From Shipping to Agriculturals and Energy
(Edited by H. Geman), Chichester (UK): John Wiley & Sons Ltd., pp. 255-291, which has been
published in final form at: http://www.wiley.com/WileyCDA/WileyTitle/productCd-0470694254.
html.
EDHEC is one of the top five business schools in France. Its reputation is built on the high quality
of its faculty and the privileged relationship with professionals that the school has cultivated
since its establishment in 1906. EDHEC Business School has decided to draw on its extensive
knowledge of the professional environment and has therefore focused its research on themes that
satisfy the needs of professionals.

2

EDHEC pursues an active research policy in the field of finance. EDHEC-Risk Institute carries out
numerous research programmes in the areas of asset allocation and risk management in both the
traditional and alternative investment universes.

Copyright © 2011 EDHEC

Risk management in commodity futures trading takes two different forms, depending on whether
trading is done for a commercial or a purely speculative enterprise.
In a commercial enterprise, the rationale for trading activity is usually to “optimise the value
of physical assets;” and the returns and risks from this activity would be expected to be a small
fraction of the enterprise’s overall profits and losses. One would include BP’s trading activity in
this category, for example.
Commercial and investment banks also engage in commodity derivatives trading, historically to
facilitate their overall business in financing natural-resource producers. This is arguably the case
historically with Canadian commercial banks.
For commercial enterprises, the important aspects of risk management are in adhering to
regulatory rules and laws, and in establishing strict operational policies and procedures over
every facet of risk-taking activity.
For a purely speculative participant, the emphasis is almost entirely on market risk-management.
The barriers-to-entry in futures trading are remarkably low: strictly speaking, a participant solely
needs a quote device to track the markets and a Futures Commission Merchant (FCM) to execute
and clear one’s trades. Arguably, the tail risk on a futures trading position is ultimately the
responsibility of an FCM.
It became ingrained in the minds of financial-market participants that should fixed-income or
equity markets ever have extreme dislocations, they could ultimately rely on a “central-bank
put” underwritten by either the Federal Reserve Board (Fed) or the European Central Bank. On
December 12th, 2007, for example, the Fed unveiled the Term Auction Facility (TAF) to enable
depository institutions in the U.S. (and indirectly in Europe) to acquire short-term funds against
a wide variety of capital. Further on March 11th, 2008, the Fed created the Term Securities
Lending Facility (TSLF) whereby primary dealers, including investment banks, could borrow
Treasury securities against a wide variety of mortgage-related securities. With these and other
actions, financial-market participants again assured at the time that the too-big-to-fail doctrine
still held.
For commodity speculators, though, there is no TCAF (“Term Commodity Arbitrageur Facility”).
Instead, commodity speculators are forced to rely on disciplined risk management. The financial
writer, Ralph Vince, goes so far as to recommend that before studying the mathematics of
money management, one should consider what would happen if the prospective trader suffered
a cataclysmic loss:
“Take some time and try to imagine how you are going to feel in such a situation. Next, try to
determine what you will do in such an instance. Now write down on a sheet of paper exactly
what you will do, who you can call for legal help … Do it now …. ” (Vince 1992).]
Many experienced traders have noted how ephemeral trading strategies are, or at least, how all
strategies have life-cycles: “Just when you think you found the key to the market, they change
the locks”, declared the late Gerald Loeb, who was a highly successful financier and founding
partner of E.F. Hutton, as quoted in Cashin (2008).
As a matter of fact, Weisman et al. (2007) have quantified one of the consequences of Loeb’s
observation. The expected draw-down for a strategy is positively related to how consistently
profitable a strategy is, if a threshold of returns is constantly demanded. In the words of
Weisman et al., the markets have “periodic market efficiency,” which is another way of saying all
strategies have a limited lifespan. The “tail loss”, when a strategy finally (and inevitably) outlives
3

its usefulness, can be found to be:
Loss = [(Demanded Returns) * Probability of the Strategy Succeeding / Probability of the Strategy
Failing].
For strategies that target an absolute level of return, the natural consequences of this demand
are that (1) losses are proportional to wins; and (2) losses are inversely proportional to their
probability of occurrence, as explained by Weisman and his colleagues.
Ethical issues do not arise when Weisman et al.’s analysis is applied to proprietary trading firms
since in this case it is the partners’ capital at risk. The partners accept that draw-downs are
endogenous to the trading strategy. As a result, they may not target absolute returns, knowing
that trading strategies are fleeting. As Eagleeye (2007) wrote, “One can manage risk … [but] one
can’t demand a threshold return from the market.” Enduring proprietary trading firms instead
typically target risk.
Now, there are severe consequences to Weisman’s analysis for investors in hedge funds, who have
historically based their investment decisions on past track records, which may not be predictive
of future results; and who pay hedge-fund traders based on short-term results, with no clawbacks of fees if the strategy suffers disastrous results.
We can take an example from the natural gas futures markets to illustrate the negative consequences
to Weisman’s observation of the “dangerous attraction” to absolute-return targeting.
Figure 1 - Natural Gas Bear-Calendar Spread P/L, 1-Month Horizon, January 2004 through August 2006

Source of Data: Bloomberg.

Figure 1 illustrates how consistent a strategy of trading natural gas bear calendar spreads was in
the spring of 2004 through the spring of 2006. A “calendar spread” consists of taking offsetting
positions during the different delivery months of a particular futures contract. A “bear calendar
spread” consists of taking a short position in a nearer-month futures contract while simultaneously
taking a long position in a later-delivery contract of the same futures market.

4

By early summer 2006, the profitability of this strategy had declined by about half of the
performance of the previous two years. If the commodity futures trader had responded by
doubling up his or her position size (to try to maintain an absolute-return target), then in
July and August of 2006, that trader would have sustained losses about twice the size of the
trader’s year-to-date profits. The significance of such a loss is that when a trader’s risk-andreturn results differ dramatically from client and/or prime-broker expectations, this can set off
a “critical liquidation cycle” where client redemptions and/or additional demands for collateral
from creditors cause a trader to liquidate positions in a distressed manner, which can then cause
further losses that imperil a fund’s survival, as both the fund’s investors and creditors lose faith
in the manager. This process is mathematically modelled in De Souza and Smirnov (2004) as being
like a short barrier put option.

Keeping Weisman’s and De Souza and Smirnov’s analyses in mind, perhaps one should accept
that individual trading strategies may not be enduring, but perhaps a trader’s risk-management
methodology can be enduring, instead.
A number of studies have indirectly verified this latter point. The fund-of-hedge-funds investor,
David Gordon, found that while pre-investment returns for managers had no predictive value, as
discussed in Gordon (2003a), it was different for risk:
“Historical standard deviation tends to be somewhat helpful in predicting future risk. The
correlation between pre-investment standard deviation … [versus] downside deviation and
maximum drawdown during the subsequent period of investment is [statistically] significant.”
(Gordon 2003b).
Further, Kat and Menexe (2003) found that the historical value of a hedge-fund manager’s
track record is precisely in its risk characteristics; they found that the standard deviation of a
manager’s returns (and the manager’s correlation to the stock market) was what persisted across
time, but not manager performance itself.
Interestingly, for institutionally-scaled hedge funds, the publicly available information on these
funds is precisely in the quality of their risk-management-and-monitoring infrastructure. This
was the message from the extensive Moody’s operational reports on Chicago-based Citadel
Investments and London-based Brevan Howard, which were both $15 billion-plus hedge-fund
institutions, as at the date of this article. These reports were made available on Moody’s website
in 2007 for accredited investors, and are listed in this article’s references section under Gains
(2007) and Lahav (2007).
So perhaps it is not controversial, after all, to state that risk management is the most important
aspect of a futures trading operation. In this article, we will discuss the risk-management lessons
from a number of recent trading debacles with the hope that the reader will thereby be able to
avoid such mishaps in their own professional lives.
In the following, we will briefly discuss the apparent risk-management lapses at three large
institutions involved in commodity derivatives trading; these lapses were mainly operational in
nature rather than market-risk problems per se. This section will then be followed by a discussion
of the market-risk lessons garnered from trading debacles that have occurred from 2005 through
the spring of 2008.

Institutional Risk Management
Regulatory and Legal Risk
In 2007, BP ran afoul of market-conduct laws and rules, as enforced by the Commodity Futures
Trading Commission (CFTC) and the U.S. Department of Justice (DOJ), for trading activities of the
previous five years. In one particular case, the civil and criminal fines far exceeded the marketrisk of the activities, illustrating where the risk-management priorities need to be for large
participants in the commodity markets.
There is a strict body of law prohibiting market manipulation by commodity traders, especially
when retail customers are put at risk. According to the CFTC (2006a, 2006b, 2007b), during
the spring of 2004, traders at BP Products North America cornered the February 2004 physical
propane market at a particular delivery location in Texas. This delivery location ultimately serves
customers throughout the Midwest and Northeast via pipeline. The largest users of propane in
5

the U.S. are from the residential and commercial heating sector as well as from the petrochemical
industry, which uses propane for creating plastics.
A 28th June 2006 CFTC complaint specifically cites recorded telephone conversations where a
senior BP trader discusses whether BP could “control … [a particular propane] market at will”. To
find out whether they could accomplish a corner, “BP employees purchased enormous quantities
of propane to establish a dominant and controlling long position” in physical propane, notes the
CFTC (2006a). The CFTC complaint also notes how senior management at BP consented to the
strategy. For example, in the CFTC (2006b) complaint, the compliance manager at the BP business
unit responsible for propane trading is quoted as approving the propane-purchasing strategy,
but told the traders “to refrain from using certain words in conjunction with … the strategy,
including the word ‘squeeze.’”
From a careful reading of the CFTC complaint, the actual trade construction of BP’s speculative
trading strategy appears to be one of being long propane for physical delivery by the end of
February 2004 while also being short propane for physical delivery by the end of March 2004. BP
had also attempted a similar strategy, in what appears to be a smaller scale in April 2003, again
according to the CFTC complaint.
According to an internal BP document that is posted to the CFTC website, BP actually lost $10
million from their speculative propane strategy. The BP traders were only able to sell a relatively
small fraction of their February 2004 position at the elevated price levels that prevailed at the
end of February, meaning that they had to close out their remaining February longs at much
lower prices prevailing in March 2004.
Because of this trading loss, the BP trading bench had put together a Powerpoint presentation,
“NGL Feb Value Trade[:] Lessons Learned.” This presentation is publicly available on the CFTC’s
website as Exhibit F of the 28 June 2006 complaint; each page of the presentation is stamped,
“BP Confidential.” Based on past historical relationships, the propane traders had expected their
trading strategy’s performance to be in the range of -$5 million to +$15 million. The document
detailed the controls that were in place, and those which needed to be put in place going
forward, to avert unexpected trading losses in the future. The report documents the Market
Value-at-Risk limits, limits on calendar spreads and plans for improved communications across
BP trading units for better information-sharing. Figure 2 summarises the key compliance risks
from the internal BP presentation.
Figure 2 - Compliance: Key Risks As Excerpted from Internal BP Presentation: “NGL Feb Value Trade Lessons Learned”

• Regulatory - No violations under current framework, but could increase the risk of regulatory
intervention;
• Legal/credit – No specific legal concerns identified, but could increase the risk of an “aggrieved
short” failing to make payment or filing a claim for damages;
• Reputational – Primary risk.
Source: Exhibit F of CFTC (2006b).

Amongst the “actions going forward” are a request for the trading unit to have training in
compliance and regulatory matters.

6

On 25 October 2007, the CFTC announced the entry of a consent order in the United States
District Court in Northern Illinois, which settled civil charges against BP. “In a related filing, the
Criminal Division, Fraud Section of the United States Department of Justice also announced the
simultaneous filing of an information and entry into a deferred prosecution agreement with BP
America Inc. based upon the same underlying conduct,” stated CFTC (2007b). The total monetary

sanction against BP was approximately $303 million, “the largest manipulation settlement in
CFTC history,” according to the CFTC (2007b), which included both civil and criminal penalties.
The order found that BP employees had “violated the Commodity Exchange Act’s prohibitions
against manipulating the price of a commodity and cornering a commodity market.”
The key risk-management lesson from this debacle is to have a clear-cut compliance and ethics
program, not just for the trading staff but also for senior management, given how the regulatory
and legal risks can outweigh market risks when engaged in large-scale commodities trading.
In fact, the 25 October 2007 consent order required BP to “establish a compliance and ethics
program, and install a monitor to oversee BP’s trading activities in the commodity markets.”
On the same date, the CFTC also announced an additional settlement against an individual BP
trader. In this case, the CFTC found that a BP gasoline trader had attempted “to manipulate the
price spread between the November and December 2002 unleaded gasoline futures contract
traded on the New York Mercantile Exchange … on October 31, 2002, the last day of trading for
the November 2002 unleaded gasoline futures contract,” according to CFTC (2007c).
The order found that even though BP had 52 more gasoline contracts than were stated as
commercially needed, the individual gasoline trader had “bought an additional 720 November
2002 unleaded gasoline contracts throughout the course of the day on October 31, 2002” in
order to influence the spread between the November and December contracts.
The individual trader was personally fined $400,000 and was prohibited from any employment
activity that required registration with the CFTC.
Again, in this case, the key risk-management lesson is operational. One would conclude that a
large-scale trading operation should have systems in place, which monitor position sizes versus
limits, particularly on the last trading date for a physically-settled contract.
As will be discussed later in this article, U.S. regulators are now more active in fining traders
personally rather than just fining their place of employment. This would lead one to the
conclusion that prospective traders entering into large-scale derivatives trading operations
should be as (or more) knowledgeable about regulatory rules and laws, as they are with
sophisticated market risk-management techniques.
Valuation Risk1
Bank of Montreal was temporarily in the headlines in 2007 for an energy trading debacle. This
case received a lot of attention in the specialty commodity press in the spring of 2007, but then
quickly faded from the headlines, as the subprime-credit-related crisis picked up steam in the
early summer of 2007.
At the end of April 2007, the Bank of Montreal, which is also known as BMO, announced trading
losses of about C$400 million: these losses were later revised upwards to C$680 million. This
sum was higher than the bank’s revenue from trading during the previous year, according to The
Desk (2007b). Unfortunately, BMO’s auditors had found that the bank’s over-the-counter (OTC)
natural-gas book had been seriously “mismarked.”
The bank’s auditors reported that they had never seen such a large discrepancy between the
marks that were used, and market value, according to Mavin (2007). Another way of framing
the significance of BMO’s natural-gas trading loss was that in its filings with the U.S. Securities
and Exchange Commission (SEC), BMO had stated that its average one-day Value-at-Risk in its
commodity book was only C$8.8 million during the quarter that ended on 31 January 2007 (See
Figure 3).
7
1 - This section is based on Till (2008a).

Figure 3 - Total Tradng and Underwriting MVE Summary ($ millions)*

*Explanatory Notes:
- MVE means "Market Value Exposure"
- MVE is an one -day measure using a 99% confidence interval. Losses are in brak ets and benefits are presented as positive numbers.
Source: Bank of Montreal (2007a).

Figure 4 excerpts from a presentation by BMO’s chief risk officer that has been available on the
Bank of Montreal’s website. This presentation is listed in this article’s references section under
McGlashan (2007).
Figure 4 - Histogram of Trading P/L As Excerpted from BMO Financial Group’s Presentation: “Q1 2007 Risk Review”
Frequency Distribution of Daily P&L For Trading and Underwriting at the BAnk of Montreal (November 1, 2006 to January 31, 2007)

Source: McGlashan (2007).

Figure 4 shows a relatively modest frequency distribution of daily trading profits-and-losses
(p/l) for the bank. Clearly, the quantitative method of summarising BMO’s trading risk had fallen
short.
A key reason why there has not been more communication from BMO on this unexpected loss, as
at the plate of this article, is because the bank has had to focus on a number of inquiries from
securities, commodities, banking and law enforcement authorities, according to an SEC filing by
the bank. The bank “is cooperating with all of these authorities”, noted Bank of Montreal (2007b).
Based on publicly available reports, can we say anything about how BMO’s trading loss
amounted to about 100 times its average VaR reported for fiscal year 2006? One cannot say
anythingconclusively since litigation involving the bank’s outsized trading loss was still ongoing,
as at the date of this article, but a number of press reports and lawsuit allegations provide some
indication of what may have gone wrong. Because a number of the facts and allegations in this
case are quite incendiary, this section of the article will carefully document the source of each
statement.

8

• Apparently, in the over-the-counter natural gas markets, it “is highly prevalent … [for] the front
office/trading … [staff to] mark curves for a) implied volatility and b) illiquid basis locations”,
noted a chief risk officer in The Desk (2007d).

• “The Deloitte report [on BMO’s OTC energy valuation] … indicated that some of the prices
used in BMO’s mismarked book of trades were provided by Valhalla-NY-based Optionable, Inc.,”
reported Mavin (2007).
• Optionable is “a brokerage [that] specialized in OTC derivatives for long maturities,” stated
Blanco and Mark (2007). The firm is a public company with its stock trading on an OTC Bulletin
Board, which trades under the ticker, OPBL.
• BMO was Optionable’s largest customer for a long time, according to The Desk (2007c).
• And correspondingly, Optionable was BMO’s principal options broker, again according to The
Desk (2007a).
• “ …. Optionable made private offers of stocks or warrants to traders in exchange for volume
guarantees,” reported The Desk (2007a).
• The lead natural gas options trader at BMO, who was regarded as the “biggest national-gas
options trader in the market, had a close personal relationship with the senior management of
Optionable,” wrote Mavin (2007).
• BMO may have been directly connected to over 80% of Optionable’s revenues, according to a
lawsuit filed in the U.S. District Court of Southern New York (2007).
• The Optionable CEO’s past included being sentenced to 30 months for a felony conviction
on credit card fraud in 1997, and six months for income tax evasion in 1993, reported Leising
(2007a).
• As shown in Figure 5, Optionable’s share price precipitously declined after the revelations of
BMO’s unexpected losses and associated fraud allegations. As Richard Oldfield said in his book
on investing, Simple But Not Easy, “Ethics is not just a county to the east of London.2 Markets
are particularly intolerant of seriously unethical behavior by management, and the revelation of
scandal is something which can be relied upon to cause a collapse in share price.”
Figure 5 - Optionable (OPBL) Share price

“Optionable, Inc., doing business as OPEX, operates a real-time electronic trade matching and brokerage system. The company provides
trading and brokerage services to brokerage firms, financial institutions, energy traders, and hedge funds nationwide.”
Source of Data: Bloomberg

2 - Explanation of Richard Oldfield’s quote: Essex is a county in the East of England.

9



Blanco and Mark (2007) conclude the following about the BMO fiasco:
• “Insufficient checks and balances in the mark-to-market process” appear to be the main reason
for this debacle.
• A contributing factor may have been “deficiencies in the bank’s pricing and risk models in terms
of incorporating the impact of implied volatility changes.”
• “The bank started experiencing heavy losses as implied volatilities came down in the first
months of 2007.”
• BMO’s “chief risk officer noticed [then] that the risk models had some deficiencies in measuring
the risk of long positions in [deep] out-of-the-money … natural gas OTC options.”
Figure 6 illustrates the dramatic swings in implied volatility that are inherent to the natural-gas
options market.
Figure 6 - Natural Gas Implied Volatility (March 1993 to December 2007) Monthly Data Implied from Call Prices

Source of Data: Bloomberg.

According to The Desk (2007c), the lessons thus far for energy-trading participants are as
follows:
• “… [A]lways get your marks from … [large,] legitimate, established brokers, publishers, or
exchanges.”
• “… [M]ake sure that one’s code-of-conduct document for traders is … [sufficient,] and that …
[all traders] have signed it.”
• Include in the code-of-conduct document what sort of broker “perks” are unacceptable for
traders, and “let nothing be left to common sense.”
• Ensure that one’s trading activity is diversified across more than one broker.
Again, like the two BP case studies, we have to conclude that for large-scale commodity-trading
efforts the complexity may not be in market-risk monitoring, but in relatively-simply described
operational controls, which must be rigorously...