Professional Documents
Culture Documents
This book covers the following general topics: development and assessment of theories for
evaluating commodity risk; the role of derivative securities in managing commodity risk; and,
an examination of actual corporate commodity risk management in specific situations. The
primary contribution of the book is the explicit development of the often overlooked connection
between risk management and speculation. The central theme is to demonstrate commodity risk
management decisions require an in depth understanding of the overall commercial strategy
of the firm. To this end, this book aims to provide a unified treatment of important concepts
and techniques that are useful in applying derivative securities in the management of risk
arising in commodity markets.
Geoffrey Poitras
Preface vii
Part 1
Commodity Risk Basics
Part 2
Commodity Risk Management Concepts
Part 3
Risk Management Applications
Notes 367
References 375
Index 393
This book aims to raise an alarm bell about the current state of commodity risk management
and to suggest some helpful improvements. In particular, the rise of commodity-related exchange
traded funds (ETFs) and other commodity-based securities traded on stock markets has extended
the scope of commodity trading to include a large class of traders not directly involved in the
cash commodity market. These largely long side traders operate under the motivation that purely
speculative commodity transactions constitute “investments” within an efficiently diversified
portfolio of assets. After an initial period of commodity price increases driven by the increase
in long side participation of these “new” traders, the underlying speculative motivation for the
commodity transaction surfaces. The inevitable collapse in prices induces dis-hoarding of the
commodity “investment” positions, resulting in sustained periods of distressed commodity prices.
Real losses are imposed on actual commodity producers created by the excess stocks and flows
of the commodity that were encouraged by the increase in prices. Historically, the commodity
markets have suffered severe systemic disruptions in the price discovery process when the hoarding
and dis-hoarding activities of traders not involved in the cash market become too significant a
proportion of commodity stocks and flows.
Despite claims to the contrary (e.g. ITFCM 2008), available historical evidence supports the
view that there has been substantive disruption in the price discovery process in specific commodity
markets from the “excessive” participation of traders not directly involved in the cash market. For
example, consider the debacle in the natural gas market precipitated by the hedge fund Amaranth
Advisors LLC in 2006 or the excess speculation in the wheat market by commodity index traders
identified by the United States Senate Permanent Subcommittee on Investigations (2009). Less
obvious is the unprecedented rise in gold prices from 2005–2010 that has coincided with the rise
of gold ETFs trading on stock markets. Globally destabilizing volatility in oil prices coincided
with the appearance and increasing use of oil ETFs. Sustaining these market developments is the
view that commodities are an “asset class.” Investment bankers, financial advisors and leading
vii
academics all recommend the inclusion of this “asset class” in a “well-diversified” investment
portfolio. Risk management programs such as those associated with the Global Association of
Risk Professionals treat “commodity risk management” as a special case of financial risk
management.
Properly defined, commodities are fundamentally distinct from equity securities and fixed
income securities. Unlike a share in General Electric, which represents a claim against real assets that
produce goods and provide employment, commodities do not earn a physical or pecuniary return.
Those not directly involved in the cash market that are purchasing a commodity for “investment”
purposes are fundamentally confused. By definition, such transactions are “speculative.” More
precisely, a transaction that is undertaken for the purposes of benefiting solely from price changes
is speculative, by definition. Attempts to portray such purchases as “risk management” through
“efficient diversification” are misplaced. Given the central role of commodities in economic life,
such views promote economically destabilizing behavior. This is not intended as a neo-Luddite
argument against the use of ETFs and other innovative exchange traded products. Quite the
contrary. Modern markets are blessed with the liquidity and sophistication to provide products
that significantly expand the commodity risk management universe. However, the ill-advised use
of such products diminishes the potential social benefits such products can provide in terms of
enhanced market liquidity and stability of the price discovery process.
With this background in mind, this book develops and assesses theories for evaluating and
managing commodity risk. This involves describing the use of derivative securities in commodity
risk management, both theoretically and by examining commodity risk management practice in
specific commercial situations. The primary academic contributions of the book are: the explicit
development of the often overlooked and misunderstood distinction between speculation and risk
management; and to demonstrate that commodity risk management decisions can be improved
with an in-depth understanding of the strategic character of decisions involving commodity prices.
This book aims to provide a unified treatment of important concepts and techniques that are useful
in the management of risk arising in commodity markets. Some of the techniques examined are
well known, such as the replication strategies associated with put-call parity arbitrage. However,
extensions to specific situations and the connection to speculative trading strategies are not. In
actual situations, commodity risk management often involves dealing with uncertainty arising
from both price and quantity, an aspect that is either ignored or given too brief a treatment in
conventional risk management texts that typically overemphasize the application aspects of
derivative security contracts and risk measurement methodologies, often treating commodity
risk management as a special topic in financial risk management.
This book is not intended to provide a comprehensive introduction to commodity risk
management. There are many excellent academic sources that contain the relevant background
material. Rather, it aims to provide a constructive critique of both the received theory of com
modity risk management and the real-world practices arising from that theory. By design, this
involves approaching the subject matter from a somewhat different perspective. Considerable
discussion revolves around the impact that uncertainty has on the optimal solution to the risk
management decision problem. More precisely, it is demonstrated that optimal risk management
decisions involve a speculative component. Following a line of thought going back at least to
Frank Knight (1921), the resolution of the uncertainty associated with the speculative element is
a fundamental source of economic profit and, as such, directly impacts the need to integrate risk
management into the corporate decision-making process. Practical illustrations are provided of
strategies employed by firms facing commodity risk in the base metal mining, airline, and oil and
gas industries.
The book is divided into three parts. The first deals with the general framework for commodity
risk management; the second focuses on the theoretical aspects of commodity risk management
decisions; the third deals with three specific practical commodity risk management situations
from publicly traded corporations in base metal mining, airlines, and oil and gas exploration and
production. A considerable amount of finance background is assumed at various points up to, say,
the level of standard introductory texts such as Bodie, Kane and Marcus’s Investments. Where
specialized mathematical knowledge is assumed, this is delegated to appendices. In general, only
basic algebra and calculus are sufficient to understand the theoretical discussion. While the bulk
of this book is aimed at providing material relevant for academic presentations of commodity risk
management, the material in Part 3 is selected with the practitioner in mind. A substantial amount
of the material in that part illustrates by practical example how commodity risk management can
involve exceedingly complicated strategic business decisions.
Throughout its long history, futures trading has been primarily associated with
commodities having major seasonal patterns of production and inventory accumu-
lation and liquidation. Prices of seasonally produced commodities are speculative.
Thomas Hieronymous (1977)
The phrase “and all other goods and articles” involving “contracts for future delivery”
includes a much larger universe of products than those listed, such as: food and fiber
products, e.g. cocoa, coffee and sugar; metals, e.g. gold, silver and copper; energy
products, e.g. crude oil, natural gas and propane; and financial securities, e.g. stock
indexes and fixed income securities.
Legally, the definition used to identify a commodity in US law provides a binding
limitation on the authority of the Commodity Futures Trading Commission (CFTC).
A decades-old conflict with the Securities and Exchange Commission (SEC) over
jurisdictional authority is founded on the definition of a commodity. Are financial
securities to be considered as commodities? Is a derivative contract on a stock index
different from such a contract for wheat or copper? Is commodity risk distinct from
financial risk? Do theoretical tools useful for financial risk management also work
for managing commodity risks? And so it goes. Failure of the CFTC effort in 1998
to extend authority over OTC trading in financial derivatives—a key element in the
financial market collapse of 2008–2009—can be attributed to a definitional dispute over
what constitute “commodity futures” and related derivative security contracts. This
failure reflects the limitations of using a legal definition to capture essential elements of
contracts associated with commodity trading activity.
The legal definition of a commodity embodied in the CEA is driven by the types of
contract associated with the trading of commodities. An alternative to the legal definition
of a commodity is to isolate the essential features that distinguish commodities and
financial securities. More precisely, in contrast to financial securities, commodities are
physical goods that are costly to store and produce. As a consequence, commodities
can generate a convenience yield for stocks of the commodity. The behavior of the
convenience yield will differ across time and commodities. In a derivative security
pricing context, factors such as costs of storage and convenience yield fundamentally
restrict the execution of long dated cash-and-carry arbitrages for physical commodities.
This has direct implications for the liquidity and price discovery properties of long dated
derivative security contracts. The limit of the definition is provided by commodities
such as gold that have near zero storage costs, a plentiful stock of gold available for
lending to short sellers, and cash and carry yields that are close to those of financial
securities. In turn, the strategic risk management problems confronting, say, a silver
mine operator are similar to those of a copper mine, not an investment bank. Something
more is required of the definition.
commodity risk management. In turn, the bulk of these sources seek to extend to non-
financial risks techniques developed for managing financial risks. The result—financial
engineering or portfolio management with physical commodities—is primarily of use
to investment bankers, market makers, and others involved in intermediating the risk
management and investment products on offer to retail and institutional investors. An
airline seeking to manage the price and volume risk of future jet fuel purchases receives
little guidance. Similarly, a corn farmer or feedlot operator worried about future price
movements for corn receives only general guidance with little practical impact. Even
the general guidance being given can sometimes provide dubious recommendations
that fail to capture key economic impacts of the risk management decision on the
underlying strategic business decision problem.
Requiring a commodity to involve costly production and storage provides essential
structure to the search for optimal risk management solutions considered in this book.
Financial “commodities” such as stock indexes, sovereign government debt securities
and Euro-currency deposits are specifically excluded from this definition. The dividing
line between financial and physical commodities is provided by the physical commodity
with the greatest financial use, gold, which is considered a physical commodity in this
book. In other words, in this book “commodity” is used as a shortened form of “physical
commodity.” Examining commodities coming within this definition reveals other
common aspects, such as being consumable, costly to produce and having variation
in supply. In more technical terms, the commodity risk management decision can
involve both price and quantity risk. There are also elements of Keynesian uncertainty
associated with the production, storage and consumption decisions involving specific
commodities that fundamentally impact the risk management decision. Included in
these elements is the increasing financial securitization of commodity transactions,
especially storage, for certain commodities that has had unintended consequences for
the individuals and firms directly involved in those markets, e.g. United States Senate
Permanent Subcommittee on Investigations (USSPS 2007, 2009).
involves two steps. First, a buyer and seller agree on a market clearing price for the goods
involved in the transaction. Second, the transaction is completed, typically with a cash
payment being made in exchange for adequate physical delivery of the goods involved.
In small rural markets and bazaars, both steps occur at the same time. In many large
commercial transactions, time can separate the pricing agreement, the cash settlement
or the delivery of goods. For example, a forward credit sale involves immediate pricing,
delivery at a later date as specified in the forward contract and settlement dates dependent
on the credit conditions extended by the seller.
The traditional subject of “risk” management as portrayed in financial economics,
e.g. Merton (1993), Tufano (1996), Poitras (2002), Chance and Brooks (2010), empha-
sizes controlling risk through diversification, hedging and insurance. Applications of
“free standing” derivative securities to risk management situations are emphasized.
This approach is somewhat misleading, as it disguises the problem of specifying the risk
management situation, and gives the appearance that the risk involved can, somehow,
be managed in a systematic and unambiguous fashion using derivative security con-
tracting. In emphasizing pro-active management techniques, methods needed for risk
avoidance and risk absorption are often not examined. The various risk management
approaches also differ in applicability to specific cases. In some situations, such as the
Tufano (1996) gold mining sample, the firms involved may have little opportunity to
exploit diversification opportunities to manage risk. Situations vary and the identifica-
tion of an optimal risk management strategy depends on the objective function speci-
fied. Similarly, the management of “risks” disguises the importance of the “uncertainty”
contained in random future events. It is difficult to formulate general rules. Even if
general rules can be derived, such rules rely crucially on information about the proper-
ties of the relevant random variables, i.e. the types of risk being managed
The “insurance principle” approach used in actuarial science provides an excellent
source of general insights into certain types of risk management problem, e.g.
Trieschmann (2005). By design, the insurance principle approach examines situations
where only the chance of loss or no loss is considered. This is a restriction on the
properties of the random variables being modeled. As such, there is only partial overlap
with the situations of commodity businesses, where the random variables involved in
risk management, such as corporate profits or changes in shareholder wealth, can take
both positive and negative values. Whereas insurance problems seek to reduce risk, it
may be desirable to increase certain commercial risks if the potential gains significantly
outweigh the possibility and size of loss. Given these qualifications, Vaughan (1982)
suggests the following insurance principle methods for handling the risks faced in
actuarial science: risk can be avoided, e.g. by forgoing the writing of a policy; risk may
be retained, e.g. by self-insuring; risk may be transferred, e.g. through hedging; risk may
be shared, e.g. through the purchase of reinsurance; and risk may be reduced, e.g. by
increasing audit surveillance.
The contrast between the actuarial science approach and that suggested by financial
economics is revealing. There are some close correspondences. Risk reduction can be
The distinction between pure and speculative risks is an important one, because
normally only pure risks are insurable. Insurance is not concerned with the
protection of individuals against those losses arising out of speculative risks.
Speculative risk is voluntarily accepted because of its two-dimensional nature,
which includes the possibility of gain.
It is apparent that the study of risk management requires a careful and detailed
discussion of the definition and classification of the types of risks that are going to be
managed.
periods of up to one year with suppliers, setting in advance the price for products
to be delivered in the future and unit pricing based on an average of commodity
prices over the corresponding period of time. We purchase a significant amount
of green coffee and typically have purchase commitments fixing the price for a
minimum of six months, and we also typically hedge against the risk of foreign
exchange at the same time. We do not generally make use of financial instruments
to hedge commodity prices, partly because of these contract pricing techniques. As
we make purchases beyond our current commitments, we may be subject to higher
commodity prices depending upon prevailing market conditions. While price
volatility can occur, which would impact profit margins, we and our franchisees
have some ability to increase product pricing to offset a rise in commodity prices,
subject to franchisee and customer acceptance, respectively.
In addition, we currently have purchase contracts in place for at least the first
half of 2010 covering key commodities such as coffee, wheat, sugar, and cooking
oils. As we have stated previously, we may be subject to higher commodity prices
depending upon prevailing market conditions and foreign exchange rates at the
time we make purchases beyond our current commitments. Higher commodity
costs could also impact earnings from our joint venture operations.
securities of those firms. Applying fundamental valuation techniques to assess the com-
modity risk management strategies of specific firms leads to consideration of the defini-
tional distinction between speculation and investment that is fundamental to the work
“value investors,” such as the prominent equity security analysts, Benjamin Graham
and Warren Buffett. This distinction has a helpful extension in commodity risk manage-
ment where, as the chapter-opening quote by Hieronymous reflects, decisions involving
commodity price risk are inherently speculative. The widespread application of finan-
cial risk management techniques to manage commodity risks ignores this distinction,
e.g. Stoll and Whaley (2010).3
The distinction between speculation and investment in fundamental analysis roughly
corresponds to the notional distinction between “speculation” and “hedging” identified
in CFTC (2008, p.2) and numerous other sources. For example, in the Intelligent Investor
(1949), Graham observes:
The distinction between investment and speculation in common stocks has always been
a useful one and its disappearance is a cause for concern. We have often said that Wall
Street as an institution would be well advised to reinstate this distinction and to
emphasize it in all dealings with the public. Otherwise the stock exchanges may
some day be blamed for heavy speculative losses, which those who suffered them
had not been properly warned against.
This statement has a direct parallel in the commodity ETFs currently offered on stock
exchanges to provide “efficient diversification of investment portfolios.” The eventual
collapse in prices precipitated by “investor” dis-hoarding contains the “seeds of
blame” for heavy speculative losses by commodity producing firms that made unhedged
capital investments based on higher “expected” prices, sustained by the implicit
assumption that investor demand would continue unabated over the life of the capital
investment.
Despite being a widely used concept, subtle and not-so-subtle differences in the
definition of speculation abound. For example, the American Heritage Dictionary
recognizes two general uses. First, there is the colloquial usage: “1. Contemplation or
consideration of a subject; meditation; 2. A conclusion, opinion, or theory reached by
conjecture; 3. Reasoning based on inconclusive evidence; conjecture or supposition.” In
addition, there is the commercial usage: “1. Engagement in risky business transactions
on the chance of quick or considerable profit; 2. A commercial or financial transaction
involving speculation.” Of these, the most relevant to commodity risk management—
“Engagement in risky business transactions on the chance of quick or considerable
profit”—fails to adequately recognize fundamental elements of speculation. Similarly,
the definition of speculation given on the popular internet site Investopedia focuses on
expanding the commercial dictionary definition: “speculation: The process of selecting
investments with higher risk in order to profit from an anticipated price movement.” The
following explanation is also provided: “speculation should not be considered purely
“Hedging” is the term used to describe the activity of someone who is using the
futures market to manage the price risks associated with the sale, purchase, or use
of a commodity. Hedging has sometimes been described as an activity undertaken
by the producer, merchant, or end-user of a commodity as opposed to a speculator
who does not produce, use, or consume the commodity. Since there are numerous
strategies and approaches to managing price risks, however, it often is impossible
to distinguish, from an economic perspective, whether a particular transaction
is, in fact, hedging or speculation. The line between minimizing risks—which is
what the term “hedge” connotes—and maximizing profits—which is what the term
“speculation” connotes—can be exceedingly difficult to draw.
Table 1.1 captures the current distinctions between different types of trader made by the
CFTC in the useful weekly Commitment of Traders (COT) report that decomposes the
Tuesday open interest.
In 2006, the CFTC undertook to make changes to the COT report. The reasons for
this review were given as follows (CFTC 2008a):
Over time, for some commodities the nature of the positions carried in the COT
reports has changed significantly. Prior to 1991, both the long and the short side of the
commercial open interest listed in the COT reports represented traditional hedgers
(producers, processors, manufacturers or merchants handling the commodity or
its products or byproducts). Since that time, trading practices have evolved to such
an extent that, today, a significant proportion of the long-side open interest in a
number of major physical commodity futures contracts is held by so-called non-
traditional hedgers (e.g. swap dealers), while the traditional hedgers may be either
net long or net short (more often, the latter). This has raised questions as to whether
the COT report can reliably be used to assess overall futures activity by persons
who are directly involved in the underlying physical commodity markets.
Percent of Open Interest Held by the Indicated Number of the Largest Traders
By Gross Position By Net Position
4 or Less Traders 8 or Less Traders 4 or Less Traders 8 or Less Traders
Long Short Long Short Long Short Long Short
All 18.9 21.4 32.6 27.9 18.1 18.1 29.6 22.3
Old 25.7 29.7 34.9 37.3 23.5 29.3 32.5 36.6
Other 26.1 18.5 38.8 26.4 25.1 15.7 36.7 20.7
11/16/2012 6:43:47 PM
Book 1.indb 13
Crude OIL, Lightsweet – New York Mercantile Exchange
Disaggregated Commitments of Traders – Futures Only, December 6, 2011 Nonreportable
Producer/Merchant/ Reportable Positions Positions
Open Processor/User Swap Dealers Managed Money Other Reportables
Interest Long Short Long Short Spreading Long Short Spreading Long Short Spreading Long Short
11/16/2012 6:43:48 PM
14 Commodity Risk Basics
Following the 2006 review of the COT, new categories for “managed money” and “swap
dealer” categories appear in the revised COT (see Table 1.1). These additional categories
represent a substantive improvement over the “legacy” COT reports that gave results
only for “Commercial,” “Non-commercial” and “Non-reportable” positions.
B. Commodity Characteristics
Overview
It is conventional to identify and categorize various risks that may be encountered in
financial and “non-financial” commercial operations. Important categories typically
include: credit risk; market risk; liquidity and funding risk; operational risk; regulatory
and legal risk; and reputation risk. While such categorization of risk is insightful
for financial firms, it is market risk that usually predominates in commodity risk
management, e.g. Bartram (2005). To be sure, as discussed in Section 1.3, there are
significant operational risks. In addition, certain commodity producers, e.g. heap-
leach gold miners or oil sands miners, face substantial regulatory and legal risk. Junior
mining companies, small grain and oil seed farmers, and junior oil and gas exploration
companies often face substantial liquidity and funding risks. And so it goes. Such risks
are specific to a particular commercial situation. What is common to commodity risk
management situations is a concern about “market risk” or, in other words, the behavior
of the commodity price over time. As a consequence, “commodity risk management”
requires understanding of the basic characteristics of commodity prices relevant to the
commercial situation at hand.
The number of commodity prices in the world is overwhelming, especially if
variation in commodity quality and location are taken into account. Some method of
classification and simplification is indicated before commodity price characteristics
can be identified. For this purpose, standard “complex” classifications employed by
the derivative security industry can be used. These classifications focus on sectors of
commodity production: the agriculture complex; the energy complex; and the metals
complex. Each “complex” contains a number of categories. For example on October 30,
2011, within the agriculture complex on the Chicago Mercantile Exchange (CME) there
is contract trading for: grains and oilseeds; livestock; dairy; forestry; soft commodities;
and commodity indexes (see Fig. 1.1). Tunneling into, say, “livestock” reveals contracts
for live cattle, feeder cattle and lean hogs. The date is relevant because the contracts
listed for trade change over time, sometimes significantly. For example, on July 18, 2011,
the CME de-listed the frozen pork bellies contract due to lack of trading volume. Yet,
during the 1960s the success of this contract was primarily responsible for the survival
of the CME as an exchange following the Congressional ban imposed on trading the
onion futures contract, e.g. Castaldo (2011).
Classifying commodities into “complexes” is not without difficulties. Commodities
within a given complex share some basic price characteristics, e.g. commodity prices
in the metals complex are usually sensitive to levels of industrial production. However,
there is sufficient heterogeneity that it is usually necessary to consider characteristics
specific to a particular commodity within a complex. For example, in the energy
complex the characteristics of natural gas prices have been distinct from those of crude
oil in the last three years (see Figs 1.2 and 1.3) due to, among other factors, rapid change
in natural gas drilling technology. Consider also the impact of geographical location.
Different infrastructure requirements for production, transportation and consumption
impact on natural gas pricing and availability differently than for crude oil (see Fig. 1.4
for oil pipeline infrastructure). With this in mind, only price characteristics of specific
commodities relevant to the discussion in Parts 2 and 3 of this book are examined in this
section. More precisely, the commodities considered are within: the energy complex—
crude oil and natural gas; and the metals complex—base metals and precious metals.
Some information on the agricultural complex—primarily US grains and oilseeds—is
provided in Section 2.3.C.5
The conversion of resources into reserves depends fundamentally on the price of the
commodity involved. Significant changes in prices can change the mineable supply
dramatically, both for individual companies and for the whole industry.
Calculation of resources and reserves for oil and gas companies follows much
the same lines as for mineral deposits. However, the codes employed are typically
those set out by the American Society of Petroleum Engineers and the World Petr-
oleum Council. Reserves can be proved, probable or possible. Proved reserves have
reasonable certainty of economic recoverability. Probable reserves are less certain than
proved reserves, but are more likely than not to be economically recoverable. Possible
reserves are associated with insufficient data about the specific reservoir. Recognizing
that oil often occurs in familiar basins and geological formations gives possible
reserves more meaning than would be the case for most mineral deposits. Contingent
resources have been discovered but are not yet economically viable, while prospective
resources are potentially recoverable but not yet subject to drilling activity. There is
also allowance for “unproved reserves,” which for technical, political or regulatory
reasons cannot be classified as proven. The relevance of these various definitions
for the strategic decision making of specific firms is apparent when evaluating the
economically important Canadian oil sands deposits.
of the clay or sand from the bitumen and obtaining sufficient viscosity in the bitumen
for extraction requires heating, which adds substantially to the extraction cost.
The oil sands “reserves” are located in the Cold Lake, Athabasca and Peace River
areas of Alberta (see Map 1.1 and Map 1.2). Only about 18 percent of these “reserves”
are surface reserves that can be mined. The surface reserves are located primarily in the
Athabasca region situated north of Fort McMurray. Even though the size of the surface
reserves is small relative to the total oil sands resource, the substantially higher recovery
rate of bitumen from surface reserves and the high yield of synthetic crude oil (SCO)
from the bitumen makes these reserves a larger exploitable portion of the oil sands
resource base.
Casual inspection of Fig. 1.7 and Table 1.2 reveals that the widely respected estimates
of world oil reserves produced annually by BP reflect the difficulties of classifying
resource or reserve value to the oil sands. A significant amount of the increase for
Canada between 1998 and 2007 in Table 1.2 is due to the addition of about 16 billion
oil sands reserves to the Canadian total, which is below that for Libya and the Sudan.
This compares to reserve estimates of around 175 billion barrels of economically
recoverable crude oil produced by the Alberta government, an estimate accepted by
the US Energy Information Administration. If accepted, this would place Canadian oil
reserves behind only Saudi Arabia. This highlights the difficulties with straightforward
assessment of the world oil market by examining supply and demand factors. The actual
supply of economically recoverable oil varies with the assessment of market price and
the costs of oil recovery. As this spread widens, previously uneconomic reserves become
feasible. This tipping point toward profitability occurred for the early Alberta oil sands
producers when WTI crude oil traded above $20 per barrel.
important to realize that Alberta’s oil sands and heavy oil deposits are not
homogeneous . . . Among the important ways in which deposits differ are: specific
gravity (some crudes are heavier than other oils), bitumen concentration (the
proportion by weight or volume which is bitumen, ranging from 1 percent to
18 percent) and depth (where shallow deposits—usually up to 75 metres deep—are
regarded as amenable to mining operations).
Bitumen is a sticky, viscous type of crude oil with low API gravity compared to other
types of crude. A light sweet crude, such as the West Texas Intermediate (WTI) that is
the deliverable commodity for the widely followed NYMEX crude oil futures contract,
has an API gravity of not less than 37º and not more than 42º, while a heavy crude, such
as that produced in Saudi Arabia, has an API gravity of 10º to 20º.6 Even with the sand,
clay and water removed, bitumen still has an API gravity well below 10º. The low-sulfur,
upgraded Syncrude sweet blend (SSB) produced by Syncrude prior to 2007 obtained
an API gravity of only 30º, with the Syncrude sweet premium blend produced since
2007 being slightly higher. Because oil from conventional drilled wells has to travel
up the well bore in order to be recovered, the API gravity determines the potential
recovery rate. Light sweet crude deposits have a recovery rate of about 30 percent, while
heavy crude deposits have recovery rates up to 20 percent using conventional recovery
methods. The API gravity of bitumen depends on the depth of the deposit. In addition,
the composition of refinery output for heavy and light crude oils results in higher-value
products being obtained from light crude (see Fig. 1.5).
Unlike conventional crude oil and heavy oil deposits, which are accessible only by
drilling, approximately 18 percent of the Alberta oil sands are at depths from the surface
down to 245 feet. Even though this surface oil sand bitumen has an API gravity of from
0º to 10º, “the first two commercial oil sands mining plants built in Alberta . . . produce
a little over 0.8 of a barrel of syncrude per barrel of bitumen” (Atkins and MacFayden
2008). Allowance made for loss of bitumen in the slurrying process, which depends on
the richness of the oil sands being mined, produces an additional loss of about 15%.
As a result, about 65 percent of the crude oil contained in the surface oil sand can be
recovered because the resource at this depth range can be open pit mined. Oil sands at
depths below 245 feet usually have an API gravity about the same as heavy crude and
have to be recovered through drilling and in situ (in the formation) methods. While
these deposits have a recovery rate of between 25 and 75 percent, the costs of loosening
up the solid oil sand for extraction makes for much higher recovery costs than for oil
sands mining, which, in turn, is considerably more expensive than conventional oil.
Though a number of extraction techniques of bitumen deposits below 245 feet are being
explored, steam assisted gravity drainage (SAGD) has proven to be the most popular to
date (see Fig. 1.8). This technology requires considerable energy to generate the steam.
There are few alternatives to the use of steam, e.g. toe-to-heel injection method used at
the Whitesands project.
At about 24 miles from the fork (of the Athabasca and Clearwater Rivers) are some
bituminous fountains into which a pole of 20 feet long may be inserted without
the least resistance. The bitumen is in a fluid state and when mixed with gum,
the resinous substance collected from the spruce fir, it serves to gum the Indians’
canoes. In its heated state it emits a smell like that of sea coal.
The first geological survey of the area was conducted in 1875 with further exploratory
efforts made in 1882 and 1889. The first attempts to commercially develop the oil
sands in the Athabasca region were initiated in 1906 by the entrepreneur Alfred
von Hammerstein. This project was based on the assumption that the surface bitumen
was originating from pools of oil deep beneath the surface. In an attempt to locate
these pools, a series of well holes were drilled in the area north of Fort McMurray
where the bulk of surface bitumen in the Alberta oil sands is located. This drilling
activity continued from 1906 to 1917, with a total of 24 wells being drilled. Due to
the faulty initial assumption, none of the drills holes was successful at finding oil.
However, the drilling activity did discover salt deposits, which became a major
industry in the Fort McMurray area for 50 years, until it was eclipsed by the oil sands
developments.
Since the early attempts at locating conventional oil deposits by drilling, there
have been ongoing attempts to identify a commercially viable method of extracting
the surface oil in the region. The process that is used in Syncrude and other heavy oil
mining projects involves the use of a hot water flotation method to separate the bitumen
from the sand (see Fig. 1.9). The development of this method for processing oil sands has
a long history. Early on-site research on this method began around 1913 when Sidney
Ells, a federal Department of Mines engineer, began a series of experiments on the
viability of applying this technique to the oil sands. Ells continued this work until 1945.
One of these experiments involved shipping mined bitumen to Edmonton for use as
road paving material. While it was demonstrated that paving material was a feasible use
for the separated oil sand, the costs of application were not competitive with imported
asphalt and the project was dropped. Another commercial attempt to develop the oil
sands using the hot water floatation method began in the 1920s when an entrepreneur,
R.C. Fitzsimmons, used the process to produce bitumen for roofing and road surfacing
at a plant near Bitumount, 80 kilometers north of Fort McMurray. Financial difficulties
eventually forced Fitzsimmons to sell the operation in 1942.
Much of the early history of oil sands development using the hot water flotation
method involves research done by the federal and Alberta governments. In addition to
the work of Sidney Ells, during the 1920s, Dr Karl Clark, a scientist with the Alberta
Research Council, also conducted experiments with a hot water flotation process,
which involved mixing oil sand with hot water and aerating the resulting slurry. This
separates the oil sand into a floating bitumen froth and a layer of sand that settles
to the bottom of the holding tank. In 1948, the Alberta government acquired the
Fitzsimmons plant to investigate the application of extraction methods, such as those
investigated by Clark, using large-scale equipment. By 1949, the plant was processing
450 tonnes of oil sand daily. While an experimental success, the plant was closed
because the Alberta government was not motivated to launch a commercial venture.
Data from the experiments were used as the basis for a major study of the viability of
commercial production. The resulting Alberta government report indicated that crude
oil production from the oil sand could be a profitable venture. Though this created some
commercial interest during the 1950s, it was not until the 1960s that major commercial
ventures began to come on-stream.
The research of the Alberta government about the possibility of commercially
viable extraction of crude oil or oil byproducts from the oil sands was not without
practical foundation. In 1936, an entrepreneur, Max Ball, founded Abasand Oils Ltd,
which used a plant west of Fort McMurray to produce diesel oil from the oil sands.
Despite the relatively small scale and commercially unproven technology, there was
a considerable interest in his project, especially during World War II when diesel oil
was in short supply. During the war, the plant was sold to the federal government, but
soon thereafter the plant burned down and the project died. Commercial oil production
from the oil sands did not come to fruition until the Alberta government launched
an initiative in 1962 that resulted in the Great Canadian Oil Sands (GCOS) project.
Though GCOS experienced a number of ownership changes after its incorporation and
prior to the construction decision, by 1963 ownership of the project resided with the
Sun Oil Company (later Suncor Energy). It was this Suncor project that came on-stream
in 1967 to become the world’s first commercial oil sands operation. The GCOS was soon
followed by the Syncrude consortium, which was formed in 1964.
bbd and a $36.76 per barrel cost. The numbers for the first three quarters of 2011 are
311,000 bbd and $36.56 in costs per barrel. Syncrude output is usually below potential
output for a number of reasons, including the need to conduct comprehensive scheduled
turnaround, and circulation problems in the coker. Costs per barrel do fluctuate. For
example, for the second quarter of 2009, COS reports: “operating costs averaged $50.23
per barrel compared with $41.92 per barrel in 2008. For the six-month period, per
barrel operating costs were $43.66 and $38.90 in 2009 and 2008, respectively” (COS,
Q2-09 report). Lower production volumes are significant because “Syncrude’s operating
costs are largely fixed, so changes in production volumes significantly impact per barrel
operating costs.”
Syncrude is not the only operator working the oil sands, but it is the largest and,
together with Suncor, one of the two oldest (see Map 1.B). The original oil sands project
run by Suncor Energy Inc. produced about 228,000 barrels per day (bpd) in 2008
compared to 235,600 bpd in 2007. Recently, a number of other on-stream bitumen
mining projects in the same area have come on-stream, including: the Athabasca
Oil Sands Project—also called Shell Albian Sands—a joint venture by Shell Canada
(60 percent), Chevron Canada (20 percent) and Marathon Oil Corp. (20 percent), with
the Muskeg River Mine entering production in 2002 and full production achieved when
the Scotford upgrader in Fort Saskatchewan, Alta., came on-stream in 2003. With an
original design capacity of about 155,000 barrels per day, there is currently a 100,000
bpd expansion under way that received regulatory approval in 2006. Other projects
just completed include the Horizons Oil Sands project owned by Canadian Natural
Resources (CNQ), which has an eventual project design capacity of about 255,000 bpd.
The initial stage has a design capacity of 110,000; it came on-stream in February 2009
at a cost of just under $10 billion and was producing an above design capacity 120,000
bpd by July 2009.
In addition to completed and currently producing projects, there are a number of
projects at various stages of the development process. Escalating costs and other factors
have contributed to a significant slowdown in development. Consider the Fort Hills Oil
Sands Project, where the original leases were owned by two junior oil and gas explorers:
78 percent by True North Energy, a fully owned subsidiary of Koch Industries; and
22 percent by UTS Energy. In April 2004, UTS agreed to acquire 100 percent control
from True North. Not long after acquiring the company, UTS was able to get companies
capable of bringing an oil sands project to completion involved by having PetroCanada
purchase a 60 percent share and Teck Cominco a 20 percent share. The initial True
North project plans were to have the Fort Hills project producing approximately 95,000
barrels per day starting in 2005, with plans to ship the bitumen to a Koch refinery
near St Paul, Minn., for processing. These plans have changed a number of times. The
current initial stage plan is for a bitumen mining operation only, with design capacity
of 160,000 bpd and use of Suncor upgrading facilities. The ultimate objective was a
280,000 bpd operation and upgrader in place for 2014. Due to an increase in costs of
50 percent or more, which took final costs to over $20 billion, the Fort Hills mining
project was delayed indefinitely in September 2008.
Oil sands mining operations are immense—easily one of the largest materials
handling operations in the world. The Syncrude operation alone has an output equal to
approximately 13 percent of Canada’s petroleum requirements. The bitumen is mined
using open pit techniques. A site is prepared for mining by removing the “overburden.”
This requires the use of “supersized” trucks and shovels to expose the oil sands. Due to
the progressive development of mining and extraction technology, original mines of
Syncrude and Suncor use somewhat different and more costly mining and extraction
techniques than more recent operations. For example, the original Base Mine at
Syncrude uses two draglines to mine the oil sand, which is piled in windrows along
the sides of the mine pit. The oil sand is dug from the windrows by bucketwheels and
placed on a conveyor system that transports the material to the extraction plant. The
North Mine was the next to be developed. The extraction technology at this mine uses
trucks and shovels in conjunction with two hydrotransport pipelines. The ore from the
truck and shovel operation is crushed and then mixed with hot water. This produces an
oil sand slurry that is screened to remove large materials and then pumped through the
hydrotransport pipeline to the extraction plant.
The mining technology used at the Base Mine during the extraction phase differs
from that used at the Aurora mine. The bitumen from the Base Mine is extracted
from the oil sand at the Base Plant where the oil sand from the Base Mine is fed into
tumblers—large horizontal rotating drums—and mixed with steam, hot water and
caustic soda in preparation for bitumen separation. The slurry from the tumbling phase
is screened to remove large rocks and other such material before being pumped into
four primary separation vessels (PSVs). At this stage, feedstock from the slurry coming
from the North Mine hydrotransport system can be added to the process. A distributor
directs the North mine slurry to any or all of the four PSVs, where it supplements feed
from the tumblers. In addition to the Base Mine extractor, there is also an extraction
plant at Aurora North. This extraction plant uses a low energy extraction process
developed by Syncrude that is designed to operate at 25° Celsius. Included in the
Syncrude technological innovations used at this plant are the hydrotransport of high-
density slurry, froth underwash, lean froth recycle, and air injection to enhance flotation.
The resultant froth from Aurora North is transported to the Base Plant via a pipeline
for further processing at the froth treatment plant. At this stage, the frothy bitumen
from the extraction plant is diluted with naphtha and cleaned using a combination of
centrifuges and inclined plate separators.
Following the bitumen extraction phase, the naptha treated bitumen froth enters
the upgrading phase, which eventually results in the output of SCO. The first step
in upgrading is naphtha recovery for recycling back to the froth treatment plant.
The naptha reduced bitumen is then fed in to two fluid cokers and one LC-Finer
hydroprocessor for upgrading. The LC-Finer breaks down bitumen by adding hydrogen
with the aid of a catalyst to produce gas oil. Gray (2002, p.53) describes the relevance of
this step: “Hydroconversion processes, such as LC-Fining . . ., use catalyst and hydrogen
to control coke formation and maximize yields.” Residuum from the LC-Finer is sent
to the fluid cokers where it is mixed with bitumen. Each of the cokers can process up
to 105,000 barrels of bitumen per day. High temperatures in the coker reactors cause
the cracking or decomposing of the bitumen molecules into various products. The
lighter products, primarily naphtha and gas oils, become the main ingredients of crude
oil. Carbon is rejected in the fluid coking process as coke, some of which is burned to
generate heat for the bitumen cracking process, while the remaining coke is stored in
coke cells. Over time, Syncrude has made improvements in the design and efficiency
of the cokers and the LC-Finer to permit the processing of a heavier feed derived from
the vacuum distillation unit (VDU). Following start-up of this unit in 1999, the gas oil
extracted now is directed to the hydrotreaters.
The products from the cokers, the LC-Finer and the gas oil from the VDU are
processed in hydrotreating units that adds hydrogen using fixed bed catalytic reactors.
This stabilizes the products, removes the nitrogen and sulfur, and reduces product
density, making SCO a highly desired feedstock for oil refineries. This is a crucial step,
as Gray (2002, p.53) observes: “The contributions of various reactions show that sulphur
and nitrogen removed in the coke have the biggest impact on product density. Higher
sulphur and nitrogen removal and lower hydrogen losses increase volumetric yield and
product gravity.” Though all this discussion of extraction and upgrading may seem
more appropriate to a chemical engineer than a commodity risk manager, the relevance
lies in the potential for improvements in SCO quality, and reducing the cost of the
extraction and upgrading process. The Syncrude project requires substantial amounts
of energy to produce steam and hydrogen for the catalytic reactors. Internally generated
fuel gas is used as the primary source of energy to generate electricity and steam, while
natural gas is used mainly to produce hydrogen. Potential energy savings or reduction
in hydrogen loss and the like by technological improvements will translate into gains in
commercial profitability.
What are the physical limits to technological improvements in the upgrading
process? Efforts to answer this question form an active area of research in chemical
engineering. Because a number of different chemical reactions are occurring at once in
the different stages of the upgrading process, it is difficult to predict precise outcomes.
The bitumen entering the Syncrude upgrading process is typically around 7º API with
4.75 percent sulfur and 0.42 percent nitrogen. The LC-Fining hydroconversion process
removes about 65 percent of the sulfur and 15 percent of the nitrogen, but produces
about 5.3 percent light ends. The delayed coking process removes only 37 percent of the
sulfur, little or no nitrogen, and produces about 2.0 percent light ends. Following Gray
(2002, p.53): “any change in the coking process that increases sulphur and nitrogen
concentration in the coke will enhance volumetric yield and product quality . . . if
hydrogen lost to light ends and to coke were halved, the volumetric yield of liquids
would increase 1.8 percent and product gravity would increase by 2º API . . . More
effective catalysts for nitrogen removal would give significant benefit in product yield.”
On balance, despite the potential for some improvement in API gravity and product
yield, there is an underlying problem that adding more hydrogen to improve sulfur
removal also tends to increase the amount of coke produced. Though some future gains
are possible, physical limitations dictate that sizable increases in API gravity or product
yield in the future do not seem too likely.
11/16/2012 6:43:53 PM
Book 1.indb 35
TABLE 1.3 World refined copper production, 2006–2011
2006 2007 2008 2009 2010 2010 2011 2011
p/
Jan-Sep Jun Jul Aug Sep
World Mine Production 14,990 15,483 15,527 15,897 15,989 11,769 11,805 1,324 1,289 1,350 1,344
World Mine Capacity 17,174 18,111 18,740 19,523 19,898 14,862 15,196 1,671 1,732 1,738 1,687
Mine Capacity Utilization (%) 87.3 85.5 82.9 81.4 80.4 79.2 77.7 79.2 74.4 77.7 79.7
Primary Refined Production 14,678 15,191 15,403 15,454 15,786 11,771 11,891 1,335 1,327 1,349 1,366
Secondary Refined Production 2,613 2,743 2,823 2,818 3,249 2,413 2,697 309 319 336 303
World Refined Production 17,291 17,934 18,226 18,272 19,035 14,185 14,588 1,644 1,646 1,685 1,669
(Secondary + Primary)
Seasonally Adjusted Refined 1,585 1,629 1,643 1,630 1,655 1,667
Production - monthly1/
World Refinery Capacity 20,555 21,784 22,723 23,625 23,908 17,881 18,364 2,021 2,096 2,103 2,042
Refineries Capacity Utilization (%) 84.1 82.3 80.2 77.3 79.6 79.3 79.4 81.3 78.5 80.1 81.7
Secondary Refined as % in Total 15.1 15.3 15.5 15.4 17.1 17.0 18.5 18.8 19.4 19.9 18.1
Refined Prod.
World Refined Usage2/ 17,034 18,197 18,039 18,108 19,386 14,613 14,758 1,676 1,625 1,722 1,682
Seasonally Adjusted World 1,608 1,625 1,607 1,592 1,747 1,649
Refined Usage - monthly1/
Refined Stocks End of Period 1,075 970 1,102 1,353 1,193 1,169 1,341 1,325 1,340 1,326 1,341
Period Stock Change 265 –105 132 251 –160 –184 148 18 14 –13 15
Refined Balance3/ 257 –263 187 165 –351 –429 –170 –32 21 –37 –13
Seasonally Adjusted Refined –210 33 36 38 –92 18
Balance4/
LME Copper Price5/ 6,727 7,126 6,952 5,164 7,539 7,097 9,265 9,045 9,619 9,041 8,315
Comex Copper Price6/ 309.42 322.84 313.18 236.54 342.66 325.82 420.69 410.91 440.14 407,83 373.78
Source: http://www.icsg.org/index.php?option=com_content&task=view&id=57&Itemid=60
11/16/2012 6:43:54 PM
36 Commodity Risk Basics
Figure 1.11 Refined Copper Usage by End-Use Sector and Region, 2009
Source: International Copper Study Group, World Copper Factbook 2010
collapse in copper prices that began in 2008. Zinc prices have also not recovered to the
levels reached in 2006, while copper prices have already exceeded previous highs, before
falling back. Though zinc is the fourth most common metal in the earth’s crust, with
more than 50 countries mining zinc ore, much of global production is concentrated
with the largest producers (see Fig. 1.16): China, Australia, Canada, Peru and the
United States. Ore grades for zinc are decidedly higher than for copper, resulting in
The contrast between non-ferrous metals and gold is striking. Figs 1.18–1.20 detail the
three-year, ten-year and twenty-four-year US$ spot gold price. The longest time series
corresponds to the time period following the collapse of the international gold standard.
As reflected in the price chart for gold since the introduction of the most popular gold
ETF (GLD) since late 2004, there is some appealing evidence to attribute the incredible
accelerating run-up in gold prices over the first decade of the 21st century to the rise
of ETF trading. The first gold ETF was listed in Australia in 2003. Originally listed in
2004, the SPDR gold ETF (GLD) is the largest and most liquid gold exchange traded
fund. At the end of 2011, GLD held 1,254.57 tonnes of gold, representing 40,335,690.64
oz, valued at US$63,484,275,822.93. Unlike some ETFs for other commodities that use
futures markets to maintain positions in the commodity, e.g. oil and wheat, almost
all gold ETFs issue shares against physical gold holdings. However, as reflected in the
commitment of traders report (see Table 1.4), there is still considerable long holding of
gold futures by managed money traders. However, the corresponding physical size of
the long position in managed money gold futures on COMEX is less than one-third the
size of the physical holdings of the SPDR gold ETF.
20.6 57.6 13.9 16.0 4.3 28.8 3.5 2.9 11.2 5.3 4.6
11/16/2012 6:43:56 PM
Book 1.indb 42
Table 1.5 Gold Demand by Country, 2010
12 months ended Q4’09 12 months ended Q4’10* Year on Year % chg
Total bar Total bar Total bar
Jewellery and coin invest Total Jewellery and coin Invest Total Jewellery and coin invest Total
11/16/2012 6:43:56 PM
Book 1.indb 43
Table 1.6 Historical Demand for Gold
Tonnes US$bn
Total bar and ETFs and Total bar and ETFs and
Jewellery coin demand similar Technology Total Jewellery coin demand similar Technology Total
2001 3,009 357 - 363 3,729 26.2 3.1 - 3.2 32.5
2002 2,662 340 3 358 3,363 26.5 3.4 0.0 3.6 33.5
2003 2,484 301 39 382 3,207 29.0 3.5 0.5 4.5 37.5
2004 2,616 352 133 414 3,515 34.4 4.6 1.7 5.4 46.2
2005 2,718 393 208 433 3,753 38.8 5.6 3.0 6.2 53.6
2006 2,298 416 260 462 3,435 44.6 8.1 5.1 9.0 66.7
2007 2,417 435 253 465 3,571 54.0 9.7 5.7 10.4 79.8
2008 2,192 860 321 439 3,812 61.4 24.1 9.0 12.3 106.9
2009 1,760 743 617 373 3,493 55.0 23.2 19.3 11.7 109.2
2010 2,060 995 338 420 3,812 81.1 39.2 13.3 16.5 150.1
Source: World Gold Council
11/16/2012 6:43:56 PM
Book 1.indb 44
Table 1.7 Gold Demand by Category 2010
Tonnes 2009 20102 Q4’09 Q1’10 Q2’10 Q3’10 Q4’10 Q1’11
11/16/2012 6:43:57 PM
The Commodity Risk Management Landscape 45
Inspection of Tables 1.5–1.7 reveals that there is more to the impressive decade-long
run of gold prices than the emergence of global trading of physical gold on stock market
venues using ETFs. Rather, a voracious demand from India and China for jewelry, bars
and coins has fuelled a large uptake in available supply. For cultural reasons, these
geographical regions have traditionally used gold as a store of value. Jewelry is typically
sold with high gold content, e.g. 22 carat, and gold dealers actively buy and sell gold
items at prices close to gold content. Unlike the base metals, where global industrial and
construction demand drive price movements, gold pricing is being driven by speculative
“investment” demand and culturally based South and East Asian demand for gold as a
store of value and sign of social status. Recent annual fluctuations in jewelry demand
from these regions is comparable to, if somewhat less than, fluctuations due to the
combination of ETF and bar and coin “investment.” Dis-hoarding by gold “investors”
associated with the eventual end to gold price appreciation contains the seeds of an
inevitable precipitous collapse in prices.
Against this demand backdrop, the supply situation is also complicated (see Figs
1.21 and 1.22, and Table 1.8). Mine output alone is incapable of meeting current levels
of demand, e.g. 3812 tonnes demanded versus 2572 tonnes produced in 2010. Mine
output is costly to increase. Marginal additions to mine output are achieved by bringing
higher-cost mines into production. This is occurring at the same time that the reserves
of lower-cost mines currently in operation are diminishing. To be sure, there are still
immense stocks of gold, primarily held by central banks and other governmental entities.
These gold stocks are sufficient to supply the market for more than a decade at current
levels of demand, with no mine output, and for three decades at current levels of mine
output with no recycling. Yet, traditional sellers of gold from reserve stocks, central
banks have recently become net buyers. Significantly, current levels of mine output are
approximately equal to the jewelry plus technology usage. The continuing additions
to private “speculative” stocks of gold through ETFs and bars and coins provides an
increasingly larger overhang of stocks that will threaten to come on to the market at a
time when prices are weakening.
Academic Studies
Most industry and government studies emphasize the real-time reporting of market
conditions and commodity characteristics. In addition, there are descriptions of risk
management practices and application of financial products and techniques that can
be used in the risk management activities of firms. In contrast, academic studies aim
to provide a “scientific” approach to commodity risk management, with mixed results.
As Mackay and Moeller (2010) observe: “What risks do firms hedge? How much do
they hedge? How far ahead do they hedge? What determines corporate hedging
policy? Should firms hedge at all? Can corporate risk management create value? As
straightforward and important as they might appear, these questions are still largely
unresolved.” At least since Smith and Stulz (1985), academic studies have been seeking
and proposing answers to such questions, e.g. Brown and Toft (2002). Yet, helpful
academic guidance is still lacking.
Theories of “optimal” commodity risk management still feature a wide range of
competing strategies. Carter et al. (2006b, p.21) summarize these notions:
Finance theorists have proposed a number of ways that hedging and, more
generally, risk management can increase corporate market values. Stated briefly,
risk management has the potential to add value by (1) reducing corporate income
taxes; (2) reducing the probability and expected costs of financial distress; and
(3) preserving management’s ability and incentives to carry out all positive-NPV
projects (incentives that can otherwise be distorted by the pressure for near-term
cash flow faced by financially troubled firms).
and Peterson (2001, p.955) about academic risk management studies in agriculture seem
well placed for other commodity sectors:
Even the optimistic review by Smithson and Simkins (2006, p.15) is able to conclude
only: “Is there any direct evidence that risk management increases firm value? The
answer is yes, but the evidence is fairly limited as yet.”
Seeking sources of homogeneous “factors” across firms, academic theories have been
incapable of explaining why firms in the same commodity industry are observed to
pursue hedging policies ranging from unhedged to conditionally hedged to operation-
ally hedged. Smithson and Simpkins (2006) hint at the strategic heterogeneity of risk
management decisions as an explanation: “is there some kind of ‘self-selection’ process
in which successful firms are more likely to have the capital and other resources needed
to run a derivatives program?” In various academic studies, there is only limited differ-
entiation between the underlying commodity markets involved in empirical work. The
search for empirical support for general theories supporting the use of commodity risk
management tools lumps gold mines from Tufano (1996) with airlines in Carter et al.
(2006a,b), with oil refineries in Mackay and Moeller (2007), and so on. As Bertus et al.
(2009, p.737) observe: “The airline industry provides motivation for our analysis but our
results may be applied to any firm with similar exposures.” This “generalist” approach
stands in contrast to more traditional, largely agriculturally inspired academic risk
management studies that examine risk management practices in specific commodity
markets without seeking to confirm the validity of general theories such as the Froot
et al. (1993) under-investment hypothesis.
A classical example of the traditional approach to managing commodity risk is
Hieronymous (1977), where detailed study of the characteristics of commodity produc-
ers, merchandisers and consumers is used to motivate risk management practices for
specific agricultural commodities, e.g. corn, wheat, cotton and soybeans. Other older
studies include Heifner (1972) for cattle feedlots, and Hayenga and DiPietre (1982) for
pork producers. Recent examples with elements of the traditional approach include:
Gemech and Struthers (2007) and Mohan (2007) for coffee producers; Wilson et al.
(2007) for bakeries; Harrington and Niehaus (2003) for United Grain Growers; van
Duren et al. (2003) for food processors; Dahlgran (2000) for cottonseed; Bielza et al.
(2007) for Spanish potato growers; Wilson (1987) for sunflowers; Bergfjord (2009)
for Norwegian aquaculture producers; Buguk and Brorsen (2005) for Turkish cotton;
Costa and Turner (2001) for peanut meal; Mohapatra et al. (2010) for strawberries; and
Menachof and Dicer (2001) for the liner shipping industry. The traditional approach
also includes studies where use of a specific risk management product is detailed or pro-
posed, e.g. Muller and Grandi (2000) and Chantarat et al. (2007) on the use of weather
derivatives; Hart et al. (2001) for livestock revenue insurance; and Turvey (2006) for
commodity-linked credit instruments.
“Commitments of Traders” report is one such publication by the CFTC. However, there
is again no specific information on risk management practices and prescriptions.
The final source of government reports and programs of relevance to commodity
risk management is international agencies and NGOs. Producers of occasional research
reports in the agricultural arena include: the Food and Agriculture Organization of
the UN; International Fund for Agricultural Development; UN Commission on Trade
and Development; the World Food Program of the UN; the World Bank; the WTO; the
International Food Policy Research Institute; and the UN High Level Task Force on
world food policy. Together with the IMF and OECD, these agencies together released
an important policy document (FAO 2011) dealing with risk management responses to
recent price volatility in food and agricultural markets that have severely impacted less
developed countries. The World Bank, the IMF and other entities occasionally publish
studies relevant to commodity risk management in regularly published periodicals such
as the World Bank Observer, World Bank Economic Review and IMF Staff Papers, e.g.
Faruqee et al. (1997); Larson et al. (2004). Though often concerned with agricultural
development in less developed countries, there are also occasional studies relating to
commodity risk management in general, e.g. Larson et al. (1998).
Industry Studies
In the digital era, every commodity sector has numerous websites that are invaluable
sources of information about “real time” commodity characteristics and, in a few cases,
also provide practical guides or policy positions concerning risk management practices
for the specific commodity at hand. An indication of the scope of available websites
applicable to farming in North America includes: American Farm Bureau website
(www.fb.org); American Feed Industry Association; American Soybean Association;
Agricultural Retailers Association; Animal Agriculture Alliance; Arkansas Rice
Farmers; Canadian Canola Growers Association; Grain Growers of Canada; National
Agricultural Marketing Association; National Corn Growers Association; United
Soybean Board; US Grains Council; and the US Meat Export Federation. Though a
useful starting point, this list is far from including all relevant sites. Being lobby and
information entities, few such agricultural sites provide substantive information about
risk management. When views on risk management are expressed, it is often to promote
a particular policy position. Fig. 1.24 is an example of a policy promotion for agricultural
revenue insurance by the Canadian Canola Growers Association.
Similar, though less numerous, websites are available for other commodities and
commodity complexes. For example, in copper the International Copper Study Group
Figure 1.24 Business Risk Management Webpage from Canadian Canola Growers
Association
55
There is, for example, the story which is told of Thales of Miletus. It is a story about
a scheme for making money, which is fathered on Thales owing to his reputation for
wisdom; but it involves a principle of general application. He was reproached for his
poverty which was supposed to show the usefulness of philosophy; but observing
from his knowledge of meteorology (so the story goes) that there was likely to be
a heavy crop of olives [next summer], and having a small sum at his command, he
paid down earnest-money, early in the year, for the hire of all the olive-presses in
Miletus and Chios; and he managed, in the absence of any higher offer, to secure
them at a low rate. When the season came, and there was a sudden and simultaneous
demand for a number of presses, he let out the stock he had collected at any rate he
chose to fix; and making a considerable fortune he succeeded in proving that it is
easy for philosophers to become rich if they so desire, though it is not the business
which they are really about.
A key point in the development of derivative security contracts is where market liquidity
was sufficient to permit the securitization of contingent claims associated with forward
delivery and the privileges of “put” and “refusal.” As early as Ehrenberg (1928), it was
recognized that this required the emergence of sufficient speculative trading to sustain
market liquidity.
Verbal contracts are made between party and party, or by means of Brokers or
Mediators, and that only by word without writing. Such are the daily buying and
selling of commodities either for ready money, or payable at some dates of payment,
wherein the mediation of a Broker is most necessary: For as it would be troublesome
to use Scrivners in every bargain; so is it commodious to use the means of Brokers,
the commodities are not only bought and sold with more credit and reputation,
but all controversies which do arise by misadventure or otherwise are sooner
determined, and a sworn Broker is taken as a double witness, if he do produce his
book, with a Memorandum of the bargain, as the same was agreed between both
parties, whereby many variances are reconciled, and differences (like to fall out)
are prevented.
This brief discussion on the use of brokers in commodities transactions follows a longer
discussion by Malynes (1622, pp.124–6) regarding the use of “notariall contracts” in
the trading business of the regulated company of “Merchant adventurers” where the
systemic use of forward contracts in their commercial transactions is apparent.4 In
modern vernacular, the 16th- and 17th-century sales contracts that Malynes describes
for the sale of English cloth goods arranged on the exchanges in Antwerp, Bruges and
other important centers were structured as non-transferable forward contracts with
multiple delivery dates and option features.
“Exchange traded” requires “exchange” to be defined. This is not as easy as might be
expected as a number of conventional and technical definitions for an “exchange” are
possible, e.g. Lee (1998, pp.322–3). Almost all modern definitions identify an exchange
with a physical location or building. One exception is Ehrenberg (1928, p.54), where the
following definition is provided:
For historical purposes, this definition is more adaptable as some exchanges, such as
the important Amsterdam exchange prior to 1611 or the London stock exchange prior
to 1773, conducted trading at different locations until moving to fixed quarters.5 This
definition is also sufficient to distinguish an exchange from, say, a marketplace selling
produce and is preferable to definitions that identify an exchange only as a physical
location where buyers and sellers meet to trade goods.6
Given this, exchange trading of a contract needs to be distinguished from a situation
where a buyer and seller meet at, say, the Royal Exchange and agree to a forward sale
of goods with a contract then drawn up by a “scrivner.” In this situation, it is the
goods that are being traded, not the forward contract. An exchange traded contract
requires transferability and an exchange clearing mechanism to settle positions. In
turn, transferability requires standardized contract terms and relatively homogeneous
deliverable commodity. A transaction where merchants meet at an exchange and agree
to a non-transferable forward contract with multiple delivery dates would correspond,
in modern terms, to an over-the-counter (OTC) derivative security transaction. As
such, a technical distinction is being made that corresponds to the modern difference
between trading on exchanges or OTC.
The beginning of exchange trading of derivative security contracts occurs when the
parties involved in the completion of the contract are different from those initiating the
contract. With this condition, a traded contract could be created for which there was
no intention of completing the underlying goods transaction; in effect, the seller may
not have possession of the goods and the buyer may not intend to take delivery. In this
case, a contract can be created for which there is no resulting delivery of goods. This
requires a clearing method for determining and settling gains and losses on contracts.
Various prerequisite conditions are required for such trading to occur. The evolution
was gradual, not dramatic, and depended on a range of informal restrictions on those
participating in the trade. Recognizing that initial trade was in the bulk commodities of
herring, whale oil and wheat—commodities that require special warehousing, grading
and handling facilities—initial trade was associated with dealers directly involved in the
bulk commodity trade willing to execute “to arrive” forward contracts for which there
were no associated goods transactions, seeking to offset the position prior to delivery or,
if necessary, cover the position in the spot market upon arrival of the fleet or delivery
of the harvest. Given the vagaries of market liquidity, in the event the contract could
not be transferred, both parties to the contract needed to be able to complete delivery.
Today banks discount the trade acceptances or the promissory notes of merchants
who are in need of credit. Such a procedure was ruled out as long as contracts
involving the payment of interest were unenforceable at law. It is true that usury
laws could be circumvented by various subterfuges. However, the easiest method
for securing short-term credit was for merchants to “take up” money by exchange
and not at interest. The result of this practice was that commercial credit was tied to
the exchange. This point, although obvious, is so fundamental that its importance
should be stressed . . . the credit system rested on the exchanges.
As such, methods for clearing bill of exchange transactions were fundamental to the
smooth operation of the international commercial and financial system.
Though the precise origin of the practice is unknown, “arbitration of exchange” first
developed during the Middle Ages. Around the time of the First Crusade (1095–99),
Genoa had emerged as a major sea power and important trading centre (Einzig 1964).
The Genoa fairs had become sufficiently important economic and financial events that
traders from around the Mediterranean were attracted. To deal with the problems of
reconciling transactions using different coinages and units of account, a forum for
arbitrating exchange rates was introduced. On the third day of each fair at Genoa, a
representative body composed of recognized merchant bankers would assemble and
determine the exchange rates that would prevail for that fair. The process involved
each banker suggesting an exchange rate and, after some discussion, a voting process
would determine the exchange rates that would apply for that fair. Similar practices
were adopted at other important fairs later in the Middle Ages. At Lyon, for example,
Florentine, Genoese and Lucca bankers would meet separately to determine rates,
with the average of these group rates becoming the official rate. These rates would
then apply to bill transactions and other business conducted at the fair. Rates typically
stayed constant between fairs in a particular location providing the opportunity for
arbitraging of exchange rates across fairs.
The actual clearing process differed from fair to fair (Parker 1974, p.546). At the
Lyons fairs, clearing involved the participation of all merchants attending the fair. At
other fairs, such as the fairs of Besançon or Medina del Campo, clearing was controlled
by a restricted group of merchant-bankers who were responsible for setting exchange
rates and for handling the book transfers between the accounts of merchants at the
various clearing member banks. Ehrenberg (1928, p.284) describes the clearing process
used in Lyons:10
Before the merchants attended the fair they entered in their “market book” . . . all
the payments due from or to them in the fair. At the beginning of the fair these
payments books were compared with one another. In the case of every entry found
correct the person from whom the payment was due made a mark which was taken
as a binding recognition of the debt; later he had to sign his whole name. The bill—
for, generally speaking, there was no question of anything but bills—was accepted
in this way. If an item was not recognized, the owner of the book would write by it
“S.P.” (sous protest).
After the acceptance of the old bills there followed the new business with
foreign markets, which originated wither at the preceding fair or as the result of
the acceptance, or otherwise. Here we meet for the first time a peculiar arrange-
ment, the settlement of an official average price for each species of bill, the so-
called Conto.
[T]he Conto in Lyons was done as follows: The bill dealers met on a certain day
and formed a circle (Faire la Ronde); the Consul of the Florentines then asked the
dealers of the different nations in turn what they thought the price ought to be. The
answers were noted and an average taken. This was the official rate for bills which
was noted in the bulletins . . . and sent abroad. The dealers themselves were naturally
not bound by this, their business was left free to bargaining. Yet the Conto at the
beginning had some meaning for the market itself, as previously many transactions
had been concluded at the average rate which had not yet been settled . . .
The payment proper closed the fair. It was affected chiefly by viremant de parties,
giro or scontro, as follows: Two persons were commissioned to collect and compare
. . . all the fair books. They then canceled the payments against one another, and
only paid the balances in cash . . . The fair payments at Lyons owe their form to the
Florentines, a fact which is clearly shown by the development of the Lyons Bourse.
Various features of the clearing process at Lyons were not only adapted for use at other
important fairs, but also had an impact on the methods later employed on the Lyons
bourse. The method of offset used in the end-of-fair settling process was later reflected
in the rescontre system adopted to settle exchange trading of joint stock shares in
17th-century Amsterdam and 18th-century England.
The Lyons fairs first assumed importance circa 1463 due to the explicit mercantilist
policies of Louis XI. As early as 1419, various French kings had granted privileges
to merchants doing business in Lyons in an attempt to counteract the success of the
fairs held in Geneva. These privileges included freedom to engage in various financial
transactions, such as manual exchange of coin and dealing in bills of exchange,
activities that were tightly regulated elsewhere in France. Even more than the economic
benefits associated with the commodities trade, the French monarchs were motivated
by the gains associated with the financial dealings of the fairs. By the 15th century, the
capital that could be raised at important fairs such as those of Geneva was substantial.
This capital was essential to securing financing for the military adventures in which
the national monarchs were, almost continually, engaged. The extension of commercial
liberties beyond the time period of the fairs contributed significantly to the emergence of
bourse trading. As early as the end of the 13th century, the dukes of Brabant encouraged
the growth of Antwerp by granting privileges to alien merchants visiting the city (van
Houtte 1966), such as not requiring that local brokers be used to transact commodity
business. Such merchants trading in Bruges, the northern centre of European commerce
during the 14th century, were required to use local brokers.
While the fairs served as an important step in the growth of trade and payments,
by the late 15th century economic activity was outgrowing the restrictions of the fixed
fair dates. A network of international merchants had established permanent offices and
warehouses throughout the key commercial centers of Europe. To support the associated
trading activities, sizable communities of foreign merchants were established. These
changes meant that liquidity was sufficient to support trading throughout the year. This
growth sustained the creation of bourses in various cities, designed to facilitate dealings
in both physical and financial commodities. The bourses were, effectively, meeting places
for merchants of various countries to transact financial and commodities business. The
use of the term “bourse” (beurs) is indicative of the historical development, the term
being taken from a square in Bruges, named for an inn on the square owned at one
time by the van Beurs family, where the Florentines, Genoese and Venetians had their
consular houses. This inn was a popular meeting place for foreign merchants. Though
exchange trading of derivative securities was yet to come, some essential characteristics
in the trading cities of Italy, [bourse trading] arose from the business which
developed at the banks of the money-changers native to the city, when the notaries
likewise had stalls in the open air . . . there arose . . . the characteristics of the
exchange business as early as the fourteenth century . . . In the countries north of
the Alps bill business . . . developed in closest connection with the factories of the
Italians. The streets and market places where they lived, and more especially where
they had their consular houses or Loggias, were the localities where the bourse
business first developed.
Bruges was geographically well situated to have the first significant bourse trading in
northern Europe. The opening of seaborne trade routes through the Straits of Gibraltar
contributed to the decline of fairs along the land trade routes, such as the Champagne
fairs. In addition, the Hansards developed important seaborne trade from northern
Europe. All this growth in seaborne commodity traffic contributed to the initial rise of
Bruges as “the greatest market of Christendom in the fourteenth century” (van Houtte
1966, p.37).
In addition to being a main seaport, Bruges was also the locale for one of the five
fairs of Flanders. The importance of Bruges peaked in the mid-1300s, the comparatively
faster subsequent growth of commercial markets and bourses in other centers being due
to two primary local factors: the silting of the waterway connecting Bruges to the ocean
and the various restrictions imposed by Bruges on foreign merchants trading there.
The international growth of trade meant that Portuguese, Spanish, South German and
Italian merchants had sufficient reason to establish permanent colonies in locales such
as Bruges and Antwerp whereas, before, these merchants sojourned to the fairs. In
addition to geographical factors, the freedoms granted to alien merchants played a key
role in determining where bourse trading was concentrated.11
From the beginnings of bourse trading there was competition between exchange
venues for business. A second factor favoring bourse trading was that fairs required
goods to be transported to the fair’s geographical location for inspection in order to
conclude specific transactions. The goods were then transported to another district to
be sold. As trade expanded, factors such as acceptable levels of standardization and the
growth of mutual merchant confidence allowed goods transactions to be made without
actual inspection of goods at the time the sale was completed. In turn, exchanges were
located geographically close to the center of the underlying bulk goods trade. Such
factors significantly reduced transactions, transport and other costs. By providing
enhanced liquidity and cheaper execution, bourse trading was an essential impetus
to the emergence of speculation in commodities which, ultimately, progressed into
exchange trading of derivative securities.
Though the transition from fairs to exchange trading was gradual, the 16th century
does provide a transition period: at the beginning of the century, the fairs still played
an important role in providing fixed dates and locations at which concentrations of
liquid capital were assembled; by the end of the century, general economic activity
was such that bourse and exchange trading predominated. During the century, the
emergence of exchange trading in Antwerp and Lyons was especially important,
though by the end of the century both these centers were in decline. Of these two
centers, Antwerp was initially most important for trade in commodities, and Lyons for
trade in bills of exchange. In 1531, Antwerp opened a new exchange building designed
exclusively for trading of commodities and bills of exchange. Tawney (1925, pp.62–5)
describes this international market of the 16th century:
My ventures are not in one bottom trusted, Nor to one place, nor is my whole
estate Upon the fortune of this present year; Therefore, my merchandise
makes me not sad.
Markowitz (1999, p.5) claims that: ‘Clearly, Shakespeare not only knew about
diversification but, at an intuitive level, understood covariance.’ However,
given the later developments in the Merchant, it is not at all clear that Antonio’s
understanding of covariance was as deep as Markowitz claims.”
owners and operators of ships as well. They were merchants with an interest in more
assured supplies of preserved fish . . . even individuals with no direct connection with
fishing can and did invest in the boats and their supplies” (Unger 1980, p.258).
That “to arrive” contracts came to be actively traded by speculators also directly
involved in trading the underlying physical commodity is not surprising. Because trans-
port by sea was a risky business and information about cargoes to arrive at a later date
could be sketchy, the quality and quantity of physical commodity available for delivery
could not be known prior to arrival of the fleet; a forward sale of such cargoes would
be inherently speculative. Such speculative trade facilitated the legitimate hedging
activity of other merchants. The concentration of speculative liquidity on the Antwerp
Exchange centered around the important merchants and large merchant houses that
controlled either financial activities or the goods trade (van der Wee 1977). The milieu
for such trading was closely tied to medieval traditions of gambling and insurance
where wagering on the safe return of ships, a rudimentary form of early insurance
(Lewin 2003), was often connected with the conclusion of commercial transactions. A
key step in the evolution of exchange traded contracts came when trading in “to arrive”
contracts involved standardized transactions in fictitious goods for a future delivery
that was settled by the payment of “differences.”12 Purchasers of such contracts would
speculate on the rise in prices before the due date. If such a rise occurred, the contract
could then be sold and the speculator pocketed the difference in price. This “difference
dealing” was also conducted by goods vendors, selling for future delivery betting that
prices would fall.
The development of difference dealing was accompanied by the emergence of
“premium contracts” where: “The buyer made a contract for future delivery at a fixed
price, but with the condition that he could reconsider after two or three months: he
could then withdraw from the contract provided that he paid a premium to the vendor
(stellegelt)” (van der Wee 1977). Little is known about the precise evolution of the
contracts used for speculative trading, but the premium contract appears well suited
to difference dealing by speculators. The “premium” form of contract for forward sale
became a staple of European trade into the 20th century, e.g. the contract for the German
prämiengeschäfte. Such contracts differ from the options traded in modern markets,
which have inherited characteristics associated with historical features of US option
trading. Following Emery (1896, p.53), the prämiengeschäfte “may be considered as an
ordinary contract for future delivery with special stipulation that, in consideration of a
cash payment, one of the parties has the right to withdraw from the contract within a
specified time.”13 As such, this option is a feature of a forward contract with a fee to be
paid at delivery if the option is exercised. Circa 1908 on the Paris and Berlin bourses,
the premium payment at maturity was fixed by convention and the “price” would be
determined by setting the exercise price relative to the initial stock or commodity price,
e.g. Courtadon (1982).
Characteristics of exchange trading of derivative securities contracts in Antwerp
formed the basis for later trading at other venues. Elements of that trade are of still of
contemporary relevance. While access to the Antwerp exchange was unrestricted, those
unconnected to the bulk commodity trade and seeking to speculate required a broker
to establish a position. Brokers could also be dealers in the commodity. The exchange
was a largely self-regulatory entity with broker-dealers clearing derivative security
trades with other broker-dealers. Rules of conduct for trading were largely governed
by merchant convention, e.g. Malynes (1622). Penalties for violations involved loss of
reputation and an ensuing inability to conduct business. The state provided official
recognition to certain “sworn brokers” and established a civil court system for settling
disputes. Physical infrastructure, and a sympathetic legal and taxation environment
were provided to promote the development of trade. Difference dealing was facilitated
by the use of premium contracts. Following traditions developed in the bill of exchange
market, the clearing of positions in difference dealing was done by brokers coordinating
with other brokers.
Figure 1.27 Quotes for Commodities and Shares (Actions) from Houghton, A Collection for
the Improvement of Husbandry and Trade, July 6, 1694
trading in modern derivative markets. By the middle of the 17th century, trading on
the Amsterdam bourse of derivative securities had progressed to where contracts with
regular expiration dates were traded (Gelderblom and Jonker 2005; Wilson 1941).14 By
the 18th century, in addition to commodities, the trade involved both Dutch joint stock
shares and “British funds.” This trading on the Amsterdam bourse is the first historical
instance of exchange trading in financial derivative securities. “With the appearance of
marketable British securities, and the application to them of a speculative technique that
was already well understood, the Amsterdam bourse became the scene of international
finance at its most abstract and most exciting—gambling in foreign securities” (Wilson
1941, p.79).
While information about derivative security trading in Antwerp is scattered and
sparse, detailed accounts of such trading in Amsterdam are available in Josef de la Vega
(1688) and Isaac de Pinto (1762). Both sources discuss trading of joint stocks; trading
in commodities is not directly examined, though following traditions developed in
Antwerp, the commodities trade was also a common source of speculative trading. As
such, exchange trading in Amsterdam marks the beginning of the distinction between
derivative securities for bulk commodities versus financial assets, in particular shares
in joint stock companies. While trading of derivative securities in the bulk commodity
trade was controlled by a network of brokers and dealers directly connected to the
underlying goods trade that made speculation difficult for “outsiders,” the same was
not the case with shares. Amsterdam is the first instance in the history of exchange
traded derivative securities where the distinction between financial assets and bulk
commodities as deliverables assumes importance. Though there was speculative trade in
bulk commodities—grain, herring, spices and whale oil—the trade in shares apparently
captured the bulk of this activity by the later 17th century.
Circa 1602, the Amsterdam Exchange was held in the open air on the New Bridge.
It was not until 1613 that trading completely moved to a building dedicated to the
Amsterdam exchange. Trading in shares was initially only a small portion of the
general activity on the exchange, which was predominantly in bills and commodities.
By the beginning of the 17th century, it was apparent that trading in Amsterdam
had become the successor to the Antwerp bourse that had fallen on hard times due
to a combination of political, geographic and economic factors. In conjunction with
the shift in trading activity, many of the traders also eventually relocated from
Antwerp to Amsterdam, and brought with them the trading techniques that had
been successfully developed on the Antwerp exchange. Included among these
techniques was speculative trading for future delivery. This technique, almost immedi-
ately, was applied to trading in Company shares. Ehrenberg (1928, pp.358–9) provides
some fundamental insight into the advantages of speculative trading in shares over
commodities:
From the beginning, the speculation in shares . . . as a means of gain depending
on taking advantage of future price changes, made it appear extremely desirable to
postpone the fulfilment of the bargains. In the case of bears, who had sold shares
which they did not possess, this was an absolute necessity.
Speculative future dealings made possible a twofold simplification of the
technique of dealing. First, speculative dealings could be realized before the date of
delivery. Secondly, settling days made it possible to use the same procedure that had
done so much in the methods of payment, namely, set off. Both together resulted in
an incalculable increase in turnover, since now only a little ready money and stock
were required for very large dealings.
Significantly, “it was speculation which made the first modern stock exchange.”
Speculators provided the liquidity essential for continuous trading and “accurate”
pricing. In turn, hedgers and traders seeking to acquire or dispose of stock positions
provided the “honest” liquidity needed to clear the market. De la Vega (p.164) suggests
that the relative composition of the speculative trading population changed over time,
whereas “formerly twenty speculators ruled the exchange . . . Today there are as many
speculators as merchants.”
Kellenbenz (1957, pp.139–42) provides a useful summary of de la Vega’s discussion
of the various types of transactions in the Amsterdam share market, which likely also
corresponded with practices in the exchange of commodities:
Trading for forward delivery was essential to the 17th-century trade in commodities
and shares on the Amsterdam bourse (Barbour 1950).15 For shares, such trading was
necessary because the delivery and settlement process was much different than the
modern process. Though shares could be transferred, the process required the seller to
be present at the company offices for the transfer and to pay a transfer fee. The practice
of same day settlement, delivery and transfer, as practiced in modern stock markets,
was not usually possible—even for trades arranged at the transfer office. Agreements
to sell shares typically included a future settlement and transfer date, which could be
months in the future, though delivery dates longer than one month in the future were
discouraged by statutes dating from 1610.
Perceived speculative abuses of the delivery process appeared almost from the start
of trade in Dutch East India Company (VOC) shares in 1602 (van Dillen et al. 2007).
This included the activities of a bear ring, formed “in early 1609 . . . to challenge the
company on the exchange. It is not clear that the ring did more than help to hold
down the already slumping prices, but the company lodged a protest with the States of
Holland and West Friesland in the summer of 1609 to have a ban placed on the sale
of shares ‘in blanco’” (De Marchi and Harrison 1994, p.51). The result was the Dutch
edict of 1610 banning short sales “in blanco,” where, at the time of the short sale, the
seller does not actually possess the shares or the commodity being sold. In addition,
the edict required that share transfers be made within one month of the sale date.
Transactions dated to correspond to a future goods delivery were conventional in
the legitimate commodities trade. The ban on short sales was not permanent and the
“occasion of renewal brought out anew sentiment for and against VOC” (p.51). Despite
opposition, the ban on “selling in the wind,” or windhandel trade, was repeated in 1624,
1630, 1636 and 1677. It is important to recognize that the de facto impact of the ban on
in blanco short selling was to make such contracts unenforceable in the courts. There
was no direct criminal penalty for entering into such contracts which provided the basis
for difference dealing.
For an event that has received such substantial attention in modern times, the Dutch
tulipmania of 1634–37 has been surprisingly misrepresented. Malkiel (1985, pp.29–32),
for example, makes the following comments:
The instruments that enabled tulip speculators to get the most for their money were
“call options” similar to those popular today in the stock market. A call option con-
ferred on the holder the right to buy tulip bulbs (call for their delivery) at a fixed
price (usually approximating the current market price) during a specified period.
He was charged an amount called the option premium, which might run from 15
to 20 percent of the current market price. An option on a tulip bulb currently worth
100 guilders, for example, would cost the buyer only about 20 guilders. If the price
moved up to 200 guilders, the option holder would exercise the right; he would buy
at 100 and simultaneously sell at the then current price of 200. He then had a profit of
80 guilders (the 100 guilders’ appreciation less the 20 guilders he paid for the option).
As happens in all speculative crazes, eventually prices had been going higher for a long
enough time that some people decided they would be prudent and sell their bulbs.
And what of those who had sold out early in the game? In the end, they too were
engulfed by the tulip craze. For the final chapter of this bizarre story is that the
shock generated by the boom and collapse led to a prolonged depression in Holland.
No one was spared.
Malkiel also relates an anecdote about a sailor unknowingly eating an expensive tulip
bulb thinking it was an onion.17 Though more detailed, elements of Malkiel’s discussion
can be found in numerous modern references to the tulipmania.
The reference to call option trading during the tulipmania, which appears in
numerous modern sources, is difficult to support. There is considerable evidence that
forward contracts—varying in form from the rudimentary to notarized contracts with
embedded option features—were the method used in trading for future delivery during
the tulipmania. Malkiel (1985, p.352) claims: “My discussions of the tulip-bulb craze . . .
rely heavily on Mackay’s description.” However, Mackay makes numerous references to
“bargains,” which was a conventional reference to forward contracting. Mackay (1852,
p.95) describes a typical trade:
Confidence was destroyed, and a universal panic seized upon dealers. A had agreed
to purchase ten Semper Augustines from B, at four thousand florins each, at six
weeks after signing the contract. B was ready with the flowers at the appointed time;
but the price had fallen to three or four hundred florins, and A refused either to pay
the difference or receive the tulips.
Malkiel (1985, p.31) also seems confused on the point, making the statement that: “Dealers
went bankrupt and refused to honour their commitments to buy tulip bulbs.” If Malkiel
is correct about option trading, this problem would only be a problem if put options
were being traded, not call options.
What actually did happen during the tulipmania? Garber (1989, 1990) and Posthumus
(1929) are academic sources that detail the events and market activities. Prior to these
studies, information about the tulipmania could be derived from various sources. Due
to the attention given by modern sources such as Malkiel (1985), Mackay (1841, 1852)
has received considerable credit for chronicling the event. Similar treatments of the
tulipmania are reflected in other sources from this period, such as Francis (1850). Garber
correctly observes that much of Mackay’s relatively brief discussion is plagiarized from
Beckmann (1846). The most essential primary source for Beckmann was the Gaergoedt
and Waermondt (GW) dialogues, which are a series of three pamphlets (1637), written
in dialogue form by a now anonymous author.18 An English translation of key parts of
the GW dialogues is contained in Posthumus (1929).
The tulip was first imported into Europe from Turkey. Early reports have tulips in
eastern Europe during the 1550s. By the later part of the 16th century, the tulip had
appeared in the northern Netherlands. The tulip trade expanded quite rapidly, being
centered around Haarlem where, even today, the tulip fields extend north and south for
40 miles or more. Though it is possible to propagate tulips from seed to flowering bulb
over a seven- to twelve-year cycle, the primary method of propagation is from bulbs.
During a growing season, which goes from September to June, the bulb that was planted
will be propagated into a new bulb, a clone of the first. If all goes well, the new primary
bulb will also have some additional buds, outgrowths referred to as excrescences. By
this process, it was possible to increase the tulip stock of normal bulbs “at a maximum
annual rate of from 100 to 150 percent.”
Trade in tulips is done with bulbs. In certain cases, excrescences can also be traded
but this is riskier. The outgrowth has to be separated from the motherbulb and,
depending on size, can take from one to three years to flower. An additional risk with
excrescences is that growing into a flowering bulb is not certain. Two general categories
of bulbs can be distinguished based on an important quality difference between various
bulbs. “Pound goods” are run-of-the-mill bulbs that were sold by weight (pounds or
thousand azen), by the bed, or by the garden.19 “Piece goods” are the rarer varieties
of tulips, which are sold by the bulb. Heavier bulbs would have more outgrowths and
would, as a consequence, be more expensive. Because the propagation process produces
clones, a rare bulb would eventually become common as more bulbs were produced
from the original bulb.
The process of creating rare bulbs created an additional source of uncertainty. The
rare bulbs originate from “breaking,” the invasion of the bulb by a virus that produces
unique coloring patterns on the flowers. Though it is now recognized that the virus is
spread by aphids, this was not known in the 17th century and there was considerable
mystery about the breaking process. What was known is that breaking could not be
replicated with seed propagation; only bulbs retained the unique color pattern. Because
breaking is due to a disease, “broken” bulbs had generally lower propagation rates and,
possibly, could fail to survive entirely. Because heavier bulbs were more likely to have
a larger number of excrescences, a heavy bulb with a unique and valued color pattern
would be a very unusual commodity. It was these bulbs that commanded seemingly
outrageous prices.
Bulbs can safely be removed from beds in June, but had to be replanted by September.
Conventional practice in the cash market for tulips was to trade physical bulbs during
the summer. In addition to cash market trading, forward trading was also common:
It often happened that the price was not fixed in money; the most heterogeneous lot
of goods was accepted in payment, such as cows, fruit, wine, yards of cloth, clothes,
silver dishes, horses and carriages, land, houses, shops, and paintings. The usual
condition was for these various goods to be delivered at once, often long before the
bulb had been taken out of the ground.
(Posthumus 1929, p.439)
In effect, the tulip trade was conducted using forward contracting methods, which
were common practice in agricultural areas, albeit adapted to the special features of the
tulip. However, a new type of “bulb trading’ appeared during the tulipmania which was,
decidedly, unconventional.
The tulipmania was precipitated by the entrance, around the end of 1634, of purely
speculative buyers into the tulip market:
People who had no connection with bulb growing began to buy after [early 1634].
Among these were weavers, spinners, cobblers, bakers, and other small tradespeo-
ple, who had no knowledge whatsoever of the subject. About the end of 1634 . . .
the trade in tulips began to be general, and in the following months the non-pro-
fessional element increased rapidly. Rumours about rising prices paid for tulips in
Paris and the North of France accelerated the movement. New ways of selling were
organized . . . Towards the boom in 1636 . . . buyers of bulbs often know that the
seller possessed none; so they did not pay or deliver their goods till they were certain
the tulip would really come into their possession. At the height of business most
transactions took place without any basis in goods. The trade in [forward positions]
had degenerated into purest gamble, the seller selling bulbs he did not have against
a counter value, mostly money at this period, which the buyer did not possess. Each
succeeding buyer tried to sell his ware for higher prices; and in the general excite-
ment, one could make a profit—at least on paper—of several thousand florins in a
few days. The craze spread rapidly with these high profits. All classes of population
ended by taking part in it—intellectuals, the middle classes, and the labourers.
(Posthumus 1929, pp.438–40)
GW traces the collapse to February 3, 1637. By the end of February 1637, there was
widespread default on forward contracts. After a short period of political and legal
wrangling, the bulk of contracts outstanding at the time of the collapse were voided on
the basis of “appeals to Frederick,” the common reference to the ban on in blanco short
selling. Where payments of differences were made, these payments were almost always
in the 1–5 percent range.
What did the price of tulips do during the tulipmania? Drawing primarily on GW,
Garber (1989, 1990) provides detailed information on certain extreme price movements
during the speculative updraft in prices. Yet, the bulbs examined by Garber are selective.
GW report the following prices for the period from 1635 to early 1637:
To mention a few out of so many, just as you know the lion by its claw. A plant
Gheele en Root van Leyden of 515 aces had been sold in the first instance for
46 gld., and then for 515 gld.; a Gouda of 4 aces first for 20 gld., later for 225 gld.;
and Admirael de Man of 130 aces first for 15 gld., then for 175 gld.; a Generalissimo
of 10 aces first for 95 gld., and then for 900 gld.; and so on with the other plants.
This only lasted for a month or six weeks; then they started selling by the thousand
ace and by the pound. A pound yellow Croonen could be bought first for 20 or 24
gld.; in a month’s time it was 1,200 gld. and over. A pound of Switzers first cost 60
gld., later 1,800 gld. A pound of White Croonen first cost 125 gld., later 3600 gld . . .
GW reports similar price behavior for various other bulbs. Garber (1989) has evidence
from a small sample of bulbs indicating that, as is common in speculative frenzies, there
was a steep increase in prices in the last couple of months prior to the collapse. In any
event, the price increases reflect the temporary social obsession that speculating in tulip
bulbs had in Holland.
The claim about widespread options trading during the tulipmania is puzzling,
especially as there is a fairly detailed record of the types of contract used. The tulip trade
during the mania period was conducted using a number of different methods, from the
“promises and vouchers” of the most speculative and uninformed traders, to the formal
notarized written contracts of tulip dealers. GW provides numerous examples of the
text of contracts. Some are quite basic, such as: “sold to N.N. a quarter of Witte Kroonen
for the sum of 525 gld. when the delivery takes place; and four cows at once, which may
be now taken from the stable and led to the seller’s house.” A more detailed example for
the forward sale with option features of a piece good is:
will remain with two praiseworthy people, who know these things and who live in
the place or town, where this transaction has taken place. And by default of payment
of the aforesaid sum, I hereby engage all my goods, movable and immovable,
submitting same in the power of all rights and magistrates; all this without arch or
cunning. Have signed this. Act in Haarlem on December 12th, 1636.
Perhaps some speculative fringe players in the tulipmania engaged in pure gambles,
which were configured as an options transaction. However, such deals, if any were
ever done, were only obscure incidents in the tulipmania.20 Evidence for such
dealings is not available in important primary sources, such as GW, or in key secondary
sources.
One interesting feature of the speculative mania was the trade that developed
among groups of speculators called “colleges” meeting in numerous taverns where
speculating was combined with eating and drinking. As the mania gained steam
an increasing number of uninformed speculators were gathered to the trade, the
cumbersome trading method of using notarized contracts became problematic.
The colleges developed a method of trading with few rules, though what rules there
were appear to be much the same from college to college. Garber (1989, p.557) observes
“the college [forward] markets suffered from a lack of internal control over the nature
of contracts . . . These markets consisted of a collection of people without net worth
making ever-increasing numbers of ‘million-dollar’ bets with each other with some
knowledge that the state would not enforce the contracts.”
What lessons can be drawn from the tulipmania? Aside from the obvious observa-
tions about the social and economic consequences of mania, there are the fundamental
insights about the relationship between forward contracting and the underlying
commodity being traded.21 The mania was largely driven by the excesses induced
by forward trading by uninformed speculators. Significantly, because the forward
contracts were traded on bulbs that were in the ground, the underlying commodity
had elements of non-storability. Insofar as there was an insufficient supply of unplanted
bulbs available for purchase during the period from October until June, there was no
possibility of doing cash-and-carry arbitrage either for piece goods or pound goods. This
permitted the forward price to be determined, almost exclusively, by the uninformed
speculators unconnected to the tulip trade who dominated the tulip market between
1635 and 1637. The upshot of this participation by uninformed speculators was severe
dislocation of the price discovery process for tulip bulbs.
While academically interesting, this claim speaks more to the difficulty of theoretical
modeling than to whether there was a tulipmania. The inability of theoretical models
to verify whether a mania happened or not does not mean that a mania did not occur.
Precisely what type of evidence Garber requires to verify the occurrence a mania is
unclear. Since David Hume (1711–76), philosophers have recognized that such skeptical
empiricism is a slippery epistemological slope: “A mania by any other name is still a
mania.”
Garber’s other claim is that there is insufficient evidence to support the hypothesis
that there was a tulipmania:
While lack of data precludes a solid conclusion, the results of the study indicate
that the bulb speculation was not obvious madness, at least for most of the 1634–7
“mania.” Only the last month of the speculation for common bulbs remains as a
potential bubble, although the nature of the market, the contractual commitments,
and the surrounding events are unclear enough that one could seriously embrace
one side of the fundamentals versus bubble dispute only on the basis of strong prior
beliefs.
Garber bases this claim on an apparently detailed analysis of the empirical evidence.
After a useful review of previous studies on the tulipmania, the institutional structure
of the tulip market is examined and the price performance of various types of tulip over
long time periods presented. From an examination of the long time period price data,
Garber concludes that: “the magnitude of prices for valuable bulbs and their patterns of
decline are not out of line with later prices for new varieties of rare bulbs.” Garber also
indicates “the absence of descriptions of economic distress in accounts of the period not
engaged in antispeculative moralizing.”
Is Garber correct about the tulipmania? Has a 350-year-old myth been exposed? The
crux of Garber’s empirical argument is that observed prices for rare tulip bulbs, so-
called “piece goods,” were consistent with typical market pricing for this type of bulb.
This conclusion is based largely on a comparison of the rate of price depreciation of
selected piece goods prices over three periods: from the peak of the mania in February
1637 to 1642, and for 1707–22 and 1722–39, neither of the 18th-century periods being
associated with a tulip speculation or crash. Observing that the average piece goods
depreciation rate of 32 percent for the 17th-century period was comparable to a 28.5
percent average for the two 18th-century periods, Garber concludes: “the crash of
February 1637 for rare bulbs was not of extraordinary magnitude and did not greatly
affect the normal time series pattern of rare bulb prices.”
There is considerable ground to cover in order to debunk Garber’s somewhat
incongruent dual hypotheses: that the tulipmania was not a real mania; and that it is
impossible, based on an examination of empirical evidence, to sustain any conclusion
about speculative manias, in general, and the tulipmania, in particular. Evaluating
whether the tulipmania qualifies to be called a “mania” or, to use a modern expression, a
speculative bubble, is complicated by the limited amount of data available. That there is
insufficient evidence about an event that happened over 350 years ago is not surprising.
As Garber recognizes, both piece goods and pound goods prices suffer from a number
of practical limitations. For example: “With the end of large-scale bulb trading after
February 1637, records of transactions prices virtually disappeared.” The 1642 prices
that Garber uses were obtained from the records of a single sample of purchases later
revealed at a 1643 estate auction.
Despite this paucity of data, Garber chooses to ignore empirical evidence that would
seem to support the possibility of a mania. Like a good prosecutor, Garber highlights
those facts that support a conviction, leaving facts that favor the defense for presentation
by the defense attorney. In particular, while the information about depreciation rates is
interesting, it isn’t the main issue with the inexplicably rapid increase and dramatic
collapse in prices for a wide range of bulbs. GW provides numerous instances of bulb
price increases from 20 gld. to 225 gld., or from 95 gld. to 900 gld., values that could
be justified in terms of Garber’s depreciation analysis, leaving reasons for the tenfold
increase and subsequent retrenchment of prices still unexplained. GW indicates that
these prices are for actual bulbs, not from trading in the colleges.
(CBT) in mid-19th century Chicago, a city that was first incorporated as a village in
1833, growing into a city of 4,107 by 1837. In order to promote commerce, the Board of
Trade of the City of Chicago was founded on April 3, 1848, with 82 members. While
this event, in itself, was not particularly noteworthy, the usefulness of boards of trade in
promotion had been recognized for quite some time. For example, around 1700, John
Law of the infamous Mississippi scheme promoted the creation of a board of trade for
the city of Edinburgh (Murphy 1997).
Inherent conflicts between the exchange as a promoter of trade and interests of
members and the exchange as a self-regulatory entity were not recognized, e.g. Lurie
(1972, p.221):
The CBT initially served as a marketplace for members of the grain trade. A system
of wheat standards was developed together with a system of inspecting and weighing
grain. In 1859, the Board of Trade was authorized by Illinois state to engage in
the measuring, weighing and inspecting of grain, effectively corn and wheat. As
Hieronymous (1977, p.73) observes: “The development of quality standards and an
inspection process and the substitution of weighing for the measurement of grain greatly
facilitated trade. The substitution of weight for volume measures made the development
of grain handling machinery possible. Increase in physical efficiency was important in
the development of Chicago as a great grain terminal.” These developments facilitated
the handling of grain in bulk, through the use of grain elevators. This permitted
interchangeable warehouse receipts to be introduced, instead of having to deal in
unstandardized, specific lots. Not unlike the bulk commodity trade in 16th-century
Antwerp and 17th-century Amsterdam, conditions in the goods market were evolving
to where standardized contracts on physical commodities could be traded.
The grain trade of that time typically involved merchants at various points along
major waterways such as the Illinois–Michigan canal purchasing grain from farmers,
which was then held in storage, often from fall or winter into spring. In this operation,
the merchants’ capital investment involved: paying the farmers for their crops at
delivery; costs of building and maintaining storage facilities; and providing funds for
shipment of grain when required. In order to avoid the risk of price fluctuation and
to satisfy bankers, merchants started to go to Chicago and make contracts for future,
spring, delivery of grain, at prices that were determined that day. While there was ad hoc
OTC-style forward trading of grains previously, the first “time contract” in Chicago was
made on March 13, 1851, calling for delivery of 3,000 bushels of corn in June at one cent
below the March 13 cash price. The time contracts called for delivery of a standardized
grade at a later delivery date. Similar contracts for wheat appeared in 1852. However,
while there were similarities to exchange traded futures contracts—the absence of other
conditions such as a clearing mechanism—the contracts were specific to the original
parties to the transaction and created with the objective of delivery. As such, the initial
trading in time contracts did not quite qualify as exchange trading.
The development of futures markets in Chicago was significant because, in the years
immediately following the introduction of time contracts, individuals not connected
to the grain trade became interested in taking positions. The resulting contracts often
changed hands numerous times before being purchased by a market participant actually
interested in taking delivery of the grain. This marks the introduction of a fundamental
feature of futures markets: the essential participation of speculators not concerned
with completing the underlying commodity transaction. Exchange trading and purely
speculative participants were characteristics not associated with trading in the often
non-transferable “to arrive” contracts and “privileges” that had characterized American
commodities trading previously (Williams 1982). This trade was concentrated primarily
in flour. To arrive contracts in wheat, corn, rye and pickled hams were also conducted
with activity centering on New York. In contrast to time bargains, to arrive contracts
typically featured short delivery dates, limited standardization of the deliverable
commodity and the expectation that delivery would be completed. While there is some
evidence of limited speculative dealings in these “to arrive” contracts and “privileges”
associated with the flour default of May 1847, participants to these transactions usually
involved merchants directly involved in the commodity business. In keeping with the
use of such contracts in Liverpool, cotton trading did employ “to arrive” contracts that
had elements of futures contracts (Forrester 1931).
The increasing interest in time contracts led the Board of Trade to introduce a
number of resolutions to curb abuses. Many of the abuses were consistent with spe
culative participation and longer delivery dates. “It seems that when time for settlement
arrived some of the contracting parties were difficult to locate” (Hieronymous p.76).
Out of the early self-regulatory process came the beginnings of formal trading rules for
futures contracts. In 1863, the Board adopted a rule that suspended the membership
of anyone failing to comply with a contract, either written or verbal. On October 13,
1865, the General Rules of the Board of Trade explicitly acknowledged futures trading
and adopted rules that included all the essential elements of a modern futures contract
including: standardized contract terms; restriction of futures contract trading to
exchange members; margin deposits to guarantee performance; and standardized
delivery procedures. Prior to this date, individual traders had been responsible for
establishment and enforcement of the terms of the contract. This development followed
a similar move in 1864 by the Liverpool Cotton Brokers’ Association introducing formal
regulations for “to arrive” contracts in cotton (Forrester 1931).
The seemingly orthodox futures contract, occasionally used before the Civil War
and an outgrowth of earlier “to arrive,” and “forward delivery” agreements, began
to receive unprecedented attention from speculators. Persons not previously con-
nected with the commodities business had been attracted, and were buying and
selling futures contracts in the central markets, especially in Chicago and New
York. The number of bushels and bales traded on the exchanges exceeded the
annual production from 1872 on and in several years toward the end of the century
amounted to sevenfold the annual crop. Prices had moved widely before the war
because of weather, economic instability, and imperfect crop information, but it
appeared that the new volatility was due to maneuvers by speculators with large
purses. Thus “speculator” became more than ever a term of opprobrium; the physio-
cratic bias against those who produced no primary products was more bitterly
asserted as the agrarian population shifted consciously to the defensive. The myste-
rious and remote commodity speculator seemed more of a parasite to the farmers
than the local physician who was holding land for appreciation. Farmers identified
the commodity speculator as the villain responsible for erratic price changes in
Chicago, Minneapolis, and New York, especially around harvest time. The stage
was set; the national crusade against the exchange speculator was about to begin.
Various efforts were made by the exchanges, as well as state and federal legislatures,
to control the perceived market manipulations. At the federal level, between 1880 and
1920, there were some 200 bills introduced in the US Congress aimed at regulating
derivative security trading, though few bills made it out of committee (Markham 1987,
pp.6–9). State legislatures that did pass bills, e.g. an 1874 Illinois statute prohibiting
“corners,” were unsuccessful in curbing such activities. Various states passed laws
prohibiting futures trading for which there was no intent to take delivery. Such laws were
voided by the courts on grounds that “pro forma assertions of interest with respect to
delivery were sufficient to preclude application of the statutes” (Markham 1987, p.6; see
also Lurie 1972).
The decline in agrarian conditions following the post-1886 droughts generated
sufficient political will to produce the Hatch-Washburn bill of 1892, the most concerted
effort at legislation prior to passage of the Grain Futures Act in 1921. Instead of outlawing
futures trading, this bill aimed to impose a prohibitive tax on speculative dealings in
futures. The Congressional debate on the issue surrounding the Hatch-Washburn bill
is an essential primary source on 19th-century views on derivative securities. The
committee meetings leading up to votes on the bill included testimony from important
agrarians, such as J.H. Brigham, Master of the National Grange and C.W. Macune of
the Farmers’ Alliance and Industrial Union. Not only farmers were in favor of the bill,
the testimony also included statements from millers, such as Charles Pillsbury, as well
as grain and hog merchants. Pillsbury held that “neither grower nor miller had as much
influence over prices as a few men around the wheat pit in Chicago. Short selling by
these few made prices erratic and unstable; opinions based upon supply and demand
were worthless in the face of this manipulation” (Cowing 1965, p.7). Pillsbury also
maintained that the use of futures to hedge would not be necessary if price volatility
due to speculation were eliminated.
In 1893, the Hatch-Washburn bill successfully passed the House, 167 to 40, and
passed the Senate, 40 to 29, though there were some amendments that had to be
returned to the House for approval.23 However, this placed the bill too far down the
calendar to be dealt with before the end of the session. A suspension of House rules was
required for the bill to become law. However, suspension of rules requires a two-thirds
majority and the vote, 172 to 124, fell short by 26 votes. The gradual return of prosperity
dampened, but did not eliminate, the drive of the anti-speculator forces. However, it
was not until after World War I that sufficient legislation, such as the Grain Futures
Act, was in place to curb the alleged abuses of the middlemen and speculators using
the exchanges. By this time, the extreme anti-speculator position of the agrarians had
faded. Though the Act did contain provisions against manipulation these were largely
ineffective. The Act was successful in bringing the futures exchanges under federal
consequence, the trade was generally conducted by a specialized group of OTC traders
catering to a relatively small clientele. Circa the end of the 19th century, trading in
privileges was conducted only in the after market and on “the curb,” as such trading was
prohibited on all US commodity exchanges. By the end of the 19th century, all US produce
exchanges had banned option trading, though some OTC trade did take place in other
venues and under various guises. Evidence for such trade in stock options is provided
by Kairys and Valerio (1997, p.1709) where an 1873–75 sample of over-the-counter
US option contacts is examined. This sample was obtained from advertisements in
the Commercial and Financial Chronicle. The prices were only ask quotes, exclusive of
bids, and were aimed at generating business from buyers of options. The option prices
were found to favor the option writer. Following the European practice, these contracts
determined prices by keeping the premium constant and adjusting the exercise price:
Whereas current option prices are quoted after fixing the strike price, the cost of
a privilege was fixed at $1.00 per share for all contracts and the strike price was
adjusted to reflect current market conditions. Furthermore, the strike price was
expressed as a spread from the current spot price of the underlying stock with the
understanding that the spread was then the “price” that was quoted for the privilege
contract.
Based on Emery (1896), this method of pricing options was also customary in the
Chicago grain markets where contract maturities varied from one day to a week. This
indicates the prevalence of European practices in the US option market at this time.
hedging and speculation. Of particular interest, as part of the process surrounding the
revisions to the Commodity Exchange Act (1974), the CFTC was required to provide a
definition of hedging in order to determine which traders would be subject to position
limits on speculative trade, and which traders would be considered as hedgers and not
subjected to limits on trading positions. This definition for a hedger was released in 1977.
Instead of attempting a precise legal definition, the CFTC opted for a long and involved
definition, derived from the economic motivations for hedging.
Though there is considerable scope in the CFTC definition for consideration on a
case-by-case basis, the CFTC definition generally requires (Leuthold et al. 1989, p.71):
(a) it must be economically appropriate to reduce risks; (b) risk must arise from
operating the commercial enterprise; (c) the futures positions normally represent
a substitute for transactions to be made later in the physical market; and (d) price
fluctuations in futures markets must closely relate to fluctuations in the cash market
value of assets, liabilities or services hedged. Thus, hedging is more than just
enumeration of specific transactions and positions—it is a process of risk reduction.
Prior to this definition, several legitimate hedging operations, especially cross and
anticipatory hedges, were not recognized as hedges because futures positions did not
meet the approximate equal and opposite requirement to the cash position.
While this explicit attempt to incorporate economic motivations into the definition
of hedging is definitely an improvement over a strict legal approach, the underlying
issues may be unresolvable. The Cargill corn case (Falloon 1998, Ch.8) provides a
classic instance of the difficulties, and associated implications, of distinguishing
hedging from speculation. The case originated from actions of Cargill, Inc., a grain
marketing company, primarily in the trading of corn futures contracts on the CBT
during 1936 and 1937. Cargill was a major player in the grain industry, at that time
handling approximately 12 percent of American grain being marketed (Broehl 1992,
p.467), arguably the largest single private grain distributor. Also at the time, Cargill was
the largest user of grain futures contracts on the CBT. Despite being a major player in
the markets, Cargill was something of an outsider in the CBT hierarchy, having only
grudgingly been admitted as a clearinghouse member in 1935. Cargill did not have a
seat on any of the primary CBT governing committees. As a result, it reacted negatively
to a series of adverse decisions from the CBT that were implemented to prevent what
was perceived as blatant market manipulation by Cargill. The resulting court cases that
originated were the first major test of the market manipulation provisions of the CEA.
As a major player in the corn market, Cargill was able to forecast tight conditions in
US corn supplies in 1936 and again in 1937. In response, during July 1936 Cargill bought
corn offshore, in Argentina, for import and placed a large long position in September
1936 corn futures on the CBT. Such activities are consistent with the role of Cargill as a
grain distributor seeking to match orders with purchases. What was questioned by both
the CBT and the regulators was whether the size of the position was fully consistent
with hedging activity. Did the position contain a significant speculative component? The
size of Cargill’s corn futures position constituted about one-quarter of open interest
and can be compared with 22.5 million bushels, which was the four-year historical
average for the “visible supply of corn in the US” at that time (Falloon 1998, p.190).
As the end of the delivery month approached, the size of this position attracted the
attention of the business conduct committee of the CBT, which held meetings on the
matter involving, on September 25, the president of Cargill. Despite assurances from
Cargill that the position would be unwound in an orderly fashion and no evidence
to the contrary, the CBT board of directors took action on September 29 allowing
an extension of the deadline for notice of physical delivery. This action precipitated a
substantial price drop in the cash corn price, adversely impacting Cargill.
Not surprisingly, the Cargill management was incensed. The atmosphere was again
poisoned on December 7, 1936, when the business conduct committee took action
regarding Cargill’s position in the December 1936 corn contract, which was deemed
to be too large to be justified by legitimate hedging activity. The committee took the
unprecedented action of ordering Cargill to reduce positions. Even though Cargill
complied, at an estimated cost of 15¢ per bushel per contract, positions were only
further hardened, setting the stage for the events of September, 1937. Again confronted
with supply shortages, during the summer of 1937 Cargill placed large long positions
in CBT corn futures, first in the July contract and then for the September contract.
The September position was about double that of the previous year, being as large as
9.4 million bushels, about one half of the contract open interest. Based on past experience,
Cargill surmised that the size of this position would come under intense scrutiny by
the CBT. To counteract such scrutiny Cargill entered into temporary futures-for-cash
exchange contracts with the Continental Grain and Uhlmann Grain companies. The
result was a reduction in Cargill’s reported position to 2.2 million bushels.
The grain business being a closely knit community, it was not possible for Cargill
to disguise the actual activities from the CBT business conduct committee, which was
only too aware of Cargill’s controlling position in the September contract open interest.
As the delivery month progressed, Cargill did little to reduce the size of its position,
resulting in it having an increasingly larger share of open interest in the contract. By
September 22, the pressure on the price of deliverable stocks was evident. Despite
apparent assurances to liquidate in an orderly fashion, Cargill did not move promptly
to reduce its position. In response, on September 23 a cease and desist order was sent to
Cargill, and on September 24 a trading halt was ordered and a settlement price for all
outstanding contracts was set at $1.10½, 2 cents below the close of September 22. The
impact on the cash market was predictable: the cash price fell, resulting in significant
losses for Cargill on the cash grain it held in company stocks, as well as on the cash
corn it had acquired in the cash-for-futures swap it had done with Continental and
Uhlmann. Cargill was incredulous, and a long series of CBT committees as well as two
court actions under the CEA were initiated. Hearings of the CBT board of directors in
March 1938 resulted in the expulsion of Cargill from the CBT. The resolution of the
CEA cases also went against Cargill, though only specific managers at Cargill were
sanctioned. In the end, Cargill, Inc. was still able to use the futures markets to facilitate
grain marketing.
Was Cargill engaged in legitimate hedging activities? Surely, there was a significant
hedging element behind some of Cargill’s futures activities? Cargill maintained that
insiders at the CBT acted to undermine the legitimate activities of a grain distributor,
with the members of the CBT business conduct committee directly benefiting from
the negative decisions made against Cargill. This accusation that the futures exchanges
act as monopolies seeking to further the interests of the members has been replayed in
other cases. For example, a similar comment applies to the Hunt silver manipulation.
The apparent evidence of manipulative activities on the part of Cargill was confirmed
in various forums, from CBT hearings to court actions under the CEA. Yet, somehow,
the arguments are not clear cut. The boundary between legitimate hedging, speculation
and speculation supported by manipulative intent is not clear cut. The hedging decision
involves a speculative component. Combined with sufficient impact in the cash market,
it is possible to rig the game. Precisely when rigging the game is happening is not as easy
to discern as might appear.
Manipulative schemers or not, Bunker and Herbert Hunt, in the summer of 1979,
had doubled their already colossal bet on the price of silver. In just their personal
accounts, including their half of IMIC and their existing holdings of bullion, they
had positions approaching 140 million troy ounces (a level they kept more or
less until the following March). At prevailing prices, the value of the silver they
controlled exceeded $1.3 billion, a large fraction of their net worth. With every
$1 movement in the price of silver, they gained or lost $140 million, an amount
substantial even to them.
Given the size of these positions, the Hunts made considerable gains from the run-up
in prices that started around August 22 and continued to September 18 – a rise from
$9.537 to $15.90.
Not unlike the Cargill grain case over four decades previously, this abrupt price
change surrounding a contract delivery triggered the oversight bodies within the
futures exchanges. On September 4, the first of a number of initial margin increases
was announced. In early October, the Comex set up the Special Silver Committee to
monitor the market and set rules as needed. Pressure was exerted on the visible longs,
primarily Conti, to facilitate an orderly liquidation of the December contracts. However,
until the December 1980 contract deliveries started to weigh on the market during
the delivery month, the principal shorts were not having difficulty locating bullion
for delivery. What did start occurring was a substantial decrease in market liquidity.
The principal commercial shorts were exiting the market, many using an exchange for
physicals (EFP) transaction.
An EFP is an off-exchange transaction in which the futures contract is settled by
delivery of a non-standard grade of the underlying commodity. An EFP is usually
motivated by a commercial transaction, e.g. a scrap copper producer can do an EFP
with a scrap supplier, where both are hedging using copper futures contracts. The
futures contract offset is bundled with the commercial transaction. During October
there were a number of large EFPs where major silver dealers (Mocatta Metals; Sharps,
Pixley; J. Aron) seemed to be delivering a large portion of physical silver inventories
to IMIC and others in exchange for cancellation of futures contracts with maturities
covering December 1979 through April 1980:
IMIC’s EFPs which supplanted most of its futures contracts, were perfectly
consistent with its avowed business purpose of acquiring physical silver. Coupled
with the deliveries already taken in September and October, IMIC had acquired
35.3 million troy ounces by mid-December, 27.8 million of that as bullion. Bunker
and Herbert Hunt themselves took delivery of 6.425 million troy ounces during the
fall of 1979. For the two Hunts, taking delivery afforded sizable tax advantages, given
the increase in price since the summer. According to US tax laws then applicable, a
liquidation of a futures position, including a rollover into a later month, triggered
a taxable event, upon which any gain would be taxed. In contrast, deliveries taken
were not a taxable event; the gain, if it existed, would be taxed only when the silver
was ultimately sold.
traded briefly above $50/oz. in the week prior, the close on Tuesday January 22 was
$34, a level that was maintained until mid-March when prices again fell precipitously
to the $17 level. In the interim, the Comex deemed that the pressure on the market
had eased sufficiently that the liquidation-only restriction on silver futures trading was
lifted. The price behavior had reversed the pressures on the commercial shorts, placing
the burden of variation margin squarely on the longs. The underlying strategy of taking
profits in bullion, through EFPs and standing for deliveries, turned on the longs with a
vengeance. The bullion, which could be used to secure financing, is declining in value
and can only be partially leveraged. Considerable cash on hand has been expended to
settle the EFPs.
Variation margin rules at the Comex and on most exchanges provide for daily
limits on the payments that have to be made to the account. This caps the daily cash
flow pressures, leaving a longer period of time for payment and the possibility that
prices will recover. Nonetheless, given a long enough time frame, the payments will
eventually be made. Englehard Mineral, an important commercial short, was reported
to have paid $1.3 billion in variation margin on silver futures up to mid-January.
Though the notional variation margin was over $1 billion in mid-March, the actual
payments required from personal sources was some $60 million per day for the Hunts
(Fig. 1.29). The cash flow pressure was such that on March 13 the Hunts and IMIC
defaulted on variation margin payments to their brokers. After a brief period during
which the brokerage houses covered unpaid variation margin balances, on March 27,
Figure 1.29 Variation Margin Payments for Hunt’s and Conti Positions, March 1980
1980, the final phase of the bubble took place, with brokers liquidating various cash and
futures positions.
Under the selling pressure of the brokerage house liquidations, the price of silver
reached $10.40/oz. The bubble had completely burst and new longs were entering
the market. The Hunts were forced to mortgage key assets in the family portfolio,
particularly Placid Oil, which secured a loan of $1.1 billion. By the end of April, the
outstanding balances for the Hunts at various brokerage houses had been paid. Though
the court cases dragged on for years, the immediate crisis was over. The usual array of
House and Senate subcommittees, regulatory reports and academic studies followed.
One key finding of the regulators was that the key brokerage houses acting for the longs,
Bache and Merrill Lynch, both acted imprudently by making large loans backed by
bullion. The solvency of the firms could have been put in jeopardy. Yet, like the Cargill
case, there is a strong case to be made regarding the lack of fairness in the exchange’s
treatment of the Hunts.
The resulting confusion associated with applying all these standards in a specific legal
situations is understandable.
The confusion surrounding manipulation extends to the jargon used to describe
the possible strategies. Some sources, e.g. Williams (1995, p.6), use the terms squeeze
and corner interchangeably. Others, e.g. Leuthold et al. (1989), require a controlling
position in both the derivative market and the cash market for there to be a corner.
For a squeeze, the trader only takes advantage of cash market shortages (oversupply)
by establishing long (short) positions in derivative contracts. Corners and squeezes
can occur from both the short and long side of the market, though most attempts in
practice are from the long side. For a long corner, the trader establishes long derivative
positions, typically well in excess of available deliverable supplies. At the same time,
the trader has attempted to obtain a controlling position in the available supply. The
process of standing for delivery on the contracts forces the short side to pay high prices
to bring available supplies on to the market. Yet, those available supplies are controlled
by the holder of the long derivative positions. The shorts are forced to go “hat in hand”
to the longs to cover their positions.
The textbook description of a long side corner is usually more involved in practice.
The Hunt silver operations leading up to the silver price peak in 1980 had elements of
a corner, but there are real questions about whether the Hunts and their confederates
had controlling positions in both the spot and futures markets. The Sumitomo copper
operations, which ended in 1995, offer a much better example of a corner. This particular
operation was spread over a long period of time, with the position in the deliverable spot
commodity growing gradually, starting around 1986 when Yasuo Hamanaka assumed
control of Sumitomo’s team of copper futures traders. In the Sumitomo case, a plausible
explanation was given for the build-up in stocks: even before Hamanaka began his
trading activities for Sumitomo, the firm was an important player in the international
copper market.
The Hunt silver operation had many of the earmarks of a traditional corner.
Swashbuckling entrepreneurs making big bets on rigged games. The evolution of the
Sumitomo copper corner has a decidedly more modern flavor. Hamanaka was a career
man at Sumitomo, with a 20-year history in the company division. Whether more
senior Sumitomo executives were aware of his activities is not clear. However, in any
event, Hamanaka was legitimately able to assume huge positions in cash and futures on
Sumitomo’s behalf. He also had signing authority over various corporate bank accounts
and access to corporate lines of credit. Hamanaka also was able to geographically
disperse his positions around the globe and exploit the laxness of regulators in specific
jurisdictions. For example, Hamanaka did a considerable amount of trading on the
London Metal Exchange, which—together with the Comex—is the most important
market for forward and futures trading of copper. Despite being vigorously warned
about possible wrongdoing by Hamanaka as early as November 1991, the LME did not
get actively involved in serious investigations of Hamanaka until the CFTC became
involved in October 1995.29
Due to filing requirements and other regulatory oversight, cornering activities in
modern markets require considerable effort to avoid detection. The elaborate network
of traders involved in the Hunt silver operations was needed to avoid the appearance
that large positions were being accumulated on one side of the market by one group
of traders. Despite the laxness of the LME, Hamanaka had to enter into arrangements
to hide the total size of Sumitomo’s position in deliverable supplies of copper. A
combination of LME regulatory laxness and careful planning permitted Hamanaka
to successfully deny involvement in a market manipulation. This despite a number
97
Hamanaka’s unauthorized trading activity was done without the awareness of superiors
at Sumitomo.
The strange tale of Yasuo Hamanaka and the Sumitomo Copper Scandal begins in
1985 when Hamanaka incurred an unreported loss of $30 million in the physical trading
of copper and tried to cover up the loss with more unreported deals. Civil and criminal
court testimony by Saburo Shimizu, Hamanaka’s boss in 1985, indicates that Shimizu
told Hamanaka the appropriate way to recoup the loss was through speculating on the
London Metal Exchange (LME) (Furukawa, 1997). In 1986 Hamanaka was chosen by
Sumitomo to lead a copper futures trading team with authority to accumulate large
spot inventory positions and signing authority over several bank accounts (Dwyer
1996). In July 1987 Shimizu resigned from Sumitomo because he did not want to accept
reassignment to its Manila office. From this point, Hamanaka assumed complete control
of the firm’s copper trading activities, greatly facilitating a decade-long pattern of rogue
trading. The ability of Hamanaka to avoid detection for such a long period speaks to the
difficulties of regulating corporate commodity trading in the modern era.
Perhaps the most troubling aspect of this debacle was the failure of regulators to
react to reports of manipulative trading that appeared well before 1995. As early as
1991, the effects of Hamanaka’s speculative trading appeared to impact price discovery
in the copper market. In particular, during 1991 copper prices on the LME rose to
levels considered too high by market observers to be justified by pricing fundamentals.
At the time, this distortion in prices was attributed to a technical supply squeeze that
resulted in a backwardation between the spot and futures prices. Because copper
for immediate delivery was in short supply late in the year, the LME became suspicious
that Hamanaka was purchasing high volumes of copper to manipulate the market.
Hamanaka denied this, arguing that Sumitomo always maintained large volumes to
satisfy its customers’ needs. In response, the exchange imposed somewhat arbitrary
limits on how far the backwardation could range (Jenkins 1996). It was later revealed
that, in November, 1991, David Threlkeld, an American copper trader based in London,
notified the LME about a request he had received from Hamanaka asking him to
backdate confirmations for a deal worth more than $400 million. The LME contacted
Sumitomo, but the corporation denied any wrongdoing, so the investigation went no
further (Fennell, 1996).
Just two years later, in 1993, another technical squeeze on the LME occurred, and
member traders complained that Sumitomo controlled much of the copper supplies
in LME warehouses and refused to release it (Gooding 1996). Once again, the LME
intervened by putting limits on prices and pressuring Hamanaka to release significant
holdings of copper back into the market. The LME justified these actions by claiming
that market fundamentals had been unnaturally thrown out of balance (Jenkins
1996). Threlkeld, who alerted the LME to Hamanaka’s trading methods in late 1991,
identified essential operating assumptions of a free market economy when he told the
Metals Weeks Copper Conference in 1992: “Markets run on trust, that they are fair and
equitable, that they protect the participants by protecting the market first, that they
maintain its financial and ethical credibility.” Ironically, it was because Hamanaka was
given so much trust by Sumitomo that he could control the copper market as much as
he did from 1986–95.
Apparent price irregularities in the copper market continued for the next two
years. Spot prices were typically higher than forward prices. Such backwardation was
historically unusual in copper because of the storage and transaction fees necessary
for forward purchasing of copper in combination with typically plentiful supplies
(Glasser and Barbash 1996). Various traders observed that Hamanaka’s trading prac-
tices were the cause of the backwardation, some accusing Sumitomo of trying to
corner the market by holding large supplies and driving up the prices. In October,
1995, the New York Mercantile Exchange notified the CFTC that copper supplies in
the LME Long Beach warehouse in California were growing significantly, indicating a
technical squeeze on the market (Dwyer 1996). Though there was little doubt who the
principal suspect was, Hamanaka insisted that his trading was influenced by “nothing
more than fundamentals” (Furukawa 1995). The CFTC contacted the LME and, by late
November 1995, the British Securities Investment Board joined the investigation. Sumi-
tomo agreed, in early April 1996, to make Hamanaka available for questioning (Dwyer
1996). In early May, 1996, the investigators notified Sumitomo that its copper trading
practices exhibited apparent abnormalities (Glasser and Barbash 1996). On May 17, after
discovering unaccounted for bank transactions, Sumitomo removed Hamanaka from
his position as the company’s chief copper trader (Burgert and Furukawa 1995). During
this period copper prices fluctuated wildly as traders first heard rumors that Hamanaka
was buying all the copper he could, and then rumors that Sumitomo had fired him.
On June 5, 1996, Sumitomo learned of copper trading losses of $1.8 billion at
the current copper price, though it was predicted that the final figure might reach
$2.5 billion if the price of copper fell. On the same day Hamanaka confessed to ten
years of rogue trading in the commodity futures market (Burgert and Furukawa 1996).
On June 6, owing to rumors about Hamanaka’s fate, the LME price for copper fell an
unprecedented 15 percent in two hours. Someone on the exchange said, “Our copper
dealer was having to quote a price every second, I thought he was going to have a heart
attack.” The price had dropped 25 percent in only six trading days. (Gooding 1996, June
15) On June 7, however, copper prices rebounded, but they remained under pressure
throughout this period, pushing the backwardation spread to $330 a ton on the LME.
The biggest losers were apparently those involved in delta-hedging with long-term put
positions. On June 13, Sumitomo made its 51.8 billion yen loss public, and the same
day the firm fired Hamanaka. Because traders feared Sumitomo would dump much
of its massive inventory of deliverable copper, despite assurances from the firm that it
would honor all its obligations and would not unload large amounts of copper, by
June 25, 1996, prices dropped to a two-and-a-half-year low on the LME and a new
low on the Comex.
Against this market backdrop, the CFTC and the British Serious Fraud Office vowed
to investigate all the companies involved in Sumitomo’s copper activities (Fennell
1996). Winchester Commodities in London and Global Minerals and Metals in New
York were investigated, but both denied any responsibility for Sumitomo’s losses.
Throughout the crisis phase of the debacle Sumitomo officials repeatedly insisted that
they had no knowledge whatsoever of Hamanaka’s trading activities (Fennell 1996).
The rest of the story involves the legal consequences of Hamanaka’s rogue trading. On
October 22, 1996, the Tokyo police arrested Hamanaka (Taylor 1996). On November
13, he was formally indicted for fraud and forgery (Furukawa 1996). Hamanaka’s trial
began in the spring of 1997, and on March 26, 1998, he was sentenced to eight years in
prison. Throughout all this time Hamanaka never revealed exactly how he had traded,
whether or not he had attempted to control the global copper market, and whether or
not Sumitomo had been aware of his activities (Shirouzu 1998).
In all the accounts of Yasou Hamanaka, there appears to be a substantial
discrepancy between the man as a person and the man as a trader. As one observer
put it: “Hamanaka is a quiet family man, who wears conservative suits and glasses,
drives a modest car, and lives in a medium-sized suburban house. But give him
a telephone and vast sums of money to invest, and suddenly he’s stomping through
world metals markets like Godzilla in downtown Tokyo” (Moffett 1996). At the time
of the scandal came to light, the 48-year-old Hamanaka had developed a number
of popular titles in addition to “Mr Five Percent,” including “Mr Copper,” “The
Hammer”—even “The Mad Jap”—because of his famous influence in the world of
copper trading (Jenkins, 1996). As another observer put it, Hamanaka was “the puppet-
master who pulled the string of traders in the ‘copper ring’ at the London Metal
Exchange . . . and bent a $1.45 trillion market to his will” (Dwyer 1996, p.28). And
yet outside the world of Sumitomo and the copper market Hamanaka was anything
but remarkable.
Hamanaka was a graduate of the Seikei University law school, initially joining
Sumitomo in 1970 and working his way up to the position of chief copper trader in 1986.
This was a highly trusted and respected position. Known as a cheerful, hard-working
employee, the head of a family with a wife and two children, Hamanaka seemed to
be the average salaried Japanese man dedicated to the conservative firm for which he
worked (Dawkins 1996). Being a notorious chain smoker was perhaps the only outward
sign of the inner tensions that Hamanaka must have felt (Dwyer 1996). At Hamanaka’s
trial, his former boss, Saburo Shimizu, confirmed Hamanaka’s claim that his motive for
engaging in speculation in the commodities and futures market was simply the desire
to recoup losses that he had already incurred (Furukawa 1997). Presumably, Hamanaka
was ashamed of these losses, so he did not report them. He did not appear to have any
intention to defraud Sumitomo for personal gain. His only intention seems to have been
recouping losses in the interest of the firm.
At trial, Hamanaka stated that his boss Shimizu first suggested that they engage
in speculation on the LME (Furukawa, 1997a). In defense of his action, Hamanaka
claimed he was only following orders—like an ordinary salaried man. However,
once Hamanaka set out to recoup the initial losses a transformation took place. To
In many instances where derivatives have been misused, egos overtook rationality.
Rather than using derivatives objectively, users became addicted to a transaction’s
profit aspect. Instead of attributing the transaction’s performance to market
conditions, they believed they had developed an innate skill to outperform the
market consistently . . . the moment fortune turns, panic often ensues.
Though Hamanaka did not show signs of panic, no matter how much money was
lost, he did display a gambler’s irrational drive to keep on betting. As Dawkins (1996,
p.8) observes: “Rogue traders are, by definition, hard to stop in any company of any
nationality.”
The scope of the Sumitomo copper scandal raises questions as to how Hamanaka
could have amassed such great losses over a ten-year period without anyone in the
company knowing what he was doing. Various market observers maintained that
Sumitomo, in spite of its typical Japanese corporate culture of trust, must have been
involved. Whether or not Sumitomo knew what Hamanaka was doing, the firm was
quick to dissociate itself from Hamanaka, publicly denouncing him as a criminal
(Furukawa 1996). If this caused Hamanaka any grief, perhaps a greater cause for
sorrow was his ultimate downfall in the market itself. “One lesson to be learned from
the Sumitomo affair is that it is clearly impossible for any player to defy market forces
for very long” (Dwyer 1996, p.29). Ten years is a long time. “No one really knows—at
least publicly—how a trader loses that much money over a decade as Sumitomo officials
claim the deviant trader did. Because the answer lies with Hamanaka and Sumitomo
officials, who have not divulged the losing strategy nor the company’s current positions,
it is difficult even for those trading copper every day to determine what went wrong”
(Kharouf 1996, p. 66).
Even after Hamanaka’s trial and subsequent imprisonment, many questions remain
that can only be partially addressed with available facts and some educated guesses.
Hamanaka traded on the LME, and “around” the LME. Sumitomo was not a ring
member of the LME. Normally trading is done with brokers or ring members and
trades are reported on the exchange. However, it is also possible to trade “on the
LME curb” (Furukawa, 1996). The bulk of Hamanaka’s trading involved such over-
the-counter transactions and third-party business not reported on the exchange
(Taylor 1996). That Sumitomo used the derivatives market in copper to compensate for
its relatively limited control of physical copper supplies (Dawkins 1996) was unusual
among Japanese firms. Hedging copper market dealings by trading derivatives to
supplement buying copper on the spot market was not common practice (Gooding
1996). In hindsight, it is surprising the amount of trading Hamanaka did that was not
reported on the LME.
Hamanaka seems to have been drawn to the LME rather than Comex in New York
and Tokom in Tokyo because the LME is basically a forward market, whereas the other
two are futures markets (Furukawa 1996). Many analysts believe that Hamanaka’s
main interest was in forward trading because this allowed him to take three-month
positions without putting up margins and to defer settlement with brokers without
losses being reported. Essentially, a forward trader, unlike a futures trader, operates
with a credit line, but once the credit line is reached, the player cannot go on (Burgert
and Furukawa 1996). Hamanaka’s ability to keep the trading scheme going for
ten years not only demonstrates ingenuity, but also the extent of Sumitomo’s credit
in the market. Hamanaka’s forward trading was combined with trading in the cash
market for physical copper. The purpose of this strategy seems to have been to corner
as much copper supplies as possible without being detected. The more certainty
Hamanaka had in controlling the price of spot price copper and the forward copper
curve, the more certainty of profiting from derivatives trades linked to copper (McGee
and Frank 1996).
An ongoing complaint that other traders brought against Hamanaka was that he
was amassing physical copper in exchange warehouses, leading to a supply squeeze that
drove up futures prices. This would also drive up the value of copper stocks, further
benefiting Sumitomo (Dwyer 1996). Considering the extent of Hamanaka’s trading and
the seemingly effective complexity of his apparent strategy, it is surprising that he kept
on losing money. One analyst has referred to Hamanaka’s “staggering incompetence” in
the market, while another referred to his kamikaze or suicidal trading practices (Jenkins
1996). No matter what Hamanaka’s strategy was, he still had to make good bets, and it
seems that he often failed to do this. As the losses continued to mount, it is likely that
Hamanaka borrowed to cover them. It seems that his reputation was so esteemed that he
was able to borrow from banks without any higher authorization. It is, however, difficult
to say how much Sumitomo executives knew about Hamanaka’s borrowing. One of the
banks, Merrill Lynch in London, claimed after the scandal came to light that Sumitomo’s
account was properly authorized.
Like various other rogue trading operations, the inability to settle ongoing losses led
eventually to the unraveling of the scheme. By 1994 Hamanaka was having difficulty
borrowing in London, so he turned to New York, where he borrowed nearly a billion
dollars from J.P. Morgan and Chase Manhattan, allegedly by falsifying and forging
documents (McGee et al. 1996). By this time Hamanaka appears to have been desperate,
for he probably saw the end of his career approaching.
Reserves, the trading agency for the State Reserve Bureau of China (SRB). Mr Liu built
a substantial short position in one- to three-month copper forward contracts traded
on the London Metal Exchange (LME), betting on a downward trend for the copper
price. The size of the position built up in the first two weeks of September was estimated
at 220,000 metric tons of copper with delivery dates mostly due on December 21 and
others reaching into 2006. Unfortunately for Liu, other copper traders and several
international speculative funds took large long positions in copper futures trying to
squeeze SRB’s short position. At the end of 2005, the rise in copper prices produced a
cash loss to the SRB estimated to be US$150 million, needing the SRB to either settle or
extend future contracts to later delivery dates, some to March 2006 and some to 2007.
While the story of Yasuo Hamanaka originates with a large Japanese conglomerate,
the SRB is not a corporate business entity but an internal department in the National
Development and Reform Commission of China (NDRC). As a government agency,
the SRB is responsible for managing national strategic material stockpiles that are
reserved for national defense purposes, for safeguarding the stability of the society,
and for mitigating unexpected disasters or catastrophes. One byproduct of the SRB’s
operations is the stabilization of market prices for reserved materials such as oil, cotton
and copper. Specifically, when the market prices of certain reserved materials, such as
copper, are considered “high,” the SRB might increase supply by auctioning part of its
stockpile. In turn, when the price is decreasing due to reduced demand, SRB might
increase inventories to add some extra support to the market price and build inventory
for use in future periods. In other words, the SRB can function, to some extent, as a
“pool” that can stabilize prices of some strategic materials.
In order to perform a price stabilization function, the SRB routinely participates in
open market operations to adjust reserved material stockpiles. However, as a government
agency, the SRB cannot directly participate in market trading, but instead has to adjust
stockpiles through another entity: the State Regulation Center for Supply Reserves
(SRCSR). This Center, although owned by the NDRC, is an independently registered
state-owned entity. It is a non-profit organization performing trading activities based
on the SRB’s authorization. The SRCSR actively participates in domestic and global
market trading, but the underlying commodities, the reserved materials, that the Center
trades are the property of SRB. In other words, the SRB, an internal department of the
NDRC, and SRCSR, an independent entity owned by NDRC, do not have a parent-
and-subsidiary relationship, but are principal and agent. In addition, the SRCSR is a
corporate body with registered capital and could be protected by bankruptcy laws if its
net liabilities (total liabilities - total assets) are greater than the capital invested by the
NDRC.
Contrary to western press reports, Mr Qibing Liu was not an employee of SRB but
a senior officer in SRCSR, e.g. Zwick and Collins (2006). This obscure administrative
detail explains why, when Liu “mysteriously” stopped trading in November 2005,
officers from the SRB repeatedly rejected any linkage between Mr Liu and SRB, e.g. Fu
Jing (2005). For example, Zwick and Collins (2006, p.14) report: “in November when
Liu abruptly stopped trading, and an SRB official was quoted as saying, ‘We do not
have such a person working for us,’ which later morphed into, ‘He was acting on his
own behalf,’ and eventually became, ‘He is on leave.’” From a legal perspective, it was
the SRCSR rather than SRB that was ultimately responsible for the short positions built
by Mr Liu. What is less clear is whether the SRB was aware that an official in the SRCSR
was trading on the LME using information obtained from the SRB about possible future
sales of spot copper.
Not unlike “Mr Five Percent,” Qibing Liu earned the nickname of “Mr Mangler” in
the Chinese copper market. Liu entered SRCSR following a bachelor’s degree in
economics. In 1995, he was sent to the LME and was trained there for half a year.
In 1999, Liu was promoted to be director of Import and Export Department in SRCSR,
and was authorized to place orders on LME and the Shanghai Futures Exchange
(SHFE). Upon becoming the chief trader of SRCSR around 2002, Liu started building
a reputation among Chinese traders by precisely forecasting a bullish copper market
while most other traders were predicting the market would not change significantly.
Starting around 2003, Liu started to arbitrage copper prices between the SHFE and
LME. Liu was typically long copper futures on the SHFE, simultaneously entering an
equivalent amount of short positions on the LME. The fundamentals underlying the
trades were institutional limitations on the SHFE: foreign investors and international
speculative funds are not allowed access to the SHFE and, as a result, there are often
disparities of prices between the Chinese and global copper markets.
Liu was apparently adept at taking advantage of the disparities between SHFE
and LME, and was successful at generating some arbitrage profits. As the major long
position taker in SHFE, SRCSR and its head trader Qibing Liu were a nightmare
for short position takers, which—as noted above—earned Liu the nickname of “Mr
Mangler” among Chinese copper traders, e.g. Le-Min Lim (2005). Unfortunately, Liu
did not keep to a pure inter-market arbitrage strategy. His trading strategy was observed
to gradually change at the end of 2004. Success in Shanghai apparently nourished an
overconfidence, making him willing to take more risks. Liu evolved, little by little, from
a market arbitrageur to a speculator. In September 2005, likely based on information
about impending spot sales of copper by the SRB, Liu believed that copper price would
turn downward in the near term and started increasing short positions on the LME. In
the last ten days of September, Liu took up as many as 8,000 copper future contracts.
For liquidity reasons, these contracts were mostly for December 21 delivery. The average
price of those contracts was about US$3,500 per metric ton.
As the LME copper future contract is for 25 tonnes of copper, the 8,000 contracts
meant that SRB had agreed to deliver approximately 200,000 tonnes of grade A copper
in three months. Such a short position, even on the LME, was a huge number. This large
exposure was noticed by experienced LME traders who apparently concluded that the
SRB did not have enough deliverable copper stockpiles to execute those contracts. This
meant the SRB would have to close these positions by taking enough long contracts to
offset the uncovered short positions. In combination with the LME traders, international
speculative funds marshaled “hot money” to buy copper futures and forwards adding to
the SRB demand, pushing the copper price to a record high.
On November 14, 2005, news broke that Qibing Liu was missing, bringing the SRB
short positions on LME into the global news spotlight. Was Liu another rogue trader in
the same vein as Nick Leeson, Yasuo Hamanaka or Jiulin Chen? While the world was
guessing where Liu was, officers from SRB said to the influential China Daily newspaper
that “we do not have such a person working for us!” and “investigations show that
Mr. Liu’s market behaviors are not authorized and the SRB should not be responsible
for the short positions” (Fu Jing 2005). Given the administrative relationship between
the SRB and SRCSR, such comments were legally correct. This dramatic unofficial
announcement worried the LME’s clearing house members and brokerage firms
with which Liu had opened accounts. A lengthy and costly lawsuit was inevitable if
SRB insisted that Liu was trading for personal, unauthorized reasons. In the event of
such fraud, the short position would be non-executable. Shortly thereafter, the SRB
took over the positions created by Liu and continued efforts to restrain market price
increases.
More precisely, on November 11 the SRB released information about auction sales
of copper that apparently benefited short position takers. A senior officer from SRB
informed reporters that the SRB copper stockpile was unexpectedly larger than
forecasted. At this time, market estimates placed the Chinese government copper
reserves at about 250,000 tones of copper. However, the SRB officially disclosed a
stockpile of 1.3 million tonnes. With such a large stockpile, SRB had the capability to
deliver the underlying copper when the futures contracts expired. This was clearly not
information the LME traders and speculative funds wanted to see. The market sus
pected the genuineness of such information; most traders did not believe it was true:
“If you believe China has these stocks, then copper is not worth what it’s at right now”
(Wall Street Journal 2005). A stockpile as large as 1.3 million tons of copper would give
SRB great flexibility to cool the heated copper market, but the SRB did not respond
until Liu had built such large short positions. Consequently, a reasonable explanation
was that SRB did not have that much copper in storage.
On November 16 and 23, the SRB launched two auctions, each time selling 20,000
tonnes of copper reserves. Although the copper price fell sharply as a reaction to the
auctions, it soon rebounded and reached an even higher level. On November 30 and
December 7, SRB again auctioned more copper to the market. In theory, such extra
spot supplies should force the market price to a lower level. However, the copper price
quoted by LME moved in an opposite direction: from US$4,100 before the first auction
to US$4,400 after the fourth. On December 21, the expiry date of SRB’s future contracts,
the market experienced a smooth day because most of the traders on the LME already
knew SRB’s choice: partial execution and partial extension. The execution was proven
by the increase of copper stocks in the LME’s warehouse in Korea. After December 21,
LME’s copper stocks in Korea increased by around 50,000 tons of copper, most of which
was transported from China. The consensus among copper traders was that the rest of
SRB’s short positions were extended until 2006 and 2007. The first round of the battle
ended with SRB’s realized loss of about US$45 million.2
The SRB and its authorized trading agency, SRCSR, both fell into the trap of
participating in unauthorized financial speculation. As the government agency manag
ing strategic reserve materials, the SRB should have avoided speculating for profit.
Unfortunately, Qibing Liu’s market activities, both market arbitrage and the more
speculative bets on LME, were apparently profit-driven behaviors. On January 24, 2006,
the head office of SRB released a public notice to its nationwide branches reiterating that
the entities involved in managing reserve materials are not allowed to participate in
trading of financial derivatives or to invest in foreign markets. Significantly, the notice
was dated October 25, 2005, a date on which Qibing Liu was already on leave but the
whole matter had not been disclosed to the public. Similar to the case of Barings Bank,
the SRB’s loss can be explained as being due to operational risk. As Nick Leeson wrote
in Rogue Trader: “It’s unbelievable that nobody comes to stop me.” Nobody came to
stop Liu either, until it was too late. For breaking various Chinese state regulations, Liu
Qibing was sentenced to seven years’ imprisonment in March 2008.
When SRCSR started its operations in the futures and forward market, it had a
trading group (the number of traders at that time is unknown). At the beginning
of 2004, only Mr Liu and another trader were left in that trading group. About six
months later, the other trader was removed from his position and Liu became the only
authorized trader with rights to place orders on both the SHFE and LME. This situation
was ongoing for more than one year until the outbreak of the event. Both SRB and
SRCSR were apparently lacking sufficient internal control systems, as demonstrated
by the short position Liu was able to establish. Both the SRB and Chinese regulators
did not play a proper role in stopping the speculative trading of Liu. Though the
trading of financial derivatives is supervised by the China Securities Regulatory
Commission (CSRC), the CSRC did not supervise the market activities of other
government agencies such as SRB. In addition, the CSRC’s supervision does not cover
trading activities conducted by the Chinese central bank and commercial banks,
although these entities are extensively involved in such trading.
To firms outside China, the Liu Qibing affair has produced some unexpected credit
risks related to Chinese customers. In the case of SRB, for example, the brokerage firms
with which Liu opened accounts were exposed to considerable credit risks. Qibing Liu
worked for SRCSR, an agency performing trading under the authorization of SRB.
However, the internal control policies and authorization systems taken for granted in
conventional commercial operations might not be properly designed or executed by
Chinese companies and administrative entities. In addition, Chinese firms apply an
accounting system that is different from the International Accounting Standard (IAS)
or American General Accepted Accounting Principles (GAAP). As illustrated in the
case of China Aviation Oil Corporation Ltd (CAO), for example, the company was
bankrupted in 2004 due to a debacle caused by the derivative trading of Jiulin Chen, the
CEO of CAO, before the losses were noticed by the public. The rapid economic growth
of China and the global economy means that more and more Chinese firms will present
themselves in the global arena. How to manage commercial risks of doing business with
Chinese firms has become an important topic in commodity risk management.
value for their end-user. By 2005, this strategy had produced acquisition of: South China
Bluesky Aviation Holding Company, Ltd; Shuidong Oil Storage Tank Farm; Shanghai
Pudong International Airport Aviation Fuel Supply Company Ltd; Compania Logistica
de Hidrocarburos, S.A.; and an acquisition agreement with the Emirates National Oil
Company for a subsidiary purchase (China Aviation Oil 2005). For the international
oil trading component, China Aviation Oil traded a wide variety of products including
crude oil, fuel oil, gas oil, gasoline, naphtha and petrochemical.
It was the jet fuel procurement component of CAO activities that was the source of
the large losses. The company’s jet fuel procurement business was heavily dependent
on the parent group as CAOSC was the sole entity authorized by the People’s Republic
of China (PRC) government to import jet fuel into the PRC (PricewaterhouseCoopers
2005). Jet fuel procurement still remains an important component of CAO’s businesses
as CAO supplies nearly 100 percent of the imported jet fuel for China’s aviation oil
industry. Despite the losses suffered in 2004, CAO continued as a publicly traded entity
on the SGX. Future prospects for CAO are excellent. Jet fuel procurement activities
alone promise approximately 10–15 percent growth in the next ten years. Though the
losses suffered in 2004 by CAO are largely in the past, the story of Jiulin Chen survives
in the chronicles of commodity risk management debacles.
Jiulin Chen was born in 1961. Despite a childhood in the turbulence of the Cultural
Revolution, Chen managed to get enough schooling to attend Beijing University in
1982, studying Vietnamese in the Department of Orient Languages for five years,
receiving a Bachelor of Arts. This was followed with a postgraduate diploma in law
from the Chinese University of Political Science and Law in Beijing. At the time of
the losses, Chen was pursuing a PhD in law at Tsinghua University. Before joining
China Aviation Oil, Chen acquired management experience with Air China. In 1993,
he started working with CAOSC as the chief negotiator and project manager on
various projects, later moving to be the managing director and CEO. In 1997, Chen
was given the assignment of establishing CAO in Singapore with an initial capital of
US$210,000. Chen successfully combined the strong demand for aviation oil in China
with previous business experience to turn the small firm into an company worth US$326
million by 2003. This success had raised his salary to S$4.9 million, making Chen one of
the best-paid executives in Singapore (CRIOnline 2004).
Before 2002, CAO used derivatives only for hedging purposes, employing crude
oil futures and swap contracts to hedge the price risk exposure associated with the jet
fuel and fuel oil cargoes. Starting in 2002, CAO began to use option contracts with
the objective of increasing firm profitability. An audit conducted by Pricewaterhouse-
Coopers found substantial evidence that CAO did not have enough knowledge and
ability to do the option trading that later caused considerable losses. Not only was it
not reported that options were used for speculative purposes, the size of the losses was
worsened by the inadequate risk management and corporate governance in place at
CAO during that time. CAO did not know how to value the option portfolio properly for
the accounting reporting purposes required of publicly traded companies in Singapore.
Finally, CAO management did not provide enough disclosure regarding the speculative
derivatives operation. The inadequate loss recognition associated with the derivative
transactions resulted in major inaccuracies in the 2003 and 2004 financial statements
(PricewaterhouseCoopers 2005).
Initially, CAO involvement in options trading featured back-to-back transactions
between the company and certain PRC airline companies. In turn, CAO then resold
options with similar terms to non-PRC counter-parties. Because CAO bought options
from PRC airlines, CAO was not required to pay premiums to the airline companies.
By reselling the option and assuming the credit risk of the airline companies, CAO
earned premium income from the sale of the option to the other counter-parties
(PricewaterhouseCoopers 2005). Starting March 28, 2003, CAO commenced “specu
lative” trading in options in addition to its trading in futures and swaps. CAO
management felt that, by executing these options, more profit could be gained than
just the premium income from reselling the options (Santini et al. 2004). These trades
were at first restricted to only Gerard Rigby, the deputy Head of Trading Division I
at CAO, and Abdallah Kharma, head of Trading Division II. However, most of the
trades were done by Rigby, who had more previous experience in option trading
(PricewaterhouseCoopers 2005).
The evolution of speculative derivative trading at CAO is described by Price-
waterhouseCoopers (2005, p.viii):
[B]y 2003, the volume of derivatives traded well exceeded the volume for physi-
cal trades. This was accompanied by an increase in the revenue generated from oil
derivatives trading from 2001 to 2003, such revenue exceeding the revenue gener-
ated from physical oil trading from as early as 2001. In keeping with these trends,
by November 2004, the Company had a total of 9 oil traders.
During the first three quarters of 2003, company trading decisions assumed a bullish
view on oil prices and pursued a strategy of “structured collars”—buying calls and
selling puts—involving approximately 2 million bbl. of oil. This assumption proved
to be correct and yielded some profit (PricewaterhouseCoopers 2005). As the oil price
rose, call options that had been purchased were exercised and profits were registered
in the CAO financial statements. On the other side, puts that CAO sold could not be
exercised since the oil price had risen, allowing CAO to book the premium as a profit.
Starting in the fourth quarter 2004, the CAO view on oil prices changed to bearish.
This resulted in a change in the options strategy to selling calls and buying puts.
Presumably to increase premium income received, extendible features were attached
to the calls being sold. Unfortunately, the assessment of the oil price trend proved to
be largely incorrect. A sharp rise in oil prices caused the mark-to-market value of the
CAO position to deteriorate. As the oil price kept rising, the counter-parties exercised
the right to extend the term to maturity, generating further profit deterioration.
PricewaterhouseCoopers (2005) describes the situation:
As prices continued to trend upwards, the options had an increasing negative MTM
value. These options were maturing in 1Q 2004. This was probably an early defining
moment in the events that were to transpire. Those managing the Company took
the view (incorrectly) that unless the losses were realised, there was no requirement
to account for them in its financial statements. To avoid realising the losses and
in the hope that the situation could be managed, the Company entered into a
restructuring of its options portfolio with J. Aron on 26 January 2004. Further, the
Company (again incorrectly) did not book the losses that were realised upon the
restructuring when closing-out the loss-making near-dated options.
This situation led to the defining moment in the CAO debacle. Losing money on the
speculations would likely have been manageable if CAO management had employed
appropriate risk management and corporate governance practices. For example, impos-
ing a stop-loss strategy would have allowed the company to limit losses to a certain
amount. Apparently, CAO already had a stop-loss policy for the speculative accounts.
However, because of the CEO’s ambition to surpass past achievements, there was no
specific control as the CEO hired himself as the CFO as well. As a result, CAO was
allowed to trade more in the first quarter of 2004 even though it had experienced losses
of $5.5 million on the option trades at that time.
options contracts. The Company then exacerbated the situation manifold in the
restructurings that followed in June and September 2004 for the same reasons.
This was an imprudent course to take and it was ultimately the immediate cause
of the Company’s predicament.
If CAO had been a wholly owned PRC state enterprise, the losses suffered would
likely have past unnoticed to public scrutiny. However, CAO was publicly traded on the
SGX and, faced with the prospect of reporting significant losses, entered into a sequence
of ill-conceived decisions designed to avoid reporting the losses in the financial
statements. Instead of closing out the option position and realizing a loss of approximately
$5.8 million from the options maturing in the first quarter of 2004, CAO entered
into a restructuring of the options portfolio. In doing so, the company incorrectly
assumed that there was no requirement to book the losses. The restructuring
was effected by the purchase of option contracts to close out the loss on the short-
dated options, and financing the cost of this exercise by selling longer call options and
buying puts with higher strike prices, with maturity dates stretching from the second
quarter of 2004 to the fourth quarter of 2005. The aim of CAO management was
to achieve a zero net cash flow by matching the cost of buying out short-dated call
options with the premium revenue from that received by selling a higher volume of
call options.
As it turns out, the management objective of using premium income from new posi
tions to cover off losses from maturing positions is unacceptable from a public accounting
perspective because it overrides the full disclosure and recognition principle. Not only
had CAO hidden recognition of the option losses from the financial statements, there
was also no information released about the option portfolio restructuring. The person
held responsible for this action was Jiulin Chen as he was responsible for signing the
manipulated-financial statement results. The inaccuracies in the financial statements
being presented were identified by PricewaterhouseCoopers (2005) as:
where PBT is “profit before tax.” As the oil price kept rising in the second quarter of
2004 following the restructuring, the mark-to-market value of the CAO option port-
folio kept deteriorating. This caused margin calls to the company in May 2004. Yet
CAO trading, which was apparently guided mostly by the opinion of Jiulin Chen, still
believed that the oil price would eventually fall. Therefore, in the second quarter 2004,
CAO did another restructuring to avoid closing out and recording losses of around $30
million. Once again, Jiulin Chen misrepresented the financial accounts by withholding
information about the option restructuring and not recognizing the loss being incurred.
Two critical issues arise in regard to the second restructuring: the risk associated
with the CAO option positions; and the mark-to-market value of the positions. The risk
associated with the CAO positions was greater than following the first restructuring
due to a further round of closing out option losses from the first quarter by issuing
an even larger volume of call options with higher strike prices and, significantly,
with options that were subject to tighter margin call requirements if there was a
substantial oil price change. In September 2004, CAO did a third major restructuring,
again aimed at achieving a zero cash flow from the options portfolio in order to avoid
recording losses from the second quarter. The same approach was used to reach zero
cash flow from the portfolio, which involved closing out short-dated options by the
issuing longer-dated call options with substantially larger underlying oil volumes.
An important difference with this restructuring was the involvement of five counter-
parties. For the third time, Jiulin Chen manipulated the quarterly financial statements
being released.
By October 2004, CAO was holding trades that involved 52 million bbl. of oil. The
options portfolio held by CAO was now subject to margin calls even if the oil price just
rose slightly. This presented a serious cash flow problem for CAO, as it had used almost
$26 million of its working capital, plus a $120 million loan and $68 million from the
proceeds of a trade receivable to finance previous margin calls. In October, 2004, the
oil price reached what was, at that time, an all-time high of $55.67/bbl. on NYMEX.
“In the 7-month period from May through November 2004, a total of approximately
$381 million was paid to meet margin calls arising from the mounting MTM losses”
(PricewaterhouseCoopers 2005, p.xiv). As a consequence, CAO had to close out the
option positions as CAO did not have the financial capability to meet margin calls. On
November 30, 2004, CAO issued a press release stating the firm was “unable to meet
some of the margin calls arising from its speculative derivative trades, resulting in the
company’s being forced to close the positions with some of its counterparties” and would
have to recognize a $550 million loss.
The revelation about the hidden losses had frightening repercussions for
those involved. The SGX immediately halted trading on the stock and appointed
PricewaterhouseCoopers to investigate the losses. Following release of the auditors’
report in March 2005, on June 7, 2005, five CAO executives, including Jiulin Chen,
were charged with various offences. During the trial, further details of the debacle were
revealed. In particular, when the only remaining avenue of rescue was for CAO man-
agement to tell the parent company, CAOHC, the situation being faced in the hope of
financial help to reconcile the losses without publicly disclosing the information in the
financial statements, Jiulin Chen, CEO of CAO, told the director of Finance, Peter Lim,
that CAOHC would help CAO. To this end, Chen presented a document signed by the
parent company sufficient to preclude disclosure of the loss. However, Jiulin Chen had
forged the CAOHC document and the associated signatures (Sawyer 2005, p.44). While
the failure to disclose information in a timely fashion was a potential source of crimi-
nal prosecution, it was for the forging of documents that Chen was later to receive the
harshest penalties from criminal conviction. On March 21, 2006, the subordinate court
of Singapore sentenced Jiulin Chen to four years and three months’ imprisonment, and
fined him for S$335,000.
In addition to a variety of obvious failures, investigation of the events in the debacle
revealed that CAO employed improper option valuation methods to determine
estimates for the accounts. CAO was unprepared to handle the valuation of the exotic
options that came to dominate the portfolio. For accounting purposes, CAO used
intrinsic value to determine the value of options being traded. Given the intrinsic value
of an option is calculated by differencing between the forward (spot) price and the strike
price, this valuation method is not accurate as it ignores the option time value. This
failure to recognize the time value was compounded by the relatively long tenure to
maturity of the options. This practice was not problematic when CAO was generating
premium income by selling back-to-back options between PRC and non-PRC counter-
party airlines in 2002. However, the practice of using intrinsic value continued when
the company moved to speculative option trading until the announcement of the big
loss in 2004.
It is significant that the CAO debacle took place within a publicly traded company
that, to all intents and purposes, appeared to have already addressed internal risk
management issues. In 2002, CAO with the help of Ernst & Young developed a risk
management manual, which was approved by the board in March 2002. In hindsight,
it does not seem the board and the audit committee had sufficient time to digest all the
contents of the risk management manual. As a consequence, the board decided to approve
internal risk management procedures on a “test run basis.” An additional complication
identified by the auditors was that the risk management manual had been developed
for swaps and futures only. When CAO began option trading in 2002, it would have
been prudent to augment the risk management manual to include option trading when
options were introduced as new products. To this end, the risk management manual
should have: had a trading limit for options; distinguished between bought and sold
options as these carry different risk; distinguished between different types of options
with varying complexity and degree of risk.
Following the announcement of losses in November 2004, big creditors moved to
make claims. The various lawsuits and criminal actions provide some information on
the names of market players. In particular, Barclays Capital, Mitsui & Co. Energy Risk
Management, J. Aron & Co., Standard Bank (London), Macquarie Bank, Sumitomo
Mitsui Banking Corporation, and Fortis Bank were the entities making the largest
claims against CAO. Several lawsuits were filed because of the non-performance and
mis-management by CAO, e.g. SK Energy Asia Pte. Ltd filed a civil lawsuit requesting
removal of CAO’s management. Creditors requested that reorganization or bankruptcy
of CAO be handled by judicial oversight to ensure CAO creditors received full and
frank information before a rescue plan could be voted. The court decided to remove
old CAO management, with new management to oversee a debt restructuring. The
debt restructuring paid 56.1 percent of the money owing over five years with deferred
interest and immediate payment of 45 percent owing, a total of approximately $275
million. This debt-restructuring plan was agreed to by most of the creditors on June 8,
and the plan was ratified by the Singaporean high court.
Other civil lawsuits that were filed included a claim by Sumitomo Mitsui Banking
Corporation that CAOHC, as the parent company, and Jiulin Chen had conspired
against the bank in hiding the derivative losses in order to obtain a loan in mid-2004.
The court decided that Sumitomo would get equal treatment with the other credi-
tors. Another related lawsuit filed by a group of creditors claimed the parent company,
CAOHC, was aware of the problems at CAO. Instead of trying to fix the problem,
CAOHC on October 20, 2004, sold 15 percent of its ownership in CAO to Deutsche
Bank (DB). At this time CAOHC did not disclose any news to DB or other potential
stakeholders. The result of this deception against DB was that CAOHC was guilty of
insider trading practices and was required to pay a total fine of S$8 million to the Mon-
etary Authority of Singapore. The criminal lawsuits charged five executives and direc-
tors of CAO with an array of crimes including forgery, insider trading and filing false
financial statements. The CEO, Jiulin Chen, was charged with 15 offences including
forgery, filing false financial statements, lack of full disclosure on notifying the SGX
of CAO losses, and conspiracy to deceive Deutsche Bank AG on the sale of a block of
CAO shares. Jia Changbin, CAO chairman and president of CAOHC, was charged with
insider trading and two other offenses. Peter Lim, CAO’s Singapore finance director,
was charged with five offenses including issuance of false financial statements, and con-
spiracy with Mr Chen to cheat and deceive Deutsche Bank. In addition to the penalties
imposed on Chen, Jia Changbin was fined S$250,000 for trading on insider information
and S$150,000 for failing to disclose information about the losses. The following year Jia
Changbin was forced to resign as chairman of CAOHC.
The reasons behind the CAO debacle can be traced to a misguided speculative view
of the oil price trend. In addition to this view, the auditors also identified the following
sources of difficulty:
QQ a desire not to disclose losses in 2004;
QQ a failure to value the options portfolio in accordance with industry standards;
QQ a failure to appropriately recognise the correct value of the options portfolio in
the Company’s financial statements;
QQ the absence of proper and stringent risk management procedures specifically
for options trading;
QQ the willingness by the management to override risk management policies that
ought to have been obeyed; and
QQ a failure on the part of the Audit Committee in particular and the Board in gen-
eral to fulfil their duties in relation to risk management and controls applicable
to the Company’s speculative derivatives trading.
Ultimately, the person that was held liable for these difficulties was Jiulin Chen, as the
CEO of CAO. It was Chen that was the source of the inaccurate prediction regarding
the future of oil prices in the fourth quarter of 2003. Faced with relatively small
losses from this incorrect prediction, Chen attempted to cover the losses with option
restructuring and not recognizing the losses incurred in the financial statements. Chen
was also responsible for ensuring the appropriate option valuation method was used.
Valuations done using conventional option valuation software were ignored when the
numbers did not conform to the values Chen was wanting. Finally, application of risk
management controls was lacking. To address this problem, the restructuring of the risk
management at CAO has produced the risk management structure given in Fig. 1.30.
to speculating in energy derivatives by MGRM, it turned out that MGRM was actually
engaged in a sophisticated long term marketing program for gasoline and heating oil.
The saga of how a firm engaged in hedging activities could incur such losses has been
told and retold, often brilliantly, by Culp and Miller (1994, 1995), Mello and Parsons
(1995), Kuprianov (1995) and Edwards (1995).
Mello and Parsons (1995) outline the background to the MGRM saga:
MGRM was involved in intermediating the spot market for oil products with the long-
term forward market. For this business strategy to work, MGRM had to be directly
involved in sophisticated commodity risk management. Though some of the risk could
be captured with longer dated OTC products, to accurately handle the risk it was
assuming for customers, MGRM also had to use oil complex futures contracts. Due to
limited contract liquidity for longer delivery dates, MGRM had to implement a rolling
stack hedging strategy, involving short-dated futures contracts.
As demonstrated in numerous sources, e.g. Culp and Miller (1995), a rolling stack
hedge can have a sizable basis risk. For the MGRM story, this basis risk was dramatically
compounded by variation margin costs and certain peculiarities of German accounting
principles. As a result, a promising business plan was destroyed by inadequate
execution. That MGRM had a business plan is apparent. The plan commenced
with the recruitment of a management team with a track record in implementing a
similar plan at Louis Dreyfus Energy Corporation. The program was featured on the
cover of the annual report of the parent corporation, MG AG. Under the forward
supply or “flow delivery” contracts MGRM had contracted to deliver approximately
160 million barrels of associated oil products, primarily heating oil and gasoline, at fixed
prices under contracts stretching out ten years. These contracts had a sell-back option
clause, permitting the counter-party to terminate early if the market price was some
threshold greater than the fixed price at which MGRM had contracted to deliver. The
counterparties in these contracts were a mix of retail gasoline suppliers, large industrial
corporations and a few government bodies.
The fixed-price contracts written by MGRM provided for a spread over current spot
market prices of from $3 to $5 per barrel, with many of the contracts being written in
the summer of 1993. This was the profit margin that MGRM had to design a hedging
strategy to protect. The unhedged risk to MGRM was that prices would rise and MGRM
would be obligated to deliver oil products at lower than market prices. To hedge this spot
position, circa late 1993, MGRM had a position of 100 to 110 million barrels in OTC
energy swaps and 55 million barrels in heating oil and gasoline futures on NYMEX. It
seems that MGRM was pursuing a long one-to-one hedge. An important complication
facing MGRM was the lack of liquidity in long-dated maturities for both futures and
swaps. Instead of implementing a relatively riskless strip hedge, which seeks to match
the maturity of the futures contract delivery with the date of the future spot transaction,
MGRM was obliged to use a rolling stack hedge. Apparently, this was considered to be
a benefit to MGRM, due to rollover gains implied by a one to one hedge when futures
prices are in backwardation.
Unfortunately for MGRM in the latter part of 1993 oil prices fell. While this would
be an excellent outcome for an unhedged MGRM, the long hedge positions started
requiring significant amounts of variation margin. In addition, futures prices went into
contango, dictating rollover losses instead of rollover gains. These negative variation
margin cash flows were not matched by offsetting mark to market gains on the long
term forward delivery contracts. Such was the business risk that MGRM assumed.
Prices fell from the $19 level to below $15, combined with the rollover losses, this meant
cash flow requirements to the hedge in the hundreds of millions of dollars. As it turns
out, German accounting principles, which were applicable to the parent corporation,
required the classification of these variation margin payments as losses. In what can
only be described as a classic case study in strategic risk management, on December 17,
1993, the supervisory board of Metallgesellschaft fired the management board chairman
and brought in new management with a mandate to liquidate both MGRM derivative
security positions and its forward supply contracts.
The end result of the supervisory board decision can be estimated at $640–$800
million on the derivatives positions alone. The cancellation of the forward supply
contracts was done without penalties, thereby releasing the counterparties from what
was a positive cash flow situation for MGRM, again losing value. The MGRM saga
has several key questions to examine. Among these points, one stands out: what were
the members of the supervisory board thinking about when they pulled the plug on
the operation? Unfortunately, the deliberations of the board, such as they were, are
hidden behind the veil of corporate secrecy. It is apparent that the hedging strategy
that was implemented was not well understood ex ante by the supervisory board. As
such, Metallgesellschaft failed to follow a tenet of strategic risk management: that the
(1999, p.40) defined the term “to refer to a variety of pooled investment vehicles that are
not registered under the federal securities laws as investment companies, broker-dealers,
or public corporations.” A similar definition appears in an SEC staff report on hedge funds
appearing in 2003 (SEC 2003), with the clarification that a hedge fund “is not registered as
an investment company under the Investment Company Act.” This recognizes ongoing
efforts by the SEC to regulate hedge funds under the Investment Advisors Act (1940),
e.g. Pekarek (2007). The continuing lack of regulatory oversight is not due to vigilance
by US regulators. Substantive changes have been implemented in July 2011 as part of
Dodd-Frank initiatives. New rules—the Private Fund Investment Advisers Registration
Act of 2010—have tightened registration, reporting and record-keeping requirements of
the IAA to include hedge funds.
SEC (2011) marks a considerable advance from PWGFM (1999, p.40), where it is
observed that: “The term ‘hedge fund’ is not defined or used in federal securities laws.”
Under Dodd-Frank initiatives, hedge funds are now loosely defined in SEC (2011,
pp.22–3), where Form PF filing requirements for “private funds” are detailed. More
precisely:
Form PF defines “hedge fund” generally to include any private fund having any one
of three common characteristics of a hedge fund: (a) a performance fee that takes
into account market value (instead of only realized gains); (b) high leverage; or
(c) short selling . . . [A] commodity pool that is reported or required to be reported
on Form PF is treated as a hedge fund.
Altogether, the seven types of private fund defined in Form PF are: (1) hedge fund;
(2) liquidity fund; (3) private equity fund; (4) real estate fund; (5) securitized asset
fund; (6) venture capital fund; and (7) other private fund.
By significantly increasing the variety of funds that would qualify as hedge funds, the
introduction of this legal definition of a hedge fund changes the landscape somewhat.
Despite repeated recommendations and attempts to regulate hedge funds dating
to the 1960s, the defining characteristic of hedge funds was until recently: “pooled
investment vehicles that are not registered under federal securities laws.” To achieve
this combination of pooling and avoidance of registration, hedge funds are organized
as limited partnerships or, in some jurisdictions, limit liability companies with shares
not publicly traded (van Berkel 2008). This approach emphasizes that hedge funds
are designed to avoid registration restrictions and reporting requirements imposed
on tradable securities. Over time, the growth of the hedge fund sector has made this
definition too narrow. In order to access public capital markets, some hedge funds have
undergone registration directly, or indirectly as part of a fund of hedge funds. By loosely
classifying hedge funds as a type of “private fund,” the SEC and CFTC are aiming to
cast a wide net in order to facilitate reporting needed to identify possible systemic risks
such funds may pose. The objective is not to provide a precise analytical description of
“hedge fund” characteristics.
Much is made in academic studies of the different hedge fund categories and that
“hedge fund investment strategies provide greater diversification opportunities and
may result in higher risk-adjusted returns for investors” (Edwards 2006, p.46). Some
even claim: “the hedge fund industry may have played more of a role in creating
liquidity and making markets efficient than the mutual fund industry” (Stulz 2007,
p.193). On balance, Stulz (2007) captures the “bullish” stance of academics on hedge
funds: “regulation should leave alone financial innovators who dream of new strategies
and find savvy well-funded investors to bet on them.” Prior to the market downturn
of 2008–09, there was even considerable progress toward retailization of “alternative
asset classes” such as hedge funds and private equity funds because such funds “can
pursue investment and speculative strategies that are not open to other institutional
fund managers, . . . avoid the costs associated with regulatory oversight, and . . . use
whatever fee structure they believe to be optimal” (Edwards 1999, p.191).
Viewed as a type of managed fund, the characteristics of classical hedge funds are:
actively managed; leveraged; regulatory free rider; and de facto investment companies
disguised as limited partnerships. Though a hedge fund does not directly issue securities,
because fund size changes with redemptions and additional investments, hedge funds
can also be classified as open-ended funds with restrictions on redemptions. In any case,
hedge funds possess essential characteristics of the types of managed funds that the ICA
and IAA were designed to stamp out. There are sound historically based rationales for
restricting highly leveraged speculative trading activities by unregulated entities. The
costs associated with regulatory oversight are important to maintaining the stability
and integrity of financial markets. Free rider funds that are able to avoid such regulatory
costs are at an advantage to funds that do pay such costs. From a historical perspective,
permitting unregulated and possibly highly leveraged financial entities that operate
in securities and commodity markets with the sole objective of making speculative
profits is ill-conceived and reckless, and results in increased potential for severe market
disruption that outweigh the potential benefits of increased market liquidity.
funds is the types of strategy the funds pursue. Given the restricted scope of other types
of funds seeking exemptions, hedge funds can exhibit considerable variation in strate-
gies. “There is no single market strategy or approach pursued by hedge funds as a group.
Rather, hedge funds exhibit a wide variety of investment types, some of which use highly
quantitative techniques while others employ more subjective factors” (PWGFM 1999).
Within each of these general group, a variety of different strategies could be pursued.
Similarly, some funds may be involved in activities covering more than one fund
category.
The diversity of hedge fund strategies extends to the types of securities traded
(PWGFM, p.9):
Many hedge funds trade equity or fixed income securities, taking either long or
short positions, or sometimes both simultaneously. A large number of funds also
use exchange-traded futures contracts or over-the-counter derivatives, to hedge
their portfolios, to exploit market inefficiencies, or to take outright positions. Still
others are active participants in foreign exchange markets. In general, hedge funds
are more active users of derivatives and of short positions than are mutual funds
and many other classes of asset managers.
However, behind all the confusion about hedge fund typology, some basic intuition
is relatively clear: hedge funds often combine long positions is certain securities with
short positions in other securities. Such “hedging” strategies can be relatively low risk
where the securities being traded are highly correlated and short and long positions are
balanced, e.g. the “on-the-run” “off-the-run” Treasury security arbitrage run by John
Merriweather, first at Salomon Brothers and subsequently at LTCM. Because the price
differences involved in achieving a profit are small, substantial leverage is required and
warranted. Such hedge fund strategies will, directly or indirectly, involve leveraging.
However, many other hedge fund strategies do not have sufficient correspondence
between the short and long positions to warrant the degree of leverage that is being
partially hidden from public view by the managed funds operating under exemptions
from securities laws designed to deter such excessive leveraging.
Hedge funds are not conventional investment vehicles. Investor liquidity is often
compromised with “lock-up periods of one year for initial investors and subsequent
restrictions on withdrawals to quarterly intervals” (Ackermann et al. 1999, p.834).
The regulatory exemptions that hedge funds work under severely restricts the ability
of hedge funds to advertise though the barrier to the public markets is increasingly
porous.4 Another atypical feature of hedge funds concern the management (Ackermann
et al. 1999):
In contrast to mutual funds which have a much longer history that has been intensively
studied, hedge funds only started to receive academic attention in the mid 1990s, though
work on managed futures funds and commodity pools, which started somewhat earlier,
is also applicable, e.g. Cornew (1988); Edwards and Ma (1988); Edwards and Park
(1996); Elton, Gruber and Rentzler (1987); Irwin and Brorsen (1985); Irwin et al. (1993).
As data have accumulated on hedge fund activities, a voluminous number of studies
has appeared on various aspects of hedge funds. Among the useful studies directly on
hedge funds are: Ackermann et al. (1999); Brown et al. (1999, 2001); Fung and Hseih
(2000, 2002); Goetzmann et al. (2003); Gregoriou (2002); Griffen and Xu (2009); Klein
and Lederman (1995); Liang (2000); Patton (2009); and Schneweiss and Spurgin (1998).
[Amaranth’s energy book] was up for the year roughly $2 billion by April, scoring a
return of 11 percent to 13 percent that month alone, say investors in the Amaranth
fund. Then . . . [the energy strategies] had a loss of nearly $1 billion in May when
prices of gas for delivery far in the future suddenly collapsed, investors added. [The
energy traders then] won back the $1 billion over the summer.
As it turns out, Amaranth’s energy book was primarily concentrated in natural gas
and had benefited substantially from natural gas price movements associated with
Hurricane Katrina in the latter part of 2005: “the double-whammy of Hurricanes Katrina
and Rita made Mr. Hunter a hero at Amaranth and a minor legend on Wall Street, as
he made $1 billion for Amaranth” (A. Davis 2006). The extent of the trading is revealed
in US Senate (2007) and Till (2008). At times during 2006, Amaranth controlled as
much as 40 percent of all the open interest on NYMEX for the winter month contracts
delivering in October 2006 through March 2007 (US Senate 2007, pp.51–2). In late
July 2006, Amaranth held a total of more than 80,000 NYMEX and ICE contracts for
January 2007, representing a volume of natural gas that equaled the entire amount of
natural gas eventually used in that month by US residential consumers nationwide
(US Senate 2007, p.52). On July 31, 2006, Amaranth’s trading in the March and April
2007 contracts represented almost 70 percent of the total NYMEX trading volume
in each of these contracts on that date. Amaranth held large positions in winter and
summer months spanning the entire five-year period from 2006–2010, e.g. Amaranth
held 60 percent of the outstanding contracts (open interest) in all NYMEX natural gas
futures contracts for 2010 (p.52). On 7/4/06, Amaranth’s futures position as a percentage
of NYMEX futures open interest in the December 2007 contract was 81 percent. On
August 28, 2006, Amaranth accounted for over 40 percent of the total volume on the
ICE and over 25 percent of the entire volume of exchange-traded futures and swaps on
NYMEX and on ICE on that date.
Even in the aftermath of the debacle, it is not possible to precisely determine the
trading strategies employed. However, evidence on Amaranth’s positions provided
in US Senate (2007) is sufficient to determine the general picture (see Fig. 1.31). In
addition, the size and timing of the losses is also known. When margin calls on the
position crossed US$3 billion, around September 2006, the fund offloaded some of
these positions, ultimately selling much of the book to JP Morgan and Citadel for US$
2.5 billion, with this amount to be credited against the losses. The fund ultimately took
a $6.6 billion dollar loss and had to be dissolved entirely. Oddly enough, Maounis said
in an interview in August 2006: “What Brian is really, really good at is taking controlled
and measured risk” (Davis 2006). Judging from the size and type of positions, those
involved in overseeing the various strategies employed by the fund were not employing
sufficient risk management tools to accurately manage the risks that were being taken
up. In addition, subsequent prosecution of Amaranth by both FERC and the CFTC
revealed illegality by the principals. Once the position began to “blow up,” Amaranth
sought to cover NYMEX losses by substantially increasing position size in the OTC
market on ICE, knowingly violating reporting requirements for speculators.
Combining the information in Till (2008) with US Senate (2007), it is apparent that
Amaranth was short shoulder season contract deliveries—mainly April and October
2006—and long winter deliveries—November, January and March 2006. There also
appears to be spreading across years as well with a short position in December, 2006,
approximately balanced by a December, 2007, long position. Till (2008) used the
movements of natural gas spreads in September to confirm that these two spread trades
could produce the losses reported. September 2006 was an unusual month historically
for the term structure of futures prices with a contango from summer to winter, and a
backwardation from March to April.5 According to publicly available information, the
fund lost $560 million on September 14, shrinking the size of the fund about 35 percent
for the week, about $3.2 billion based on a $9.2 billion net asset value. At the same time,
the March/April spread on September 14 fell by $6000 per spread, and in the whole
week the spread dropped by $30,650 per spread combination. On the other hand, the
spread of the short summer/long winter combination decreased by $3720 on September
14 and was down by $48,950 per spread for the whole week.
Evidence of the sophistication of the trading strategy employed by Hunter is
provided by Till (2006), which demonstrates that, from June to August 2006, the
same type of trades generated $1.2 billion profits, close to the $1.3 billion reported
by Amaranth. Even though the position sizes throughout the months from May
to September were not the same, the derived position sizes were similar to what
Amaranth’s energy portfolio was holding in summer 2006. In effect, Amaranth
was taking the same intrinsic view about the seasonal calendar spread in natural
gas, profiting from events such as exceptionally cold winters or summers with big
hurricanes. Collins (2006) quotes Erk Hinrichsen, senior managing director of Energy
Arbitrage Management: “It looks like they put the same position on over and over.”
The short December, 2006–long December, 2007 spread seems particularly misguided
due to supply considerations that were readily apparent in the natural gas market
at that time. However, speculative traders often do take incorrect views and the key
risk management failure was identified by Till (2006): “These strategies were and
are economically defensible, but the scale of their position-sizing relative to their capital
base clearly was not.”
Till (2006, p.97) also makes a useful observation:
it is likely that physical natural gas market participants were the ultimate risk
takers on the other side of Amaranth’s trades, and so benefited from the temporary
dislocations that ensued from the fund’s distress. In other words, it does not appear
that the commercial natural gas industry was damaged by this financial crisis; in
fact, commercial suppliers of natural gas likely benefited. Natural gas commercial
hedgers would have earned substantial profits had they elected to realize their
hedging windfall during the three months that followed the Amaranth debacle.
The large position size established by Amaranth was unsound relative to the relatively
illiquid market for natural gas futures (see Table 1.9). Unlike liquid financial futures
contracts for the S&P 500 index or Eurodollar deposits, activity in natural gas futures
relies on commercial participants initiating trades for hedging purposes. Commodity
futures traders in relatively illiquid markets recognize a major part of a trading strategy
is how to exit positions when necessary. The size of Amaranth’s trades were so large
there was no counter-parties capable of providing sufficient liquidity for Amaranth to
exit positions in a week or two. The counter-parties to many of Amaranth’s trades were
physical-market participants that had locked in attractive value for production and
storage. There was little economic incentive to provide the liquidity for Amaranth to
unwind trades.
The final aspect of the Amaranth debacle concerns the disposition of the main
players, Hunter and Donohoe. Actions were initiated by both the CFTC, for attempted
manipulation of NYMEX contracts, and FERC, for actual manipulation of the natural
gas market. The enforcement action in July 2007 was the first major initiative by FERC
under the expanded powers granted by the Energy Policy Act (2005). In particular, Till
(2008, 95):
In August 2009, in an action coordinated with the CFTC, FERC imposed a $7.5 million
fine on Amaranth and Donohoe. In addition to stifling attempts by Hunter to relaunch
as a commodity hedge fund manager, in April 2010 a FERC administrative judge handed
down a $30 million civil fine against Hunter for violating anti-manipulation rules. As
for the scary prospect of Hunter continuing as a hedge fund manager, Christopher Holt,
founder of hedge fund consultant Holt Capital Advisors Ltd, observed: “As long as you
can explain what happened, why it happened, and why it won’t happen again, there’s no
reason that a top trader couldn’t attract new backers” (Willis 2007).
. . . human decisions affecting the future, whether personal or political or economic,
cannot depend on strict mathematical expectation, since the basis for making such
expectations does not exist.
John Maynard Keynes (The General Theory, Ch. 12)
129
It is sometimes said that hedging is the opposite of speculation. This is not so.
They are different kinds of the same thing. The thing that is usually identified as
speculation—that is, long or short positions in futures contracts—is speculation
in price level. The thing that we identify as hedging—that is, long cash and short
futures or vice versa—is speculation in price relationships . . . Thus hedging and
speculation are not opposite; in fact, they are conceptually similar. They are just
different kinds of speculation.
Thomas Hieronymous (1977)
131
The evolution of methods for the identification, assessment and management of risk
have played a central role in the progress of civilization. In ancient times, religious
beliefs were important in reconciling the risks confronting a society. Appeals to the
gods by the priesthood, prophecies from the oracle, chanting by the shaman were all
methods of passively dealing with risks encountered. The development of scientific,
mathematical and probabilistic methods during the Enlightenment permitted risk to be
more actively identified and assessed. This advancement encountered a philosophical
quandary concerning subjective and objective interpretations of probability. More
precisely, the objective interpretation views probability as inherent in nature. Logic,
scientific investigation and statistical analysis can be used to discover objective
probabilities. In contrast, subjective probabilities quantify an individual’s belief in the
truth of a proposition or the occurrence of an event, and are at least partially revealed
in an individual’s choice behavior. Such probabilities can vary across individuals due,
say, to differing degrees of ignorance about the event of interest.
Debate over subjective versus objective probability reached a peak around the time
that Frank Knight (1885–1972) introduced a distinction between risk—where the
objective probability of an event is at least measurable—and uncertainty—where the
probability is not knowable and has to be determined subjectively. This terminological
distinction between risk and uncertainty has now faded from common usage as the
subjectivist approach has gained prominence, supported by seminal contributions from
Frank Ramsey (1903–30), Bruno di Finetti (1906–85) and Leonard Savage (1917–71).
Attention has shifted to whether subjective beliefs derive from intuition or are only
realized in choice behavior. The intuitive approach leads to a focus on the perception of
risk, a concept often employed in psychometric and sociological research. Development
of the choice-theoretic approach to subjective probability was facilitated by the expected
utility function introduced by John von Neumann (1903–57) and Oskar Morgenstern
(1902–76) in a classic work of social science, The Theory of Games and Economic Behavior
(1944). The choice-theoretic approach has sustained the modeling of decision making
under uncertainty that is a central component of modern economic theory, in general,
and financial economics, in particular.
Prior to von Neumann and Morgenstern, neoclassical economic theory was based
on certainty or perfect foresight, though consideration of risk in decision making
could be found in the less formal approaches of Frank Knight, John Maynard Keynes
(1883–1946) and Irving Fisher (1867–1947). These individuals contributed to a range
of future contributions and perspectives on the impacts of risk in economics. Precisely
how to model predictions for random variable outcomes, e.g. using the conditional
distribution, raises deep philosophical questions, variants of which have been debated
for centuries. For example, Thomas Bayes (1701–61) suggested that the conditional
(posterior) distribution is determined by combining prior beliefs with available
empirical evidence. In the 20th century, both Keynes and Knight advanced the notion
that the variation in future outcomes is a combination of a measurable component, risk,
and an unmeasurable component, uncertainty. At the time, this was an intellectual
sufficient number to make it possible to tabulate enough like it to form a basis for
any inference of value about any real probability in the case we are interested in.
In effect, Knight is saying that in many instances involving commercial risks there
is insufficient empirical information to accurately form the subjective probabilities
needed to implement the expected utility approach. Risk is associated with objectively
measured probabilities, while uncertainty requires subjective probability assessments
that are difficult to determine in important practical situations. The economic rents
to business ownership or, for that matter, commodity risk management arise from
correctly anticipating uncertain outcomes.
As for methods of dealing with uncertainty, Knight (p.239) recognizes four general
approaches:
We may call the two fundamental methods of dealing with uncertainty, based
respectively upon reduction by grouping and upon selection of men to “bear” it,
“consolidation” and “specialization,” respectively. To these two methods we must
add two others . . . (3) control of the future, (4) increased power of prediction.
Knight recognizes the complementarity among the different approaches for dealing
with uncertainty. For example, increased specialization permits more firm resources
to be devoted to data collection and analysis, which increases power of prediction.
Writing in 1921, Knight has little to say about the use of derivative securities to “control
the future.” Other than occasional references, Knight also does not deal with specific
aspects of financial risk and uncertainty. What Knight does say very clearly is that
the randomness associated with commercial decisions, such as strategic business risk,
is composed of “risk,” which is measurable in an objective sense, and “uncertainty,”
which is measurable only subjectively. It is in dealing correctly with uncertainty that
“entrepreneurs” earn value.
Keynes and Fisher represent two alternative approaches to uncertainty. By expli
citly recognizing the “caution coefficient” that measures the difference between the
mathematical expectation and the price that will be paid for a gamble, Irving Fisher laid
the foundation for later contributions in mean-variance portfolio theory. This approach
assumes sufficient empirical information for subjective probabilities to be specified. In
contrast, the numerous contributions by Keynes on the impact of uncertainty range
from the Treatise on Probability (1921) to the General Theory of Employment, Interest
and Money (1936). As the chapter-opening quote identifies, Keynes followed Knight in
maintaining there was insufficient empirical information to determine the subjective
probabilities needed to determine the mathematical expectations associated with
means and variances. Disciples of Keynes, such as George L.S. Shackle (1903–92),
argue against the use of probability theory to model decision making under uncert
ainty. Poitras (2011, Ch.5) demonstrates the failings of the ergodicity assumption in
applications of probabilistic notions to model financial decisions.
Given the diverse approaches to risk generated by Knight, Keynes and Fisher, the
application of expected utility in economic theory requires the use of preference
orderings over state contingent commodities. Kenneth Arrow (b. 1921) and Gerard
Debreu (1921–2004) extended the neoclassical economics of Stanley Jevons (1835–82),
Leon Walras (1843–1910) and Alfred Marshall (1842–1924) to include decision making
under uncertainty. This development follows naturally from using the choice-theoretic
approach to subjective probability developed by von Neumann and Morgenstern.
The utility of a certain outcome is replaced by the expected utility, calculated using
known probabilities and the associated utilities for a set of random outcomes. The
known probabilities can be notionally determined by direct observation of previous
choice behavior. Using this approach, there is no formal distinction between risk
and uncertainty, e.g. Penello (2009). Risk is associated with the variability of random
outcomes and uncertainty with randomness. Sensitivity to risk is measured by com
paring a certain outcome to a random outcome with the same expected value. Risky
outcomes are measured in income, dollars or returns, and can take both positive and
negative values.
In financial economics, the expected utility framework has been applied to
determining solutions to: the optimal combination of individual securities in a
portfolio of securities; and, as discussed in Sec. 2.2, the optimal hedge ratio to use in
risk management decisions. The initial application involved using an expected utility
function specified over the expected (portfolio) return and variance of (portfolio) return.
Harry Markowitz (b. 1927) and William Sharpe (b. 1934) were able to demonstrate that
the variability or risk of a portfolio can be further divided into two components: firm-
specific risk, which is diversifiable and non-systematic; and market-related risk, which is
systematic and not diversifiable. Applying this to the tradeoff between risk and return,
it is demonstrated that only increases in the systematic risk of an individual security
will be rewarded with higher expected return. Hence, it is only that portion of the total
variability of a security’s return that cannot be diversified away that warrants higher
expected return. A measure of systematic risk—the beta of a security—is provided.
Beta can be calculated as the slope coefficient in a least squares regression of individual
security return on market return: the ratio of the covariance between the individual
security return and the market return divided by the variance of the market return.
More recently, a variety of risk measures have been developed to deal with perceived
limitations of variance of return and beta in dealing with downside risk. These new
measures include: value at risk; expected regret; conditional value at risk; and expected
shortfall. In many applications, losses have a different impact than gains, if only
because losses may increase the risk of bankruptcy or lead to the delay of investment
projects. Risk measures such as variance, standard deviation or beta give similar weight
to upside and downside movements. “New” risk measures such as value at risk and
expected shortfall are concerned only with the properties of downside risk. While this
enhances the identification of the potential impact of downside risk, the connection
with overall firm profitability is obscured. The raises the well-known quandary facing
risk management: whether risk associated with losses or overall profitability is the
ultimate objective.
fact and value appears as a distinction between “positive economics” and “normative
economics” (p.4):
Much of Friedman (1953) is concerned with the issue of whether a theory with unreali
stic assumptions, even “wildly inaccurate descriptive representations of reality,” can be
“important and significant.” For Friedman, the ultimate test of a theory was “whether
it yields sufficiently accurate predictions,” not whether the assumptions are realistic.
The concern of Friedman (1953) with the form of the theory being examined is
consistent with the evolution of positivist epistemology. Initially, positivism placed
heavy reliance on the inductive process of collecting facts. Spurred by the remarkable
successes of the natural sciences during the late 19th and early 20th centuries, this view
evolved into logical positivism, an epistemology that placed emphasis on theories and
the logical deduction of hypotheses to test those theories as well as the collection of
facts. The epistemology of logical positivism allows only two grounds for truth: there
are deductive truths such as those in mathematics and formal logic, e.g. 12 – 3 = 9; and
inductive statements that match reality precisely. As a consequence, truthful statements
have to be verifiable in order to be meaningful. In logical positivism, statements have
meaning relative to the conditions under which the statement can be verified. Friedman
adapts this approach to where the test of verification for a hypothesis is the ability to
predict. This is consistent with the tenet of logical positivism that a statement that does
not describe an “experiential proposition” carries no significance, i.e. it is not knowledge.
Friedman (1953, p.7) clearly reflects these tenets of logical positivism in what Boland
(1979, 1991) has termed economic positivism: “theory has no substantive content; it
is a set of tautologies . . . Factual evidence alone can show whether the categories of
the ‘analytical filing system’ have a meaningful empirical counterpart, that is, whether
they are useful in analyzing a particular class of concrete problems.” Statements that
are verifiable provide a basis for building a science. Under positivism, science is the
source of knowledge. As such, both positivism, in general, and economic positivism,
in particular, share a fundamental commitment to empiricism, an epistemology
where claims that have no empirical consequences are without meaning. Economic
positivism extends empiricism by arguing that science can also seek to build
theories to describe the regularities of cause and effect in order to explain the world.
This requires theories to be expressed as a set of axioms or, less formally, basic assum
ptions. These theories have rules to systematically link the predictions with objective
measurements of the real world. The connection to Friedman (1953), von Neumann
and Morgenstern (1947) and innumerable other projects in positivist economics
is apparent.
This discussion begs the question: so what is wrong with economic positivism? There
are a number of different answers. At this point, it is relevant to observe that positivism
maintains that science is the only way to create knowledge and allow individuals to
understand the world well enough to predict and control outcomes. In the positivist
framework, the objective world is viewed as deterministic, operated by laws of cause and
effect that can be identified if the unique approach of the scientific method is applied
correctly. Science is conceived as a mechanistic operation. It is possible to use deductive
reasoning to postulate theories that can be empirically tested. Based on the results
of these empirical tests, it is determined whether a theory “fits the facts” or whether the
theory needs to be revised in order to provide better predictions of reality. Ultimately,
there is an objective reality that can be discovered if there is sufficient empirical
information available to verify the “true” deductive hypotheses.
Criticisms of economic positivism are numerous. One type of criticism focuses on the
misunderstanding of the process by which science is conducted. Is there really a unity
of science? Are the procedures used in physics and chemistry directly applicable to eco-
nomics or psychology? Do scientists really develop deductive hypotheses that are then
“verified” on empirical data? Another related criticism observes that economic positiv-
ism says little or nothing about how axioms (or Friedman’s assumptions) are translated
into possible testable hypotheses. In other words, positivism has no substantive insight
into the process by which knowledge is created. Positivism is only interested in specify-
ing the scientific process, without recommending criteria for selecting among permit-
ted ideas. This leads to Friedman (1953) and the criteria of predictive ability. But, this
leads to the problem of measuring predictive ability. The distinction between ex ante
and ex post predictability is one key example of this type of problem in modern finance.
Positivism proposes that there is a unity of science. Certain developments in
epistemology after positivism deny this proposition. As such, schools of thought have
emerged that are concerned specifically with the epistemological problems arising in
the human sciences. One such epistemology is “critical realism,” a school that observes
all measurement is fallible is some way, e.g. Bhaskar (1978), Lawson (1997). For example,
critical realists maintain that all observations are theory-laden and that individuals in
general, and scientists in particular, are inherently biased by their cultural experiences,
world-views, and so on. Friedman (1953, pp.4–5) recognizes this issue but does not
view it as a basis for “a fundamental distinction” between economics and the natural
sciences. For critical realists the challenge is how to move from a notion of objectivity
that is inherently a social phenomenon to the identification of knowledge. If objectivity
is not perfect, then how are these separate and imperfect individual interpretations of
reality to be combined?
Friedman (1953) provides a window to the 20th-century development of the
philosophy of the social sciences. In this development, words like “hermeneutics” and
“ontology” are essential to the discussion, though references to notions such as “the
questionableness of romantic hermeneutics” require knowledge of the philosophical
developments to be correctly interpreted. Hermeneutics has a long history in
philosophy, starting with problems of biblical exegesis. During the 18th and early 19th
centuries, hermeneutics evolved into a more general theory of textual interpretation,
aiming to provide a set of rules for accurate interpretive practice applying to a wide
range of subject matter. Taking hermeneutics as the relevant method for the recovery
of meaning, Wilhelm Dilthey (1833–1911) broadened hermeneutics to represent a
methodology for the recovery of meaning that is central to understanding knowledge
within the “human” or “historical” sciences.
Strongly influenced by Martin Heidegger (1889–1976), Hans-Georg Gadamer (1900–
2002) is “the decisive figure in the development of twentieth century hermeneutics”
(Stanford Encyclopedia of Philosophy). Gadamer is part of a long line of thought that
questions the ability to apply techniques of the natural sciences to the human sciences,
e.g. (p.6): “the real problem that the human sciences present to thought is that one has
not properly grasped the nature of the human sciences if one measures them by the
yardstick of an increasing knowledge of regularity. The experience of the socio-historical
world cannot be raised to a science by inductive procedure of the natural sciences.”
Though Gadamer’s notion of the human sciences may seem to have more applicability
to, say, political science or sociology, it is difficult to evade the observation that the
prices of commodities and common stock are set in markets, and are the outcome of
a social interaction. In turn, commodity risk management practices form an integral
part of the process by which commercial firms are operated and the securities of those
firms are valued.
Unlike the natural sciences, the human sciences have to allow for prejudice derived
from authority. In contrast, methodologically disciplined use of reason cannot
accept arguments based on authority for that involves not using one’s reason to reach
conclusions: “If the prestige of authority takes the place of one’s own judgment, then
authority is in fact a source of prejudices.” But the approach toward the human sciences
proposed by Gadamer (1960, p.249) does not view prejudice either negatively or
positively. As such, authority as a positive prejudice provides a basis for knowledge:
the recognition of authority is always connected with the idea that what authority
states is not irrational or arbitrary, but can be seen, in principle, to be true. This is
the essence of the authority claimed by the teacher, the superior, the expert. The
prejudices that they implant are legitimized by the person who presents them. But
this makes them then, in a sense objective prejudices, for they bring about the same
bias in favor of something that can come about through other means. e.g. through
solid ground offered by reason.
time, the same is not true about the human sciences where great achievements of the
past “hardly ever grow old.”
For Gadamer, the interpreter is an essential component of knowledge in the human
sciences: “the object appears truly significant only in the light of him who is able to
describe it to us properly. Thus it is certainly the subject that we are interested in, but the
subject acquires its life only from the light in which it is presented to us.” Subjects appear
historically “under different aspects at different times or from a different standpoints”
(p.252). Insightful interpretations require the past to be echoed in the present. As such,
the human sciences are involved not only in the accumulation of empirical results but
in the transmission of an important source of authority: tradition. “That which has
been sanctioned by tradition and custom has an authority that is nameless, and our
finite historical being is marked by the fact that always the authority of what has been
transmitted—and not only what is clearly grounded—has power over our attitudes and
behavior” (p.249).
Gadamer sees an essential role for tradition in the human sciences (pp.251–2):
“That there is an element of tradition active in the human sciences, despite the
methodological nature of its procedures, an element that constitutes its real nature,
and is its distinguishing mark, is immediately clear if we examine the history of
research and note the difference between the human and natural sciences with regard
to their history.” For Gadamer: “the natural scientist writes the history of his subject
in terms of the present stage of knowledge. For him errors and wrong turnings are
of historical interest only, because the progress of research is the self-evident criterion
of his study . . . the human sciences cannot be described adequately in terms of
this idea of research and progress.” Knowledge in the human sciences does not proceed
by distancing and freeing ourselves from what has been transmitted through tradition.
Rather, the problem is to find the relationship of the present with the traditions of
the past.
The positivist foundation of academic studies depends on the premise that know
ledge in the subject is obtained solely from the methodology of the natural sciences.
Somehow, increasingly greater knowledge is obtainable about the natural phenomena
of stock or commodity markets, especially commodity prices, as increasingly larger
amounts of data are examined or more precisely mathematical theories are derived. The
historical evolution of markets is unimportant. The views of writers in the past, such
as Holbrook Working or Nicolas Kaldor or J.M. Keynes, are only of historical interest,
useful illustrations of how far knowledge has progressed since that time. Gadamer,
and other philosophers of his ilk, would argue that this approach is predicated on the
supposition that the subjects of finance, in general, and commodity risk management, in
particular, can be a natural science. However, the objects of interest in commodity risk
management, especially commodity prices, are the result of human interactions and,
as such, belong in the realm of the human sciences. If correct, knowledge of the subject
could be substantively increased by proceeding beyond the scientific search for sources
of homogeneity in commodity risk management across firms to incorporate the notions
Under Item 305 of SEC Regulation S-K, “Quantitative and Qualitative Disclosures
about Market Risk,” companies must disclose in their 10-Ks information about their
exposures to fluctuations in variables such as interest rates, foreign exchange, and
commodity prices. While disclosure is mandatory, companies have the discretion
to choose among three alternative methods: sensitivity analysis, value-at-risk (VaR),
and the so-called “tabular” method.
The increased attention to risk management has led many firms to reform the process
by which risk management is integrated into the hierarchy of managerial control. The
VaR technique is, in and of itself, not much different than risk management techniques
that have been used for many years by more sophisticated firms. The VaR revolution
is associated more with the system-wide adoption of these techniques by depository
institutions and other financial intermediaries. However, it is not clear this system-
wide introduction of VaR has resulted in a corresponding reduction in systemic risk in
financial markets.
On balance, the VaR revolution has been profound for financial firms, e.g. Jorion
(2006). Non-financial firms pose a somewhat different risk management problem
(Oxelheim and Wihlborg 1997, p.21):
For a non-financial firm the primary risk would be its commercial risk—i.e. its
uncertainty about the value of cash flows that can be generated by its physical assets
producing output. Its liquidity risks are secondary in the sense that they merely
enhance or modify the primary risk. The importance of a specific kind of risk can
shift depending upon the situation.
As such, the value at risk (VaR) revolution is somewhat narrowly confined to financial
firms, especially firms making markets in derivative securities and other leveraged
instruments such as bond portfolios financed using repurchase agreements. Yet, VaR
can be of importance for non-financial firms, particularly multinational firms, seeking
to assess and control the financial risk which is associated with activities such as
currency and interest rate risk management.
Wilmott (1998, p.547) provides a useful definition for value at risk (VaR):
Value at risk is an estimate, with a given degree of confidence, of how much one can
lose from one’s portfolio over a given time horizon.
The VaR calculation is aimed at making a statement of the following form: “We
are X percent certain that we will not lose more than V dollars in the next N days.”
The variable V is the VaR of the portfolio. It is a function of two parameters: N, the
time horizon, and X, the confidence level. One attractive feature of VaR is that it
is easy to understand. In essence, it asks the simple question: “How bad can things
get?” In calculating a bank’s capital, regulators use N = 10 and X = 99. They are,
therefore, considering losses over a 10-day period that are expected to happen
only one percent of the time. The required capital for market risk is, at the time of
writing, three times the 10-day 99% VaR.
Though the VaR methodology can conceptually be applied to a wide range of situations,
applications have focused on situations involving market risk: the potential for changes
in the value of a position resulting from changes in market prices.
An important initial impetus to the spread of VaR was the widespread availability
of software and technical material to support the implementation. In particular, the
RiskMetrics group that was formerly associated with JP Morgan/Reuters was an
important early promoter of the VaR methodology providing detailed technical
publications, such as the RiskMetrics manual (JP Morgan 1996), and daily data sets for
important financial variables that were free of charge on the JP Morgan web page. The
RiskMetrics manual describes a set of methodologies outlining how risk managers can
compute VaR on a portfolio of financial instruments where commodities are included
as an “asset class.” RiskMetrics paid close attention to modeling the VaR for positions
containing options. Nonlinear payoffs associated with options can pose problems
for the VaR methodology. The RiskMetrics group was first established in 1989 when
Sir Dennis Weatherstone was chairman of JP Morgan. In 1993, JP Morgan publicly
launched the RiskMetrics methodology. In January 1998, the RiskMetrics Group was
spun off from JP Morgan and the RiskMetrics Group was listed on the New York Stock
Exchange (NYSE: RISK). Finally, in June 2010, RiskMetrics was acquired by Morgan
Stanley Capital International.
uranium, then an estimate for the price has to be obtained from a survey of dealers
and brokers specializing in the commodity or from an appropriate pricing model. For
financial institutions subject to BIS-style rules, the relevant pricing models have to
conform to certain requirements, e.g. Hendricks and Hirtle (1997). In-house models are
acceptable, perhaps preferable, as long as the resulting prices are, on average, accurate.
From this data, the relevant statistical parameters can be calculated using conventional
formulas appropriate for the probability distribution selected. However, the situation
is less transparent for non-financial firms where the physical “assets” involved are
typically non-traded and the impact of commodity price change on the value of the
asset is difficult to determine.
Originally introduced by the Riskmetrics group, Andren et al. (2005) give the
following description of Cash Flow at Risk (CFaR).
CFaR . . . is the cash flow equivalent of Value-at-Risk, or VaR, which is widely
used as the basis for risk management systems within financial institutions.
Whereas VaR-based systems specify the maximum amount of total value a
firm is expected to lose under most foreseeable conditions (for example, with a
95% confidence level), CFaR-based systems determine the maximum shortfall
of cash the firm is willing to tolerate (again, with a specified level of statistical
confidence). CFaR is gaining in popularity among industrial companies for
much the same reasons VaR has succeeded with financial firms: it sums up
all the company’s risk exposures in a single number that can be used to guide
corporate risk management decisions. It is this number—the maximum
shortfall given the targeted probability level—and the fact that it can be
directly compared to the firm’s risk tolerance that are the uniquely attractive
features of both VaR and CFaR.
In effect, CFaR is a sophisticated form of optimal hedge ratio calculation. Can such a
“scientific” approach to risk management be successful?
To illustrate the CFaR methodology, consider the cash flow generated by a real asset
producing a single commodity being sold at a random price. Let rt = ln [1 + Rt] where
Rt = (CFt - CFt-1)/CFt-1 and CF is the cash flow at a given time which changes due to
fluctuations in “market risk” associated with the commodity price. This transformation
to a rate of return calculation is done for statistical reasons which will be discussed
shortly. Due to commercial factors, such as the level of fixed costs, the impact of
commodity price changes on cash flow may not be linear. Define the probability density
associated with r as Φ[r]. With this density it is possible to obtain the probability that a
future value of r will take a value less than r*:
Price changes
Crude oil – WTI US$ 1. 00/bbl (1)
Excluding financial derivatives $ 128 $ 0.12 $ 99 $ 0.09
Including financial derivatives $ 128 $ 0.12 $ 99 $ 0.09
Natural gas – AECO C$0. 10/Mcf(1)
Excluding financial derivatives $ 34 $ 0.03 $ 25 $ 0.02
Including financial derivatives $ 38 $ 0.04 $ 29 $ 0.03
Volume changes
Crude oil – 10,000 bbl/d $ 175 $ 0.16 $ 104 $ 0.10
Natural gas – 10 MMcf/d $ 9 $ 0.01 $ 1 $
Foreign currency rate change
$0.01 change in US$ (1)
Including financial derivatives $ 101–103 $ 0.09 $ 40–41 $ 0.04
Interest rate change – 1% $ 9 $ 0.01 $ 9 $ 0.01
11/16/2012 6:44:09 PM
148 Commodity Risk Management Concepts
r*
Prob[rt +1 < r *] = ∫ Φ[r] dr = c
−∞
Cash flow at risk can now be determined by evaluating CFt r and (CFt r*).
For short time horizons, such as weekly CFaR, it can often be assumed that μ = 0 to
produce the result:
CFaR = ( S ) (µ − 2.33 σ )
In VaR calculations, some presentations of the calculation of the return form have an
additional time scale factor to account for differences between the time scale used to
estimate the volatility and the time horizon for the VaR, e.g. if a weekly VaR is desired
and the volatility estimate is for annual returns, scaling by (1/52)½ is required, e.g.
Hamidieh and Ensor (2010). This adjustment is unnecessary if the sampling frequency
of the data used to estimate the parameters is the same as the horizon for the VaR
forecast.
range of potential solutions has been proposed to deal with limitations of the normality
assumption. The basic idea of many approaches is, somehow, to adjust the normal
distribution to accurately reflect the true tail density.
The problem of empirically fitting a distribution to a time series of past data that
is useful for predicting future values of the time series is not without difficulties but
may be manageable in many situations. A common statistical approach to fitting non-
normal data is to use a series approximation, such as an Cornish-Fisher or Edgeworth
expansion, to model the true distribution. This would typically result in higher moments,
such as skewness and kurtosis, being estimated and used to adjust the tail densities, e.g.
Erb and Harvey (2006). Regarding the use of a Cornish-Fisher expansion, Fuss et al.
(2010, p.263) observe:
Although the Cornish-Fisher (CF) VaR adjusts the critical values of the standard
normal distribution, neither normal VaR nor CF-VaR react sufficiently to changes
in the return process, which can be problematic for forward-looking investment
decisions.
the underlying market price). And here again the bottom-up approach cannot
provide such a measure.
[Stein et al. (2001)] instead apply a “top-down” approach in which the focus
is on overall cash flow volatility. In place of bottom-up estimates based on
a company’s historical data and line managers’ projections, the top-down
method pools cash flow data for a large number of comparable companies to
estimate a pooled cash flow distribution. The advantage of such an approach
is its ability to provide a historical average exposure estimate that reflects the
collective experience of many firms under a variety of market conditions.
Bur this approach also has an obvious limitation in that the firm in question
could be very different from the “average” company in the sample. Moreover,
the top-down approach does not provide an estimate of CFaR conditional on
market risk, nor can it be easily adapted to do so.
In particular, Fuss et al. (2010) demonstrate that unconditional VaR estimates based
on normal distributions for commodity price changes are improved using estimators
that admit non-normality such as the GARCH-VaR, e.g. Giot and Laurent (2003), and the
conditional auto-regressive value at risk, e.g. Engle and Manganelli (2004). Kuester et al.
(2006) provide a survey and comparison of many of the available methods of adjustment.
Yet, while the problems associated with the normality assumption in determining VaR
estimates for financial applications are in the realm of the theoretically rectifiable,
determination of CFaR for non-financial firms is decidedly more complicated because
the CFaR approach is concerned with cash flows being generated by the firm and not
with value changes resulting from changes in financial prices. Even if the sources of
“market risk” could be meaningfully identified, the time series of past cash flows for an
individual firm are insufficient to generate a meaningful “bottom-up” distribution for
use in CFaR measurement. Similarly, the heterogeneity of firms prevents a meaningful
“top-down” distribution from being identified.
A final theoretical concern with CFaR involves a well-known problem for calculating
VaR for a portfolio: many securities traded in financial markets, such as options, have
nonlinear payoffs, e.g. Jorion (2000). Similarly, many real assets also contain various
types of option that can have a nonlinear impact on firm valuation. Nonlinear payoffs
substantially complicate the problem of dealing with non-normality, e.g. Hull and White
(1998). To address these problems in VaR models, the alternative delta and delta-gamma
approaches have been proposed, e.g. Byers (2005), Duffie and Pan (1997), Riskmetrics
(1996). Much like the use of duration and convexity in fixed income analysis, these
approaches use a Taylor series expansion (see Poitras 2005) to approximate the nonlinear
payoff function for an instantaneous change in the random variable. In fixed income
analysis, the payoff function is usually assumed to be convex, to reflect the inverse
relationship between price and yield. Such a convexity assumption may not be valid for
situations involving real options impacting the cash flows of a non-financial firm.
ρ A + B ≤ ρ [ A ] + ρ [ B]
ρ A + B + C ≤ ρ [ A ] + ρ [ B] + ρ [C ]
Consider all the possible non-Z outcomes for a “portfolio” company that combines the three
real assets:
Observing that whenever Z occurs, the change from the initial value will be negative,
from these calculations it follows that .729 + .23085 + .02436 + .0008574 = .9805 is
the total probability associated with non-Z events. All value changes below a .9805
probability have a VaR > 0 (Remember VaR is a positive number when there is a loss).
Therefore, VaR is not sub-additive at the 99 percent level.
ρ A + B ≤ ρ [ A ] + ρ [ B]
A super-additive risk measure satisfies:
ρ A + B ≥ ρ [ A ] + ρ [ B]
and so on for larger collections of asset cash flows. When only equality holds, the
risk measure is additive. The variance (var[A]) of two random variables with positive
covariance (cov[A, B] > 0) is a super-additive risk measure:
sd [ A + B ] ≤ sd [ A] + sd [ B ]
To see this, consider the case where sd[A] = sd[B] = 1 and cov[A,B] = .5. In this case:
1 + 1 + 2 ( .5 ) = 3 ≤ 1 + 1
Only in the perfectly, positively correlated case will equality hold. Unlike VaR, the
standard deviation is a two-sided risk measure.
Types of Basis
In market terminology, a basis refers to the difference between two prices. The study
of basis relationships is fundamental to understanding cash markets as well as futures
and forward markets. Various types of basis relationships are of interest. One basis
relationship that is of theoretical interest is the maturity basis, the difference between
the delivery date price of a futures contract and the corresponding spot price. It is
often theoretically convenient to assume that the maturity basis is zero, implying that
F(T,T)=S(T). Though the maturity basis is almost always zero for forward contracts
that are written with delivery in mind, for the futures maturity basis to be zero, it is
necessary but not sufficient that the spot and futures prices both refer to the deliverable
commodity. When the maturity basis for a deliverable spot commodity deviates from
zero, a profit opportunity is provided for delivery arbitragers operating on the futures
exchange. Lack of convergence in the futures basis in the wheat futures market was a
central concern in US Senate (2009) (see Fig. 2.2).
Evaluation of the maturity basis when the spot commodity is not the same as that
deliverable on the futures contract is often complicated by the grade and location
characteristics of the spot commodity. Even for futures contracts, the deliverable
standard grade specified in the futures contract often permits multiple delivery grades
or locations. There are also embedded options associated with time of delivery within
the delivery month, deferral of delivery to future months and so on that are commodity
and exchange specific. For example, non-ferrous London Metals Exchange contracts can
allow for delivery in ports such as Bristol or Hamburg. In addition to the complications
this presents to delivery arbitragers, the presence of multiple delivery specifications in a
futures or forward contract requires the cheapest deliverable commodity to be identified
in order to determine the precise commodity grade and location that is being traded for
future delivery.2 Increasingly, futures exchanges are moving to financial settlement as a
method of dealing with problems of multiple delivery specification.
The quality basis is a development on the maturity basis. The quality basis relates to
the difference between prices for different grades of a commodity. Consumers encounter
various types of quality basis decisions on a daily basis. For example, automobile
drivers have to decide whether to buy regular or premium gas when tanking up. The
difference between these two prices is a quality basis. When the maturity basis is not
zero, the difference between spot and futures prices will be a quality basis. Numerous
examples of the quality basis are provided in the cash market price quotes provided in
the financial press. For example, on August 8, 1994, the Wall Street Journal reported
that, for New York delivery, Brazilian coffee was selling for $1.74/lb, while Columbian
coffee was selling for $1.84/lb (Poitras 2002, p.193). Engelhard refined industrial and
fabrication quality platinum was selling for $408 and $508 per troy ounce. Copper
cathodes and copper scrap were selling for $1.11 and $.87 per pound. On the same date,
Arabian heavy and light oil in Rotterdam were selling for $15.05 and $16.25 per barrel,
respectively.
Price differences due to variations in quality are determined by market con
siderations and are not typically constant. This is apparent from the time series
behavior of the Brent-WTI (West Texas Intermediate) crude oil spread. This variation
will be of concern to hedgers, both for determining the hedge position for the spot
commodity and identifying the appropriate deliverable commodity to use for the
calculating a cash-futures basis. Because of possible variation in the quality basis, futures
contracts with multiple delivery specifications must provide an acceptable method to
account for changes in the quality basis of deliverable commodities. In determining
the quality basis, considerations of location basis can often be relevant. For example,
the Brent-WTI quality basis also incorporates a location basis differential because WTI
prices are for Cushing, OK, delivery and Brent is for Rotterdam delivery (see Fig. 2.3).
More precisely, the location basis refers to the difference between the price of more
or less the same commodity at two different locations, e.g. live hogs in Des Moines and
Chicago, winter wheat in Topeka and Kansas City. An example of a location basis on
August 8, 1994, is provided in Poitras (2002, p.195), where No. 1 Canola in Vancouver is
$481.30/tonne and in Thunder Bay $473.60/tonne. For many of the physical commodities,
including canola, transportation costs have an important impact on the location basis.
In addition to transport costs, a number of other factors can impact the location basis,
such as local supply and demand considerations. Information on the behavior of the
location basis is important for hedgers to determine hedging strategies. For futures
hedgers, the relevant location basis is the difference between the local price of the relevant
spot commodity being hedged and the price of the commodity deliverable against the
futures contract. Where multiple delivery locations are permitted, the location basis
can also be important for determining which location is the cheapest for purposes of
making delivery. For many commodities, grade and location basis are combined, as is
the case for WTI crude oil in Cushing, OK, and Brent Sea crude in Rotterdam, which
feature somewhat different API gravity and sulfur content.
When referring to the basis, without further adjectives, reference is usually being
made to the difference between an appropriate forward or futures price and the cash
price: F(0,T) - S(0), e.g. Leuthold and Peterson (1983). This form of basis is also referred
to as the cash basis or the cash-futures/cash-forward basis. Certain markets have
specialized terminology for the basis, e.g. in FX markets the basis is referred to as the
swap rate or swap points. Comparing the conventionally quoted cash prices with the
futures prices for grain and oil reveals that, when the price of the deliverable spot
commodity is correctly identified, the nearby futures contract price is often almost
identical to the spot price of the appropriate deliverable commodity. As noted previously,
failure of the basis for wheat to converge to zero at maturity was a major element in US
Senate (2009). To make the relevant comparisons needed to evaluate the basis using
futures contracts, it is necessary to identify the deliverable commodity associated with
the futures contract of interest by referencing the contract specifications from the
relevant exchange website. This exercise will reveal that, even though it is not always
possible to precisely reconcile cash market quotes with futures markets quotes, various
commodities such as gold, silver, crude oil and soybean meal do typically exhibit a near
zero maturity basis.
Having established the correspondence between the prices of the cash and nearby
futures contract, it is important to determine the behavior of prices for futures or
forward contracts as delivery dates get progressively more deferred. Significant
deviations between spot and futures prices can be observed. For example, soybean
futures prices exhibit a pattern related to the soybean harvest cycle (see Table 2.2). In
Table 2.3, heating oil exhibits futures prices that are seasonal, highest in the inventory
buildup going into winter than in summer. The precise pattern of the futures price
deviations for different delivery dates varies considerably across commodities. A classical
“contango” relationship, where prices increase with the time to delivery, is observed
in the gold market (see Table 2.5), where futures prices increase monotonically with
interest carrying charges. The basis relationship between the prices of futures contracts
for different delivery dates is the futures basis. A “backwardation” in the futures basis
occurs when prices decrease monotonically with time to delivery. In general, the
basis and the futures basis are theoretically determined by cash-and-carry arbitrage
considerations.
F(t,T): the forward or futures price observed at time t for delivery at time T.
S(t) = St: the cash or spot or physical price of the deliverable commodity observed at
time t.
Typical subscripts that will be used are t=0 and t=1 with N for contracts that are
nearby or closer to delivery, and T for contracts that are deferred or farther from deli-
very. For consistency, it has to be that T ≥ N ≥ t. In much of what follows, the assum
ption, F(T,T) = S(T) is made in order for the price of a futures contract observed on
the delivery date t=T to be equal to the price of the deliverable commodity. Unless
otherwise stated, the spot commodity is taken to be the deliverable commodity on the
derivative security contract. This condition is often satisfied for forward contracts, but
requires assuming away the possibility of cross-hedging if futures contracts are involved.
As the quote from Tim Horton’s, Inc. in Sec. 1.1 illustrates, contracting methods
are an essential component of practical commodity risk management activities. In
turn, contracting methods employed depend on the economic characteristics of the
commodity industry involved. For example, the iron ore industry—where about two-
thirds of seaborne trade originates from three large suppliers—does not feature exchange
traded derivative securities to manage price risk. Instead, by setting prices on a quarterly
basis, the industry employs short dated forward contracts tailored to the requirements
of suppliers and the end consumers—primarily large steel mills and iron foundries.
A similar situation can be found in the market for metallurgical coal. In contrast, the
copper industry features a larger number of ore suppliers—many of which do employ
forward contracting methods to manage price risk—and a consumption chain that
involves a sufficient number of refiners and end users to sustain active exchange trading
in derivative securities for copper. The agricultural commodities feature different
variations on forward contracting and exchange traded derivative securities.
While useful, this brief summary disguises important features of futures trading. For
example, one of the significant limitations of forward contracting is the requirement
of precise specification of the grade and quantity of the commodity be determined
by the parties to the contract. This procedure raises the problem of how to ascertain
whether the commodity delivered meets the grade and quantity requirements. Because
forward contracts typically require delivery of the commodity, this procedure is an
essential feature of forward contracting. Because futures markets deal in a standardized
commodity for which delivery can be avoided by taking an offsetting position prior to
delivery, futures contracting avoids this problem.
Like futures, options have a specialized nomenclature. To understand this jargon, the
essential notions of an option contract for a commodity need to be identified:
An option contract is an agreement between two parties in which one party, the
writer, grants the other party, the purchaser, the right, but not the obligation, to
either buy or sell a given commodity at a future date under stated conditions.
Options almost always involve the purchaser making some type of premium payment
to the writer. The timing and form of the premium payment depends on the specifics
of the contract. For exchange traded options and many OTC options, the premium is
paid up front, when the option agreement is initiated. It is essential to recognize that
an option does not represent an ownership claim. Rather, an option is a claim against
ownership under prespecified conditions. While it is not necessary that options be
exchange traded, many option contracts do originate on exchanges.
Given this, two types of commodity option can identified:
A call option gives the option buyer the right to purchase the underlying commodity
from the option seller at a given price.
A put option gives the option buyer the right to sell the underlying commodity to
the option seller at a given price.
The seller of the option is often referred to as the option writer. An option purchaser
makes a payment to the option writer referred to as the option premium. Once the
premium has been paid, the purchaser has no further liability.
Various features for exchange traded commodity option contracts can be identified.
Some or all of these features may apply to other types of option transaction. In order
to be accurately specified, option contracts require an exercise or strike price as
well as an expiration date on which the right is terminated. The exercise price is the
contractually specified price at which the purchaser is allowed to buy (for a call) or
sell (for a put) the underlying commodity. When the exercise price is below (above)
the current underlying price, the call (put) option is said to be in the money. When
the exercise price is above (below) the underlying price, the call (put) option is out
of the money. An at the money option has the exercise price and underlying asset or
commodity price approximately equal. Exercising an option involves completion of
the relevant transaction specified in the option contract. Options that can be exercised
prior to the stated expiration date are referred to as American options; to be contrasted
with options that can be exercised only on the expiration date, commonly referred to
as European options.3 Depending on the type of option, either a spot delivery (physical
settlement) or a net dollar value transaction (cash settlement) may be required to
satisfy the conditions of exercise. Finally, the option contract will typically contain
other adjustment provisions, e.g. handling of variation in quality and location of
delivery.
funds. Where the purchase of a commodity is involved, the long cash-and-carry arbitrage
is executed by borrowing funds at the current interest rate and using those funds to
purchase the spot commodity, simultaneously contracting to sell the commodity at
some future date at the current futures price. Hence, the cash-and-carry arbitrage
provides a relationship between the futures and spot prices, which depends on the net
price of carrying the commodity until delivery. Similarly, other derivative securities
such as options on futures contracts also provide instances of arbitrage trades, which
may or may not involve purchase of the physical commodity. In this fashion, arbitrages
determine relationships among current prices of spot commodities and derivative
securities.
In practice, the execution of cash-and-carry arbitrages for commodities and the
associated behavior of the basis depends on the type of commodity under consideration.
Various questions must be addressed to identify the details involved in a specific
arbitrage. Is the commodity storable, storable with loss, or not storable? What costs
are associated with storage? Is the commodity harvestable? What is the appropriate
borrowing rate? Is there an off-setting carry return? To illustrate how arbitrage trading
determines the basis for a specific commodity, it is convenient to develop the trades
assuming that markets are perfect. In other words, there are no transaction costs, either
in the form of commission or bid/offer spreads. Other components of perfect markets
include no taxes or other regulatory restrictions, and equal lending and borrowing at
riskless rate of interest. Relaxing these assumptions converts the cash and carry equality
conditions to be derived into upper and lower bounds on forward prices associated with
the long and short arbitrage conditions.
while the June 2011/Jun 2012 contracts (= (1421.3/1410.7) give an implied interest rate of
0.751 percent. This is roughly consistent with the rates offered on three-month Euro-US
deposit futures that, on February 25, 2011, were trading at 99.625 (= 37.5 basis points)
for June 2011 and 98.735 (= 1.265 percent) for June 2012 delivery. Examination of the
futures price structures on February 25, 2011, for other commodities reveals a range
of different relationships. Silver, for example, is in backwardation on this date with
July 2011 at $32.91, July 2012 at $32.713 and July 2015 at $31.928. Soybeans exhibit a
backwardation for the July 2011 and July 2012 contracts at 1382 ¾ ¢ and 1310 ½ ¢. The
price structure of copper is also inverted, with prices for deferred June 2012 delivery at
$4.4275 being lower than the nearby June 2011contracts at $4.46. Due to the impact of
the harvest cycle on supply available for storage, most of the agricultural commodities
exhibit some form of kinking or reversal in the direction of futures prices as delivery
dates get more distant, e.g. the CME/CBT wheat contract. The diversity of futures price
spread behavior should be apparent. Full carry relationships are the exception and
situations can vary over time. For example, until the recent emergence of backwardation
in silver, all the precious metals futures complex prices exhibited contango.
Because each futures contract involves both a long and a short position, the
arbitrage relationship between cash and futures prices involves two trading strategies.
In addition to the long-the-cash strategy described above, which involves combining
a fully leveraged purchase of the spot commodity with a short futures position, it is
also possible to combine a short position in the cash commodity with a long futures
position, a short cash-and-carry arbitrage trade. Hence, cash-and-carry arbitrage
strategies for futures contracts are two-sided, having both a long and a short arbitrage
trade to be satisfied. While there are differences across commodities, in practice, due to
restrictions on the ability to short the cash commodity, execution of the short or reverse
cash-and-carry arbitrage can be substantially more difficult than the long arbitrage,
which involves only leveraged purchase of the spot commodity and costs associated
with carrying the commodity to delivery. In cases where sufficient commodity supply
is unavailable for lending to short sellers, the cash-futures basis cannot be determined
by the execution of cash-and-carry arbitrages.
Consider the extreme case of gold, where central bank stocks have traditionally
provided a significant supply of gold for short selling. Payments from short sellers
increase the central bank return on required holdings of gold inventory. In addition,
there are funds that hold physical gold stocks to earn the carry return by selling for future
delivery. Deviations from full carry in gold induce dis-hoarding, which puts pressure on
the spot price. As a consequence of limited restrictions on execution of the reverse cash-
and-carry arbitrage and the activities of funds that hold physical stocks of gold, gold
futures prices are at full carry. Another less extreme example would be a grain storage
company holding inventory over the crop cycle. At harvest, prices for future delivery
dates determine the return to storage. If the futures prices at harvest fall “too far” below
a level consistent with an adequate return to storage, then grain inventories would not
be purchased for storage, and the required grain inventory needed to maintain physical
delivery commitments would be acquired using long futures positions. This acts to
restore an adequate return to storage until the next harvest when a new storage cycle
begins. Table 2.6 illustrates that the return to storage for corn from December 2011
until July 2012 is 5.35 percent (annualized).
To execute the short cash-and-carry arbitrage for gold, the funds received from the
sale of the borrowed gold will be invested at a different, probably lower, rate of interest
than the long arbitrage. Taking i(0,T) to be the all-in lending rate and again ignoring
incidental costs, the short arbitrage is described in Table 2.7. Recalling that absence
of arbitrage implies π ≤ 0 and observing that Q > 0 gives the short cash-and-carry
Mar 2011 722’0 +10’0 712’0 712’4 723’4 706’6 30.690 999999’0 1:30:00 PM CST
-99999’0 2/28/2011
May 2011 730’2 +8’2 722’0 723’4 731’6 716’2 169.744 752’0 1:30:00 PM CST
692’0 2/28/2011
JUl 2011 732’0 +5’2 726’6 727’0 734’0 720’6 71.106 756’6 1:30:00PM CST
696’6 2/282011
Sep 2011 648’6 +2’6 646’0 646’0 650’0 641’0 4.229 676’0 1:30:00 PM CST
616’0 2/28/2011
Dec 2011 605’0 +3’2 601’6 601’4 607’0 597’4 17.585 631’6 1:30:00 PM CST
671’6 2/28/2011
Mar 2011 614’4 +4’2 610’2 610’0 615’4b 606’6 1.162 640’2 1:30:00 PM CST
580’2 2/28/2011
May 2012 621’0 +4’2 616’6 615’6 621’0 612’6 245 646’6 1:30:00 PM CST
686’6 2/28/2011
JUl 2012 624’6 +4’0 620’6 621’0 625’6 618’0 340 650’6 1:30:00 PM CST
690’6 2/28/2011
Sep 2012 565’2 +3’4 561’6 556’2 564’6b 556’2 47 591’6 1:30:00 PM CST
531’6 2/28/2011
Dec 2012 521’2 +1’4 519’6 516’2 522’4 515’2 982 549’6 1:30:00 PM CST
436’6 2/28/2011
Mar 2012 529’0 .0’6 529’6 529’0 530’0 526’4 33 559’6 1:30:00 PM CST
499’6 2/28/2011
May 2013 539’4 +3’6 635’6 – – – 65 565’6 1:30:00 PM CST
505’6 2/28/2011
Jul 2013 543’0b +1’2 541’6 546’4 546’4 542’2 2 571’6 1:30:00 PM CST
511’6 2/28/2011
Sep 2013 – +2’8 532’6 – – – 0 582’6 1:30:00 PM CST
502’6 2/28/2011
Dec 2013 526’4 – 523’6 526’0 – – 35 553’6 1:30:00 PM CST
493’6 2/28/2011
Jul 2014 – – 538’6 527’2 523’6 0 566’6 1:30:00 PM CST
506’6 2/28/2011
538.6
11/16/2012 6:44:17 PM
Measuring Risk and Exposure 165
arbitrage restriction on the gold futures price: F(0,T) ≥ S(0) {1 + i(0,T)}. If this lower
bound inequality is violated the short arbitrage can be profitably executed.
The combination of the long and short arbitrage conditions imposes upper and lower
boundaries on the gold futures price:
S( 0 ) { 1 + r( 0 ,T )} ≥ F ( 0 ,T ) ≥ S( 0 ) { 1 + i( 0 ,T )}
In an idealized world where r(0,T) = i(0,T) = r, the equality condition is binding and
F(0,T) = S(0) {1 + r}: the futures price will be fully determined at t=0 by the current spot
price and the cost of financing. This idealized result requires that: lending and borrowing
rates be equal; there are no short sale costs or restrictions; the commodity earns no
pecuniary or non-pecuniary carry return and is costless to store; and transactions costs
such as those associated with the bid/offer spread are ignored. When the futures prices
for gold obey this condition, the commodity is at full carry.
Given the idealized full carry model for gold, it is useful to evaluate what happens to
basis behavior as maturity of the gold contract approaches. Defining Δ X ≡ X(1) – X(0)
and substituting X(1) ≡ X(0) + Δ X permits the profit function for a one-to-one long
spot/short futures gold position to be expressed as:
This result is useful for interpreting the profitability of inventory hedges. Recalling that
r(i,j) is not annualized but, rather, reflects the actual interest cost over the i to j holding
period, r(0,T) will typically be larger than r(1,T). This is due to the one period reduction
in the number of days until maturity for both the loan and the futures contract. Hence,
even if ΔS = 0 and the level of annualized interest rates is unchanged, there is a time
decay in the basis associated with Δr < 0 which is fundamental to understanding hedge
profit determination. There is a “time clock” at work in a hedge, acting to reduce the
difference between the spot and futures price. The time clock continues to wind down
to the point where r = r(T,T) = 0 and F(T,T) = S(T). Significantly, this time decay is not
present in the futures basis.
Because both the cash price and the futures prices of a commodity vary with time,
the basis varies with time as well. In addition, at any particular time, there may be
multiple different cash prices for a commodity, depending upon the precise grade
of the commodity, the location of the commodity, and the costs of transporting the
commodity to market. The basis will vary, therefore, according to the commodity’s
attributes and location. At any given time, the same crop could give rise to a variety
of basis calculations, depending upon specific variables in the crop and the market
where it is to be sold. A buyer or seller of a crop must take into account each of those
variables when computing the basis for a particular sale or purchase.
An important feature of this Senate investigation was the failure of the delivery basis
for wheat futures prices to converge to the spot price of deliverable wheat, substantially
undermining the price discovery and hedging function of wheat futures markets.
Non-storability can be associated with rapid spoilage or prohibitive storage costs.
There is a tradeoff between “non-storability” and length of time to delivery. For non-
storable commodities, neither the short or long cash-and-carry arbitrage can be
executed for contracts more than a few days or weeks in the future. Commodities in
this group have included potatoes, eggs and onions. Poultry and sugar beets also have
elements of non-storability, e.g. “Chickens cannot be shipped far before losing value,
due to both the direct costs of transport or extra feed and the indirect costs from the
birds losing value due to stress, weight loss, or death during transport, or aging during
additional feeding. Similarly, sugar beets lose value quickly, and transport costs are still
quite high” (Macdonald and Korb 2011, pp.3–4).
When available, the price performance of longer dated futures contracts for “non-
storable” commodities is decidedly more erratic when compared to such contracts
for storable commodities. Without the ability to do cash-and-carry arbitrages, price
determination for futures and forward contracts relies on expectations about future
spot prices. At times, this can create market clearing problems, due to lack of liquidity
on one side of the market. Empirically, while a number of different types of contracts
for non-storables have been offered over the years, there are currently only a few thinly
traded futures contracts for truly non-storable commodities. Where contracting for
future delivery occurs, short dated forward contracting is the norm, e.g. Macdonald
and Korb (2011, p.4). A number of the older studies on non-storables can be found in
A. Peck (1977).
In addition to the truly non-storables, certain other commodities may not satisfy the
strict requirements of storability. For example, feeder cattle associated with delivery
on the CME futures are required to be 650 to 849 pound steers. Given weight gain
over time, heavier steers eligible for delivery in one contract month cannot be carried
for delivery against the next contract as would be the case for, say, gold or silver.
However, in these cases, non-storability does not significantly affect cash-futures price
determination because there are feeder cattle available that can be used to do the cash-
and-carry arbitrage. There are opportunities to purchase feed stock that is at a point in
the growth cycle that the stock will qualify for future delivery. Similarly, agricultural
crops are only profitably storable from harvest time until the beginning of the next
harvest. As illustrated in Table 2.6, there is a break in the corn basis from the September
to December contracts associated with the new supply created by the corn harvest.
On balance, it is safe to say there is considerable diversity across commodities in the
execution of the long cash-and-carry arbitrages.
Examining the long grain elevator arbitrage in Table 2.8 reveals the importance of
storage companies in arbitrage trade execution. Grain elevators and similar storage
operators are able to provide storage facilities at cheapest cost and, as a consequence, are
appropriately situated to exploit long arbitrage opportunities. Henderson and Fitzgerald
(2008) describe the grain marketing process:
Since their emergence in the mid-1800s, grain elevators have earned income by
collecting, storing, and readying grain for transportation. Smaller, country grain
elevators collect grain from farmers, hold it in storage, and coordinate transportation
to final end users or larger terminal elevators, which coordinate larger shipments to
other domestic or international users. The grain held in storage is either owned by
the elevator or by the farmers, who pay storage fees.
Table 2.8 Profit Function for a Long Grain Elevator Cash and Carry Arbitrage
DATE Cash Position Futures/Forward Position
Table 2.9 Profit Function for a Short Grain Elevator Cash and Carry Arbitrage
DATE Cash Position Futures/Forward Position
t=0 Sell {QA} units of grain @ S(0) Long QA units at F(0,N)
and invest proceeds at @ r(0,N)
saving grain storage costs @ sc(0,N)
t=T Use maturing investment funds to take delivery of the QA units of grain
(warehouse receipts) against the long position.
Assuming all costs of storage are variable, the arbitrage profit function is:
Grain storage companies also control the bulk of supply available for short selling. In
contrast to gold where short sales involve borrowing of physical gold, short arbitrage
trade execution in grains involves selling activities of storage operators. This process
is decidedly more complicated than for gold (see Table 2.9). Because the bulk of grain
in storage at a given time has already been contracted for future delivery, the supply
available in storage for sale in the spot market between harvests is restricted. Costly
movement of grain from other storage locations can be difficult due to transport costs and
uncertain supply conditions in other locations. Committed grain for deferred delivery
that is sold needs to be repurchased for a nearby delivery date that will accommodate
the completion of previously initiated deferred delivery contracts.
Consideration of the short cash and carry grain arbitrage exposes a fundamental
distinction between financial “commodities” and the physical “commodities.” Shorting
a financial product such as a common stock index or a foreign exchange rate is
straightforward. The long and short arbitrage bounds on forward and futures prices for
financial “commodities” are tight. Due to limitations of physical supply, short selling
of commodities is problematic. As a consequence, the short arbitrage bound on the
basis for commodities is not binding and convenience yield is introduced to explain
deviations from the cash-and-carry arbitrage bound. More precisely, ignoring fixed
costs the general cash and carry condition for forward prices is:
F ( 0 ,T ) = S( 0 ) { 1 + cc( 0 ,T ) − cr( 0 ,T )}
where cc(0,T) is variable carry costs, expressed as a rate and cr(0,T) is the variable
pecuniary and non-pecuniary returns to holding the commodity, also expressed as a
rate. Decomposing cc and cr into components gives:
F ( 0 ,T ) = S( 0 ) { 1 + r( 0 ,T ) + oc( 0 ,T ) − d( 0 ,T ) − cy( 0 ,T )}
where r(0,T) is the interest rate portion of carry costs, cc(0,T) – r(0,T) ≡ oc(0,T) is
that portion of variable carry costs not attributable to interest carrying costs,
d(0,T) is the pecuniary interest or dividend return associated with carrying the
commodity, cy(0,T) = cr(0,T) – d(0,T) is that portion of cr not associated with pecuniary
interest or dividend return.
Unlike financial products where attention centers on r(0,T) and d(0,T), physical
commodities do not earn a pecuniary carry return—d(0,T) = 0—and elements in oc(0,T),
such as storage costs and spoilage, can have greater importance than r(0,T) . Recognizing
the essential role of storage operators in determining commodity forward prices, the
diversity of arbitrage execution across commodities can be captured by generalizing the
profit function for the cash-and-carry arbitrage to include both fixed and variable carry
costs and carry returns. More precisely, for a transaction starting at t=0 and ending at
t=T, the profit function for the marginal storage operator would produce:
where CC(0,T) is the fixed cost component of carry costs, e.g. regular maintenance and
capital costs of storage facilities; and CR(0,T) is any fixed return earned for being able to
hold the commodity, e.g. payments associated with “take or pay” storage contracts with
grain consumers. Though this approach allows for all elements of cost and return to be
introduced, the specification of CC and CR can be problematic. In many cases, CR = 0
and CC will vary inversely with the amount of the commodity in storage.
For notational convenience, it is sometimes expedient to define the “implied carry”
as ic(0,T) where ic(0,T) ≡ cc(0,T) – cr(0,T) combining the carry cost and carry return
elements. It is also typical to ignore fixed costs and to work with the non-pecuniary
variable, cy(0,T), to explicitly allow for potential returns the commodity may provide
during the “arbitrage” period. Recalling that the pecuniary return d(0,T) = 0, it follows:
F ( 0 ,T ) = S( 0 ) { 1 + cc( 0 ,T ) − cy( 0 ,T )}
Following the discussion of the long spot/short futures inventory hedge for gold, this
formulation can be directly applied to an inventory storage hedge profit function
involving a general commodity. In this case, the profit function can be written:
p /Q = { F ( 0 ,T ) − S( 0 )} − { F ( 1,T ) − S( 1)}
= [ S ( 0 ) { cc( 0 ,T ) − cy( 0 ,T )}] − [ S( 1){ cc( 1,T ) − cy( 1,T )}]
Again, defining Δ X ≡ X(1) – X(0) and using X(1) ≡ X(0) + Δ X permits this form of the
profit function to be expressed as:
%DVLV
,QYHQWRU\
'
Analytically, the notions of convenience yield and supply of storage have direct
implications for explaining the behavior of a key variable in the speculator and hedger
profit functions: F(0,T) - E[S(T)].6 However, the connection between the forecasting
accuracy of the futures or forward price and the notions of convenience yield and
supply of storage is not immediately obvious. Building on empirical work in Working
(1949), modern studies model the supply of storage as a relationship between the basis
(F(0,T) – S(0)) and commodity inventory levels. Because convenience yield and the
supply of storage are concerned with properties of the physical commodity, some early
approaches ignore the role of the futures market and examine E[S(T)] - S(0). The two
period, two agent equilibrium model of the supply of storage of Brennan (1958) is a case
in point. Supply and demand functions are derived for a consumer-merchant market.
Brennan describes the market this way:
During any period there will be firms carrying stocks of a commodity from
that period into the next. Producers, wholesalers, etc. carry finished inventories
from the periods of seasonally high production to the periods of low production.
Processors carry stocks of raw materials. Speculators possess title to stocks held
in warehouses. These firms may be considered as supplying inventory stocks or,
briefly, supplying storage . . . On the other hand, there will be groups who want
to have stocks carried for them from one period . . . to another period . . . These
consumers may be regarded as demanding storage.
In this case, the supply and demand functions for storage are behavioral, dependent on
both the spread between the expected future spot price and the current spot price as
well as on the levels of stocks being held. The upshot is an identified supply of storage
function which provides a (potentially nonlinear) monotonically increasing relationship
between physical inventory levels and E[S(T)] – S(0).
The development of the partial equilibrium supply of storage model to include futures
markets was provided initially by Weymar (1966, 1968) and extended by Turnovsky
(1983). In Weymar’s model, three agents are identified: merchants, manufacturers
and speculators. Futures markets provide cash market participants with an additional
method of carrying inventories. Equilibrium in the futures market is directly specified
and a supply of storage function is derived. Much as in Brennan’s case, there is a
monotonically increasing relationship between physical inventory levels and E[S(T)] –
S(0). Using a more sophisticated, but similar model, Turnovsky is able to show:
with risk averse behaviour, the current futures price is a weighted average (with
weights summing to less than unity) of the current spot price and the expected
future spot price. Only if . . . producers and speculators are risk neutral . . . does
F(0,T)= E[S(T)] and the futures price become an unbiased predictor of the future
spot price. Otherwise, the futures price is a biased predictor, with the direction of
the bias depending on the magnitude of the (relevant) cost parameters.
For stocks which are hedged on an active futures market the price spread relevant to
a decision about storage levels is the difference between a futures and a spot price.
Arbitrage between cash and futures markets will insure that the cash price expected
to exist in a future period is accurately reflected in the current quotation of the
futures price for delivery in that period.
The transition from unobserved expected prices to observed futures prices leads to more
recent empirical studies of the supply of storage where the observed basis (F(t ,T) – S(t))
at a given point in time is attributed to three factors: the costs of storage; the convenience
yield; and a risk premium to account for the substitution of the futures price for the
expected spot price in empirical estimations. From this point, empirical studies have
examined the implications of different definitions of “inventory” and alternative
estimation methods that do not directly rely on inventory levels. In addition, theoretical
studies of convenience yield have developed using option valuation methodology. This
approach inspired the influential Gibson and Schwartz (1990) and Schwartz (1997).
Carter and Ghia (2007, pp.864–5) give the following helpful overview of studies
evolving from the seminal Working (1949).
There are two main explanations for the particular shape of the Working curve.
The first is the “supply of storage” explanation (Working 1949), which includes
the concept of convenience yield to explain storage under backwardation
. . . However, the supply of storage approach has been criticized for taking
convenience yield as given, and then incorporating it into modern commodity
models through an ad hoc functional form (see Deaton and Laroque 1996,
and Brennan, Williams, and Wright 1997). The second explanation views the
Working curve as an artifact of an aggregation problem (Wright and Williams
1989; Brennan, Williams, and Wright 1997). Holbrook Working aggregated
U.S. wheat stocks from different domestic locations and for different wheat
grades, and then plotted these stocks against intertemporal Chicago price
spreads. According to this second explanation, if the Working curve had been
drawn using single location prices and stocks, storage under backwardation
would not have been found, and the intertemporal theory of prices would be
complete without any need to appeal to the convenience yield idea.
Gao and Wang (2005, p.401) describe the indirect method of testing for the supply
of storage proposed by Fama and French:
is higher than the variability of forward prices during those periods when
the interest-adjusted basis is negative (indicating a low inventory period) and
the variability of spot and forward prices tends to be equal when the basis is
positive (indicating a high inventory period). Furthermore, they found that
the variability of basis is greater when the basis is negative than when the basis
is positive.
and Schwartz (1990) and Schwartz (1997), a stochastic process for the convenience
yield is assumed in these studies and combined with assumptions about the stochastic
process for spot prices for a two factor model plus interest rates for a three factor model.
The two factor specification of Gibson and Schwartz (1990) and Hilliard and Reis (1998)
takes the stochastic differential equation form:
F( 0 , N ) = S( 0 ) ( 1 + ic( 0 , N )) F ( 0 ,T ) = S( 0 ) ( 1 + ic( 0 ,T ))
It follows that:
F ( 0 ,T ) = F ( 0 , N ) ( 1 + ic( 0 ,T − N ))
the short position at F(0,T). This sequence of transactions can be used to specify ic(0,T-N)
as the implied carry cost, reflected in F(0,T) and F(0,N) at time t=0, for a cash-and-carry
arbitrage which will begin at t=N and end at t=T. Observe that the actual implied carry
earned on the cash and carry transaction between N and T, cc(N,T) - cr(N,T), will not be
the same as that reflected in futures prices at time t=0.
179
risk management is reflected in the annual reports of major banks. For both financial
and non-financial firms, the financial risk management approach requires each of these
risks to be assessed for the specific firm involved. Decisions are then made on which of
these risks will be assumed and which will be managed. Beyond this general intuition,
things get more difficult as it is not possible to deal with all the various aspects in detail.
Further focus is required. As a consequence, there is a myriad of different possible
approaches to corporate risk management. Some treatments, e.g. Dowd (1998), Jorion
(2006), emphasize the measurement of market risks using VaR; others, e.g. Andren
et al. (2005), Oxelheim and Wihlborg (1997), emphasize the use of CFaR to provide an
integrated treatment of business risks. Still others, e.g. Smith, Wilford and Smith (1994),
examine the methods for handling commodity or financial risk in specific situations.
The modern, integrated approach to corporate risk management is a utopian ideal. It
is conventional, if not essential, to treat risks in isolation in order to better conceptualize
the methods of managing the risk. In certain situations, such as for financial firms
making markets in securities, the risks being managed are primarily market risks,
and utopian models, such as those derived from VaR, can be used effectively. These
situations stand in stark contrast to cases for non-financial firms where the risks are less
amenable, e.g. Procter & Gamble or Coca-Cola seeking to manage firm-wide business
risks across different geographical markets. Even where the risks are perceived to be
primarily market risks, the complexity of the risks being faced can defy adequate
treatment (see Sec. 1.3). In all of this, the traditional distinction between hedging and
speculation seems misplaced. This is unfortunate, as there are useful lessons contained
in the earlier discussions of risk management, which typically structure the discussion
as a partial equilibrium problem in hedging specific transactions.
The traditional approach to commodity risk management predates the modern
renaissance of derivative securities. Products for managing risks were limited, both
by legislation and market practices. Financial derivatives, so important in modern
financial risk management, were largely traded OTC and were not widely used. Various
types of risks, such as the volatility in market prices associated with the introduction of
commodity ETF trading or variation in cash flows due to the impact of flexible exchange
rates, were still on the horizon. In this simpler world, commodity risk management was
typically associated with hedging using agricultural forward or futures contracts, where
risks were treated in isolation and transactions involved in the hedge were emphasized.
Considerable attention was dedicated to clarifying the distinction between hedging
(risk management) and speculation. This distinction seems to have been lost in the
modern approach to risk management. There is a modern belief that the engineering of
risk is a precise enough science to make this distinction irrelevant.
transactions in a trading schematic. The basic profit function is then specified from the
schematic. For simple trading strategies, such as a naked speculation, the profit function
and schematic are not too useful, as the basic insights can be obtained without much
analysis. However, for more complicated trades, the profit function can be an invaluable
aid. Once the basic profit function is specified, it is possible to do manipulations and
substitutions which can be used to identify relevant features of the trading strategy. One
important substitution that is often used is to replace the deferred contract prices with
the cash-and-carry arbitrage conditions.
To complete the illustration of what determines the profitability of long and short
positions, let Q be the number of units of the commodity purchased. At t=0, the two
parties to the futures contract for delivery of Q units at t=T agree to a price of F(0,T).
Consider what happens if F(1,T) > F(0,T), i.e. that futures prices for the commodity rise.
The short position who agreed to sell Q units at F(0,T) now is faced with a situation
where the value of the commodity to be delivered is Q F(1,T) versus Q F(0,T) the previous
period. Pretending for the moment that t=1 was actually the delivery date T, then the
short would have to cover by going into the cash market, purchasing the appropriate
number of units of the commodity and delivering. This would require a larger outlay
of funds, Q F(1,1) = Q S(1), than would be received from the sale of the commodity
through the futures markets, Q F(0,T). The opposite type of example would hold for the
long position. This produces the elementary result: short positions benefit from price falls
while long positions benefit from price rises.
While it is straightforward to illustrate the conditions under which long and short
positions are profitable, the analysis is decidedly more complex when more involved
positions are being considered. For this reason, it is instructive to illustrate how to
convert the previous discussion into algebraic terms. Table 2.10 demonstrates that the
profit function for the long futures position is: π(1,T) = Q {F(1,T) – F(0,T)}. It follows that
π > 0 when futures prices rise between t=0 and t=1. Observing Q > 0 by assumption, a
similar profit function can be defined for short positions:
It is useful to recognize that the form of the profit function will depend on how the
units, Q, are specified. Some presentations will vary the sign of Q such that when a long
position is indicated then Q > 0 and, for a short position, Q < 0. In this case, what is
referred to above as the long profit function will apply in both cases. If prices fall, and
Q < 0 for a short, then profit will be positive. While this approach is somewhat tidier to
use in presenting basic concepts, in the analysis of more complicated trading strategies
this convention will lead to variations in the signs of some terms when compared to
derivations based on Q > 0 everywhere. Assuming Q > 0 throughout facilitates use
of the rule for calculating profit: “what you sold it for minus what you bought it for.”
While in many cases the difference is between using Q > 0 and Q of varying sign is
transparent, there are instances where some care is required when comparing results
given using the different specifications of Q.
The speculative profit function for a long or short position is relatively simple when
compared to the profit function for a hedger, where account has to be taken of both
the futures and cash positions. In addition, details for the hedge will vary depending
on the specific hedge under consideration. To illustrate the hedger’s profit function,
consider the hedging problem confronting the stylized grain elevator operator of the
19th century. Grain would be hauled by land to the riverside where the grain elevator
was situated. The elevator operator would pay the farmer the prevailing cash price and,
in the case at hand, would store the grain through the winter until the river thaw in the
spring. The elevator operator was concerned that grain prices would move adversely
between the harvest and springtime. To offset this risk the farmer would engage in a
futures hedge with the traders at the Chicago Board of Trade. The relevant transactions
are described in Table 2.11. This profit function applies only to the hedge transaction, it
does not take account of other profits associated with the actual business. For example,
there may be spoilage or loss or grain to vermin such that QA(0) ≠ QA(1), or the grain
may be processed and sold in a different form.
The intricacies of hedging can be illustrated by extending the grain elevator hedge
example in Table 2.11.7 The futures position in the example implicitly assumes that
the elevator operator has grain for sale which is not deliverable against the futures
Table 2.11 Profit Function for a Grain Elevator Hedge using Futures Contracts
DATE Cash Position Futures Position
position because it does not conform to the standardized grade. The possibility that the
relationship between prices for different grades will change over the life of the hedge
is a source of risk in the hedge. If the elevator operator has a deliverable grade of grain
then the futures hedge can be completed by making delivery. In this case, the futures
hedge profit function takes the form of a profit function for a forward sale, a hedge that
is done using forward contracts. Because T =1 and F(1,1) = S(1) in the case, the forward
sale profit function where QA = QH is:
p ( 1) = QA { F ( 0 ,1) − S( 0 )}
In this case, the costs of financing, storing the commodity and making the delivery at
maturity are ignored for simplicity.
The profit function for the grain elevator hedge using futures explicitly recognizes
that the size of the hedge position in futures may be different than the size of the cash
position. The precise relationship between the size of the cash and futures positions can
be formulated as an optimization problem from which an optimal hedge ratio can be
determined. However, it is still revealing to assume that the hedge is one-to-one (QA =
QH), which gives, after some manipulation:
p /Q = { F( 0 ,T ) − S( 0 )} − { F( 1,T ) − S( 1)}
The profitability of the hedged position depends on the change in the difference
between the spot and futures prices. In a positive carry market, if this difference
narrows, the hedge will have π > 0.
The analytical usefulness of the hedge profit function is also apparent when the
futures spread trades are considered. For example, the basic intra-commodity spread
trade, also called a calendar spread, involves offsetting, long and short, positions in
different contract delivery months for the same commodity. If the spread is in different
commodities, an inter-commodity spread, the delivery months involved have less
importance. Intra-commodity trades can be done for different reasons, which are
discussed in Poitras (2002, Ch.3). Recalling the use of N and T for the nearby and
deferred deliveries of amount QN and QT , this trade is depicted in Table 2.12. One
immediate interpretation of spread behavior from this profit function is to assume
Table 2.12 Profit Function for an Intra commodity Futures Spread Position
DATE Nearby Position Deferred Position
that the spread is one-to-one and intra-commodity, i.e. let QN = QT = Q, which gives,
after some manipulation:
p /Q = { F ( 1,T ) − F ( 1, N )} − { F ( 0 ,T ) − F ( 0 , N )}
The one-to-one intracommodity spread which is short the nearby and long the
deferred will be profitable if the difference between the deferred and nearby prices
widens, i.e. becomes more positive or less negative. The opposite would be true for the
alternative spread, long the nearby and short the deferred.8
where: Wt+1 is wealth at time t+1 and Wt is the known level of initial wealth; A is the
fixed initial size of the asset, e.g. acres planted for a farmer; Yt+1 is the possibly random
quantity per unit or yield per unit of the asset observed at t+1; Pt+1 is the random spot
price at t+1; C(A) is the given cost function associated with producing or purchasing A;
r is the risk-free interest rate.10 Manipulation gives the more conventional form of the
wealth process for a single risky asset:
Wt +1 = Wt ( x( 1 + R ) + ( 1 − x )( 1 + r )) = Wt (( 1 + r ) + x( R − r )) = Wt + p t +1
where πt+1 is the profit defined by the wealth process realized at time t+1, x is (C(A)/Wt)
the given fraction of initial wealth invested in the risky asset, and (1+R) is [(A Yt+1Pt+1)/
C(A)] one plus the rate of return on the risky asset. For simplicity of exposition, it will be
assumed that x > 0 in what follows.11
The basic specification for the decision maker’s terminal wealth function requires
some additional terms if there is a need to capture the payoffs associated with,
say, introducing a put option. While the terminal wealth function derived from
the wealth process, with put options included, follows appropriately, some motivation
is required. In particular, in the absence of some form of put option to provide
asset insurance, there is a natural minimum on R, the rate of return on the invest-
ment. Either a complete catastrophic loss occurs where Yt+1=0, or a spot price of
zero occurs at time t+1, both cases corresponding to the result (1+R)=0. Significantly,
three possible variants of put option payout are possible, each aimed at dealing
with the different types of risks faced by the risk manager. More precisely, put
option pay-outs can depend on the deviation of price, yield or revenue from a stated
exercise value. Payouts based on revenue provide protection against Yt+1Pt+1 falling
below a given floor. In contrast, payouts based on yield or price cannot guarantee a
minimum return higher than (1+R) = 0. For the farmer example, while put payouts
based on revenue set a lower level for farm income, put option payouts guaranteeing a
price of $K per bushel cannot prevent a 100 percent loss due to crop failure, nor can a
put payout based on yield providing for, say, Y bushels an acre prevent the future spot
price falling to zero. However, put payouts based on either price or yield do reduce
the probability of the total return attaining low values and, as a result, do alter the
distribution for terminal wealth.
In practice, conventional exchange traded put options are structured with payouts
based on price. Other types of put options, such as multiple peril crop insurance
schemes, are a type of yield insurance. Still other types of put options, such as some types
of real options or embedded options in forward contracts, provide revenue or income
protection. The case where the put payout is based on revenue insurance produces a
wealth process similar to the yield insurance case. Introduction of a put option based
on price produces:
QP K − Pt +1 z
= Wt x(1 + R) + (1 − x )(1 + r ) + z t max 0 , −
Wt Pt Pt
z
= Wt (1 + r ) + x(R − r ) + γ max[0 , − R p ] −
Pt
where K is the exercise price on the put option which is assumed to be “at the money”
(where K = Pt ), z is the price per unit of output of the put, Qz is the number (in output
units) of puts purchased, with the ratio γ being the asset value covered by the option
position divided by initial wealth.12
This specification can be contrasted with that for put option payouts based on yield
where, instead of the number of options to purchase, it is the fraction of A to insure
which is the decision variable:
Wt y+1 = AYt +1Pt +1 + (Wt − C( A))(1 + r ) + Q y (Pt +1 max[0 , Y − Yt +1 ] − L)
where L is the price (put premium) per unit of A for the yield put option, Qy is the
number of units covered by the yield put option and Y is the yield floor provided by the
put option or insurance plan. Defining the optimization problem by allowing the risk
manager to choose the fraction of A to insure leads to:
where l equals (LA/C(A)), λ = (Qy/A) is the fraction of A, e.g. the total planted acreage,
covered or insured with the physical yield put option and RR = {Pt+1 Y A}/C(A).13
Assuming actuarially fair pricing requires insurance to impact the decision problem
through its effect on downside risk and skewness.
This basic structure can be readily adjusted to account for other derivatives, such
as futures or forward contracts. For example, if it is assumed that the only hedging
instrument available for a farmer is futures contracts then the underlying wealth
dynamics can be specified:
Wt +1 = A Yt +1Pt +1 + [Wt − C( A )]( 1 + r ) + Q f ( f t +1 − f t )
where: Q f is the quantity of futures contracts sold (-) or bought (+); and f t+1 = F(t+1,T)
and f t = F(t,T) are the futures prices observed at t+1 and t respectively. Manipulation
gives:
Wt +1 = Wt ( x( 1 + R ) + ( 1 − x )( 1 + r ) + HR f )
= Wt (( 1 + r ) + x( R − r ) + HR f )
= Wt + p * t +1
where: π*t+1 is the profit for the futures hedge realized at time t+1, x is, again, (C(A)/Wt)
the fraction of initial wealth invested in the crop production, H is the value (f t times Q f)
of the hedge position divided by initial wealth (not the value of the spot position), Rf is
(f t+1 - f t)/f t and (1+R) is, again, [(A Yt+1Pt+1)/C(A)] one plus the rate of return on planting
for a farmer.
EU [ x ] = ∑ θ j U j [x] where ∑θ j =1
j =1 j =1
where EU[x] is the expected utility of x; S is the number of possible futures states of
the world; θj is the probability that state j will occur; and U[xj] is the utility associated
with the amount of x received in state j. The EU function ranks risky prospects with an
ordering which is unique up to a linear transformation. While there are a number of
possible selections for x, in what follows either terminal wealth or terminal profit will
typically be used. One of the key difficulties with this approach—the non-additivity of
subjective probability—was identified during the “years of high theory” from 1926–39
(Shackle 1967) that produced a number of insightful approaches to resolving the difficult
quandaries surrounding the distinction between risk and uncertainty. Knight (1921)
identified the ability of entrepreneurs to resolve uncertainty as a source of economic
profit. Keynes (1921) used subjective probability to motivate the distinction. Shackle
Altering the context slightly from aggregate investment to commodity risk manage-
ment, the Shackle “potential surprise” approach appears capable of explaining the
empirical observation of a more violent downdraft of prices with a bear market compared
to the gradual upswing of a bull market. This implies value for hedging programs that
protect against downside risk.
example, it is not apparent how to determine the probabilities θj in the expected utility
function. The axiomatic foundation cannot say much more than that the probabilities
are subjective. General equilibrium models often proceed by assuming that expecta-
tions are homogeneous or that individual agents are homogeneous. Such assumptions
permit the derivation of market equilibrium conditions, such as the “the speculative effi-
ciency hypothesis” or the CAPM. However, general equilibrium concerns are of little use
in commodity risk management applications. The decision problems encountered are
partial equilibrium. The theoretical results apply to speculators and hedgers confronted
with a parametric world of atomistic competition where their activities will not impact
prices. In this process, the expected utility function can be an invaluable analytical tool.
This can be readily demonstrated by applying an essential tool from functional analysis:
the Taylor series expansion.
To see this, consider the problem of determining the cost of risk. The solution to
this problem would be useful in analyzing whether to buy insurance or to invest in a
risky capital project. While there are a number of possible methods to extract the cost
of risk, consider the following solution. Let the expected value of terminal wealth be:
E[Wt+1] = Ω. Observe that Ω is a parameter that permits the certainty equivalent income
of a risky prospect to be defined as Ω - C, where C is the cost of risk. It follows from
the expected utility axioms that the cost of risk, C, can be calculated as the difference
between the expected value of the risky prospect and the associated certainty equivalent
income:
S
U[Ω − C] = ∑θ
i=1
i U[Wi ] = EU[Wt +1 ]
It is now possible to expand U[Ω – C] in a Taylor series and estimate the cost of risk by
manipulating the first and second order approximations.
More precisely, expanding the deterministic function U[Ω – C] around Ω the first
order approximation is:
U[Ω − C] = U[Ω] + U /[Ω] (Ω − C − Ω) = U[Ω] − U /[Ω] C
Similarly, a second order Taylor series approximation for the deterministic function
U[Wt+1], expanded again about Ω provides:
1 //
U[Wt +1 ] = U[Ω] − U [Ω] (Wt +1 − Ω ) + U [Ω] (Wt +1 − Ω )2
/
2
1
→ EU[Wt +1 ] = U[Ω] + U / /[Ω] var [Wt +1 ]
2
Using U[Ω – C]=EU[Wt+1], observing var[Wt+1] = var[(1+R) Wt] = Wt 2 var[(1 + R)] and
manipulating gives:
U / / [Ω] C U / /[Ω] Wt
C = − var[Wt +1 ] → = − var[1 + R]
2 U / [Ω] Wt 2 U / [Ω]
This demonstrates theoretically that the cost of risk will vary across utility functions,
providing a foundation for widely used theoretical measures of the cost of risk. The
measures of absolute risk aversion, -{U''/U'}, and relative risk aversion, -{U'' Wt }/U' are
now textbook concepts, e.g. Elton and Gruber (1995).
As with other results that rely on a Taylor series expansion to derive a theoretical
result, e.g. Poitras (2005, Ch.5), subtle changes occur when a higher order of the
expansion is used. First order expansions only provide a linear approximation to the
function. Such approximations are helpful when the functions are approximately linear
or the change in the variable—in this case the cost of risk—is small. For convex and
other nonlinear functions, sizable changes in the variable will produce inaccurate
approximations indicating a second order approximation is in order. Consider what
happens when a second order approximation is used:
1
U[Ω − C] = U[Ω] + U / [Ω](Ω − C − Ω) + U / /[Ω] (Ω − C − Ω )2
2
= U[Ω] − U /[Ω]C + U / / [Ω] C 2
Using U[Ω-C]=EU[Wt+1] no longer provides a convenient solution. The solution to the
resulting quadratic equation now has two solutions, as opposed to one for the linear
approximation. Neither solution has the appealing form of the commonly used measures
of absolute and relative risk aversion.
U / / [Ω] U / / /[Ω]
U[Wt +1 ] = U[Ω] + U /[Ω](Wt +1 − Ω ) + (Wt +1 − Ω )2 + (Wt +1 − Ω)3 + .....
2! 3!
Exploiting this type of expansion requires certain technical conditions be satisfied. For
example, convergence of the power series within the interval of interest is needed. In
addition, desirable properties for utility functions require: U'[W] > 0, non-satiation;
U''[W] < 0, risk aversion; and U'''[W] > 0, preference for positive skewness.
Employing what are typically viewed as relatively weak distributional restrictions,
e.g. Hassett et al. (1985), the Taylor series representation of U[W] can be transformed
into an approximation for a general expected utility function based on the moments of
the conditional distribution for Wt+1. The relevant approximation is derived by taking
conditional expectations at time t and ignoring terms associated with moments higher
than the second, for a mean-variance approximation, and moments higher than the
third, for a mean-variance-skewness approximation. The general notation EU[ ⋅ ] will
be used to denote such a moment preference functional. Taking expectations for the
mean-variance-skewness case gives:
U / /[Ω ] U / / /[ Ω ]
EU MVS [Wt +1 ] ≅ EU MVS = U [Ω] + 0 + var[Wt +1 ] + skew[Wt +1 ]
2! 3!
for positive skewness may undertake projects with skewed payoff distributions that
appear to be unfair gambles.” Horowitz (1998) correctly takes exception to the Prakash
et al. claim, arguing that there is no underlying utility function that is consistent with
the central theoretical condition which Prakash et al. use, i.e. 3U''/U''' > skew[W1]/
var[W1]. Horowitz refutes the Prakash et al. claim by demonstrating that it is not possible
for an expected utility function to conform to the Prakash et al. restrictions.
Theoretical exploration of conditions required for mean-variance solutions led Ross
(1978) to propose the family of separating distributions. Chamberlain (1983) further
characterizes the set of distributions from which the mean-variance expected utility
function may be derived as distributions with spherical symmetry, a special case of
elliptical distributions. Wei et al. (1999) extend these results to a “linear conditional
expectation” (LCE) relation. A well-known distribution that admits LCE is the
elliptical distribution class, of which the normal is a special case. Significantly, LCE
is also consistent with the Pearson system of distributions. In turn, the skew elliptic
distributions, e.g. Genton et al. (2004), can be used to motivate solutions using the
mean-variance-skewness expected utility function. Chiu (2010) demonstrates such a
generalization is justified. In effect, restrictions on the subjective distributions of the
ex ante random variable(s) can be used to justify the selection of a particular expected
utility function for use in determining the optimal solution to problems of decision
making under “uncertainty.” As with other arguments based on Taylor series, further
properties of the expansion can raise significant complications.
More generally, continuing with the procedure used to determine the Taylor series
approximation for the mean-var-skewness case to include more terms produces:
EU[Wt +1 ] = U[Ω] − b var[W ] + c skew[W ] + d kurt[W ] + e σ5 + f σ6
where: σ5 and σ6 are the fifth and sixth central moments of the distribution for
terminal wealth; kurt[W] is the kurtosis, or fourth central moment; and d, e and f are
coefficients depending on the fourth, fifth and sixth utility function derivatives with
respect to wealth.16 Though seemingly arcane, introducing these additional terms
into the expansion raises substantive issues about using moment preference in approx-
imating expected utility. Recalling that an ex ante subjective distribution is used to
determine conditional expected utility, Poitras (2011) examines the limitations of the
ergodicity hypothesis to conclude that as the ex ante decision time interval increases,
the ensemble of future time paths for the random variables determining future wealth
could become more bi-modal than unimodal. The fifth and sixth central moments of
the distribution assume significance when the density is bimodal.
The immediate theoretical implication of bi-modality in the ex ante probabilities is
nonlinear dynamics for the mean of the process, e.g. Poitras (2011, Ch.5) gives numerous
references. The inability to reject bimodality of the ex ante time path is a consequence
of the fundamental quandary arising from the use of non-experimental data to make
decisions about the future, as is the case with expected utility modeling: there is only
one ex post time path available to estimate the theoretically infinite ensemble of possible
ex ante future time paths for the random variable(s) of interest, such as the spot price at
harvest for a farmer or metal price over the life of a mine for a miner. The implications of
nonlinear dynamics in the mean is complicated for conventional academics studies, e.g.
Wang and Tomek (2007). Various theoretical and empirical estimation approaches have
been introduced to deal with the failure of the normality assumption while still retaining
essential features of unimodality. For example, Poisson jump processes are added to mean-
reverting processes to capture manageable random changes in the mean, e.g. Bernard
et al. (2008), Bertus et al. (2009). Markov switching models and long memory models with
structural change can also be employed, e.g. Coakley et al. (2011). Some approaches, such
as the use of GARCH estimation methods, model changes in the variance rather than
the mean. The limitations of non-experimental data imply that problems of identifying a
“best methods” “scientific” approach may not be resolvable on empirical grounds.
where b (> 0) measures the sensitivity of expected utility to changes in risk. The optimal
speculative position for this objective function can now be identified.
To construct a mean-variance solution for the optimal speculative position, consider
the profit function for a futures or forward contract speculator who is either long the
actual (Q > 0) or short the actual (Q < 0):
p ( 1) = Q{F ( 1,T ) − F ( 0 ,T )}
Assuming that EU is depends only on the choice of Q, the resulting expectation and
variance of profit lead to the following:17
dEU
= {E[F (1,T )] − F (0 ,T )} − 2b Q σ2f = 0
dQ
E[F (1,T )] − F (0 ,T )
⇒ Q* =
2b σ2f
where b (> 0) is a parameter which measures the sensitivity of mean-variance expected
utility to changes in risk. The optimal speculative position size is seen to depend on three
elements: the expected change in the futures price; the conditional ex ante variance
of futures prices; and the speculator’s subjective sensitivity to risk. It is instructive to
consider the different solutions which are associated with varying these elements.
The solution depends on a combination of the trader’s attributes: subjective probability
assessments of the trader about future states of the world; the trader’s degree of risk
aversion; and the trader’s ability to forecast. Under certain conditions, the solution to the
speculative trader’s optimization problem can be aggregated to get implicit indications
about the nature of market equilibrium. Given this, if, in aggregate, speculators behave
as though they were risk neutral, then the speculators’ offer curve for Q in terms of
{F(1,T) - F(0,T)} would be (theoretically) infinite at E[F(1,T)]=F(0,T). To see this, observe
what conditions the numerator of the optimal speculative position must satisfy for Q
to be finite (required for markets to clear) and b goes to zero. If this were the case,
hedgers would not pay a risk premium to speculators in the form of a systematic bias
in the forecasting accuracy of the futures price. This is because speculators are already
willing to participate when the futures price is an unbiased forecast. Given that the
futures markets are designed to facilitate the participation of a wide range of traders, it
is possible that b may vary across market environments, from risk loving to risk neutral
to risk averse. Analysis of this situation could be explored by appropriate differentiation
of the optimal speculative solution with respect to b.
Given the solution to the optimal speculative position, it is now possible to develop a
solution to the optimal hedging problem. In most practical situations, the hedger is faced
with the question of what ratio of cash to futures positions should be selected. This can be
translated into questions about optimizing behavior. As for the speculator, optimality has to
be defined using the maximization of expected utility. This objective includes minimizing
the variance (risk) of the hedged position as an important special case of the slightly more
general mean-variance expected utility function. If risk is taken to be variance, then the
objective of minimizing risk can be reformulated in expected utility form as:
EU[p ] = −var[p ]
which is the general mean-variance objective function with E[π] = 0 and b = 1. This
variant of the mean-variance objective function can be used to address the issue of
whether hedgers are minimizers of risk or maximizers of expected utility (or both).
Over time, considerable academic attention has been given to the solution of the
fundamental question: what ratio of spot to futures positions is most appropriate to
maximize the expected utility of end-of-period profit? Early studies include Ederington
(1979), Heaney and Poitras (1991), Herbst et al. (1989), Hill and Schneeweis (1982),
Johnson (1960) and Toevs and Jacob (1986). Much of this research has focused on
estimating hedge ratios using an ordinary least squares (OLS) regression of spot prices
on futures prices; the “optimal” hedge ratio being the estimated slope coefficient. This
result can be derived from the stylized short (long) hedger trading profile in Table 2.13,
where short refers to the hedger’s position in futures. In addition to the Rahgozar and
Najafi (2003) and Chen et al. (2003) surveys, there are a number of recent approaches
that employ more advanced estimation methods, e.g. Bertus et al. (2009), Jin (2007),
Moschini and Myers (2002), Park and Jei (2010).
Given the variance of trade profit, the optimal hedge ratio follows by solving the first
order conditions for max EU (with EU = - var[π]) using QH as the choice variable:
dEU QH σSf σS
= 2QH σ2f − 2QS σSf = 0 ⇒ * = 2 = ρSf
dQH QS σf σf
where ρ is the correlation coefficient between S and F. A number of observations can be
made about this solution. Most importantly, it identifies the minimum variance hedge
ratio as the OLS slope coefficient in a bivariate regression of spot on futures prices.
This is the operational result that makes the minimum variance hedge ratio empirically
attractive and accounts for its widespread use. However, despite the popularity of
the approach, there are significant analytical limitations on its use and unanswered
questions about its validity. For example, one important limitation is the dependence
of the optimal solution on one choice variable, the size of the futures position. The size
of the cash position is taken as fixed and certain. No allowance is made for leveraging
to purchase the spot commodity or for hedging situations where the size of the cash
position is uncertain, e.g. the farmer who faces stochastic output. Before addressing
these issues, it is important to address unanswered questions about its validity: does
EU = -var correspond to optimal solutions for other, more theoretically plausible,
expected utility functions?
To see this, consider the optimal hedge ratio that is associated with max EU =
E[π] – b var[π]. Observing that for the short hedge E[π] = QS {E[S(1)] – S(0)} + QH
{F(0,T) – E[F(1,T)]}, the following problem and solution can be posed:
max
EU[π] = E[π] − b var[π]
QH
dEU
= (F (0 ,T ) − E [ F (1,T )) − b (2QH σ2f − 2QS σSf ) = 0
dQH
Q σSf
* F (0 ,T ) − E[F (1,T )]
⇒ H
= +
QS σ 2
f 2 b QS σ2f
The mean-variance optimal solution is composed of two parts: the minimum variance
hedge ratio and the optimal speculative position. While the minimum variance com
ponent depends on the ratio of statistical parameters, the speculative component depends
on the hedger’s risk attitudes as reflected in b. Hedgers who are “less risk averse” will
have lower b (ceteris paribus) and, as a result, will be more willing to take speculative
positions in the form of over or under hedges. In addition, because the futures price
variance enters in the numerator of the “speculative” term, as the perceived volatility
increases the hedger will be less willing to take positions over or under the minimum
variance hedge. More precisely, variances as well as expectations are conditional on the
information available on the hedge date. These values are derived from the subjective
probability assessments of the hedger. Hence, the less capable or willing the hedger is to
make forecasts, the less important is the speculative component of the hedge. Similarly,
if F(0,T) = E[F(0,T)] due to the willingness of speculators to provide sufficient liquidity
at such prices, the incentive to engage in “active” hedging – where speculation about
future commodity price behavior impacts the risk management decision – is severely
dampened.
incorporate the fact of cash flow variability across months into a decision about
whether and how much to hedge. When the time pattern of cash flows matters—as
it typically does for corporations looking to hedge—a smaller size hedge may be
preferable to a one-for-one rolling stack or other strategy that actually increases
initial cash flow variability.
In effect, MGRM used a transaction hedging approach when the situation was
theoretically better suited to using an optimal hedge due to the substantial amount of
basis risk facing MGRM. However, how an optimal hedge would have been estimated
in the MGRM case is not obvious.
The airline industry provides another illustration of the contrast between an optimal
hedge and a transactions hedge. Based on information contained in annual reports, this
industry is characterized by a range of risk management activities using both exchange
traded and OTC derivative securities to manage jet fuel price risk. Some firms, e.g.
Thai Airways, engage in only limited use of derivatives, e.g. for hedging specific FX
transactions. Other firms, e.g. Singapore Airlines, have a sophisticated program that
covers the range of market risks, especially the most volatile cost component – jet fuel
prices. A transactions hedge of this commodity price risk would fully hedge the quantity
of jet fuel to be purchased over the planning horizon. For example, based on estimated
passenger volume and other factors, the airlines will quote ticket prices for flights up
to one year in advance. An optimal hedge would attempt to estimate the impact of
changes in jet fuel prices on changes in firm net cash flows. In Part 3, industry practice
among airlines using derivative security strategies is determined by examining airline
company financial statements to find what fraction of jet fuel purchases are hedged
(see Sec. 3.3).
Grain elevators usually purchase grain from farmers with cash purchases or
forward contracts which set a specified date in the future for the delivery of the
commodity . . . Many grain elevators, particularly co-operatives owned by farmers,
also sell seed, fertilizer and other items that farmers need. In order to protect
themselves from the risk of falling crop prices, elevators usually hedge their cash or
forward purchases by entering into futures contracts on the futures exchanges to
sell the grain at a price they expect will cover their expenses. Grain elevators that
possess grain in storage are said to be “long” in the cash market; when they enter
into futures contracts to sell that grain in a future month, they are said to be “short”
in the futures market. Once the purchase of a cash crop is hedged with a futures
contract, any decline in the value of the crop in the cash market should be offset
with a gain in the futures market.
In contrast to the profit function described in Table 2.14, grain elevators are averse to
taking unhedged positions for grain in storage.
Application of the definition for conditional variance (where the conditioning
notation has been dropped for convenience), var[ ⋅ ] = E{(π – E[π])2} provides the result
that the conditional variance for an unhedged spot position is:
In the case of unhedged cash positions, risk depends on the size of the position and the
volatility of the cash price. Despite some stylized textbook treatments to the contrary,
hedging does not typically eliminate all the risk of cash price fluctuations. To see this,
recall the profit function for the one-to-one grain elevator hedge from Sec. 2.2.A:
π = Q {F ( 0 ,T ) − S( 0 )} − {F ( 1,T ) − S( 1)}
It was remarked that the profitability of the hedged position depends on the change in
the basis. This discussion can now be extended to observe that the conditional variance
for the hedged position is:
In other words, a transaction hedge substitutes basis risk for price risk.
to profit from storage. The merchant seeks to profit from changes in the basis
relationship, rather than price level changes.
Operational Hedge: An operational hedge facilitates merchandising or process-
ing operations. A merchant hedges to establish the price of an input or
output, often ignoring changes in the basis. The hedge protects the merchant
against rapid change in price while a product is being processed or transported.
Typical examples of an operational hedge include a flour miller, buying wheat
futures to offset a forward sales contract of flour to a baker, or a shipper, exporter
or importer selling futures against a cash purchase. These hedges act as temporary
substitutes and are liquidated as soon as the trader takes a corresponding cash
position.
Active or Selective Hedge adjusts the hedge according to price expectations. The
holder of the commodity adjusts the hedge if prices are expected to fall or rise.
Selective hedging introduces an additional speculative element to hedging as
traders hedge position changes with price expectations. This common hedging
procedure is often done to prevent large losses, or variation margin difficulties.
Such hedging may be related to an optimal hedging strategy.
Anticipatory Hedge: An anticipatory hedge is usually not matched or offset by a
contemporaneous goods or merchandising commitment. The anticipatory hedge
serves as a temporary substitute for merchandising to be done later, that is, an
expected future cash purchase or sale. The anticipatory hedge involves either the
purchase of futures contracts against raw material requirements, or the sale of
contracts by producers in advance of the completion of production. For example,
flour millers and soybean processors may buy futures contracts in anticipation of
subsequent purchases of wheat and soybeans, respectively. Livestock feeders may
sell live cattle and live hog futures long before the animals are ready for market.
Similarly, grain farmers forward sell their crops before harvest . . .
Cross Hedge involves a derivative security position opposite an existing cash
position, but in a different commodity. Typically, there is no active futures
contract in the commodity corresponding to the cash position, so the trader must
select a related commodity for hedging. To be effective, the prices and commodity
values of the cash commodity and futures contract must have a fairly high positive
correlation. Examples include hedging corporate bonds in the Treasury bond
market, grain sorghum in corn, and boneless beef in live cattle.
example, a refiner seeking to hedge the price of purchasing copper scrap using the high
grade copper contract will want to know the basis relationship between the grade of
refined copper being produced from the scrap as well as the approximate amount
of refined copper which can be produced per unit of scrap. With this information,
appropriate adjustments can be made to Q. When will the hedge completely eliminate
price risk? This will occur when F(1,T) = S(1). Accomplishing this result requires a delivery
hedge where the commodity being hedged is the same as the commodity specified in
the futures contract and T = 1. In this case, F(1,1) = S(1) and the futures position is
satisfied by delivery of the commodity. This result is typically easier to achieve with the
use of forward contracts, which can be tailored to match the size, grade, location and
other factors that can impact the basis relationship.
another example where a hedge using related commodities that are available can be
constructed with combinations of option contracts.
The objective of increasing the number of different derivative security contracts used
in the hedge is, ultimately, to improve hedge performance. At some point, this will
be a fruitless exercise because there would be so many hedging contracts to monitor
and transactions costs would increase accordingly. The obvious question is: how
to optimally construct “multivariate hedges” using combinations of derivative sec
urities? Theoretically, it would be most appropriate to develop an equilibrium model
explaining the relationship between the commodity being hedged and the commodities
underlying the contracts being used to construct the hedge. The approach used here
is to assume that such a model has been specified and to work with the general profit
function for a multivariate hedge position. For simplicity it is assumed that only futures
contracts are being used to construct the hedge of a commodity position.
Two problems will be considered, the minimum variance hedge and the mean-
variance optimal speculative solution. These two solutions will then be combined
to produce the mean-variance optimal multivariate hedge solution. As with the
univariate hedge solutions, it is assumed that there is one random variable to be hedged,
the price risk of the cash commodity position. The size of the cash position is fixed
and the relevant components of perfect markets are adopted. Given this, consider
the general profit function for a hedge involving k futures contracts and a fixed cash
position:
π(1) = QS (S1 − S0 )
where the bar on Qs indicates that the size of the cash position is not a choice variable.
Like the previous univariate hedge where it was assumed that the cash position
was long and that the futures position was short when Q > 0, this formulation also
permits futures to be short or long with the Qi > 0 solution denoting a short futures
position matched with a long cash position and Qi < 0 denoting a long futures position
matched with a long cash position. It is possible for some of the ΔF to represent option
prices.
The key step in transforming the general profit function into a form suitable for
estimating a minimum variance solution is to reformulate the problem in vector space
form:
y = X β + u
where y = Qs ΔS, X = (1, ΔF1, ΔF2, . . . .,ΔFk), β = (α, Q1, Q2, . . . .,Qk) and u is an equation
error that captures the unexplained variation in y not accounted for by Xβ and α is the
β = ( X T X )−1 X T y
and the individual σij refer to variances and covariances for changes in the relevant
futures prices. It follows that:
Q*
Qmv = = ∑ −1 cov[F , s]
QS
This result is identical to the estimation solution, with the proviso that the equilibrium
solution involves ex ante conditional population parameters while the estimated
solution involves ex post estimates of the population parameters determined using T ex
post observations of data on the ΔFi and ΔS.
As an example of the multivariate solution, consider the case of a minimum variance
hedge using two futures contracts. In this case:
1 σ2 −σ12 σ s ,1
2
−1
∑ = X yT = QS
det −σ12 σ12 σs ,2
These results are intuitively the same as those given previously, with the proviso that the
sign of β has positive coefficients reflecting short hedge positions combined with long
cash positions and negative coefficients representing long futures combined with a long
cash position.
As in the univariate hedge case, the optimal mean-variance solution is a combination
of the minimum variance solution and the mean-variance optimal speculative solution.
To derive the optimal speculative solution, observe that the objective is to maximize
expected utility which, in this case, is specified using the mean and variance of
speculative profit:
E[∆F1 ]
E[∆F ]
2
= QT . − b{QT ∑ −1 Q}
.
.
Differentiating the objective with respect to the choice variables QT produces the
solution:
Q1* E[∆F1 ]
Q * E[∆F ]
2 1 2
Q* = . = ∑ .
−1
2b
. .
. .
where * denotes an optimum value.
Solving the general mean-variance optimal speculative solution for the case of two
futures positions gives:
where det is the same as that given in the minimum variance hedge example given
above. Solving for the specific case of Q1*:
where, using the definitions from the previous derivation of the equilibrium minimum
variance solution:
where, as before, the * indicates an optimum value. Solving Q1* involves determining
the solution of the other k-1 positions.
Proceeding to determine all k derivatives and expressing the solution in matrix
form:
Q* E[∆F ]
∑Q = −
2b Q s
+ cov[F , s]
s
= −
Qs σ22 σ12 − σ1,22 2b Qs (σ12 σ22 − σ12,2 )
From this it follows that if the different futures contracts involved in the hedge have
price changes that are uncorrelated (σ1,2 = 0), then the individual hedge ratios will be
equal to the univariate solutions. The dependence of the optimal solution on the utility
parameter b casts doubt on the general validity of the numerous ex post empirical studies
that have taken the minimum variance solution to be the optimal solution.
The two period structure of the profit function leads to a number of other qualifications
about the generality of the optimal mean-variance solution. Extending to a multi-period
framework raises the possibility of readjusting the hedge position over time, producing
a different hedge ratio at each hedge readjustment date. In addition, the solution to an
inter-temporal, multi-period optimization problem will not necessarily produce the
decomposition of the hedge ratio into the minimum variance and optimal speculative
solution (Heaney and Poitras 1991). Among other reasons, this is because as prices evolve
over the life of the hedge, risk propensities will affect the desire to adjust to observed data.
Wt +1 = Wt ( 1 + Rs ( t + 1) − h1 R f ( t + 1))
where: ht is defined to be the ratio of the values (price times quantity) of the spot and
futures positions at time t, Rs(t+1) and Rf(t+1) are the t to t+1 returns to holding spot
and futures and t ε [0. . ..T], i.e. RJ(t+1) = (Jt+1 - Jt)/Jt where J is either the spot or futures
price at t and t+1. By construction, this specification for the wealth dynamics restricts
the problem in order to derive implementable solutions. This particular wealth dynamic
assumes a single period decision framework with no potential for variation in the
quantity of the spot commodity being held.
Another significant feature of this approach to wealth dynamics is the absence of
portfolio theoretic considerations. In particular, by not incorporating lending and
borrowing into the specification of the profit function, the size of the cash position
has been fixed. In effect, the resulting hedging optimization assumes away the portfolio
decision by having the hedger fully invested in the spot commodity. A number of
sources argue that “farmers can manage their risk exposure through adjusting their
leverage, obviating the need for hedging instruments,” e.g. Pannell et al. (2008),
Simmons (2002). If lending and borrowing is admitted, considerations of leveraging to
buy the spot commodity and short-selling the spot to invest in the riskless asset enter
the hedger’s decision process. Theoretically, this is translated to the leveraged wealth
dynamics:
Wt +1 = Wt ( 1 + xt Rs ( t + 1) + ( 1 − xt )r( t + 1) − H t R f ( t + 1))
where: x is the fraction of total wealth invested in the spot commodity, H is the value
(price times quantity) of the hedge position divided by initial wealth (not the value of
the spot position) and r is the riskless rate. In turn, the leveraged wealth dynamics can
be used as the argument in the hedger’s optimization problem. In practice, the primary
advantage of using the basic wealth dynamics specification over the leveraged dynamics
specification is analytical simplification: the optimal hedge ratio requires only a joint
probability distribution and a utility function. The addition of lending and borrowing
results in the introduction of an additional choice variable.
Given this, the conventional hedge ratio optimization problem can be generalized
to admit any type of well-behaved utility function, Using basic wealth dynamics, the
optimization problem can be specified:
where the conditional expectation E[ ⋅ ] has been formally defined using an appropriately
specified conditioning information set. In this form, the joint conditional probability
distribution for S(t) and F(t) associated with the expectation is Ψ[S ,F] and the profit
function is either π(t+1) = Wt (Rs(t+1) - htRf(t+1)) or, allowing leveraging, π(t+1) = Wt
(xt Rs(t+1) + (1-xt) r(t+1) - Ht Rf(t+1)). In practice, there are restrictions on the types of
commodities for which the EU optimization using the basic wealth dynamics is the
appropriate hedging problem. For example, because no allowance has been taken of
unexpected variation in quantity of the spot commodity, the hedged portfolio would
not fully capture the wealth dynamics associated with many harvestable crops.
In terms of solutions, under some strong distributional assumptions, the minimum
variance solution leads to an optimal hedge ratio which equals the slope coefficient in
an ordinary least squares (OLS) regression of spot on futures prices, e.g. Lien (2005).
By construction, OLS depends fundamentally on the selection of joint probability
distributions which are constant over time. This assumption results in equality of
conditional and unconditional parameters. When time variation in the joint probability
distributions is permitted, e.g. due to ARCH errors, the decision problem can be more
complicated. In this case, specification of the optimal hedging problem typically takes
on a more complicated form and has to be solved using some dynamic optimization
procedure, e.g. dynamic programming, which takes account of the state variable time
paths. The resulting solutions are potentially intractable and difficult to interpret.
However, in the special case of log utility, the dynamic solution will reduce to a sequence
of one-period solutions (Samuelson 1969). This important simplification permits the
introduction of certain types of temporal variation in the conditional variances and
covariances without significantly complicating the solution.
In addition to complications arising from non-constant distributional parameters,
when the structure of the optimal hedging problem is altered by the introduction of
riskless lending and borrowing, variation in the size of the spot position means that the
hedge ratio cannot be determined by choosing the relative size of the futures position.
There are now two choice variables, the fraction of initial wealth invested in (borrowed
using) the riskless asset and the size of the hedged position. Again, while there has been
explicit recognition of riskless lending and borrowing, analysis has been restricted to
special cases, particularly mean-variance (e.g. Bond and Thompson 1986, Turvey and
Baker 1989). In certain special cases (e.g. Poitras 1989), the resulting optimal hedge ratio
has been shown to be independent of hedger risk preferences, depending solely on the
parameters of the (un)conditional joint distribution of returns.
Given this background, it is possible to derive two propositions corresponding to the
two different formulations of the “myopic” optimal hedge ratio problem, where myopia
is a direct consequence of the single period specification of the optimization problem.
The first formulation is based on the conventional approach which omits lending and
borrowing from the portfolio decision, where the basic wealth dynamics are the basis
of the objective function. The second approach uses leveraged wealth dynamics thereby
admitting riskless lending and borrowing. In this analysis, “myopia” dictates that
future time paths of the conditioning variables are ignored; the trade is initiated at
time t and profits are taken at t +1. This permits use of the unconditional distributions.
Given this, Proposition 2.1 extends the conventional constant distributional parameter
solution to include a general expected utility function. It is shown that the optimal
hedge ratio can be decomposed into the OLS-based hedge ratio (hOLS) and a utility
function dependent term. Proposition 2.2 incorporates riskless lending and borrowing
to determine a market equilibrium hedge ratio which is shown to be independent of the
utility function selected.
Conventionally, hOLS has been the foundation of empirical estimation of hedge
ratios. Hence, it is important to know the relationship between specific solutions to the
expected utility optimization problem and the associated OLS estimate. More precisely,
the linkage between the optimization problem and the minimum variance hedge ratio
is given by the following Propositions (see Heaney and Poitras 1991 for proofs).
h = hOLS +
/
var[R f ] Wt E[U (⋅ )]
//
In words, Proposition 2.1 demonstrates that, for myopic agents, the optimal hedge
ratio can always be decomposed into a sum of the OLS-based hedge ratio and an
additional term that is fully determined by statistical parameters and the risk aversion
propensity of the selected utility function.
The primary upshot of Proposition 2.1 is that in addition to hOLS consideration must
be given to the variance-deflated expected return on the futures position. When the
expected return is non-zero, the properties of the particular expected utility function
assumed, i.e. the inverse of the coefficient of relative risk aversion, takes on importance.
Examining the effect of the statistical parameters, an important general corollary
follows: when the current futures price is an unbiased predictor of the distant futures
price (E[Rf] = 0) or when var[Rf] → ∞ , hOLS is optimal. Hence, results that apply for
specific utility functions (e.g. Benninga et al. 1984; Poitras 1989) can be generalized
to any type of admissible utility function, albeit under the restriction of bivariate
normality. However, for many commodities, E[Rf] = 0 or var[Rf] → ∞ does not always
hold ex ante in which case the issue of selecting an appropriate utility function is raised.
For the important specific power utility case of log utility, U = ln(π), the solution reduces to:
E[R f ] E[π−1 ]
h*ln = hOLS −
var[R f ] Wt E[π−2 ]
From these results it follows that a given optimal hedge ratio depends on parameters of
both the conditional distribution and the expected utility function.
Significantly, Proposition 2.1 demonstrates that, when E[Rf] ≠ 0 and var[Rf] < ∞, it is
not “optimal” to use OLS hedge ratios without making further assumptions about the
return and profit generating processes and the form of expected utility. In practice, given
specific distributions for the relevant processes, the h* of interest can be approximated
with quasi-maximum likelihood estimation or other numerical methods, e.g. Haigh
and Holt (2002), Moschini and Myers (2002). Following Baillie and Myers (1991), it is
also possible to empirically model and estimate the conditional hedge ratio, e.g. Chang
et al. (2011), Haigh and Holt (2000), Kroner and Sultan (1993), Park and Jei (2010).
However, such applications typically result in a substantive increase in the complexity
of the estimation problem. Unfortunately, it is not possible to establish, theoretically,
whether there will be corresponding increases in the value of the resulting hedge ratio
estimates. For example, Alexander et al. (2012) observe:
The value-added of more direct specification of the both the expected utility function
and return generating processes is still largely an unresolved empirical issue.
Within the myopic model the introduction of riskless lending and borrowing alters
the objective function such that the leveraged wealth dynamics process is used to
specify the optimization problem. The general importance of leveraging and financial
constraints in the risk management decision is gradually being recognized in academic
studies, e.g. Adam et al. (2007), Haushalter et al. (2007). Solution of the stylized
optimization problem with leveraging leads to Proposition 2.2, which holds for the case
of bivariate normality of the spot and futures price. Because this result depends only
on parameters of the joint (un)conditional distribution, the optimal hedge ratio with
riskless borrowing and lending is independent of both the specification of the expected
utility function and initial wealth. Such separation results are relatively rare and require
investigation.
While not immediately apparent, the relationship between Propositions 2.1 and 2.2
can be seen by evaluating H* in Proposition 2.2 where E[Rf] = 0. In this case, H* = β[S,F]
= hOLS. Significantly, as is the case without lending and borrowing, hOLS is optimal when
the current futures price is an unbiased predictor of the distant futures price. Hence,
the introduction of lending and borrowing into the hedger’s optimization problem does
not alter the general result that the optimal hedge ratio is decomposable into hOLS and
another term which depends on statistical parameters. However, admitting the ability
to short-sell and leverage eliminates the need to consider the risk aversion properties
of the selected utility function. The upshot is that, in practice, optimization problems
based on Proposition 2.2 may produce more implementable solutions than those based
on Proposition 2.1.
On a more critical level, as demonstrated by Poitras (1988), Kroner and Sultan (1993),
Lien (2009a) and many others, hedge ratio estimation is complicated because both the
means and the variance-covariance matrix of the relevant variables are not typically
constant through time as required for ordinary least squares (OLS) regression. In
addition, because the minimum variance hedge ratio does not take the mean return on
the hedge portfolio into account, regression-based hedge ratios will not necessarily be
optimal for all types of hedger expected utility functions. Improved ex post estimation
of ex ante hedge ratios suggests the use of sophisticated estimation procedures such
as GARCH, e.g. Lien (2009a). However, straightforward implementation of ARCH
methods assumes that only minimum variance solutions are relevant to the hedger.
In an intertemporal model which permits changing means and variances, Heaney and
Poitras (1991) demonstrate that minimum variance techniques will only be valid for
log utility. Lee and Yorder (2007) rationalize minimum variance by assuming that the
prices of futures contracts used in the hedge follow a martingale. Empirical evidence
indicating that introducing more advanced estimation methods, such as the Markov
switching techniques used by Lee (2009), substantially improves the practical hedge
performance over the use of OLS-based hedge ratio estimates is mixed. Similarly, while
it is possible to incorporate information on the mean returns of the hedge portfolio into
the estimation procedure, this also leads to considerably more complicated estimation
procedures.
Fig. 2.6 provides an example from Moschini and Myers (2002) comparing OLS and
GARCH hedge ratios for corn. The conventional unconditional hedge ratio is the OLS
estimate, which is constant over the sample at 0.9. The conditional hedge is a GARCH
estimate of the hedge ratio. The relevance of conditional volatility models such as ARCH
and GARCH for hedge ratio estimation has been examined in numerous sources going
back at least to Hsieh (1988). Because a bivariate regression hedge ratio is determined
as the ratio of a covariance between the dependent and independent variables divided
by the variance of the independent variable, if these parameters change conditionally
as time evolves, the associated hedge ratio will also change. This is the basis for the
conditional hedge which is reported in Fig. 2.6. Unfortunately, evaluation of the
relative performance of the conditional volatility and OLS hedge ratios is difficult.
Figure 2.6 Conditional versus Unconditional Hedge Ratio for Corn, January 1976 to June
1997
More importantly, such a performance evaluation is largely irrelevant to the actual risk
management practices of non-financial firms.
The numerous academic studies that recommend the use of more sophisticated
estimation methods than OLS to determine hedge ratios typically employ statistical
criteria such as minimum mean squared error in assessing performance. Following
Leitch and Tanner (1991) comparing different hedge ratios using statistical criteria,
e.g. by examining the in-sample or out-of-sample minimum mean square forecast
error, fails to account for the profitability of the hedge. Significantly, for both corn and
currencies, estimates for the OLS hedge ratio in Figure 2.6 is close to one and conditional
hedge ratios fluctuate around one providing some support for the use of full hedges,
as opposed to no hedging. While attention focuses almost exclusively on the “fit” of
the estimate, the level of hedge ratio estimates goes largely unmentioned. In particular,
while hedge ratio estimates—OLS and otherwise—often indicated close to full hedging
is optimal, such estimates do not correspond to observed hedging behavior. In many
cases, risk management directives from the Board of Directors prevent hedging beyond
some fraction of total exposure, presumably to reduce the possibility of corporate
speculation.
∂2 J ∂2 J
E | X ( t ) ≡ E[ JWs ] E | X ( t ) ≡ E[ JWf ]
∂Wt +1 ∂Rs ,t +1 ∂Wt +1∂R f ,t +1
In addition to hols now being a conditional estimate, the role of preferences in the
intertemporal optimal hedge ratio is more complicated than in the myopic case. More
significantly, hols is no longer generally optimal when E[Rf] = 0.
Specifically, the additional preference-dependent terms arise from expected changes
in the state variables affecting the marginal utility of wealth. In this situation, utility
function selection takes on added importance. For example, log utility posses the
important simplifying property that Jws = Jwf = 0, which allows the intertemporal solution
to correspond directly to the myopic case, with the caveat of potential inequality of
conditional and unconditional statistical parameters. Empirical estimation proceeds
by assuming a specific form of conditional distribution, e.g. GARCH (Moschini and
Myers 2002; see Fig. 2.6). Significantly, other important types of utility functions such
as quadratic and power are not so well behaved. Estimation of intertemporal optimal
hedge ratios for these types of functions can be problematic.
Similar complications arise when riskless lending and borrowing is admitted. In this
case, the wealth specification is:
The first point to appreciate is that all sensible approaches [to risk management]
have the same first step, i.e. we formulate a corporate risk management philosophy
to impose some guidelines on risk management decision-making. This tells us what
kinds of risks we wish to bear, what risks we want to avoid, what sort of options
we will consider to manage our risks, and so forth. Usually, we will readily bear
those risks that we have some particular expertise in handling (e.g. risk unique
to our particular line of business), but there will also be other risks that we will
usually wish to avoid (e.g. the risk of our factory burning down). This philosophy
217
should also give us some indication of what attitude we should take towards the
many other types of risks that we might face—when we should bear them, when we
should not bear them and the like.
Strategic risk management and enterprise risk management are terms often used to
describe the process of formulating and implementing a corporate risk management
philosophy.1 Though these terms are often used synonymously, it is possible to define
distinct notions. Making this distinction identifies a vital step in the commodity risk
management process that receives relatively little attention in conventional academic
studies that tend to focus on risk measurement and techniques of hedging, diversifying
and insuring.
The multifaceted importance of the risk management function has led to the
creation of corporate risk management offices run by “risk management officers.”
Mikes (2008) examines the role of chief risk offices within the corporation. Such
offices engage in a whole range of activities which are of little or no direct relevance to
commodity risk management. Monitoring of Occupational, Health and Safety rules,
internal security, handling of various insurance plans for fire, theft and the like, these
types of activities could be localized in the risk management office. Depending upon
the specific corporation, it is possible that the risk management officer may have little
or no financial expertise. Commodity risk management decisions could be made by
marketing, purchasing and treasury departments. By design, narrowing the focus to
corporate commodity risk management abstracts from the integrated risk management
process. This runs the risk of failing to make an adequate connection between the
implementation of a risk management program and the overall risk management
philosophy of the firm. Arguably, the failure to make such a connection has contributed
to a number of recent and not so recent debacles, e.g. the Metallgesellschaft losses.
“Enterprise risk management” is now a popular, arguably essential, concept in
discussions of corporate risk management practice, e.g. Hampton (2009); Moeller (2007).
Unfortunately, this terminology can cover a range of possible notions, creating some
semantic confusions. Recognizing lack of precision in these concepts, various notions
of “strategic risk management” and “enterprise risk management” have been adopted
enthusiastically by numerous non-financial corporations. For present purposes, a rough
description of enterprise risk management applicable to all the various notions can be
formulated as:
This definition of enterprise risk management is similar to that given by the Casualty
Actuarial Society: “the discipline by which an organization in any industry assesses,
controls, exploits, finances and monitors risk from all sources for the purposes of
increasing the organization’s short- and long-term value to its stakeholders.” The goal
of enterprise risk management is to deal with the risks facing the firm in a systematic
and enterprise-wide fashion, instead of relying on the ad hoc and independent risk
management functions which often characterize traditional firm activities surrounding
risk.
Enterprise risk management is a multi-faceted concept. For airline companies, a
leading concern of enterprise risk management is catastrophic risk of plane crashes and
other repair and maintenance issues. Similarly, enterprise risk management for open
pit mining companies will be concerned with adherence to environmental standards.
While commodity risk management is only one part of the larger task of enterprise risk
management, the resulting decision have fundamental strategic implications for actual
firm performance. To recognize this key aspect of enterprise risk management, the
concept of “strategic risk management” can be introduced to narrow the focus to only
include market risk that impacts on the business strategy of the firm such as the ability
to: finance future projects and acquisitions; maintain or increase the level of dividend
payments; improve firm profitability; and reduce the possibility of bankruptcy. As such,
strategic risk management is squarely aligned with management of business risks facing
corporations and other firms operating in commodity producing and consuming
industries.
The importance of the drive to strategically manage risks has not been lost on the
management consulting industry. Yet, it is the perceived need for an enterprise risk
management function that has led to the major players in that industry to develop
programs for implementing the appropriate “business organization and management
structures, geographic, regulatory and reporting matrices, and the mandates which
underwrite these.” The now defunct accounting firm Arthur Andersen (1998, p.9) gives
the following description of enterprise risk management:
philosophy for the firm. This requires an initial evaluation of the range of risks facing
the firm. Decisions are then made about the exposures the firm wants to manage and
what types of systems will be used to manage those risks. This step in the strategic risk
management process is referred to as developing a philosophy because there is much
that is subjective and intuitive for non-financial firms. The correct method of identifica-
tion and handling of risks is not obvious. Loosely put, philosophy has to do with ways
of looking at the world. What risks are relevant and how these risks are to be handled
depends on the managers’ view of the world. This stage in the process is top-down, with
senior management being an integral part of the decision making process in corporate
situations. It is likely that those senior managers responsible for risk management will
be an essential cog in the process, due to potentially limited knowledge about specific
risk management matters by those at the most senior levels of the firm.
A number of academic studies, many originating from the strategic management
area, e.g. Ahn and Falloon (1991), Andersen (2006), Andersen and Schroder (2010),
Damodaran (2008), Oxelheim and Wihlborg (1997), Rawls and Smithson (1990), identify
a specific type of risk management philosophy with “strategic risk management.” This
essence of this approach can be illustrated by examples. Consider Gallo Wines, a
company that produces and sells the bulk of its outputs in the USA. Cash flows and
assets are denominated primarily in US dollars. Does this firm need to manage FX
risks arising from changes in the US dollar? From a transactions hedging perspective,
seemingly no; but if it is recognized that the major competitors for Gallo are situated
offshore with cost structures denominated in foreign currency, then Gallo’s exposure
to FX changes becomes apparent. What about the range of other macroeconomic risks?
Oxelheim and Wihlborg (1997) examine the issues surrounding the management of
macroeconomic risks, and do a detailed analysis of Volvo Cars. Shapiro (2010) gives
numerous other examples, from US ski resorts to Monsanto. These “soft” risks can be
contrasted with the “hard” risks arising from pure financial decisions, such as funding
of debt or investing in marketable securities. These types of issues can be considered as
the conceptual aspect of the identification phase of strategic risk management.
The process of formulating a risk management philosophy also involves an empirical
aspect. The conceptual aspect requires detailed empirical data about the various risks
facing the firm. These data have to be collected, processed and evaluated. Decisions have
to be made about which variables to include, the relevant sample periods to examine and
the types of technique to use in measuring and evaluating the risks. There is feedback
between the conceptual and empirical aspects. In a corporate context, whereas senior
management is primarily involved in the conceptual aspect, the empirical aspect has to
have wider involvement, with data inputs being collected and processed in the various
risk management units within the firm. Once the basic empirical results have been
obtained, decisions have to be made about the appropriate risk management techniques
to use for managing the risks. The data may require a fresh look at the firm’s approach to
risk management, a rethinking of the conceptual aspect, and a retooling of the empirical
aspect.
Judging from the commodity risk management problems at various firms, e.g.
Metallgesellschaft, China Aviation Oil, the Chinese State Reserve Bureau, Amaranth
Advisors, the costs of ignoring the implementation phase of strategic risk management
can be considerable. The first step of the implementation process is to determine the
relevant chain of command, ensuring that each level in the chain understands the
risk management philosophy and subscribes to it. Implementation also requires
putting decision making systems in place to adequately manage risk. For non-financial
firms, there have been some efforts to apply value at risk techniques, e.g. Godfrey and
Espinosa (1998). Others, such as Culp, Miller and Neves (1998, p.34), suggest that non-
financial firms are more concerned with cash flow volatility than financial firms. In
such situations, firms “are better off eschewing VaR altogether in favor of a measure of
cash flow volatility.” With all these competing, ad hoc approaches to risk management
on the landscape, conceptual guidance is needed.
[T]here is generally no difference in firm values between firms that hedge and firms
that do not hedge.
Jin and Jorion (2004)
[R]isk management can add value when revenues and costs are nonlinearly
related to prices . . . For a sample of 34 oil refiners, we find that hedging concave
revenues and leaving concave costs exposed each represent between 2% and 3% of
firm value.
Mackay and Moeller (2007)
The problem of risk management for the corporation has been well studied using
techniques adapted from traditional corporate finance where managers act as agents
for the owners of the firm, the common stockholders. Following conventions from
traditional corporate finance, the appropriate primary objective is to maximize the
expected utility of the end-of-period wealth of stockholders. Achieving this objective is
complicated by the inability of managers to observe the expected utility functions of
individual shareholders leading to potential agency problems. Yet, under reasonable
conditions, the primary corporate objective for a non-dividend paying firm can
be reformulated as maximizing the long run value of the firm’s common stock. Given
Keynes’s observations about the formation of prices in stock markets, this objective
is not without difficulties, e.g. (Poitras 1994). In addition, in ignoring the relevance of
dividend payout restrictions, a potentially important motive for risk management is
not captured. Proceeding on the assumption that long run common stock prices will
correctly reflect firm value, the market value of the firm can be determined as the
sum of the net present values (NPV) of the firm’s ventures. Observing the variables
in the fundamental NPV equation, it follows that NPV increases due to corporate risk
management activities can arise from: reductions in discount rates; increases in net
cash flows; and increases in the real option value of projects.
Corporate managers facing exposure to commodity risks must address a natural
question: when are risk management actions such as hedging commodity price risk using
derivative securities consistent with the primary corporate objective? Consistent with
the observed behavior of non-financial firms that do not hedge commodity price risk,
a number of persuasive arguments have been made against hedging commodity price
risk and other such market risks, e.g. corporate foreign exchange (FX) risk. Academic
arguments against hedging commodity price risk are specific applications of more
general arguments which claim, in general, that using derivative securities for corporate
risk management activities will not be value enhancing for common stockholders. When
such risk management activities are too costly to implement, monitor and execute, firms
are generally recommended to forego the direct use of derivative securities to manage
commodity risk. In such situations, risk can be managed by indirect methods such
as: holding significant cash balances; allowing other entities to forward contract, e.g.
small farmers selling to cooperatives; or embedding forward contracting provisions in
marketing contracts, e.g. metallurgical coal producers contracting annually with steel
mills to set forward delivery prices.
The general content of these academic arguments is that, in perfect markets, the role
of derivative securities in the risk management policy of the firm is irrelevant to the
market valuation of the common stock, e.g. Siegel and Siegel (1990, pp.146–9). Such
early theoretical arguments provide the foundation for subsequent empirical studies
that explore implications of specific perfect market assumptions, e.g. bankruptcy risk,
agency costs, taxes, financial constraints. More recent studies generally support the
use of derivative securities to manage commodity price risk, e.g. Aretz et al. (2007),
Servaes et al. (2009), Smithson and Simpkins (2005). In addition, some recent studies
have progressed beyond questions related to whether risk management is relevant to
examine the interaction between corporate strategy, which is impacted by product
market dynamics, and risk management decisions, e.g. Adam et al. (2007), Haushalter
et al. (2007). In the process, the connection between risk management and the level of
cash holdings has been recognized.
In earlier studies, a variety of theoretical arguments were proposed that attempted to
demonstrate that the hedging policy or, more generally, the use of derivative securities
to manage firm risks, is irrelevant, e.g. Dufey and Srinivasulu (1983), Levi and Sercu
(1991). Are the arguments that belie the importance of such corporate risk management
activities correct? To determine the answer to this question, it is helpful to classify
the important arguments against such corporate risk management into the following
groups: perceived shareholder preference; Modigliani-Miller (MM); CAPM (capital
asset pricing model); and market efficiency (expected value).2 There is a complementarity
among irrelevance arguments that depend on perfect markets assumptions. Because,
in practice, the use of derivative securities involves an expenditure of firm resources
which would not be required if derivatives were not used, it is argued that the use of
derivatives is impractical if that use is not value enhancing. The firm is better off not
using derivative securities at all.
Shareholder preference arguments against the use of commodity price risk
management arise in many practical situations. An excellent example is Barrick Gold
(see Sec. 3.1.B), once a leading advocate of gold price hedging, which now states (Barrick
Annual Report, 2010):
Barrick’s revenues are primarily derived from the sale of gold and the market
price of gold can fluctuate widely due to macroeconomic factors that are beyond
our control. Consequently, the market price of gold is one of the most significant
factors in determining the profitability of our operations. All of our future gold
production is unhedged, providing full leverage to changes in the market gold
price. To maximize our realized gold price, we have a corporate treasury function
which monitors the gold market and is responsible for our gold sales.
Significantly, Barrick is involved in managing interest rate and FX risk, but not
commodity price risk. Canadian Oil Sands (see Sec. 3.2.A), which receives a relatively
predictable amount of synthetic crude oil at the Syncrude plant gate reflects a similar
view, but does identify a possible situation where hedging of commodity price risk
would be considered (Annual Report, 2010):
Canadian Oil Sands prefers to remain un-hedged on crude oil prices; however,
during periods of significant capital spending and financing requirements,
management has in the past, and may again, hedge prices and exchange rates to
reduce revenue and cash flow volatility to the Corporation. Canadian Oil Sands
did not have any crude oil price hedges in place for 2010 or 2009. Instead, a strong
balance sheet was used to mitigate the risk around crude oil price movements. As
at February 23, 2011, and based on current expectations, the Corporation remains
un-hedged on its crude oil price exposure.
COS also does not hedge FX risk. The often stated rationale for such risk management
inaction is that shareholders want exposure to commodity price risk. Hedging such
exposure would frustrate the objectives of shareholders. Presumably such perceptions
are filtering through from the Board of Directors where significant shareholdings are
usually represented.
In summarizing the remaining irrelevance arguments, it is conventional to start
with the MM arguments. The gist of the MM argument is captured by Levi and Sercu
(1991): “It is a well-accepted principle of finance that managers of a firm will not increase
the firm’s value by doing anything the shareholders of the firm can do themselves at
the same or lower cost.” This argument is an extension of the MM arguments from
traditional corporate finance that propose the financial policies of the firm are irrelevant
in determining the market value of the firm. The original MM arguments focused on
demonstrating that the capital structure and dividend policies of the firm have no
implications (are irrelevant) for the market value of the firm.3 Value is determined by
the asset side of the balance sheet. The extension to the international arena is that, as a
financial decision of the firm, the use of derivative securities to implement corporate FX
risk management decisions is irrelevant for the same reasons as outlined by Levi and
Sercu: the market will not increase the value of the firm for engaging in practices that
can be done directly by investors.
The MM irrelevance argument relies on perfect market assumptions. Within
the MM framework, violations of key assumptions can dramatically change the
results. For example, when corporate taxes are admitted and tax deductibility of
interest payments on the debt is allowed, then instead of debt irrelevance, the simple
MM model indicates that all debt financing is the optimal method to maximize
the market value of the firm. Though introducing taxes can also provide a rationale for
the use of derivative securities, e.g. Graham and Smith (1999), this type of motiv-
ation does not appear to be widespread in practice, e.g. Graham and Rogers (2002).
Siegel and Siegel, (1990, pp.150–1) provide an illustration of when taxes could
provide a motive for hedging. More importantly, the MM argument is not exempt
from the implications of relaxing other perfect market assumptions such as no
bankruptcy costs. If the market value of the firm is affected by bankruptcy risk, then
by reducing the total variability of cash flow, hedging and other risk management
activities can increase the market value of the firm by lowering the default premium
and, thereby, lowering the discount rate in the long run NPV calculation, e.g.
Purnanandam (2008).
Following Frank Knight, firms can earn economic rents from correctly handling
uncertainty. Measurable risks, which can be handled by conventional risk management
techniques such as purchasing insurance, are part of the cost structure, not a source of
economic value added. In many cases, firms that do not accurately handle measurable
risks will, in the long run, suffer the consequences of the market place. This is all too
apparent from the MGRM debacle. Yet, if risk management activities are aimed at
increasing expected net cash flows then it does not follow that the firm will also be
able to reduce discount rates by reducing the variability of future cash flows. Using
risk management to increase expected cash flow may also increase the firm’s cash flow
variability. As demonstrated in the debacles of Sec. 1.3, risk management activities
can be a source of economic profit, i.e. increased net cash flow, only by moving out of
the realm of measurable risks and into the grayer area of uncertainty. Ultimately, the
optimal risk management solution is composed of a risk minimizing component and
a speculative component. Optimal risk management may increase bankruptcy risk to
achieve an expected speculative gain.
Being derived using perfect market assumptions, CAPM arguments have many
similarities with the MM arguments. A version of the CAPM argument can be found in
the Levi and Sercu (1991):
will be given by: SU(0) = E[S(1)]/(1 + E[Rs,1]), where the discount rate is determined by the
CAPM.
To demonstrate that hedging is irrelevant to the market value of the firm, consider
the value of a share if the firm decides to fully hedge and uncertainties in production are
ignored: SH(0) = F(0,1)/(1 + Rf ). Because the output price has been locked in by hedging,
the discount rate will be lower than for the unhedged firm. The CAPM argument
involves making sufficient assumptions to ensure Su(0) = SH(0). This requires E[S(T)] to
be higher that than F(0,T) by precisely the amount needed to offset the difference in the
discount rates. Observing that if silver is typically near full carry, then it is possible to
assume that Rf = r(0,1) = ic(0,1). Similarly, because this company is a pure silver mining
play, the discount rate for this company can be assumed to be the same as the CAPM
discount rate for holding spot silver, E[Rs,1] = E[Ri,1]. Given this, Su(0) = SH(0) = S(0)
and the CAPM argument is validated. The market price of the firm’s stock will be equal
to the current spot price of silver. This is consistent with the company being a pure play
on the deterministic stock of silver which will be marketed in one year’s time.
Evidence on motivations for non-financial firm hedging from surveys and other
sources reveals that the important determinant of firm hedging activity was the desire
to reduce the volatility of firm cash flows, e.g. Bodnar et al. 1996, 1998; Servaes et al.
2009. By exhibiting less volatile cash flows firms can potentially lower the cost of capital.
However, the CAPM argument maintains that this motivation is fictional. Any decrease
in the cost of capital from hedging is exactly offset by a decrease in the expected cash
flows of the firm. This follows from the equilibrium underlying the determination of
forward prices. The forward price will differ from the expected spot price by just the
amount needed to offset the gain associated with the reduction in the cost of capital. Yet,
among other problems, the CAPM argument imposes unrealistic empirical conditions
on the basis and the futures basis. In addition, the full hedge assumption implies that
the motivation for hedging is to only to reduce the volatility; no attempt is made to
identify the optimal hedge and to examine the associated valuation implications of
employing such a hedge.
Much like the MM argument, the CAPM argument can be criticized by demonstrating
that relaxation in the underlying assumptions substantively changes the results. Under
the perfect markets assumptions required for the CAPM to hold, the CAPM argument
could have considerable validity. However, the CAPM assumptions are relatively severe.
Of particular interest are the assumptions which would make total instead of systematic
variability a concern. No bankruptcy costs is, again, a key CAPM assumption. The basic
hedging framework explicitly identifies risk afford ability as an essential element in
establishing a hedging program. While there are numerous possible examples, two that
could be used are: the case of AMR just prior to filing for bankruptcy in 2011 where
the change in jet fuel prices could more than eliminate firm cash resources needed
to avoid bankruptcy; and the FX exposure of the (now defunct) Vancouver Grizzlies
who earned the bulk of revenues in C$ but were obliged to incur expenses, including
but not limited to salaries, in US$ at a time when the US$/C$ was falling. Among
other things, the presence of bankruptcy costs can affect a firm’s cost of and accessibility
to capital.
Another difficulty with the CAPM argument is the significant restrictions on
diversification associated with real assets. The CAPM framework was developed to
explain optimal portfolio selection, where the assets involved are highly liquid and
divisible. However, the non-financial firms involved in the production, transportation
and consumption of commodities use real assets that are typically “lumpy” and not
easily divisible. The alternative assets needed to adequately diversify may not be available
for purchase. Where such assets are available, capital constraints and other factors may
prevent their acquisition. In short, in commodity risk management situations it is not
usually possible to construct “an efficiently diversified portfolio” of real assets sufficient
to achieve a CAPM solution. Again, factors such as the lumpiness of assets and the
inability to adequately diversify, means there is an element of uncertainty in business
decisions that cannot be reduced to the type of measurable risk argument which
underlies the CAPM.
Another group of arguments against the use of derivatives in risk management can
be classified as “expected return” arguments. These arguments make the empirical
observation that risk management activities involving derivatives will be, on average,
a zero expected value operation. Due to the costs associated with initiating, monitoring
and executing a derivatives program, practical concerns dictate that such activities
are unnecessary. To see this, consider the case of a futures hedge. Assume that
futures prices are unbiased predictors of future spot rates, then a policy of continuous
hedging will just reflect back the price changes. Sometimes hedges will make money,
sometimes hedges will lose money. On balance, the gains and the losses will net out
and the hedged position will have the same expected value as the unhedged position.
In effect, the expected values of the returns on the hedged and unhedged positions
will be equal.
As it turns out, this argument may have validity in specific theoretical settings.
For example, in perfect markets with risk neutral participants, traders are indifferent
between risky prospects with the same expected values, independent of the variance of
the prospect. However, the theoretical assumptions required to generate risk neutrality
results are associated with models that do not have a well defined general equilibrium.
In a world with risk averse participants, the dependence solely on expected value
omits one of the primary reasons for hedging: controlling the total variability of the
firm’s cash flows due to changes in commodity prices, exchange rates, interest rates,
and equity prices. A practical objective of this could be to cause fluctuations in the
firm’s market value to be due solely to changes in the firm’s business activities, not
(random) changes in market prices. As in the MM and CAPM arguments, this could
enhance long run share prices by reducing bankruptcy cost thereby reducing the cost
of capital.
It could be further argued that the expected return arguments against using derivative
securities for risk management do not fully develop the implications of observing that
such activities have zero expected value. For example, consider the statement that “it is
just as likely to be surprised on a foreign exchange hedge as on the cash position.” Given
the additional costs associated with having a derivatives program, an expected return
argument would conclude this is a reason for not hedging. However, situations when
the hedge loses money, i.e. provides an unanticipated “surprise,” will likely be situations
where a windfall gain would have occurred as a result of favorable exchange rate changes.
In order to reduce volatility, the hedge using futures or forward contracts trades off both
downside and upside changes in the cash prices being hedged. This allows the firm
to concentrate on production problems without having to worry about complications
related to unexpected changes in market prices. The important question is whether, in
this situation, firms that successfully pursue active commodity risk management will
have a substantive competitive advantage over firms that are continuously fully hedged
or do not hedge at all.
While there is empirical support for these rationales of hedging at the firm level, the
evidence is only modestly supportive, suggesting alternative explanations.
In another recent survey article, Smithson and Simpkins (2005) come to a similar
conclusion:
Is there any direct evidence that risk management increases firm value? The answer
is yes, but the evidence is fairly limited as yet. A number of more recent studies
show a clearly positive correlation between higher share values and the use of
derivatives to manage foreign exchange rate risk and interest rate risk. And one study
provides fairly compelling evidence that the use of commodity price derivatives
by commodity users increases share values. But studies of hedging by commodity
producers provide no clear support for the argument that risk management adds
value. At a minimum, whether hedging adds value appears to depend on the types
of risk to which a firm is exposed.
As hinted at in the last sentence in the quote, a possible explanation for such tepid
results is the heterogeneity of commodity risk management situations which defies the
search for “one-size-fits-all” explanations.
Employing Tufano’s method of classifying studies extends the conventional corporate
finance classifications which are based on the elements of the NPV calculation. Aretz
et al. (2007, p.434) identify the various theories supporting risk management activities
associated with reductions in the discount rate, increases in expected net cash flow or
increases in real option value of projects:
When there are imperfections in capital markets, corporate hedging can enhance
shareholder value through its impact on agency costs, costly external financing,
direct and indirect costs of bankruptcy, as well as taxes. More specifically, corporate
hedging can alleviate underinvestment and asset substitution problems by reducing
the volatility of cash flows, and it can accommodate the risk aversion of undiver-
sified managers and increase the effectiveness of managerial incentive structures
through eliminating unsystematic risk. Lower volatility of cash flows also leads to
lower bankruptcy costs.
markets assumption will have to be altered. Along this line, Froot et al. (1994, p.98)
maintain that by stabilizing cash flows, firms can use derivative securities to align
the internal supply and demand of funds. By stabilizing cash flows, corporate risk
management permits the firm to participate in investment opportunities that may
arise at inopportune times: “Managers who adopt our approach should ask themselves
two questions: How sensitive are cash flows to risk variables such as exchange rates,
commodity prices, and interest rates? and How sensitive are investment opportunities
to those risk variables? The answers will help managers understand whether the supply
of funds or the demand for funds are naturally aligned or whether they can be better
aligned through risk management.”
According to Froot et al. (1994), by stabilizing cash flow, risk management permits
firms to undertake some positive NPV projects that would be avoided in the absence
of such activities. This hypothesis could be targeted at any of the elements in the
NPV calculation. By stabilizing cash flows, the firm is better able to access sources
of internal financing, which are cheaper to use than external financing. This will
lower the discount rate. By using derivative securities to avoid under investment, risk
management increases expected future cash flow by increasing the number of positive
NPV projects. Finally, stabilizing cash flows can permit the firm to exercise real
options, such as the development option, thereby increasing the value of these options
to shareholders. Various empirical studies provide mostly favorable evidence on the
under-investment hypothesis, e.g. Adam (2002), Carter et al. (2006), Gay and Nam
(1998). For example, using a sample of 325 firms using derivatives combined with 161
firms not using derivatives, Copeland and Copeland (1999, p.74) estimated that “firms
with enhanced investment opportunities, lower liquidity, and low correlation between
investment expenditures and internally generated cash flows tend to be more likely
users of derivatives.”
Other promising explanations for corporate risk management have been advanced
in academic studies. Key factors in these explanations include: the ownership structure
of the firm (Smith 1995); resolving conflict between firms by enhancing the contracting
relationship between firms (Pennings and Leuthold 2000); risk shifting within the
firm (Smith 1995); and lowering expected tax costs (Smith and Stulz 1985). Ownership
structure can be related to both managerial incentives and shareholder wealth
maximization. “Managers whose human capital and wealth are poorly diversified
strongly prefer to reduce the risk to which they are exposed. If managers judge that it
will be less costly (to them) for the firm to manage this risk than to manage it on their
own account, they will direct their firms to engage in risk management” (Tufano 1996,
p.1109). Similarly, concentrated ownership, whether in the hands of management or
not, likely means that owners do not have well-diversified portfolios, again providing
an incentive for the firm to engage in risk management. As in the argument against the
CAPM approach, real assets can be lumpy; it is not easy to hold an efficiently diversified
portfolio.
The theoretical rationales for corporate risk management using derivatives provide
some foundation for the discussion of practical issues involved in commodity risk
management decisions and activities. Translating academic discussion into practice
is facilitated by detailing some heuristic guidelines. The development of a “risk
management philosophy” for a specific firm is an essential step in developing an
effective risk management program. This process can be motivated by identifying
a number of basic considerations to be addressed in order to determine the type of
derivatives trading program to be undertaken, e.g. Fisher and Kumar (2010), Poitras
(2002), Powers and Vogel (1981). It is in formulating answers to the various questions
that essential elements of a risk management policy become apparent. To this end,
consider the following sequential list of questions that are of particular relevance for
a firm considering the implementation of derivatives trading program for commodity
risk management purposes.
share price to be fully exposed to changes in metal prices. Affordable for management
is not necessarily consistent with affordable for shareholders which is not necessarily
consistent with affordable for bond holders. Similarly, firms may want to use hedging
to protect the ability to make future dividend payments to common and preferred
stockholders.
The risk afford ability issue is also difficult to determine for government enterprises
where, ultimately, afford ability is determined by the ability to raise general tax revenue,
borrow against future tax revenue or levy increased user rates. For example, BC Ferries,
the government monopoly ferry service in British Columbia is able to arbitrarily add
surcharges to regulated ferry fares when fuel costs are higher than expected, instead
of hedging the full expected amount of fuel usage over the period the regulated fares
are mandated. Some government-linked Canadian electric utilities, such as BC Hydro,
have a substantial portion of debt denominated in US dollars while the bulk of cash
flows are in Canadian dollars. Large, adverse changes in the exchange rate which would
be sufficient to eliminate the net asset value of a private company, would not have the
same impact on the government-owned utility. This situation is further complicated
for some Canadian utilities, such as BC Hydro, that do considerable business selling
surplus power to US consumers.
Can the risks be hedged? Are other methods of risk management applicable?
This is the problem of hedge design, a topic that is the central concern of Poitras (2002,
Ch.6). There may be a variety of possible hedging techniques that have to be considered.
An important practical concern is whether there are derivative contracts available
which qualify as feasible hedging instruments. In many commodity markets, forward
contracts are available that allow the cash position to be matched with the commodity
underlying the forward contract. In some cases, no forward contracting method
is available and a cross hedge can be executed using exchange traded derivative
securities. Even in cases where forward contracting is available, the pricing on the
forward contract may be considered to be expensive relative to doing a cross hedge. For
example, an airline may undertake a cross hedge using NYMEX heating oil futures in
lieu of doing a short dated jet fuel swap in the OTC market because the cost of the swap
is deemed to be too expensive relative to doing the NYMEX hedge and absorbing the
basis risk.
Cross hedging involves managing a specific commodity risk using a derivative which
is written for a commodity which differs from the cash commodity, e.g. Borger (2009).
For example, a copper scrap dealer can cross hedge using copper futures contracts which
feature copper cathodes as the deliverable commodity or a wheat farmer in S. Dakota
can cross hedge with Chicago-deliverable wheat futures. Cross hedges can sometimes
involve quite different commodities, such as hedging brass scrap with a combination of
aluminum, copper and lead futures/forward contracts. Cross hedging raises questions
about the appropriate size of the derivative position relative to the cash position (an
optimal hedging problem) and whether the hedge will be effective.
What are the tax and accounting implications of hedging with derivative
securities and other risk management activities?
The relevant issues involved here are discussed in other sources, e.g. Okochi (2008),
Zhang (2009). These issues are not incidental and will have to be determined in order
to precisely calculate the costs and feasibility of risk management activities such as
hedging with derivative securities. In particular, the introduction of FAS 133 and IAS
139 raises a host of questions and queries that lie outside the confines of the present
inquiry. Small to medium sized non-financial firms often rely on outside consultants,
e.g. Accenture or Deloitte, to handle the preparation of relevant derivative related tax
and accounting items. Emm et al. (2007) provide a valuable discussion of accounting
disclosure issues relevant for risk management. Pincus and Rajagopal (2002) examine
the impact of hedging on accrual management.
foreign currency. In the absence of indexing, these factors cannot be readjusted when
unanticipated changes in the nominal exchange rate occur. Hence, it is possible for the
real exchange rate to be unchanged and still have substantive changes in economic
behavior. Similarly, it is possible for the nominal exchange rate to be unchanged and
for changes in relative inflation rates to occur that will have substantive economic
implications. Shapiro (1992, pp.228–9) provides an illustration of this happening
1979–82 in Chile where a government attempt to fix the value of the Chilean peso led
to a significant erosion in international competitiveness which had a disastrous impact
on the Chilean economy.
A useful Canadian example of how economic currency exposure can affect firm
profitability is the hotels and related businesses at the Whistler/Blackcomb ski resort
in BC.6 Even though virtually all revenues and costs are in Canadian dollars, revenues
are indirectly dependent on competition from overseas ski resorts. In effect, Whistler/
Blackcomb is operating in a global market for skiing and other vacation services. Changes
in the Canadian dollar will change the relative value of overseas ski vacations, for both
domestic and foreign vacationers. More generally, even though a firm does not have any
direct foreign currency exposure, the presence of foreign competition in either the input
or output market means that there could be substantial economic currency exposure.
Another Canadian example of corporate currency exposure is provided by the
Canadian mining industry. Because the price of metals is set in global markets in US
dollars, mining company US dollar revenues will not be affected by changes in the
Canadian dollar, assuming the price of sales in Canadian dollars is allowed to change
to reflect the US dollar price. While US dollar revenues will not change, changes in the
value of the Canadian dollar will alter the US dollar cost of Canadian labor and supplies
used in the production of metals. This type of situation occurs in many other Canadian
cases, where the product being produced is being priced on the international market in
terms of US dollars. This is the case with the grains such as wheat, and energy products
such as oil, natural gas and hydroelectricity.
As a final example of corporate foreign exchange exposure, consider Toyota, an
automobile manufacturer, where both revenues and costs are affected by exchange
rate changes. On the revenue side, Toyota sells the bulk of its production overseas,
concentrating on the USA. Changes in the value of the yen will force a pricing policy
decision. For example, in the face of an appreciation of the yen, to maintain market
share the US dollar price has to be held constant, reducing yen revenues because the
yen price per unit has fallen. If the yen price is held constant, market share will be
reduced because of higher US dollar prices. On the cost side, Toyota is a purchaser of
commodities required in car production which are priced on international markets.
Changes in costs will tend to offset changes in revenues, though not one-for-one. In the
case of Toyota, because such a large component of revenues is in US dollars while only
a relatively small portion of costs is not yen determined, the impact of appreciation in
the yen is negative. Hence, there are numerous ways in which currency exposure can
impact a given firm.
nature of the product being sold in the USA. In effect, real exchange rate changes
made competition at the low end of the market unprofitable. As a result, these com-
panies have made long-term product adjustments by offering higher-priced auto-
mobiles targeted at middle to upper middle income consumers. In contrast to
persistent real exchange rate changes, temporary exchange rate changes will usually
not require substantial adjustments to product offerings. However, temporary
depreciations may provide timing opportunities for firms seeking to penetrate
foreign markets. This is an important point because the high fixed startup costs
associated with overseas expansion are often incurred in the initial stages of establishing
a market presence. Favorable, if temporary, exchange rate changes can partially offset
these costs.
Perhaps the most widely recognized method for multinational firms to manage
currency risks is to create natural hedges through appropriate plant location and
input purchase decisions. Firms that have similar production facilities in areas with
different currencies can, potentially, shift production to plants where production
is least expensive. Where it is not possible to establish production facilities in the
appropriate locales, then a similar result can be achieved by spreading sources for
inputs across countries in different currency areas. In practice, the benefits associated
with multinational sourcing and production facilities have to be balanced against the
costs associated with plant redundancy and loss of economies of scale. In a corporate
context, this requires managerial decision makers to incorporate forecasts of exchange
rate changes into company strategies. Hence, there is an element of active management
in adjusting to currency exposure. In addition to the natural hedges provided by
plant shifting and alternative sourcing, it is also possible to react to currency related
changes in competitive conditions in a more traditional fashion, i.e. by raising domestic
productivity.
The final important method for corporations to use natural hedges to manage
currency exposure is by adjusting the capital structure of the firm. Conventionally,
this involves taking advantage of the natural hedge provided by financing real assets
with foreign debt. Where the cash flows of the real assets have an identifiable currency
exposure, either because of foreign competition or dependence on foreign markets for
inputs or sales, changes in operating cash flows arising from exchange rate changes
will be met by offsetting changes in debt service costs. As with any type of hedging
situation, there will be situations when the hedge position is unprofitable, i.e. where the
domestic currency value of the foreign borrowing increases. In these cases, it would
be desirable to finance real assets with domestic debt. Because it may not be possible
to adjust borrowing programs to keep pace with the numerous exchange rate changes,
once again the natural hedge decision depends on active management of currency
exposure to achieve the highest return. If management is not able to forecast or has a
high degree of risk aversion, the optimal solution will be to establish a natural hedge
that matches the foreign currency exposure.
Ri , t = α + β1 RM , t + β2 RFX , t + ui , t
PPP Arguments
Though the roots of PPP can be found in Adam Smith and early 19th century classical
political economy, the PPP theory is usually credited to Gustav Cassels, writing in the
1920s. The earliest versions of PPP took the form of the Law of One Price: assume a one
good world with no transactions or transportation costs, then the price of that good
denominated in different currencies will sell at the same price:
∗ Pt
Pt St = Pt ⇒ St =
Pt ∗
where P* and P are the foreign and domestic prices of the good with S being the spot
exchange rate. Norman (2010, p.919) observes: “The concept of purchasing power parity
(PPP) is one of the most empirically well studied theories in international economics,
perhaps because evidence of its existence has been so elusive.”
Extending the Law of One Price using price indices instead of individual prices is
known as Absolute Purchasing Power Parity (APPP). Even in the unlikely event that the
Law of One Price holds for each good individually, the APPP extension may be invalid
if the index weights are not the same for both economies. The problem of traded and
untraded goods also creates significant difficulties. Nevertheless, ignoring the various
possible problems with APPP, substituting price levels p* and p into the Law of One
Price and taking logs produces:
1 dS 1 dp 1 dp *
ln S = ln p − ln p * ⇒ = − ⇒ S = p − p *
S dt p dt p * dt
Hence, APPP holds that foreign exchange rate changes are determined by the difference
between foreign and domestic inflation rates. One implication of this appealing
interpretation of exchange rate changes is that predicting domestic and foreign inflation
rates will permit exchange rate changes to be forecasted accurately.
A more popular form for PPP to take is Relative Purchasing Power Parity (RPPP). This
is the version used to define the real exchange rate as the nominal exchange rate adjusted
for changes in the relative purchasing power of each currency since some base period. In
a one period framework, the relative form of the PPP condition can be expressed:
St +1 Pt +1 / p * t +1 1 + p t , t +1 1 + p * t , t +1
= = ⇒ St = St +1
St Pt / pt * 1 + p * t , t +1 1 + p t , t +1
where p is the appropriate price level index, the inflation rate and * denotes a foreign
value. The real exchange rate (st) notion is an attempt to convert observed exchange
rates back to some base period. Starting from some base year where S0 = s0, then:
1 + p * 0 , 1 1 + p *0 , n
s1 = S1 ⇒ sn = Sn
1 + p 0 , 1 1 + p 0 , n
The multiperiod form of st involves compounding the inflation term over the time
between the selected base year and the desired date. Some evidence on the historical
behavior of nominal and real foreign exchange rates is given in Poitras (2002, Table
4.2.1). Casual examination of this Table reveals that real exchange rates for many
currencies do deviate significantly from the PPP requirement that the real exchange
rate is relatively constant over time.
The basic approach of the PPP arguments against hedging currency risk is to attack
the notion of exchange risk. This follows from the PPP implication that, in the long run,
exchange rate changes will offset price level changes.7 Take the example of a Canadian
sugar refiner selling output in C$ but purchasing sugar in US$. The PPP argument
indicates that a deterioration in the FX rate will be compensated for in price level
increases. If, say, the US$/C$ increased by 50 percent (C$/US$ falls), causing the cost
of raw sugar inputs to increase proportionately, then PPP dictates that the Canadian
inflation rate will be such that the price of refined sugar in Canada will increase to
completely offset the Canadian dollar increase in input costs. When appropriate
assumptions are satisfied, PPP holds and the real foreign exchange rate is unchanged.
In this case, there are no real implications to nominal foreign exchange rate changes.
The argument that PPP holds and, hence, corporate currency hedging is unnecessary
has a number of obvious and not-so-obvious shortcomings. A list of these would include
the following.
(a) Empirical applicability of PPP: there is a sizable literature on the empirical validity of
PPP. the long lead-lag time period for the relationship to hold makes PPP inconsistent
with the typical types of business decision time frames; the applicability of PPP to
tradables more so than non-tradables creates complications if the hedger is interested
in non-tradables. The persistence of deviations from PPP in the face of a volatile real
exchange rate led Rogoff (1996) and others to identify the “Purchasing Power Parity
Puzzle.” Rogoff (1996) estimates the half lives of shocks that generate deviations from
PPP to be in the order of three to five years which “seem[s] to be extremely long even
when PPP is viewed as a long run concept” (Norman 2010, p.919).
(b) The slippage created between the price index which underlies PPP and the specific
prices that are of interest to the hedger. It is relative, not aggregate, prices which are of
interest.
(c) The presence of financial and operating contracts, which are fixed in nominal terms,
i.e. cash flows which do not adjust when the aggregate price level changes.
Agricultural production is still quite concentrated with less than 6% of the farms
in the United States producing 75% of the value of production . . . Most U.S. farms
still produce undifferentiated commodities for markets where production is
characterized by relative ease of entry and exit. And farming is still a risky business.
For the metals and energy complex commodities, information about commodity risk
management practices is relatively thin, available risk management tools are limited,
and capital costs and other factors can pose significant barriers to entry, e.g. ICME
(2001). In contrast, agricultural risk management features a variety of commodity risk
management tools, including crop insurance, exchange traded derivatives, forward
contracts with elevators, disaster and other income assistance, production contracts,
crop diversification and cash on hand, e.g. Harwood et al. (1999). The vast scope of
studies on agricultural risk management extends into the Third World where various
NGOs and international agencies have produced numerous studies on the topic, e.g.
UNCTAD (1997), World Bank (1999).
On the CME, commodities within the agricultural complex are divided into: grains
and oil seeds; livestock and dairy; forestry; and the soft commodities, coffee, sugar,
cocoa and cotton. Of these, grain and oil seeds exhibit the most seasonality due to
the harvesting cycle. As illustrated in Figs 2.7 and 2.8, the volatile price behavior of
Figure 2.8 Monthly Percentage Price Changes for Corn, Wheat and Soybeans
Source: Macdonald and Korb (2011)
Based on numerous studies by the USDA, such views are somewhat misleading as it is
the limited amount of interest in using futures contracts that is being observed, not the
overall use of risk management tools. USDA sources such as the Agricultural Contracting
Update, e.g. Macdonald and Korb (2008, 2011) and Macdonald et al. (2004), combined
with results from some academic studies provide a much clearer picture of the changing
risk management landscape.
Even if an elevator is completely hedged—so that the elevator will have “locked
in” a gain regardless of the direction of the market—a steeply rising market
can impose significant additional costs upon the elevator operator. In a rising
market, grain elevators and merchants that have hedged by selling futures
may be subject to margin calls from the exchanges to cover the loss in value
of their “short” positions. These margin calls, which are made at the end of
each trading day, require payments by the grain elevator or other party to the
futures exchanges into a margin account. The amounts in the margin account
are not recovered by the elevator until the short position is closed out – in
this case, until the elevator sells its grain and terminates the hedge. If a grain
elevator cannot make the requisite margin payments, the exchange will close
out its position at the current market price, possibly causing further losses. In
2008, rising grain prices in the cash markets, together with rising margin calls,
required many grain elevators to make much larger cash outlays than normal.
The National Grain and Feed Association estimated that a typical grain
elevator faced a 300% increase in hedging costs in 2008, compared to 2006.
It stated that “recent commodity price increases have led to unprecedented
borrowing by elevators – and unprecedented lending by their bankers – to
finance inventory and maintain hedge margins.”
The characteristics of farming differ by crop and geographical region, e.g. Parkinson
(2000). Despite this, Harwood et al. (1999) examine the comprehensive 1996 US
Agricultural Resource Management Study (ARMS) data to discover:
Keeping cash on hand for emergencies and good buys was the number one strategy
for every size farm, for every commodity specialty, and in every region.
Table 2.15 provides further examination of results from the ARMS data, identifying both
farm characteristics and use of marketing contracts for risk management. Examining
the results for the use of marketing contracts, raises some issues when the following
from Mark et al. (2008, p.22) is considered:
Goodwin and Schroeder (1994) found in a sample of Kansas producers that only
11% hedged any of their grain using futures. Schroeder [et al.] (1998) summarized
several studies that consistently showed that more producers used forward contracts
than used futures hedges. These studies showed that 42–74% of producers used
forward contracts to price any of their grain.
Comparing Table 2.15 with Table 2.16, which was calculated from a survey of a smaller
number of farms, reveals that the ARMS data for “marketing contracts” in wheat at just
under 10 percent likely relates to the use of futures contracts. However, sugar beets, a
relatively non-storable commodity for which there is no actively traded futures contract
available has 87 percent usage of marketing contracts. As a consequence, some care is
needed when using ARMS data and the related Census of Agriculture data (www.nass.
usda.gove/census/) to compare usage of different types of derivative security contracting
across field crop producers, e.g. agreements covering the sale of harvested commodities
in storage are not defined as agricultural contracts in ARMS.
Cash forward contract 82.2% Catastrophic coverage 42.1% Younger than 25 years 0.7% Over $1,000,0000 16.5%
Basis contracts 42.2% Crop revenue coverage 49.6% 25–29 years 4.4% $999,999–$500,000 25.9%
Futures contracts 40.4% Only hail insurance 21.4% 30–34 years 12.8% $499,999–$4000,000 13.7%
Put options 37.0% Group risk plan (GRP) 8.9% 35–39 years 21.2% $399,999–$3000,000 15.4%
Hedge-to-arrive contracts 20.6% Income protection (IP) 5.8% 40–44 years 20.0% $299,999–$200,000 17.3%
Minimum price contracts 13.2% Revenue assurance (RA) 5.3% 45–49 years 18.0% $199,999–$100,000 9.9%
50–59 years 18.8% $99,999–$50,000 1.1%
60–64 years 2.7% Less then $50,000 0.1%
65 years and older 1.4%
Crop acreage (planted annually) Com Sorghum Soybean Wheat Cotton Rice Hay
Over 2,000 acres 4.5% 1.1% 2.9% 9.1% 2.2% .4% 5.2%
1,999–1,500 acres 16.3% 1.5% 10.9% 14.7% 3.7% 1.3% 3.1%
1,499–1,000 acres 42.3% 3.0% 34.2% 16.3% 4.7% 1.8% 5.4%
999–500 acres 7.9% 5.1% 14.4% 8.0% 1.5% 1.1% 7.1%
499–300 acres 6.9% 8.3% 9.9% 13.3% .6% .8% 14.9%
Under 300 acres 2.9% 6.6% 4.6% 12.4% .4% .1% 21.3%
No acres 19.3% 74.5% 23.1% 26.2% 87.0% 94.6% 42.9%
Note: The sample consists of 1,105 U.S. crop producers in the Midwest, Southeast, and Great Plains. The crop producers' age, gross annual farm sales,
and crop acreage were obtained from accounting data. Data on price risk management instruments and insurance products were measured during
the survey and reflects usage during 1999 and 2000.
Source: Pennings et al. (2008, p.35)
11/16/2012 6:45:19 PM
248 Commodity Risk Management Concepts
For grains and oilseeds, one common method for hedging to be implemented is for
grain elevators to provide forward price quotes to farmers. Henderson and Fitzgerald
(2008) describe the process:
As the various USDA studies on agricultural contracting detail, this seemingly simple
process that has characterized commodity risk management in this sector since the 19th
century has been changing as the landscape of agricultural production has changed:
Farm production is shifting from smaller to larger family farms and from spot (or
cash) markets to contracts. Technological developments may underlie much of
the shift to larger farms, but expanded use of production and marketing contracts
supports that shift by reducing financial risks for farm operators. For farm operators,
contracts provide benefits from reduced risks, but also result in loss of managerial
control and reduced autonomy.
annual sales)” (Macdonald and Korb 2011, p.9). Nevertheless, it cannot be overlooked
that: “Most transactions for US agricultural products are conducted through spot
market exchanges in which commodities are bought and sold for immediate delivery”
(Macdonald and Korb 2008, p.iii).
In addition to farm size, there considerable diversity in commodity risk management
practices across sectors in the agricultural industry, e.g. Musser and Patrick (2002).
Certain agricultural segments—especially hogs and poultry—use production contracts
as a mechanism for dealing with price and production risks. For example, Dimitri
et al. (2009) examine the transition from cash trading to contracts in the broiler industry.
Such contracting methods (Macdonald and Korb 2011, pp.1–2):
specify services provided by a farmer for a contractor who owns the commodity
while it is being produced. The contract covers (1) the services provided by the
farmer, (2) the manner in which the farmer is to be compensated for the services,
and (3) the specific contractor responsibilities for provision of inputs. For example,
farmers provide labor, housing, and equipment under livestock and poultry
production contracts, while contractors provide such other inputs as feed, veterinary
and livestock transportation services, and young animals. The farmer’s payment
resembles a fee paid for the specific services provided by the farmer, instead of a
payment for the market value of the product. Since contractor-provided inputs may
account for a large share of production costs, the fee paid to the farmer may be a
small fraction of the commodity’s value.
In situations where the farmer also has obtained financing for capital investments, and
especially if the financing was obtained from the contractor, such arrangements can lead
to significant problems. For example, the length of time required to repay the financing
on the capital investment can be longer than the term of the pricing agreement leading
to “holdup” costs where the contractor is able to extract a much lower fee to the farmer
when the pricing arrangement is renegotiated at the end of the term. In general, farms
with more debt also tend to be greater users of contracts (Key 2004).
The relevance of hedging farm revenue instead of examining price and quantity
risk separately has received considerable attention, e.g. Acs et al. (2009), Guinvarc’h
et al. (2004), Hart et al. (2001), Poitras (1993). Various sources have also identified the
amount of farm leverage as an important element in hedging decision, e.g. Shapiro and
Brorsen (1988).
The decision problem is motivated by the stylized farming situation, at planting
time the farmer must estimate the size of the future crop in order to determine the
size of the hedge. More generally, this problem also applies to hedging situations where
future inputs to the production process have to be estimated. For the farmer’s problem,
because QS is not fixed at t = 0, it must be estimated. At t = 0 the farmer plants with
the objective of reaping ^q where ^q is generated by some production function, Q[K , L ,
Land], and offered for sale at some expected price, E[S(1)], where K is the capital stock
used in production and L is the labor input. Even at this basic level, the complexities of
the farmer’s problem are apparent. To make the problem more manageable, a number
of simplifying assumptions can be used.
In the analysis of the minimum variance hedge ratio, it was shown that EU = -var
was the associated objective function. This functional form can be compared to the
mean-variance EU. Given the similarities in the mean-variance and minimum variance
optimal hedge ratios, it is possible to rationalize the minimum variance approach by
assuming that hedgers, in this case farmers, do not forecast spot prices. In effect, the
price in the future is expected to be the same as it is today. This form of myopia permits
use of the minimum variance hedge ratio by eliminating the need to consider expected
price change. Another simplifying assumption is the requirement that the hedge
position put on at t = 0 is held and not changed until harvest at t = 1. More formally, no
dynamic hedge adjustment is permitted. Given these two assumptions, two situations
can be considered: where ^q = QH = Q, i.e. there is no cross hedging of estimated output;
and ^q and QH are allowed to differ, i.e. there is cross hedging. If farm output realized at
t =1 is taken to be q̃, then it is possible to define the profit function for the farmer doing
a delivery hedge (where no cross hedge is involved).
Creating a new random variable q̃(1) S(1) = Y(1), which effectively represents future
farm income, it is possible to define an associated variance of profit function which can
be used to solve for the minimum variance hedge ratio:
d var[π] = 2[Q σ2 − σ ] = 0
F YF
dQ
σ
Q * = YF2
σF
While interesting, this form of the optimal hedge ratio has little practical value, if only
because it is difficult to determine σYF. Expanding the solution to allow for QH to differ
from the expected size of the crop position does not substantially change the practicality
of the solution. The primary analytical difficulty is the presence of a random variable,
farm income, which is the product of two random variables. Only one of these variables,
spot prices, is hedgable. With this in mind, it is possible to reconstruct the optimal
problem into the form used in determining the optimal myopic hedge ratio and the
quasi-separation of the hedge ratio.
The basic model is discrete. Farmers have access to a variety of possible risk
management instruments to hedge production decisions. The representative grain
farmer plants a crop at time t and harvests it at time t+1. Both the price at harvest and
the quantity harvested are unknown at time t, the date the relevant risk management
and planting decisions are initiated. As conceived here, in addition to choosing the
usage of hedging instrument(s), the farmer’s optimization problem also involves
choosing the amount of initial wealth to invest in crop production. Hence, the
production decision is treated in a portfolio context. As a result, the costs associated
with planting the given acreage are also determined. Starting from a given initial level
of wealth, the farmer’s objective is to maximize the value of terminal wealth assuming
that the balance (possibly negative) of initial wealth which is not allocated to planting
costs will earn (pay) the risk-free rate of interest.
Given this basic structure, it will initially be assumed that the only hedging instrument
available is forward contracts with a grain elevator. In this case, the underlying wealth
dynamics can be specified:
Wt +1 = AYt +1Pt +1 + [Wt − C( A )] ( 1 + r ) + Q f ( ft +1 − f t )
where Wt+1 is wealth at time t+1 and Wt is the known level of initial wealth; A is the
number of acres planted; Yt+1 is the random yield per acre observed when the crop is
harvested at t+1; Pt+1 is the random spot price at t+1; C(A) is the known cost function
associated with planting the A acres; r is the risk-free interest rate; Q f is the quantity of
futures contracts sold (–) or bought (+); and f t+1 and f t are the futures prices observed
at t+1 and t respectively. Manipulation gives terminal wealth used in Proposition 2.4:
Wt +1 = Wt ( x( 1 + R ) + ( 1 − x )( 1 + r ) + HR f )
= Wt (( 1 + r ) + x( R − r ) + HR f )
= Wt + p t +1
where πt+1 is the profit realized at time t+1, x is (C(A)/Wt ) the fraction of initial wealth
invested in the crop production, H is the value (f t times Q f) of the hedge position divided
by initial wealth (not the value of the spot position), Rf is (f t+1 – f t)/f t and (1+R) is [(A
Yt+1Pt+1)/C(A)] one plus the rate of return on planting.
Given this, the farmer’s optimal risk management decision problem is to choose x
and H such that the expected utility of terminal wealth is maximized. As previously,
the decision problem is modeled with a general expected utility function. However, in
order to achieve analytically concise results, joint normality of R and Rf is invoked. This
leads to the following:
Significantly, while the derivation of Proposition 2.3 reveals that the individual
optimal solutions (denoted by *) to the farmer’s risk management problem (x*, H*)
depend on preferences, the ratio (H*/x*) only involves ex ante statistical parameters.
The portfolio-theoretic intuition behind the proposition is as follows: the farmer faces
two problems, one involving hedging, the other involving the scale of production. To
determine the fraction of the crop to hedge, the farmer must solve a portfolio problem
involving two risky “assets” with returns (R - r) and Rf. From mean-variance portfolio
theory, it is well known that if asset returns are jointly normal and riskless borrowing
and lending is permitted then all investors, regardless of preferences, hold the same
portfolio of risky assets. In addition, the ratio of any two assets in an optimal portfolio
will be independent of risk preferences. Since the farmer’s choice of the fraction of
initial wealth to invest in crop production (x) is unconstrained, as long as returns
are independent of the scale of production—and the other assumptions relevant to
Proposition 2.4 are satisfied—(H*/x*) will not involve preferences.
When used to analyze (H*/x*), the practical implication of Proposition 2.4 is
that the fraction of the investment in crop production to be hedged (Q f f t /C(A)) is
independent of the size of the crop. As a consequence, observed differences in the use
of hedges and other contracting methods by large and small farmers revolve around
the determination of R. Further, when the futures or forward price is unbiased (E[Rf]
= 0), only joint normality of R and Rf is required to motivate ordinary least squares as
the optimal hedge ratio estimation technique. Though similar types of conditions have
been derived for related problems, e.g. Benninga, et al. (1984), Heaney and Poitras (1991),
this result has not been recognized as applying to the farmer’s hedging problem with
both price and production uncertainty. On balance, Proposition 2.4 is of theoretical
significance because it establishes a connection between the results of portfolio theory
and the farmer’s hedging problem further clarifying a possible motivation for the use of
regression analysis to estimate the farmer’s optimal hedge ratio.
From these early beginnings, crop insurance in the USA and Canada has expanded
to where it is a key federal agricultural support program. In Canada, the Crop Insurance
Act (1959) was intended as a major step to stabilize farm incomes against production
related risks. In contrast to other regions such as Europe and Australia where private
markets write actuarially priced insurance, mostly for hail related damage, crop
insurance in Canada and the USA is subsidized and covers a wide range of crops
and perils.31 Recognizing that agriculture is a joint federal-provincial responsibility
in Canada, the extent and evolution of the subsidy in Canada is given in Agriculture
Canada (1998, p.3):
The CI Act of 1959 enabled the federal government to assist provinces in making
CI available to producers at a 60% coverage level. Originally the federal
government’s share of total premiums was 20%, with a 50% share of administrative
expenses. In 1964, the Act was amended to incorporate general provisions for a
reinsurance agreement between the provinces and the federal government . . .
amendment to the Act, in 1973, provided two options for the federal-provincial-
producer cost-sharing arrangements. In one option, the federal and provincial
governments each contributed 25% of total premiums and 50% of administrative
costs. In the other option, the federal government contributed a total of 50% of
premiums and the provinces paid all administrative costs. In the 1990 amendment,
the maximum coverage was increased to 90% for low risk crops. Furthermore,
the single cost-sharing formula was adopted, where the federal government and
provinces each pay 25% of total premiums and 50% of administration costs.
Other changes included waterfowl crop damage compensation, and regulations
concerning self-sustainability and actuarial soundness requirements.
policies for more than 100 crops. Policies typically consist of general crop insurance
provisions, specific crop provisions, policy endorsements and special provisions . . .
Policies are available for most commodities.” In contrast to Canadian programs that
focus on production insurance, the USA features policies with a variety of payout
features. In particular, a partial list of the policies the RMA offers involve the following:
actual production history; actual revenue history; adjusted gross revenue; “dollar plan”
crop yield; specified revenue protection; and specified yield protection. For example, a
crop insurance policy based on actual production history can be described as follows:
Actual Production History policies insure producers against yield losses due to
natural causes such as drought, excessive moisture, hail, wind, frost, insects, and
disease. The producer selects the amount of average yield to insure; from 50–75
percent (in some areas to 85 percent). The producer also selects the percent of the
predicted price to insure; between 55 and 100 percent of the crop price established
annually by RMA. If the harvested plus any appraised production is less than the
yield insured, the producer is paid an indemnity based on the difference. Indemnities
are calculated by multiplying this difference by the insured percentage of the price
selected when crop insurance was purchased and by the insured share.
Specific revenue protection policies have a different procedure for determining the price
used in calculating the indemnity. It is possible for farms to insure using individual or
area wide variables to determine the indemnity.
Analysis of commodity risk management in agriculture is complicated by the array
of government programs and subsidies aimed at mitigating significant production and
revenue risks. These various programs have a complementarity that make it difficult to
ascertain the commodity risk management activities of farmers. While the introduction
of the Federal Crop Insurance Act of 1980 generated academic discussion surrounding
the failure of private markets to provide a risk pooling solution, e.g. Knight and
Coble (1997), Myers and Oehmke (1988), the period leading up to passage of the Crop
Insurance Reform Act (CIRA) (1994) was characterized by a reluctance by many farmers
to participate in the crop insurance program. This reluctance coincided with Congress
authorizing disaster relief payments in eight of the ten years from 1985–94. The CIRA
required farmers to purchase a minimum amount of catastrophe insurance to qualify
for disaster relief programs. The CIRA was successful in doubling the number of acres
insured under crop insurance plans from 1993 to 1998.
Changes to US crop insurance since CIRA have focused on increasing the diversity of
policy types and increasing private-sector participation. These changes have coincided
with the appearance of academic studies concerned with empirically identifying farmer
attitudes to crop insurance, e.g. Ginder et al. (2009), Mishra and El-Osta (2002), Sherrick
et al. (2003), and to identifying sound insurance policy designs, e.g. Deng et al. (2007),
Mahul (1999), Mahul and Wright (2003), Miranda and Glauber (1997). Since 2000, when
the government increased the subsidy from about 42 percent of the premium to about
60 percent, farmers not only increased purchases of crop insurance, but also tended
to choose more expensive revenue insurance policies and at higher coverage levels. As
a consequence, US crop insurers paid out a record $9.1 billion in indemnities in 2011,
a total that could top $10 billion when all claims are settled. These claims originated
primarily from damage due to drought, flooding and freezing weather. The previous
record was $8.7 billion in 2008. This has led to pressure to introduce cost reduction in
the farm bill due in 2012.
Risk pooling mechanisms in agriculture can be traced back to the rural, agrarian
economies of antiquity. At least since Evans-Pritchard (1940), it has been recognized
that concerns surrounding economic and social security played an important role
in the institutions and practices traditional rural societies (Platteau 1997, p.764). In
turn, this has stimulated interest in “insurance” arrangements in the village economies
of the Third World, e.g. Jalan and Ravallion (1999), Ligon et al. (2002), Little et al. (2001),
Townsend (1995). Because these primitive “mutual insurance” schemes are second best
compared to a market solution, considerable attention has been given to developing crop
insurance programs that would be feasible for low-income, rural village economies, e.g.
Sakurai and Reardon (1997), Skees and Enkh-Amgalan (2002), Skees et al. (1999, 2002).
As with the US experience, the success of such programs depends on the level and type
of subsidization and the presence of competing government programs also aimed at
farm income support.
Sources of Information
Changes to accounting rules for publicly traded companies have dramatically altered the
availability of information on corporate risk management activities. Previous to these
259
Overview
Academic studies of commodity risk management usually find that activities such as
hedging are value increasing for firms. Some studies even report a point estimate for the
percentage value increase in value attributed to risk management activities. Other studies
suggest the use of active hedge adjustment based on hedge ratios that are empirically
estimated using advanced econometric techniques. In the process, theoretically derived
rationales for commodity risk management, such as the under-investment hypothesis,
are tested. The general message is that commodity risk management activities such as
hedging are beneficial and firms that do not employ such methods are, somehow, not
maximizing value for shareholders. Yet, detailed examination of the relevant sections of
the financial statements for major commodity producing and consuming corporations
reveals a different picture. In the process of detailing commodity risk management
practices at major firms, questions decidedly more relevant to the commodity risk
management decisions of non-financial firms become apparent.
Whether commodity risk management activities are “value enhancing” raises
the fundamental ex ante/ex post quandary that plagues “scientific” commodity risk
management. Completely accurate ex ante forecasts will produce optimal ex post
operating and financial results. However, complete accuracy is a difficult target to
achieve and some firms may opt to avoid the possibility of negative ex post outcomes. This
quandary is aptly recognized below by Southwest Airlines Company, an active hedger
of jet fuel price risk. In addition, management may correctly feel that shareholders want
exposure to the underlying commodity price and hedging the commodity price would
defeat this objective. Details for such a firm—Canadian Oil Sands Limited—are given
below. All firms examined below are acutely aware of the difficulties associated with
forecasting essential commodity prices but there is considerable diversity regarding the
approach to managing the commodity price risk. This diversity is reflected in the range
of instruments used to hedge the risk, including: OTC forward contracts; exchange
traded futures contracts; commodity swaps; “costless collars”; out-of-the money puts
(for producers) and calls (for consumers); leveraging; and basis swaps.
copper mining operations require significant energy, principally electricity, diesel, coal
and natural gas. Energy costs approximated 20 percent of our consolidated copper
production costs in 2010 and 2009, and included purchases of approximately 215
263
million gallons of diesel fuel; 6,100 gigawatt hours of electricity at our North America,
South America and Africa copper mining operations (we generate all of our power at
our Indonesia mining operation); 800 thousand metric tons of coal for our coal power
plant in Indonesia; and 1 million MMBTU (million British thermal units) of natural
gas at certain of our North America mines. For 2011, we estimate energy costs will
approximate 20 percent of our consolidated copper production costs.
Where derivative securities are employed, this usage is often associated with purchase
and sale contracts that have embedded forward contracting provisions. FCX is an
excellent example of these observations. The 2010 10-K for FCX gives the following
description of the company:
We are one of the world’s largest copper, gold and molybdenum mining companies
in terms of reserves and production. Our portfolio of assets includes the Grasberg
minerals district in Indonesia, significant mining operations in North and South
America, and the Tenke Fungurume (Tenke) minerals district in the Democratic
Republic of Congo (DRC). The Grasberg minerals district contains the largest
single recoverable copper reserve and the largest single gold reserve of any mine in
the world based on the latest available reserve data provided by third-party industry
consultants. We also operate Atlantic Copper, our wholly owned copper smelting
and refining unit in Spain.
Prior to the 2007 acquisition of Phelps Dodge, the primary asset of FCX was the
Grasberg mining operation. The current corporate structure and mine locations are
given in Figs 3.1 and 3.2. The structure of consolidated mining output to copper, gold
Figure 3.1 Freeport McMoRan (Ticker Symbol: FCX) Corporate Structure (2009)
and molybdenum, and reliance of FCX revenues, income and production to each
mineral are given in Table 3.1.
Copper (recoverable)
Production (millions of 3,908 4,103 4,030 3,884 3,639
pounds)
Production (thousands of 1,773 1,861 1,828 1,762 1,651
metric tons)
Sales, excluding 3,896 4,111 4,066 3,862 3,630
purchases (millions
of pounds)
Sales, excluding 1,767 1,865 1,844 1,752 1,647
purchases (thousands
of metric tons)
Average realized price $ 3.59 $ 2.60 $ 2.69 $ 3.22 $ 2.80
per pound
Gold (thousands of
recoverable ounces)
Production 1,886 2,664 1,291 2,329 1,863
Sales, excluding 1,863 2,639 1,314 2,320 1,866
purchases
Average realized price $ 1,271 $ 993 $ 861 $ 682 $ 566
per ounce
Molybdenum (millions of
recoverable pounds)
Production 72 54 73 70 68
Sales, excluding 67 58 71 69 69
purchases
Average realized price $ 16.47 $ 12.36 $ 30.55 $ 25.87 $ 21.87
per pound
(Continued)
FCX is acutely aware of the exposure of company operations to price of copper and,
to a lesser extent, gold and molybdenum (see Fig. 3.3 from 2010 10-K). The following
information is provided in Item 7A, Quantitative and Qualitative Disclosures about
Market Risk of the 10-K:
FCX does not purchase, hold or sell derivative financial instruments unless there
is an existing asset or obligation or if it anticipates a future activity that is likely
to occur and will result in exposure to market risks and FCX intends to offset or
mitigate such risks. FCX does not enter into any derivative financial instruments
for speculative purposes, but has entered into derivative financial instruments in
limited instances to achieve specific objectives. These objectives principally relate
to managing risks associated with commodity price, foreign currency and interest
rate risks. The fair values of FCX’s financial derivative instruments are based on
widely published market closing prices.
Though this statement has some of the typical boilerplate common in similar statements
made by mining companies also not engaged in management of commodity price
risk with derivative security contracts, FCX does make substantive use of forward
contracting provisions in purchase and sale contracts associated with “an existing
obligation.” FCX describes these contracting provisions as:
For 2010, 52 percent of our mined copper was sold in concentrate, 26 percent as
cathodes and 22 percent as rod (principally from our North America copper mines).
Substantially all of our copper concentrate and cathode sales contracts provide final
copper pricing in a specified future period (generally one to four months from the
shipment date) based primarily on quoted LME monthly average spot prices.
In effect, FCX receives a market price that is based on prices in a specified future period.
It records revenues and invoices customers at the time of shipment based on then-
current LME prices. In the accounting process, the resulting embedded derivative in
the forward sale contract is adjusted to fair value through earnings each period, using
the period-end forward prices, until the date of final pricing “some one to four months
from the shipment date.” FCX recognizes that the embedded derivative will impact
earnings:
certain FCX copper concentrate, copper cathode and gold sales contracts provide
for provisional pricing primarily based on LME or COMEX prices (copper) and the
London Bullion Market Association price (gold) at the time of shipment as specified
in the contract. Similarly, FCX purchases copper and molybdenum under contracts
that provide for provisional pricing (molybdenum purchases are generally based
on an average Metals Week Molybdenum Dealer Oxide price) . . . the contracts
do not allow for net settlement and always result in physical delivery. Sales and
purchases with a provisional sales price contain an embedded derivative (i.e. the
price settlement mechanism that is settled after the time of delivery) that is required
to be bifurcated from the host contract. The host contract is the sale or purchase of
the metals contained in the concentrates or cathodes at the then-current LME or
COMEX price (copper), the London Bullion Market Association price (gold) or the
average Metals Week Molybdenum Dealer Oxide price (molybdenum) as defined
in the contract. Mark-to-market price fluctuations recorded through the settlement
date are reflected in revenues for sales contracts and in cost of sales as production
and delivery costs for purchase contracts.
The actual amount of metal production covered by such embedded derivative contracts
is indicated in Table 3.2. As of the end of 2010, FCX had no outstanding positions or
credit exposure in financial derivative securities for either foreign exchange or interest
rates. The bulk of outstanding long-term debt is fixed-rate US dollar and, despite having
a sizable revolving credit facility, only a small amount of the facility has been drawn
down.
Embedded derivatives in
provisional sales contracts:
Copper (millions of pounds) 663 $ 3.84 $ 4.36 May 2011
Gold (thousands of ounces) 297 1,382 1,411 March 2011
Embedded derivatives in
provisional purchase contracts:
Copper (millions of pounds) 210 3.82 4.37 April 2011
Molybdenum (thousands of 245 15.28 15.71 January
pounds) 2011
Table 3.3 Capstone Mining Metal Sales (as of Sept. 10, 2010)
Comparative Current Comparative
Metal sales Current Quarter Quarter Period Period
Copper – pounds
Cozamin 7,567,970 8,745,375 22,699,391 26,505,256
Minto 15,558,755 15,847,081 39,266,595 43,211,052
23,126,725 24,592,456 61,965,986 69,716,308
Realized copper price $3.39 $2.65 $3.33 $2.19
Average copper price $3.29 $2.64 $3.25 $2.12
Zinc – pounds
Cozamin 3,826,946 2,908,375 10,173,557 10,364,604
Lead – pounds
Cozamin 1,757,434 2,800,766 7,927,178 6,561,040
During October 2011 the Company entered into additional copper forward purchase
contracts at the corporate level to offset its remaining outstanding copper forward
sales contracts. This decision was made to allow the Company to participate in
any future copper price increases. As at October 31, 2011, 100% of the outstanding
copper forward sales contracts had been offset. The offsetting copper forward
purchase contracts locked in an approximate $1.0 million gain on an equivalent
number of copper forward sales contracts and provide the Company exposure to
any copper price movement going forward. The locked in gain will be included in
the results of operations for the three months ended December 31, 2011.
In addition to rolling out of all longer dated forward sales contracts leaving full exposure
to commodity price risk, the company is rapidly expanding equity capital to eliminate
long-term debt and acquire new development assets with the purchase of Far West
Mining Ltd in June 2011:
Far West Mining Ltd (“FWM”), a 70% owned Canadian subsidiary, owns 100% of
Minera Lejano Oeste, which is advancing the Santo Domingo project, a large scale
copper-iron-gold project (“Santo Domingo”) in Chile to a production decision. In
addition FWM owns active exploration properties in Australia. Kutcho Copper
Corp. (“Kutcho Copper”), a wholly-owned Canadian subsidiary of the Company,
is advancing the Kutcho copper-zinc-silver-gold project (the “Kutcho Project”) in
British Columbia towards a production decision.
Capstone has no substantial interest rate risk to hedge. The considerable FX risk is
related to the US$, C$ and Mexican peso. Capstone does not hedge FX risk.
1. Non-GAAP financial measure – see pages 78–85 of the 2010 Financial Report.
2. In July 2010, Barrick increased its dividend by 20% to $0.12 per share on a quarterly basis;
based on converting the previous semi-annual dividend of $0.20 per share to a quarterly
equivalent.
Perhaps the most widely examined of such hedge book changes occurred with
Barrick Gold, a company that, in the early 1990s, claimed great success for a hedging
program that was ultimately unwound at considerable cost. While the FCX Grasberg
operation can lay claim to being the largest gold mine currently in production, Barrick
Gold (ABX) can lay claim to being the single largest gold producer, with the largest
market capitalization and proven reserves. ABX became a public company in May 2,
1983, when it was listed on the Toronto Stock Exchange. With 140 million ounces of
Figure 3.4 Barrick Mines in North America and South America, 2010
proven and probable gold reserves as of December 31, 2010, ABX also has 6.5 billion
pounds of proven and probable reserves of copper and 1.07 billion ounces of proven
and probable reserves of silver as of December 31, 2010. The headquarters of ABX are in
Toronto, Ontario. Currently Barrick operates in Australia, Africa, North America and
South America with a portfolio of 26 operating mines (see Figs 3.4 and 3.5).
The majority of Barrick’s current production comes from North and South America
(see Fig. 3.6).
Based on Barrick reserves it appears that the percentage of Barrick’s total production
coming from South America will increase in the future (see Figures 3.6 and 3.7).
output was formally announced in November 2003 but was apparent to the market as
early as February 2002. “In 2003 we adopted a no hedge policy” (ABX Annual Report
2003). At the time of the formal announcement, the accumulated Barrick hedge book
was 16.3 million oz., reduced from the hedge book entering 2003 of 18.1 million oz. By
comparison, in 2003, Barrick produced 5.51 million oz. of gold at a total cash cost of
$189/oz. Foster (2003) estimates the global hedge book for gold to be 71.6 million oz.,
making this reversal an omen for “the death of hedging” by gold producers. In 1998,
a year where Barrick produced 3.2 million oz. of gold, the Annual Report gives the
following description of hedging activities using forward sales:
As part of its gold hedging program, the Company has entered into spot deferred
contracts with several major financial institutions to deliver 11.5 million ounces of
gold. A spot deferred contract represents a forward sale on which contango accrues
until the intended delivery date of the contract. The rate at which contango accrues
is determined by LIBOR interest rate less the gold lease rate existing at the time of
each rollover. The contracts have an average price of $357 per ounce at December
31, 1998. The Company’s expected gold production is fully hedged over the next two
years and it has further contracts in place designated from 2001 to 2004. Delivery
under these spot deferred contracts can be deferred at the Company’s option for up
to 15 years.
Barrick’s gold hedging strategy extended beyond forward sales contracts to include
options trading. For example, in 1998 the following positions were in place:
In addition, the Company has entered into long-term written gold call options in
respect of 1.7 million ounces. The call options have an average strike price of $380
per ounce and expire at various dates over the period from 2000 to 2007. In the
event that they are exercised at their maturity date, the Company has the intent and
ability to convert them into spot deferred contracts at the strike price.
In 2009, Barrick used $3.4 billion of proceeds from a stock sale to buy back all of its
fixed-price contracts and the majority of its floating spot price contracts. From 2004
to 2009, Barrick’s hedge book liabilities more than doubled, rising to $5.6 billion from
$1.9 billion. Moving into an unhedged position allowed Barrick’s gross margins
per ounce to be more leveraged to the price of gold.
This is followed by a discussion that identifies the key risks facing the company. The
first such risk is self-evident and of interest, given the past track record for Barrick gold
hedging:
Significantly, Barrick follows this with a discussion of other significant risks that
does not include foreign currency exchange rates, interest rate risks, diesel fuel, and
the price risks associated with byproduct minerals, mostly silver and copper. Instead,
the enterprise-wide risk management at Barrick identifies three other sources of
risk: re-licensing risk, project development risk and global economic conditions. On
these items, the Barrick Annual Report for 2010 identifies re-licensing risk with more
traditional risk management concerns in mining:
Re-licensing risk
Maintaining our social license to operate is critical for Barrick to operate our existing
mines and develop our projects around the world. Some of the risks to our social
license include: compliance with environmental laws and regulations; community
relations and human rights issues; and the health and safety of our employees. To
manage these risks and maintain our social license, we have developed global
environmental standards which, in many cases, exceed regulatory requirements
and represent industry best practice. We have a globally coordinated community
relations strategy that utilizes our corporate and local expertise to improve
relations in the communities in which we operate. We have recently joined the
Voluntary Principles on Security and Human Rights and are undertaking two new
corporate social responsibility (“CSR”) initiatives to further strengthen our CSR
performance . . . Additionally, we have an extensive Safety and Health program,
committed to the protection of our employees and the residents of communities in
proximity to our operations.
The enterprise risk management solution for re-licensing risk is focused on expanding
the corporate social responsibility profile of Barrick gold.
Another element of traditional risk management in mining is a concern with project
development risk:
the ultimate operating cost of the relevant project. Our Capital Projects group is
responsible for completing relevant studies, obtaining the necessary approvals
and managing construction. We utilize a formal system to govern advancement of
projects as they progress from scoping through the execution and commissioning
stages.
Finally, in the discussion of global economic conditions, the present elements of Barrick
commodity risk management activities are identified:
As it turns out, Barrick is still an active hedger of commodity risk associated with
commodity inputs, interest rates and foreign exchange rates.
At present, our interest rate exposure mainly relates to interest receipts on our cash
balances ($4.0 billion at the end of the year); the mark-to-market value of derivative
instruments; the fair value and ongoing payments under US dollar interest-
rate swaps; and to the interest payments on our variable-rate debt ($1.0 billion at
December 31, 2010). Currently, the amount of interest expense recorded in our
Currently, Barrick has foreign currency contracts for non-US expenditures, in total
of $4,488 million AUD, $364 million CAD, 211 billion CLP and 35 million EUR.
These are designated as cash flow hedges of anticipated operating, administrative,
sustaining capital and project capital spend.
Significantly, Barrick reports financial results in US$ (not C$). As illustrated in Table
3.8, Barrick hedges a significant amount of expected currency risk. Barrick describes
the general risk as:
The rationales for hedging currency exposure are identified as reduced US$ volatility for
expected operating and capital expenditures in that currency during the hedge period.
For the two largest FX positions, the A$ and C$, Barrick reports:
Barrick also observes that the currency hedging strategy has produced positive results:
Our currency hedge position has provided benefits to us in the form of hedge gains
recorded within our operating costs when contract exchange rates are compared
to prevailing market exchange rates as follows: 2010—$146 million; 2009—
$27 million; and 2008—$106 million. For 2010, we also recorded currency hedge
gains in our corporate administration costs of $33 million (2009—$7 million loss
and 2008—$11 million gain).
While Barrick seeks full gold price risk exposure, at the same time it hedges commodity
inputs such as diesel, electricity, propane and natural gas, as well as its production
outputs for both copper and silver. Regarding diesel fuel: “On average we consume
11/16/2012 6:45:29 PM
Book 1.indb 281
Table 3.10 Fair Value of Derivatives ($ millions)
Fairs Values of Derivative Instruments
Asset Derivatives Liability Derivatives
Consolidated Fair Value at Fair Value at Fair Value Consolidated Fair Value at Fair Value at Fair Value
Balance Sheet September December at January Balance Sheet September December at January
Classification 30, 2011 31, 2010 1, 2010 Classification 30, 2011 31, 2010 1, 2010
11/16/2012 6:45:29 PM
Book 1.indb 282
Table 3.11 Cash Flow on Hedges ($ millions)
Cash Flow Hedge Gains (Losses) in OCI
Commodity price hedges Currency hedges Interest rate hedges
Gold/ Operating Administration/ Capital Long-term
Silver Copper Fuel costs other costs expenditures debt Total
11/16/2012 6:45:29 PM
Mining Companies 283
approximately 3.8 million barrels of diesel fuel annually across all our mines. Diesel fuel
is refined from crude oil and is therefore subject to the same price volatility affecting
crude oil prices.” The hedging strategy for diesel fuel is described as follows:
Volatility in crude prices has a significant direct and indirect impact on our
production costs. To mitigate this volatility, we employ a strategy of combining the
use of financial contracts and our production from Barrick Energy to effectively
hedge our exposure to high oil prices. We currently have financial contracts in
place totaling 4.7 million barrels, which represents 56% of our total estimated direct
consumption in 2011 and 34% of our total estimated direct consumption over the
following two years. Those contracts are primarily designated for our Nevada-
based mines, and have average prices below current forward prices. In 2010, we
recorded hedge losses in earnings of approximately $28 million on our fuel hedge
positions (2009: $97 million loss; 2008: $33 million gain). Assuming market rates
at the December 31, 2010 level of $91 per barrel, we expect to realize hedge gains of
approximately $20 million in 2011 from our financial fuel contracts.
To this hedge and other energy related expenses, Barrick was engaged in the following:
During the year, we entered into 480 thousand barrels of WTI/ULSD crack spread
swaps, 1,222 thousand barrels of MOPS forwards, 228 thousand barrels of WTB
forwards, 228 thousand barrels of JET forwards, and 19 million gallons of propane
designated against forecasted fuel purchases for expected consumption at our mines.
The designated contracts act as a hedge against variability in market prices on the
cost of future fuel purchases over the next four years.
In this case, the cost of downside protection is partially offset by having a larger volume
of copper associated with the ceiling than with the floor. However, the ceiling is further
out of the money than the floor resulting in a net collar expense, i.e. the hedge is not
constructed as a “costless collar” commonly observed in oil and gas producer hedging
strategies.
Silver production is incidental to Barrick gold operations. More precisely, “silver
prices have a significant impact on the overall economics and expected gold total cash
costs for our Pascua-Lama project, which is currently in the construction phase. Silver
prices do not significantly impact our current operating earnings, cash flows or gold
total cash costs.” Given, there are two items of interest in the Barrick treatment of
commodity risk management of silver. The first concerns the hedging method used:
In the fourth quarter, utilizing zero-cost option collar strategies, we took advantage
of high spot silver prices and attractive option pricing by adding hedge protection
on three million ounces per year of expected silver production from 2013 to 2017,
inclusive, with a floor price of $20 per ounce and an average ceiling price of $55 per
ounce.
Presumably, by taking advantage of “high” silver spot prices, Barrick is not averse to
selective hedging. Another interesting commodity risk management feature is the
agreement reached with Silver Wheaton:
In 2009, we entered into a transaction with Silver Wheaton Corp. (“Silver Wheaton”)
whereby we sold 25% of the life-of-mine Pascua-Lama silver production from the
World-class assets
Production
Total Petroleum production (millions of barrels of 159.38 158.56 137.97
oil equivalent)
Alumina (’000 tonnes) 4,010 3,841 4,396
Aluminium (’000 tonnes) 1,246 1,241 1,233
Copper (’000 tonnes) 1,139.4 1,075.2 1,207.1
Nickel (’000 tonnes) 152.7 176.2 173.1
Iron ore (’000 tonnes) 134,406 124,962 114,415
Manganese alloys (’000 tonnes) 753 583 513
Manganese ores (’000 tonnes) 7,093 6,124 4,475
Metallurgical coal (’000 tonnes) 32,678 37,381 36,416
Energy coal (’000 tonnes) 69,500 66,131 66,401
Aluminium
Ref Country Asset Description Ownership
Base Metals
Ref Country Asset Description Ownership
Uranium(b)
Ref Country Asset Description Ownership
Iron Ore
Ref Country Asset Description Ownership
Manganese
Ref Country Asset Description Ownership
Metallurgical Coal
Ref Country Asset Description Ownership
Energy Coal
Ref Country Asset Description Ownership
governance given in the general description of group risk management in BHP Billiton’s
2011 annual report:
Risk management
We believe that the identification and management of risk is central to achieving
the corporate objective of delivering long-term value to shareholders. Each year,
the Board reviews and considers the risk profile for the whole business. This risk
profile covers both operational and strategic risks. The Board has delegated the
oversight of risk management to the Risk and Audit Committee. In addition, the
Board specifically requires the CEO to implement a system of control for identifying
and managing risk. The Directors, through the Risk and Audit Committee, review
the systems that have been established for this purpose and regularly review their
effectiveness.
Tables 3.16 and 3.17, and Fig. 3.8, describe the risk management methods and risk
governance structure of BHP. Significantly, accountability for risk management
Aluminium 7 30 21 26
Copper 111 102 83 84
Zinc 2 2 29 19
Lead 6 8 40 26
Silver 18 27 4 9
Nickel 25 13 47 36
Iron ore 2 5 – 2
Energy coal 16 41 21 31
Metallurgical coal – – – 2
Petroleum 4 24 – 33
Gas 129 52 111 31
Freight 24 7 38 13
Other – – – 3
Total 344 311 394 315
Comprising:
Current 189 232 241 223
Non-current 155 79 153 92
activities lies at the Group Management committee and related sub-committees that
report to the CEO and, through the CEO, to the Board of Directors. This is in contrast
to alternative risk management structures that make risk management a subcommittee
of the Board with the CEO as a member, reporting directly to the Board.
Another significant aspect of Table 3.16 is the explicit use of CFaR and VaR methods
to measure risk. Precisely how the CFaR methodology is employed and critical values
obtained is not discussed. Rather, a general discussion of the methodology is provided:
A Cash Flow at Risk (CFaR) framework is used to measure the aggregate and
diversified impact of financial risks upon the Group’s financial targets. The principal
measurement of risk is CFaR measured on a portfolio basis—which is defined as
the worst expected loss relative to projected business plan cash flows over a one-
year horizon under normal market conditions at a confidence level of 95 per cent.
The CFaR framework includes Board-approved limits on the quantum of the CFaR
relative to the Group’s financial targets.
Describing the risk management methods employed, BHP observes: “In executing the
strategy, financial instruments are potentially employed in three distinct but related
activities.” These activities are outlined in Table 3.16 as: risk mitigation; economic
hedging; and strategic transactions to benefit from market mis-pricing.
BHP provides a conventional classification of risks into: market risk, liquidity risk
and credit risk:
The financial risks arising from the Group’s operations comprise market, liquidity
and credit risk. These risks arise in the normal course of business, and the Group
manages its exposure to them in accordance with the Group’s Portfolio Risk
Management Strategy. The objective of the strategy is to support the delivery of the
Group’s financial targets while protecting its future financial security and flexibility
by taking advantage of the natural diversification provided by the scale, diversity
and flexibility of the Group’s operations and activities.
Of these risks, the most important to firm performance is market risk. BHP describes
market risk as:
BHP Billiton manages risk using portfolio risk management strategies and it
operates within the overall cash flow at risk (CFaR) limits imposed by the board of
directors. BHP does not actively engage in currency and commodity hedging because
it believe that this provides no long-term benefits to its shareholders. Instead BHP
employs a financial risk management strategy that incorporates three distinct
activities.
(emphasis added)
In effect, BHP does not actively hedge commodity price or currency risk because it has
a view that such activities do not add value for shareholders.
Contracts for the sale and physical delivery of commodities are executed
whenever possible on a pricing basis intended to achieve a relevant index target.
Where pricing terms deviate from the index, derivative commodity contracts
are used when available to return realised prices to the index. Contracts for the
physical delivery of commodities are not typically financial instruments and
are carried in the balance sheet at cost (typically at nil); they are therefore
excluded from the fair value and sensitivity tables . . . Accordingly, the
financial instrument exposures . . . do not represent all of the commodity price
risks managed according to the Group’s objectives. Movements in the fair value
of contracts . . . are offset by movements in the fair value of the physical
contracts, however only the former movement is recognised in the Group’s
income statement prior to settlement. The risk associated with commodity prices
is managed as part of the Portfolio Risk Management Strategy and within the
overall CFaR limit.
The relevance of omitting of contracts for physical delivery from reporting of com-
modity risk management activities is apparent in the process used for pricing
iron one, the largest commodity produced by BHP. Together with Vale and Rio Tinto,
BHP controls about two-thirds of the seaborne trade in iron ore. Until the first half
of 2010, for about four decades iron ore was priced using fixed price contracts that
were set on an annual basis.2 In 2010, led by the vocal campaign of BHP Billiton
CEO Marius Kloppers, the pricing window for selling iron ore to China, the largest
consumer of iron ore, was changed to quarterly using observed market price indexes
from the month prior to the beginning of a new quarter. Market reports indicate that
BHP, reacting to observed discrepancies between spot prices and the quarterly price
settings, may have moved during 2011 to monthly and spot pricing on some iron
ore contracts. These activities, which are fundamentally related to commodity risk
management, are not examined in the commodity price risk management part of the
Annual Report.
The commodity price risk reporting that is available in the Report is described
following Table 3.16 as:
All such instruments are carried in the balance sheet at fair value.
Not included in [Table 3.18] are provisionally priced sales volumes for which
price finalisation, referenced to the relevant index, is outstanding at balance date.
Provisional pricing mechanisms embedded within these sales arrangements have
the character of a commodity derivative and are carried at fair value as part of trade
receivables.
This, presumably, includes the reporting for iron ore and other commodity contracts that
“involve physical delivery.” Based on Table 3.18, sensitivity analysis for these commodity
price exposures are reported in Table 3.19. The calculation of these exposures is
described as follows:
The sensitivities in [Table 3.18] have been determined as the absolute impact on
fair value of a 10 per cent increase in commodity prices at each reporting date,
while holding all other variables, including foreign currency and exchange rates,
constant. The relationship between commodity prices and foreign currencies is
complex and movements in foreign exchange can impact commodity prices. The
sensitivities should therefore be used with care.
Given the method of determining items that are considered relevant for “commodity
risk management,” the items identified in Tables 3.18 and 3.19 are small compared to the
overall scope of BHP corporate activities as illustrated in Table 3.20, which details the
cash from operations for BHP.
BHP Billiton is exposed to exchange rate transaction risk on foreign currency sales
and purchases. It faces translational exposure in respect of non-functional currency
monetary items and transactional exposure in respect of non-functional currency
expenditure and revenues.
Operating activities
Profit before taxation 31,255 19,572 11,617
Adjustments for:
Non-cash exceptional items (150) (255) 5,460
Depreciation and amortisation expense 5,039 4,759 3,871
Net gain on sale of non-current assets (41) (114) (38)
Impairments of property, plant and equipment, 74 35 190
financial assets and intangibles
Employee share awards expense 266 170 185
Financial income and expenses 561 459 543
Other (384) (265) (320)
Changes in assets and liabilities:
Trade and other receivables (1,960) (1,713) 4,894
Inventories (792) (571) (116)
Trade and other payables 2,780 565 (847)
Net other financial assets and liabilities 46 (90) (769)
Provisions and other liabilities 387 (306) (497)
Cash generated from operations 37,081 22,246 24,173
Dividends received 12 20 30
Interest received 107 99 205
Interest paid (562) (520) (519)
Income tax refunded 74 552 –
Income tax paid (6,025) (4,931) (5,129)
Royalty related taxation paid (607) (576) (906)
Net operating cash flows 30,080 16,890 17,854
other than the functional currency of an operation are periodically restated to US dollar
equivalents, and the associated gain or loss is taken to the income statement.”
Other than conventional depletion, exploration and project development risks
associated with mining, the key element in the strategic risk of BHP appears as the
second item in the list of risk factors identified in the 2011 Annual Report:
Reduction in Chinese demand may negatively impact our results. The Chinese
market has become a significant source of global demand for commodities. In
CY2010, China represented 59 per cent of global seaborne iron ore demand, 39 per
cent of copper demand, 38 per cent of nickel demand, 41 per cent of aluminium
demand, 42 per cent of energy coal demand and 10 per cent of oil demand. China’s
demand for these commodities has been driving global materials demand over the
past decade.
The central importance to BHP of the Asian commodities market, in general, and the
China trade in particular, is given in Table 3.22. By moving to monthly and, possibly,
daily price settings for iron ore and other commodities, BHP is increasing, not
decreasing, exposure to commodity price risk associated with Chinese consumers. Given
Functional currency of
Group operation
US dollars – (4,344) 187 23 (1,414) (5,548)
Australian dollars (1) – – – – (1)
UK pounds sterling (3) – – – – (3)
(4) (4,344) 187 23 (1,414) (5,552)
Functional currency of
Group operation
US dollars – (1,398) 90 31 (942) (2,219)
Australian dollars (1) – – – – (1)
UK pounds sterling 4 – – – – 4
3 (1398) 90 31 (942) (2,216)
Financial risks
Consistent with conventional financial risk management, BHP identifies and discusses
separately and in some detail liquidity risk, credit risk and interest rate risk (see Tables
3.16 and 3.23):
BHP Billiton is exposed to interest rate risk on its outstanding borrowings and
investments from the possibility that changes in interest rates will affect future
cash flows or the fair value of fixed interest rate financial instruments. BHP
Billiton enters into interest rate swaps and cross currency interest rate swaps to
convert most of the centrally managed debt into US dollar floating interest rate
exposures.
Given a strong corporate financial position, BHP is able to access the long-term fixed-
rate debt market and achieve a variable interest rate expense gain by doing fixed to
floating interest rate swaps. In addition, BHP describes liquidity risk as:
Liquidity risk for BHP Billiton arises from the possibility that it may not be able to
settle or meet its obligations as they come due. Additionally, liquidity risk arises on
debt related derivatives due to the possibility that a market for derivatives might
not exist in some circumstances. However, Moodys Investors service has long-term
credit rating of A1 for BHP Billiton and Standard & Poors has a long-term credit
rating of A+ for BHP Billiton.
In effect, BHP views strength of credit rating and the associated ability to access
short-term borrowing as the solution to liquidity risk. The possibility of liquidity
considerations impacting the ability of customers to make contracted payments is not
considered relevant enough to mention.
Canadian Oil Sands prefers to remain un-hedged on crude oil prices; however,
during periods of significant capital spending and financing requirements,
management has in the past, and may again, hedge prices and exchange rates to
reduce revenue and cash flow volatility to the Corporation. Canadian Oil Sands
did not have any crude oil price hedges in place for 2010 or 2009. Instead, a strong
balance sheet was used to mitigate the risk around crude oil price movements. As
at February 23, 2011, and based on current expectations, the Corporation remains
un-hedged on its crude oil price exposure.
(Canadian Oil Sands Annual Report 2010)
Company Description
Though the Syncrude consortium was formed in 1964, the history of Canadian Oil
Sands Limited is more recent. The company began operation on November 30, 1995,
when an operating subsidiary, Athabasca Oil Sands Investment Inc., acquired an 11.74
percent interest in the Syncrude project. On June 26, 1996, the company’s other operating
subsidiary, Canadian Oil Sands Investments, Inc., acquired a 10 percent interest in
Syncrude from Pan-Canadian Petroleum Limited. In a one-for-one exchange of units,
301
on July 5, 2001, these two subsidiaries were merged into the single entity Canadian Oil
Sands Limited (COS), which then held 21.74 percent of Syncrude.1 Shortly thereafter,
the current President and CEO of COS, Marcel Coutu, was appointed by the board
of directors. Prior to this, Coutu had accumulated more than 20 years of oil industry
experience, though little of this had been at the senior executive level. Coutu came to
COS after a two-year stint as CFO at Gulf Canada, which was taken over by Conoco
just prior to his departure. Prior to being at Gulf Coutu worked at Trans-Canada
Pipelines (TRP) where he attained a position at the head of the international unit. In
February 2003, COSL successfully completed the purchase of an additional 10 percent of
the Syncrude project from EnCana with an option to buy EnCana’s remaining
5 percent share, which was subsequently exercised. Coutu is still the CEO and public
face of COS.
COS was initially organized as a unit trust but, as a result of a change in federal tax
policy, converted back to a corporation on December 31, 2010. One shortcoming of the
unit trust structure often identified by critics was the weak management and governance
structures of such entities. Whether this is the case at COS, which is effectively a non-
operating company, is unclear. In addition to Coutu, the management team at COS
is relatively small. Given that the activities of COS are not directly involved with the
production of oil at Syncrude, this is not surprising. However, the management of COS
does have significant activities that relate primarily to the running of the trust and
in the marketing of SCO. There is a Chief Operations Officer on staff with extensive
knowledge of bitumen mining and heavy oil upgrading operations. The implementation
of administrative and general management activities for the trust within COS began
in 2002. Prior to this time, this aspect of COS operations was conducted under an
Administrative Services Agreement with EnCana (and its predecessors). While
moving these activities within COS did result in some cost savings to unit holders,
the demand on management was such that: “In 2006 Syncrude Canada Ltd. entered a
Management Services Agreement with Imperial Oil Resources. It provides Syncrude
with operational, technical and business management services” (COS Annual Report
2008). The responsibility for the marketing of SCO by COS stems from the Syncrude
joint venture partnership agreement where Syncrude is responsible for delivering SCO
to each consortium member “at the plant gate.”
Being initially organized as a unit trust, the impact that the management of COS had
on the income that was generated for unit holders came largely through participation in
shaping the development of the Syncrude project. The passive character of the business
was well suited to the classical unit trust, making it difficult for trust management to
expand the underlying business with available resources due to a stated desire to pay
out a significant portion of operating cash flow net of capital expenditures. While there
was some scope for oil and gas royalty trusts to issue units or use cash flow to purchase
additional properties to offset depletion in current properties, the legal intention of the
trust structure was that trusts would be passive investors. As the trust sector expanded,
especially into business trusts, the notion of passive management was increasingly
ignored. Corporate tax and other advantages associated with unit trust issues meant
There are a number of risks that could impact Canadian Oil Sands’ net income
and cash from operating activities and, therefore, the dividends ultimately paid
to Shareholders. Cash from operating activities is highly sensitive to a number of
factors including: Syncrude production; sales volumes; oil and natural gas prices;
price differentials; foreign currency exchange rates; operating, administrative, and
financing expenses; non-production costs; Crown royalties; and regulatory and
environmental risks. Dividends may also be impacted by Canadian Oil Sands’
financing requirements for capital expenditures.
(COS Annual Report 2010, p.33)
Implications of the COS no-hedging policy appear from comparing results for 2010–11
with those for 2008–09 (see Tables 3.25 and 3.26). Comparing 2008Q1 with 2009Q1 is
particularly compelling. There were considerable cash costs to shareholders from the
management decision to not lock in the higher prices that prevailed during 2007–08, or
buying downside protection with puts. In addition to the no-hedging policy for the oil
price, COS provides the following statement about hedging of natural gas, an important
commodity input to production of SCO:
Sales1 ($ millions) $ 989 $ 1,045 $ 1,016 $ 912 $ 692 $ 843 $ 734 $ 863
Net income ($ millions) $ 242 $ 346 $ 324 $ 575 $ 193 $ 245 $ 176 $ 96
Per Share, Basic & Diluted $ 0.50 $ 0.71 $ 0.67 $ 1.19 $ 0.40 $ 0.51 $ 0.36 $ 0.20
Cash flow from operations2 ($ millions) $ 512 $ 544 $ 478 $ 398 $ 230 $ 379 $ 225 $ 366
Per Share2 $ 1.06 $ 1.12 $ 0.99 $ 0.82 $ 0.48 $ 0.78 $ 0.46 $ 0.76
Dividends ($ millions) $ 145 $ 145 $ 97 $ 242 $ 242 $ 242 $ 170 $ 169
Per Share $ 0.30 $ 0.30 $ 0.20 $ 0.50 $ 0.50 $ 0.50 $ 0.35 $ 0.35
Daily averages sales volumes3 (bbls) 109,260 102,938 120,894 114,739 96,477 118,569 99,286 119,287
Realized SCO selling price ($/bbl) $ 97.89 $ 111.00 $ 93.04 $ 83.97 $ 77.94 $ 78.07 $ 82.06 $ 78.67
Operating expenses4 ($/bbl) $ 37.19 $ 37.07 $ 35.53 $ 35.81 $ 37.97 $ 30.86 $ 37.94 $ 30.18
Purchased natural gas price ($/GJ) $ 3.51 $ 3.62 $ 3.59 $ 3.45 $ 3.44 $ 3.68 $ 4.95 $ 4.33
West Texas Intermediate5 (avg $US/bbl) $ 89.54 $ 102.34 $ 94.60 $ 85.24 $ 76.21 $ 78.05 $ 78.88 $ 76.13
Foreign exchage rates ($US/SCdn)
Average $ 1.02 $ 1.03 $ 1.02 $ 0.99 $ 0.96 $ 0.97 $ 0.96 $ 0.95
Quarter-end $ 0.96 $ 1.04 $ 1.03 $ 1.01 $ 0.97 $ 0.94 $ 0.98 $ 0.96
11/16/2012 6:45:32 PM
Canadian Oil and Gas Exploration and Development 305
(1) The Trust’s volumes differ from its production volumes due to changes in inventory, which are
primarily in-transit pipeline volumes, and are net of purchased crude oil volumes.
(2) Pricing obtained from Bloomberg.
(1) The Trust’s sales volumes differ from its production volumes due to changes in inventory,
which are primarily in-transit pipeline volumes, and are net of purchased crude oil volumes.
(2) Pricing obtained from Bloomberg.
The no-hedging policy for outputs is also applied to commodity inputs. Fortunately for
COS, the “de-linking of crude oil and natural gas price movements” has worked to the
advantage of COS due to natural gas prices falling both absolutely and relative to crude
oil prices.
In addition to a no-hedging policy for commodity inputs and outputs, the no-hedging
policy extends to currency hedging:
In the past, the Corporation has hedged foreign currency exchange rates by entering
into fixed rate currency contracts. The Corporation did not have any foreign
currency hedges in place during the first nine months of 2011 or 2010, and does not
currently intend to enter into any new currency hedge positions. The Corporation
may, however, hedge foreign currency exchange rates in the future, depending on
the business environment and growth opportunities.
What COS does do is to finance $1.275 billion in long-term debt in US dollars, which
provides a significant partial natural hedge when compared to the just over $7 billion
book value of assets:
Our revenue exposure [to changes in the US to Canadian dollar exchange rate]
is partially offset by US dollar obligations, such as interest costs on US dollar
denominated long-term debt (Senior Notes) and our share of Syncrude’s US dollar
vendor payments. In addition, when our US dollar Senior Notes mature, we have
exposure to US dollar exchange rates on the principal repayment of the notes. This
repayment of US dollar debt acts as a partial economic hedge against the US dollar
denominated revenue payments we receive from our customers.
Canadian Oil Sands’ net income and cash flow from operations are impacted by U.S.
and Canadian interest rate changes because our credit facilities and investments are
In addition to having no foreign currency transaction hedges, COS does not have “a
significant exposure to interest rate risk.” Following Emm et al. (2007), COS provides
sensitivity analysis for measuring risk exposure (sees Tables 3.27 and 3.28). COS does
engage in two substantive risk management practices for credit risk: “Canadian Oil
Sands carries credit insurance to help mitigate a portion of the impact should a loss
occur and continues to transact primarily with investment grade customers.”
1 An opposite change in each of these variables will result in the opposite cash from operating
activities impacts.
Canadian Oil Sands may become subject to minimum Crown royalties at a rate of 1 per cent of
gross bitumen revenue.
The sensitivities presented herein assume royalties are paid at 25 per cent of net bitumen
revenue.
1 An opposite change in each of these variables will result in the opposite cash flow from opera-
tions impacts. Canadian Oil Sands anticipates $nil current taxes in 2011. As such, the sensitivities
in the table above do not reflect any impact for current taxes.
Penn West considers price hedging of oil and natural gas production to be a useful
tool of risk management. Its uses include protecting planned capital budgets,
safeguarding the economics of acquisitions and providing downside cash flow
protection to support planned distributions.
(Penn West Annual Report 2010)
Company Overview
Penn West Exploration is based in Calgary, Alberta, with operations in North East BC,
Central Alberta, Southern Saskatchewan and Manitoba. In 2011, Penn West output was
63 percent oil and 37 percent natural gas (see Fig. 3.9). Prior to conversion from an
income trust to petroleum exploration and development company in January 2011, Penn
West Exploration operated under the trade name Penn West Petroleum. Penn West’s
capital expenditure focus for 2011 was concentrated almost exclusively on enhanced oil
recovery in the Western Sedimentary Basin (see Fig. 3.10). As indicated in the operating
cost of production per barrel of more than C$15, the company has a somewhat higher
cost of production than other Canadian oil and gas producers operating in the Western
Sedimentary Basin, such as Crescent Point Energy Corp. at about C$11 per barrel for
light oil obtained mostly by drilling for new deposits and Canadian Natural Resources
at about $12 per barrel for mostly Canadian heavy oil production. Penn West describes
the overall business as:
Penn West’s assets include ownership within many of the largest pools discovered
in western Canada. The majority of the resources-in-place are not extractable
using conventional primary production methods. The future opportunities
are increasingly coming through the application of various enhanced recovery
techniques to access and produce more of the resource. This is a major focus for
Penn West. The elements of the enhanced recovery strategy include:
As illustrated in Fig. 3.10 and Table 3.29, Penn West is an exploration and development
company that needs to fund on-going capital expenditures to replace depleting fields:
Penn West as at
Light and Natural
December 31, 2010
Medium Gas
Reserves Estimates Oil Heavy Oil Natural Liquids boe
Category (mmbbls) (mmbbls) Gas (bcf) (mmbbls) (mmboe)
Proved
Developed producing 203 51 735 21 398
Developed non-producing 5 2 47 1 16
Undeveloped 51 1 83 2 68
Total Proved 259 54 865 24 481
Probable 94 14 370 9 180
Total proved plus probable 353 68 1,235 33 661
A key business objective is to add reserves and production from our existing
resource base, thereby creating incremental future cash flows available for
reinvestment and distribution to Penn West’s investors. The proportion of our
capital spending directed towards enhanced oil recovery is planned to be carefully
increased as we go forward based on the opportunities that we currently recognize.
Although the magnitude of the annual capital commitment is managed in
accordance with commodity prices and the capital available for reinvestment, the
overall strategy points to growth.
The 2010 cash flow statement and 2011-Q3 income statement (see Tables 3.30 and 3.31)
give an indication of the level of capital expenditures that need to be maintained is well
over a billion dollars per year and the significant size of risk management gains and
losses relative to the overall size of the business.
Hedging Philosophy
Penn West describes the firm hedging policy as follows:
Penn West considers price hedging of oil and natural gas production to be a useful
tool of risk management. Its uses include protecting planned capital budgets,
safeguarding the economics of acquisitions and providing downside cash flow
protection to support planned distributions.
In effect, Penn West identifies three essential elements guiding commodity risk
management decisions. In addition to the protection of capital budgets and
acquisitions – motivations for hedging often identified in academic studies, e.g. Froot et
al. (1993, 1994) – Penn West also identifies the “downside” protection of distributions
by hedging to ensure that capital expenditure commitments do not impinge on cash
distributions which, in January 2012 were approximately 5.2 percent with a common
stock price just over C$20.3 The impact of risk management activities on smoothing the
Operating activities
Net income (loss) $ 226 $ (144)
Depletion, depreciation and accretion (note 5) 1,338 1,556
Future income tax recovery (note 11) (319) (378)
Unit-based compensation (note 13) 47 52
Unrealized risk management (gain) loss (note 10) (25) 593
Unrealized foreign exchange gain (82) (186)
Asset retirement expenditures (53) (65)
Change in non-cash working capital (note 17) 85 (27)
1,217 1,401
Investing activities
Additions to property, plant and equipment (1,187) (688)
Acquisitions of property, plant and equipment (637) (32)
Dispositions of property, plant and equipment 1,148 401
Change in non-cash working capital (note 17) 155 (79)
(521) (398)
Financing activities
Decrease in bank loan (1,101) (687)
Proceeds from issuance of notes (note 7) 460 238
Issue of equity 557 280
Distributions paid (591) (799)
Redemption of convertible debentures – (4)
Repayment of acquired facilities (note 18) (21) (31)
(696) (1,003)
Change in cash – –
Cash, beginning of year – –
Cash, end of year $ – $ –
Interest paid $ 147 $ 147
Income taxes recovered $ (1) $ (3)
prices for crude oil and natural gas for Penn West and on Penn West earnings is given
in Tables 3.32 and 3.33.
Business Risks
A detailed discussion of the key business risks faced by Penn West is included in the
Management’s Discussion and Analysis for the year ended December 31, 2010, under
the heading “Business Risks,” where Penn West provides a conventional identification
of business risks facing the firm:
We are exposed to normal market risks inherent in the oil and natural gas business,
including, but not limited to, commodity price risk, foreign currency risk, credit
Expenses
Operating 260 239 765 701
Transportation 7 8 22 25
General and administrative 38 36 112 104
Share-based compensation 13 (66) 23 16 77
expense (recovery)
Depletion and depreciation 6 292 263 850 875
Gain on dispositions – (368) (151) (1,082)
Exploration and evaluation 5 1 – 5 1
expense
Unrealized risk management loss 10 11 13 (2) 10
(gain)
Unrealized foreign exchange loss 136 (46) 91 (27)
(gain)
Financing 7,8 47 46 142 131
Accretion 9 11 10 33 30
737 224 1,883 845
(1) Gross revenues include realized gains and losses on commodity contracts
risk, interest rate risk, liquidity risk and environmental and climate change risk. We
seek to mitigate these risks through various business processes and management
controls and from time to time by using financial instruments.
From this conventional listing of risks, Penn West singles out commodity price risk:
Commodity price fluctuations are among our most significant exposures. Crude oil
prices are influenced by worldwide factors such as OPEC actions, world supply and
demand fundamentals, and geopolitical events. Natural gas prices are influenced
by the price of alternative fuel sources such as oil or coal and by North American
natural gas supply and demand fundamentals including the levels of industrial
activity, weather, storage levels and liquefied natural gas activity.
In accordance with policies approved by our Board of Directors, we may, from time
to time, manage these risks through the use of swaps, collars or other financial
instruments up to a maximum of 50 percent of forecast sales volumes, net of
royalties, for the balance of any current year plus one additional year forward and
up to a maximum of 25 percent, net of royalties for one additional year thereafter.
Such Board directives to limit the allowable percentage of output that can be hedged
is commonplace across non-financial firms. For example, Canfor Corp., an integrated
forest products company headquartered in Vancouver, BC, has a Board directive to
hedge no more the 15 percent of lumber sales and 5 percent of pulp sales. Similarly,
BC Ferries, the monopoly provider of ferry services in BC, has a Board determined
70 percent restriction on hedging fuel expenses.
As illustrated in Tables 3.34 and 3.35, Penn West is an active user of derivative
security contracting to manage risks associated with the price of crude oil, natural
Crude oil
WTI Collars 35,000 bbls/d Jan/11 – Dec/11 US$80.06 to $ (70)
$91.98/bbl
WTI Collars 15,000 bbls/d Jan/12 – Dec/12 US$83.67 to (17)
$96.32/bbl
Electricity swaps
Alberta Power Pool 90 MW Jan/11 – Dec/11 $63.16/MWh (10)
Alberta Power Pool 45 MW Jan/12 – Dec/12 $53.02/MWh (2)
Alberta Power Pool 30 MW Jan/12 – Dec/13 $54.60/MWh (1)
Alberta Power Pool 20 MW Jan/13 – Dec/13 $56.10/MWh –
Alberta Power Pool 50 MW Jan/14 – Dec/14 $58.50/MWh (2)
Interest rate swaps
$500 Jan/11 – Dec/11 1.61% (1)
$600 Jan/11 –Jan/14 2.71% (13)
$50 Jan/11 – Jan/14 1.94% –
Foreign exchange forwards
19-month term US$378 Jan/11 – Dec/11 1.06085 CAD/USD 23
8-year term US$80 2015 1.01027 CAD/USD 1
10-year term US$80 2017 1.00016 CAD/USD 1
12-year term US$70 2019 0.99124 CAD/USD 1
15-year term US$20 2022 0.98740 CAD/USD –
Cross currency swaps
10-year term £57 2018 2.0075 CAD/GBP, (28)
6.95%
10-year term £20 2019 1.8051 CAD/GBP, (6)
9.15%
10-year term €10 2019 1.5870 CAD/EUR, (2)
9.22%
Total $ (126)
2012 2013
2012 2013
gas, electricity, foreign exchange and borrowing costs. For the commodities, the size of
the hedge position does change over time. Observing that Penn West had production
of 164,000 barrels of oil equivalent (boe)/day from natural gas, light oil and heavy oil
with 80,000 bbls. from light oil, the size of the hedge position for oil appears to hover
close to Board determined maximum percentage. Comparison of Tables 3.34 and 3.35
reveals the size of hedge positions do fluctuate through time. For example, consider the
situation for natural gas hedges in place reflected in the 2010 annual report:
In 2010, the AECO Monthly Index averaged $4.12 per mcf compared to
$4.13 per mcf in 2009. Our corporate average natural gas price before the impact
of the realized portion of risk management was $4.20 per mcf for 2010. We
currently have no natural gas risk management contracts for 2011 or beyond.
Yet, by the end of Q3-2011, Penn West had established a swap hedge for natural gas of
50,000 mcf. Given the weak pricing environment for natural gas, hedging natural gas
output at this juncture has considerable justification.
Prices received for crude oil are referenced to or denominated directly in US dollars,
thus our realized oil prices are impacted by Canadian dollar to US dollar exchange
rates. A portion of our debt capital is denominated in US dollars, thus the principal
and interest payments in Canadian dollars are also impacted by exchange rates.
When we consider it appropriate, we may use financial instruments to fix or collar
future exchange rates to fix the Canadian dollar equivalent of crude oil revenues or
to fix US denominated long-term debt principal repayments.
At December 31, 2010, we had US dollar denominated debt with a face value of
US$1.2 billion (2009—US$0.9 billion) on which the repayment of the principal
amount in Canadian dollars was not fixed.
collar interest rates. As at December 31, 2010, none of our long-term debt instru
ments were exposed to changes in short-term interest rates (2009—14 percent).
As at December 31, 2010, Penn West had a total of $1.7 billion of fixed interest rate
debt instruments and $0.3 billion of convertible debentures outstanding. On the fixed
interest rate debt the average remaining term was 7.2 years (2009—7.7 years) with an
average interest rate of 5.7 percent (2009—4.6 percent), including the effects of interest
rate swaps.
Finally, Penn West uses sensitivity analysis to provide measurement of exposure to
change in oil prices and volumes, natural gas prices and volumes, the US$/C$ exchange
rate and the level of interest rates. This approach leads to statements such as: “Based
on December 31, 2010 pricing, a $1.00 change in the price per barrel of liquids would
change the pre-tax unrealized risk management gain by $15 million.” In comparison
to more sophisticated methods of measuring exposure, the limitations of sensitivity
analysis are apparent. No information is provided about the distribution of previous
price, rate or volume changes. Sensitivity estimates are, at best, point estimates and take
no account of convexities and other nonlinearity effects that occur when the variables
of interest experience large changes. Only the impact on “funds flow” is considered.
Other measures of firm performance are not considered.
Company Overview
Canadian Natural Resources Limited is headquartered in Calgary, Alberta, with
operations in western Canada, the North Sea off Scotland, and West Africa. The
independent company has grown rapidly since 1989 when production was approximately
1400 boe per day. At the end of 2010, the company had 4,671 employees of which 371
were employed in the international segment of the business. In 2010, combining crude
and natural gas production, the company produced 643,000 boe per day (see Tables
3.37 and 3.38) of which 425,000 bbls./day were due to crude oil production. The bulk
of this production is located in North America, with about 20 percent in light oil,
35 percent in heavy oil and bitumen and 14 percent from the Horizon oil sands mining
(1) Per common share amounts have been restated to reflect a two-for-one common share split
in May 2010.
(2) Adjusted net earnings from operations is a non-GAAP measure that the Company utilizes to
evaluate its performance. The derivation of this measure is discussed in the Management’s
Discussion and Analysis (“MD&A”).
(3) Cash flow from operations is a non-GAAP measure that the Company considers key as it
demonstrates the Company’s ability to fund capital reinvestment and repay debt. The derivation
of this measure is discussed in the MD&A.
(4) Includes the current portion of long term debt.
(1) Net of transportation and blending costs and excluding risk management activities.
PRODUCTION EXPENSE – EXPLORATION AND PRODUCTION
2010 2009 2008
(1) Amounts expressed on a per units basis and based on sales volumes.
and upgrading project. For each type, costs of production are well below the selling
prices. With initial construction beginning in 2005 and expected production startup
in mid-2010, due to production interruptions since the project came on-stream, the
Horizon project is yet to deliver a full year at expected production capacity of 200,00
plus bbl./day. While the earlier history of CNQ revolved around conventional light oil
and natural gas production, Table 3.39 demonstrates that the future productive capacity
of CNQ depends fundamentally on bitumen, oil sands mining and in-situ heavy oil
deposits.
The extent of CNQ commercial activities and the relative impact of hedging on the
financial results is reflected in the earnings and cash flow statements (see Tables 3.40
and 3.41). While the earnings statement records the net result of realized and unrealized
gains and losses from risk management activities, the cash flow statement for CNQ
records both. Table 3.42 further attributes the realized and unrealized gains and losses
to risk management for three categories of hedges: sales of crude oil and natural gas
liquids; purchases of natural gas; and FX and interest rates. The size of the realized
and unrealized gains and losses for 2008 and 2009 is remarkably large compared to
cash flow from operations or earnings. For example, the unrealized gain in 2008 was
almost half the size of cash from operations. The net impact on earnings was about
20 percent from realized plus unrealized gains and losses in 2008. While 2008
Operating activities
Net earnings $ 1,697 $ 1,580 $ 4,985
Non-cash Items
Depletion, depreciation and amortization 4,036 2,819 2,683
entirement obligation accretion 107 90 71
Stock-based compensation expense (recovery) 294 355 (52)
Unrealized risk management (gain ) loss (25) 1,991 (3,090)
Unrealized foreign exchange (gain) loss (180) (661) 832
Deferred petroleum revenue tax expense (recovery) 23 15 (67)
Future Income tax expense (recovery) 364 (99) I,607
Other (7) 5 2S
Abandoment expenditures (179) (48) (38)
Net change in non-cash working capital (note 14) 149 (235) (189)
Cash from operations 6,284 5,812 6,757
Financing activities
Repayment of bank credit facilities, net (472) (2,021) (623)
Replacement of bank credit facilities, net 400 – –
Replacement of senior unsecured notes – (34) (31)
Issue of US dollar debt securities – – (1,215)
Issue of common shares on exercise of stock options 170 24 13
(Continued)
Table 3.42 CNQ Risk Management Activities, Unrealized and Realized Gains and Losses
RISK MANAGEMENT ACTIVITIES
The Company utilizes various derivative financial instruments to manage its commodity price,
foreign currency and interest rate exposures. These derivative financial instruments are not
intended for trading or speculative purposes.
($ millions) 2010 2009 2008
Complete details related to outstanding derivative financial instruments at December 31, 2010
are disclosed in note 12 to the Companies consolidated financial statements.
The cash settlement amount of commodity derivative financial instruments may vary materially
depending upon the underlying crude oil and natural gas prices at the time of final settlement, as
compared to their mark-to-market value at December 31, 2010.
Due to changes in crude oil and natural gas forward pricing and the reversal of prior period
unrealized gains and losses, the Company recorded a net unrealized gain of $25 million ($16 mil-
lion after-tax) on its risk management activities for the year ended December 31, 2010 (2009 –
$1,991 million unrealized loss, $1,437 million after-tax; 2008–$3,090 million unrealized gain,
$2,112 million after-tax).
exhibited a dramatic improvement to earnings and cash flow, the bulk of these gains
were returned in 2009.
Crude oil
Crude oil price Jul 2011 – Dec 2011 50,000 bbl/d US$70.00 – US $ WTI
collars 102.23
Crude oil puts Jul 2011 – Dec 2011 100,000 bbl/d US$70.00 WTI
The cost of outstanding put options and their respective periods of settlement are as follows:
Q3 2011 Q4 2011
Cost ($ millions) US$27 US$27
Crude oil
Crude oil Oct 2008 – Dec 2008 20,000 bbl/d US$50.00 – US$65.53 Mayan
price collars Heavy
Oct 2008 – Dec 2008 50,000 bbl/d US$60.00 – US$75.22 WTI
Oct 2008 – Dec 2008 50,000 bbl/d US$60.00 – US$76.05 WTI
Oct 2008 – Dec 2008 50,000 bbl/d US$60.00 – US$76.98 WTI
Oct 2008 – Dec 2008 25,000 bbl/d US$70.00 – US$112.63 WTI
Jan 2009 – Dec 2009 25,000 bbl/d US$70.00 – US$111.56 WTI
Crude oil puts Oct 2008 – Dec 2008 50,000 bbl/d US$ – 55.00 WTI
Jan 2009 – Dec 2009 92,000 bbl/d US$ – 100.00 WTI
Figure 3.11 Crescent Point Energy Corp. Crude Oil Hedge Positions, 2010-Q3
that CNQ has substantively reduced both the size and term of the collar position.
Whereas collars in 2008 had maturities beyond one year in the future, the 2011-Q2
term is only six months. The size of the out of the money put position has also
been reduced, though the 100,000 bbl./day size of the 2011-Q2 position is comparable
to the 92,000 bbl,/day put with the $100 exercise price in 2008 that applied to pricing
for 2009. The premium cost of maintaining the 2011-Q2 put position is $27 million
per quarter. As illustrated in Figure 3.11, CNQ is not the only Canadian oil and gas
company pursuing a hedging strategy that combines different types of derivative
securities. Among others, Crescent Point Energy combines puts, collars and swaps to
manage crude oil price risk.
As with other non-financial firms, CNQ provides considerable reporting on risk
management associated with FX and borrowing costs (see Tables 3.45 and 3.46). In
particular, for borrowing costs:
The Company enters into interest rate swap contracts to manage its fixed to floating
interest rate mix on certain of its long-term debt. The interest rate swap contracts
require the periodic exchange of payments without the exchange of the notional
principal amounts on which the payments are based.
The impact of interest rate risk management on the overall performance of CNQ is
incidental. The notional amount of the total is relatively small compared to the more
than $10 billion outstanding of long-term debt and other long-term liabilities.
Regarding FX risk management, CNQ observes:
The Company is exposed to foreign currency exchange rate risk in Canada primarily
related to its US dollar denominated long-term debt and working capital. The
Company is also exposed to foreign currency exchange rate risk on transactions
conducted in other currencies in its subsidiaries and in the carrying value of
Table 3.45 CNQ Interest Rate and Cross Currency Swaps, Dec. 2010
Amount
Remaining term ($ millions) Fixed rate Floating rate
(1) During 2010, the Company unwound US$350 million of 4.9% interest rate swaps for
proceeds of US$54 million.
(2) During 2010, the Company unwound C$300 million of 1.0680% interest rate swaps for
nominal consideration.
(3) Canadian Dealer Offered Rate.
Cross currency
Swaps(1) Jan 2011–Jul 2011 US$150 .0.999 6.70% 7.70%
Jan 2011–Aug 2016 US$250 1.116 6.00% 5.40%
Jan 2011–May 2017 US$1,100 1.170 5.70% 5.10%
Jan 2011–Mar 2038 US$550 1.170 6.25% 5.76%
(1) Subsequent to December 31, 2010, the Company entered into cross currency swap contracts
for US$50 million with an exchange rate of $0.994 (US$/C$) and average interest rates of 6.70%
(US$) and 7.88% (C$) for the period January to July 2011.
Cross currency swap contracts are periodically used to manage currency exposure
on US dollar denominated long-term debt. The cross currency swap contracts
require the periodic exchange of payments with the exchange at maturity of notional
principal amounts on which the payments are based.
In addition to cross currency swaps, CNQ also employs forward FX contracts for cash
management purposes:
Price changes
Crude oil – WTI US$1.00/bbl(1)
Excluding financial $ 128 $ 0.12 $ 99 $ 0.09
derivatives
Including financial $ 128 $ 0.12 $ 99 $ 0.09
derivatives
Natural gas – AECO C$0.10/
Mcf(1)
Excluding finding derivaties $ 34 $ 0.03 $ 25 $ 0.02
Including financial $ 38 $ 0.04 $ 29 $ 0.03
derivatives
Volume changes
Crude oil – 10,000 bbl/d $ 175 $ 0.16 $ 104 $ 0.10
Natural gas – 10 MMcf/d $ 9 $ 0.01 $ 1 $ –
Foreign currency rate
change
$0.01 change in US$(1)
Including financial derivatives $ 101 – 103 $ 0.09 $ 40 – 41 $ 0.04
Interest rate change – 1% $ 9 $ 0.01 $ 9 $ 0.01
In addition to the cross currency swap contracts in Table 3.45, at year end 2010, CNQ
had US$1,162 million of foreign currency forward contracts outstanding, with terms of
approximately 30 days or less.
The final element of the risk management reporting provided by CNQ is provided by
the use of sensitivity analysis to report exposures. As reported by CNQ, the sensitivity
analysis in Table 3.46:
summarizes the annualized sensitivities of the Company’s net earnings and other
comprehensive income to changes in the fair value of financial instruments out-
standing as at December 31, 2010, resulting from changes in the specified variable,
with all other variables held constant. These sensitivities are prepared on a different
basis than those sensitivities disclosed in the Company’s other continuous disclo-
sure documents, and do not represent the impact of a change in the variable on
the operating results of the Company taken as a whole. Further, these sensitivities
are theoretical, as changes in one variable may contribute to changes in another
The airline industry provides a wealth of information available from academic, indus-
try and government studies. Given the economic importance of the airline industry
both domestically and globally, it is not surprising that most academic studies examine
issues other than risk management. For example, there are numerous studies on market
329
organization and pricing behavior, e.g. Borenstein (2011), Chi and Koo (2009), Cliberto
and Williams (2010), Doganis (2002), Gerardi and Shapiro (2009), Hofer and Ergolu
(2010). Some studies focus on pricing dynamics, e.g. Chi and Koo (2009), Cliberto and
Williams (2010), while others consider the impact of competition, e.g. Gerardi and
Shapiro (2009). There are also other studies that aim to explain the poor performance
of the US airline industry since the Airline Deregulation Act (1978). Borenstein (2011,
p.233) describes the situation:
The industry lost $10 billion from 1979 to 1989, made $5 billion in the 1990s and
lost $54 billion from 2000 to 2009 (all figures in 2009 dollars). To put these figures
in context, at the end of 2000, after six consecutive profitable years, the entire book
value of US passenger carriers’ assets was $159 billion and shareholder equity was
$40 billion.
Even if the financially disastrous first decade of the 21st century is ignored: “This dismal
financial record isn’t what economists, analysts, or industry participants predicted in
1978.”
In addition to studies on the economics of the airline industry as a whole, there
are also studies focusing on characteristics that drive business performance for
firms within the airline industry, e.g. Busse (2002), Flouris and Walker (2005), Gritta
et al. (2006), Malighetti et al. (2011). Government studies and industry studies
provide impressive detail on various aspects of individual airline performance.
For example, the Bureau of Transportation Statistics in the USA provides detailed
statistics on airline traffic and financial performance obtained from required regulatory
filings (see Tables 3.47 and 3.48). Other useful information is available from: the
Air Transport Association (ATA), the primary association of US airline companies;
the International Air Transport Association (IATA), representing some 240 inter-
national airlines with 84 percent of total air traffic; and the Boeing Corporation which
produces a number of useful publications including the Current Market Outlook,
2011–2030.
In contrast to the considerable information about airline economics and traffic
statistics, there are comparatively few studies directly on commodity risk management
for airlines. In particular, the increase in the level and volatility of jet fuel prices requires
airline companies to be fundamentally concerned with managing the strategic risk
associated with jet fuel prices. As such, the strategic need for airlines to manage jet fuel
price risk is relatively recent development (see Table 3.49). The upswing in crude oil
prices between 2001 and 2008 resulted in jet fuel costs becoming the largest and most
volatile component of expenses. As illustrated Figs 3.12 and 3.13, this runup in jet fuel
price level and volatility continues, especially when prices are denominated in US$.
Prior to the emergence of this jet fuel risk, labor costs and capital expenditure programs
were more central to airline strategy. Financial risk management associated with FX
and interest costs were typically more significant than managing jet fuel risk. Against
Table 3.47 Revenue Passenger Miles, US Airlines with more than 1 billion miles,
Jan.–Sept. 2011
Description Revenue Passenger Miles
this backdrop, risk management for airlines also has had to address fundamental issues
associated with catastrophic loss and related security issues.
Some studies that do examine risk management for airlines emphasize the disaster
element of airline risk management, the need to maintain safety and prevent crashes,
e.g. Nomura (2003). The most exhaustive and useful studies directly concerned with
commodity risk management in the airline industry are Carter (2006a, b), though
Morrell and Swan (2006), Rao (1999) and Weiss and Maher (2009) are also helpful.1
Bertus et al. (2009) provide a “scientific” approach to hedge ratio estimation for airlines
using the most advanced available econometric techniques to conclude “airlines
should hedge their positions in jet fuel using a horizon-sensitive model that directly
Revenue Passenger Enplanements (000) 730,660 6,336 736,996 712,206 5,935 718,141
Revenue Passenger Miles (000) 814,625,747 11,154,853 825,780,600 786,283,364 11,081,573 797,364,937
Available Seat Miles (000) 993,859,218 19,139,392 1,012,998,610 960,788,737 19,352,643 980,141,380
Passenger Load Factor (%) 81.97 58.28 81.52 81.84 57.26 81.35
Revenue Freight Ton Miles (000) 6,950,989 112,344 7,063,333 7,024,072 65,191 7,089,263
Total Revenue Ton Miles (000) 89,112,798 1,227,862 90,340,660 86,350,734 1,173,355 87,524,089
Available Ton Miles (000) 143,881,386 3,124,663 147,006,049 139,112,693 3,112,022 142,224,715
Ton Miles Load Factor (%) 61.93 39.3 61.45 62.07 37.7 61.54
Revenue Departures Performed 9,505,525 182,311 9,687,836 9,496,502 168,858 9,665,360
Revenue Aircraft Miles Flown (000) 7,047,527 98,905 7,146,432 6,876,795 97,878 6,974,673
Revenue Aircraft Hours (Airborne) 16,687,946 264,338 16,952,284 16,352,623 261,304 16,613,927
11/16/2012 6:45:38 PM
Airlines and Jet Fuel Hedging 333
accounts for movements in the jet fuel, crude oil spread.” Hallerbach and Menkveld
(2004) employ a “Component Value at Risk” methodology to examine downside risk
assessment and management in the airline industry. Such sources provide a variety of
conflicting viewpoints on commodity risk management for airlines.
Carter et al. (2006a, 2006b) take a strong empirically based stance in favor of
hedging jet fuel price risk. Given the dramatically increasing importance of jet fuel
costs in airline expenditures (see Table 3.49) combined with the thin profit margins
(see Figs 3.14 and 3.15), the importance of commodity risk management activities
such as hedging is not surprising. The key questions revolve around how to implement
and how much risk management is appropriate. Both costless collars and out-
of-the-money calls involve insurance solutions, whereas swaps, forwards and futures
contracts expose the firm to losses from adverse movements on the hedge. There is
little information on which approach is appropriate in a particular situation. The
airlines examined below—SIA and Southwest Airlines—use a combination of strate-
gies. In addition, while hedge ratio estimates from academic studies are typically
close to a full hedge, firms in practice tend to hedge substantially less. Some
airlines, e.g. SIA, do not always report the exact amount of total fuel purchases that
are hedged.
While taking a generally bullish view of airline hedging, the empirical results
provided in Carter et al. (2006a, 2006b) do not distinguish between various types of
risk management activity. An out-of-the-money call strategy is lumped in with the use
of jet fuel swaps. This general overview approach to academic advice on hedging is not
restricted to Carter et al. For example, Morrell and Swan (2006, p.729) take an agnostic
position on fuel price hedging:
A fuel price hedge would create exceptional value when an airline is on the
edge of bankruptcy, but when this occurs an airline does not have the liquidity
to buy oil derivatives. On the other hand, hedges probably did make sense
when airlines were state-supported, and variable levels of hedging can be
useful in transferring profits from one quarter to another. Finally, hedging
may be a zero cost signal to investors that management is technically alert and
perhaps this is the most compelling argument for airline hedging. However, it
lies more in the realm of the psychology of markets than in the mathematics of
economics.
after controlling for trend, seasonality, and persistence of shocks, hedging has the
potential to reduce the unexplained volatility of the average airline’s quarterly
income by over 20 percent. Thus, the results suggest that the usefulness of hedging
is not restricted to protecting weak airlines incapable of withstanding an increase
in fuel prices. Also, airlines should not eschew hedging merely because of the
possibility of incurring opportunity costs if fuel prices go down rather than up;
hedging appears to pay off in the long run by providing a more stable earnings
stream.
Carter et al. (2006a, pp.32–3) provide the following description concerning the
value of airline hedging practices:
In contrast to views in favor of fuel price hedging, Weiss and Maher (2009, p.362) claim:
These operational elements of natural hedging can be used to explain the less than full
hedging of fuel price risk that is widespread practice among airline companies.
Academic studies are largely concerned with whether hedging jet fuel is a desirable
strategy for airline companies. This is consistent with the general types of question asked
in academic studies of commodity risk management across industries. While somewhat
helpful, the types of question being asked miss the main questions confronting “real
time” commodity risk managers, e.g. what are the appropriate methods to use in
managing a particular commodity price risk? Airlines are not the only companies in
the transportation business that are impacted by jet fuel prices. For example, air cargo
companies such as FedEx face similar challenges. Yet, in the FedEx case, jet fuel prices
are managed using fuel surcharges built into FedEx billing procedures. Though the fuel
surcharge approach has been used at times in the airline industry, this approach poses
80.5% SIA Engineering Company Limited 100% Aircraft Maintenance Services Australia Pty Ltd
100% SilkAir (Singapore) Private Limited 100% Tradewinds 100% NexGen Network (1) Holding Pte Ltd
Tours & Travel
Private Limited
56% Abacus Travel Systems Pte Ltd 5% Abacus Travel 100% NexGen Network (2) Holding Pte Ltd
Systems Pte
Ltd
100% Singapore Aviation and General 100% SIAEC Global Pte Ltd
Insurance Company (Pte) Limited
100% SIA Properties (Pte) Ltd 20% PT Purosani Sri 100% SIA Engineering (USA), Inc.
Persada
100% Singapore Flying College Pte Ltd 65% Singapore Jamco Private Limited
100% Sing-Bi Funds Private Limited 65% SIA Engineering (Philippines) Corporation
100% Singapore Airline Cargo Pte Ltd 51% Cargo 100% Cargo 51% Aerospace Component
Community Community Engineering Services Pte Ltd
Network Pte (Shanghai) 51% Aviation Partnership (Philippines) Corporation
100% SIA (Mauritius) Ltd Ltd Co. Ltd 50% International Engine Component Overhaul Pte Ltd
76% Singapore Airport Duty- Free 25% Great Wall 50% Singapore Aero Engine Services Pte Ltd
Emporium (Private) Limited Airlines
Company
Limited
50% Service Quality (SQ) Centre Pte 49% Combustor Airmotive Services Pte Ltd
Ltd
49% Virgin Atlantic Limited 49% Eagle Services Asia Private Limited
33.7% Tiger Airways Holdings Limited 49% Fuel Accessory Service Technologies Pte Ltd
20% Ritz-Carlton, Millenia Singapore 49% Pratt & Whitney Airmotive International Ltd
Properties Private Limited
49% Safran Electron Asia Pte Ltd
49% PT JAS Aero - Engineering Services
47.1% Pan Asia Pacific Aviation Services Limited
45% Jamco Aero Design & Engineering Private Limited
40% Goodrich Aerostructures Services Center-Asia Pte Ltd
40% Messier Services Asia Pte Ltd
39.2% Asian Surface Technologies Pte Ltd
33.3% International Aerospace Tubes-Asia Pte Ltd
24.5% Asian Compressor Technology Services Company Ltd
24.5% Turbine Coating Services Pte Ltd
11/16/2012 6:45:40 PM
Airlines and Jet Fuel Hedging 339
significant complications when people, not air cargo, are being transported. Another
example of the diversity of possible methods for handling fuel price risk appears
with railroads where, despite fuel costs being as much as 20 percent of total expenses,
fuel hedging is uncommon. This could be due to the more severe impact of fuel
costs on important competitors, such as trucking companies. In turn, major trucking
companies, such as Marten Transport, employ partial fuel surcharges to deal with fuel
price risk.
Company Overview
Singapore Airlines (SIA) is, arguably, the most successful airline company in
the world. This success is reflected in various performance metrics: profitability;
customer satisfaction; quality of fleet and airport infrastructure. SIA focuses on the
higher-end segment of the passenger market with extensive first class, and business
class, capacity. As of March 30, 2011, SIA had a fleet of 108 aircraft with an average
age of 6 years, 3 months making SIA “one of the world’s youngest and most fuel
efficient” airlines in the world (see Table 3.52). SIA evolved from Malay Airlines,
a company founded May 1, 1947. In May 1966, the airline became Malaysia-Singapore
Airlines. In 1972, Malaysia-Singapore Airlines divided into two separate companies:
Malaysia Airlines and Singapore Airlines. Currently, the biggest shareholder in SIA
is the Singapore government, which owns 54.95 percent of Singapore Airlines stock
through its investment company Temasek Holdings Ltd (see Table 3.51). SIA Group
has over 20 different subsidiaries including SIA cargo, SIA engineering company, and
Silkair (see Fig. 3.17). The SIA fleet of 108 aircraft is expanded to 137 if SIA Cargo and
Silkair are included. The fleet consists of both Boeing and Airbus aircraft, with 68
owned and 40 on operating lease.
SIA is a Singapore-listed company filing subject to rules in that jurisdiction. As
with previous discussion of Australian, US and Canadian firms, difference in filing
jurisdiction has some impact on the type and quality of the results presented. Being
primarily a long-haul carrier, the dramatic importance of jet fuel expense to SIA is
captured in Table 3.53 and Fig. 3.18. Over one-third of total expenditures depend
on this one item. Yet, in contrast to Canadian reports where risk management
activities appear as specific items in the cash flow statement, for SIA it is more difficult
to identify information specifically concerned with risk management due to accounting
treatment of this item. As indicated in Table 3.54, no direct information about risk
management is provided in the cash flow statement. Rather, relevant information that is
available is provided in the “Financial Instruments” section of the notes to the financial
statements. More detail is provided for interest rate and liquidity risk than for jet fuel
price risk.
The Group operates globally and generates revenue in various currencies. The
Group’s airline operations carry certain financial and commodity risks, including
the effects of changes in jet fuel prices, foreign currency exchange rates, interest
rates and the market value of its investments. The Group’s overall risk management
approach is to moderate the effects of such volatility on its financial performance.
The Group’s policy is to use derivatives to hedge specific exposures.
Singapore Airlines:
B747-400 6 1 7 375 11 y 6 m 1
B777-200 7 1 8 288 13 y 1 m 1
B777-200A 6 1 7 323 9y7m
B777-200R 9 2 11 266 7y6m
B777-200ER 6 3 9 285 8y5m
B777-300 4 7 11 332 9y5m 4 1
B777-300R 1 1 284 9y6m
B777-300ER 19 19 278 3y7m
A340-500 5 5 100 7y1m
A380-800 5 6 11 471 2y4m 8 6
A330-300 19 19 285 1y4m
B787-9 20 20
A350-900 XWB 20 20
SIA Cargo:
B747-400F 4 4 3 11 N.A. 9y2m 2
SilkAir:
A319-100 4 2 6 128 5y2m
A320-200 7 5 12 150 6y4m 6 4
Sub-total 11 7 18 N.A. 5 y 11 m 6 4
11/16/2012 6:45:41 PM
342 Risk Management Applications
The Group
The Group
As derivatives are used for the purpose of risk management, they do not expose
the Group to market risk because gains and losses on the derivatives offset losses
and gains on the matching asset, liability, revenues or costs being hedged. Moreover,
counterparty credit risk is generally restricted to any hedging gain from time to
time, and not the principal amount hedged. Therefore the possibility of a material
loss arising in the event of non-performance by a counterparty is considered to be
unlikely.
The first “financial” risk listed is jet fuel prices. This treatment of commodity
risk management activities in the notes to the financial statements can be contrasted
with the prominent treatment of “Risk Management” in the Annual Report. In
the main body of the report, the “Statement on Risk Management” provides the
following:
This “enterprise risk management” structured approach to risk reflects the classification
of risks in the airline industry given in Sec. 2.2. More precisely, the “Statement of Risk
Management” continues with the types of major event “risk” encountered:
Expenditure of fuel before hedging was $679 million higher because of:
$ million
26.3% increase in weighted average fuel price from 79.5 USD/BBL to 100.4 + 840.8
USD/BBL
3.6% increase in volume uplifted from 27.2 M BBL to 28.2 M BBL + 100.6
6.5% weakening of USD against SGD from US$1=S$1.425 to US$1= – 262.8
S$1.332
+ 678.6
The listing of risks continues without mention of jet fuel price risk.
This discussion is not meant to imply that SIA ignores jet fuel price risk. To the
contrary, the objective in the discussion is only intended to identify the method of
presenting the risk—in the notes to “financial instruments” section of the financial
statements—and the commodity risk management methods used to manage the
risk. Conventional sensitivity analysis is also provided for interest rates and FX rates.
Sensitivity analysis for jet fuel in the main body of the annual report is given as follows:
Table 3.56 is provided in support. Details about the approach to jet fuel hedging does
not appear until the notes to the financial statements provide the following:
The Group’s earnings are affected by changes in the price of jet fuel. The Group’s strat-
egy for managing the risk on fuel price, as defined by Board Executive Committee,
aims to provide the Group with protection against sudden and significant increases
in jet fuel prices. In meeting these objectives, the fuel risk management programme
allows for the judicious use of approved instruments such as swaps and options
with approved counterparties and within approved credit limits.
As illustrated in Table 3.57, the analysis compares jet fuel sensitivity with sensitivities
for interest rates and exchange rates. In addition, the notes to the financial statements
also provide the sensitivity analysis provided in Table 3.58.
In contrast to the Canadian oil production companies where detailed information on
specific hedges is provided, SIA is less forthcoming:
The Group manages this fuel price risk by using swap and option contracts and
hedging up to 15 months forward using jet fuel swap and option contracts. The
Group no longer enters into new gasoil hedges. Existing gasoil swap contracts will
be rolled up into jet fuel equivalents by hedging in the gasoil-jet fuel regrade closer
Asset*
Currency hedging contracts 2.5 20.7 1.6 15.8
Fuel hedging contracts 54.2 5.7 43.9 5.2
Cross currency swap – 7.7 – –
contracts
Interest rate cap contracts 17.6 23.9 17.6 23.9
74.3 58.0 63.1 44.9
Liabilities†
Currency hedging contracts 57.3 17.3 48.5 14.5
Fuel hedging contracts – 112.9 – 93.6
Cross currency swap 63.3 43.5 – –
contracts
Interest rate swap contracts 15.3 29.9 – 13.0
135.9 203.6 48.5 121.1
* Included under trade debtors
† Included under trade creditors
The Group
Notes payable 949.4 22.6 22.6 22.6 319.3 568.7 1,905.2
Finance lease 71.1 73.4 74.5 57.5 51.5 44.4 372.4
commitments
Trade and other creditors 2,725.7 – – – – – 2,725.7
Derivative financial – – – – – –
instruments:
Currency hedging 57.3 – – – – – 57.3
contracts
Cross currency swap 63.3 – – – – – 63.3
contracts
Interest rate swap 15.3 – – – – – 15.3
contracts
3,882.1 96.0 97.1 80.1 370.8 613.1 5,139.2
The Company
Note payable 949.4 22.6 22.6 22.6 319.3 568.7 1,905.2
Trade and other creditors 2,161.8 – – – – 2,161.8
Amounts owing to 1,529.0 – – – – 1,529.0
subsidiary companies
Derivative financial – – – – – –
instruments:
Currency hedging 48.5 – – – – – 48.5
contracts
4,688.7 22.6 22.6 22.6 319.3 568.7 5,644.5
to maturity. As at 31 March 2011, all gasoil contracts have matured. The Group has
applied cash flow hedge accounting to these derivatives as they are considered to
be highly effective hedging instruments. A net fair value loss before tax of $321.6
million (2010: $458.9 million), with a related deferred tax credit of $93.8 million
(2010: $116.9 million), is included in the fair value reserve in respect of these
contracts.
The only further information provided about the jet fuel swap and option contract
positions in provided in Note 37 on “Financial Instruments:”
The fair values of jet fuel swap contracts are the mark-to-market values of these
contracts. The fair values of jet fuel option contracts are determined by reference
to available market information and the Black-Scholes option valuation model.
As the Group hedges its jet fuel requirements in Mean of Platts Singapore Jet
Kerosene (“MOPS”) and that the majority of the Group’s fuel uplifts are in MOPS,
the MOPS price (2011: USD133.22/BBL, 2010: USD89.59/BBL) is used as the input
for market fuel price to the Black-Scholes option valuation model. Consequently,
the annualised volatility (2010–11: 26.85%, 2009–10: 23.46%) of the jet fuel swap
and option contracts is also estimated with daily MOPS price. The continuously
compounded risk-free rate estimated as average of the past 12 months Singapore
Government Securities benchmark issues one-year yield (2010–11: 0.38%, 2009–10:
0.41%) was also applied to each individual jet fuel option contract to derive their
estimated air values as at the end of the reporting period.
The fair values of gasoil and regrade swap contracts are also determined by
reference to available market information and are the mark-to-market values of
these swap contracts. As the Group hedges in InterContinental Exchange (“ICE”)
gasoil and MOPS jet-fuel-ICE gasoil regrade, the ICE gasoil futures contract price
and the MOPS price are used as the mark-to-market prices.
Without more precise information about the structure of the option and swap contracts,
the point estimate reporting of sensitivity analysis to assess risk is inadequate due to the
lack of information about the convexity associated with option positions.
Inspection of Tables 3.57 and 3.58 reveals that jet fuel risk is not the only significant
risk faced by SIA. Being a Singaporean company reporting in S$, SIA has significant FX
exposure. This exposure is described in note 38 as follows:
The Group is exposed to the effects of foreign exchange rate fluctuations because of
its foreign currency denominated operating revenues and expenses. For the finan-
cial year ended 31 March 2011, these accounted for 63.5% of total revenue (2009–10:
62.4%) and 64.0% of total operating expenses (2009–10: 58.6%). The Group’s largest
exposures are from USD, Euro, UK Sterling Pound, Swiss Franc, Australian Dollar,
New Zealand Dollar, Japanese Yen, Indian Rupee, Hong Kong Dollar, Chinese Yuan,
Korean Won and Malaysian Ringgit. The Group generates a surplus in all of these
currencies, with the exception of USD. The deficit in USD is attributable to capital
expenditure, fuel costs and aircraft leasing costs—all conventionally denominated
and payable in USD. The Group manages its foreign exchange exposure by a policy
of matching, as far as possible, receipts and payments in each individual currency.
Surpluses of convertible currencies are sold, as soon as practicable, for USD and
SGD.
The connection between jet fuel prices and the US$ illustrated Figure 3.12 goes
unrecognized. The methods employed for handling of FX risk are described as follows:
The Group also uses forward foreign currency contracts and foreign currency
option contracts to hedge a portion of its future foreign exchange exposure. Such
contracts provide for the Group to sell currencies at predetermined forward rates,
buying either USD or SGD depending on forecast requirements, with settlement
dates that range from one month up to one year. The Group uses these currency
hedging contracts purely as a hedging tool. It does not take positions in currencies
with a view to making speculative gains from currency movements.
The currency risk associated with US$ exposure associated with future lease payments
has been handled separately:
In addition, the Group has cross currency swap contracts in place with notional
amounts ranging from $30.1 million to $109.6 million (2010: $35.9 million to $128.6
million) where it pays SGD and receives USD at USD/SGD exchange rates ranging
from 1.3085 to 1.6990 (2010: 1.3085 to 1.6990). These contracts are used to protect
the foreign exchange risk exposure of the Group’s USD-denominated finance lease
commitments. The maturity period of these contracts ranges from 21 August 2015
to 14 February 2018.
SIA reports sensitivity analysis results for both FX and interest rate exposures. In
addition to interest rate swaps, the interest rate exposure also involves interest rate caps
“maturing in 6 to 7 years, to hedge against risk of increase in aircraft lease rentals.”
The payoffs on such instruments are not well captured by the point exposure estimates
provided by sensitivity analysis.
The final items in the financial statements relevant to commodity risk management
concerns “liquidity risk” and reporting of the fair value of derivative positions (see Table
3.58). While possible cash requirements for derivative security positions is substantive,
these risks are small relative to the overall operations of the company. SIA is more
concerned with the liquidity implications of having to meet the cost of firm aircraft
deliveries, especially over the next fiscal year. “Due to the necessity to plan aircraft
orders well in advance of delivery, it is not economical for the Group to have committed
funding in place at present for all outstanding orders, many of which relate to aircraft
which will not be delivered for several years.” In conjunction with liquidity risk, SIA
also provides a statement about “Derivative financial instruments” in the section on
classification of financial instruments (see Table 3.58).
Company overview
Southwest Airlines Co. (LUV) is legendary in the aviation industry as the pioneer of
low-cost point-to-point air travel. The LUV website provides the following company
description for 2011:
(1) A
s discussed further in Note 7 to the Consolidated Financial Statements, 128 of the Com-
pany’s aircraft have been pledged as collateral.
Firm Orders, Options and Purchase Rights for Boeing 737-700 and 737-800 Aircraft
The Boeing Company
Purchase Previously-
-700 Firm -800 Firm Options Rights Owned 700 Total
2011 17 – – – 2 19
2012 – 20 – – – 20
2013 19 – 6 – – 25
2014 21 – 6 – – 27
2015 14 – 1 – – 15
2016 17 – 7 – – 24
2017 – 17 – – 17
Through 2021 – – 98 – 98
Total 88* 20 37 98 2 245
* The Company is evaluating substituting 737-800s in lieu of 737-700 firm orders currently
scheduled for 2013 through 2016.
LUV ended 2010 with 548 Boeing 737 aircraft serving 69 cities in 35 states throughout the
United States and is the largest airline in the USA measured using number originating
passengers boarded (see Table 3.60). Similarly, the 2010 Annual Report provides the
following description of the business model:
The use of a single type of aircraft combined with operations out of non-hub airports
are central elements in the success of the LUV point-to-point low price business model,
e.g. Adler and Smilowitz (2007).
The overall scope of LUV operations are captured by the income and cash flow
statements (see Tables 3.61 and 3.62). Though the importance of fuel costs in total
expenditures will be discussed in more detail below, examination of the income
statement permits the level and variation of fuel expense to be compared to other
important items such as salaries, wages and benefits or maintenance and repairs over
a three-year period. Table 3.50 details the evolution of fuel expenditures over time.
The cash flow statement reveals the dramatic implications that risk management
activities can have for commercial operations. The $2.24 billion cash payment to fuel
derivative counterparties could pose significant liquidity problems throughout the firm
OPERATING REVENUES:
Passenger $ 11,489 $ 9,892 $ 10,549
Freight 125 118 145
Other 490 340 329
Total operating revenues 12,104 10,350 11,023
OPERATING EXPENSES:
Salaries, wages, and benefits 3,704 3,468 3,340
Fuel and oil 3,620 3,044 3,713
Maintenance materials and repairs 751 719 721
Aircraft rentals 180 186 154
Landing fees and other rentals 807 718 662
Depreciation and amortization 628 616 599
Other operating expenses 1,426 1,337 1,385
Total operating expenses 11,116 10,088 10,574
OPERATING INCOME 988 262 449
OTHER EXPENSES (INCOME):
Interest expense 167 186 130
Capitalized interest (18) (21) (25)
Interest income (12) (13) (26)
Other (gains) losses, net 106 (54) 92
Total other expenses (income) 243 98 171
INCOME BEFORE INCOME TAXES 745 164 278
PROVISION FOR INCOME TAXES 286 65 100
NET INCOME $ 459 $ 99 $ 178
NET INCOME PER SHARE, BASIC $ .62 $ .13 $ .24
NET INCOME PER SHARE, DILUTED $ .61 $ .13 $ .24
Cash dividends declared per $ .0180 $ .0180 $ .0180
common share
if such a cash flow drain were not properly anticipated. Fortunately, LUV was able to
tap previous cash reserves plus a $1 billion long-term debt issue and a $400 million
revolving credit draw to deal with the liquidity implications. Though LUV has been able
to maintain an impressive record of only reporting positive annual earnings results,
the cash flow picture is much different. All cash flows associated with purchasing and
selling derivatives are classified as operating cash flows in the Consolidated Statement
of Cash Flows.
In contrast to SIA, treatment of fuel costs receives considerable and detailed attention
in the 2010 10K. Even before recognition under Item 1A, in Item 1 where the business is
described, after listing Table 3.59 on fuel costs, LUV provides the following significant
statement:
The Company has historically entered into fuel derivative contracts to manage
rising fuel costs; however, because energy prices can fluctuate significantly in a
relatively short amount of time, the Company must also continually monitor and
adjust its fuel hedge portfolio and strategies to address fuel price volatility.
For example, during 2008, market “spot” prices for crude oil peaked at a high
of over $147 per barrel and hit a low price of under $35 per barrel—both within
a period of approximately five months. This led to the Company’s decision in
late 2008 and early 2009 to significantly reduce its net fuel hedge position in place
for 2009 through 2013. As a result of these activities, the Company effectively
locked in some hedging-related losses for 2009 through 2013. Since early
2009, the Company has continued to adjust its fuel hedge portfolio in an
attempt to economically layer back in some protection in the event of a significant
surge in market prices. Fuel costs continued to be volatile during 2010, with
market spot prices ranging from a low of $68 per barrel to a high of $91 per barrel.
Therefore, the Company continues to actively manage its fuel hedge portfolio to
address volatile fuel prices and, in particular, to mitigate the impact of significant
increases in energy prices, while maintaining an objective to manage derivative
premium costs. The Company’s fuel hedging activities are discussed [further in the
10-K] under “Risk Factors,” “Management’s Discussion and Analysis of Financial
Condition and Results of Operations,” and Note 10 to the Consolidated Financial
Statements.
The amount of information given in the additional items on “fuel hedging activities”
is too much to report here. It is more than apparent that management of jet fuel risk
is a central component of the lowest-cost-airline business strategy at LUV.
The discussion associated with jet fuel price risk given under Item 1A of the
10-K is revealing about the perception of LUV regarding factors that drive jet fuel
pricing:
Fuel price volatility continues to present one of the Company’s most significant
challenges, as (i) the cost of fuel, which has been at historically high levels over the
last few years, is largely unpredictable; and (ii) airlines are inherently dependent
upon energy to operate; therefore, even a small change in market fuel prices can
significantly affect profitability. Fuel prices are unpredictable, in part, because of
many external factors that are beyond the Company’s control. For example, fuel
prices can be impacted by political and economic factors, such as (i) dependency
on foreign imports of crude oil and the potential for hostilities or other conflicts
in oil producing areas; (ii) limited refining capacity; (iii) changes in governmental
policies on fuel production, transportation, and marketing; and (iv) changes in
exchange rates.
Likewise, the Company’s ability to react to fuel price volatility can be affected
by factors outside of its control. For example, the Company’s profitability is
affected in part by its ability to increase fares in reaction to fuel price increases;
however, fare increases are difficult to implement in difficult economic environments
when low fares are often used to stimulate traffic. The Company’s ability to increase
fares can also be limited by factors such as its low fares reputation, the portion
of its Customer base that purchases travel for leisure purposes, the competitive
nature of the airline industry generally, and the risk that higher fares will drive a
decrease in demand.
Table 3.63 Fuel Hedges, Hedge Ratio and Fuel Pricing Impact of Hedges
Percent of estimated fuel consumption covered
by fuel derivative contracts*
Average Crude Oil Price per barrel 2011-Q1 Full Year 2011
* Estimated fuel consumption for 2011 and beyond excludes any potential impact of the Com-
pany’s proposed acquisition of AirTran Holdings. Inc.
This is followed by a statement that characterizes the ex post/ex ante quandary facing
those involved in making commodity risk management decisions:
In other words, there are no guarantees that, ex post, the fuel hedging program
will be able to achieve the potential benefits that the program was expected to achieve,
ex ante.
In addition to detailed discussion under Items 1 and 1A of the 10-K filing and
reporting of specific risk management related amounts in the financial statements,
fuel hedging is also examined in considerable detail under: Item 7, Management
Discussion and Analysis; Item 7A, Quantitative and Qualitative Disclosures about
Market Risk; and in Note 10 to the Financial Statements. This impressive detail on
jet fuel hedging stands in contrast to SIA where the actual hedge ratio was difficult
to determine and comparatively more detail was provided on the compensation
and stock holdings of board members and senior management. In particular, in
Item 7 LUV reports detailed information on the hedge ratio for current and future
periods (see Table 3.64) as well as a sensitivity analysis that is reported over a range of
jet fuel prices, not the conventional “if prices change by $1 then the impact on earn-
ings is $x.” The reported response of LUV hedged fuel costs to changes in the fuel
price level reflects the importance of out-of-the-money calls to manage large increases
in fuel costs.
Table 3.64 Sensitivity Analysis for Jet Fuel Prices and Operating Expenses, LUV 2010
Estimated difference in Southwest
economic jet fuel price per gallon,
compared to unhedged market prices,
including taxes*
Avg crude oil price per barrel First quarter 2011 Full Year 2011
Increase Percent
2010 2009 (decrease) change
The sensitivity analysis is supplemented with information about hedging beyond the
next year:
Beyond 2011, the Company has coverage of approximately 60 percent of its estimated
fuel consumption in 2012; approximately 50 percent in 2013; and approximately 45
percent in 2014, all at varying price levels.
Fuel and oil expense increased $576 million, or 18.9 percent, and on a per-ASM
basis increased 18.3 percent versus 2009. Both the dollar and the per-ASM increase
were driven primarily by an 18.4 percent increase in the average price per gallon
for jet fuel, including the impact of fuel derivatives used in hedging, and including
related taxes. As a result of the Company’s fuel hedging program and inclusive of
the impact of the accounting guidance for derivatives and hedging, the Company
recognized net losses totaling $324 million in 2010 in Fuel and oil expense relating
to fuel derivative instruments versus net losses of $467 million recognized in Fuel
and oil expense in 2009. These totals are inclusive of cash settlements realized
from the expiration/settlement of fuel derivatives, which were $153 million paid to
counterparties in 2010 versus $245 million paid to counterparties for 2009. However,
these totals exclude gains and/or losses recognized from hedge ineffectiveness,
which are recorded as a component of Other (gains) losses, net.
In effect, over the reporting period the jet fuel hedging program lost money. This is
followed by the recognition of “frozen” losses from previous periods:
LUV goes on to report: “the Company’s jet fuel costs per gallon are expected to exceed
market (i.e. unhedged) prices during some of these future periods. This is based primarily
on expected future cash settlements associated with fuel derivatives, but excludes any
impact associated with the ineffectiveness of fuel hedges.” These statements regarding
losses do not reflect the actual aggregate position of the fuel hedging program given in
Item 7A.
Oddly enough, the discussion of fuel hedging in Item 7 is longer and more detailed
than the discussion provided in Item 7A where such items typically receive such
attention. The Item 7 reporting continues with discussion classified under “Financial
derivative instruments.” Of interest, LUV reports the scope of the hedging program:
“At December 31, 2010, the Company was a party to over 600 financial derivative
instruments, related to its fuel hedging program, for the years 2011 through 2014.” The
aggregate value of the hedging program is then reported:
The fair value of the Company’s fuel hedging financial derivative instruments
recorded on the Company’s Consolidated Balance Sheet as of December 31, 2010,
not considering the impact of cash collateral deposits provided to counterparties, was
a net asset of $142 million, compared to a net liability of $477 million at December
31, 2009. The change in fair value primarily was due to an increase in energy prices
throughout most of 2010, the expiration (i.e. settlement in which the Company paid
cash to counterparties) of approximately $153 million in fuel derivative instruments
that related to 2010, and the purchase of new derivative positions that will settle in
future periods. Although the Company’s fuel derivative portfolio was in a net asset
position at December 31, 2010, the positions that are expected to settle or expire
during 2011 currently consist of a net liability of approximately $62 million.
Because the mark-to-market value of the hedge portfolio will not necessarily generate
corresponding cash flows, LUV observes that: “Changes in the fair values of these
instruments can vary dramatically based on changes in the underlying commodity
prices, as has been evident in recent years.” Regarding the types of instruments employed,
LUV observes: “The financial derivative instruments utilized by the Company primarily
are a combination of collars, purchased call options, call spreads, and fixed price swap
agreements. The Company does not purchase or hold any derivative instruments for
trading purposes.” Significantly, after making references to the loss of hedge accounting
when gasoline and other non-jet fuel commodities were used to form hedges, the
complications associated with cross-hedging leads LUV to report that: “The Company
enters into financial derivative instruments with third party institutions in ‘over-the-
counter’ markets. Since the majority of the Company’s financial derivative instruments
are not traded on a market exchange, the Company estimates their fair values.”
The reporting in Item 7 concludes with the following description of the analytical
process involved in determining jet fuel hedges:
The Company continually looks for better and more accurate methodologies in
forecasting expected future cash flows relating to its jet fuel hedging program. These
estimates are an important component used in the measurement of effectiveness
for the Company’s fuel hedges. The current methodology used by the Company
in forecasting forward jet fuel prices is primarily based on the idea that different
types of commodities are statistically better predictors of forward jet fuel prices,
depending on specific geographic locations in which the Company hedges. The
Company then adjusts for certain items, such as transportation costs, that are
stated in fuel purchasing contracts with its vendors, in order to estimate the actual
price paid for jet fuel associated with each hedge. This methodology for estimating
expected future cash flows (i.e. jet fuel prices) has been consistently applied during
2010, 2009, and 2008, and has not changed for either assessing or measuring hedge
ineffectiveness during these periods.
In this, there are hints that methodologies, such CFaR, that may be “better and more
accurate . . . in forecasting expected future cash flows” could be explored. There are
definite hints of a regression based approach using “different types of commodities
[that] are statistically better predictors of forward jet fuel prices”.
Given the extent and detail of reporting given in Item 7, information on commod-
ity risk management activities under Item 7A of the 10-K, Quantitative and Qualitative
Disclosures About Market Risk, is relatively sparse. Though uncommon, the detailing of
risk management items under Item 7, rather than in Item 7A, does happen. For example,
investment banks such as Goldman Sachs forgo reporting Item 7A altogether, giving all
risk management items in Item 7. This is more than appropriate for financial intermedi-
aries where involvement in risk management is the primary commercial activity of the
firm. In turn, treatment of jet fuel hedging is strategically important to the commer-
cial activities of LUV and reporting under Item 7 is, again, appropriate. Given this, the
discussion in Item 7A includes a statement of “commodity risk hedging philosophy”:
Hedging
The Company utilizes financial derivative instruments, on both a short-term and a
long-term basis, as a form of insurance against the potential for significant increases
in fuel prices. The Company believes there is significant risk in not hedging against
the possibility of such fuel price increases.
This is followed by a precise statement of the risk exposure, indicating the method of
measurement, together with a traditional sensitivity analysis:
the Company expects to consume approximately 1.5 billion gallons of jet fuel in
2011. Based on this usage, a change in jet fuel prices of just one cent per gallon
would impact the Company’s Fuel and oil expense by approximately $15 million
per year, excluding any impact of the Company’s derivative instruments.
The use of active risk management and the current stance on future fuel pricing is then
given:
The Company may increase or decrease the size of its fuel hedge based on its
expectation of future market prices, as well as its perceived exposure to cash
collateral requirements contained in the agreements it has signed with various
counterparties. In 2010, the Company added to its fuel hedging position related to
expected future fuel purchases.
if market prices for the commodities used in the Company’s fuel hedging activities
were to decrease by 33 percent from market prices as of December 31, 2010, given
the Company’s fuel derivative portfolio, its aircraft collateral facilities, and its
investment grade credit rating, it would have to provide an additional $497 million
in cash collateral to its current counterparties.
While LUV does have sufficient cash on hand to meet such as demand, at some point
cash draws would impinge on other demands for cash, e.g. to make payments on plane
deliveries. The final substantive information in Item 7A concerns the approach to
operational risk:
Due to the significance of the Company’s fuel hedging program and the emphasis
that it places on utilizing fuel derivatives to reduce its fuel price risk, the Company
has created a system of governance and management oversight and has put in place
a number of internal controls designed so that procedures are properly followed
and accountability is present at the appropriate levels. For example, the Company
Fair value of fuel derivatives …………………………. $ 114 $ (238) $ (3) $ 79 $ 189 $1* $ 142
Cash collateral held by CP …………………………... (60) 125 — — — — 65
Aircraft collateral pledged to CP…. — 113 — — — — 113
If credit rating is investment grade, fair value of fuel
derivative level at which:
Cash is provided to CP.. 0 to (300) 0 to (125) >(75) >(75) >(75)
or >(700) or >(535)
Cash is received from CP .......………………........ >40 >150 >200*** >125*** >250
Aircraft is pledged to CP ........ (300) to (700) (125) to (535) N/A N/A N/A
If credit rating is non-investment grade , fair value of
fuel derivative level at which:
Cash is provided to CP...........……………….......... 0 to (300) 0 to (125) ** ** **
or >(700) or >(535)
Cash is received from CP.......……………….......... ** ** ** ** **
Aircraft is pledged to CP........………………........... (300) to (700) (125) to (535) N/A N/A N/A
* Sum of counterparties with fair value of fuel derivatives <$5M and no risk of the Company posting collateral.
** Cash collateral is provided at 100 percent of fair value of fuel derivatives contracts.
*** Thresholds may vary based on changes in credit ratings within investment grade.
11/16/2012 6:45:44 PM
362 Risk Management Applications
has put in place controls designed to: (i) create and maintain a comprehensive risk
management policy; (ii) provide for proper authorization by the appropriate levels
of management; (iii) provide for proper segregation of duties; (iv) maintain an
appropriate level of knowledge regarding the execution of and the accounting for
derivative instruments; and (v) have key performance indicators in place in order
to adequately measure the performance of its hedging activities. The Company
believes the governance structure that it has in place is adequate given the size and
sophistication of its hedging program.
While this does reflect a concern with operational risk, information provided by other
companies on the corporate governance structure for risk management, in general, and
commodity risk management, in particular, e.g. Figure 3.8 for BHP, is not given.
Fuel contracts
Airline operators are inherently dependent upon energy to operate and, therefore,
are impacted by changes in jet fuel prices. Furthermore, jet fuel and oil typically
represents one of the largest operating expenses for airlines. The Company endeav-
ors to acquire jet fuel at the lowest possible cost and to reduce volatility in operating
expenses through its fuel hedging program. Because jet fuel is not widely traded on
an organized futures exchange, there are limited opportunities to hedge directly in
jet fuel. However, the Company has found that financial derivative instruments in
other commodities, such as crude oil, and refined products such as heating oil and
unleaded gasoline, can be useful in decreasing its exposure to jet fuel price volatil-
ity. The Company does not purchase or hold any financial derivative instruments
for trading purposes.
The Company has used financial derivative instruments for both short-term and
long-term time frames, and typically uses a mixture of purchased call options,
collar structures (which include both a purchased call option and a sold put option),
call spreads (which include a purchased call option and a sold call option), and fixed
price swap agreements in its portfolio. Generally, when the Company perceives that
prices are lower than historical or expected future levels, the Company prefers to
use fixed price swap agreements and purchased call options.
However, at times when the Company perceives that purchased call options have
become too expensive, it may use more collar structures and call spreads. Although
the use of collar structures and swap agreements can reduce the overall cost of
hedging, these instruments carry more risk than purchased call options in that
the Company could end up in a liability position when the collar structure or swap
agreement settles. With the use of purchased call options and call spreads, the
Company cannot be in a liability position at settlement.
This discussion hints at a constant question in the structure of commodity risk hedges:
whether to use an insurance approach, where premium costs are locked in upfront, or to
use hedges with little or no up front costs but that may require cash flow contributions in
the future. LUV seeks to answer such questions by using an active hedging strategy that
switches between derivative instruments based on the perceived costs, as determined by
a “view” on future jet fuel prices.
A fundamental question in commodity risk management that was skirted in the
discussion of “risk management philosophy” in Part 2 concerns the impact of derivative
accounting on the incentive to hedge. At numerous point in the 10-K, LUV given copious
detail on the accounting methods used and the implications of accounting rules on the
selection of particular derivative securities to use in the jet fuel hedging program.
The Company evaluates its hedge volumes strictly from an “economic” standpoint
and does not consider whether the hedges qualified or will qualify for hedge
accounting. The Company defines its “economic” hedge as the net volume of
fuel derivative contracts held, including the impact of positions that have been
offset through sold positions, regardless of whether those contracts qualify for
hedge accounting. For 2010, the Company had fuel derivatives in place related to
approximately 40 percent of its fuel consumption. As of December 31, 2010, the
Company had fuel derivative instruments in place to provide coverage on a large
portion of its 2011 estimated fuel consumption
Table 3.66 provides information about the accounting treatment of various derivative
security positions. The desired accounting treatment for LUV is described as follows:
Upon proper qualification, the Company accounts for its fuel derivative instruments
as cash flow hedges. All derivatives designated as hedges that meet certain
requirements are granted hedge accounting treatment. Generally, utilizing the
hedge accounting, all periodic changes in fair value of the derivatives designated as
hedges that are considered to be effective, are recorded in AOCI until the underlying
jet fuel is consumed . . .
* Represents the position of each trade before consideration of offsetting positions with each counterparty and does not include the
impact of cash collateral deposits provided to or received from counterparties.
11/16/2012 6:45:44 PM
Airlines and Jet Fuel Hedging 365
Unfortunately, not all derivative security hedges are ensured to qualify as cash flow
hedges. The following discussion associated with the requisite accounting treatment is
a classic illustration of the deterring affect that complicated accounting rules can have
on legitimate commercial risk management activities:
The Company is exposed to the risk that periodic changes will not be effective,
as defined, or that the derivatives will no longer qualify for hedge accounting.
Ineffectiveness results when the change in the fair value of the derivative instrument
exceeds the change in the value of the Company’s expected future cash outlay to
purchase and consume jet fuel. To the extent that the periodic changes in the fair value
of the derivatives are ineffective, the ineffective portion is recorded to Other (gains)
losses, net in the Consolidated Statement of Income. Likewise, if a hedge ceases to
qualify for hedge accounting, any change in the fair value of derivative instruments
since the last period is recorded to Other (gains) losses, net in the Consolidated
Statement of Income in the period of the change; however, any amounts previously
recorded to AOCI would remain there until such time as the original forecasted
transaction occurs, at which time these amounts would be reclassified to Fuel and
oil expense. When the Company has sold derivative positions in order to effectively
“close” or offset a derivative already held as part of its fuel derivative instrument
portfolio, any subsequent changes in fair value of those positions are marked to
market through earnings. Likewise, any changes in fair value of those positions
that were offset by entering into the sold positions are concurrently marked to
market through earnings. However, any changes in value related to hedges that
were deferred as part of AOCI while designated as a hedge, would remain until the
originally forecasted transaction occurs. In a situation where it becomes probable
that a hedged forecasted transaction will not occur, any gains and/or losses that
have been recorded to AOCI would be required to be immediately reclassified into
earnings.
This complex accounting discussion is followed with the proviso: “The Company did
not have any such situations occur during 2010, 2009, or 2008.”
The implications of this dense accounting treatment on actual commodity risk
management activities at LUV is apparent in the following:
prices, the number of derivative positions the Company holds, significant weather
events affecting refinery capacity and the production of refined products, and the
volatility of the different types of products the Company uses in hedging.
These qualifications are followed by the assuring statement: “even though derivatives
may not qualify for hedge accounting, the Company continues to hold the instruments
as management believes derivative instruments continue to afford the Company the
opportunity to stabilize jet fuel costs.” However, even in a firm the size of LUV, the
complications associated with current “hedge accounting” rules involve considerable
effort to satisfy:
In effect, by using “regression and other statistical analyses” LUV appears to engage
in a particular type of optimal hedge ratio estimation, albeit for largely accounting
purposes.2
367
other crudes, such as UK Brent, at a discount. The common view, that the NYMEX contract is for WTI, is
due to the contract delivery point being any pipeline or storage facility at Cushing, Oklahoma, a delivery
site that tends to favor delivery of WTI. Gray (2002) discusses the calculation of API gravity.
7. The following discussion is derived from information provided on the Syncrude website: www.syncrude.
com. Chastko (2004) is a helpful source on the history of the Alberta oil sands.
8. Shares in the Syncrude project do change hands infrequently and companies merge and change names.
Nexen was formerly called Canadian Occidental and Mocal was formerly called Mitsubishi Oil (fully
owned by Nippon). Canadian Oil Sands increased ownership share in 2003 by buying out the share of
EnCana, which, in 2002, was formed from the merger of Pan-Canadian Energy Corp and Alberta Energy
Company Ltd. Prior to this merger, the EnCana share was held by Alberta Energy Company Oil Sands
Partnership.
merchants of the Staple in 1559, when it was agreed that the latter should be free of brokers when buying.
It was asserted in 1562 that in most foreign countries no ‘stranger’ bought or sold except through a sworn
broker, and the English Statute Book contains a number of regulations of similar import. Such arrange-
ments were general, being due to the universal prejudice against foreigners.” Buckley (p.591) also makes
another observation which is indicative of the pervasiveness of brokers at Gresham’s time: “Dealings in
Bills of exchange without the intervention of a broker were exceptional.”
12. The identification of this early trade as “futures” contracting is found in Gelderblom and Jonker (2005).
This approach is at variance with the conventional view that futures trading began in Chicago in the 19th
century or the less conventional view that such trading began in the 17th-century Japanese rice market
(Schaede 1987; West 2000).
13. Emery (1896, pp.51–3) provides a number of references to late 19th-century German and French sources
on options trading. The connection between German and English terminology is also discussed (p.91).
Courtadon (1982) examines option trading practices on the Paris bourse.
14. The acronym VOC is a reference to the English to Dutch translation of the Dutch East India Company,
as the Verenigde Oostindische Compagnie.
15. The primary documentation associated with the Dutch Edict of 1610, which removed legal protection for
“windhandel” contracts, contains an important memoir, probably written by Isaac le Maire, which outlines
arguments in favor of retaining short sales (De Marchi and Harrison 1994; van Dillen 1930; van Dillen
et al. 2007). A number of arguments draw on the similarity of the trade in shares to the trade in goods:
“the authors proceed from free trade in goods (perfectly conventional from a common weal point of view),
move on to the freedom to make forward purchases of commodities (accepted practice for at least several
decades), and end with the freedom to trade in shares. This bundling, as well as the progression itself,
may have been intended to persuade the reader that (all) share trading practices should unquestionably be
regarded as no different in principle from trade in goods” (De Marchi and Harrison 1994, p.55).
16. This section is based on Poitras (2000, Ch.10). Early sources on the tulipmania, such as Francis (1850) and
Mackay (1841), refer to the “tulipomania.”
17. The source of this anecdote is Mackay. Garber (1989, p.540, n.12) casts considerable doubt on the validity of
Mackay’s account.
18. Waermondt and Gaergoedt translate loosely as True-Mouth and Greedy-Goods. The dialogue format
was popular in the 16th and 17th centuries. This approach was used in a number of other important finan-
cial works of this period, such as de la Vega (1688) and Wilson (1572). As for the primary literature, Posthu-
mus (1929, p.436) reports that: “At least fifty booklets written by defenders and opponents were published,
as well as a great number of prints and caricatures.” Posthumus states that the best source of information
is the GW dialogues.
19. The aas or ace (plural azen) is a Dutch unit of weight, which equals about one-twentieth of a gram.
20. The basic mechanics of tulip production argue against widespread option trading for those directly involved
in the tulip trade. Tulip growers wanted to sell bulbs for future delivery. Due to potential and actual limita-
tions in the supply of bulbs, other potential market participants were not in a position to quote call option
prices from created hedged positions.
21. There is little support for Malkiel’s point about the devastating economic consequences of the mania.
Though there were a few traders who lost large down payments which had been made, most of the contracts
resulting from the crash were either cancelled or settled with nominal payments. Posthumus (1929, p.448)
concludes that “socially the losses had been very small. The growers had been affected most of all by the
crisis, having grown and sold their bulbs, without getting any money in return.”
22. Of the various recent exchange mergers, in 1999, the NYMEX merged with the COMEX to form the New
York Board of Trade, which subsequently merged with the CME. In another important development, the
CBT merged with the CME in 2007. Markham and Harty (2008) have a detailed discussion of more recent
mergers and developments.
23. The Senate debate included a vote on the George amendment, which aimed to ban futures trading
altogether. This amendment was prompted by the concern of Southern members about the use of tax-
to-destroy as a method of dealing with the anti-speculation arguments of the agrarians and populists.
This amendment was defeated by 51 to 19. However, as it turns out, the Southern supporters of the
George amendment held the balance in the House vote to suspend rules, which led to the defeat of the
Hatch-Washburn bill.
24. Williams (1995) is a helpful source for more in-depth discussion of the issues surrounding the definition
and legal application of manipulation. The following discussion draws liberally from that source.
25. Fay (1982) and Williams (1995) are excellent sources on this topic. The Sumitomo copper corner is similar
in many ways to the Hunt silver dealings, though there were some significant differences, e.g. the Sumitomo
losses were the result of a trading operation within a larger corporate entity.
26. The Hunt family fortune was founded by the eccentric H.L. Hunt, who left three sets of children (Hurt
1981). Bunker and Herbert were from the first of H.L. Hunt’s families. This first family also includes Lamar
Hunt, owner of the Kansas City Chiefs. Circa 1980, the centerpieces of the Hunt family fortune were
Penrod Drilling, an oil drilling company, and Placid Oil, the holder of large oil reserves and leases. The two
companies, together with the family’s other assets, were controlled through an elaborate network of over
200 companies and trust funds (Williams 1995, p.20).
27. Johnson (1981, p.97) reports: “In fact, its quite rare for there to be manipulation cases. There are, perhaps,
not more than a half dozen manipulation cases of any true significance that have been reported in the
courts.”
28. General Foods v. Brannan, 170 F.2d 220 (7th Circuit, 1948), p.231, quoted in Williams (1995, p.5).
29. In November 1991, David Threlkeld, a US copper broker operating on the LME, received a letter requesting
him to backdate trade confirmation dates for a fake deal worth $425 million. This letter was apparently
from Hamanaka. Recognizing the illegality of the request, Threlkeld passed the letter along to the head
of the LME. The LME’s view on the letter was, more or less, that Threlkeld was well advised to keep quiet
over the matter to avoid getting sued. At this point, it is not clear whether the LME did anything to follow
up on the Threlkeld complaint.
6. As before, E[.] is the conditional mathematical expectation of S(T) given the information available at time
t=0. For notational simplicity the conditioning information is dropped because, in virtually ever case
encountered in the analysis of derivative securities, expectations are conditional.
equivalent to the slope coefficient in a bivariate normal regression of spot on futures prices. Early discus-
sion of the OLS result focuses on whether the price variables should be expressed in levels, changes or rates
of return, e.g. Myers and Thompson (1989); Toevs and Jacob (1986); Witt et al. (1987). More recent applica-
tions, e.g. Bertus et al. (2009), Godbey and Hilliard (2007), have used alternative estimation methods to
OLS to determine the minimum variance solution. Following Schwartz (1997), using stochastic differential
equations to specify factors for the convenience yield, the spot price and, possibly, an interest rate process,
the Kalman filtering technique can be used to determine the minimum variance solution. As illustrated in
Lien (2005, 2012), which particular technique is “optimal” to determine the hedge ratio in a practical sense
is still an unresolved issue.
13. Outside of agricultural applications, the insurance implications to risk management receive limited atten-
tion. Available studies typically focus on the relationship between insurance and firm financing decisions,
e.g. Grace and Rebello (1993), MacMinn (1987), Mayers and Smith (1982) and Rebello (1995). Paulson et al.
(2008) discuss the insurance approach to risk management in ethanol production.
14. This section follows Poitras (2002, Secs 6.H to 6.J).
15. Because h t enters with a minus sign, this defines a short futures position to be a positive quantity.
16. The transformation from terminal wealth, Wt+1, to terminal profit, πt+1, follows because the expectation of
U[Wt] reduces to a constant which does not affect the optimization. In addition, it will always be assumed
that the only state (conditioning) variables of interest are R s and Rf . However, in general, this need not be
true.
17. When employed in common usage, the terms strategic risk management and enterprise risk management
are often used interchangeably. However, enterprise risk management has become more commonly used
since the actuarial societies in the USA and the UK adopted “enterprise risk management” as a recognized
area of specialization. For example, the Society of Actuaries in the USA now offers a Chartered Enterprise
Risk Analyst designation. Enterprise risk management is also somewhat broader description involved in
integrating risk into “business strategy” and, perhaps more importantly, dealing with legal requirements
for internal control mechanisms. In this vein, Section 404 of the Sarbanes-Oxley Act requires US publicly
traded corporations to utilize a “control framework” for internal control assessments. In addition, since
2007 the SEC has, in conjunction with the Committee of Sponsoring Organizations of the Treadway Com-
mission (COSO), placed increasing scrutiny on top-down risk assessment, especially fraud risk assessment.
In this context, strategic risk management focuses on the “business strategy” component of risk assess-
ment while enterprise risk management encompasses both strategic risk management and internal control
assessment, e.g. Moeller (2007). As issues of internal control receive little attention in the following discus-
sion, reference is made to strategic, as opposed to enterprise, risk management.
Insurance Co. and Agrinational Insurance Co., a branch of Archer-Daniels-Midland. In 2009, the top five
writers of multi-peril crop insurance in the United States were: Wells Fargo Insurance Group (16.5 percent
market share); Ace INA Group (16.2 percent); NAU Country Insurance Co. (10.6 percent); Great American
Property & Casualty Insurance Group (9.2 percent); and Allianz of America (8.1 percent).
The Company has floating-to-fixed interest rate swap agreements associated with its $600 million
floating rate term loan agreement and its $332 million term loan agreement that are accounted for as
cash flow hedges. These interest rate hedges have fixed the interest rate on the $600 million floating-
rate term loan agreement at 5.223 percent until maturity, and for the $332 million term loan agreement
at 6.64 percent until maturity. The ineffectiveness associated with these hedges for 2010 and 2009 was
not material.
Ackerman, C., R. McNally and D. Ravenscraft (1999) “The Performance of Hedge Funds: Risk, Return and
Incentive,” Journal of Finance 54: 833–74.
Acs, S., P. Berensten, R. Huirne and M. van Asseldonk (2009) “Effect of yield and price risk on conversion
from conventional to organic farming,” Australian Journal of Agricultural & Resource Economics
53: 393–411.
Adam, T. (2002) “Do Firms Use Derivatives to Reduce their Dependence on External Capital Markets,”
European Finance Review 6: 163–87.
Adam, T. and C. Fernando (2006) “Hedging, Speculation and Shareholder Value,” Journal of Financial
Economics 81: 283–309.
Adam, T. and C. Fernando (2008) “Can Companies Use Hedging Programs to Profit from the Market?
Evidence from Gold Producers,” Journal of Applied Corporate Finance 20: 86–97.
Adam, T., S. Dasgupta and S. Titman (2007) “Financial Constraints, Competition, and Hedging in Indus-
try Equilibrium,” Journal of Finance 62: 2445–73.
Adam-Müller, A. and K. Wong (2003) “The Impact of Delivery Risk on Optimal Production and Futures
Hedging,” European Finance Review 7: 459–477.
Adler, M. and Dumas, B. (1984) “Exposure to Currency Risk: Definition and Measurement,” Financial
Management (Summer): 41–50.
Adler, N. and K. Smilowitz (2007) “Hub-and-Spoke Network Alliances and Mergers: Price-Location
Competition in the Airline Industry,” Transportation Research: Part B: Methodological 41: 394–409.
Agriculture and Agri-Foods Canada (1998) “The Federal-Provincial Crop Insurance Program, An Inte-
grated Environmental-Economic Assessment” (October), Ottawa.
Ahn, M. and W. Falloon (1991) Strategic Risk Management. Chicago: Probus.
Al Janabi, M. (2009) “Commodity Price Risk Management: Valuation of Large Trading Portfolios Under
Adverse and Illiquid Market Settings,” Journal of Derivatives & Hedge Funds 15: 15–50.
Alexander, C., M. Prokopczuk and A. Sumawong (2012) “The (De)Merits of Minimum-Variance Hedging:
Application to the Crack Spread,” SSRN Working Paper (1986047, Jan.).
Alexander, G., W. Sharpe and J. Bailey (2000) Fundamentals of Investment (3rd edn). Englewood Cliffs,
NJ: Prentice-Hall.
Allayannis, G. and J. Weston (2001) “The Use of Foreign Currency Derivatives and Firm Market Value,”
Review of Financial Studies 14: 243–276.
Anand, A. (2000) “The Impossibility of Full Carry Markets,” Derivatives Quarterly 7: 31–48.
375
Andersen, T. (2006) Global Derivatives: A Strategic Risk Management Perspective. New York: Financial
Times/Prentice Hall.
Andersen, T. and P. Schrøder (2010) Strategic Risk Management Practice: How to Deal Effectively With
Major Corporate Exposures. Cambridge, UK: Cambridge University Press.
Andren, N., H. Jankensgård and L. Oxelheim (2005) “Exposure-based Cash-Flow-at-Risk: An Alternative
to Value-at-Risk for Industrial Companies,” Journal of Applied Corporate Finance 17: 76–87.
Aretz, K., B. Söhnke and G. Dufey (2007) “Why Hedge? Rationales for Corporate Hedging and Value
Implications,” Journal of Risk Finance 8: 434–49.
Arthur Andersen (1998) “Financial Risk Management,” mimeo.
Artzner, P., F. Delbaen, J.M. Eber and D. Heath (1999) “Coherent Measures of Risk,” Mathematical
Finance 9: 203–228.
Atkins, F. and A. MacFadyen (2008) “A Resource Whose Time Has Come? The Alberta Oil Sands as an
Economic Resource,” Energy Journal (Special Issue) 29: 77–97.
Ayer, A.J. (1936) Language, Truth and Logic. London: Gollancz (2nd edn, 1946).
Baillie, R. and R. Myers (1991) “Bivariate GARCH Estimation of the Optimal Commodity Futures Hedge,”
Journal of Applied Econometrics 6: 109–24.
Bank of International Settlements, Basle Committee on Banking Supervision (1996) Amendment to the
Capital Accord to Incorporate Market Risks. Basle, Switzerland: BIS.
Barbour, V. (1950) Capitalism in Amsterdam in the 17th Century, Ann Arbor, Mich.: University of Michi-
gan Press.
Barboza, D. (2010) “Rio Tinto Goes to Quarterly Pricing on Iron Ore,” New York Times (April 9): B7.
Bartram, S.M. (2005) “The Impact of Commodity Price Risk on Firm Value—An Empirical Analysis of
Corporate Commodity Price Exposures,” Multinational Finance Journal 9: 161–87.
Battermann, H.L. and Broll, U. (2004) “Price Uncertainty, Futures Markets and Correlation,” Interna-
tional Economic Journal 18: 237–44.
Beckmann, J. (1846) A History of Inventions, Discoveries, and Origins (4th edn), 2 vols. London: Bohn.
Bell, A., C. Brooks and P. Dryburgh (2007) “Interest Rates and Efficiency in Medieval Wool Forward
Contracts,” Journal of Banking and Finance 31: 361–80.
Bell, D. (1995) “Risk, Return and Utility,” Management Science 41: 23–30.
Benninga, S., R. Eldor and I. Zilcha (1984) “The Optimal Hedge Ratio in Unbiased Futures Markets,”
Journal of Futures Markets 4: 155–59.
Bergfjord, O. (2009) “Risk perception and Risk Management in Norwegian Aquaculture,” Journal of Risk
Research 12: 91–104.
Bernard, J., L. Khalaf, M. Kichian and S. McMahon (2008) “Forecasting Commodity Prices: GARCH,
Jumps and Mean Reversion,” Journal of Forecasting 27: 279–91.
Bertus, M., J. Godbey and J. Hilliard (2009) “Minimum Variance Cross Hedging Under Mean-Reverting
Spreads, Stochastic Convenience Yields, and Jumps: Application to the Airline Industry,” Journal of
Futures Markets 29: 736–56.
Bhaskar, R. (1978) A Realist Theory of Science (2nd edn). Brighton, UK: Harvester.
Bielza, M., A. Garrido and J.M. Sumpsi (2007) “Finding Optimal Price Risk Management Instruments:
The Case of the Spanish Potato Sector,” Agricultural Economics 36: 67–78.
Blaug, M. (1992) The Methodology of Economics: Or, How Economists Explain (2nd edn). Cambridge, UK:
Cambridge University Press.
Bodnar, G.M., G.S. Hayt, R.C. Marston and C.W. Smithson (1996) “Wharton Survey of Derivatives Usage
by US Non-Financial Firms,” Financial Management 25: 113–33.
Bodnar, G.M., G.S. Hayt and R.C. Marston (1998) “1998 Wharton Survey of Derivatives Usage by US
Non-Financial Firms,” Financial Management 27: 70–91.
Boland, L. (1979) “A Critique of Friedman’s Critics,” Journal of Economic Literature 17: 503–22.
Boland, L. (1991) “Current Views on Economic Positivism,” in D. Greenaway, M. Bleaney and I. Stewart
(eds) Companion to Contemporary Economic Thought: 88–104.
Bollman, K., P. Garcia and S. Thompson (2003) “What Killed the Diammonium Phosphate Futures Con-
tract?” Review of Agricultural Economics 25: 483–506.
Bond, G. and S. Thompson (1986) “Optimal Commodity Hedging within the Capital Asset Pricing
Model,” Journal of Futures Markets 6: 421–31.
Borenstein, S. (2011) “Why Can’t US Airlines Make Money?” American Economic Review 101: 233–7.
Börger, R., Á. Cartea, R. Kiesel and G. Schindlmayr (2009) “Cross-Commodity Analysis and Applications
To Risk Management,” Journal of Futures Markets 29: 197–217.
Borovkova, S. (2006) “Detecting Market Transitions and Energy Futures Risk Management Using Princi-
pal Components,” European Journal of Finance 12: 495–512.
Bowden, R. and J. Zhu (2008) “The Agribusiness Cycle and its Wavelets,” Empirical Economics 34: 603–22.
Brennan, D., J. Williams and B. Wright (1997) “Convenience Yield Without the Convenience: A Spatial-
Temporal Interpretation of Storage under Backwardation,” Economic Journal 107: 1009–22.
Brennan, M. (1958) “The Supply of Storage,” American Economic Review 47: 50–72.
Bresnahan, T. and P. Spiller (1986) “Futures Market Backwardation Under Risk Neutrality,” Economic
Inquiry 24: 429–41.
Brockett, P. and Y. Kahane (1992) “Risk, Return, Skewness and Preference,” Management Science (June):
851–66.
Broehl, W. (1992) Cargill: Trading the World’s Grain. Hanover, NH: University of New England Press.
Brooks, C., A. Clare, J. Dalle Molle and G. Persand (2005) “A Comparison of Extreme Value Theory
Approaches for Determining Value at Risk,” Journal of Empirical Finance 12: 339–52.
Brown, G. and K. Toft (2002) “How Firms Should Hedge,” Review of Financial Studies 15: 1283–324.
Brown, G., P. Crabb and D. Haushalter (2006) “Are Firms Successful at Selective Hedging?” Journal of
Business 79: 2925–49.
Buckley, H. (1924) “Sir Thomas Gresham and the Foreign Exchanges,” Economic Journal XXXIV: 589–601.
Buguk, C. and W.B. Brorsen (2005) “Is a Futures Market Viable in Turkey? The Case of a Cotton Futures
Market,” Journal of International Food & Agribusiness Marketing 17: 135.
Bullock, D.W., W.W. Wilson and B.L. Dahl (2007) “Strategic Use of Futures and Options by Commodity
Processors,” International Review of Economics and Finance 16: 578–91.
Burgert, P. and T. Furukawa (1996) “Sumitomo’s Copper Crisis Deals Aftershocks: Company Says it’s
Standing Behind Obligations,” American Metal Market (June 17).
Burton, K. and M. Leising (2006) “Amaranth Says Funds Lost 50% on Gas Trades This Month,” Bloomberg
News (September 18).
Busse, M. (2002) “Firm Financial Condition and Airline Price Wars,” RAND Journal of Economics 33:
298–318.
Byers, J.W. (2005) “Risk Management and Monetizing the Commodity Storage Option,” Natural Gas &
Electricity 21: 1–8.
Carter, C. and C. Ghia (2007) “The Working Curve and Commodity Storage Under Backwardation,”
American Journal of Agricultural Economics 89: 864–72.
Carter, D., D. Rogers and B. Simkins (2006a) “Hedging and Value in the US Airline Industry,” Journal of
Applied Corporate Finance 18: 21–33.
Carter, D., D. Rogers and B. Simkins (2006b) “Does Hedging Affect Firm Value? Evidence from the US
Airline Industry,” Financial Management 35: 53–86.
Casassus, J. and P. Collin-Dufresne (2005) “Stochastic Convenience Yield Implied from Commodity
Futures and Interest Rates,” Journal of Finance 60: 2283–331.
Castaldo, J. (2011) “Pork Belly Futures (1961–2011),” Canadian Business (8/16/2011) 84(14): 11.
Chamberlin, G. (1983) “A Characterization of the Distributions that Imply Mean-Variance Utility Func-
tions,” Journal of Economic Theory 29: 185–201.
Chan, K., P. Gray and B. van Campen (2008) “A New Approach to Characterizing and Forecasting Elec-
tricity Price Volatility,” International Journal of Forecasting 24: 728–43.
Chance, D. (1998) An Introduction to Derivative Securities (4th edn). New York: Dryden.
Chance, D. and R. Brooks (2010) An Introduction to Derivatives and Risk Management (8th edn). Mason,
OH: Thomson-Southwestern Publishing.
Chance, D. and. M. Hemler (1993) “The Impact of Delivery Options on Futures Prices: A Survey,” Journal
of Futures Markets 13: 127–56.
Chang, C.-L., M. McAleer and R. Tansuchat (2011) “Crude Oil Hedging Strategies Using Dynamic Multi-
variate GARCH,” Energy Economics 33: 912–23.
Chantarat, S., C. Barrett, A. Mude and C. Turvey (2007) “Using Weather Index Insurance to Improve
Drought Response for Famine Prevention,” American Journal of Agricultural Economics 89: 1262–8.
Chastko, P. (2004) Developing Alberta’s Oil Sands: from Karl Clark to Kyoto. Calgary: University of Cal-
gary Press.
Chen, S.-S., C.F. Lee and K. Shrestha (2003) “Futures Hedge Ratios: A Review,” Quarterly Review of Eco-
nomics & Finance 43: 433–66.
Chi, J. and W. Koo (2009) “Carriers’ Pricing Behaviors in the United States Airline Industry,” Transporta-
tion Research: Part E: Logistics and Transportation Review 45: 710–24.
Chincarini, L. (2008) “A Case Study on Risk Management: Lessons from the Collapse of Amaranth
Advisors LLC,” Journal of Applied Finance 18: 152–74.
Chiu, D. and S. Foerster (1997) “Using Derivatives to Manage Risk,” Business Quarterly 61(3): 57–64.
Chiu, W.H. (2010) “Skewness Preference, Risk Taking and Expected Utility Maximisation,” Geneva Risk
and Insurance Review 35: 108–129.
Cho, D. and G. McDongall (1990) “The Supply of Storage in Energy Futures Markets,” Journal of Futures
Markets 10: 611–21.
Chowdhry, B. and J.T.B. Howe (1998) “Corporate Risk Management for Multinational Corporations:
Financial and Operational Hedging Policies,” European Finance Review 2: 229–46.
Chung, S. (2009) “Out-of-sample Hedge Performances for Risk Management in China Commodity
Futures Markets,” Asian Economic Journal 23: 349–72.
Ciliberto, F. and E. Tamer (2009) “Market Structure and Multiple Equilibria in Airline Markets,”
Econometrica 77: 1791–828.
Claessens, S. and P.N. Varangis (1993) “Implementing Risk Management Strategies in Costa Rica’s Coffee
Sector,” Managing Commodity Price Risk in Developing Countries. Baltimore and London: Johns
Hopkins University Press for the World Bank.
Coakley, J., J. Dollery and N. Kellard (2011) “Long Memory and Structural Breaks on Commodity Futures
Markets,” Journal of Futures Markets 31: 1076–113.
Cole, S. and B. Kirwan (2009) “Between the Corporation and the Household: Commodity Prices, Risk
Management, and Agricultural Production in the United States,” American Journal of Agricultural
Economics 91: 1243–9.
Collins, D. (2006) “A Billion Here, a Billion There . . .,” Futures 35: 58.
Commodity Futures Trading Commission (2008) “Staff Report on Commodity Swap Dealers and Index
Traders with Commission Recommendations,” September, Washington, DC: CFTC.
Commodity Futures Trading Commission (2008a) “Commission Actions in Response to the
‘Comprehensive Review of the Commitments of Traders Reporting Program’,” (Dec. 5).
Conejo, A., F. Nogales, M. Carrion and J. Morales (2010) “Electricity Pool Prices: Long-Term Uncertainty
Characterization for Futures-Market Trading and Risk Management,” Journal of the Operational
Research Society 61: 235–45.
Copeland, T. and M. Copeland (1999) “Managing Corporate FX Risk: A Value-Maximizing Approach,”
Financial Management 28: 68–75.
Cornell, B. (1997) “Cash Settlement when the Underlying Securities are Thinly Traded: A Case Study,”
Journal of Futures Markets 17: 855–71.
Cornew, R. (1988) “Commodity Pool Operators and Their Pools: Expenses and Profitability,” Journal of
Futures Markets 8: 617–37.
Costa, E.F. and S.C. Turner (2001) “Price Risk Management for Peanut Meal,” Journal of International
Food & Agribusiness Marketing 13: 99–110.
Courtadon, G. (1982) “A Note on the Premium Market for the Paris Stock Exchange,” Journal of Banking
and Finance 6: 561–5.
Cowing, C. (1895) Populists, Plungers, and Progressives: A Social History of Stock and Commodity
Speculation, 1890–1936. Princeton, NJ: Princeton University Press (1965 reprint).
Culp, C. and M. Miller (1995) “Metallgesellschaft and the Economics of Synthetic Storage,” Journal of
Applied Corporate Finance 7: 62–76.
Culp, C. and M. Miller (eds) (1999) Corporate Hedging in Theory and Practice. London: Risk Books.
Culp, C., M. Miller and A. Neves (1998) “Value at Risk: Uses and Abuses,” Journal of Applied Corporate
Finance 10: 26–38.
Dahlgran, R.A. (2000) “Cross-Hedging the Cottonseed Crush: A Case Study,” Agribusiness 16: 141–58.
Damodaran, A. (2008) Strategic Risk Taking: A Framework for Risk Management. Upper Saddle River, NJ:
Wharton School Publishing.
Danielson, J., B. Jorgenson and C. deVries (1998) “The Value of Value-at-Risk: Statistical, Financial and
Regulatory Considerations,” Federal Reserve Bank of New York, Economic Policy Review: 1–23.
Danthine, J. (1978) “Information, Futures Prices and Stabilizing Speculation,” Journal of Economic Theory
17: 418–43.
Davidson, W., P. Chandy and M. Cross (187) “Large Losses, Risk Management and Stock Returns in the
Airline Industry,” Journal of Risk and Insurance 54: 162–72.
Davis, A. (2006) “How Giant Bets on Natural Gas Sank Brash Hedge-Fund Trader,” Wall Street Journal
(Sept. 19): A1.
Dawkins, W. (1996) “Sumitomo Loses its Way After Nearly 400 Years,” Financial Post (June 15): 8.
Deaton, A. and G. Laroque (1996) “Competitive Storage and Commodity Price Dynamics,” Journal of
Political Economy 104: 896–923.
de la Vega, J. (1688) Confusion de Confusiones, reprinted in M. Fridson (ed.) (1996) Extraordinary Popular
Delusions and the Madness of Crowds; and, Confusion de Confusiones (reprints of classic texts). New
York: Wiley.
Deloitte (2006) The China Aviation Oil Debacle. Belgium: Deloitte Touche Tohmatsu (January).
De Marchi, N. and P. Harrison (1994) “Trading ‘in the Wind’ and with Guile: The Troublesome Matter of
the Short Selling of Shares in Seventeenth-Century Holland,” in N. de Marchi and M. Morgan (eds)
Higgling: Transactors and their Markets in the History of Economics, Annual Supplement to History
of Political Economy 26.
De Marzo, P. and D. Duffie (1991) “Corporate Financial Hedging With Proprietary Information,” Journal
of Economic Theory 53: 261–86.
Deng, X., B. Barnett and D. Vedenov (2007) “Is There a Viable Market for Area-Based Crop Insurance?”
American Journal of Agricultural Economics 89: 508–19.
de Roover, R. (1948) Banking and Credit in Medieval Bruges. Cambridge, Mass.: Harvard University Press.
de Roover, R. (1949) Gresham on Foreign Exchange. London: Harvard University Press.
de Roover, R. (1954) “New Interpretations in the History of Banking,” Journal of World History: 38–76;
reprinted in J. Kirshner (ed.) (1974, Ch.5).
Dia, M. and D. Zéghal (2008) “Fuzzy Evaluation of Risk Management Profiles Disclosed in Corporate
Annual Reports,” Canadian Journal of Administrative Sciences 25: 237–54.
Dimitri, C., E.C. Jaenicke and A.B. Effland (2009) “Why Did Contracts Supplant the Cash Market in
the Broiler Industry? An Economic Analysis Featuring Technological Innovation and Institutional
Response,” Journal of Agricultural and Food Industrial Organization 7: 1
Doege, J., M. Fehr, J. Hinz, H. Lüthi and M. Wilhelm (2009) “Risk Management in Power Markets:
The Hedging Value of Production Flexibility,” European Journal of Operational Research 199:
936–43.
Doganis, R. (2002) Flying Off Course: The Economics of International Airlines (3rd edn). London: Routledge.
Dominguez, K. and L. Tesar (2006) “Exchange Rate Exposure,” Journal of International Economics 68:
188–218.
Dowd, K. (1998) Beyond Value at Risk, The New Science of Risk Management. New York: Wiley.
Dufey, G. and S. Srinivasulu (1983) “The Case for Corporate Management of Foreign Exchange Risk,”
Financial Management 12(4): 54–62.
Duffie, D. and J. Pan (1997) “An Overview of Value at Risk,” Journal of Derivatives 5: 7–49.
Dunbar, N. (2000) Inventing Money, The Story of Long-Term Capital Management and the Legends Behind
It. New York: Wiley.
Dutt, H., J. Fenton, J. Smith and G. Wang (1997) “Crop Year Influences and Variability of the Agricultural
Futures Spreads,” Journal of Futures Markets 17: 341–67.
Dwyer, P. (1996) “Descent into the Abyss: How the Copper Trading Affair Engulfed Sumitomo,” Business
Week (July 1): 28–9.
Ederington, L. (1979) “The Hedging Performance of the New Futures Markets,” Journal of Finance 34:
157–70.
Edwards, F. (1995) “Derivatives can be Hazardous to Your Health: The Case of Metallgesellschaft,”
Derivatives Quarterly 1 (Spring): 8–17.
Edwards, F. (1999) “Hedge Funds and the Collapse of Long Term Capital Management,” Journal of
Economic Perspectives 13: 189–210.
Edwards, F. (2006) “Hedge Funds and Investor Protection Regulation,” Economic Review Federal Reserve
Bank of Atlanta, Quarter 4: 35–48.
Edwards, F. and C. Ma (1988) “Commodity Pool Performance: Is the Information Contained in Pool
Prospectuses Useful?” Journal of Futures Markets 8: 589–616.
Edwards, F. and J.M. Park (1996) “Do Managed Futures Make Good Investments?” Journal of Futures
Markets 16: 475–517.
Ehrenberg, R. (1928) Capital and Finance in the Age of the Renaissance (trans. from the German by H.M.
Lucas). London: Jonathan Cape.
Einzig, P. (1970) The History of Foreign Exchange. London: Macmillan.
Elton, E. and M. Gruber (1984) Modern Portfolio Theory and Investment Analysis (2nd edn). New York:
Wiley.
Elton, E. and M. Gruber (1995) Modern Portfolio Theory and Investment Analysis (5th edn). New York:
Wiley.
Elton, E., M. Gruber and J. Rentzler (1987) “Professionally Managed. Publicly Traded Commodity Funds,”
Journal of Business 60: 175–200.
Emery, H. (1896) Speculation on the Stock and Produce Exchanges of the United States. New York: Columbia
University Press; reprinted by AMS Press, New York (1968).
Emm, E.E., G.D. Gay and L. Chen-Miao (2007) “Choices and Best Practice in Corporate Risk Management
Disclosure,” Journal of Applied Corporate Finance 19: 82–93.
Engle, R.F. and S. Manganelli (2004) “CAVar: Conditional Autoregressive Value at Risk by Regression
Quantiles,” Journal of Business & Economic Statistics 22: 367–81.
Erb, C. and C. Harvey (2006) “The Tactical and Strategic Value of Commodity Futures,” Financial
Analysts Journal 62: 69–97.
Evans-Pritchard, E. (1940) The Nuer: A Description of the Modes of Livelihood and Political Institution of a
Nilotic People. Oxford, UK: Clarendon Press.
Falloon, W. (1998) Market Maker: A Sesquicentennial Look at the Chicago Board of Trade. Chicago: CBT.
Fama, E. and K. French (1987) “Commodity Futures Prices: Some Evidence on Forecast Power, Premiums
and the Theory of Storage,” Journal of Business 60: 55–74.
Fama, E. and K. French (1988) “Business Cycles and the Behavior of Metal Prices,” Journal of Finance 43:
1075–93.
Faruqee, R., J.R. Coleman and T. Scott (1997) “Managing Price Risk in the Pakistan Wheat Market,”
World Bank Economic Review 11: 263–92.
Fay, S. (1982) Beyond Greed. New York: The Viking Press.
Feder, G. et al. (1980) “Futures Markets and the Theory of the Firm Under Price Uncertainty,” Quarterly
Journal of Economics 94: 317–28.
Fennell, T. (1996) “Denting Copper: How Far Will Japan’s Massive Trading Scandal Spread?” MacLeans
(July 1): 28–9.
Fioretti, G. (2009) “Evidence Theory as a Procedure for Handling Novel Events,” Metroeconomica 60:
283–301.
Fisher, B. and A. Kumar (2010) “The Right Way to Hedge,” McKinsey Quarterly (4): 97–100.
Flouris, T. and T. Walker (2005) “The Financial Performance of Low-Cost and Full-Service Airlines
in Times of Crisis,” Revue Canadienne des Sciences de l’Administration/Canadian Journal of
Administrative Sciences 22: 3–20.
Fong Chan, K. and P. Gray (2006) “Using Extreme Value Theory to Measure Value-At-Risk for Daily
Electricity Spot Prices,” International Journal of Forecasting 22: 283–300.
Food and Agriculture Organization (FAO) together with the IFAD, IMF, OECD, UNCTAD, WFP, the
World Bank, the WTO, IFPRI and the UN HLTF (2011) Price Volatility in Food and Agricultural
Markets: Policy Responses (June 2), United Nations.
Ford, J.L. (1993) “G.L.S. Shackle (1903–1992): A Life with Uncertainty,” Economic Journal 103: 683–97.
Forrester, R. (1931) “Commodity Exchanges in England,” Annals of the American Academy of Political and
Social Sciences 155: 196–207.
Foster, K. (2003) “The Death of Hedging?” American Metal Market 111(51), (Dec. 22): 8.
Francis, J. (1850) Chronicles and Characters of the Stock Exchange (1st US edn). Boston: Crosby and
Nichols.
Frechette, D. (2000) “The Demand for Hedging and the Value of Hedging Opportunities,” American
Journal of Agricultural Economics 82: 897–907.
Frésard, L., C. Pérignon and A. Wilhelmsson (2011) “The Pernicious Effects of Contaminated Data in Risk
Management,” Journal of Banking & Finance 35: 2569–83.
Fridson, M. (ed.) (1996) Extraordinary Popular Delusions and the Madness of Crowds; and, Confusion de
Confusiones (reprints of classic texts). New York: Wiley.
Friedman, M. (1953) “The Methodology of Positive Economics,” in M. Friedman, Essays in Positive
Economics. Chicago: University of Chicago Press.
Froot, K., D. Scharfstein and J. Stein (1993) “Risk Management: Coordinating Corporate Investment and
Financing Policies,” Journal of Finance 48: 1629–58.
Froot, K., D. Scharfstein and. J. Stein (1994) “A Framework for Risk Management,” Journal of Applied
Corporate Finance 7: 22–32.
Fu Jing (2005) “Copper Trader Racks up Loss on Massive Short Position,” China Daily (Beijing) (Nov. 17).
Furukawa, T. (1997) “Hamanaka’s Ex-Boss has Day in Court,” American Metal Market (June 9).
Furukawa, T. (1997) “Hamanaka Testimony: He Followed His Leader,” American Metal Market (July 10).
Furukawa, T. (1995) “Hamanaka Sees Copper’s Strength Easing in Late ’96,” American Metal Market
(November 7).
Furukawa, T. (1996) “Sumitomo Calls Hamanaka’s Actions Criminal, Copper Trading Losses Increase to
$2.6 Billion,” American Metal Market (September 20).
Füss, R., Z. Adams and D. Kaiser (2010) “The Predictive Power of Value-At-Risk Models in Commodity
Futures Markets,” Journal of Asset Management 11: 261–85.
Gadamer, H.-G. (1960) Truth and Method. New York: Seabury Press (1975 translation of the second
German edition of 1965).
Gao, A. and G. Wang (2005) “Asymmetric Volatility of Basis and the Theory of Storage,” Journal of Futures
Markets 25: 399–418.
Garber, P. (1989) “Tulipmania,” Journal of Political Economy 97: 535–60.
Garber, P. (1990) “Famous First Bubbles,” Journal of Economic Perspectives 4 (Spring): 35–54.
Gardner, B. (1989) “Rollover Hedging and Missing Long-Term Futures Markets,” American Journal of
Agricultural Economics 71: 311–18.
Gardner, L.A. and J.T. Schmit (1996) “College and University Programs in Risk Management,” Financial
Practice & Education 6: 68–77.
Gay, G. and J. Nam (1998) “The Underinvestment Problem and Corporate Derivatives Use,” Financial
Management 27 (Winter): 53–69.
Géczy, C., B. Minton and C. Shrand (1997) “Why Firms Use Currency Derivatives,” Journal of Finance
52: 1323–54.
Géczy, C.C., B.A. Minton and C. Schrand (2006) “The Use Of Multiple Risk Management Strategies:
Evidence From the Natural Gas Industry,” Journal of Risk 8: 19–54.
Gelderblom, O. and J. Jonker (2005) “Amsterdam as the Cradle of Modern Futures and Options Trading,
1550–1630,” Ch.11 in W. Goetzmann and K. Rouwenhorst (eds) The Origins of Value. Oxford, UK:
Oxford University Press.
Geman, H. and S. Ohana (2008) “Time-Consistency in Managing a Commodity Portfolio: A Dynamic
Risk Measure Approach,” Journal of Banking & Finance 32: 1991–2005.
Gemech, F. and J. Struthers (2007) “Coffee Price Volatility in Ethiopia: Effects of Market Reform
Programmes,” Journal of International Development 19: 1131–42.
Genton, M. (ed.) (2004) Skew-Elliptical Distributions and Their Applications: A Journey Beyond Normality.
London: Chapman and Hall.
Gerardi, K. and A. Shapiro (2009) “Does Competition Reduce Price Dispersion? New Evidence from the
Airline Industry,” Journal of Political Economy 117: 1–37.
Gibson, R. and E.S. Schwartz (1990) “Stochastic Convenience Yield and the Pricing of Oil Contingent
Claims,” Journal of Finance 45: 959–76.
Ginder, M., A. Spaulding, K. Tudor and R. Winter (2009) “Factors Affecting Crop Insurance Purchase
Decisions by Farmers in Northern Illinois,” Agricultural Finance Review 69: 113–25.
Giot, P. (2003) “The Information Content of Implied Volatility in Agricultural Commodity Markets,”
Journal of Futures Markets 23: 441–54.
Giot, P. and Laurent, S. (2003) “Market Risk in Commodity Markets: A VaR Approach,” Energy Economics
25: 435–57.
Glasser, J. and F. Barbash (1996) “Sumitomo News Hits Copper: Price Tumbles; Criminal Investigation
Under Way,” Montreal Gazette (June 15): G1.
Godbey, J. and J. Hilliard (2007) “Adjusting Stacked-Hedge Ratios for Stochastic Convenience Yield: A
Minimum Variance Approach,” Quantitative Finance 7: 289–300.
Godfrey, S. and R. Espinosa (1998) “Value-at-Risk and Corporate Valuation,” Journal of Applied Corporate
Finance 11: 108–15.
Goesch, T., A. Hafi, S. Thorpe, P. Gooday and O. Sanders (2009) “Climate Change, Irrigation and Risk
Management,” Australian Commodities 16: 3–20.
Goldstein, D.G. and N.N. Taleb (2007) “We Don’t Quite Know What We Are Talking About,” Journal of
Portfolio Management 33: 84–6.
Gooding, K. (1996) “LME Waits for the Sumitomo Aftershocks,” Financial Post (June 15): 9.
Goodwin, B.K. and T. Schroeder (1994) “Human Capital, Producer Education Programs, and the
Adoption of Forward-Pricing Methods,” American Journal of Agricultural Economics 76: 936–47.
Grace, M. and M. Rebello (1993) “Financing and the Demand for Corporate Insurance,” Geneva Papers on
Risk and Insurance Theory 18: 147–72.
Graham, B. (1949) The Intelligent Investor: A Book of Practical Counsel. New York: Harper and Row (4th
edn, 1973).
Graham, B. and D. Dodd (1934) Security Analysis. New York: McGraw-Hill.
Graham, B., D. Dodd and S. Cottle (1962) Security Analysis. New York: McGraw-Hill.
Graham, J. and D. Rogers (2002) “Do Firms Hedge in Response to Tax Incentives?” Journal of Finance 57:
815–39.
Graham, J. and C. Smith (1999) “Tax Incentives to Hedge,” Journal of Finance 54: 2241–62.
Gray, M. (2002) “New Technique Defines the Limits of Upgrading Heavy Oils, Bitumens,” Oil and Gas
Journal 7: 50–4.
Gray, R. (1981) “Economic Evidence in Manipulation Cases,” Research on Speculation Seminar Report,
Chicago Board of Trade: 108–14.
Gray, R. and A. Peck (1981) “The Chicago Wheat Futures Market: Recent Problems in Historical
Perspective,” Food Research Institute Studies 18: 89–115.
Greene, N., A. Aziz and G. Liersaph (2007) “Hedge Fund Organizational Decisions and How They Affect
the Sponsor’s Compliance Needs,” Journal of Securities Compliance 1: 247–59.
Greenspan, A. (1996) “Remarks at the Federation of Bankers Associations of Japan,” Tokyo, Japan.
Gritta, R., B. Adams and B. Adrangi (2006) “Operating, Financial and Total Leverage and the Effects on
US Air Carrier Returns, 1990–2003,” Journal of the Transportation Research Forum 45: 57–68.
Grootveld, H. and W. Hallerbach (1999) “Variance vs Downside Risk: Is There Really That Much
Difference?” European Journal of Operational Research 114: 304–19.
Guay, W. and S.P. Kothari (2003) “How Much Do Firms Hedge With Derivatives?” Journal of Financial
Economics 70: 423.
Guinvarc’h, M.V., J. Janssen and J.E. Cordier (2004) “Agricultural Finance Revenue Futures Contract,”
International Journal of Theoretical & Applied Finance 7: 85–99.
Haigh, M. and M. Holt (2002) “Crack Spread Hedging: Accounting for Time Varying Volatility Spillovers
in the Energy Futures Markets,” Journal of Applied Econometrics 17: 269–89.
Haigh, M. and M. Holt (2000) “Hedging Multiple Price Uncertainty in International Grain Trade,”
American Journal of Agricultural Economics 82: 881–89.
Haines, L. (2001) “Alberta’s Heavy Oils,” Oil & Gas Investor 21 (October): 30–41.
Hamidieh, K. and K. Ensor (2010) “A Simple Method for Time Scaling Value-At-Risk: Let the Data Speak
For Themselves,” Journal of Risk Management in Financial Institutions 3: 380–91.
Hampton, J. (2009) Fundamentals of Enterprise Risk Management: How Top Companies Assess Risk,
Manage Exposures, and Seize Opportunities. New York: American Management Association.
Harrington, S. and G. Niehaus (2003) “United Grain Growers: Enterprise Risk Management and Weather
Risk,” Risk Management & Insurance Review 6: 193–208.
Hart, C., B. Babcock and D. Hayes (2001) “Livestock Revenue Insurance,” Journal of Futures Markets
21: 28.
Harwood, J., R. Heifner, K. Coble, J. Perry and A. Somwaru (1999) “Managing Risk in Farming: Concepts,
Research, and Analysis,” Economic Research Service Agricultural Economic Report 774, Washington,
DC: US Department of Agriculture.
Hassett, M., S. Sears and G. Trennepohl (1985) “Asset Preference, Skewness and the Measurement of
Expected Utility,” Journal of Economics and Business 37: 35–47.
Haushalter, G. (2000) “Financing Policy, Basis Risk and Corporate Hedging: Evidence from Oil and Gas
Producers,” Journal of Finance 55: 107–52.
Haushalter, G., S. Klasa and W.F. Maxwell (2007) “The Influence of Product Market Dynamics on a Firm’s
Cash Holdings and Hedging Behavior,” Journal of Financial Economics 84: 797–825.
Hayek, F. (1955) The Counter-Revolution of Science. London: Collier-Macmillan.
Hayenga, M.L. and D.D. DiPietre (1982) “Cross-Hedging Wholesale Pork Products Using Live Hog
Futures,” American Journal of Agricultural Economics 64: 474–751.
Heaney, J. and G. Poitras (1991) “Estimation of the Optimal Futures Hedge, Expected Utility and Ordinary
Least Squares Regression,” Journal of Futures of Markets 11: 603–12.
Heaney, R. (2002) “Approximation for Convenience Yield in Commodity Futures Pricing,” Journal of
Futures Markets 22: 1005–17.
Heifner, R.G. (1972) “Optimal Hedging Levels and Hedging Effectiveness in Cattle Feeding,” Agricultural
Economic Research 24: 25–36.
Heinkel, R., M. Howe and J. Hughes (1990) “Commodity Convenience Yields as an Option Profit,” Journal
of Futures Markets 10: 519–33.
Henderson, J. and N. Fitzgerald (2008) “Can Grain Elevators Survive Record Crop Prices?” Main Street
Economist (Issue III), Federal Reserve Bank of Kansas City.
Henderson, J. and R. Quandt (1980) Microeconomic Theory: A Mathematical Approach (3rd edn). New
York: McGraw-Hill.
Hendricks, D. and B. Hirtle (1997) “Bank Capital Requirements for Market Risks: The Internal Models
Approach,” Federal Reserve Bank of New York Economic Policy Review (December): 1–12.
Herbst, A., D. Kare and S. Caples (1989) “Hedging Effectiveness and Minimum Risk Hedge Ratios in the
Presence of Autocorrelation,” Journal of Futures Markets 9: 185–97.
Hicks, J.D. (1961) The Populist Revolt: A History of the Farmers’ Alliance and the People’s Party. Lincoln,
NB: University of Nebraska Press.
Hieronymous, T. (1977) The Economics of Futures Trading (2nd edn). New York: Commodity Research
Bureau.
Hill, J. and T. Schneeweis, (1982) “The Hedging Effectiveness of Foreign Currency Futures,” Journal of
Financial Research 5: 95–104.
Hilliard, J. and J. Reis (1998) “Valuation of Commodity Futures and Options Under Stochastic
Convenience Yields, Interest Rates and Jump Diffusions in the Spot,” Journal of Financial and
Quantitative Analysis 33: 61–86.
Hofer, C. and C. Eroglu (2010) “Investigating the Effects of Economies of Scope on Firms’ Pricing Behavior:
Empirical Evidence from the US Domestic Airline,” Transportation Research: Part E: Logistics and
Transportation Review 46: 109–19.
Horowitz, I. (1998) “Assume a Can Opener,” Decision Sciences 29: 517–20.
Houghton, J. (1692–1703) A Collection for Improvement of Husbandry and Trade. London: Taylor,
Hindmarsh, Clavell, Rogers and Brown; reprinted by Gregg International Publishers (1969).
Hsieh, D. (1988) “Comments on ‘Hedging Canadian Treasury Bills with US Money Market Futures’,”
Review of Futures Markets: 192–5.
Hull, J. (2000) Options, Futures and Other Derivatives (4th edn). Prentice-Hall.
Hull, J. and A. White (1998) “Value at Risk when Daily Changes in Market Variables are not Normally
Distributed,” Journal of Derivatives 6: 9–18.
Hurt, H. (1981) Texas Rich. New York: Norton.
Ingersoll, J. (1987) The Theory of Financial Decision Making. Totawa, NJ: Rowan and Littlefield.
Interagency Task Force on Commodity Markets (ITFCM) (2008) “Interim Report on Crude Oil,” (July)
Washington, DC.
International Council on Metals and the Environment (ICME) (2001) Risk Assessment and Risk
Management of Non-Ferrous Metals Realizing the Benefits and Managing the Risks. Ottawa,
Canada.
Irwin, S. and W. Brorsen (1985) “Public Futures Funds,” Journal of Futures Markets 5: 149–72.
Irwin, S., T. Krukemyer and C. Zulauf (1993) “Investment Performance of Public Commodity Pools:
1979–1990,” Journal of Futures Markets 13: 799–820.
Jacks, D. (2007) “Populists Versus Theorists: Futures Markets and the Volatility of Prices,” Explorations
in Economic History 44: 342–62.
Jalan, J. and M. Ravallion (1999) “Are the Poor Less Well Insured? Evidence on Vulnerability to Income
Risk in Rural China,” Journal of Development Economics 58: 21.
Jenkins, H. (1996) “Business World: Never Mind the $1.8 Billion, London’s the Place to Be,” Wall Street
Journal (July 23): A23.
Jin, H.J. (2007) “Heavy-Tailed Behavior of Commodity Price Distribution and Optimal Hedging Demand,”
Journal of Risk & Insurance 74: 863–81.
Jin, Y. and P. Jorion (2004) “Firm Value and Hedging: Evidence from US Oil and Gas Producers,” Journal
of Finance 61: 893–919.
Johnson, L. (1960) “The Theory of Hedging and Speculation in Commodity Futures,” Review of Economic
Studies 27: 139–51.
Johnson, P. (1981) “Commodity Market Manipulation,” reprinted in R. Gray (ed.) Research on Speculation.
Chicago: Chicago Board of Trade.
Jorion, P. (1995) Big Bets Gone Bad, Derivatives and Bankruptcy in Orange County. New York: Academic
Press.
Jorion, P. (1997) Value at Risk (2nd edn). New York: McGraw-Hill.
Jorion, P. (2006) Value at Risk (3rd edn). New York: McGraw-Hill.
JP Morgan (1996) Riskmetrics Technical Model. New York: JP Morgan.
Ju, X. and N. Pearson (1999) “Using Value-at-Risk to Control Risk Taking: How Wrong Can You Be?”
Journal of Risk 1: 5–36.
Kairys, J. and N. Valerio (1997) “The Market for Equity Options in the 1870s,” Journal of Finance 52:
1707–23.
Kaldor, N. (1939) “Speculation and Economic Stability,” Review of Economic Studies 7: 1–27.
Kamara, A. and A. Siegel (1987) “Optimal Hedging in Futures Markets with Multiple Delivery
Specifications,” Journal of Finance 42: 1007–21.
Kat, H.M. and R.C.A. Oomen (2007) “What Every Investor Should Know About Commodities Part II:
Multivariate Return Analysis,” Journal of Investment Management 5: 16–40.
Katchova, A. and M. Miranda (2004) “Two-Step Econometric Estimation of Farm Characteristics
Affecting Marketing Contracts Decisions,” American Journal of Agricultural Economics 86:
88–102.
Kellenbenz, H. (1957) “Introduction” to de la Vega, Confusion de Confusiones; reprinted in Fridson
(1996) Extraordinary Popular Delusions and the Madness of Crowds; and, Confusion de Confusiones
(reprints of classic texts). New York: Wiley.
Key, N. (2004) “Agricultural Contracting and the Scale of Production,” Agricultural and Resource
Economics Review 33: 255–71.
Keynes, J. (1921) A Treatise on Probability. London: Macmillan (1963 reprint).
Keynes, J. (1936) The General Theory of Employment, Interest and Money. New York: Harcourt (Harbinger
edn, 1964).
Kharouf, J. (1996) “The Copper Trader Who Fell From Grace,” Futures 25(10): 66–9.
Klir, G.J. (2006) Uncertainty and Information: Foundations of Generalized Information Theory. Hoboken,
NJ: John Wiley & Sons.
Knight, F. (1921) Risk, Uncertainty and Profit. New York: Houghton Mifflin.
Knight, T. and Coble, K. (1997) “A Survey of Literature on US Multiple Peril Crop Insurance Since 1980,”
Review of Agricultural Economics 19: 128–56.
Kraus, A. and R. Litzenberger (1976) “Skewness Preference and the Valuation of Risk Assets,” Journal of
Finance 31: 1085–100.
Kroll, Y., H. Levy and H. Markowitz (1984) “Mean-Variance versus Direct Utility Maximization,” Journal
of Finance 39: 47–61.
Kroner, K. and J. Sultan (1993) “Time Varying Distributions and Dynamic Hedging with Foreign
Currency Futures,” Journal of Financial and Quantitative Analysis 28: 535–52.
Kuester, K., S. Mittnik and M. Paollella (2006) “Value-at-Risk Prediction: A Comparison of Alternative
Strategies,” Journal of Financial Econometrics 4: 53–89.
Kuprianov, A. (1995) “Derivatives Debacles,” Federal Reserve Bank of Richmond, Economic Quarterly
(Fall): 1–38.
Larson, D., P. Varangis and N. Yabuki (1998) “Commodity Risk Management and Development,”
Development Research Group, World Bank Policy Working Paper #1963 (August) Washington, DC.
Larson, D.F., J.R. Anderson and P. Varangis (2004) “Policies on Managing Risk in Agricultural Markets,”
World Bank Research Observer 19: 199–230.
Lawson, T. (1997) Economics and Reality. London: Routledge.
Layard, R. and A. Walters (1978) Microeconomic Theory. New York: McGraw-Hill.
Lechner, L. and Ovaert, T. (2010) “Techniques to Account for Leptokurtosis and Asymmetric Behavior in
Returns Distributions,” Journal of Risk Finance 11: 464–80.
Lee, H.-T. and J. Yorder (2007) “A Bivariate Markov Regime Switching GARCH Approach to Estimate
Time Varying Minimum Variance Hedge Ratios,” Applied Economics 39: 1253–65.
Lee, R. (1998) What is an Exchange? Oxford, UK: Oxford University Press.
Leitch, G. and J. Tanner (1991) “Economic Forecast Evaluation: Profits Versus the Conventional Error
Measures,” American Economic Review 81: 580–61.
Leuthold, R. and P. Peterson (1983) “The Cash-Futures Price Spread for Live Hogs,” North Central Journal
of Agricultural Economics 5: 25–9.
Leuthold, R., J. Junkus and J. Cordier (1989) The Theory and Practice of Futures Markets. Lexington,
Mass.: Lexington Books.
Levi, M. and P. Sercu (1991) “Erroneous and Valid Reasons for Hedging Foreign Exchange Rate Exposure,”
Journal of Multinational Financial Management 1: 25–37.
Levy, H. and H. Markowitz (1979) “Approximating Expected Utility by a Function of Mean and Variance,”
American Economic Review 69 (June): 308–17.
Lewin, C. (2003) Pensions and Insurance Before 1880: A Social History. East Lothian, Scotland: Tuckwell
Press.
Lewis, N. (2009) “Is There a Role for Commodities in Long-Term Wealth Accumulation?” Journal of
Wealth Management 12: 130–7.
Lien, D. (1989a) “Cash Settlement Provisions on Futures Contracts,” Journal of Futures Markets 9: 263–70.
Lien, D. (2005) “A Note on the Superiority of the OLS Hedge Ratio,” Journal of Futures Markets 25: 1121–6.
Lien, D. (2009) “Timing the Value-at-Risk Hedge,” Research in Finance 25: 333–41.
Lien, D. (2009a) “A Note on the Hedging Effectiveness of GARCH Models,” International Review of
Economics and Finance 18: 110–12.
Lien, D. (2012) “A Note on Utility-Based Futures Hedging Performance Measure,” Journal of Futures
Markets 32: 92–7.
Lien, D. and Y.K. Tse (2002) “Some Recent Developments in Futures Hedging,” Journal of Economic
Surveys 16: 357–96.
Lien, D. and L. Yang (2008) “Asymmetric Effect of Basis on Dynamic Futures Hedging: Empirical
Evidence from Commodity Markets,” Journal of Banking & Finance 32: 187–98.
Ligon, E., J. Thomas and T. Worrall (2002) “Informal Insurance with Limited Commitment: Theory and
Evidence from Village Economies,” Review of Economic Studies 69: 36.
Lim, K.-G. (1989) “A New Test of the Three Moment Capital Asset Pricing Model,” Journal of Financial
and Quantitative Analysis 24(2): 205–16.
Lim, L.-M. (2005) “From Highflier to Scapegoat: Copper Trader Scandal Shows How Risk Controls Have
Not Kept Pace With Boom,” The Standard—China’s Business Newspaper (Hong Kong) (Dec. 1).
Little, P.D., K. Smith, B.A. Cellarius, D.L. Coppock and C.B. Barrett (2001) “Avoiding Disaster:
Diversification and Risk Management Among East African Herders,” Development and Change 32:
33.
Loistl, O. (1976) “The Erroneous Approximation of Expected Utility by Means of a Taylor’s Series
Expansion: Analytic and Computational Results,” American Economic Review 66(5): 904–10.
Longstaff, F. (1995) “How Much Can Marketability Affect Security Values?” Journal of Finance 50: 1767–74.
Luehrman, T. (1990) “The Exchange Rate Exposure of a Global Competitor,” Journal of International
Business Studies 21: 225–42.
Lurie, J. (1972) “Private Associations, Internal Regulation and Progressivism: The Chicago Board of
Trade, 1880–1923, as a Case Study,” American Journal of Legal History 16: 215–38.
MacDonald, J. and P. Korb (2008) Agricultural Contracting Update: Contracts in 2005/EIB-35, Economic
Research Service, USDA.
MacDonald, J. and P. Korb (2011) Agricultural Contracting Update: Contracts in 2008/EIB-72, Economic
Research Service, USDA.
Mackay, C. (1852) Extraordinary Popular Delusions and the Madness of Crowds (2nd edn), reprinted by
New York: Bonanza Books (1980); first edition (1841).
Mackay, P. and S. Moeller (2007) “The Value of Corporate Risk Management,” Journal of Finance 62:
1379–419.
Mackay, P. and S. Moeller (2010) “Corporate Risk Management: The Hedging Footprint,” SSRN Working
Paper.
MacMinn, R. (1987) “Insurance and Corporate Risk Management,” Journal of Risk and Insurance 54:
658–77.
Mahul, O. (1999) “The Design of an Optimal Area Yield Crop Insurance Contract,” Geneva Papers on Risk
and Insurance Theory 24: 13.
Mahul, O. and B. Wright (2003) “Designing Optimal Crop Revenue Insurance,” American Journal of
Agricultural Economics 85: 580–9.
Malighetti, P., M. Meoli, S. Paleari and R. Redondi (2011) “Value Determinants in the Aviation Industry,”
Transportation Research: Part E: Logistics and Transportation Review 47: 359–70.
Malkiel, B. (1985) A Random Walk Down Wall Street (4th edn). New York: Norton.
Malliaris, A. and W. Brock (1982) Stochastic Methods in Economics and Finance. Amsterdam: North
Holland.
Malynes, G. (1622) Consuetudo, vel Lex Mercatoria or the Ancient Law Merchant Divided into Three Parts
according to the Essentiall Parts of Trafficke (1st edn) London: A. Islip; reprinted by Norwood NJ:
Walter Johnson, Inc. (1979).
Manfredo, M. and R. Leuthold (1999) “Value-at-Risk Analysis: A Review and the Potential for Agricultural
Applications,” Review of Agricultural Economics 21: 99–111.
Mann, S. (2008) “Too Far Over the Hedge: Why the SEC’s Attempt To Further Regulate Hedge Funds
Had to Fail and What, if any, Alternative Solutions Should be Considered,” St John’s Law Review
82: 315–57.
Manzur, M., A. Hoque and G. Poitras (2010) “Currency Option Pricing and Realized Volatility,” Banking
and Finance Review 2: 73–86.
Mark, D., B. Brorsen, K. Anderson and R. Small (2008) “Price Risk Management Alternatives for Farmers
in the Absence of Forward Contracts with Grain Merchants,” Choices 23 (2nd Quarter): 22–5.
Markham, J. (1987) The History of Commodity Futures Trading and its Regulation. New York: Praeger.
Markham, J. (2002) A Financial History of the United States. Armonk, NY: M.E. Sharpe.
Markham, J. and D. Harty (2008) “For Whom the Bell Tolls: The Demise of Exchange Trading Floors and
the Growth of ECNs,” Journal of Corporation Law 33: 865–939.
Markowitz, H. (1999) “The Early History of Portfolio Theory,” Financial Analysts Journal (July/August):
5–16.
Matz, A. (1978) “Maximum Likelihood Parameter Estimation for the Quartic Exponential Distribution,”
Technometrics 20: 475–84.
Mayers, D. and C. Smith (1982) “On the Corporate Demand for Insurance,” Journal of Business 55: 281–96.
McCormack, J., R. LeBlanc and C. Heiser (2003) “Turning Risk into Shareholder Wealth in the Petroleum
Industry,” Journal of Applied Corporate Finance 15: 67.
McDermott, E. (1979) “Defining Manipulation in Commodity Futures Trading: The Futures ‘Squeeze’,”
Northwestern University Law Review 74: 202–25.
McGee, S. and S. Frank (1996) “Manipulation may be Hard to Prove in Sumitomo Case,” Wall Street
Journal (July 15): C1.
McGee, S., S. Frank and N. Shirouzu (1996) “Japanese Authorities Arrest Key Figure in Sumitomo Copper-
Trading Debacle,” Wall Street Journal (October 23): C1.
Mello, A. and J. Parsons (1995) “Hedging a Flow of Commodity Deliveries with Futures: Problems with a
Rolling Stack Hedge,” Derivatives Quarterly 1: 16–19.
Menachof, D.A. and G.N. Dicer (2001) “Risk Management Methods for the Liner Shipping Industry: The
Case of the Bunker Adjustment Factor,” Maritime Policy & Management 28: 141–55.
Merton, R. (1993) “Operation and Regulation in Financial Intermediation: A Functional Perspective,”
in Peter Englund (ed.) Operation and Regulation of Financial Markets. Stockholm: Ekonomiska
Radet.
Mian, S. (1996) “Evidence on Corporate Hedging Policy,” Journal of Financial and Quantitative Analysis
31: 419–39.
Mikes, A. (2008) “Chief Risk Officers At Crunch Time: Compliance Champions or Business Partners?”
Journal of Risk Management in Financial Institutions 2: 7–25.
Milonas, N. and S. Thomadakis (1997) “Convenience Yield and the Option to Liquidate for Commodities
With a Crop Cycle,” European Review of Agricultural Economics 24: 267–83.
Miranda, M.J. and J.W. Glauber (1997) “Systemic Risk, Reinsurance, and the Failure of Crop Insurance
Markets,” American Journal of Agricultural Economics 79: 10.
Mishra, A.K. and El-Osta, H.S. (2002) “Managing Risk in Agriculture through Hedging and Crop
Insurance: What Does a National Survey Reveal?” Agricultural Finance Review 62: 135–48.
Moeller, R. (2007) COSO Enterprise Risk Management: Understanding the New Integrated ERM
Framework. Hoboken, NJ: Wiley.
Moffett, S. (1996) “Culture Shock: Sumitomo Debacle Shows Japan Management Flaw,” Far Eastern
Economic Review (June 27): 69–70.
Mohan, S. (2007) “Market-Based Price-Risk Management for Coffee Producers,” Development Policy
Review 25: 333–54.
Mohapatra, S., R. Goodhue, C. Carter and J. Chalfant (2010) “Effects of Forward Sales on Spot Markets:
Pre-Commitment Sales and Prices for Fresh Strawberries,” American Journal of Agricultural
Economics 92: 152–63.
Morrell, P. and W. Swan (2006) “Airline Jet Fuel Hedging: Theory and Practice,” Transport Reviews 26:
713–30.
Moschini, G. and D. Hennessy (2001) “Uncertainty, Risk Aversion and Risk Management for Agricultural
Producers,” Handbook of Agricultural Economics, vol. IA, Agricultural Production; 2001,
Amsterdam: Elsevier Science. Edited by B. Gardner and G. Rausser.
Moschini, G. and R. Myers (2002) “Testing for constant hedge ratios in commodity markets: a multivariate
GARCH approach,” Journal of Empirical Finance 9: 589–604.
Muller, A. and M. Grandi (2000) “Weather Derivatives: A Risk Management Tool for Weather Sensitive
Industries,” Geneva Papers on Risk and Insurance: Issues and Practice 25: 273–87.
Munro, J. (2000) “English ‘Backwardness’ and Financial Innovations in Commerce with the Low
Countries, 14th to 16th Centuries,” pp.105–67 in P. Stabel, B. Blondé and A. Greve (eds) International
Trade in the Low Countries (14th—16th Centuries). Garant: Leuven-Apeldoorn.
Murphy, A. (1997) John Law, Economic Theorist and Policy-Maker. Oxford: Clarendon Press.
Musser, W. and G.A. Patrick (2002) “How Much Does Risk Really Matter to Farmers?” Comprehensive
Assessment of the Role of Risk in US Agriculture. Norwell, Mass.: Kluwer. Edited by R.E. Just and R.
Pope.
Muth, J. (1961) “Rational Expectations and the Theory of Price Movements,” Econometrica 29: 315–35.
Myers, R. and J. Oehmke (1988) “Instability and Risk as Rationales for Farm Programs,” Agricultural
Stability and Farm Programs: Concepts, Evidence, and Implications, Boulder, Colo.: Westview.
Edited by D. Sumner.
Myers, R. and S. Thompson (1989) “Generalized Optimal Hedge Ratio Estimation,” American Journal of
Agricultural Economics 71: 858–68.
Nakamura, M., T. Nakashima and T. Niimura (2006) “Electricity Markets Volatility: Estimates,
Regularities And Risk Management Applications,” Energy Policy 34: 1736–49.
Nam, J., A. Tucker and J. Wei (2005) “Pricing Hedging with Local and Aggregate Quantity Risk,” Journal
of Derivatives 13: 49–69.
Ng, V. and S. Pirrong (1994) “Fundamentals and Volatility: Storage, Spreads and Dynamics of Metal
Prices,” Journal of Business 67: 203–30.
Niemeyer, G. (2008) “Statement of Garry Niemeyer, National Corn Growers Association, Agricultural
Markets Roundtable, Commodity Futures Trading Commission,” April 22.
Nomura, K. (2003) “Managing Risks in Airline Industry,” Japan and the World Economy 15: 469–79.
Norman, S. (2010) “How Well Does Non-Linear Mean Reversion Solve the PPP Puzzle?” Journal of
International Money and Finance 29: 919–37.
O’Brien, T. (1997) “Accounting versus Economic Exposure to Currency Risk,” Journal of Financial
Statement Analysis 2: 21–9.
Ocran, M.K. and N. Biekpe (2007) “Forecasting Volatility in Sub-Saharan Africa’s Commodity Markets,”
Investment Management and Financial Innovations 4: 91–102.
Officer, L. (1976) “The Purchasing-Power-Parity Theory of Exchange Rates,” IMF Staff Papers 23: 1–60.
Okochi, J. (2008) “Commodity Hedging: Best Practices For Modelling Commodity Exposures,” Journal of
Corporate Treasury Management 2: 42–6.
Ormiston, M. and J. Quiggin (1994) “Two-Parameter Decision Models and Rank-Dependent Expected
Utility,” Journal of Risk and Uncertainty 8: 273–82.
Overdahl, J.A. (1987) “The Use of Crude Oil Futures by the Governments of Oil Producing States,” Journal
of Futures Markets 7: 605–17.
Oxelheim, L. and C. Wihlborg (1997) Managing in the Turbulent World Economy: Corporate Performance
and Risk Exposure. New York: Wiley.
Pannell, D.J., G. Hailu, A. Weersink and A. Burt (2008) “More Reasons Why Farmers Have So Little
Interest in Futures Markets,” Agricultural Economics 39: 41–50.
Park, S. and S. Jei (2010) “Estimation and Hedging Effectiveness of Time Varying Hedge Ratio: Flexible
Bivariate GARCH Approaches,” Journal of Futures Markets 30: 71–99.
Parker, G. (1974) ‘The Emergence of Modern Finance in Europe 1500–1730’, in C. Cipolla (ed.) (1974,
Ch.7).
Parkinson, P.M. (2000) “Statement to the US House Subcommittee on Risk Management, Research, and
Specialty Crops of the Committee on Agriculture,” Federal Reserve Bulletin 86: 577.
Parsley, D. and H. Popper (2006) “Exchange Rate Pegs and Foreign Exchange Exposure in East and
Southeast Asia,” Journal of International Money and Finance 25: 992–1009.
Patnaik, I. and A. Shah (2010) “Does the Currency Regime Shape Unhedged Currency Exposure?” Journal
of International Money and Finance 29: 760–9.
Paulson, N.D., B.A. Babcock, C.E. Hart and D.J. Hayes (2008) “An Insurance Approach to Risk
Management in the Ethanol Industry,” Agricultural and Resource Economics Review 37: 51–62.
Pearson, T. and J. Pearson (2007) “Protesting Global Financial Market Stability and Integrity:
Strengthening SEC Regulation of Hedge Funds,” North Carolina Journal of International Law and
Commercial Regulation 33: 1–82.
Peck, A. (ed.) (1977) Selected Writings on Futures Markets. Chicago: CBT.
Peck, A. (2001) “The Development of Commodity Exchanges in the Former Soviet Union, Eastern Europe
and China,” Australian Economic Papers 40: 437–60.
Pirrong, C. (2010) “Energy Market Manipulation: Definition, Diagnosis and Deterrence,” Energy Law
Jounral 31: 1–20.
Pekarek, E. (2007) “Hogging the Hedge? ‘Bulldog’s’ 13F Theory May Not Be So Lucky,” Fordham Journal
of Corporate and Financial Law 12: 1079–81.
Penello, W. (2009) “Eliminating Uncertainty,” Risk Management 56: 38–41.
Pennings, J. and R. Leuthold (2000) “The Role of Farmers’ Behavioral Attitudes and Heterogeneity in
Futures Contracts Usage,” American Journal of Agricultural Economics 82: 908–19.
Pennings, J., O. Isengildina-Massa, S.H. Irwin, P. Garcia and D.L. Good (2008) “Producers’ Complex Risk
Management Choices,” Agribusiness 24: 31–54.
Pincus, M. and S. Rajgopal (2002) “The Interaction of Accrual Management and Hedging: Evidence from
Oil and Gas Firms,” Accounting Review 77: 127–60.
Pirrong, C. (2010) “Energy Market Manipulation: Definition, Diagnosis and Deterrence,” Energy – Law
Journal 31: 1–20.
Platteau, J.P. (1997) “Mutual Insurance as an Elusive Concept in Traditional Rural Communities,” Journal
of Development Studies 33: 764–96.
Poitras, G. (1988) “Hedging Canadian Treasury Bill Positions with US Money Market Futures,” Review of
Futures Markets 7: 176–91.
Poitras, G. (1989) “Optimal Futures Spread Positions,” Journal of Futures Markets 9: 397–412.
Poitras, G. (1993) “Hedging and Crop Insurance,” Journal of Futures Markets 13: 373–88.
Poitras, G. (1994) “Shareholder Wealth Maximization, Business Ethics and Social Responsibility,” Journal
of Business Ethics 13 (Feb.): 125–34.
Poitras, G. (2000) The Early History of Financial Economics, 1478–1776; From Commercial Arithmetic to
Life Annuities and Joint Stocks, Aldershot, UK: Edward Elgar.
Poitras, G. (2002) Risk Management, Speculation and Derivative Securities. New York: Academic Press.
Poitras, G. (2005) Security Analysis and Investment Strategy. Oxford, UK: Blackwell.
Poitras, G. (2008) “Risk,” in William Darity (ed.) International Encyclopedia of the Social Sciences (2nd
edn). Farmington Hills, MI: Macmillan Reference, vol. 7: 251–3.
Poitras, G. (2009) “From Antwerp to Chicago: The History of Exchange Traded Derivative Security
Contracts,” Revue d’Histoire des Sciences Humaines (Journal for the History of the Social Sciences)
20: 11–50.
Poitras, G. (2011) Valuation of Equity Securities: History, Theory and Application. Singapore: World
Scientific Publishing.
Poitras, G. and J. Heaney (1999) “Skewness Preference, Mean Variance and the Optimal Demand for Put
Options,” Managerial and Decision Economics 20: 327–42.
Posthumus, N. (1929) “The Tulipmania in Holland in the Years 1636 and 1637,” Journal of Economics and
Business History 1: 434–66.
Powers, M. and R. Vogel (1983) Inside the Financial Futures Markets. New York: Wiley.
Prakash, A., C. Chang, S. Hamid and M. Smyser (1996) “Why a Decision Maker May Prefer a Seemingly
Unfair Gamble,” Decision Sciences 27: 239–53.
President’s Working Group on Financial Markets (1999) Hedge Funds, Leverage and the Lessons of Long
Term Capital Management. Washington, DC: US Government Printing Office.
PricewaterhouseCoopers (2005) “Statement of Phase I Findings, China Aviation Oil” (Mar. 29).
Purnanandam, A. (2008) “Financial Distress and Corporate Risk Management: Theory and Evidence,”
Journal of Financial Economics 87: 706–39.
Rahgozar, R. and H. Najafi (2003) “Effect of Diversification on Managing Revenue and Risk: An Empirical
Analysis Using Crude Oil Futures Markets,” Derivatives Use, Trading & Regulation 9: 133, 17p.
Rao, V. (1999) “Fuel Price Risk Management Using Futures,” Journal of Air Transport Management 5:
39–44.
Rawls, S. and C. Smithson (1990) “Strategic Risk Management,” Journal of Applied Corporate Finance 2:
6–18.
Rebello, M. (1995) “Adverse Selection Costs and the Firm’s Financing and Insurance Decisions,” Journal
of Financial Intermediation 4: 21–47.
Rees, G. and J. Wiseman, J. (1969) “London’s Commodity Markets,” Lloyds Bank Review 91: 22–45.
Reynolds, R. (2005) “The Economics of Oil Definitions: The Case of Canada’s Oil Sands,” OPEC Review
29: 51–73.
Sherrick, B., P. Barry, G. Schnitkey, P. Ellinger and B. Wansink (2003) “Farmers’ Preferences for Crop
Insurance Attributes,” Review of Agricultural Economics 25: 415–29.
Shirouzu, N. (1998) “Former Sumitomo trader sentenced to eight years for fraud and forgery,” Wall Street
Journal (March 26): 1.
Siegel, D. and D. Siegel (1990) Futures Markets. New York: Dryden.
Simmons, P. (2002) “Why Do Farmers Have So Little Interest in Futures Markets?” Agricultural Economics
27: 1–6.
Skees, J. and A. Enkh-Amgalan (2002) “Examining the Feasibility of Livestock Insurance in Mongolia,”
World Bank Working Paper 2886, Washington, DC.
Skees, J., P. Varangis, D. Larson and P. Siegel (2002) “Can Financial Markets be Tapped to Help Poor
People Cope with Weather Risks?” Policy Research Working Paper 2812, Washington, DC: World
Bank.
Skees, J., P. Hazell and M. Miranda (1999) “New Approaches to Crop Yield Insurance in Developing
Countries,” Environment and Production Technology Division Discussion Paper 55, Washington,
DC: International Food Policy Research Institute.
Smith, C. (1995) “Corporate Risk Management: Theory and Practice,” Journal of Derivatives 2: 21–30.
Smith, C. and R. Stulz (1985) “The Determinants of Firms’ Hedging Policies,” Journal of Financial and
Quantitative Analysis 20: 391–405.
Smith, C., D. Wilford and C. Smithson (1994) Managing Financial Risk: A Guide to Derivative Products,
Financial Engineering, and Value Maximization. Irwin Professional.
Smithson, C. and B. Simkins (2005) “Does Risk Management Add Value? A Survey of the Evidence,”
Journal of Applied Corporate Finance 17: 8–17.
Smithson, C., C. Smith and D. Wilford (1995) Managing Financial Risk. Irwin.
Sorensen, C. (2002) “Modeling Seasonality in Agricultural Commodity Futures,” Journal of Futures
Markets 22: 393–426.
Srinivasan, P. (2009) “Price Discovery in NSE Spot and Futures Markets of Selected Oil and Gas Industries
in India: What Causes What?” IUP Journal of Financial Risk Management 6: 22–37.
Stein, J., S. Usher, D. La Gatutta and J. Youngen (2001) “A Comparables Approach to Measuring Cashflow-
at-Risk for Non-financial Firms,” Journal of Applied Corporate Finance 13: 100–9.
Stoll, H. and R. Whaley (2010) “Commodity Index Investing and Commodity Futures Prices,” Journal of
Applied Finance 20: 7–46.
Stulz, R. (2007) “Hedge Funds: Past, Present and Future,” Journal of Economic Perspectives 31: 175–94.
Tao, L. and M. Shahidelphour (2007) “Risk-Constrained Generation Asset Arbitrage in Power Systems,”
IEEE Transactions on Power Systems, 22: 1330–1339.
Tawney, R. (1925) “Introduction” to A Discourse Upon Usury by T. Wilson (1572); reprinted London:
Frank Cass (1962).
Taylor, C. (1917) History of the Board of Trade of the City of Chicago. Chicago: Robert O. Law.
Taylor, K. (1996) “Non-Stick Metal: London Metal Exchange Survives Sumitomo Crisis,” Far Eastern
Economic Review (October 31): 44–53.
Tchernitser, A. and D.H. Rubisov (2009) “Robust Estimation of Historical Volatility and Correlations in
Risk Management,” Quantitative Finance 9: 43–54.
Thompson, S. (1986) “Returns of Storage in Coffee and Cocoa Future Markets,” Journal of Futures Markets
6: 541–64.
Thraen, C.S. (1999) “A Note: The CSCE Cheddar Cheese Cash and Futures Price Long-Term Equilibrium
Relationship Revisited,” Journal of Futures Markets 19: 233–44.
Till, H. (2006) “EDHEC Comments on the Amaranth Case: Early Lessons from the Debacle,” EDHEC
Risk and Asset Management Research Centre.
Till, H. (2008) “Amaranth Lessons Thus Far,” Journal of Alternative Investments 10: 82–98.
Till, H. and J. Eagleeye (2005) “Commodities: Active Strategies for Enhanced Return,” Journal of Wealth
Management 8: 42–61.
Toevs, A. and D. Jacob (1986) “Futures and Alternative Hedge Ratio Methodologies,” Journal of Portfolio
Management 12: 60–70.
Tomek, W. and H. Peterson (2001) “Risk Management in Agricultural Markets: A Review,” Journal of
Futures Markets 21 (10): 955–85.
Townsend, R. (1995) “Consumption Insurance: An Evaluation of Risk-Bearing Systems in Low-Income
Economies,” Journal of Economic Perspectives 9 (3): 83–102.
Trieschmann, J. (2005) Risk Management and Insurance (12th edn). Mason, OH: Thomson/South-Western.
Tufano, P. (1996) “Who Manages Risk? An Empirical Examination of Risk Management Practices in the
Gold Mining Industry,” Journal of Finance 51: 1097–138.
Turnovsky, S. (1983) “The Determination of Spot and Futures Prices with Storable Commodities,”
Econometrica 51(5): 1363–88.
Turvey, C. (2006) “Managing Food Industry Business and Financial Risks with Commodity-Linked
Credit Instruments,” Agribusiness 22: 523–45.
Turvey, C. and T. Baker (1989) “Optimal Hedging Under Alternative Capital Structures,” Canadian
Journal of Agricultural Economics 37: 135–43.
UNCTAD (1997) Integrated Risk Management in Commodities. UNCTAD/ITCD/COM/8, 26 November,
Geneva.
Unger, R. (1980) “Dutch Herring, Technology and International Trade in the Seventeenth Century,”
Journal of Economic History 40 (June): 253–80.
United States Senate Permanent Subcommittee on Investigations, Committee on Homeland Security and
Government Affairs (2009) “Excessive Speculation in the Wheat Market,” Staff Report (June 24).
US Senate Permanent Subcommittee on Investigations, Committee on Homeland Security and
Government Affairs (2007) “Excessive Speculation in the Natural Gas Market,” Staff Report (June
25).
van Berkel, S. (2008) “Should Hedge Funds be Regulated?” Journal of Banking Regulation 9: 196–223.
van der Wee, H. (1977) “Monetary, Credit and Banking Systems,” in E.E. Rich and C.H. Wilson (eds) The
Cambridge Economic History of Europe. Volume V: The Economic Organization of Early Modern
Europe, Cambridge: Cambridge University Press: Ch.V.
van Dillen, J. (1930) “Isaac Le Maire en de handel in actien der Oost-Indische Compagnie,” Economisch-
Historisch Jaarboek: 1–165.
van Dillen, J., G. Poitras and A. Majithia (2007) “Isaac Le Maire and the Early Trading in Dutch East
India Company Shares,” in G. Poitras (ed.) Pioneers of Financial Economics (vol.1). Cheltenham,
UK: Edward Elgar: Ch.2.
van Duren, E., D. Sparling, C. Turvey and L. Lake (2003) “An Assessment of the Strategies and Strengths
of Medium-Sized Food Processors,” Agribusiness 19: 115–32.
van Houtte, J. (1966) “The Rise and Decline of the Market of Bruges,” Economic History Review XIX:
29–47.
Varangis, P. and D.F. Larson (1996) “Dealing with Commodity Price Uncertainty,” Policy Research
Working Paper 1667, Washington, DC: World Bank.
Vaughan, E. (1982) Fundamentals of Risk and Insurance. New York: Wiley.
Volmer, T. (2011) “A Robust Model of the Convenience Yield in the Natural Gas Market,” Journal of
Futures Markets 31: 1011–51.
Wall Street Journal (2005) “Mystery after China Copper Trader Vanishes Following Heavy Losses” (Nov
15).
Wang, D. and W. Tomek (2007) “Commodity Prices and Unit Root Tests,” American Journal of Agricultural
Economics 89: 873–89.
Wang, H. (2004) “The Impact of US Commodity Programmes on Hedging in the Presence of Crop
Insurance,” European Review of Agricultural Economics 31: 331–52.
Wang, H.H., L.D. Makus and X. Chen (2004) “The Impact of US Commodity Programmes on Hedging in
the Presence of Crop Insurance,” European Review of Agricultural Economics 31: 331–52.
Weber, M. (1894) “Stock and Commodity Exchanges?,” trans. S. Lestition (2000) Theory and Society 29:
305–38.
Weber, M. (1924) “Commerce on the Stock and Commodity Exchanges,” trans. S. Lestition (2000) Theory
and Society 29: 339–71.
Wei, J., C.-F. Lee and A. Lee (1999) “Linear Conditional Expectation, Return Distributions, and Capital
Asset Pricing Theories,” Journal of Financial Research 22: 471–89.
Wei, S. and Z. Zhu (2006) “Commodity Convenience Yield and Risk Premium Determination: The Case
of the US Natural Gas Market,” Energy Economics 28: 523–34.
Weiss, D. and M. Maher (2009), “Operational Hedging Against Adverse Circumstances,” Journal of
Operations Management 27: 362–73.
West, M. (2000) “Private Ordering at the World’s First Futures Exchange,” Michigan Law Review 98:
2574–615.
Weymar, H. (1966) “The Supply of Storage Revisited,” American Economic Review 56: 1226–35.
Weymar, H. (1968) Dynamics of the World Cocoa Market. Cambridge, MA: MIT Press.
Widell, M. (2005) “Some Reflections on Babylonian Exchange during the End of the Third Millennium
BC,” Journal of the Economic and Social History of the Orient 48: 388–400.
Williams, J. (1982) “The Origin of Futures Markets,” Agricultural History 56: 306–25.
Williams, J. (1995) Manipulation on Trial: Economic Analysis and the Hunt Silver Case. Cambridge, UK:
Cambridge University Press.
Willis, A. (2007) “Fallen Amaranth Trader Preparing New Fund,” Globe and Mail (Mar. 23): B1.
Wilmott, P. (1998) Derivatives. New York: Wiley.
Wilson, C. (1941) Anglo-Dutch Commerce and Finance in the Eighteenth Century, reprinted London:
Cambridge University Press (1966).
Wilson, T. (1572) A Discourse Upon Usury with an Historical Introduction by R. Tawney, reprinted
London: Frank Cass (1962).
Wilson, W. (1987) “Price Discovery and Hedging in the Sunflower Market,” Journal of Futures Markets
9: 377–91.
Wilson, W., W. Nganje and C. Hawes (2007) “Value-at-Risk in Bakery Procurement,” Review of Agricultural
Economics 29: 581–95.
Witt, H., T. Schroeder and M. Hayenga (1987) “Comparison of Analytical Approaches for Estimating
Hedge Ratios for Agricultural Commodities,” Journal of Futures Markets 7: 135–46.
Wood, D. (2011) “Is the Oil and Gas Industry Adequately Handling Exposure to Extreme Risks?” World
Oil 232: 113–18.
Working, H. (1933) “Price Relations Between July and September Wheat Futures at Chicago since 1885,”
Wheat Studies of the Food Research Institute 9: 187–238.
Working, H. (1949) “Theory of the Price of Storage,” American Economic Review 39: 1254–62.
World Bank (1999) Dealing With Commodity Price Volatility in Developing Countries: A Proposal for a
Market-Based Approach; International Task Force on Commodity Risk Management in Developing
Countries. Washington, DC: World Bank.
Wright, B. and J. Hewitt (1994) “All Risk Crop Insurance: Lessons from Theory and Experience,” in D.
Hueth and W. Furtan (eds) Economics of Agricultural Insurance: Theory and Evidence. Boston:
Kluwer.
Wright, B. and J. Williams (1989) “A Theory of Negative Prices for Storage,” Journal of Futures Markets
9: 1–13.
Ye, Y. and M. Yeh (1995) “A Comparative Evaluation of Yield Risk Reductions with Alternative Crop
Insurance Programs in Manitoba,” Canadian Journal of Agricultural Economics 43: 57–71.
Zakamouline, V. and S. Koekebakker (2009) “A Generalisation of the Mean-Variance Analysis,” European
Financial Management 15: 934–70.
Zhang, H. (2009) “Effect of Derivative Accounting Rules on Corporate Risk-Management Behavior,”
Journal of Accounting & Economics 47: 244–64.
Zulauf, C., H. Zhou and M. Roberts (2006) “Updating the Estimation of the Supply of Storage,” Journal of
Futures Markets 26: 657–76.
Zwick, S. and D. Collins (2006) “Yes, China. There is a Liu Qibing. Will China Honor Bad Trades?” Futures
(January): 14–15.
text to come
393