Play Markowitz for Me
Ferrell Capital's new trust allocates by volatility instead of asset
By J.C. Louis
These days, more and more proprietary traders at investment banks are
being paid not on their returns but on how much risk they took to get those
returns. Although the sell-side is rapidly embracing the concept of risk-adjusted
returns, the buy-side may not be making full use of the advantages of portfolio
theory proposed by Harry Markowitz more than 40 years ago.
Ferrell Capital Management, an investment management and risk advisory
service, is preparing to launch an investment trust for institutional investors
that it hopes will push investors from asset allocation to risk allocation.
Prospective investors select from a series of volatility levels-5 percent,
10 percent or 15 percent. The trust then seeks the highest possible returns
from within that targeted volatility through the active management of multiple-asset
This offensive risk-allocation aggregates manager styles, asset allocations and the actual security positions on a single measure of risk-adjusted return-the
Sharpe ratio. (The Sharpe ratio measures risk-adjusted return as the difference
between the gross return and the risk-free rate, divided by the volatility.)
"Everyone talks about managing risk relative to a benchmark, but
the benchmark is where all the risks are," says Bluford Putnam, president
of CDC Investment Management. In other words, if the benchmark is wildly
volatile, and you are trying to beat the benchmark, your returns on a risk
adjusted basis might still be poor. Instead, the trust uses volatility itself
as its benchmark.
In a declining-rate environment, for example, a fixed-income client in
Europe may feel the need for more yield but may be hesitant to use equities,
which carry twice the risk of bonds. "What they need," observes
Putnam, "is a volatility targeted vehicle that can diversify across
many strategies, including equities, but with a volatility targeted to the
European bond market."
The approach is a radical departure from conventional pension management. "In conventional pension investment, allocations are shaped around
investment guidelines that allocate dollars based upon duration constraints,
credit restrictions, style limitations," says William Ferrell, president
of Ferrell Capital Management. "These portfolios are then rebalanced
once per quarter, diversified often with illiquid or exotic securities and
calibrated with historical 'ex-post' correlation estimates of questionable
Instead of this conventional long-only investment posture, the trust
tries to reduce correlations with the equity or fixed-income benchmarks
by using long/short strategies. "Long-only investing is like playing
the piano with the white keys, but not the sharps and flats," Ferrell
says. "Long/short investing takes more advantage of Markowitz than
traditional portfolio management." It also manages the portfolios more
dynamically with more liquid securities using more sophisticated correlation
and diversification techniques.
The trust implements the strategy by using a number of different managers in the same asset class who have different styles. Managers who use a long/short
approach to growth stocks, for example, have relatively low correlation
with managers who use a long/short approach to value stocks. "If you
have a volatility target, you can diversify to achieve that target,"
says Ferrell. "You can use multiple asset classes, countries and management
styles to diversify and lower the volatility of the portfolio."
The trust then manages the managers by closely monitoring their positions using a Value-at-Risk calculation corresponding to the appropriate volatility
level. The trust will look at the asset manager's underlying strategy, style
and its allocations and raw positions, asset by asset, and then aggregate
them. This close daily "ex-ante" position monitoring will reveal
and signal an adjustment, for example, if too many managers are exposed
to one country or asset type.
"We look at what they are doing every day," says Ferrell. "When correlations go up, we adjust exposures. To target volatility, we can change
allocations, have the managers change their position sizes, or simply overlay
a hedge. If the Deutsche mark is volatile, we can dial down the amount traded
from a $70 million position to, say, $50 million." In this way, says
Ferrell, "we're playing Markowitz with three or four traders and perhaps
The fund will also use leverage to optimize a portfolio around a target
volatility. "If you find the right mix of asset classes and styles
that give you the top return for a 5 percent preselected volatility,"
asserts Ferrell Capital chief investment officer Richard Zecher, "the
best way to maximize return for a 15 percent target volatility may be to
leverage the 5 percent portfolio mix to the higher target."
Ferrell's incentive fee is also pegged to the net risk-adjusted performance, measured by the Sharpe ratio. "Hedge funds take a ton of risk without
telling you what they are doing," says CDC's Putnam. "If Ferrell
makes a fortune on a position that has a ton of risk, they do not get paid.
The difference is extraordinary."
Commodity Users Struggle With Disclosure
By J.C. Louis
Companies that use financial instruments to manage currency and interest rate risk are all too familiar with the problems associated with the financial
disclosure of derivatives instruments. Most have watched carefully over
the past few months as the Securities and Exchange Commission moved to require
increased quantitative disclosure of derivatives instruments in annual reports.
The new disclosure requirements, however, have caught many commodity
users by surprise. "It's the first time that companies with commodity
hedging exposures have had to report on their activity," says Elizabeth
Glaeser, director of the corporate markets group at Deloitte & Touche.
In the past, companies with commodity exposures were not required to disclose
fair value of their hedge positions. Although inventories and transactions
settled by physical delivery are excluded from regulation, companies that
use commodity contracts settled in cash or with other financial instruments
such as 90-day futures face the same reporting requirements that larger
financial institutions face.
The SEC gave registrants a range of possible quantitative representations-tabular listings of fair value information and contract terms, sensitivity analysis,
or a Value-at-Risk measure. But a number of groups protested extending the
disclosure requirements to commodity transactions altogether. The Financial
Executives Institute (FEI), for example, opposed the inclusion of commodity
instruments settled by physical delivery in the new regulations. These instruments,
it maintains, are no different from purchase orders or physical inventory.
"The fact that a company enters into avariably priced purchase order
for a commodity," declared a March 1997 FEI letter to the SEC, "and
then uses the commodity market to fix the price...via a futures contract...is
the same as entering into a fixed-price purchase order with its supplier."
The commodity requirements seem particularly irksome to these users because most commodity positions are deeply integrated into the procurement, production,
sales and marketing chain. "Why does the commission believe that users
are better served by knowing that an enterprise has purchased a copper forward
by means of an exchange-traded instrument, but have never felt the information
to be sufficiently interesting to request analysis of open purchase orders?"
asked Philip Ameen, vice president and comptroller of General Electric,
in a letter to the SEC before the announcement.
The Treasury Management Association (TMA) went further, objecting to
the qualitative discourse rules as biased against firms engaged in financial
instrument transactions. It noted that a grain miller that purchased a farm
as a hedge against its wheat purchases is not subject to reporting requirements.
Such "vertical integration hedges" can go unreported while a partially
hedged producer using futures is required to produce lengthy disclosures
of its primary risks. "Such requirements will only create the perverse
result of discouraging hedging activities...and encouraging...unhedged risk
Critics also argued that the tabular method of disclosure, one of the
SEC's three disclosure alternatives, would give away far too much competitive
information on commodity holdings. "The tabular method is point-specific,"
says Deloitte & Touche's Glaeser. "It shows underlying exposures
that could disclose competitive commodity positions. There is a huge amount
of sourcing activities involving futures that has never been part of financial
disclosure." Bruce Domash, senior manager of the audit department of
the Chicago Board of Trade, agrees. "Competitors could glean what direction
the firm is going in the market," he says.
Before the SEC decision, Hershey, a commodity-oriented company with a
powerful stake in the outcome, campaigned actively in favor of a disclosure
methodology that would protect proprietary information concerning the sizing,
hedging and pricing of its cocoa positions. Robert Rosenbaum of the law
firm of Arnold and Porter, which represented Hershey, argued that the company
"demonstrated that you could easily reverse-engineer sensitivity analysis
or Value-at-Risk disclosures, and thereby determine from them the position
that the company has in the commodity market." This was particularly
true for a company heavily concentrated in one commodity.
The SEC responded by permitting quantitative disclosure of market risk
based on a netting of the futures position with the underlying commodity
or physical procurement. For example, if the company procured $1 million
worth of cocoa and purchased $600,000 in futures contracts as a hedge, the
company would make a Value-at-Risk calculation based on the $400,000 of
unhedged procurement contracts at risk. Since disclosure is made solely
on the difference, it would be difficult or impossible to determine underlying
positions. "Because the futures market and competitors do not have
access to both sides of the equation," says Rosenbaum, "it is
more difficult to reverse-engineer the disclosures and derive proprietary
Some consultants, however, have tried to point out something overlooked
in the debate-that the SEC disclosures exercise could help commodity companies
better understand the risks they are exposed to. Bill Foote, senior manager
of risk management and regulatory practice at Ernst & Young, notes that
commodity firms with wholly or majority-owned commodity trading arms ancillary
to the core business should benefit from the mark-to-market and Value-at-Risk
calculations. "One purpose in SEC disclosure is to provide better information
to assess the financial quality of a firm," he says. "A firm that
lowers its risk by effective hedging can utilize the disclosure mandated
by the SEC to convey a clearer picture of the company's financial strength.
The company's bond rating can improve and the cost of capital can decline,
enhancing shareholder equity."
But others doubt whether the SEC mandates will serve any useful purpose
for commodity companies-or anybody else. "Nothing has stopped commodity
users from rigorous risk analysis of their positions to date," says
Ken Lehn, professor of business administration at the University of Pittsburgh
"They don't need the SEC mandates to do that. The power of SEC disclosure
to lower the cost of capital is overrated. Investors are unlikely to gain
valuable information from Value-at-Risk or fair-value disclosures. These
will probably not achieve the SEC's goals of rendering disclosures between
The Great Franco-American Tech Swap
By J.C. Louis
Exchanges have spent fortunes developing systems to trade and clear the irproducts, and each has developed separately, forcing member firms to deal
with each exchange on its own technological terms.
A major technology swap between American and European exchanges, set
to be completed this summer, promises a new level of standardization and
operational harmony that may reduce clearing costs and streamline front-
and middle-office management for clearing member firms.
Last month, the Chicago Mercantile Exchange (CME) and the New York Mercantile Exchange (NYMEX) officially adopted the Paris Bourse's NSC trading system,
which has been modified for derivatives trading to serve as MATIF's after-hours
trading system. In return, the Paris markets adapted Clearing 21, a clearing
system developed jointly by the CME and the NYMEX. The swap plans to breathe
new life into the CME's GLOBEX system, which was flirting with technological
obsolescence. (See box.)
The key change occurred when the MATIF agreed to adapt the Paris Bourse's NSC system for after-hours derivatives trading. "The Bourse always
had a strong expertise in electronic trading focused on the equity market,"
says Francois Guy, deputy managing director of MATIF in charge of information
technology. "We had the back-end engines for matching transactions
and only had to develop it for derivative products."
The NSC's open architecture is a harbinger of a new standard for electronic trading. Under the older systems, a trader wanting to access different markets
needed a separate workstation for each separate market. Open architecture
allows traders to work within the same graphical interface on any market
in which they have authorization. Another hallmark of this new standardization,
says Ellen Resnick, a spokesperson for the CME, will be dial-up access that
will let traders use any desktop or notebook computer to access the system.
"We're very excited by the idea of offering members a single platform with other exchanges," says Nacamah Jacobovits, a NYMEX spokesperson.
"We like the idea of not having to train staffs in different systems,
which will certainly lower expenses for member firms." NYMEX and the
other exchanges view increased screen trading as a good way to support certain
low-volume contracts during the trading day, build liquidity and provide
a nurturing environment for new products from different time zones.
The exchanges are also eager to tout the cost-saving advantages of Clearing 21, the CME's flexible, electronic real-time clearing system that was developed
jointly with NYMEX. The Merc's 81 clearing member firms are already using
Clearing 21, which will be fully operational at the Merc by the end of the
year and at NYMEX in early 1998.
Clearing 21 allowed the CME to give up the antiquated mix of main-frame
computer systems and manual processing used to clear trade and position
data. The onset of after-hours electronic trading only underscored existing
inefficiencies. But the momentum behind the parallel Bourse-MATIF and CME-NYMEX
joint ventures could carry well beyond the present technology swap. Once
a standard is in place, notes Jacobovits, it becomes much more efficient
and less costly for other exchanges to adapt the new standard than to maintain
a separate, less compatible system. "Just as there is no need to redevelop
electronic trading on the American side, there is no need to redevelop the
clearing system on the European side," says the MATIF's Guy. "Once
you can trade and clear derivatives with the functionality of the equity
market, you have a world system in place that will enable anyone on any
platform to access the complete range of products in any capital market."
Bourse Deal Saves GLOBEX
By Simon Boughey
The technology swap agreement between the Chicago Mercantile Exchange
(CME) and the Paris Bourse did not come a moment too soon for the CME and
GLOBEX. The exchange's agreement with Reuters, GLOBEX's creator, is due
to expire in April 1998. If the CME had not signed a new deal with the Bourse
to hook up to the NSC, it would have been facing "technological oblivion,"
as one systems analyst put it.
Under the terms of the arrangement, the CME and the New York Mercantile
Exchange (NYMEX) are to adopt the Bourse's NSC system for after-hours trading.
In exchange, the Paris market will use the Clearing 21 system developed
jointly by the CME and NYMEX, so no money actually changes hands. The NSC
system is due to be implemented by November 1998, and the name GLOBEX is
still to be used, even though it will bear no resemblance to GLOBEX as it
was originally conceived.
GLOBEX has had a famously checkered history. It was born in 1987 in a
star-crossed arrangement between Reuters, the CME and the Chicago Board
of Trade. Reuters agreed to provide all the cash, hardware and software,
and the exchanges the products. Four years later, when the system was finally
launched, it was already technologically out of date.
The Chicago Board of Trade (CBOT), pulled out in 1994, dealing a serious blow to the project. The Board then went on to develop Project A, its own
after-hours system, which has become increasingly successful in recent months.
MATIF also joined the consortium, but when it decided to pull out early
in 1997, the system was clearly in its death throes. It never succeeded
in attracting other exchanges, and consequently never generated the planned
revenues for Reuters.
The mess in which the CME found itself has led to a reversal of the fortunes of chairman Jack Sandner. CME sources say that Sandner was never really
interested in after-hours systems, as they threatened the brokers in his
power base who viewed any automated system with suspicion. Without support
from the top, GLOBEX foundered, and some members of the exchange grew dissatisfied
with Sandner's leadership.
The end result? The return of Leo Melamed, who first promoted the GLOBEX system a decade ago. "We started out in first, but all of a sudden
we fell behind technologically," says one CME heavyweight. "The
membership was aware of that and pushed Leo back into a leadership role."
Competing EMU Models for Euroskeptics
By Robert Hunter
Not since the days of the "Who Shot J.R.?" craze has a more
suspenseful international drama unfolded than the countdown to European
Economic and Monetary Union. Pundits around the world are voicing their
opinions daily on the likelihood of EMU by January 1, 1999. Germany's obsession
with the mark, England's obsession with its colonial past and France's obsession
with "Frenchness" have provided ample grist for the mainstream
But these stories are for financial lightweights. For a quantitative
assessment of the likelihood of EMU, JP Morgan and Paribas are offering
two competing sets of statistical models that claim to offer unbiased, market-based
approaches to EMU predictions. Making predictions is always a shaky proposition,
but these models try to crunch out a vision of the future without a crystal
JP Morgan was first out of the gate when it unveiled the EMU Calculator, an index that measures the statistical likelihood of EMU based on market
probabilities. The probabilities are based on the assumption that long-term
interest rate spreads between the European countries involved are influenced
by the market's expectations of EMU, which are pegged at either 100 percent
probability or 0 percent probability. At any given time, the markets will
trade between these two probabilities, and the closeness with which the
market approaches either of these extremes indicates the relative probability
The model uses Germany as its baseline, assuming (perhaps incorrectly
in light of recent events) that the probability of Germany and one other
country joining EMU on January 1, 1999 is 100 percent. The probability of,
say, France joining EMU is calculated by comparing three spreads-today's
post-1999 forward swap spread between French francs and German marks, the
market's 100 percent probability of EMU, and the 0 percent probability.
The mathematics are straightforward: EMU probability = (actual price level
- no-EMU price level) / (EMU price level - no-EMU price level). One slight
problem: the model can produce probabilities higher than 100 percent.
Since spreads are determined only by the expected changes in exchange
rates, if the market is 100 percent certain today that EMU will happen by
1999, today's post-1999 forward swap spread between, say, lira and marks
would be zero. The no-EMU swap spread is a bit more complicated. JP Morgan
looks at a period when expectations of EMU were known to be quite low and
performs a statistical regression of swap spreads, focusing on the difference
between actual, estimated and EMU levels.
Paribas criticizes the model in four main areas:
- Forward swap spreads indicate a great deal more than simply the probability
that a country will join EMU. Global currency trends, such as a strong
dollar or yen, will make spreads converge as well, and the Morgan model,
which Paribas calls the "Simple Model," would automatically attribute
smaller spreads to an increase in the probability of joining EMU.
- The Morgan model uses historical spreads from the 1980s, but today's
global economic climate is much different.
- The model is unnecessarily rigid in its assumptions that countries
either will join Germany in EMU in 1999 or will not, ignoring the possibility
of a country joining after 1999. This rigidity pushes the probability results
higher than they would be if other scenarios were considered.
- The model relies too heavily on historical averages, which can vary
significantly depending on which time period is chosen.
As an alternative, the bank offers what it calls modestly "The Superior Model." Like the EMU Calculator, it looks at forward swap spreads,
but assumes that countries could join in 2000 or in any year after that.
The model looks at one-year rates, since the countries who will be joining
in 1999 will have identical one-year forward rates, and these will be lower
than those of the countries not joining in 1999. The spread of one-year
rates between each country and Germany can be used to determine the market's
belief in a country joining EMU in 1999. The key to the model is that it
assumes that the spread is normally distributed around its mean or average
spread. The probability is calculated based on the size of the area on the
distribution curve beneath Morgan's EMU Calculator spread.
It's important to remember one thing: predictive models, no matter how
complex, can often be utterly useless. Like that famous episode of "Dallas,"
no one will really know how EMU will play out until it does.