Main Contents Page • Part 1 • Part 2 • Part 3 • Part 5 • About the book

This document contains updates to my book, Global Financial Markets (Houghton Mifflin, 1994). It also contains information on developments and research in the international financial markets that might be of interest to students and professionals.

Contents

Chapter 12: The International Bond Market

The Eurobond Market in 1993; The Eurobond Market in 1994; Domestic bond markets--Japan; International Bond Market Computations

Chapter 13: Currency and Interest-Rate Swaps

Innovative swaps; Procter's Diff Swap Gamble

Chapter 14: International Equity Markets and Portfolio Diversification

Emerging Stock Markets; Performance attribution for international portfolios; Global market capitalization; New research results

Chapter 15: Global Commodity Markets

Commodity futures and interest-rate futures; The Convest-Paribas oil swap; Metallgesellschaft; New research results

Main Contents Page Back to the top of this section

The Eurobond Market in 1993

In 1993, the dividing line between Eurobonds, international and domestic bonds--and between bonds and medium-term notes--became increasingly irrelevant, as both issuers and investors broadened their horizons to take advantage of opportunities wherever they appeared. The global bond--one issued in the Euro and the domestic market simultaneously--became more common.

The Eurobond Market in 1994

In 1994, the U.S. dollar remained overwhelmingly the most popular currency of Eurobond issues, with the yen in second place. The most prominent of the banks in the market were as follows.

Domestic bond markets--Japan

Many domestic corporate bond markets remain far behind the Eurobond market in liquidity and efficiency. An example is Japan, where issuance costs are twice as high as in the Euromarket.

Until 1993, investors in Japan's corporate bond market were insulated from both excessive risk and excessive reward. Regulations initiated by the Ministry of Finance disallowed issuance of bonds with less than a single-A rating; now BBB bonds are allowed. But it seems that having been sheltered for such a long time, issuers and investors are reluctant to face the market pricing of risk in the primary bond market. Until 1992, there was no difference in yield between single-A and triple-A bonds. Riskier issuers are still reluctant to pay a premium. As a result, a year after the ban was lifted, only one BBB-rated issue had been made.

Until recently, the fiction of all-bonds-bear-the-same-yield was possible because the secondary market for corporate bonds was almost nonexistent. Now that bonds are becoming more freely traded, and yields are starting to reflect the issuer's riskiness, investors are disinclined to buy primary issues at rates that cause the paper to immediately lose value in the secondary market.

Even so, identical primary-market yields for issuers of quite different ratings remain the official norm, with underwriters absorbing substantial proportions or making deals with recalcitrant investors to sell the bonds at special discounts from the offer price. Most lower-rated issuers evidently prefer to remain aloof from the market rather than pay high yields.

International Bond Market Computations

On page 359, the formula for Eurobond valuation should be replaced.

The ISMA calculation is as follows:

Where

P = Market price of the bond, relative to 100.

CPN = Coupon rate corresponding to the time interval implied by the coupon frequency; for example, 10% per annum paid semi-annually means CPN = 5.

FACE = Price at redemption; normally 100

i = Internal rate of return, or yield to maturity

n + f = Remaining life of the bond in coupon periods. n is an integer, and f the fractional part.

Main Contents Page Back to the top of this section

Innovative swaps

Tax-rate swap. In August 1994, Morgan Grenfell, the British merchant bank, engineered a unique "tax swap" deal. The swaps market has long been used by companies to provide tax benefits. Since the tax authorities are constantly on the lookout for tax-reducing ploys, companies are usually unwilling to publicize tax-driven deals. However, the Morgan Grenfell deal involved a straightforward transfer of the risk of a U.K. tax rate increase, so there was no effect on net tax revenues to Britain's Inland Revenue.

In the deal, an international bank agreed to pay a counterparty if U.K. corporate tax rate went up, and to receive a net payment if tax rates fell. The bank was willing to take this position because of the specific structure of a leasing transaction which meant that the bank would lose if tax rates fell, and vice-versa. The swap covered GBP15 million worth of profits over 4 years, deferred for 1 year (because a tax rate increase was only considered likely if a Labour government was elected the following year). Morgan Grenfell had no trouble finding a counterparty for the other side of the deal, since many corporations would be hurt if tax rates rose.

Somewhat less happy is the story of how Procter & Gamble, an American consumer-goods company, lost $102 million after tax in 1994. Here's the story.

Procter's Diff Swap Gamble

Procter & Gamble, the country's leading maker of soaps and diapers, announced on April 12, 1994 that it would take a $102 million charge because of losses resulting from a complex swap done with Bankers Trust. This was one of a series of corporate derivatives-related losses that have been reported recently, at companies the likes of Mead, Gibson Greetings, Cargill and Metallgesellschaft A.G. The P&G-Bankers Trust swap, according to press reports, was a differential swap, or diff swap in market parlance.

A diff swap pays the difference between a floating rate in one currency and a floating rate in another currency, less (or plus) a spread, with all payments denominated in a single currency and no exchange of principal.. Diff swaps are typically used by sophisticated financial market players such as international portfolio managers. For example, a U.S. money manager, Steinhardt Partners, which is receiving three-month U.S. dollar LIBOR from a portfolio of money-market instruments, would, in a diff swap, pay that dollar LIBOR on an agreed notional principal amount to the arranging bank, in exchange for Deutschemark LIBOR minus a spread that reflects the swap-rate differential between the two currencies. The bank would be a derivatives market maker, such as Credit Suisse Financial Products (CSFP), one of a handful with the ability and will to trade these complex instruments. The diff swap works by exploiting the difference between the relative shape of two yield curves and the interest rate expectations of the buyer of the diff. The transaction is settled entirely in dollars so no exchange of principal is necessary. The buyer of a diff swap, in other words, faces no currency risk. The seller, in this example CSFP, creates the diff swap using two interest rate swaps, one in dollars and one in Deutschemarks; but since the customer bears no exchange risk, the bank is faced with a complex option-type hedging task, one which must be hedged with "quanto" options. Quantos provide a hedge against an uncertain amount of foreign currency exposure, where the exposed amount depends on some market variable or variables (such as the difference between two interest rates).

At this point the precise nature of the diff swap that Bankers Trust did with P&G is not public knowledge. It is known that the swap involved U.S. and German interest rates, and that it was a "leveraged" transaction. What follows is an example of how the two parties may have engaged in such a swap, and how it may have gone wrong.

The swap appears to have been a 5-year contract whose payoffs depended on the relation between 3-year German and 3-year U.S. interest rates. As part of the swap, P&G reportedly converted its fixed-rate U.S. dollar funding into a floating rate, to take advantage of the low short term rates that prevailed in the U.S. in recent years. Normally, this would be done by exchanging fixed for floating payments at the going market rate for interest-rate swaps. To subsidize its floating cost of funds, however, P&G agreed to an adjustment that equaled some multiple of the difference between the German and U.S. 3-year interest rate. P&G's benefit came from the supposition that German rates would fall less, and U.S. rates would rise less, that was suggested by the implied forward rates in the two countries' yield curves. The structure might have looked something like this:

As it happened, 3-year rates converged much faster that was implied by the yield curves of a year or so ago. According to one report, every 1/100 of a percentage point convergence cost P&G $400,000! So the losses on the deal started piling up in late 1993, and reached a critical point following the rate hike in the early months of 1994. At this point, presumably, P&G's management decided to bail out, albeit at a much higher cost than they could have done earlier.

One issue that has been discussed is whether the risks were fully explained and fully understood. Most people familiar with the two institutions feel that it was a transaction done between consenting adults, with the terms, and the potential risks, explicitly spelled out. No doubt the swap agreement specified the formula for the cash flows in every detail, allowing P&G to run it through a spreadsheet to predict what would happen under various interest rate scenarios.

What such an agreement cannot do, however, is to spell out the potential liquidation value of a complex derivative instrument such as a diff swap. To do so would be to give away trade secrets, even assuming P&G could replicate Banker Trust's computations and assumptions. There is, therefore, an important lesson to be learned from the event. Complex derivatives, especially illiquid ones, may have mark-to-market value fluctuations far greater than the cash-flow fluctuations typically shown in breakeven analyses. The mark-to-market or liquidation value of a complex instrument may take on importance if the corporation may, by choice or necessity, liquidate the contract. This could happen, for example, when the original exposure being hedged disappears, or when the company undertakes a change in foreign-exchange policy. If there is any chance that the deal may be reversed prior to maturity, it is the liquidation value that counts. To the extent possible, corporates should ask their bankers to simulate the liquidation values, as well as the payouts, under different market conditions.

Main Contents Page Back to the top of this section

Emerging Stock Markets

See Chapter 11, Emerging Markets

Performance attribution for international portfolios

Measuring the performance of an international portfolio requires dissecting the relative influences of currency, local market performance, and other factors, in order to evaluate the value added by management, relative to a well-accepted benchmark. In a classic article on international performance attribution, Brinson and Fachler (1985) illustrated a simple analytic framework for evaluating the returns of non-U.S. equity returns, from a U.S. standpoint. Their approach broke returns into

1. market selection,

2. stock selection, and

3. cross-product effects.

This approach was recently extended by Ankrim and Hensel (1994) who retained the simplicity and intuitive appeal of Brinson and Fachler's approach but also accounted for currency exposure, by breaking the returns due to currency exposure into two components. One component recognizes the opportunity cost of returns available in forward exchange markets, while a second measures the currency return attributable to being less than fully hedged, both in the portfolio and in the benchmark. Put differently, the returns to currency exposure in an international portfolio can be broken down into two parts--the forward premium, and the additional or "surprise" change. Because the value of active management lies in its ability to forecast the uncertain sources of return, performance attribution should focus on the ability to capture positive returns due to currency surprise.

Hence the four major components of any differential between the manager's total performance and that of some benchmark are:

1. Security selection effect: how much the manager's security selection decision within each add to the portfolio's performance differential.

(Portfolio return - Benchmark return) x (Benchmark weight)

2. Allocation effect: the impact of selecting countries in proportions that differ from those in the benchmark.

(Portfolio weight - Benchmark weight) x (Benchmark country return - Total return)

3. Forward premium effect: the return attributable to the forward contracts in place at the beginning of each period

(Portfolio weight - Benchmark weight) x (Expected currency return - average premium in benchmark portfolio)

4. Currency management effect: the effect that differential currency exposure (from the benchmark) has on return performance.

[(Portfolio weight -Benchmark weight) x (Currency surprise - Total benchmark currency surprise)] + (Forward contract adjustment)

An approach such as this one will help investors to measure more accurately the value added by active management of individual stocks, of countries, and of currency hedges in an international portfolio.

References:

G.P. Brinson and N. Fachler, "Measuring Non-U.S. Equity Portfolio Performance," Journal of Portfolio Management, Spring 1985.

E.M. Ankrim and C.H. Hensel, "Multicurrency Performance Attribution," Financial Analysts Journal, March-April 1994, 29-35.

Global market capitalization

As of November 30, 1993, the value of all shares traded around the world was $14,167 billion. The breakdown by major markets was as follows.

Market segment	Percentage of total
United States Japan Britain Other developed markets Emerging markets	37 23 9 19 12

Source: The Economist January 15, 1994.

The structure of international portfolios, 1994

The Economist's portfolio poll of July 1994 showed how ten prominent international money-management houses allocated their ideal portfolio at that time. A new feature is the "tracking error" computed for each portfolio by BARRA, an investment consultancy firm.

Holdings by instrument, %
	O	A	B	C	D	E	F	G	H	I	J
Equities Bonds Cash		55 35 10	50 40 10	60 30 10	65 30 5	65 35 0	53 42 5	48 44 8	38 54 8	50 45 5	40 48 12
Equity holdings by area, %
	O	A	B	C	D	E	F	G	H	I	J
Americas Europe Far East	38 27 35	30 30 40	48 28 24	37 22 41	28 34 38	35 31 34	35 22 43	33 38 29	33 41 26	38 25 37	37 33 30
Tracking error, equities
	0	1.6	4.9	1.1	1.4	2.1	1.8	2.9	3.3	0.6	2.2
Bond holdings by currency, %
	O	A	B	C	D	E	F	G	H	I	J
Dollar Yen Sterling DM Ffr Others	37 20 5 11 6 21	46 19 5 7 4 19	48 6 7 5 7 27	40 22 12 12 12 2	45 10 10 13 9 13	50 15 5 10 8 12	25 0 10 0 15 50	40 0 10 15 15 20	54 6 6 10 6 18	48 10 4 11 6 21	36 13 6 12 10 23
Tracking error, bonds
		0.6	1.7	1.1	1.1	0.8	2.7	2.1	1.7	1.2	0.7

Key:

O Neutral weighting for equities: Morgan Stanley Capital International world equity index

For bonds: Salomon Brothers world government bond index

A Merrill Lynch

B Lehman Brothers

C Nikko Securities

D Daiwa Europe

E Credit Agricole

F Robeco Group Asset Management

G Bank Julius Baer (Zurich)

H UBS International Investment

I Commerzbank International Capital Management

J Credit Suisse Asset Management

The "tracking error" refers to how far each portfolio's performance is likely, with 66% confidence, to deviate from the neutral portfolio's performance. For example, Robeco's bond portfolio has a 66% change of under- or out-performing the Salomon bond index by as much as 2.7%. By contrast, the Merrill Lynch portfolio is unlikely to differ from the Salomon index by more than 0.6%.

Source: The Economist, July 9, 1994.

New research results

"The Pricing of Risk in Common Shares" by MYRON GORDON,

International Review of Financial Analysis, Volume 2, Number 3, 1993.

Abstract

The capital asset pricing model predicts positive correlation between the expected return on a share and its systematic risk or beta. A number of papers in which some variant on average realized holding period return is used as a proxy for expected returns finds no significant correlation between the variables. These results cast doubt on the fundamental theorem of finance, that investors price risk. This paper uses a better estimate of expected return, and the data provides strong evidence that investors price risk. In addition the relation between beta and other risk variables is clarified.

"A Review of Recent Developments in International Portfolio Selection" by BRIAN HATCH and BRUCE G. RESNICK

Department of Finance, School of Business,

Indiana University, Bloomington, IN 47405

(812) 855-8568

Abstract

This paper begins with a foundation for the inspection of the potential gains from international diversification by citing the results of the seminal works in the area. From that pint, the paper reviews recent ex-post studies, adaptation to currency risk, consideration of bond investment, the development of ex-ante strategies, and the consideration of market imperfections. The prevailing impression from this review is that international investment can potentially provide superior performance to solely domestic investment. However, it is a matter of developing the correct strategy to exploit the opportunities.

"International Portfolio Choice and Asset Pricing: An Integrative Survey" by RENE STULZ

College of Business

Ohio State University

1775 College Road

Columbus, OH 43210

614-292-1970

Abstract

In general, theories of portfolio choice and asset pricing let investors differ at most with respect to their preferences, their wealth and possibly their information sets. If there are multiple countries, however, the investment and consumption opportunity sets of investors depend on their country of residence. International portfolio choice and asset pricing theories attempt to understand how the existence of country-specific investment and consumption opportunity sets affect the portfolio held by investors and the expected returns of assets. In this paper, we review these theories within a common framework, discuss how they fare in empirical tests, and assess their relevance for the field of international finance.

"The World Ex Ante Risk Premium: An Empirical Investigation" by BARBARA OSTDIEK

Fuqua School of Business

Duke University

Durham, NC

919-660-7788 27706

ostdiek@dukefsb

Abstract

Using multiple conditional inequality tests, this paper provides evidence against the nonnegativity of the world ex ante risk premium. The tests are conducted with several information sets used in previous tests of the conditional CAPM. The evidence indicates that the ex ante world risk premium as measured by the MSCI dollar-denominated world portfolio can be negative. If this proxy portfolio is conditionally positively correlated with the market portfolio, this is a rejection of a necessary condition of the CAPM. Previous tests of the linearity constraint of the CAPM using this portfolio likely involve violations of this necessary condition of the model. The expected risk premium on this portfolio, however, compounds the expected risk premium in the underlying assets and the expected change in exchange rates. Conducting the same set of tests on the MSCI local currency-denominated portfolio, which represents only the returns on the underlying securities, provides no evidence against the null hypothesis of a positive ex ante market risk premium.

"Excess Volatility and Closed-End Funds" by JEFFERY PONTIFF, April 1994.

Contact: Pontiff, pontiff@uwavm.u.washington.edu, School of

Business Administration, University of Washington, Seattle,

WA 98195, Tel (206) 543-4773, Fax (206) 685-9392.

Abstract

The average closed-end fund's return is shown to be 65% more volatile than its assets. Unlike variance-bound tests, this facilitates an excess volatility test that does not rely on strong assumptions about discount rates or dividend streams. This finding can not be attributed to non-synchronous trading, or distorted net asset values. Although largely idiosyncratic, 15% of the average fund's excess risk is explained by market risk, small firm risk, and risk that affects other closed-end funds. For discounted funds, excess risk is also related to the book-to-market risk.

Market Microstructure and Asset Pricing: On the Compensation

for Market Illiquidity in Stock Returns, London Business

School Institute of Finance and Accounting Working Paper

190

BY: MICHAEL BRENNAN and AVANIDHAR SUBRAHMANYAM

CONTACT: AVANIDHAR SUBRAHMANYAM

E-MAIL: asubrahm@agsm.ucla.edu

POSTAL: Anderson Graduate School of Management,

University of California, Los Angeles, 405

Hilgard Avenue, Los Angeles, CA 90024-1481,

USA

PHONE: (310) 825-3587

FAX: (310) 825-3587

OTHER: Contact: Michael Brennan -

mbrennan@agsm.ucla.edu

REF: WPS94-277

Models of price formation in securities markets suggest that

privately informed investors are a significant source of

market illiquidity. Since illiquidity increases the

round-trip trading cost of an investor, this implies that

uninformed investors will demand higher rates of return from

securities in which informational asymmetries are more

severe. In this paper we derive a simple relationship

between expected stock returns and market illiquidity in a

model with a single representative investor. Using CRSP data

for the period 1984-1992, and ISSM intraday data for the

year 1988, we investigate the empirical relation between

stock returns and measures of market illiquidity. We find a

significant relation between required rates of return and

our measure of market illiquidity using two types of test.

First, following Amihud and Mendelson (1986), we control for

the effects of firm size and systematic risk, as well as the

quoted spread; and secondly, following Fama and French

(1993), we adjust for risk factors related to the overall

market, firm size, and the book-to-market ratio.

"The Nestle' Crash" (August 1994)

BY: CLAUDIO LODERER and ANDREAS JACOBS

CONTACT: CLAUDIO LODERER

E-MAIL: lodere@ifm.unibe.ch

POSTAL: Universitat Bern, Institut fur

Finanzmanagement, Sennweg 2, 3012 Bern,

Switzerland

PHONE: 41 31 631 3775

FAX: 41 31 631 8421

REF: WPS94-304

On November 17, 1988, the board of directors of Nestle' AG

decided to allow foreign investors to hold Nestle'

registered stock, reversing a longstanding practice. This

decision had a tremendous impact on the prices of the firm's

three classes of common stock, as well as on the prices of

several other corporations traded on the Zurich stock

exchange. These price changes can be explained by the

hypothesis that demand curves slope down.

"On Selectivity and Market Timing Ability of U.S.-Based

International Mutual funds: Using Refined Jensen's Measure"

GLOBAL FINANCE JOURNAL, Vol 5 No 1, Spring 1994

BY: SON-NAN CHEN and HOYOON JANG

CONTACT: HOYOON JANG

E-MAIL: eadtyoon@ube.ub.umd.edu

POSTAL: Department of Economics and Finance,

University of Baltimore, Merrick School of

Business, Baltimore, MD 21201-5779

PHONE: (410) 625-3255

FAX: (410) 752-2821

REF: APS94-119

This paper evaluates the performance of 15 U.S.-based

international mutual funds for the period 1980-89.

Selectivity and timing skills of mutual fund managers are

the primary criteria for performance evaluation. The

technique used here is the one developed by Treynor and

Mazuy and refined by Lee and Rahman. We find that many of

the international mutual funds outperformed the U.S. market

benchmark, perhaps due to the expanded diversification

opportunities that they provide. When a world market index

is used as the benchmark, fund managers show relatively poor

performance in terms of selectivity skills. However, there

is strong evidence that some managers rely rather heavily on

timing skills in international capital markets.

"Investment Restrictions and the Pricing of Korean

Convertible Eurobonds" (May 1994)

BY: WARREN BAILEY, Y. PETER CHUNG, and JUN-KOO KANG

CONTACT: WARREN BAILEY

E-MAIL: bailey@jgsm2.gsm.cornell.edu

POSTAL: Johnson Graduate School of Management, Cornell

University, Malott Hall, Ithaca, NY

14853-4201

PHONE: Not Available

FAX: Not Available

OTHER: Y. Peter Chung: ychung@ucrac1.ucr.edu

Jun-Koo Kang: jkang@uriacc.uri.edu

REF: WPS94-358

Corporations in developing countries with foreign investment

restrictions have begun to issue convertible bonds overseas.

Given low covariance of emerging market equity returns with

global economic risks, foreign investors should place a

relatively high value on these bonds. We use price data for

convertible Eurobonds of four Korean corporations to measure

the implicit premium foreigners offer for Korean equities

relative to prices in the domestic Korean stock market. We

find large, volatile premiums which vary widely across firms

and differ from the premium on the closed-end Korea Fund

listed on the New York Stock Exchange, particularly after

foreign access to the Korean stock market was liberalized.

Our results have several implications for the pricing of

emerging market securities.

Do Bulls and Bears Move Across Borders? International

Transmission of Stock Returns and Volatility by WEN-LING LIN

and ROBERT F. ENGLE, and TAKATOSKI ITO, Review of Financial

Studies, Volume 7, Number 4, 1994.

Contact: Lin, E-mail: ecwlin@macc.wisc.edu, Department of

Economics, University of Wisconsin - Madison, 1180

Observatory Drive, Madison, WI 53706, Tel (608) 263-3858 Fax

(608) 263-0477.

This paper investigates empirically how returns and

volatilities of stock indices are correlated between the

Tokyo and New York markets. Using intradaily data that

define daytime and overnight returns for both markets, we

find that Tokyo (New York) daytime returns are correlated

with New York (Tokyo) overnight returns. We intrepret this

evidence that information revealed during the trading hours

of one market has a global impact on the returns of the

other market. In order to extract the global factor from the

daytime returns of one market, we propose and estimate a

signal model with GARCH processes.

Main Contents Page Back to the top of this section

Commodity futures and interest-rate futures

Replace the formula on page 459 with the following:

where Ft = Commodity futures for delivery at t

RFt,t+n = Interest rate futures for period t to t+n

The following is a good example of the use of a commodity swap by a producer of oil.

The Convest-Paribas oil swap

The use of oil swaps increased sharply in late 1990 and early 1991, following Iraq's invasion of Kuwait. An example is the swap arrangement between Convest Energy Corp., a Houston-based company that produces about 500,000 barrels of crude a year, and Banque Paribas.

In January 1991, Convest arranged to receive a guaranteed price of $23.80 a barrel of crude for about 25% of its 1991 output. The guarantee, which was settled in cash, worked as follows. Each month, Paribas calculated the monthly average settlement price of the nearby futures contract of light sweet crude oil traded on the New York Mercantile Exchange. At the end of each month, if the average settlement price was above $23.80, Convest paid Paribas the difference. When it was below, Paribas paid Convest the difference. The arrangement covered 11,500 barrels a month and lasted for one year.

Questions:

(1) What is the advantage or disadvantage of this method versus using futures, from Convest's point of view?

(2) Does Convest face basis risk?

(3) Does Convest face counterparty risk?

Answers:

(1) Prior to this deal, Convest had used futures to some extent. However, in the words of Convest's chief financial officer, Scott O'Keefe, "The thing that's beneficial to an agreement like this vs. going out onto the Nymex and executing contracts is that...you have to post a tremendous amount of margin out there. In this arrangement we don't have to put out any cash."

However, if Convest did get into a theoretical loss, they had agreed to post a letter of credit with Paribas. Also, the Paribas swap was certainly less liquid than futures.

(2) Basis risk, the difference in the movement of the underlying instrument being hedged and the movement of the hedging contract, is present in any such cash settlement deal. However according to Convest, the average price of the variety of grades produced by the company was close to that of the light sweet crude contract traded on the Nymex.

(3) Yes, Convest would lose if the price of crude fell below $23.80 and Paribas defaulted on their obligation. The loss would equal the replacement cost of the swap.

Metallgesellschaft

Metallgesellschaft A.G. is a German company that incurred huge losses on commodity trading in 1993-94. (In February, 1994, the New York Times reported that MG had incurred losses of $1.33 billion on the New York Mercantile Exchange.) MG, a steel fabricating company, had diversified into energy products. To win market share, they apparently entered into delivery contracts that provided buyers with attractive fixed future prices. According to press accounts, MG had extended long-term delivery contracts (5-10 years) on oil-related products to many customers, including fuel-oil distributors and even individual filling stations. (This in itself carried significant performance risk, because if the cash market moved against the company, it could not be sure these counterparties would honor the contracts.)

Instead of hedging these delivery commitments with over-the-counter forward contracts, MG hedged the price guarantees using NYMEX futures, which have maximum maturities of only 36 months in the crude oil contract. This left them with a gap or timing mismatch between their physical positions, the promised fuel delivery, and the hedged positions (many in 3 month contracts).

MG's traders seem to have thought that they could undercut the market, supporting long-term forward deals with short-term futures hedges. Effectively, the company took a position in oil akin to borrowing short and lending long. Nor, it appears, did they build in a cushion sufficient for the market and credit risks they faced. While the long-term delivery contracts (to sell) theoretically went up in value as the oil market fell, they did not result in a positive cash flow to offset the cash losses on the hedges. Meanwhile, some counterparties declined delivery of oil, leaving MG with open loss positions on its futures contracts.

An important consequence of the Metallgesellschaft losses is that they highlighted the practical differences between managing foreign-exchange and interest-rate risks, and managing commodity risks.

1. Credit and/or performance risks are much higher in physical transactions. The parties involved are often not financially robust, and force majeure or intentional delivery delays are not uncommon.

2. The markets in commodity derivatives are not as deep or as liquid as interbank financial markets.

3. Commodity markets are more prone to influence or manipulation by dominant or collaborating participants. Reportedly, because of the size of MG's activities, the locals at NYMEX knew that MG was going to have to roll over the short-dated futures contracts to cover their long-dated cash contract positions. Trading against MG at the roll-over points, commodity traders could profit at MG's expense.

4. Margin calls on commodity futures contracts generally require immediate cash settlement (with the exception of the London Metal Exchange). Putting up the cash against losses in volatile markets makes for costly financing of even fully hedged physical positions.

New research results

"Indexed Commodity Futures and the Risk and Return of

Institutional Portfolios", (OFOR Working Paper Number 94-02)

BY: KENT G BECKER and JOSEPH FINNERTY

CONTACT: JOSEPH FINNERTY

E-MAIL: finnerty@ux1.cso.uiuc.edu

POSTAL: College of Commerce, University of Illinois,

1206 South Sixth Street, Champaign, IL 61820

PHONE: (217) 333-2815

FAX: (217) 244-3118

REF: WPS94-291

Abstract

We examine the risk and return properties of equity/bond

portfolios before and after inclusion of a diversified

portfolio of long commodity futures contracts. Inclusion of

the commodities, which are proxied by the CRB and GSCI

indices from 1970 to 1990, enhances the risk and return

characteristics of the overall portfolio. However, the

improvement of the risk/return characteristic is superior

for the decade of the seventies than for the decade of the

eighties. This result is driven by the high-inflation 1970s

in which commodity futures serve as an inflation hedge. In

addition, commodity futures prices are shown to have modest

inflation predictive ability.

"Does Futures Speculation Stabilize Spot Prices? Evidence from Metals Markets" by A. ENIS KOCAGIL

Contact: Kocagil, E-mail: kocagil@ems.psu.edu, Department of

Mineral Economics, The Pennsylvania State University, 221

Eric A. Walker Building, University Park, PA 16802-5010, Tel

(814) 863-0810 Fax (814) 863-7433.

Abstract

The effects of speculation in futures markets have been a topic of a long lasting debate in both empirical and theoretical literature. This study attempts to empirically test the relationship between speculation in futures markets and spot price volatility. The contribution of this study is two-fold: the theoretical framework which was developed by Driskill, McCafferty and Sheffrin (1991) is generalized and an empirical test of the hypothesis that futures speculation decreases spot price volatility is conducted using data on some metals markets (i.e., copper, gold, silver and aluminum) Regression equations are estimated based on weekly futures and spot price series. Monte Carlo simulation methods are employed to test the accuracy of the estimated coefficients. The results, which are based on these four metals markets, for the period 1980-1990, reject the hypothesis that an increase in futures speculation intensity tends to decrease the spot price volatility, and thus, stabilizes spot markets.