An Investment Linked to Commodity Futures
In November 1991, the Swedish Export Credit Corporation, an AAA credit, issued a three- year zero coupon note whose prospective returns were linked to the returns from commodity futures. The total size of the issue was $100 million and the notes matured in November, 1994. Institutional investors participated in the issue, buying the notes for par in denominations of $100,000 and $50,000 (Exhibit 1).
Unlike typical zero coupon notes, the Swedish Export issue had no fixed interest rate. Instead, the prospective return on the notes was tied exclusively to the future value of the Goldman Sachs Commodity Index (GSCI). The GSCI was computed on the basis of the cumulative total returns from a fully collateralized investment in an index of commodity futures. At maturity, investors would redeem the notes according to the following formula:
Par x [.9557 x GSCI End/GSCI Begin]
GSCI Begin represented the value of the commodity index – 2590.81 – on the note's issuance date of November 7, 1991. GSCI End represented the unknown future value of the commodity index on the note's maturity date of November 7, 1994. According to the redemption formula, an index level of 2710.90 on the ending date would retire the notes at 100% of par value (Exhibit 2).
The commodity-linked notes could not be called by the issuer nor could they be redeemed by the investor prior to the maturity date. The notes, however, did trade in the secondary market. Since their issuance, they had been trading at prices ranging between 80 and 100 % of par value.
Goldman Sachs and the Commodity Business
Goldman Sachs, founded in 1869, was a leading investment banking and securities firm with offices in the United States, Europe, and Asia. Goldman performed a full range of activities in the equity and fixed-income markets for institutions and individuals. Goldman also managed operations in corporate finance, municipal finance, merchant banking, and real estate.
The creation of the Goldman Sachs Commodity Index was an extension of Goldman's expertise in commodities. Through its J. Aron Currency and Commodities Division, Goldman acted as major dealer in the commodity futures market. J. Aron, which Goldman acquired in the early 1980s, also had a significant presence in the spot commodity market. Aron purchased crude oil and
A. Nancy Donohue, MBA '92, Professor Kenneth A. Froot, and Professor Jay O. Light prepared this case as the basis for class discussion rather than to illustrate effective or ineffective handling of an administrative situation.
Copyright © 1992 by the President and Fellows of Harvard College. To order copies, call (617) 495-6117 or write the Publishing Division, Harvard Business School, Boston, MA 02163. No part of this publication may be reproduced, stored in a retrieval system, used in a spreadsheet, or transmitted in any form or by any means— electronic, mechanical, photocopying, recording, or otherwise—without the permission of Harvard Business School.
293-017 An Investment Linked to Commodity Futures
petroleum from major oil producers to supply refiners and industrial users worldwide. It sourced metals from mines around the world, arranged for their refining, and supplied them to industry. Aron advised over 50 central banks in the management of gold reserves, was one of the world's largest coffee merchants, and was a merchant in grains and oil-seed products.
The Design of the Goldman Sachs Commodity Index
Goldman Sachs introduced the GSCI in 1991 with the objective of opening the commodities market to greater institutional investor participation. The proprietary index measured the total return from investing in a basket of collateralized commodity futures in the energy, livestock, agricultural, industrial metal, and precious metal sectors (Exhibit 3).
Goldman Sachs believed that the weights incorporated in its commodity futures index reflects the relative impact each commodity has on the worldwide economy. The design of the GSCI was analogous to that of a capitalization-weighted stock market index which reflected the relative impact each individual stock had on the overall market. Specifically, commodity futures in the GSCI were weighted (arithmetically) according to the dollar value of their physical quantity of worldwide production. This gave heavy representation to energy futures, which in 1992 comprised almost 48% of the index's value. Livestock futures, meanwhile, represented 23% of the index, and agricultural futures represented 21%. The remaining sectors, industrial and precious metals, comprised 6% and 2% of the index, respectively.
Each commodity future in the index had to pass strict tests for liquidity. Only commodities with active futures contracts denominated in U.S. dollars and traded on exchanges in major industrialized countries were included. The composition of the index had become broader over time because fewer futures had been traded in the past. Between 1970 and 1980 the index was over 80% weighted in livestock and agricultural futures. During the 1980s, as new futures contracts were introduced and trading activity in metals and energy futures increased, the index became more diversified.
The GSCI was calculated from the total returns of futures contracts, rather than from the spot prices for each commodity. Unlike spot prices (or common stocks), there were a number of alternative futures contracts for every commodity, each representing a different futures maturity date. The index computation, therefore, required a set of decision rules for selecting futures contracts in individual commodities. In general, the decision rule was to hold the most active (the nearest) contract month until the contract had one month until expiration. At that time, the contract was rolled over into the next nearest contract. The rolling process entailed selling the future with only one month remaining and purchasing the next nearest futures contract. The most actively traded commodity futures (crude oil, unleaded gasoline, heating oil, aluminum, and zinc) had contracts maturing each month and, therefore, were rolled every month. Other commodity futures had bimonthly maturities and were thus rolled approximately six times per year. Futures with only four maturities per year (sugar, cotton, and platinum) were usually rolled only four times a year. To avoid delivery of the commodity, positions in any contract scheduled for delivery in the next month were rolled forward in the month prior to maturity.
Goldman Sachs normalized the GSCI index to a value of 100 on January 2, 1970. On the last day of trading in 1991, the index stood at a level of 2291.90. (Over this same time, a comparably weighted index of spot prices, called the Goldman Sachs Spot Index, rose from 100 to 176.92.) Cumulative returns for the period 1970–1990 are plotted on a logarithmic scale in Exhibit 4. The returns on the S&P 500 and U.S. Treasury bonds are plotted for comparison. Over the period, the returns from the GSCI collateralized futures portfolio were clearly attractive.
2
An Investment Linked to Commodity Futures 293-017
Measuring the Returns on the GSCI
The returns on the GSCI were generated so as to replicate the returns that would have been earned on a fully collateralized portfolio of commodities futures. In particular, the returns assumed that an investor posted collateral consisting of U.S. Treasury bills equal to the value of the commodity futures prices.1
The total returns on this portfolio consisted of two components, the yield from the collateral Treasury bills and the return from the futures. The return from the futures could be broken into two components: the spot return and the roll yield. The spot return came from the changes in the commodities' spot prices. The roll yield represented the additional gain or loss captured when rolling forward the commodity futures in a portfolio as described previously. As each futures contract approached maturity, its price approached the spot price. It was this movement in price that allowed an investor to earn a return from the roll yield. When a commodity's price structure was contango (futures prices greater than spot prices), the roll yield was negative, because the futures' price decreased toward the spot price as the future matured. When a commodity's price structure was in backwardation (futures prices below spot prices), then the roll yield was positive, because the futures' price increased toward the spot price as the future matured. In general, then, the total return on the GSCI had three components:
GSCI total return = Spot return + T-bill yield + Roll yield
Exhibit 5 shows the year-by-year total return of the GSCI from 1970 to 1991. Note that the contribution of the spot return was relatively small on average (2.59%) and extremely variable. The contribution of the T-bill yield was substantial and relatively less volatile. The contribution of the roll yield was positive though somewhat volatile.
The annualized average returns and standard deviations of return for the three components of the GSCI are shown in Table A for two different periods: the inflationary 1972–1980 period and the disinflationary 1981–1990 period. An important contribution to the total return in the 1981–1990 period came from the roll yield. Since the mid-1980s, the GSCI had held substantial weightings in energy futures (particularly oil) and livestock futures. In both of these types of commodity futures, the price structure was often in backwardation, which produced a positive roll yield. Some experts estimated that since crude oil futures first opened for trading in 1983 on the New York Mercantile Exchange, their prices had been in backwardation 80 to 85% of the time. The crude oil future had been included in the GSCI since 1984.
Table A Historical GSCI Returns and Standard Deviations
Spot Return
Roll Yield
T-bill Yield
Total Return
1972–1980
Annualized Return
12.6%
1.8%
7.3%
21.6%
Standard Deviation
24.0
6.1
0.8
23.2
1981–1990
Annualized Return
-0.9%
5.0%
8.7%
12.8%
Standard Deviation
15.2
4.4
0.8
15.4
1 This collateralized futures index portfolio is completely analogous to a fully collateralized synthetic S&P 500 index fund where the investor takes a long position in an S&P 500 future and collateralizes it fully with U.S. Treasury bills.
3
293-017 An Investment Linked to Commodity Futures
Commodities and Inflation
An attractive feature of the GSCI, or any investment tied to commodity prices, was its ability to act as a hedge against inflation (Exhibit 6). Both the Goldman Sachs Spot Index2 and the Goldman Sachs Commodity Index had a strong positive correlation with the two most widely used measures of inflation: the Consumer Price Index (CPI) and the Producer Price Index (PPI).
Commodities had long been used by investors to hedge portfolios with large amounts of inflation risk. Gold, most notably, was a safe haven because it retained value in times of economic and political instability. To preserve wealth, many European investors held gold after the world wars, when confidence in currencies decreased and hyperinflation sometimes prevailed.
The use of commodities as an inflation hedge worked especially well for two reasons. First, dramatic increases in commodity prices often triggered inflationary surges in an economy. The oil shocks of 1973–74 and 1978–79 reverberated throughout the worldwide economy. The supply shock increased not only gasoline and home heating oil prices for consumers but also raised raw material prices for industry. This in turn pushed consumer prices and, ultimately, wages into an inflationary spiral.
The second reason that commodity investments worked well as hedges against inflation was that commodity prices could react quickly to perceived monetary policy changes. For ple, the Federal Reserve (or other central banks around the world) could pursue an expansionary monetary policy by lowering interest rates. The necessary increase in money supply, many believed, would raise inflation — actual and expected — as more paper currency would soon be chasing after the same number of goods. These expectations would then be incorporated more or less instantaneously into spot and futures commodity prices around the world. Indeed, because of their extreme responsiveness, commodity prices were a good leading indicator of inflation. Economists often used the behavior of commodity prices to forecast future inflation rates.
Asset Allocation and an Investment in Commodities
A fully collateralized portfolio of commodity futures, such as that represented by the GSCI, would have been an interesting investment in its own right in the 1970–1991 period as these average returns and standard deviations shown in Table B suggest. More importantly, though, an investment in the GSCI had the potential to decrease the risk of the portfolio as well. Using the GSCI to decrease the risk of an investment portfolio worked well because returns on the GSCI were negatively correlated with returns on stocks and bonds.
Table B Comparative Returns and Standard Deviations, 1970–1991
GSCI S&P 500
Long-term U.S. Govt. Bonds
Annualized Return:
14.2%
11.0%
8.9%
Standard Deviation:
18.6
16.3
6.4
Over the 1970–1990 period, Goldman Sachs estimated that the correlation of quarterly GSCI returns with the S&P 500 was -.32, and the correlation of quarterly GSCI returns with U.S. Treasury
2 The Goldman Sachs Spot Index differed from the GSCI in that it measured the spot prices of the commodities in the GSCI.
4
An Investment Linked to Commodity Futures 293-017
bonds was -.19. In contrast, the correlation of the S&P 500 with U.S. Treasury bonds was +.39. Because the calculated returns on the GSCI were high and the correlation of the GSCI with stocks and bonds was negative, an addition of GSCI to an investment portfolio would not only have increased overall return but also would have decreased risk. For ple, by reallocating just 5% of a 60/40 stock/bond portfolio to an investment in the GSCI, the average returns over the 1970–1990 period would have increased from 9.6% to 9.8%, and the standard deviation of returns would have decreased from 12.1% to 11.5%. Exhibit 7 shows how the expected returns and standard deviations of a portfolio composed of the S&P 500 and the GSCI would vary for different sets of future return assumptions.
A look at portfolio theory, a quantitative optimization tool for determining how alternative asset classes might interact with each other in a portfolio, shows how an investment in the GSCI might be perceived. Standard computer-based portfolio optimization techniques tend to select for investment assets whose returns are assumed to be large and whose correlations are negative (or slightly positive). Obviously, if the GSCI were offered as a possible investment with its 1970–1990 returns, standard deviations, and correlations, any portfolio optimizer would have found it an extremely attractive asset class and would have recommended a very large position. Indeed, as long as the negative correlation of returns was assumed to continue into the future, a portfolio optimizer would include the GSCI as an attractive investment, even if the future returns were assumed to be very low (Exhibit 7).
An Alternative: Managed Futures Funds
There were, of course, other ways to participate in the commodity futures markets. Managed futures funds had grown in popularity with retail investors during the 1980s, as futures markets around the world expanded and as investors gained sophistication in using derivative instruments. A few pioneering institutional investors had begun to participate in these funds in recent years. Organized as limited partnerships, the managed funds invested in commodity as well as financial futures. The funds also took both long and short positions in the futures markets. Most fund managers used technical analysis (charting) to follow and attempt to ride the trend in any given market. Other fund managers had creative approaches based on weather forecasting, for instance, or used complicated computer modelling systems. In general, the portfolio positions of those funds changed dramatically from period to period. They were really more like trading pools that happened to use commodity futures as their main vehicle rather than an indexed investment in commodities.
In any given time period, performance for a handful of these managed funds seemed very impressive; for ple, some firms had generated 1991 returns of 30% or higher. Underlying these figures, however, were a maze of fees and charges amounting to over 15% of capital invested in some cases. Other complications included separating the good futures managers from the bad. There were literally hundreds of firms competing in the business, some registering 1991 performance records of worse than -60% return.
There were very few careful studies of the past performance of these funds. However, one well-regarded study (which ined a sample of these funds from 1980 to 1988) found that the average fund returned 2.36%. This could be compared with the GSCI's return of 7.66% over the same period (Exhibit 8).
293-017 An Investment Linked to Commodity Futures
Alternatives Linked to the Goldman Sachs Commodity Index
In addition to the zero coupon notes described in this case, Goldman Sachs offered a number of other products related to the index. For investors in the European market Goldman designed and issued a very similar product, a $100 million three-year zero coupon note issued by the Finnish Export Credit Corporation.
For investors interested in a more leveraged transaction (which would free up more of their principal), Goldman Sachs designed swap transactions against LIBOR (Exhibit 9). In these deals investors would pay Goldman Sachs LIBOR. In exchange, Goldman Sachs would pay investors the return on the GSCI minus 150 basis points. The 150 basis points covered transaction costs and tracking error in managing investments in the GSCI's commodity futures. Goldman Sachs conducted swap transactions on principal amounts of $5 million and more.
As of the spring of 1992, Goldman Sachs was continuing to innovate in the commodities market. It had begun to design and trade options on the GSCI and was also seeking to list GSCI- based futures and options on a U.S. exchange.
An Alternative Linked to the CRB Index
The Commodity Research Bureau (CRB) maintained an index called the CRB Futures Index. The New York Futures Exchange (NYFE) designed and offered futures and options contracts based on the value of this index. The index itself was an equal-weighted combination of the performance of
21 commodities futures, with relatively heavy representation from agricultural commodities, particularly grains.
Investing with this index was difficult, however, because the NYFE contracts had proved unpopular and illiquid. Also, the returns from this index had been substantially less in recent years than those from the GSCI. Because of its relatively heavy representation of grains (rather than energy futures), the roll yields in this index had often been quite negative. Exhibit 10 shows the 1984–1990 comparisons. While, in principle, some firms might design various collateralized commodity securities replicating the performance of the CRB Index (just as Goldman Sachs had done for the GSCI), there were no known efforts under way to do this.
6
An Investment Linked to Commodity Futures 293-017
Exhibit 1 Prospectus for the Zero Coupon Notes
image.png
293-017 An Investment Linked to Commodity Futures
Exhibit 2 GSCI Zero Coupon Note Redemption Values
Redemption Value = Par x [.9557 x GSCI End/GSCI Begin]
Redemption
