|
 Cotton
prices climbed to unexpected levels in February and March 2008
and the 2007/08 Cotlook A Index is now forecast at 74 US cents
per pound. In September 2007, the ICAC Secretariat forecast an
average Cotlook A Index of 68 US cents per pound for 2007/08. In
July 2007, the World Bank forecast the calendar year 2008
average Cotlook A Index at 59.5 US cents per pound. In June
2007, the Australian Bureau of Agricultural and Resource
Economics forecast an average Cotlook A Index of 59.2 US cents
per pound for 2007/08, while FAPRI's forecast, as of January
2007, amounted to 66 US cents per pound. The present article
attempts to shed some light on the causes of the unexpected
increase in cotton prices by analyzing (1) the flow of
investment funds to the cotton futures market, and (2) the
composition of the cotton futures and options markets by
traders, and their relation with cotton prices.
Investment Funds
 Although
the Standard & Poor's Goldman Sachs Commodity Index (S&P GSCI)
is not an index fund, several institutional investor funds, such
as the Oppenheimer Real Asset Fund or the iShares S&P GSCI
Commodity-Indexed Trust, try to replicate or beat the returns
calculated with the index. The S&P GSCI is designed as a
benchmark for investment in commodity markets and as a measure
of commodity market performance over time. The S&P GSCI is
calculated primarily on a world production-weighted basis and is
comprised of the principal physical commodities that are the
subject of active, liquid futures markets. The quantity of each
commodity in the index is determined by the average quantity of
production during the most recent five years of available data.
The production weights are designed to reflect the relative
significance of each of the constituent commodities in the world
economy while preserving the tradability of the index. The
composition of the index is reviewed on a monthly basis.
Cotton is included alongside wheat, corn, soybeans, sugar,
coffee and cocoa in the “agricultural commodities” category of
the S&P GSCI. The other categories comprise energy, industrial
metals, precious metals and livestock. Data on cotton lint
production are obtained from the Food and Agriculture
Organization of the United Nations (FAO), and the average of
cotton production for 2007 is calculated over the 2000-2004
period (Standard&Poor's S&P GSCI™ Report, June 30, 2007).
 The
estimated notional of benchmarked assets to the S&P GSCI, i.e.
the estimated institutional investor funds tracking the index,
grew from $60 billion in 2006 to $85 billion in 2007, and some
analysts project they will grow to over $100 billion in 2008.
The average share of the funds' assets channeled to cotton
futures traded in the Intercontinental Exchange (ICE) is
approximated in this article through the reference percentage
dollar weight (RPDW) [The reference percentage dollar weight of
cotton is the quotient of the reference dollar weight of cotton
divided by the sum of the reference dollar weights of all
commodities included in the index. The reference dollar weight
is the product of the contract production weight multiplied by
the average contract reference price. (S&P GSCI™ Index
Methodology 2008, available online at http://www2.
Standardandpoors.com/spf/pdf/ index/SP_GSCI_Index_Methodology_Web.pdf]
of cotton disclosed on the S&P GSCI Index Methodology annual
publications. Funds' investment in cotton futures is roughly
approximated as the product of the RPDW multiplied by the
benchmarked assets to the S&P GSCI. Funds' investment in cotton
futures approached $522 million in 2006, increased by $243
million to $765 million in 2007, and might climb to $950 million
in 2008. In order to put these figures in perspective, they are
compared to estimates of the value of U.S. cotton futures
contracts and options, which are approximated by the margin
value of the maximum total open interest in cotton futures
contracts and options traded in the ICE during the calendar
year. The margin value is calculated as the product of the
maximum total open interest in cotton futures and options over
the year, multiplied by the price of the futures contract (in US
cents/lb) on the date the maximum was registered, multiplied by
the size of the contract (50,000 lbs), multiplied by 0.13
(Margins are assumed to represent, on average, 13% of the price
of the nearby futures contracts). The margin value of cotton
futures and options more than doubled from $839 million in
 2006
to $1.8 billion in 2007, and it has increased to $2.1 billion
through February 26, 2008. Therefore, the ratio of funds'
investments in cotton futures contracts to the margin value of
cotton futures and options amounted to about 62% in 2006, but
due to a significant increase in the margin value of cotton
futures and options, it decreased to 42% in 2007. However, the
ratio might increase in 2008. It must be noted that these
figures only serve as a first approximation to the magnitude of
the flow of investment funds to the cotton futures and options
markets. In the next section, a more rigorous analysis of the
effects of speculators on those markets and on the price of
cotton is conducted. Finally, it should also be noted that
although the S&P GSCI might be the commodity index with more
benchmarked assets, it is not the only one. Other important
benchmark commodity indexes are the Rogers International
Commodity Index®, the Reuters/Jefferies CRB Index and the Dow
Jones - AIG Commodity IndexSM, in which cotton represents,
respectively, about 4%, 2.5% and 5% of all commodities included
in the indexes (RICI® Handbook 2007, available online at http://www.worldcommodityfunds.com/files/RICI_Index_Manual.pdf.
The Dow Jones - AIG Commodity IndexSM 2008 Commodity Index
Percentages, available online at http:// www.djindexes.com/mdsidx/index.cfm?event=showAigWeightings.
CRB Reuters/Jefferies Calculation Supplement 2005, available
online at http://www.jefferies.com/pdfs/RJCRB_Index_Calculation_Supplement.pdf)
Cotton Futures and Options Markets
In
order to get a broader picture of the cotton futures and options
markets, we next analyze the evolution of open interest in the
ICE disaggregated by traders' long positions. Data and
definitions were obtained from the Commitments of Traders (COT)
Reports by the U.S. Commodity Futures Trading Commission (CFTC)
[Available online at http://www.cftc.gov/marketreports/
commitmentsoftraders/index.htm]. Open interest is the total of
all futures and option contracts entered into and not yet offset
by an opposite transaction, nor fulfilled by delivery. Open
interest does not include open futures contracts against which
notices of deliveries have been stopped by a trader or issued by
the clearing organization of an exchange. Open interest held or
controlled by a trader is refe-rred to as that trader's
posi-tion. Clear-ing members, futures commission merchants, and
foreign brokers (collectively called reporting firms) file daily
reports with the CFTC. If, at the daily market close, a
reporting firm has a trader with a position at or above specific
reporting levels set by the CFTC, the Commission reports that
trader's entire position in all futures and options expiration
months in that commodity, regardless of size. The aggregate of
all traders' positions reported to the Commission (the
“reportable positions”) usually represents 70% to 90% of the
total open interest in any given market. The other 10% to 30%
are referred to as “non reportable positions.” When an
individual reportable trader is identified to the Commission,
the trader is classified either as “commercial” or
“noncommercial” according to specific regulations set by the
CFTC. All of a trader's reported futures positions in a
commodity are classified as commercial if the trader uses
futures contracts in that particular commodity for hedging. A
third category of traders under the reportable positions is the
“index traders” category. These traders are drawn from the
non-commercial and commercial categories, and include positions
of managed funds, pension funds, and other investors that are
generally seeking exposure to a broad index of commodity prices
as an asset class in an unleveraged and passively-managed manner
(non-commercial category); as well as positions for entities
whose trading predominantly reflects hedging of over-the-counter
transactions involving commodity indices (commercial category),
such as pension funds. Finally, in the Supplemental Reports to
the COT Reports, traders are classified in one of the three
categories: index traders, commercial traders, or non-commercial
traders. Therefore, we will refer to commercial traders as
“hedgers”, to non-commercial traders as “non-index-traders
speculators” and will keep the denomination of “index traders”
in this report.
Open interest in cotton futures and option contracts more than
doubled between January 2006 and February 2008. The average
daily open interest during 2006 amounted to 209,040 contracts,
with a standard deviation of 30,890 contracts. The average daily
open interest during 2007 was 59% higher than during the
previous year, amounting to 332,561 contracts, with a standard
deviation of 48,239. Finally, during the first two months of
2008, the daily average open interest amounted to 429,942
contracts, 29% higher than during 2007, with a standard
deviation of 34,843 contracts. The average share of long
positions held by speculators (index- and non-index traders) in
total open interest increased from 69% in 2006 to 73% in 2007
and to 74% in 2008, indicating not only an increase in the
absolute number of long positions held by speculators, but also
a more rapid increase in speculators' long positions than in
hedgers' long positions.
 Further
insight is gained by disagg-regating speculators' positions into
index traders' and non-index-traders speculators' positions.
Although both types of traders increased their long positions
over the last two years, non-index-traders speculators' long
positions grew faster than index traders' long positions. While
the average share of non-index-traders speculators' long
positions in total speculators' long positions was 50% in 2006,
it climbed to 65% in 2008. This suggests that index traders have
not been the main force behind the increase in open interest for
cotton futures and options, but non-index-traders speculators
have. Furthermore, using the spread as a measure of the extent
to which each speculator holds equal long and short positions,
it can be inferred that non-index-traders speculators were
increasingly bullish in recent months, since the average share
of spreads in total long positions for non-index-traders
speculators decreased from 64% in 2006 and 2007 to 59% in 2008.
Put it another way, non-index-traders speculators increased
their share of long positions not matched with short positions
over the last months. Another indicator of the sentiment of
speculators is the net position, which is calculated as the
difference between total long and total short positions. The
Figure illustrates the evolution of speculators' net positions.
Net positions held by index traders remained positive and more
than doubled between January 2006 and February 2008. On the
other hand, net positions held by non-index-traders speculators
showed greater variability and turned from positive to negative
in March 2006, to remain negative for most of the time until
June 2007, to return to positive and increasing values
thereafter. Therefore, non-index-traders speculators have been
instrumental in increasing total net speculative positions
contributing to bullish sentiments after June 2007.
 Using
a simple linear regression between nearby futures prices and net
speculative positions, we find that these measures are
positively correlated: a 1% weekly increase (decrease) in total
speculators' net speculative positions was associated with a
0.18% weekly increase (decrease) in nearby futures prices.
Finally, given the
positive and significant correlation between nearby futures
prices and the Cotlook A Index observed between January 2006 and
February 2008 (A 10% increase (decrease) in the nearby futures
price is associated with a 4.4% increase (decrease) in the
Cotlook A Index, according to a linear regression in
differential natural logarithms over the observed period with
weekly observations (R Squared = 0.398; Durbin Watson statistic
= 2.34). Furthermore,
 Granger
causality tests between the Cotlook A Index and the nearby
futures price in levels suggest that changes in the nearby
futures prices cause changes in the Cotlook A Index, and not the
other way around), we conduct an analysis of the effect of the
net speculative positions on the Cotlook A Index. A simple
linear regression indicates that a 1% weekly increase (decrease)
in total speculators' net speculative positions was associated
with a 0.14% weekly increase (decrease) in the Cotlook A Index.
Furthermore, using Granger causality tests on the levels of the
variables, it can be inferred that the direction of causality
goes from the change in the net speculative positions to the
change in the Cotlook A Index (over short periods of time, i.e.
weekly), and not the other way around (The null hypothesis “The
Cotlook A I ndex
does not Granger cause Net Speculative Posit-ions” cannot be
rejected at the 10% significance level with 1-6 lags. The null
hypothesis “Net Speculative Posit-ions do not Granger cause the
Cotlook A Index” is rejected at the 5% level of significance
with 2, 4, 5 and 6 lags, and at the 10% level of significance
with 1 and 3 lags). Since changes in total net speculative
positions were dominated by changes in non-index-traders
speculators' net positions after June 2007, we can infer that
the bullish sentiments among non-index-traders speculators have
contributed to pushing cotton prices up, above and beyond what
cotton supply and demand fundamentals suggest. According to the
ICAC Price Model 2007, which forecasts the Cotlook A Index based
on expected and past stocks-to-mill use ratios, the average
index for 2007/08 would be 67 US cents per pound. However, the A
Index climbed to 70 US cents per pound in December 2007, to 80
cents in February 20 08,
and to 90 cents in March, resulting in an average to-date of 71
US cents per pound (as of March 10, 2008). Given these observed
values of the Cotlook A Index, it is highly likely that the
forecast of the ICAC Price Model 2007 will fall below the actual
season-average for 2007/08.
Conclusions
Speculation has been a major contributor to the formation of
cotton prices in the short run over the last two years. In
particular, non-index-traders speculators seem to have played a
greater role than index traders in driving cotton prices higher
since June 2007. Therefore, the evolution of cotton prices has
not corresponded to what was expected from the available
information on cotton production, consumption, trade and stocks.
¨ |