IFCP : Features : Cotton Polyester And Oil Prices

Revisiting the Relationship between Cotton, Polyester and Oil Prices

By Alejandro Plastina1. Originally published in Cotton: Review of the World Situation 63(3), Jan–Feb 2010.

Introduction

Textile fiber prices were not the exception to the commodity boom experienced in 2008. In particular, the prices of cotton and polyester, which account for 85% of the world textile fiber market (Fiber Economics Bureau, ICAC 2009) increased by 37% and 11%, respectively, during the 12 months previous to reaching their peaks2, and declined by 36% and 20% over the following 12 months. This event renewed the interest of the cotton and textile sectors in understanding the relationships between textile fiber prices and oil prices (World Bank 2008, Plastina 2008, 2009, 2010; Harri, Nalley and Hudson 2009).

Previous studies provide mixed conclusions regarding the price relationships between textile fibers and oil. A report by the Food and Agriculture Organization of the United Nations (FAO 2002) analyzes the relationships between oil and cotton spot prices, and between oil and polypropylene spot prices, and concludes that, at most, only weak links exist between oil prices and textile fiber prices.

Baffes and Gohou (2005) found a strong co-movement between cotton and polyester spot prices, and a significant effect of oil spot prices on polyester spot prices. The study also found that changes in the price of polyester are more rapidly transmitted to cotton prices than vice-versa: about half of the effects of a shock in the polyester market is transmitted to the cotton market within a 5-month period, while a similar transmission would take about 22 months in the other direction. The Baffes and Gohou article suggests that the difference in the speed of adjustment is a consequence of cotton being a primary commodity subject to both demand and supply shocks, while polyester is an industrial product subject mainly to demand shocks; and that while cotton prices are determined in the futures market, polyester prices are determined through contractual agreements.

Baffes (2007) analyzes the contemporaneous relationship between oil spot prices and the spot prices of other commodities with annual data for 1960–2005. The elasticity of cotton price with respect to oil price is found to be 14%, and the elasticity with respect to inflation 89%. The study also finds that the long-run relationship between annual-average oil and cotton prices is stable.

Fadiga and Misra (2007) found that (a) no long-run relationship between cotton and polyester prices exists, (b) no short-term direct relationship between cotton and polyester prices exists, (c) polyester prices respond to past polyester prices and oil prices, and that (d) cotton prices depend on past cotton prices and changes in cotton stocks. However, Fadiga and Misra (2007) also found that cotton and polyester prices have synchronous short-term cycles (i.e., the correlation between their cycle disturbances is high), so oil prices indirectly affect cotton prices (through polyester prices) in the short run.

Harri, Nalley and Hudson (2009), using monthly futures prices and cointegration techniques, found that between June 2004 and September 2008 a stable relationship existed between the levels of crude oil and cotton futures prices.

The objective of the present article is to explore the relationship between oil, polyester and cotton world spot prices and identify regularities that can be used for price risk management in the cotton and textile sectors.

Data

Figure 1. Real Prices of Cotton, Polyester and Oil

The data used in the present analysis consists of monthly observations of polyester, cotton and oil prices over the period January 1991 through November 2008. The selection of the sampling period was based on data availability.

Polyester prices are the midpoints of a range of prices reported by PCI Fibres3 as the world average of polyester staple prices, 1.5 denier (North America, Western Europe, and Asia, weighted by annual production volumes). Oil prices are the simple average of crude oil spot prices of Dated Brent, West Texas Intermediate, and Dubai Fateh, reported by the International Monetary Fund (IMF 2009). Cotton prices are the monthly average of the A Index, reported by Cotlook Ltd. All prices are expressed in U.S. dollars, and deflated by the U.S. consumer price index (1982–84=100). This price index was chosen on grounds of availability of monthly data. By deflating the series, the effect of inflation is removed from nominal prices, and the resulting real prices are expressed in currency units directly comparable through time. Monthly U.S. inflation ranged from -1.9% to 1.2% over the sample period, with an average of 0.2%. However, the accumulated inflation between April 1991 and November 2008 amounted to 57%. The analysis is conducted on real prices expressed in natural logarithms. The study follows standard time series econometric techniques for systems of equations (Johansen 1991, 1998).4

Results5

The following results are derived from the econometric analysis:

There is no stable relationship between the levels of polyester, cotton, and oil prices;
There is no stable relationship between the levels of cotton and oil prices;
There is no stable relationship between the levels of polyester and oil prices;
Changes in oil prices are not affected by changes in polyester or cotton prices;
A stable relationship between the levels of cotton and polyester exists: if the price of cotton in a given month amounts to more (less) than 104% of the price of polyester in that same month, then cotton prices will tend to adjust down (up) over the following months in order to return to the equilibrium. The speed of adjustment is 4% per month, i.e. 4% of the gap between the actual and the equilibrium level is closed each month;
Monthly changes in cotton prices depend on past changes in cotton prices and the adjustment term described in the previous paragraph;
Monthly changes in polyester prices depend on past changes in polyester, oil and cotton prices.

Figure 2. Response of Real Cotton Prices to Alternative Shocks

The effects of alternative price shocks on the prices of cotton and polyester are analyzed below. The estimated parameters from the econometric model are used to calculate the magnitude and speed of propagation of price shocks from a starting hypothetical situation in which real cotton and polyester prices have been normalized at 60 U.S. cents per pound and 51.3 U.S. cents per pound, respectively.

If oil prices increase by 10% in a given month, then:

Polyester prices will increase by 1.5% and 1.4% in the following two months, by 0.1% monthly over the following 13 months, they will decline by 0.2% monthly over the next two months and stabilize at a level 3% higher than the pre-shock level;
Cotton prices will increase at an average rate of 0.1% over the first 4 months, 0.2% over the following 5 months, 0.1% over the following 6 months, and they will stabilize at a level 2% higher than the pre-shock level.

Note that 88% of the total adjustment in polyester prices to a shock in oil prices occurs over the first two months, while the adjustment in cotton prices is gradual.

If polyester prices increase unexpectedly by 10% in a given month, then:

Polyester prices will increase by 2.9% in the following month, they will decline by 0.8% in the next month, they will increase by 0.3% monthly over the next 12 months, they will decline by 1.5% and 0.8% over the next 2 months, and they will stabilize in an oscillating pattern afterwards towards a level about 15% higher than the pre-shock level;
Cotton prices will increase by an average rate of 0.7% over the first 14 months, decrease by 0.2% monthly over the following 3 months and stabilize at a level 10% higher than the pre-shock level.

Note that shocks to polyester prices are self-perpetrating and the adjustment to their new long-run equilibrium is fast, while the adjustment of cotton prices to their new long run equilibrium is gradual and does not overshoot it.

If cotton prices increase unexpectedly by 10% in a given month, then:

Cotton prices will increase by an additional 5.2% and 0.6% in the following 2 months, and decline afterwards, to return – in an oscillating pattern – to levels about 2% higher than pre-shock levels 24 months after the shock;
Polyester prices will increase by 1.6% in the following month, by 0.2% monthly over the following 9 months (in an oscillating pattern), and decline over the following 10 months to stabilize at a level 1% lower than pre-shock levels.

Note that although a small proportion of the initial shock in cotton prices is transmitted to the new set of long-run equilibrium prices, the adjustment is gradual, and prices initially overshoot the new equilibrium.

Conclusions

Following the recent commodity boom, the relationships between textile fiber prices and oil prices received renewed attention. This article analyzes the relationships between cotton, polyester and oil prices with robust time series methods and no a priori restrictions on the estimating model.

Cointegration tests indicate that there is no stable relationship between the levels of cotton, polyester and oil prices, as well as no stable relationship between the levels of oil and cotton prices or between the levels of oil and polyester prices;
As expected, monthly changes in oil prices are found not to depend on polyester or cotton prices.
Monthly changes in cotton prices depend on past changes in cotton prices and the gap between the level of cotton prices and the level of polyester prices in the previous month (the equilibrium ratio is 1.04:1), but they do not significantly depend on past changes in polyester or oil prices;
A shock in oil prices shifts the equilibrium polyester price and this, in turn, shifts the equilibrium cotton price. However, only a small fraction of the shock in oil prices is transmitted to polyester and cotton prices. In particular, a 10% increase in oil prices results in permanent increases in cotton and polyester prices of 2% and 3%, respectively, and while the adjustment in polyester prices is fast, the adjustment in cotton prices is gradual and occurs over several months;
Shocks to polyester prices are rapidly propagated and shift the long-term equilibrium prices of cotton and polyester;
Shocks to cotton prices also shift the long-term equilibrium prices, but the propagation process is gradual and occurs over several months. This result suggests that recent shocks in the cotton futures and option markets that possibly originated from non-fundamental factors (Plastina 2008, 2010) – given their rapid transmission to the spot market (Plastina 2009) – might continue to affect spot prices in 2010.

One important implication for the cotton industry is that while price competitiveness can temporarily increase industrial demand for cotton, the gains from this strategy are rapidly eroded by their tendency to maintain equilibrium with polyester prices. Therefore, cotton promotion and demand enhancement efforts on the demand side, along with research and extension efforts to improve cotton yields and cotton fiber quality on the supply side, are the most promising strategies to increase the market share of cotton in textile fiber consumption in the long run.

Several caveats apply to the present study. First, the analysis is conducted on prices only and it does not explain the relationship between fundamental factors and prices. Second, the analysis is conducted in real terms, and any extrapolation to nominal terms must include the effects of inflation. Third, monthly data smoothes out price variability, and results could differ if higher frequency data were used.

References

Baffes, J. 2007. “Oil Spills on Other Commodities.” Policy Research Working Paper 4333, The World Bank.

Baffes, J., and G. Gohou. 2005. “The Co-movement between Cotton and Polyester Prices.” Policy Research Working Paper 3534, The World Bank.

Fadiga M. and S. Misra. 2007. “Common Trends, Common Cycles, and Price Relationships in the International Fiber Market.” Journal of Agricultural and Resource Economics 23(1): 154–168.

Fiber Economics Bureau, various issues. “Fiber Organon.” Washington, DC. Food and Agriculture Organization of the United Nations. 2002. “Oil Prices and Agricultural Commodity

Prices.” In Commodity Market Review 2001–2002. Rome. Harri, A., Nalley L. and D. Hudson. 2009. “The Relationship between Oil, Exchange Rates, and

Commodity Prices.” Journal of Agricultural and Applied Economics 41(2): 501–510.

International Cotton Advisory Committee. 2009. World Textile Demand. Washington DC.

International Monetary Fund. 2009. Primary Commodity Prices. Washington DC.

Johansen, S. 1991. “Estimation and Hypothesis Testing of Cointegrating Vectors in Gaussian Vector Autoregressive Models.” Econometrica 59: 1551–80.

---. 1988. “Statistical Analysis of Cointegration Vectors.” Journal of Economic Dynamics and Control 12: 231–54.

PCI Fibres. Fibres Report, various issues, available at .

Plastina, A. 2010. “Hedge Funds and Cotton Prices.” Cotton: Review of the World Situation 63(3): 15–22.

---.2009. “Cotton : Spot Prices Become more Responsive to Futures Prices.” Cotton: Review of the World Situation 63(1): 9–11.

---. 2008. “Speculation and Cotton Prices.” Cotton: Review of the World Situation 61(4): 8–12.

The World Bank. 2008. Global Economic Perspectives 2009: Commodities at the Crossroads. Washington DC.

Footnotes

1) Economist, International Cotton Advisory Committee. Email: alejandro@icac.org.

2) Cotton prices peaked in March 2008, and polyester prices in July 2008.

3) PCI Fibres is a consultancy firm to the fibers and related industries, specializing on the major manufactured fibers and raw materials for acrylic, nylon, polyester and viscose as well as related products.

4) The methodology is described in detail in the full version of the article, available online at www.icac.org/cotton_info/speeches/english.html.

5) Results in this section differ from a preliminary version of this study presented at the 2010 Beltwide Cotton Conference.