Modelling conditional correlations in the volatility of Asian rubber spot and futures returns

https://doi.org/10.1016/j.matcom.2010.07.004Get rights and content

Abstract

Asia is presently the most important market for the production and consumption of natural rubber. World prices of rubber are subject to not only to changes in demand, but also speculation regarding future markets. Japan and Singapore are the major future markets for rubber, while Thailand is one of the world's largest producers of rubber. As rubber prices are influenced by external markets, it is important to analyse the relationship between the relevant markets in Thailand, Japan and Singapore. The analysis is conducted using several alternative multivariate GARCH models. The empirical results indicate that the constant conditional correlations arising from the CCC model lie in the low to medium range. The results from the VARMA-GARCH model and the VARMA-AGARCH model suggest the presence of volatility spillovers and asymmetric effects of positive and negative return shocks on conditional volatility. Finally, the DCC model suggests that the conditional correlations can vary dramatically over time. In general, the dynamic conditional correlations in rubber spot and futures returns shocks can be independent or interdependent.

Introduction

Natural rubber is one of the most important agro-based industrial raw materials in the world. Rubber is produced entirely in developing countries. Asia is the largest producing region, accounting for around 96.6% of output in 2007, and Thailand is one of the world's biggest rubber producers. However, rubber prices are determined in the Singapore and Japanese markets. The factors involved in setting Thailand's rubber prices are quite interesting. According to the relevance of Thailand's rubber price to the Japanese and Singapore markets, it is important to examine the relationship between the Thai spot market and the three major global rubber futures markets, namely Tokyo Commodity Exchange (TOCOM), Singapore Commodity Exchange and Agriculture Futures Exchange (SICOM), and Osaka Mercantile Exchange (OME). In particular, volatility spillover effects will be considered across and within these markets.

Recent research has used the GARCH specification to model volatility spillovers across futures markets. The volatility transmission between futures and cash markets has received considerable attention in finance. Shocks in one market may not only affect the volatility in prices and returns in its own market, but also in related markets. Apergis and Rezitis [1] investigated volatility spillover effects across agricultural input prices, agricultural output prices and retail food prices, using GARCH models. Feng et al. [5] examined the inter-temporal information transmission mechanism between the palm oil futures market and the physical cash market in Malaysia.

Despite the recent developments in the multivariate GARCH framework, most of the research in agricultural futures markets has been confined to univariate GARCH specifications. It is well known that the univariate GARCH model has two important limitations: (1) it does not accommodate the asymmetric effects of positive and negative shocks of equal magnitude; and (2) it does not permit interdependencies across different assets and/or markets. Modelling volatility in a multivariate framework leads to more relevant empirical models than using separate univariate models in financial markets, wherein volatilities can move together over time and across assets and markets.

To date, few papers have paid attention to analyzing volatility spillovers across futures markets and physical cash markets in the context of multivariate GARCH models for agricultural commodity future markets. For example, Kim and Doucouliagos [6] examined volatility spillover effects by fitting a multivariate model to realized volatility and correlations. The dynamic relationships and causality among the volatilities and correlations of three grain futures prices, namely corn, soybean and wheat, were investigated by conducting impulse response analysis based on the vector autoregressive model.

The purpose of this paper is to (1) to model the multivariate conditional volatility in the returns on rubber spot and futures price in three major rubber futures markets, namely TOCOM, OME and SICOM and two rubber spot markets, Bangkok and Singapore, using several recent models of multivariate conditional volatility, namely the CCC model [2], DCC model [4], VARMA-GARCH model [7], and VARMA-AGARCH model [9], and (2) to investigate volatility transmissions across these markets.

The remainder of the paper is organized as follows. Section 2 discusses the econometric methodology. Section 3 explains the data used in the empirical analysis, and presents some summary statistics. The empirical results are analysed in Section 4. Some concluding remarks are given in Section 5.

Section snippets

Econometric methodology

This section presents models of the volatility in rubber spot and futures prices returns, namely the CCC model [2], VARMA-GARCH model [7], VARMA-AGARCH model [9], and DCC model [4] (see [8] for a comprehensive discussion and comparison of these models in terms of their mathematical and statistical properties). The typical specifications underlying the multivariate conditional mean and conditional variance in returns are given as follows:yt=EytFt1+εtεt=Dtηtwhere yt = (y1t, …, ymt)′, ηt = (η1t, …, η

Data

The alternative multivariate GARCH models are estimated using data on daily closing prices of spot and futures returns, and are expressed in local currencies for the period 23 September 1994 to 13 March 2009, giving a total of 3755 observations. All data are obtained from Reuters. The data set comprises 2 daily RSS3 spot prices, namely RSS3 F.O.B. spot price from Bangkok (TRSS3: Bath/kg.), RSS3 Noon spot price from Singapore (SRSS3: Singapore cent/kg.), and three daily RSS3 futures from

Empirical results

The empirical results of the unit root tests for all sample returns in each market are summarized in Table 1. The Augmented Dickey–Fuller (ADF) and Phillips–Perron (PP) tests are used to explore the existence of unit roots in the individual series. Both tests have the same null hypothesis to check for non-stationarity in each time series. The results show that all returns series are stationary. In order to see whether the conditional variances of the return series follow the ARCH process, the

Conclusion

Asia is presently the world's most important market for the production and consumption of natural rubber. World prices of rubber are subject to changes in demand and also to speculation regarding future markets. Japan and Singapore are the major futures markets for rubber, while Thailand is one of the world's largest producers of rubber. As rubber prices are influenced by external markets, it is important to analyse the relationship between the relevant markets in Thailand, which produces

Acknowledgements

The authors wish to thank Felix Chan and Abdul Hakim for providing the computer programs. For financial support, the first author wishes to thank the National Science Council, Taiwan, the second and fourth authors are most grateful to the Faculty of Economics, Maejo University, Thailand, and the third author acknowledges the Australian Research Council, National Science Council, Taiwan, and a Visiting Erskine Fellowship, College of Business and Economics, University of Canterbury, New Zealand.

References (9)

  • N. Apergis et al.

    Food price volatility and macroeconomic factor volatility: “heat waves” or “meteor showers”?

    Appl. Econ. Lett.

    (2003)
  • T. Bollerslev

    Modelling the coherence in short-run nominal exchange rate: a multivariate generalized ARCH approach

    Rev. Econ. Stat.

    (1990)
  • T. Bollerslev et al.

    Quasi-maximum likelihood estimation and inference in dynamic models with time-varying covariances

    Econ. Rev.

    (1992)
  • R.F. Engle

    Dynamic conditional correlation: a simple class of multivariate generalized autoregressive conditional heteroskedasticity models

    J. Business Econ. Stat.

    (2002)
There are more references available in the full text version of this article.

Cited by (0)

View full text