Printer Friendly

Pricing efficiency in the Chinese NGM and GM soybean futures market.

SOYBEAN FUTURES MARKET IN CHINA

China is the world's largest non-genetically modified (NGM) soybean producer and importer. (1) In China's domestic market, soybeans are a very influential agricultural commodity used as a staple for human consumption, an important animal feed ingredient and a key material for the production of human consumable oil. The northeastern provinces of Heilongjiang, Jilin, Liaoning and the Inner Mongolia region comprise China's largest production base, supplying 60 per cent of domestic soybeans in China. Among these regions, Heilongjiang produced 7.48 million metric tons in 2005, about 40 per cent of the country's overall production. (2) As the government set a price floor for wheat and corn to make the relative domestic prices of these commodities more profitable than soybeans, farmers have expanded the wheat and corn acres and reduced soybean acres. This decline of soybean acres has been documented since 2004 (3) with production falling from 18.4 million metric tons to below 15.5 million metric tons from 2005 to 2007, leading to larger imports of soybeans from the United States and other countries. Statistics show that with the rapid growth of domestic soybean processing capacity and other pressures, the demand for soybeans in China has increased dramatically, pushing soybean imports up to 70 per cent of China's total consumption. (4)

Chinese regulators believed that separating the trading of genetically modified (GM) from NGM soybeans is a viable strategy for protecting domestic NGM production, offering NGM soybean growers a higher price and facilitating better marketing assistance. In China, NGM soybeans are commonly perceived to be healthier than GM soybeans in that GM soybeans are not perfect substitutes for NGM in either consumption or processing demand. In addition, the goal of the new trading system was to ensure a certain level of domestic soybean production capacity and a higher level of food security.

The Dalian Commodity Exchange (DCE) (5) is the largest agricultural futures exchange in China and the second-largest soybean futures market in the world after the Chicago Board of Trade (CBOT, a member of the CME Group). Given China's critical role in world soybean supply and demand, the DCE has grown into the largest NGM soybean futures market in the world; its soybean futures price is one of the leading indicator prices in world soybean markets. (6) The DCE, China's only soybean futures market, was established in February 1993. From 1993 to 2002, it was dominated by a small group of trading participants who were mostly speculators rather than hedgers. The low pricing efficiency of the DCE has long been a serious concern of the Chinese government, and the DCE was mandated by the government to implement many significant changes to improve efficiency. (7) With a low transaction volume, China's soybean futures market has experienced a remarkable transformation since the country joined the World Trade Organisation (WTO) in 2001. In 2002, the first year after the soybean contract was restructured, the annual trading volume of NGM soybeans exceeded 25.3 million contracts. The NGM trading volume rebounded to 226 million contracts in 2008 with a good harvest of soybeans in northeastern China and increased imports from the US. However, given Chinese farmers' average income and small cultivation land, few participate directly in soybean futures trading, although they generally obtain useful price information from the futures market and benefit indirectly from soybean futures trading.

In order to better understand the efficiency of the DCE, the focus of this research was to study the NGM price relationship between the DCE futures prices and forthcoming cash prices realised at the Zhengzhou Grain Wholesale Market (ZGWM) (8). To date, most published analyses about China's soybean futures and cash prices have focused primarily on trading footprints before the new system. That is, the period when there was no distinction between GM and NGM soybeans in the market and the two bean types were treated as perfect substitutes even though they were destined for distinctly different uses. (9,10) Williams et al. (11) discussed the development and efficiency of China's mung bean futures market in Zhengzhou Commodity Exchange. We did not test the embedded location option as proposed by Pirrong, Kormendi and Meguire (1994) (12) because the ZGWM is the biggest cash market and is of most interest to traders. Though the aforementioned studies agreed that futures from the DCE could be used to predict cash prices under different conditions, during certain periods and at different locations, these results do not represent the new price relationship and cannot be extrapolated to the NGM futures market. These studies do not include longer lags, such as the futures prices of the previous five to six months, even though those prices are particularly important for predicting futures cash prices in the DCE. This is because at five to six months before maturity, the contracts are most active and are thus more frequently traded. It is also important to point out that the efficiency of the markets with longer lags is of key importance to producers because it represents pre-planting deferment for key contract delivery points. Thus, the futures price can be used as a prediction tool to help make planting decisions whereas shorter lags, such as one or two months, represent updated expectations but cannot alter planting decisions as the seeds would have already been planted at this point. It is also important to note that previously published studies do not indicate a short-run relationship between the DCE futures price and the nationwide ZGWM cash price--the most important and influential futures-to-cash interaction in China. This then leads to an important question: what is the price relationship between the DCE and ZGWM cash market in the long and short run? Since all 15 delivery warehouses are located in the Dalian region, traders from other places must rely on railways to transport and deliver their soybeans to those warehouses. As the railway system is under the government's control, it is difficult for traders to anticipate whether they can secure freight space on trains to fulfil delivery, and how this transportation issue could affect the convergence of the futures price and cash price.

This study is timely because it supplements previously published studies by encompassing the new classification of GM and NGM contracts. In fact, since 2002, the Chinese soybean industry has undergone dramatic changes that restructured its size, management structure and information posting, which in turn create new challenges for industry participants. Institutional adjustment is carried out accordingly by the government to meet these structural changes. As a result, the futures market has become more active and the transaction volume has increased gradually. According to Liu, (13) the Chinese futures markets are seeking more influence to provide the cash markets with a potentially useful forecast indicator. The efficiency of soybean trading can serve as a benchmark to the trading of other commodities that have not been implemented. If soybean futures trading is successful and efficient, the government will be more confident and motivated in planning the trading of other important commodities. This study can thus provide policy implications of price efficiency for other agricultural commodities, such as corn and hogs. The management structure of the soybean futures exchange can also be applied to the trading of other strategic commodities including corn and hogs.

MARKET EFFICIENCY REVISITED

If a futures market is efficient--that is, the futures price is an unbiased predictor of upcoming cash price--then agents are able to mitigate potential losses by using appropriate hedging instruments. If the futures price is an unbiased predictor, then soybean producers can use the expectation signals embedded in the futures price to adjust their planting, marketing and risk management activities. An efficient futures market also allows processors and traders to better manage inventory decisions and hedge against market risks. An understanding of the implications of futures and cash price integration enables policy-makers to more effectively monitor speculation and hedging behaviour in order to enhance market liquidity and efficiency, guide market participation, implement better agricultural policies and boost infrastructure investments. With China's expanding influence in the global soybean market, an insight into the efficiency of the DCE futures market helps guide international investors and exporters in their trading decisions and provides essential price information.

Given the fact that futures markets have the ability to provide producers and other industry participants with relevant price signals, previous attempts to analyse the properties of agricultural commodity futures markets have focused on the effectiveness of these markets as a price discovery mechanism and hedging instrument. However, the differences in institutional organisations and government regulations between developed and developing economies could give rise to different outcomes in terms of market performance and outcomes of theory-based analyses in our current application to China.

According to the literature, the futures price is viewed as such a reliable indicator of forthcoming cash prices that market participants use the futures market to shift price risk to others or use the futures price signals to forecast some cash market prices. The significance of both contributions depends upon the convergence of the futures and cash prices. A strong futures cash relationship has been detected for storable commodities using data gathered from developed countries. (14,15,16) The development of efficient market hypothesis studies has focused primarily on the efficiency of futures markets (17,18) and many studies developed dynamic models to understand the futures-cash price linkages. These studies generally concluded that current futures prices and cash prices can converge. (19,20,21,22,23,24,25,26,27,28)

Cointegration between cash and futures prices is a necessary condition for market efficiency when the futures and cash prices series are stochastic. (29) However, studies about futures markets vary in their conclusions because of the different data sources and price level used, types of futures market and analytical methods. (30,31,32,33,34,35,36)

For example, Zulauf et al. examined the forecasting performance of December corn and November soybean futures contracts using price-level and a percentage-change model and found conflicting results from the two models. Empirical evidence remains unclear about the relationship between the long and short-run future cash price movement. (37,38)

Although copious literature confirms the efficiency of Western futures markets, these results might not be true in developing country scenarios, such as China and the DCE, given the differences in market structure, level of participation and government regulation or oversight. Also, it is expected that in the early stages of market development there may be relatively low efficiency.

Employing classic ordinary least squares (OLS), cointegration and error-correction model (ECM) methods to establish a testing framework, this study aimed to understand the following questions: (a) whether NGM soybean futures prices at the DCE were integrated with local cash prices; (b) how strong the integration between the two prices was; and (c) which price series was more likely to react to an external price shock to restore market equilibrium. The remainder of this article is divided into four parts as follows: the methodology and hypothesis testing, specifics of data sources, discussions of the results, and conclusions.

METHODS

A simple regression model was estimated to understand the equilibrium relationship between futures and cash prices and to test the unbiasedness of futures price series to predict cash prices:(39,40,41)

[C.sub.t] = [[beta].sub.0] + [[beta].sub.1][F.sub.t-i] + [[zeta].sub.t] (1)

where [C.sub.t] denotes the cash price at time t when the futures contract matures, [F.sub.t-i] denotes the published futures price for the time t contract, i denotes the months before maturity, and [[zeta].sub.t] denotes an error term. Futures markets are considered efficient if the futures price is an unbiased forecast of the forthcoming cash price, that is, if [[beta].sub.0] and [[beta].sub.1] are estimated to be zero and unity, respectively. The R2 from Equation (1) is the percentage of variation in the cash price explained by the futures price and is an indication of forecast efficiency.

Following Motamed et al., (42) and Carter and Mohapatra, (43) we then used a vector of prices in logarithmic form [P.sub.t] = ([P.sub.1t], [P.sub.2t] ... [P.sub.pt]), which is assumed to be generated by a [k.sub.th]-order vector auto-regression (VAR) model:

[P.sub.t] = [[beta].sub.1][P.sub.t-1] + ... + [[beta].sub.k][P.sub.t-k] + [mu] + [[epsilon].sub.t] (2)

The vector [P.sub.t] contains both the logarithms of cash and futures prices over the selected time horizon, where p = 1 in our study, and [mu] denotes a vector of constants. We used the natural logarithm of prices to correct for any skewing in the distribution of prices that often arises from the non-negativity of prices. In addition, the differences of the logarithms can be interpreted as returns. (44) Stationarity of Equation (2) implies that [[epsilon].sub.t] has constant first and second moments over time series. If the first-order difference ([P.sub.t] - [P.sub.t-1]) is stationary, this time series is said to be integrated of order one, often denoted by 7(1). According to Lai and Lai, (45) the cointegration between cash and futures prices is a necessary condition for market efficiency along with the cointegration vector [[beta].sub.1], ... [[beta].sub.k] and vector of constant [mu] which equals to unity and zero, respectively. However, one could imagine a circumstance where nonstationary transactions costs that are not accounted for in a model could result in a model of Equation (2) that does not identify cointegration while the time series are in fact cointegrated. In the case where cash and futures prices are not cointegrated, [[epsilon].sub.t] is nonstationary and the time series may deviate apart without bound (46) and no long-run equilibrium exists between the two prices. Such long-run relationships usually reflect the profit or utility-taking incentives of economic agents. In the case of futures and cash prices, when they deviate substantially, a point of disequilibrium will eventually be reached where producers, processors, and/or speculators are motivated by profit opportunities to buy or sell one or both of the assets; by doing so, the two prices would converge towards their long-run equilibrium.

Before conducting the cointegration test, each individual price series should be tested for stationarity. Commonly used tests include the augmented Dickey-Fuller (ADF) test and the Phillips-Perron unit root tests. (47,48,49) This study uses ADF test and Phillips and Perron tests to examine the stationarity of cash and futures prices. If the test confirms that futures prices and cash prices are not stationary, and that the first difference of each individual series is stationary, then the time series are integrated of order 1 (I(1)), and we then choose Johansen's methodology to test cointegration. (50) The VAR representation can be expressed in first order differences and lagged levels:

[DELTA][P.sub.t] = [[GAMMA].sub.1][DELTA][P.sub.t-1] + ... + [[GAMMA].sub.K-1][DELTA][P.sub.t-k+1] + [mu] + [[epsilon].sub.t] (3)

When the rank of the matrix [PI] takes the value of r, the system contains r cointegration relations, where 0 < r < p, with p denoting the number of time series involved; in this case p = 2 denotes the cash and a series of lagged futures prices. In our test of the futures and cash price cointegration, the null hypothesis should be tested for r = 0 and r = 1. If r = p - 1 = 1, the evidence points to a cointegration relationship between cash and futures prices. If r = 0, no long-run relationship appears to unite futures and cash relationship and the two price series move independently. To find r, the eigenvalues of [PI] were computed to identify the rank at which the eigenvalues were no longer statistically different from zero, using the approach of Johansen and Juselius. (51) The two test statistics to test the null hypothesis that there are at most r cointegration vectors were based on the trace and maximum eigenvalues of [PI]:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (4)

where T represents the sample size, and [??]; is the estimated value for the [i.sub.th] ordered eigenvalues from the [PI] matrix. In addition, [PI] = [alpha][beta]' where [alpha] represents the short-run speed adjustment coefficients and [beta] represents the long-term cointegrating vector. To select lag length for the system, Akaike information criterion was used. Specification of trends in the VAR and the error-correction portion of the model were based on graphs of price series over time that suggested no trending behaviour. Thus, trend was omitted.

While the Johansen cointegration procedure was used to test for long-run market unbiasedness, the short-run price dynamics had to be analysed by an error-correction model (ECM). The Granger representation theorem states that a time-series model of a cointegrated series can be rewritten in an error-correction form. Such a transformation renders the series stationary, and allows for normal hypothesis testing. The ECM, which incorporates the long-run relationship between the two price series while allowing for short-run dynamics, was specified as follows:

[DELTA][C.sub.t] = [lambda] + [rho][e.sub.t-i] + [beta][DELTA][F.sub.t-i] + [[summation].sup.m.sub.i=2][[beta].sub.i][DELTA][F.sub.t- i] + [[summation].sup.k.sub.j=i] [[psi].sub.j] [DELTA][C.sub.t-j] + [v.sub.t] (5)

In Equation (5), A is defined as the change or difference in a variable from one period to the next; [e.sub.t-i] denotes the error-correction term; [rho] denotes the speed adjustment coefficient; and [v.sub.t] denotes a stationary series.

Long-run unbiasedness implies that [[beta].sub.0] = 0 and [[beta].sub.i] = i in the cointegrating regression as specified in Equation (1). The assumption is that [[beta].sub.0] = 0 and [[beta].sub.1] = 1 can be tested using the Johansen multivariate cointegration procedure. This approach estimates the likelihood ratio tests for restrictions on the parameters of the cointegrating regression. The Engle-Granger (52) two-step cointegration procedure cannot be used to test these restrictions, as the test procedure does not have well-defined limiting distributions. Using the Johansen multivariate procedure, rejection of the hypothesis of [[beta].sub.0] = 0 and [[beta].sub.1] = 1 would imply either market inefficiency or the possible presence of a risk premium in the futures market. Futures markets containing a risk premium would be biased, but could still be efficient. In other words, such markets would impound information efficiently while simultaneously taking into account the risk premium. If a futures market is deemed to be unbiased in the long run, the concepts of short-run unbiasedness and market efficiency are synonymous. If the long-run unbiasedness assumptions that [[beta].sub.0] = 0 and [[beta].sub.1] = 1 are held, short-run unbiasedness requires the restrictions [rho] = -1, [beta] = 1 and [[beta].sub.i] = [[psi].sub.i] = 0 (53) from Equation (5) to hold. So the final ECM equation should be derived as follows:

[DELTA][C.sub.t] = [lambda] - [rho][e.sub.t-1] + [beta][DELTA] [F.sub.t-1] + [v.sub.t] (6)

DATA

The cash prices used here are NGM soybean prices from ZGWM and include weekly average data from March 2003 to January 2010. ZGWM was China's first and is the largest soybean trading cash market located in Zhengzhou, Henan Province where 40 per cent of China's soybeans are produced. Soybean cash prices published by ZGWM are widely used by various local marketers to adjust their cash prices.

The soybean future prices used here were from the DCE. Given that the termination of NGM soybean contracts occurs on the 10th trading day of each maturity month of January, March, May, July, September and November, we used the closing price on the second Friday of that month (each year there are six NGM soybean contracts traded in DCE: January, March, May, July, September and November). If an observation is included from a regular trading month but not a contract maturity month, the observation was the closing price on the second Friday of that month. This way, we could link the observations over time and all observations were equally apart in time. We started with observations from March 2003, the second year after the new contract, in order to minimise the impact from the initial market adjustments in 2002. We included futures prices one week prior to the cash price to analyse the impact of nearby futures on cash movement as well as futures prices taken one to six months prior to the cash price to understand longer-run price relationships. For China's soybean futures market, the most heavily traded contract is always more than six months ahead. For example, if it is June now, the most heavily traded contract in CBOT/CME is usually the July contact. But in the DCE, the most heavily traded contract is A1201, which will expire in January 2012. So the Chinese market roll-over happens six months to one year before the maturity month. But in the US, the market roll-over always happens about eight working days before the expiration dates. In the DCE market, roll-over happens over a much longer period than the US, usually more than one month. Thus, the price change caused by roll-over over a longer time in China does not cause very significant price variation.

This study aims to examine the change of futures price over a longer period of six months and its impact on cash movement. We used eight price series of cash, one-week lag futures and one to six-month lag futures prices with each series containing 42 observations. Although GM soybean futures contracts also began trading since 2009 as the "Yellow bean #2" contract in the DCE, the trading volume was so low that the prices of those GM soybean futures trading provided little reliable information, and thus not were not used in our analysis. Comparisons of NGM ("Yellow bean #1") and GM ("Yellow bean #2") soybean futures contracts are listed in Table 8.

RESULTS

We begin with a descriptive analysis of the relationship implied by Equation (i) using data on futures and cash prices from 2003 to 20i0. In Figures i and 2, the average of published cash prices are plotted and compared with earlier published futures prices. A visual inspection of the figures shows that the lagged futures prices appeared to converge with the ZGWM cash prices. Before July 2007, the futures prices and cash prices moved in tandem. In some months, cash prices rose slightly above futures. For example in May 2004, cash prices were about USD1.7 per bushel above the historical futures prices, which were revealed three and five months earlier. In some other months, such as November 2005, historical futures prices five months earlier were about USD1.7 per bushel higher than realised cash prices. The gap widened from November 2007 to March 2008 when cash prices were on average USD3.8 per bushel higher than futures prices published three and five months before, when both the US and China experienced severe drought in 2007. More recently, starting from March 2009, the two prices moved very closely, which may suggest that soybean futures prices are probably better predictors of cash prices for closer months. Summary statistics of selected cash and futures prices are listed in Table 1.

To formally test the unbiasedness of futures prices, we used Equation (1) and assumed that both price series were stationary. As indicated earlier, if futures prices and cash prices converge, then the estimated values for (30 and b1 should be zero and unity. We tested these hypotheses using individual and joint F-tests and found that for the seven estimations, the slope coefficients were not statistically different from one, and the intercept estimates were not statistically different from zero (Table 2). Thus, we concluded that the soybean futures price was a useful price forecast for cash prices up to six months later.

The [R.sup.2] for the one-week period was high and decreased as the forecast horizon increased, indicating that the predicting power declined the further the futures price deviated from the cash price (which is as expected), but despite that, the futures price published six months before had reasonably strong explanatory power.

Before conducting the Johansen cointegration test, two tests were conducted based on the VAR model of Equation (3) to examine the nonstationary properties of the prices: the augmented Dickey-Fuller (ADF) test and the Phillips-Perron test (Table 3). The ADF test uses additional lags of the first-differenced variable to account for serial correlation and the Phillips-Perron uses Newey-West bandwidth standard errors and a Bartlett kernel spectral estimation method to control serial correlation. Preliminary data examination did not show a significant trend of the variable so we tested the model with a constant but no trend. We used the Akaike information criterion to determine the number of lags to use. Our results showed that the cash prices and the lagged futures prices one week and one to six months prior contained unit roots. However, the first-differenced prices were stationary at statistically significant levels (1 per cent). We can thus conclude that our price series were I(1) (i.e., each series was integrated of order one).

As the price series at various lag levels could not be shown to be stationary and the unit root test was used to check the integration of each price series, we then used the Johansen cointegration test to examine the cointegration relationships between the futures prices and cash prices. Table 4 presents the results of the trace and maximum eigenvalue test statistics. Our results suggest that futures prices taken at one week, or one to six months prior to cash prices are cointegrated with the upcoming ZGWM average cash prices at a statistically significantly level ([alpha] = 0.1), which implies an integration between ZGWM cash and DCE futures prices. A high cash deposit of 50 per cent is required right before delivery when the exchange has already finished matching buyers with sellers, the contemporaneous futures price is less likely to influence cash price in this case and that lagged futures are more reasonable predictors of upcoming cash prices. Thus we used models that associate lagged futures with current cash prices. The DCE has very strict requirements for delivery. Only institutional traders are allowed to hold the contract to participate in the delivery. Non-institutional traders are not allowed to participate in delivery at all. Thus, given the exclusion of all non-institutional traders from trading in the maturity month, transaction volume becomes very low before delivery. Larger institutional traders are able to manipulate the price more easily. One of the purposes of this study was to test whether cash and futures price will converge in the maturity month, as this reflects the efficiency level of the market.

Having established that the DCE futures prices are cointegrated with ZGWM local cash prices, we turned to test the long-run coefficients, [beta]', to examine whether each futures price series could form a long-run relationship with the later revealed cash prices (Equation (5)). We referred to the Johansen method and used normalised cointegration vectors on futures prices (54) to test if the futures prices were unbiased forecasts of forthcoming cash prices (i.e., whether the cash price was equal to futures prices and a constant). We found that cointegration coefficients on the futures price series were estimated to be close to -1. We then tested the model for exclusion restriction, that is, whether cointegrating coefficients are zero, and we detected that they deviate significantly from zero and all the seven selected futures price series are linked to the upcoming ZGWM cash prices in the long run-up to six months (Table 5).

Finally, we tested the unbiasedness hypothesis (Equation (6)). The estimates for the ECM model are tabulated in Table 6. The results showed that the coefficient of error correction term, p, was greater than 0.90 (in absolute value) for the one-week lag estimation only. In other equations, the estimated coefficients were between -0.50 and 0.07 (null hypothesis: [rho] = -1). Both independent test and joint tests showed that the null hypothesis could not be rejected at a 5 per cent level for the one-week lag estimation, indicating that there is no bias between futures prices and cash prices in the short term. In other words, short-term futures price had significant influence on the cash price revealed a week after. The hypothesis was rejected for all other estimations, suggesting futures price had no significant influence on cash price such that using futures to predict cash price was biased for longer-term estimations. The value of p also indicated the speed of futures price adjusting back to equilibrium. We found that the futures prices taken one week prior appeared to adjust to external shocks most quickly ([alpha] = -0.90).

Using sample data from 2003 to 2010, we detected that futures prices taken one to six months prior to cash prices responded rapidly to exogenous price shocks with a magnitude of adjustment of more than 49 per cent of the previous month's equilibrium error. This may be due to the difference in time lags between cash and futures. We found that a shock in the past was felt only by the futures (not cash) prices. Given the fact that the cash price was observed later when the shock has abated, the shock affects the cash price through its adjustments to the previous futures prices published one period ahead. Thus, the cash and futures prices were able to eventually converge to reach the market equilibrium. Our results suggest that China's futures trading has grown in influence to affect the cash movement. This may be attributed to the implementation of a number of futures market reforms, which have led to more trading volume and a more efficient futures market. Cash traders have started using the signals released from the futures market to form price expectations.

CONCLUSION

This study found that the NGM soybean futures market is integrated with the most important cash market in central China. Though an active cash market is key to the existence of a successful futures contract, (55) China's market has grown extensively to facilitate an effective futures market. The Chinese government has been actively consulting academic expertise to gain an understanding of the soybean futures cash linkages and the results in this study will provide important policy implications. Williams et al. (56) highlighted that the "Chinese exchanges have had to contend with a central government often concerned about individual commodities and retaining a healthy scepticism about the role of commodity exchanges more generally". With the experience of regulating soybean trading, the Chinese government and DCE should be able to respond to issues on delivery problems, prevention of manipulation, and violations of position limits. The DCE can also better monitor and control excessive speculations so that the futures market will be regarded as trustworthy by various potential participants. A previous study showed that China's soybean market suffers from over-speculation, suggesting an inefficiency of China's soybean futures trading. (57) Our results revealed that the efficiency of soybean trading has improved in recent years and that the futures prices provide useful signals to the formation of forthcoming cash prices. The proven efficiency of the soybean market can provide government planners concrete evidence and greater confidence to initiate the launch of futures trading for other commodities. China has about 20 years of experience in trading soybeans, soybean meal and soybean oil, which are strategic agricultural commodities. Thus, soybean trading is very closely mandated by the government and tracked by the industry traders and growers. If the soybean trading is successful and efficient, the government will be more confident and motivated to carry out the trading of other important commodities including hogs, fruits and coal.

The government may want to consider allowing foreign traders and investors to trade in the DCE given the fact that it has established an effective price relationship with the local cash market, and that foreign entities could provide additional liquidity and risk transfer potential to China's commodity markets.

For millions of Chinese soybean growers, the DCE's efficient NGM soybean futures market provides an accessible and useful price signal to guide production, marketing and risk management decisions. The trading of standardised contracts for soybeans to be delivered at a later date is still a new concept to many Chinese growers. China's agricultural extension programmes have provided extensive education to help growers understand the mechanics of futures markets and how futures prices could be linked with recent cash prices. Establishing the efficiency of the NGM futures market is important to engender an understanding of basis variation, which could be an important future research topic. Basis variation is linked directly to supply and demand factors in the marketplace, and with the use of historical information, producers may establish reasonable expectations about their local basis.

Chinese soybean traders and processors should view an efficient future price as a timely price forecast of local prices when making inventory management and marketing decisions. For international soybean growers, traders and processors, the DCE's efficient soybean futures and cash relationship will generate a stronger interest in Chinese futures trading as a mechanism to hedge international transactions and against variations in their local markets which may arise from growing Chinese demand for imported soybeans. As the world's largest soybean importer, China's high demand means that many foreign growers cannot ignore price signals from China when making important production and marketing decisions.

Finally, an efficient soybean futures market will help foreign governments, researchers and investors better understand Chinese soybean markets. The US Department of Agriculture (USDA) keeps a very close watch on Chinese price signals to provide US producers with useful information about demand forecasts.

An efficient DCE futures market may help the USDA to make better predictions. It may also help US producers and marketers make better and informed decisions with confidence.

ACKNOWLEDGEMENTS

This research was supported by the Fundamental Research Funds for the Central Universities and the Research Funds of Renmin University of China (Mingde training programme, no. 10XNJ020), and National Natural Science Foundation of China (70773115). The authors are grateful to Professor Ken Foster, Dr. Xu Pei, Fengyu Hu, Linjie Peng and Mo Li for their assistance in data collection, descriptive and econometric analysis, as well as other constructive suggestions.

(1) Futures Industry Association, Annual Report 2008, at <http://www.futuresindustry.org> [30 July 2010].

(2) China Futures Association, Annual Report 2008, at <http://www.cfachina.org> [12 July 2010].

(3) Farming Network, "The Impact of Low Soybean Production in China", at <http://www.yz006.com/ news/201007/11289_26.html> [8 Apr. 2011].

(4) Xinhua News Agency, "China to Restrict Soybean Imports", 2010, at <http://news.xinhuanet.com/ english2010/business/2010-04/01/c_13233763_3.htm> [1 July 2010].

(5) To date, soybean futures investors include nearly 200,000 companies and 189 corporate members. See Dalian Commodity Exchange, 2010, at <http://www.dce.com.cn/portal/info?cid=1261730307128&iid= 1277106779100&type=CMS.STD> [20 Oct. 2013]. For NGM soybean contracts, this small group of corporate clients accounts for about 10 per cent of total transactions and 42 per cent of total positions. Corporate clients are characterised by low transaction frequency but long position holdings; see Liu, "Develop Futures Markets to Gain a Bigger Say in Bulk Commodity Markets".

(6) J. Areddy, "China Nurtures Futures Markets in Bid to Sway Commodity Prices", The Wall Street Journal, Economy, 12 Oct. 2009, at <http://online.wsj.com/article/SB125529874012778991.html> [10 Dec. 2009].

(7) (a) Trading rules and clearing rules are better regulated to enhance transaction efficiency; (b) the soybean futures margin was redefined to be 5 to 20 per cent of total contract value and a flexible system to manage the margin was introduced; (c) a more flexible rolling delivery was launched. During the delivery month, the seller holding a standard warehouse receipt offers first to sell his position and then the exchange arranges the seller and buyer to complete the delivery within the stipulated time; (e) the commission charge was reduced from 6 to 4 Chinese yuan (or USD0.92 to USD 0.62; where USD1 = 6.5 Chinese yuan, as of 3 Apr. 2011) per transaction, in order to encourage participation; (g) the exchange's use of warehouse facilities was expanded to 15 NGM soybean warehouses and nine GM soybean warehouses; and (h) software and hardware facilities were greatly improved.

(8) Zhengzhou Grain Wholesale Market, 2011, at <http://www.czgm.com> [1 Apr. 2012].

(9) Wang Z.Q., Xu F.Y. and Zhu L.H., "Efficiency Tests of Futures Prices in Dalian Commodity Futures Market in China" (in Chinese), Research on Financial Economics Issues, no. 2 (1998): 25-33.

(10) Wang H.H. and Ke B.F., "Efficiency Tests of Agricultural Commodity Futures Markets in China", The Australian Journal of Agricultural and Resource Economics 49 (2005): 125-41.

(11) J. Williams, J.A. Peck, A. Park and S. Rozelle, "The Emergence of a Futures Market Mungbeans on the China Zhengzhou Commodity Exchange", Journal of Futures Market 18 no. 4 (1998): 427-48.

(12) S.C. Pirrong, R. Kormendi and P. Meguire, "Multiple Delivery Points, Pricing Dynamics, and Hedging Effectiveness in Futures Markets for Spatial Commodities", The Journal of Futures Markets 14, no. 5 (1994): 545-73.

(13) Liu, "Develop Futures Markets to Gain a Bigger Say in Bulk Commodity Markets".

(14) W.G. Tomek and R.W. Gray, "Temporal Relationships among Prices on Commodity Futures Markets: Their Allocation and Stabilizing Roles", American Journal of Agricultural Economics 52 (1970): 372-80.

(15) B.L. Gardner, "Futures Prices in Supply Analysis", American Journal of Agricultural Economics 58 (1976): 81-4.

(16) Kenneth D. Garbade and William L. Silber, "Price Movements and Price Discovery in Futures and Cash Markets", The Review of Economics and Statistics 65, no. 2 (1983): 289-97.

(17) E.F. Fama, "Efficient Capital Markets: A Review of Theory and Empirical Work", Journal of Finance 25 (1970): 383-417.

(18) E.F. Fama, "Efficient Capital Markets II", Journal of Finance 46 (1991): 1575-617.

(19) J. Chavas, R. Pope and R. Kao, "An Analysis of the Role of Futures Prices, Cash Prices, and Government Programs in Acreage Response", Western Journal of Agricultural Economics 8 (1983): 27-33.

(20) C.C. Cox, "Futures Trading and Market Information", Journal of Political Economy 84 (1986): 1215-37.

(21) J.S. Eales, B.K. Engel, R.J. Hauser and S.R. Thompson, "Grain Price Expectations of Illinois Farmers and Grain Merchandisers", American Journal of Agricultural Economics 72 (1990): 701-8.

(22) D. Kenyon, E. Jones and A. McGuirk, "Forecasting Performance of Corn and Soybean Harvest Futures Contracts", American Journal of Agricultural Economics 75 (1993): 399-407.

(23) T.R. Fortenbery and H.O. Zapata, "An Examination of Cointegration Relations between Futures and Local Grain Markets", The Journal of Futures Markets 13, no. 8 (1993): 921-32.

(24) S.E. Beck, "Cointegration and Market Efficiency in Commodities Futures Markets", Applied Economics 26, no. 3 (1994): 249-57.

(25) K. Foster, A. Havenner and A. Walburger, "System Theoretic Time-Series Forecasts of Weekly Live Cattle Prices", American Journal of Agricultural Economics 77, no. 4 (1995): 1012-23.

(26) S. Daniel, T. Schroeder and K. Dhuyvetter, "Forecasting Performance of Storable and Non-Storable Commodities", Proceedings of the NCR-134 Conference on Applied Commodity Price Analysis, Forecasting, and Market Risk Management, 1999, Chicago, IL, at <http://www.farmdoc.uiuc.edu/nccc134> [10 July 2010].

(27) A.M. McKenzie, B. Jiang, H. Djundaidi, L. Hoffman and E. Wailes, "Unbiasedness and Market Efficiency Tests of the U.S. Rice Futures Market", Review of Agricultural Economics 24, no. 2 (2002): 474-93.

(28) C.A. Carter and S. Mohapatra, "How Reliable are Hog Futures as Forecasts?", American Journal of Agricultural Economics 90, no. 2 (2008): 367-78.

(29) K.S. Lai and M. Lai, "A Cointegration Test for Market Efficiency", Journal of Futures Markets 11 (1991): 567-75.

(30) E. Elam, "Reduction in Hedging Risk from Adjusting for Autocorrelation in the Residuals of a Price Level Regression", Journal of Futures Markets 11, no. 3 (1988): 371-84.

(31) C. Zulauf, S. Irwin, J. Ropp and A. Sberna, "A Reappraisal of the Forecasting Performance of Corn and Soybean New Crop Futures", Journal of Futures Markets 19, no. 5 (1999): 603-18.

(32) Robin J. Brenner, and Kenneth F. Kroner, "Arbitrage, Cointegration, and Testing the Unbiasedness Hypothesis in Financial Markets", Journal of Financial and Quantitative Analysis 30 (1995): 23-42.

(33) M.T. Holt and A.M. McKenzie, "Quasi-Rational and Ex Ante Price Expectations in Commodity Supply Models: An Empirical Analysis of the US Broiler", Journal of Applied Econometrics 18, no. 4 (2003): 407-26.

(34) He Dequan and Matt Holt, "Efficiency of Forest Commodity Futures Markets", American Agricultural Economics Association 2004 Annual Meeting, 1-4 Aug. 2004, Denver, CO.

(35) J. Li, K. Hanrahan and S. McErlean, "The Efficiency of the Futures Market for Agricultural Commodities in the UK", American Agricultural Economics Association 2004 Annual Meeting, 1-4 Aug. 2004, Denver, CO.

(36) R.R. Rucker, N.T. Walter and J.K. Yoder, "Estimating the Structure of Market Reaction to News: Information Events and Lumber Futures Prices", American Journal of Agricultural Economics 87, no. 2 (2005): 482-500.

(37) Andrew M.F. McKenzie and Mathew T. Holt, "Market Efficiency in Agricultural Futures Markets", Applied Economics 34, no. 12 (2002): 1519-32.

(38) D.F. Kenourgios and A.G. Samitas, "Testing Efficiency of the Copper Futures Market: New Evidence from London Metal Exchange", Global Business and Economics Review Anthology (2004): 261-7.

(39) D. Bigman, D. Goldfarb and E. Schechtman, "Futures Market Efficiency and the Time Contents of the Information Sets", Journal of Futures Markets 3 (1983): 321-34.

(40) L. Martin and P. Garcia, "The Price-Forecasting Performance of Futures Markets for Live Cattle and Hogs: A Disaggregated Analysis", American Journal of Agricultural Economics 63 (1981): 209-15.

(41) T.A. Kofi, "A Framework for Comparing the Efficiency of Futures Markets", American Journal of Agricultural Economics 55 (1973): 584-94.

(42) M. Motamed, K. Foster and W. Tyner, "Applying Cointegration and Error Correction to Measure Trade Linkages: Maize Prices in the United States and Mexico", Agricultural Economics 39, no. 1 (2008): 29-39.

(43) Carter and Mohapatra, "How Reliable are Hog Futures as Forecasts?".

(44) C. Brooks, Introductory Econometrics for Finance (Cambridge: Cambridge University Press, 2008), p. 365.

(45) Lai and Lai, "A Cointegration Test for Market Efficiency".

(46) Foster, Havenner and Walburger, "System Theoretic Time-Series Forecasts of Weekly Live Cattle Prices".

(47) D.A. Dickey and W.A. Fuller, "Distribution of the Estimators for Autoregressive Time Series with a Unit Root", Journal of the American Statistical Association 74 (i979): 427-31.

(48) P.C.B. Phillips and P. Perron, "Testing for a Unit Root in Time Series Regressions", Biometrika 75 (1988): 335-46.

(49) Lai and Lai, "A Cointegration Test for Market Efficiency".

(50) Soren Johansen, Likelihood-based Inference in Cointegrated Vector Autoregressive Models (New York: Oxford University Press, 1995).

(51) Soren Johansen and Katarina Juselius, "The Full Information Maximum Likelihood Procedure for Inference on Cointegration with Applications", Oxford Bulletin of Statistics and Economics 52, no. 2 (1990): 169-211

(52) R.F. Engle and C.W.J. Granger, "Cointegration and Error Correction: Representation, Estimation, and Testing", Econometrica 55 (1987): 251-76.

(53) The Lagrange Multiplier test was used to test if the restriction b = y = 0 holds. Our results showed that we cannot reject [[beta].sub.i] = [[psi].sub.j] = 0.

(54) Carter and Mohapatra, "How Reliable are Hog Futures as Forecasts?".

(55) B.W. Brorsen and N.F. Fofana, "Success and Failure of an Agricultural Futures Contracts", Journal of Agribusiness 19 (2001): 129-45.

(56) Williams, Peck, Park and Rozelle, "The Emergence of a Futures Market Mungbeans on the China Zhengzhou Commodity Exchange", p. 430.

(57) Wang and Ke, "Efficiency Tests of Agricultural Commodity Futures Markets in China".

Zheng Shi (zhengshil974@ruc.edu.cn) is Associate Professor in the School of Agricultural Economics and Rural Development at Renmin University of China. He obtained his PhD in Agricultural Economics from Purdue University. His research interests include agricultural markets, industrial organisation and futures market.

Wang Zhigang (ohshigo@yahoo.com.cn) is Professor in the School of Agricultural Economics and Rural Development at Renmin University of China. He obtained his PhD in Agricultural Economics from Kyushu University. His research interests include industrial organization, food safety and food policy.

TABLE 1

SUMMARY STATISTICS OF SOYBEAN PRICES IN THE DCE AND ZGWM (US DOLLAR
PER BUSHEL)

Time           Mean    Median   Maximum   Minimum   Standard
                                                    Deviation

ZGWM Average   13.51   12.26     22.01     10.08      3.36
1 week lag     13.26   12.11     22.42     9.56       3.25
1 month lag    13.15   12.23     22.71     9.64       3.07
2 month lag    13.24   12.06     22.77     9.91       3.27
3 month lag    13.07   12.36     22.83     9.85       3.31
4 month lag    13.04   12.17     21.92     9.21       3.23
5 month lag    12.79   12.03     20.73     8.76       2.94
6 month lag    12.74   11.86     20.39     8.80       2.92

TABLE 2

OLS REGRESSION RESULTS OF ALL CONTRACTS USED: [C.sub.t] = a +
b[F.sub.t-1] + [[epsilon].sub.t]

                                   1           1           2
                               week lag    month lag   months lag

a (a)                           0.0441      0.0133       0.1122
                               (0.0478)    (0.0588)     (0.0761)
b (a)                          0.9677 **   0.9979 **   0.8970 **
                               (0.0467)    (0.0575)     (0.0698)
R-squared                       0.9168      0.8853       0.8091
Adjusted R-squared              0.9147      0.8823       0.8043
F (b) (a = 0, b = 1)            1.6033      0.7062       2.4847
                               (0.2129)    (0.4058)     (0.1230)
F (b) (b = 1)                   0.0411      3.8375      5.1146 *
                               (0.8403)    (0.0575)     (0.0295)
S.E. of regression              0.0726      0.0852       0.1099
Sum of squared residuals        0.2055      0.2833       0.4713
Log likelihood                  50.3937     43.8077     33.3731
Prob (F-statistic)              0.0000      0.0000       0.0000
Mean dependent variable         1.0062      1.0062       1.0062
S.D. dependent variable         0.2485      0.2485       0.2485
Akaike information criterion    -2.3607     -2.0394     -1.5304
Schwarz criterion               -2.2771     -1.9558     -1.4468
Hannan-Quinn criterion          -2.3302     -2.0090     -1.5000
Durbin-Watson statistic         1.8355      1.3146       0.8323

                                   3            4
                               months lag   months lag

a (a)                            0.1334     0.2245 **
                                (0.0882)     (0.1006)
b (a)                          0.8752 **    0.7839 **
                                (0.0886)     (0.0979)
R-squared                        0.7265       0.6218
Adjusted R-squared               0.7195       0.6121
F (b) (a = 0, b = 1)             2.3038      5.0108 *
                                (0.1371)     (0.0310)
F (b) (b = 1)                  15.8246 **   44.1851 **
                                (0.0003)     (0.0000)
S.E. of regression               0.1316       0.1547
Sum of squared residuals         0.6753       0.9339
Log likelihood                  25.9997      19.3532
Prob (F-statistic)               0.0000       0.0000
Mean dependent variable          1.0062       1.0062
S.D. dependent variable          0.2485       0.2485
Akaike information criterion    -1.1707      -0.8465
Schwarz criterion               -1.0871      -0.7629
Hannan-Quinn criterion          -1.1403      -0.8161
Durbin-Watson statistic          0.7150       0.4947

                                   5             6
                               months lag   months lag

a (a)                           0.2470 *     0.3098 **
                                (0.1239)     (0.1334)
b (a)                          0.7627 **     0.6983 **
                                (0.1213)     (0.1304)
R-squared                        0.5034       0.4237
Adjusted R-squared               0.4907       0.4089
F (b) (a = 0, b = 1)             3.9794      5.4414 *
                                (0.0531)     (0.0249)
F (b) (b = 1)                  53.6809 **   975.5260 **
                                (0.0000)     (0.0000)
S.E. of regression               0.1773       0.1910
Sum of squared residuals         1.2263       1.4231
Log likelihood                  13.7700       10.7186
Prob (F-statistic)               0.0000       0.0000
Mean dependent variable          1.0062       1.0062
S.D. dependent variable          0.2485       0.2485
Akaike information criterion    -0.5741       -0.4253
Schwarz criterion               -0.4906       -0.3417
Hannan-Quinn criterion          -0.5437       -0.3949
Durbin-Watson statistic          0.6118       0.5082

Notes: (a) Standard errors of a and b parameter estimates are
provided in parentheses; (b) F is the joint F test of rejecting the
null hypothesis of a = 0 and b = 1 (p value of F test is given below
F-statistic in parentheses); * and ** indicate significance at the 5
and 1 per cent levels, respectively.

TABLE 3
UNIT ROOT TESTS ON PRICES SERIES

Prices                               ADF Tests

                           Levels                Differenced

                   Statistics   P-values   Statistics    P-values

Cash               -1.4754      0.5300     -4.9767 ***   0.0002
Futures            -1.7028      0.4222     -4.4157 ***   0.0011
(lag = 1 week)
Futures            -2.0165      0.2789     -4.6289 ***   0.0006
(lag = 1 month)
Futures            -2.0135      0.2801     -4.4849 ***   0.0009
(lag = 2 months)
Futures            -2.3058      0.1751     -4.7627 ***   0.0004
(lag = 3 months)
Futures            -2.3250      0.1694     -3.7691 ***   0.0065
(lag = 4 months)
Futures            -2.3635      0.1582     -4.3762 ***   0.0012
(lag = 5 months)
Futures            -1.7924      0.3789     -5.6830 ***   0.0000
(lag = 6 months)

Prices                          Phillips-Perron Tests

                          Levels                 Differenced

                   Statistics   P-values   Statistics    P-value

Cash               -1.7823      0.3838     -4.9604 ***   0.000
Futures            -1.6401      0.4535     -0.4811 ***   0.000
(lag = 1 week)
Futures            -1.7617      0.3936     -4.5523 ***   0.000
(lag = 1 month)
Futures            -1.7131      0.4173     -4.4586 ***   0.001
(lag = 2 months)
Futures            -1.7972      0.3767     -4.5841 ***   0.000
(lag = 3 months)
Futures            -1.7836      0.3831     -3.8211 ***   0.005
(lag = 4 months)
Futures            -1.8009      0.3749     -4.3205 ***   0.001
(lag = 5 months)
Futures            -1.9346      0.3137     -5.6795 ***   0.000
(lag = 6 months)

Note: Asterisk (***) denotes variables significant at the 1 per cent
level.

TABLE 4
JOHANSEN COINTEGRATION TESTS RESULTS

                  [[lambda].sub.trace]       [[1ambda].sub.max]

                 [H.sub.0]:   [H.sub.0]:   [H.sub.0]:   [H.sub.0]:
                 r = 0        r = 1        r = 0        r = 1

1 week lag        41.26 *       1.96        39.29 *       1.96
1 month lag       35.21 *       2.76        32.45 *       2.76
2 months lag      32.01 *       4.02        27.98 *       4.02
3 months lag      33.08 *       3.46        29.62 *       3.46
4 months lag      33.17 *       5.00        28.17 *       5.00
5 months lag      38.66 *       3.16        35.49 *       3.16
6 months lag      48.92 *       3.23        35.66 *       3.23
Critical value    19.96         9.24        15.67         9.24
  (5%)
Critical          24.60        12.97        20.20        12.97
  value (1%)

Notes: * : null hypothesis is rejected at 1 per cent level. Critical
values are from Table 1, see M. Osterwald-Lenum, "A Note with
Quantiles of the Asymptotic Distributions of the Maximum Likelihood
Cointegration Ranks Test Statistics: Four Cases", Oxford Bulletin (of
Economics and Statistics (1992): 461-72, p. 467 for case 1.

TABLE 5

EXCLUSION RESTRICTION TESTS ON INDIVIDUAL WEEK/MONTH PRICES

Lagged time   Estimated [beta]              Hypothesis
                (Normalised
               cointegration
               coefficients)

1 week             -1.03         [H.sub.0]: [[beta].sub.1week = 0
1 month            -1.13         [H.sub.0]: [[beta].sub.1month = 0
2 months           -1.05         [H.sub.0]: [[beta].sub.2month = 0
3 months           -1.19         [H.sub.0]: [[beta].sub.3month = 0
4 months           -1.12         [H.sub.0]: [[beta].sub.4month = 0
5 months           -1.28         [H.sub.0]: [[beta].sub.5month = 0
6 months           -1.21         [H.sub.0]: [[beta].sub.6month = 0

Lagged time   LR statistic   P-value

1 week           42.16          0
1 month          34.44          0
2 months         21.86          0
3 months         24.93          0
4 months         25.09          0
5 months         33.48          0
6 months         36.42          0

TABLE 6
ERROR-CORRECTION MODEL RESULTS

                          1           1           2            3
                      week lag    month lag   months lag   months lag

[lambda]               17.5011     42.4250     59.7172      15.4985
                      (39.8117)    (0.1900)   (48.7460)    (51.8488)
[rho]                  -0.9065     -0.4913     -0.4148      -0.0332
                       (0.2056)    (0.1776)    (0.1570)     (0.1792)
[beta]                  0.9718      0.7900      0.0535       0.4406
                       (0.1483)    (0.1366)    (0.2163)     (0.1939)
Wald test               0.2244      8.4568    118.5049      33.4634
[beta] = 1, [rho] =    (0.6383)    (0.0060)    (0)          (0)
  -1
Wald test               0.0573      3.3795     56.5675      19.5664
[beta] = 1             (0.8121)    (0.0736)    (0)          (0)
Wald test               0.3274     11.6788    243.9967      68.0024
[rho] = -1             (0.5705)    (0.0015)    (0)          (0)
[R.sup.2]               0.5390      0.4763      0.3914       0.2236
Adjusted [R.sup.2]      0.5141      0.4480      0.3585       0.1816
MSE                   250.7       267.21      288.05       325.35

                          4            5            6
                      months lag   months lag   months lag

[lambda]              -57.8493      37.9879      36.1081
                      (55.8621)     58.1500)    (58.4598)
[rho]                  -0.3473       0.1611       0.0702
                       (0.1758)      0.1471      (0.1206)
[beta]                  0.7094      -0.2068      -0.1697
                       (0.2704)     (0.2393)     (0.2225)
Wald test              25.4482      66.4977      82.1371
[beta] = 1, [rho] =    (0)          (0)          (0)
  -1
Wald test               3.4597      58.7639      48.8250
[beta] = 1             (0.0704)     (0)          (0)
Wald test              39.7268     142.5760     137.9564
[rho] = -1             (0)          (0)          (0)
[R.sup.2]               0.1614       0.0315       0.0158
Adjusted [R.sup.2]      0.1148      -0.0208      -0.0374
MSE                   342.76       363.37       366.31

Notes:

1. Standard errors are shown in parentheses below the coefficient
estimates for [lambda], [rho] and [beta];

2. Wald statistics and p value (in parentheses) are shown for
individual hypotheses [rho] = -1 and [beta] = 1;

3. P value for joint hypothesis: [rho] = -1 and [beta] = 1, are shown
in parentheses below the Wald-statistic distributed [x.sup.2.sub.2]
(chi-square with two degrees of freedom).

TABLE 7

TRADING RULES AND DELIVERY LOCATIONS OF CHINA'S SOYBEAN FUTURES
CONTRACT

Commodity            soybean number 1 (NGM Soybeans)
Trading Unit         10 metric ton/contract
Maturity Month       January, March, May, July, September and
                       November
Margin account       5% of the contract value
Transaction fee      4 Chinese yuan/contract (USD0.6/contract)
Last trading day     The 10th trading day of the maturity month
Delivery date        Before the ending of the seventh date after
                       trading
Delivery locations   15 delivery locations all located in Dalian
Storage fee          November 1-April 30: 0.4 Chinese yuan/day
                     May 1-October 31: 0.5/Chinese yuan/day

Source: Dalian Commodity Exchange website: http://www.dce.com.
cn/portal/cate?cid=1261730307127 [20 Oct.2013].

TABLE 8

COMPARISON OF DELIVERY QUALITY OF NGM AND GM SOYBEANS

Contract type            Yellow bean #1 contract

Bean type        Specific standard NGM

Standard         Quality required for food processing
                 purposes, such as tofu

Main content     Grain solid content

Moisture         Less than 13% for November, January,
                 March delivery; less than 13.5% for
                 May, July and September contract. Price
                 deduction for lower quality beans. Price
                 bonus for higher quality beans

Foreign matter   Less than 1%

Purity           100% yellow beans

Contract type           Yellow bean # 2 contract

Bean type        Specific standard GM and NGM

Standard         Quality required for making human
                 consumable oil

Main content     Fat and oil content

Moisture         Less than 13% for November, January,
                 March delivery; less than 13.5% for
                 May, July and September contract. Price
                 deduction for lower quality beans. No
                 price bonus for higher quality beans

Foreign matter   Less than 2%

Purity           Allow mix colour beans

Source: Dalian Commodity Exchange website: http://www.dce.com.
cn/portal/cate?cid=1261730307127 [20 Oct.2013].
COPYRIGHT 2013 East Asian Institute, National University of Singapore
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2013 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:RESEARCH; non-genetically modified
Author:Zheng, Shi; Wang, Zhigang
Publication:China: An International Journal
Article Type:Report
Geographic Code:9CHIN
Date:Dec 1, 2013
Words:9129
Previous Article:China's technology policy change: how effective has it been?
Next Article:Green ICT: a strategy for sustainable development of China's electronic information industry.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters