Agribusiness Review - Vol. 7 - 1999

Paper 11
ISSN 1442-6951

Testing Rice Market Integration in China:
The Case of Guangdong and Jiangxi 1

Zhang-Yue Zhou
Asian Agribusiness Research Centre, Orange Agricultural College
The University of Sydney, Orange NSW 2800 Australia
Email: zhou@oac.usyd.edu.au

Wei-Ming Tian
College of Economics and Management, China Agricultural University
Beijing China 100294
Email: tianwm@public.intercom.com.cn

and

Guang-Hua Wan
Department of Agricultural Economics, The University of Sydney
Sydney NSW 2006 Australia
Email: g.wan@agec.usyd.edu.au

• Abstract
• Introduction
• Conceptual Considerations
• Grain Production and Markets in Guangdong and Jiangxi
• Guandong Province
• Jiangxi Province
• Cross-border Grain Trading
• Testing Techniques and Data
• Results and Discussion
• Summary and Concluding Remarks
• References
• Appendix
• Footnotes

Abstract

Recent studies that examine the integration of grain markets in China reveal that, for inter-provincial markets, there is generally a lack of integration. A logical question to ask is: are markets that are close to each other in two neighbouring provinces with one being grain deficit and the other being grain surplus more integrated? This study examines the integration of rice markets between two neighbouring provinces: Guangdong (grain deficit) and Jiangxi (grain surplus). It is found that there exist inter-market price relationships exist between markets in these two provinces. Policy implications are discussed.

Introduction

Studies of spatial market integration can provide important information on how the markets work. Such information may help a government to decide the extent to which it should promote market development. In recent years increased attention has been paid to testing for spatial grain market integration in China. Such empirical analyses include Wu (1994) , Cheng and Wu (1995) , and Zhou, Wan and Chen (1997, 1998) , and Yu and Huang (1998) .

However, available studies, all relying on aggregated provincial data, suggest that in general there is a lack of integration in the Chinese grain market. Major reasons include existing heavy government intervention and lack of transportation facilities. In these studies, a whole province is treated as a single market, or one single market from a province is chosen to represent the whole province. These "markets" may be at too high an aggregate level or may be too "local" to represent the many regional markets of a province. In this study, we use data from several regional markets in neighbouring provinces to determine whether there exist integration relationships between markets across two such provinces, as well as within individual provinces. The two provinces chosen are Guangdong, a grain-deficit province, and Jiangxi, a grain-surplus province. Cointegration technique is used for testing market integration with monthly price data of indica rice as observed in 12 regional markets of the two provinces.

Conceptual Considerations

Under regularly assumed restrictions on preferences and technologies, a competitive market equilibrium exists and is efficient in the Paretian sense. In general, this also holds for the spatial competitive equilibrium of an economy consisting of a set of regions among which trade occurs at fixed transport costs. Such an equilibrium has the property that, if any trade takes place between any two regions, price in the importing region equals price in the exporting region plus the unit transport cost. Under this circumstance, the markets can be said to be spatially integrated. Market integration can thus be defined as the extent to which changes in prices in one market lead to changes in prices in another (Wyeth 1992, p. 2) .

Given the above, it can be said that any factors which affect grain trade between markets affect market integration. Availability of price information and access to transport are considered to be the most stressed exogenous factors affecting price behaviour. Price information involves not only its quality but also the speed of obtaining it. The former significantly affects the outcomes of arbitrage while the latter determines the time needed to arbitrage between markets, and hence the speed of market integration.

Availability and condition of roads and transport facilities are important. Good roads reduce the time needed for grain movement, while advanced transport facilities (e.g., bulk handling for grains) make grain transfer more efficient. In China, transport capacity used to be quite tight. Railways were the major means of connecting markets, but getting approval for rail transport was not an easy matter. The situation has been gradually changing. In the past few years, there has been rapid construction of highways in China, allowing people to rely less upon railway freight, particularly for short distance transportation. Advanced transport facilities, however, are still not readily available. Grains still have to be bagged before transportation.

The distance between markets is also relevant. In principle, distance should not be an obstacle in this context though it affects the speed of market integration. That is, it will take longer for traders to arbitrage between two markets which are further apart. Therefore, lower frequency price data (e.g., monthly or fortnightly) may suffice for examination of integration between distant markets. Higher frequency data (e.g., weekly or even daily) should be used for markets which are closer to each other. In the latter case, price deviation may disappear quickly due to speedy arbitrage, and thus low frequency data may not contain much information regarding short-run integration.

Government intervention in the grain market is another vital factor which affects market integration. Intervention may reduce integration, and in some cases it may even completely cut off price relationships between markets. Consequently, arbitrage between markets just cannot take place even when price information and transport are readily available.

Other factors, as identified by Palaskas and Harriss-White (1993) with reference to India, include the nature of trading contact, the structure of commodity systems, the nature of contracts, and so on. In the Chinese context, the size and development of the private grain sector may be of high relevance to the degree of market integration.

Grain Production and Markets in Guangdong and Jiangxi

To put our market integration discussion into context, a brief discussion on grain production and markets of the two chosen provinces should be beneficial. (A summary of grain production, markets, and other related information is given in Table A1 in the Appendix .)

Guandong Province

Guangdong has a land area of 177900 square kilometres with 21 prefecture-level cities. In 1996 it had a population of 69 million, of which 69 percent was agricultural. Per capita arable land was 0.03 ha (cf. national average of 0.08 ha). Chemical fertiliser usage was 816 kg/ha in 1996 (cf. national average of 399 kg/ha). The southern part of the province has been rapidly industrialised in the past two decades, whereas in the northern part, agriculture is still largely dominant.

The major grain crop is rice (indica), accounting for around 86 percent of total grain output. A few other minor grain crops include wheat, tuber crops, and dry land grain crops. In 1996, total grain output of the province was 18.9 million tonnes, of which a little over 16 million tonnes was rice.

Per capita grain output varies significantly between areas within Guangdong, ranging from almost nil to as high as 590 kg per capita. However, it is a major grain-deficit province in China with an average grain output of 274 kg per person in 1996 (cf. national average of 412 kg). In addition, Guandong has a large number of migrant workers who also demand grains. Guangdong has been importing grains from other provinces as well as from international markets.

There is little potential for Guangdong to increase grain output as yields are already high, and returns are relatively low. Industrialisation and urbanisation will take up more agricultural land and thus Guangdong's grain production is likely to decline further. This will lead Guangdong to relying even more on imported grains.

Guangdong shares borders with Guangxi, Hunan, Jiangxi and Fujian (see Map 1) . As Guangxi and Fujian are themselves grain-deficit provinces, it is logical for Guangdong to seek imports from Hunan and Jiangxi. In addition, the majority of rice produced in Hunan and Jiangxi is indica, which is the rice customarily consumed in Guangdong. In the past, however, there have been some disputes in relation to grain trade between Guangdong and Hunan, which has on occasions imposed blockages at its border to stop grains being moved to Guangdong. This has led to Guangdong seeking grains from Jiangxi and Hubei, which is north of Hunan (Li Ding, personal interview, 12 January 1999).

Map 1. Position of Guangdong and Jiangxi Provinces and the Markets Under Investigation

As a grain-deficit region, Guangdong has been pragmatic in its response to central government requests to intervene in the grain market. Local governments were ordered not to block the transportation of grains to Guangdong. Interviews with officials of the grain bureau in Puning area (one market included in this study) by one of the authors lend further support to the pragmatic practice of Guangdong. Private traders in Guangdong are active and some maintain close contacts with their counterparts in Jiangxi. When needed, grains can often be brought over to markets in Guangdong in a short time period. During a visit to rice markets in the Puning area in April 1998, one of the authors was told by the private merchants that, due to their business contacts in Jiangxi, they could easily have rice shipped to them from Jiangxi within 24 hours.

Jiangxi Province

Jiangxi has a land area of 169000 square kilometres with six prefecture-level cities and five prefectures. In 1996 it had a population of 41 million, of which almost 80 percent was agricultural. Per capita arable land was 0.06 ha. Chemical fertiliser usage averaged 491 kg/ha in 1996 (cf. Guangdong 816 kg/ha and national average 399 kg/ha). Agriculture is dominant in its economy, with grain production being its major agricultural industry.

The major grain crop in Jiangxi is rice (indica), accounting for around 95 percent of the total grain output. Other minor grain crops include wheat, maize, pulses, and some tuber crops. In 1996, total grain output of the province was 17.7 million tonnes, of which 16.8 million tonnes was rice. Per capita grain output was 432 kg.

Jiangxi has a promising potential to increase its grain output through improved grain yield. At present its yield is low, being about 6450 kg/ha, but research shows that a much higher grain yield is achievable (Rural Survey Sections 1996) . Two major reasons for the low yield are the lack of research and extension and the lack of incentives to invest in the land.

Jiangxi is a grain-surplus or, more accurately, a rice-surplus province. Jiangxi's surplus rice has been subject to a heavy central government quota and is usually exported to other provinces. It is the second largest rice exporter among all rice-exporting provinces in China. Grains exported from Jiangxi make up about 9 percent of the national pool. However, the majority of the grains are exported to other provinces at prices set by the central government. These prices are generally low and, on occasions, even lower than production costs. As a consequence, Jiangxi is rich in grain but poor in finance. Up to the early 1990s, of the 43 major grain-producing counties, 32 were financially poor and had to rely on fiscal subsidies (relief) (Pei and Lan 1996) .

Jiangxi is surrounded by Guangdong, Fujian, Zhejiang, Anhui, Hubei and Hunan (see Map 1) . The latter three are at least self-sufficient in grains. In recent years, however, the former three have changed from from grain-surplus to grain-deficit provinces as a result of rapid economic development, thus opening up new markets for Jiangxi. Jiangxi's proximity to these provinces provides it with an enormous geographical advantage to sell its surplus rice to these markets.

At present, government intervention has largely deprived Jiangxi of its ability to benefit from its surplus grain. In years of poor harvests, much of the surplus grain is required by the central government for export to other provinces, but at low prices. In years of good harvests, Jiangxi may find it difficult to dispose of its surplus grain.

Private grain traders have also emerged in recent years in Jiangxi. Some of them purchase grains from local markets and even from markets beyond the local area, process them, and then sell them to the markets in other provinces (for examples, see Xie and Song 1997) . However, local governments in Jiangxi are generally less "pragmatic" compared to those in their neighbouring Guangdong province. Some tend to follow commands from their higher-level administrators more closely and treat private grain traders more harshly, making business patchy for them. When there is plenty of grain, business may be all right for the private traders; otherwise, it may be significantly affected as, on such occasions, grain markets may be closed and the traders not allowed to buy grains for any arbitrage activities (Ding and Chen 1996) .

In early 1995 the ‘provincial governor responsibility system' was introduced. As a result, grain markets in Jiangxi became subject to even heavier government control due to the need for local government officials to fulfil the quota procurement assigned to them. 2
It is not unusual for local officials to close local grain markets before fulfilling their quota, thus making the business of private traders very difficult. Ding and Chen (1996) , in a report based on their survey in Jiangxi, pointed out that some private traders even find it difficult to survive. For example, in Nanchang county, there are about 200 grain processing firms with an annual capacity of 300-350 thousand tonnes of rice. In 1995, after the fulfilment of the government quota, the surplus left in the market was very limited and the amount of rice processed by those firms was less than 100 thousand tonnes. A company in Taihe county has a processing capacity of 20 thousand tonnes of rice, but processed only some two thousand tonnes in 1995 (Ding and Chen 1996) .

Since the introduction of the "provincial governor responsibility system", local government officials have treated private traders' action of purchasing grains directly from farm gates almost as illegal, as this practice disturbed the fulfilment of their procurement quota. In early 1998, the central government officially made it illegal for any non-government grain companies or traders to buy grains from farmers. Government grain companies were given a monopolistic status to procure grains from farmers.

As a major rice-surplus region, Jiangxi exports rice mainly in three directions: Shanghai (via Jiujiang or Nanchang), Fujian (via the Yingtan-Fuzhou railway) and Guangdong (via Nanchang/Changsha before the Jing-Jiu railway opened). Jiujiang is one of the four famous national rice assembly markets (the other three being Wuhan in Hubei, Wuhu in Anhui, and Wuxi in Jiangsu). There is also a major grain wholesale market in Nanchang.

Cross-border Grain Trading

State grain companies are the most important player in the market. They handle the majority of the tradable rice and have advantages over private traders in their access to loans, transport and storage facilities. However, in general they are less efficient. For a given price margin between two markets, they may incur a loss if they arbitrage between the two, while private traders may earn a profit. 3 Unfortunately, despite their own inability to arbitrage successfully, they do not generally support private traders' arbitrage due to conflict of interest; instead, they tend to make it difficult through their control of, or influence on, essential facilities such as transport, storage, and sources of credit.

Despite government intervention and difficulties in grain arbitrage, grain trading by private agents between Jiangxi and Guangdong is not completely impossible. According to Ding's surveys conducted in Jiangxi and Guangdong, grain movements by private traders between the two provinces do exist (Li Ding, personal interview, 12 January 1999). Two major reasons may be responsible for this. First, the borders of Guangdong and Jiangxi are mountainous, making it almost impossible to ensure a complete blockage. Second, some local officials are not strongly motivated to completely block grain movements, as this would only increase local grain disposal problems and reduce farmers' income.

Given the production and market situations in Guangdong and Jiangxi and based on the conceptual framework, we conclude that there exist inter-market price relationships between grain markets in Guangdong and Jiangxi.

Testing Techniques and Data

Since Lele (1967) , much attention has been given to testing spatial market integration (e.g., Blyn 1973 ; Hurd 1975 ; Harriss 1979 ; Ravallion 1986 ; Goodwin and Schroeder 1991 ; Palaskas and Harriss-White 1993 ; Alexander and Wyeth 1994 ; Dercon 1995 , and Baulch 1995) . Besides the variance analysis method which is less common (Hurd 1975) , four techniques for testing market integration have evolved in the past three decades, namely, (1) the correlation method, (2) the Ravallion procedure, (3) the cointegration approach, and (4) the parity bounds model. (For a brief discussion of each of the four techniques and examples of studies using them, see Zhou, Wan and Chen 1997 .)

In this study, the cointegration technique is used. When carrying out cointegration analysis, the initial step is to establish the order of integration of every price series (It is noted that when the word "integrated" or "integration" is related to the discussion of a time series, it is an econometric term and has a different meaning from that used in the discussion of integration of markets). If individual price series are found to be integrated to the same order, the cointegration approach can be used.

Let i, j index markets, t index time and P denote commodity price. A simple regression model can be specified as:

P_i = d + a P_j + v_t

If P_i and P_j are each integrated to the same order and v_t is integrated to degree zero, P_i and P_j are said to be cointegrated. This, in turn, implies long-run market integration. The econometric computer package used is Eview, in which the "Johansen Cointegration Test" is performed.

Monthly market prices of indica rice from 6 cities in each province were obtained from Provincial Statistical Bureaus of Guangdong and Jiangxi. The time series for Jiangxi extends from January 1992 to July 1996, while that for Guangdong from January 1994 to July 1997. The six markets included for Guangdong are Boluo, Puning, Zencheng, Pingyuan, Zhongshan and Dianbai. The six markets included for Jiangxi are Nanchang, Jiujiang, Jingdezhen, Pingxiang, Yingtan and Xinyu. See Map 1 for their locations. The selection of the markets is dictated by data availability.

Results and Discussion

The six price series of each province were deflated by the monthly Consumer Price Indices of the corresponding province, published in China Monthly Statistics , to eliminate possible co-movements between prices. Inflation can be an important cause of movements in prices. Using nominal prices may lead to the conclusion that cointegration exists between two time-series, even if there is no actual market relation between them. That is, prices between the two markets may show co-movement, but this is caused by a common inflationary trend.

Using Eview, the "Johansen Cointegration Test" was applied to the six price series to examine the existence of inter-market price relationships within a province, and to all the 12 price series to check the existence of any price relationships between markets across the two provinces. The time spans of the price series of the two provinces are not the same and they have been trimmed (with the loss of some observations) to form price series of the same time length when analysing market integration across provinces. The cointegration test results are reported in Tables 1 to 3 .

Table 1 . Cointegration among Rice Price Series, Guangdong

Notes: 1. - no significant cointegration.
2. * cointegration significant at the 5 percent level.

Table 2 . Cointegration among Rice Price Series, Jiangxi

Notes: 1. - no significant cointegration.
2. * cointegration significant at the 5 percent level.

Table 3 . Cointegration among Rice Price Series, Guangdong and Jiangxi

Notes: 1. - no significant cointegration.
2. * cointegration significant at the 5 percent level.

It is interesting to note from Tables 1 and 2 that, while there exist clear inter-market price relationships between markets within Guangdong province, such relationships are virtually non-existent between markets within Jiangxi province. 4 Most grain markets in Guangdong face shortages in local supply. Traders (mainly from other provinces) are likely to look for more favourable prices among the markets. The resultant arbitrage activities would help result in more closely linked markets. Further, transportation conditions in Guangdong are also relatively good.

When there is an overall grain surplus in a province, profitable price margins between markets may be negligible. As a consequence, there would be little incentive for traders to exploit such a small margin between the markets which are all exporting grains. Clearly, lack of arbitrage activities within Jiangxi explains why significant cointegration relationships do not exist between most pairs of its markets. Per capita grain output in Table A2 also lends support to such a finding: inter-market price relationships tend to be weak between markets where grain is in surplus but strong between markets where grain is in short supply locally. Table 1 shows that within Guangdong province, only Boluo and Pingyuan do not have a significant price relationship. Per capita grain output is high in both of these two areas, 439 kg and 472 kg, respectively.

Table 3 above shows that there exist inter-market price relationships between markets in the neighbouring provinces Guangdong and Jiangxi. Jiangxi has surplus grains to sell while Guangdong needs the surplus to meet its shortage. While Guangdong opens its door and encourages grain movements into its territory, traders in Jiangxi do capitalise on this opportunity.

Non-existence of price relationship between Pingyuan in Guangdong and all the other markets in Jiangxi further endorses the finding that inter-market price relationships tend to be weak between markets where grain is not in short supply. The lack of significant relationship between Boluo in Guangdong and Jingdezhen, Pingxiang and Xinyu in Jiangxi may be due to the relatively easy availability of grains in Boluo itself (Table A2) . Jingdezhen in Jiangxi and Dianbai in Guangdong are geographically the furthest apart among all the markets under investigation. Pingxiang in Jiangxi has a relatively low per capita grain output but it is situated next to a major grain production prefecture (Yichun, with a per capita grain output of 628 kg) and is on a railway line. Jiujiang has the lowest per capita grain output in Jiangxi but it is one of the four major national rice assembling markets. Rice from Hubei, Anhui and Jiangxi is assembled here and distributed to many destinations in other provinces.

In the discussion above, per capita grain output is used as a rough indicator of grain surplus/deficit in an area. It may be argued that, for Guangdong, what matters is per capita rice output rather than grain output, as rice is customarily the staple food. In addition, Guandong has a large floating population which is not included in local statistics, and therefore per capita grain output may be no longer a reliable indicator of grain surplus/deficit status in a particular area. However, considering that the majority of the floating population is in fact located in the Pearl Delta area rather than some county areas such as Pingyuan (Boluo is in a developed region but with a relatively strong grain sector), as far as per capita grain (for which rice accounts over 85% at the provincial level) availability is concerned, it would be much lower than calculated from statistics in the Pearl Delta area, but it would be less affected in some other areas. Thus, per capita grain output can serve as a surrogate for gauging the grain surplus/deficit status of an area in Guangdong.

There are limitations in this study. The absence of markets in the southern part of Jiangxi and most markets in the northern part of Guangdong was a major data deficiency. It would be valuable to include the markets in southern Jiangxi, which are closer to Guangdong markets and are in poorer areas. It would also have been beneficial to include more markets in the less developed areas of northern Guangdong. The short length of the time series and the low frequency of the data limited the scope of our analysis. Unfortunately, the markets which could be included in this study and the length of the time series were dictated by data availability. Obtaining crop specific price data at the local level with higher frequency in China is by no means an easy task. Further studies, using longer time series and including more markets, are called for.

Summary and Concluding Remarks

In this study, we attempted to examine whether rice markets in two neighbouring provinces, one with a grain deficit and the other with a grain surplus, are integrated and whether rice markets within a province are integrated. Owing to the limited availability of data, we were not able to include the markets in southern Jiangxi and northern Guangdong (in poorer areas of both provinces) in this study. This led us to being cautious in drawing any firm conclusions about market integration in Jiangxi and Guangdong provinces. Nonetheless, the results of this study are largely in accordance with expectations and may serve as groundwork for further studies in this area.

The results of this study suggest that:

• there exist inter-market price relationships between markets within a grain-deficit province;
• there exist weak, if any, inter-market price relationships between markets within a grain-surplus province; and
• there exist inter-market price relationships between markets across a province with a grain surplus and a province with a grain deficit.

Further, the results suggest that the tighter the grain availability from local sources, the stronger the inter-market price relationships between grain-deficit markets, and vice versa for grain-surplus markets.

The existence of a significant price relationship between markets in Guangdong and Jiangxi seems to be related to the pragmatic approaches adopted by local governments in both provinces. In addition, geographical proximity and relatively well developed telecommunication facilities are also contributing factors.

Thus, reduced government intervention in grain markets encourages cross-border grain movements. This in turn promotes the integration of markets and benefits both grain-importing and exporting regions. More integrated grain markets will help producers and consumers realise the gains from grain market reforms, as correct price signals can be transmitted down the marketing chain. Consequently, consumers in some markets will not have to pay higher prices, and farmers will be able to specialise according to their comparative advantages. This leads to more efficient use of resources.

Deserving particular attention by the Chinese government is the fact that Jiangxi is a major surplus grain provider with great potential to increase its grain production. Jiangxi's advantage in grain production has now become even more important for China in view of the recent changes in grain production and trade patterns in the country (i.e., southern China used to have surplus grain, and exported to the north; but now surplus grain has to be exported from northern China to the south). Given Jiangxi's immediate proximity to grain-deficit provinces in southern China, it is of significant strategic importance for the central government to encourage Jiangxi to maintain its grain production advantage.

However, unless Jiangxi can gain financially from producing more grain, Jiangxi may not have the necessary means to maintain such an advantage (e.g., investment in research and extension and capital construction). Farmers will also have little interest in investing in land to produce more grains. Hence, reduced central government intervention is essential. This will not only assist in market integration between Jiangxi and importing provinces, but most importantly, the right market signals will be available to Jiangxi farmers for them to produce according to the market.

The government may wish to reduce its control over the private grain sector. Private grain traders seem to operate more efficiently than government agents. They generally offer a higher price to farmers, and hence provide increased incentives for farmers to produce grain. They are even able to arbitrage between markets where government agents find it economically unviable to do so. This in turn helps to dispose of grains from a surplus region.

To enable private traders to arbitrage between markets, they need to be allowed to buy grains with few restrictions and to have access to transport facilities. This requires significant reduction in government intervention in the grain market. In particular, government control of 70-80 percent of commercial grains and the monopoly purchase of grains from farmers by state agents should be relaxed.

Finally, we note once more that, due to data limitations, particularly the non-inclusion of some markets in poorer areas of the two provinces, one needs to be cautious in interpreting the results of this study. Further studies using longer time series and including more markets would certainly be useful. A similar study that would be interesting and worthwhile is to test the integration of rice markets between Guangdong and Hunan provinces.

References

Alexander , C. and Wyeth, J. (1994), ‘Cointegration and market integration: an application to the Indonesian rice market', Journal of Development Studies , 30:303-328.

Baulch , B. (1995), ‘Transfer costs, spatial arbitrage and testing food market integration', Institute of Development Studies, University of Sussex, Working Paper No. 20.

Blyn , G. (1973), ‘Price series correlation as a measure of market integration', Indian Journal of Agricultural Economics , 28:56-59.

Cheng , E.J. and Wu, Y.R. (1995), ‘Market reform and integration in China in the early 1990s: the case of maize', Chinese Economy Research Unit, the University of Adelaide.

China Monthly Statistics , China Statistical Information and Consultancy Service Centre, Beijing. Various issues.

Dercon , S. (1995), ‘On market integration and liberalisation: method and application to Ethiopia', Journal of Development Studies , 32:112-143.

Ding , L. and Chen, L.B. (1996), ‘A survey on economic development issues of major grain-producing regions in Jiangxi', report to the Ministry of Agriculture, Beijing.

Editorial Board of Almanac of China's Domestic Trade (1997), Almanac of China's Domestic Trade 1997, published by the Edipolar Board of Almanac of China's Domestic Trade, Beijing.

Ge , C.Y. (1996), ‘A problem related to the provincial governor responsibility system', China Grain Economy , No. 1, p. 27.

Goodwin , B. and Schroeder, T. (1991), ‘Cointegration tests and spatial price linkages in regional cattle markets', American Journal of Agricultural Economics , 73:452-464.

Guangdong Statistical Bureau , Statistical Yearbook of Guangdong , various issues, China Statistical Press, Beijing.

Guo , Z. (1996), ‘A viewpoint about the provincial governor responsibility system', China Grain Economy , No. 1, p. 26.

Harriss , B. (1979), ‘There is method in my madness: or is it vice versa? Measuring agricultural market performance', Food Research Institute Studies , 17:197-218.

Hurd , J. (1975), ‘Railways and the expansion of markets in India, 1861-1921', Explorations in Economic History , 12:263-288.

Jiangxi Statistical Bureau , Statistical Yearbook of Jiangxi , various issues, China Statistical Press, Beijing.

Lele , U.J. (1967), ‘Market integration: a study of sorghum prices in Western India', Journal of Farm Economics , 49:147-159.

Palaskas , T.B. and Harriss-White, B. (1993), ‘Testing market integration: new approaches with case material from the West Bengal food economy', Journal of Development Studies , 30:1-57.

Pei , D.A. and Lan, M.H. (1996), ‘Issues concerning the development of Jiangxi' grain production and strategies to handle', Rural Development Forum , No. 2, pp. 2-5.

Ravallion , M. (1986), ‘Testing market integration', American Journal of Agricultural Economics , 68:102-109.

Rural Survey Sections of State Statistical Bureau and Jiangxi Province (1996), ‘A study on Jiangxi's grain demand-supply equilibrium: Present and prospect', Institute of Statistical Sciences, State Statistical Bureau, Beijing.

Wang , J.H. and Liu, J. (1995), ‘Interest conflicts in grain's production, marketing and distribution', China Grain Economy , No. 4, pp. 25-27.

Wyeth , J. (1992), ‘Measure of market integration and applications to food security policies', Institute of Development Studies, Discussion Paper, No. 314, The University of Sussex, England.

Wu , Y.R. (1994), ‘Rice markets in China in the 1990s', Chinese Economy Research Unit, the University of Adelaide, Working Paper, No. 94/10.

Xie , F.W. and Song, Z.H. (1997), ‘Emerging professional grain traders', Farmers Daily, 28 March, p. 4.

Yu , W. and Huang, J.K. (1998), ‘Rice market integration and grain market reforms in China', Economic Research , No. 3, pp. 50-57.

Zhou , Z.Y., Wan, G.H. and Chen, L.B. (1997), ‘Integration of rice markets: The case of northern China', Asian Journal of Agricultural Economics, Vol. 2, pp. 158-76.

Zhou, Z.Y., Wan, G.H. and Chen, L.B. (1999), ‘Integration of rice markets: The case of southern China', Contemporary Economic Policy. Accepted for publication.

Appendix

Table A1 . Grain Production and Markets in Guangdong and Jiangxi (1996)

Sources:
1. Guangdong Statistical Bureau, Statistical Yearbook of Guangdong , various issues.
2. Jiangxi Statistical Bureau, Statistical Yearbook of Jiangxi , various issues.
3. Editorial Board of Almanac of China's Domestic Trade (1997), Almanac of China's Domestic Trade 1997, pp. II167-174.

Table A2 . Per Capita Grain Output in Selected Areas - Guangdong and Jiangxi , 1996

a At the corresponding city or county level.
b At the corresponding prefectual city level.

Sources:
1. Guangdong Statistical Bureau (1997), Statistical Yearbook of Guangdong 1997 , pp. 562-565.
2. Jiangxi Statistical Bureau (1997), Statistical Yearbook of Jiangxi 1997 , p. 247.

Footnotes

1 - Revised version of a paper presented at the 43rd Annual Conference of the Australian Agricultural and Resource Economics Society, Christchurch, New Zealand, 19-22 January 1999. We are most grateful to the journal's referee and editor for their constructive advice and suggestions in revising this paper. We also wish to thank Marjorie Wilson for her editorial assistance and the participants in the session of the conference for their valuable comments on the paper. The project was funded by the University of Sydney's research grant scheme.

2 - The "provincial governor responsibility system" was introduced in early 1995 in response to the difficult grain situations in late 1993 and in 1994. Its major purpose was to achieve a regional balance of grain supply and demand. Among other expectations, a provincial governor is required to procure 70-80 percent of marketed grains, to organise grain supply, and to keep prices under control within their provinces. This system, however, hinders regional grain trade. It was reported that both grain-deficit and surplus provinces have implemented blockages one way or the other since its introduction (Wang and Liu 1995). Its introduction has caused concerns about the danger of increased regional blockage and market separation (e.g., Ge 1996; Guo 1996).

3 - It is noted that, in principle, government agents and private traders would have different goals: price smoothing and assurance of supply on the public side, and profit-driven arbitrage on the private side. In reality, however, state grain departments have also been expected to earn a profit in recent years in order to reduce central government's subsidy burden and to pay the heavy debt owed by the government grain sector. The major task of separating state agents' policy function from their market operation has been put forward by the government in a move to further reform the state grain sector, but is likely to remain difficult.

4 - With the exception being the significant cointegration relationships between price series of Yingtan and Nanchang and Jiujiang, which require further investigation.