Excerpt
Table of Contents
DEDICATION
ACKNOWLEDGMENT
ABSTRACT
ACRONYMS
CHAPTER ONE
1.0 INTRODUCTION
1.1 The Ugandan Agricultural Sector
1.2 Rice Production in Uganda
1.2.1 Rice production trends
1.2.2 Rice supply and demand
1.3 Economic importance of rice
1.4 Rice consumption in Uganda
1.5 Problem statement
1.6 General Objectives
1.6.1 Specific objectives
1.6.2 Hypotheses of the Study
1.7 Scope of the study
CHAPTER TWO
2.0 LITERATURE REVIEW
2.1 Market linkage and integration
2.1.1 The concept of Integration
2.1.2 Market integration
2.1.3 Measures of Market integration
2.1.4 Empirical Studies of market integration
2.2. Lag order determination Information Criteria
CHAPTER THREE
3.0 METHODOLOGY
3.1 Field Methods
3.1.1 Study area
3.1.2 Data Collection and handling
3.2 Analytical Methods
3.2.1 Diagnostics and testing
3.3 Testing for Co integration
3.3.1 Testing for long run co integration
3.3.2 Determination of strength of market integration
3.3.2a. Testing for bivariate co integration
3.3.2b. Testing for Granger causality relationships
3.3.2c Testing for impulse response functions of the market system
CHAPTER FOUR
RESULTS AND DISCUSSIONS
4.1 Status and extent of market integration
4.2 Co integration among price series
4.2.1. Testing for co integration by bivariate analysis
4.2.2 Determination of direction of causality relationship among markets
4.3 Results and discussion of multivariate analysis
4.3.1 Determination of co integrating market sets
4.3.2 Determination of dynamic adjustments by Impulse Response Functions
CHAPTER FIVE
5.0 SUMMARY, CONCLUSIONS AND RECOMMENDATIONS
5.1 Summary
5.2 Conclusion
5.3 Recommendations
REFERENCES
Appendix I: Map of Uganda Showing NAADS Program Expansion from
2001/02 to 2007/08
Appendix II: IRF graphs of Kampala market shocks against other markets
Appendix III: Proportional Difference test for market integration
Appendix IV: Bivariate analysis results for rice market co integration
DEDICATION
This thesis is dedicated to my parents, Mr. S. Waluube and Mrs. Alice Lovisa Nabirye who struggled to lay my foundation on whose sweat I have based to achieve higher strides.
To my dear wife, Winnie, and children; Humphrey, Penelope and Henry who were by my side through the thick & thin during the execution of this project.
ACKNOWLEDGMENT
I am highly indebted to my supervisors to Dr. B. Kiiza and Dr. Elepu Gabriel for their tireless supervision and guidance throughout this period of conducting research. Their guidance has been paramount to the completion of this postgraduate program. I wish to acknowledge the contribution of the entire staff of the Department of Agricultural Economics & Agribusiness, Faculty of Agriculture, Makerere University to the success of this research and postgraduate course in general.
Special acknowledgment goes to the Belgium Technical Cooperation, which sponsored my postgraduate program including research through its local scholarship grant program. I am thankful to I@Mak for their additional financial support towards this research.
I would like to express my thankfulness to my parents, family members especially Brothers Joel and John and Sisters Monica, Janet, Milly and Grace for their indispensable financial, material, and moral support particularly their prayers for my success.
To friends; Habajja Samuel, Samson Katengeza, Douglas, Fred Lugojja, Natanga Patrick, Kwizera Musaba, Robert Anyang, Dr. Isabirye Moses and Alice Amoding for their invaluable support and encouragements during this period.
Finally, I wish to acknowledge the assistance of the staff of International Food Policy Research Institute especially Dr. Todd Benson and David Luwandagga for their assistance in data re-organization from the FoodNet project documents.
ABSTRACT
It is crucial that formulation of market enhancing policies to improve performance hinges on a better understanding of the local market functioning. Comprehensive market performance is better understood by studying the extent of spatial market integration which in turn is affected by communication factors like telephony growth and extension services within the marketing system. Extension services are important in influencing levels of production and dissemination of trade information of both the inputs to and output from the enterprise. Unfortunately very little was known about how rice price transmission takes place. Given the importance of rice (as a food security crop & source of income) to the rural poor, many organizations have promoted it among rural farmers for improving their welfare, it was imperative that its price movements is studied to inform policy makers & implementers on the status of its spatial market integration.
The primary objective of this study was to evaluate the spatial market integration of rice markets under improving communication infrastructure particularly extension services (NAADS program) and telecommunication industry growth. This study specifically evaluated the degree of spatial market integration and strength of integration using bivariate analysis, Johansen Multiple test, and granger causality relationships. It also determined the impact of NAADS program and telecommunications on spatial rice market integration. Impulse response functions were generated to determine how long shocks in Kampala market can be eliminated in peripheral markets. Weekly prices data from sixteen markets from 2000 to 2006 were divided into three phases; full sample, pre – NAADS expansion and post NAADS program expansion to enable the study to capture the impact of these communication infrastructure.
Results from this study indicate that the markets are spatially integrated and with an incredible speed over the study period. Comparatively, there was more co integration and interdependence of rice markets in the post-NAADS expansion phase than the pre – NAADS expansion phase. A further analysis of the markets using multivariate analytical approach revealed a growing rice economic market constituting eight markets; Kampala, Jinja, Iganga, Lira and Rakai including Gulu, Mbale and Soroti markets with which growth began in the pre-expansion period. Overall the results reveal strong spatial market integration of the rice commodity in the post expanded phase of the communication infrastructure.
The study recommends the continuity of the NAADS program and improved telecommunications, continuous promotion of marketing associations and networking since the overall results depicted improved market integration. It also recommends the NAADS program to incorporate provision of timely agricultural and marketing information to strengthen the spatial integration of the rice markets. It also recommends related marketing studies to consider other structural determinant factors of market integration, and similar promoted enterprises to compare these findings.
Key terms : NAADS program, telecommunications growth, rice, markets, prices, spatial market integration, co integration, Granger causality and impulse response functions.
ACRONYMS
illustration not visible in this excerpt
CHAPTER ONE
1.0 INTRODUCTION
1.1 The Ugandan Agricultural Sector
Agriculture is still the mainstay of Uganda’s economy, accounting for 42% of GDP, over 85% of export earnings and providing about 80% of employment to the Ugandan population, 90% of who live in the rural areas (Anon, 2004). Food crops are predominant in the sector, contributing approximately 50% of agricultural GDP in 2003/4 with cash crops constituting 17% of this GDP (UBOS, 2004). FAO (2006) reports that food crops in Uganda are produced in two seasons, with estimated figures putting cereals production to 2.657 tones. With satisfactory food supplies, a large number of people are reported to have limited access to food due to low purchasing power.
1.2 Rice Production in Uganda
Rice growing in Uganda is mainly by smallholder farmers concentrated in the eastern and northern parts notably in the districts of Gulu, Iganga, Tororo, Mbale, Pallisa (Mukiibi, 2001), Lira, Kitgum, and Kumi (Imanywoha, 2001). Its production is in extensive swampy and high potential areas found around Lake Kyoga (Oryokot, 2001). The average land holdings range from 0.38 to .97 hectares (A.R. Ochollah et. al., 1997). However, since 1992 total acreage of rice in Uganda has marginally increased (MAAIF, 2000). The total area under rice production increased from 50,002 ha in 1992 to 72,000ha in 2000 while the total production increased from 68,000 tons to 108,000 tons with a unit increase in yield from 1.4 tons to1.5 tons per ha as indicated in the following table;
Table 1: Production statistics of rice in Uganda
illustration not visible in this excerpt
Source: MAAIF Statistics Department, 2000
According to UBOS (2002), the area planted to rice in Uganda in 2001 was 76,000 hectares and the volume of paddy rice production was about 114,000 tones. This showed a significant increase in production compared to 1996 in which planted area was 58,000 hectares and yield to be 82,000 tones. This represented a percentage increase of 31% and 39% for area planted and rice production respectively.
1.2.1 Rice production trends
The rice crop in Uganda was introduced on a larger scale only in early 1960s and by these periods, the annual harvested acreage was just above 2,000ha. This added up to less than 3,000 Metric tones per annum. The land area cultivated under rice rose substantially in the 1970s due to government interventions; however, this was followed by a production stagnation reducing the area planted with the crop by about 1,000ha from 18,300 ha in the 1970s to 17,000 ha in the 80s. This was later followed by a minimal rise in the production from 18,000 tons to 22,100 tons (UBOS 2005, ADC/IDEA Project 2001, P.8). The Africa Rice Center (WARDA, 2007) reported Uganda’s rice production between 2001 and 2005 to be at an average of 85.57 thousands of tones. This trend reveals that this crop enterprise is ever gaining importance through production and acreage.
1.2.2 Rice supply and demand
In Uganda, the domestic rice production can no longer meet the rising demands of especially all urban dwellers. Therefore, Uganda has become heavily dependent on imports of the rice food, making it one of her largest imports. Though rice production registered a steady growth between 1999 and 2003, averaging 114,000 MT, imports yet peaked during these years at more than 50,000 MT with an average of 41,000 MT arriving in the country each year. This implied that for Uganda to be able to meet the increasing demands, more than a quarter at times even a third, of the national rice supply had to be imported in the last five years as tabulated hereinafter. According to Jane Ininda, (2005) African rice consumption exceeds production and only 54 percent of sub-Saharan Africa rice consumption is supplied locally.”
Table 2: Uganda’s rice import demands against production from 1999 to 2003
illustration not visible in this excerpt
Sources: Uganda Bureau of Statistics, 2005 and FAOSTAT, 2004b
According to statistics, while the national production growth averaged 6%, the import growth rate was ten times higher i.e. 62%.
Regionally, Tanzania is the largest rice producer averaging over 650,000 tones per annum of unshelled rice from 2000 to 2004 but it is also the largest consumer of this grain (table 3). In contrast, Kenya is the smaller producer and largest importer of rice in the region (APEP, 2006).
Table 3: Selected East Africa’s Production and imports for rice from 2000 to 2004
illustration not visible in this excerpt
Source: FAOSTAT, online
From the available statistics, all countries in the region are deficit markets that have over the years resulted in informal cross border trade amongst them. For instance, informal rice exports to Kenya from the region are pronounced. Tanzania’s rice exports to Kenya had been increasing since 2000 and were worth UGX 13 million in 2004. This implies that a huge opportunity exists for the Uganda’s rice industry in the region (RATIN, 2005).
1.3 Economic importance of rice
Among cereals, rice and wheat share equal importance as leading food sources for humankind (www.ggs.com/history/rice in human life). Rice is a staple food for nearly one-half of the world’s population. In Africa, it is rice or wheat, followed by maize, yams and cassava that constitute the main food crops. In Uganda rice is increasingly becoming an important source of income for rural households in Uganda (Agricultural Policy Committee, 1998) and it is the fourth important crop after maize, finger millet and sorghum (Oryokot, 2001). It is second to none in economic return to peasant farmers on the basis of labour per annum per day per Ha and the main consumers of Uganda’s rice are mainly the urban populations. This was revealed further by the national household survey conducted UBOS in 1999/2000 (UBOS, 2002).
1.4 Rice consumption in Uganda
Rice consumption in Africa has a high-income elasticity, and its demand is linked to urbanization and economic growth (African Crops, 2004 & WARDA, 2005c). In Uganda the urban population more than doubled between 1980 and 1991 (Table 4), reaching almost two million persons while urban population rose by 11% (UBOS, 2002).
Table 4: Trends of urbanization in Uganda (1969 – 2002)
illustration not visible in this excerpt
Source: Uganda Bureau of Statistics, 2002
1.5 Problem statement
After the launch of the Plan for Modernization of Agriculture, PMA in 2000 (MAAIF/MFPED, 2000), improved dissemination of agricultural information was embarked on by introducing National Agricultural Advisory Services (NAADS) program and liberalization of telecommunication infrastructure which had started in the mid 90’s. Among the important agricultural information are the prices of agricultural commodities and products. Such information is easily transmitted through an efficient extension service system like NAADS and telephony usage by marketers. Therefore an improvement in telecommunications and expansion of NAADS program is supposed to bring out strong spatial market integration.
Attempts to understand the impact of such interventions that have lasted for almost ten years are important to inform stakeholders on policy impact and related issues. Available information reveals rice as the fourth important crop after maize, finger millet and sorghum (Oryokot, 2001) which enterprise importance has been increasing as source of income for rural poor households (APC, 1998) that can alleviate poverty among them and a food security crop (Odogola R. Wilfred, 2006). A study of the rice market performance in form of market integration plays a significant role as one way of achieving such an objective.
The Uganda government has tried to improve the marketing infrastructure since a poor infrastructure negatively affects market integration as reported by Goletti & Christinas-Tsigas, (1995). For instance with low telephone usage fewer marketers can easily share price information in time, like wise with NAADS program operating in a few districts its contribution to price dissemination, promotion of production is reduced. The government introduced the NAADS program, alongside the liberalized telecommunication industry to sharing of agricultural information including prices of products or commodities. In the rice industry specific efforts included promotion of rice production and marketing through NAADS program, Vice President’s Poverty eradication projects where rice seed was widely distributed as an in-kind credit (Yoko Kijima, 2008).
Various factors contribute to the efficiency of a commodity market. Among them are marketing infrastructure (like rural road network, telecommunications), the food policy, the degree of dissimilarity in production in the markets. The more diverse the markets are in terms of rice production, the more incentive it is for them to trade together, hence the higher the likeliness for spatial integration (Goletti, Ahmed & Farid, 1995). On this foundation, spatial market integration was used as surrogate for improved market efficiency.
Market integration is a precondition for effective reform in many of the former centrally planned economies (Baulch, 1997). Therefore, when integrated markets are identified carefully through study, a regional balance between surplus and deficit regions (Goletti et .al., 1995; Kherallah, 2000) can be planned whereby closely integrated are grouped together and designing market liberation, credit and transportation policies (Goletti, Ahmed and Farid, 1995). This avoids duplication of government interventions, and aid in making forecasts.
With such a study, the various institutions can be equipped with information to wisely inform the agents in the rice industry on the best ways of objectively improving or profitably benefit from the rice enterprise, contributing to poverty reduction. Because the study can measure the extent to which demand and supply shocks in one location are transmitted to other locations (Negassa et al., 2003) and respective price transmission signal durations, it helps policy makers to know the response of markets to various shocks and the adjustment periods necessary for a given policy success. This is important in informing policy makers on to not only design but also improve food security initiatives and poverty reduction interventions through agriculture.
Unfortunately, due to almost no research on rice markets and marketing, there was limited information on the spatial market integration. Previous studies on rice enterprise in Uganda focused mostly on finding ways of increasing productivity by analyzing productivity factors. The need to have studies to generate information on the market integration across the markets in the country was glowing and a detailed study on the degree of spatial integration of the rice markets was imperative. There was a further need to attempt to review the extent of this spatial integration following the commencement of the NAADS program under the increasing telecommunication improvement. Notably, it would depict the NAADS program and other marketing infrastructure interventions influence contributions and progress in enhancing market integration. In so doing, the study results contribute to strategic and tactical decision making in the agricultural sector and provide technical ground for further re-engineering their strategies of implementation.
1.6 General Objectives
The main objective of this study was to examine spatial rice market integration in Uganda following the growth and expansion of the telephony and road network in the country.
1.6.1 Specific objectives
The specific objectives were:
a) To determine the degree of spatial rice market integration in Uganda.
a) To determine causal relationships among the rice markets in Uganda
a) To determine the dynamic adjustments of rice markets to shocks in the long-run
1.6.2 Hypotheses of the Study
1) Spatial integration of the rice markets has improved over time.
1.7 Scope of the study
The study was aimed at reviewing the performance of the marketing system of the rice sub sector. With an intention of obtaining the entire country rice-market integration, sixteen districts were purposively selected from all regions namely eastern, northern, and western and central/Buganda regions. To achieve the objectives, secondary weekly rice prices data for use were collected from FoodNet project of IITA.
CHAPTER TWO
2.0 LITERATURE REVIEW
2.1 Market linkage and integration
2.1.1 The concept of Integration
In a market economy, the marketing system serves as mechanism to transmit to market participants’ information that is useful in decision making. Transparent, accurate and timely price signals play a leading role in shaping the conduct and performance of an efficient marketing system. In a competitive market environment, the pricing mechanism enables the transmission of orders and gives directions that determine the flow of market activities. These signals guide and regulate production, consumption and marketing decisions over time, form and place (Kohls and Uhl, 1990). Identifying the causes of differences in prices in interregional or spatial markets has therefore become an important economic analytical tool used to provide a better understanding of markets including agricultural product markets. The next section explores this concept.
In his explanation using population data as an example, Clive W. J. Granger (2004) explained that market price data could be available annually or less frequently. He explained that when such data series is smooth, moving with local trends or with long swings but swings are not regular, then that data that is not smooth as required by standard statistical procedure is said to be integrated or nonstationary. The difference between a pair of integrated series can be stationary and this property is known as Co integration. For co integration, a pair of integrated or smooth series must have the property that a linear combination of them is stationary. Most pairs of integrated series will not have the property of stationarity, so that co integration should be considered as a surprise when it occurs.
2.1.2 Market integration
Integrated markets are those whose prices for a given commodity or commodities in different localities have a stable long run relationship (Goletti & Christinas-Tsigas, 1995). Spatial market integration refers to co-movement of prices, and more generally, to the smooth transmission of price signals and information across spatially separated markets (Goletti, Ahmed & Farid, 1995). This phenomenon is also explained by correlations of price series of different markets in which it is ideological that integrated markets exhibit prices that move together although in some cases, parallel movements in prices can occur for several other reasons other than the integration of markets.
Market integration can also be defined as a measure of the extent to which demand and supply shocks in one location are transmitted to other locations (Negassa et al., 2003). Barrett (1996) distinguished market integration into (i) vertical market integration involving different stages in marketing and processing channels, spatial integration relating to spatially distinct markets, and (ii) inter-temporal market integration which refers to arbitrage across periods. This study intends to examine spatial market integration.
Gonzalez-Rivera and Helfand (2001) noted that while there is a general agreement that market integration somehow relates to the flow of goods & information across space, time and form, getting a widely acceptable definition of integration proved to be elusive. They therefore proposed another definition that emphasizes trade and information. Accordingly, Gonzalez-Rivera et. al. (2001) defined integrated market as a market with a set of locations that share both the same traded commodity and same long run information. In the co integration framework, the second condition requires the existence of one and only one co integrating factor common to all price series within the market. These criteria are important in identifying the sets of locations that are spatially linked directly or indirectly by trade. Under this co integration study, market integration analysis was extended from bivariate to multivariate framework, which looks at market integration in terms of a continuum of degrees of integration (Shahidur, 2002). The underlying idea is that for a set of spatially separated locations, not all locations belong to the same economic market and among all those locations that belong to the same market, some are more integrated than others.
One of the main consequences of poor market integration is the difficulty with which information and trade flows occur in spatially separated markets (Goletti & Chritinas – Tsigas, 1995). However, knowledge on market integration is important for grouping closely integrated and designing market liberation or price stabilization, credit and transportation policies (Goletti, Ahmed and Farid, 1995), to avoid duplication of government interventions, and making forecasts to guide other stakeholders. This had been emphasized earlier by Goodwin and Schroeder (1991) that markets that are not integrated may convey inaccurate price information that distort producer marketing decisions and contribute to inefficient product movements resulting in poor market performance. Market integration knowledge is a precondition for effective reform in many of the former centrally planned economies (Baulch, 1997).
Proper market linkage is one way of improving performance of product markets to benefit farmers and traders (arbitragers). For instance if there is an uneven distribution of food supplies in the country, assurance of regional balance between deficit and surplus regions (Goletti et. al., 1995; Kherallah et. al., 2000 and Istiqomah et. al., 2005) is achievable through proper knowledge of markets generated through market integration studies. Therefore, measurement of market integration can be viewed as a basic tool for an understanding of how markets work (Ravallion, 1986). It is upon such information that important implications for economic development (Gonzalez G. Rivera and Helfand, 2001) can be drawn for a country.
2.1.3 Measures of Market integration
Goletti and Christinas –Tsigas (1995) explained the various methods of measuring markets integration namely; descriptive analysis of market network, use of correlation and co-integration tests of the prices of given commodities in different markets, dynamic adjustment and composite indices. Use of correlation coefficients is based on comparison of correlation of prices of markets. This is because it is assumed that integrated markets have prices that move together over time.
The study of market integration based on correlation coefficients, however, has faced criticism because according to Goletti et. al., (1995) it does not address factors like price inflation, seasonality, population growth and procurement policy for government managed procurements of the commodity whose prices are considered. This is why in some studies it is not applied. Criticisms are raised because of its failure to address the influences of the above factors particularly in agricultural commodity markets where peak and deficit seasons often occur at the same time results in spurious correlations. This is why such specious correlations are eliminated by considering price differences instead of price levels in computations of correlation coefficients. Spurious correlations are not the only problems in measuring market price correlations; other serious problems are related to the non stationary nature of prices, and the failure to determine the direction of integration among the markets.
Use of market network analysis is a good method of studying integration but necessary data for example on trade flows and marketing costs is not readily available and tends to be unreliable.
In some studies, understanding market integration is a sequence of studies for instance bivariate analysis of integration is conducted to identify existence of equilibrium relationships between market pairs over both the medium and long term (K. Weber and D. Lee, 2006). Multivariate co integration is then conducted to establish the boundaries of the continuous market, which is also called the economic market (Gonzales Rivera and Helfand, 2001). In the final study, markets with stable long run relationships are the ones that are subjected to impulse response analysis to draw insight into the speed at which such markets react to deviations from the equilibrium to restore the long run relations.
2.1.4 Empirical Studies of market integration
Various authors and researchers have used the co integration analysis in the study of market integration. Others extend their analyses by conducting the effect of shocks in markets to the rest of the markets through impulse response analysis using impulse response functions. This sub section reviews some studies that have applied the concept of co integration with the aim to compare various ways the concept is used in relation to spatial market analysis.
In their paper on regional cattle markets, Goodwin & Schroeder (1991) aimed at evaluating the co integration and spatial linkages among 11 regional cattle slaughter markets by determining the effect of several market characteristics on the co integration of the markets. On running several spatial price relationship tests, they found that very many markets were not co integrated between 1980 and 1987. They concluded by suggesting that trade and information costs would decrease and packers could coordinate the price behavior across regions as the market concentration increases. Based on the unexpected results obtained, it was argued based on Ardeni (1989) that conventional approaches of testing for spatial integration had misrepresentation or ignorance of time series properties of regional price data.
Goletti, Ahmed & Farid (1995) studied 64 coarse rice markets at district level in Bangladesh with an aim of determining the main factors responsible for market integration. They hypothesized that marketing infrastructure, volatility of government intervention and degree of self sufficiency are the major determinants of market integration. They addressed this hypothesis by application of a two-stage approach in which four measures of market integration were used namely; correlation coefficients, co integration coefficients to capture the long term relation among the price series, long term multipliers that express the cumulative response of one market to price shocks originating from another market. They lastly conducted the fourth measure by estimating the speed of adjustment of markets to the long term multipliers.
Results from their study indicate that different measures respond differently to market integration structures, and the level of rice market integration for the period under study was moderate. The results further revealed that distance negatively affects market integration while production shocks positively affect market integration of the given markets. Once distance between markets under consideration is less than the critical distance of 250 kilometers from each other, the level of market integration is higher. It is noted that to get results that are more meaningful price and structural data that cover longer period are necessary.
While studying South African apple markets within the fresh produce markets, David I. Uchezuba (2005) used various methods to measure market integration of apples on South African fresh produce markets. He used ADF and Johansen tests to investigate long run relations in the markets. He then fitted threshold vector error correction models to investigate the dynamic relationship among apple markets and applied regime switching and impulse response functions to evaluate switches among regimes and responses to shocks by various markets. Major findings in his study were that market integration in the markets was evident with both unidirectional and bidirectional causality among selected markets. He discovered that there was sufficient arbitrage among the South African fresh produce markets, which is consistent with market integration.
In Uganda, increasing need for regional agricultural production specialization has been agitated for since early 2000 under the Plan for Modernization of Agriculture. This was done through zoning the whole country into twelve agro-ecological zones (NARS, 2002). Under such circumstances, some enterprises are assumed to perform better in some regions than in others. With such policy arrangements, market integration studies are very important to inform policy on which way to advise farmers on regional basis. A similar scenario study was conducted by K. Weber and D. Lee (2006) in which they used market integration and response analysis to examine whether by comparative advantage, markets in the U.S. and Mexico adhered to the principles of specialization accordingly.
In their study, ten rice markets in U.S. and Mexico were analyzed by examining market integration through price convergence and co-movement through a sequential process. Through use of bivariate and multivariate analyses, they established the boundary of the continuous market and applied impulse response analysis techniques to get insight on how the speed of deviations from the equilibrium could be corrected. Similar methods including Johansen sequential testing procedure were used by Titus O. Awokuse (2007) on the China’s rice markets. Their results were that under the long run equilibrium most of the Mexican markets are bound to the U.S. markets. Awokuse (2007) study further revealed that the U.S. and Chinese rice grain markets were co integrated with continuity. This means that some markets may be continuously co integrated while others are segmented from the continuous integration boundary.
Theingi Myint and Siegfried Bauer (2005) used co integration method for integration analysis of Myanmar rice market. In their explanation, they stated that integration means that if two variable series say Abbildung in dieser Leseprobe nicht enthaltenand Abbildung in dieser Leseprobe nicht enthaltenare each non stationary in levels but stationary in first differences, that is Abbildung in dieser Leseprobe nicht enthaltenandAbbildung in dieser Leseprobe nicht enthalten, there exists a linear combination between the two series that is stationary. They emphasized that in an efficient market system, prices move together in that trade takes place if prices in the importing region equals prices in the exporting region plus the unit transport cost incurred by moving between the two (Ravallion, 1986).
The result of efficient trade and arbitrage activities is that prices at different market places cohere (P.J. Dawson and P.K Dey, 2002). This is why co integration analysis allows verifying whether markets have efficient trade or not. In their study, co integration analysis was conducted using weekly prices of pairs of markets Mandalay-Yangon and Yangon – Mandalay for one type of rice product (pawsan). The most striking feature of the results is that the rice prices were highly co integrated between Yangon and Mandalay market during the 2001 and 2004.
To provide an insight understanding of the effect of trade liberalization on the rice market in Java provinces, Indonesia, Istiqomah, Manfred Zeller and Stephan von Cramon – Taubadel (2005) applied the multivariate Johansen maximum likelihood method to comparatively analyze the volatility and degree of integration in two situations i.e. in before and after –trade liberalization. In their method, it is evident that co integration analysis is conducted after confirming that price series are integrated to first order in the first differences. The results of this study indicated that among the five markets studied there were four co integrating vectors among prices series of five markets on Java in the pre liberalization and only two co integrating vectors in the post liberalization period. They concluded that the less integration could have been due to delayed response of markets to the new policy of liberalization.
Vietnam is one of the world’s largest rice producing and supplier countries but at the same time, the country has been undergoing various reforms. These included rice market liberalization after 1989 and with the various initiatives involved in these reforms, better market performance was expected. To evaluate this transition policy, Clemens Lutz et. al., (2006) used the Johansen maximum likelihood estimators to test for multiple co integrating vectors to assess the long run market integration in the country. The results of this study showed that all market places under study were integrated in the long run. The prices in Ho Chi Minh (HCM) city and the markets in the Mekong River Delta were strongly correlated even in their short run. On application of the VECM, all markets, except for the export price series, reacted strongly to deviations from at least one of the long run co integrating relations.
In their study of the food market integration with focus on the Vietnamese paddy market, Le Dang Trung et. al, (2007), used transfer costs to generalize the well known model of spatial market integration due to Ravallion to allow for the possibility of threshold effects. On application of the unrestricted version of this model using monthly paddy prices for eight markets between 1993 and 2006, a weak evidence of market integration was found in the north and south of Vietnam without any threshold effects. Their further analysis of paddy markets, revealed threshold effects and stronger forms of spatial market integration within the north and within the south of Vietnam.
Shahidur Rashid (2004) studying Uganda’s maize market integration in the post liberalization period discussed the various groups of co integration analyses. These included Ravallion’s radial market integration model in a co integration framework. The test for co integration is carried out as a test for unit roots on the saved residuals in which case if residuals are stationary, the markets will be co integrated. The second group of studies is one that uses Johansen’s multivariate co integration method to draw inference about the strength of integration among regional markets. The idea behind this method comes from Stock and Watson (1987) who demonstrated that if a set of n economic variables is co integrated with exactly n-1 co integrating vectors, these variables must therefore share a common trend. Shahidur pointed out that full integration of markets requires exactly n-1 co integrating vectors and therefore any number that is less than n-1 implies weak integration. In this study, the results indicated that markets that were not integrated in the earlier period of liberalization became strongly integrated in the subsequent years.
In Brazil, rice is one of the major cereals produced in the country but this production is concentrated in small number of states. Gonzalez-Rivera et. al., (2001) used this Brazilian rice market to illustrate that the multivariate approach is better than bivariate analysis in studying market integration. To improve on the knowledge of market integration two features namely searching for boundaries of the continuous market and use of persistence profiles to study the degrees of integration of locations that belong to the market were introduced through his study. This means that even among the integration the levels of integration differ by varying degrees. To get the continuous market, they proposed a sequential procedure based on Johansen (1988, 1991) to search for the single common factor.
In a continuous market, one and only one integrating factor must be common to all the price series. Given n-locations in a market, there must be n-1 co integrating vectors. When normalized, the n-1 co integrating vectors with respect to given a location, we find that all the locations were integrated pair-wise. Under this arrangement, it is very difficult to determine which locations belong to the same market using the bivariate approach because of all the Abbildung in dieser Leseprobe nicht enthaltenpair wise combinations only n-1 combinations would be relevant. Under this circumstance, it becomes complicated and gives inconclusive results contrary to the multivariate analysis. Bivariate analysis limits each equation in the VECM to only one error correction term and lags from only two locations considered. This grossly miss-specifies the model for some special market structures. The integrating factor is eliminated when the co integrating relation is estimated yet a common long run trend that gives rise to co integration has to be found.
Gonzalez-Rivera et. al., (2001) stressed that to find which locations belong to the same market; we begin with a maximum set of locations, n, and testing for n -1 co integrating vectors. This is done by performing Johansen Likelihood ratio test based on the trace statistic. If the number of co integrating vectors is less than n -1, there is need to identify those locations that must be removed from the system by performing a sequential procedure. The procedure is started with the core set of m (m<n) and test for the number co integrating vectors, if the number is equal to m – 1 co integrating vectors, then we add another location. With m+1 locations, either the new one shares a common trend with the previous m locations or it does not. It is mentioned that in the first case, we should find m co integrating vectors, in the second we should continue to find m -1 co integrating vectors thus adding a second common trend to the m+1 co integrating vectors. If one common trend is found, the procedure is repeated through addition of another location at a time. In one of their investigations, they found that fifteen states belonged to the same economic market while four did not appear to belong to this market.
[...]