Analysis on Spatial Integration of Thailand and Vietnam Rice Market in Indonesia

Transcription

1 ISSN: Impact Factor 2012 (UJRI): ICV 2012: 6.03 Analysis on Spatial Integration of Thailand and Vietnam Rice Market in Indonesia By Dyah Ayu Suryaningrum Wen-I Chang Ratya Anindita

2 Research Article Analysis on Spatial Integration of Thailand and Vietnam Rice Market in Indonesia Dyah Ayu Suryaningrum 1*, Wen-I Chang 2, Ratya Anindita 3 1 Master Student, Department of Agribusiness Management, National Pingtung University of Science and Technology, 1st Shuefu Road, Neipu, Pingtung 912 Taiwan. 2 Assistant Professor, Department of Agribusiness Management, National Pingtung University of Science and Technology, 1st Shuefu Road, Neipu, Pingtung 912 Taiwan. 3 Professor, Department of Agriculture Economics, University of Brawijaya, Jl. Veteran No. 1 Malang, Indonesia, dearafrica@gmail.com, Phone: ratyaa@ub.ac.id, Phone: *Corresponding Author s suryaningrum@hotmail.com ABSTRACT This study investigates spatial market integration of Thailand and Vietnam rice market in Indonesia by using real price monthly data expanded from January 2000 to December Johansen co-integration test approach was employed to examine the long-run price relationship. To examine the short-term relationship among prices, we adopt Vector Error Correction Model (VECM). The results show that there is the existence of long-run relationship among Thailand, Vietnam, and Indonesian rice in the domestic market. While, VECM indicates that price change in Indonesia is mainly influenced by its seasonal dummy and import ban in Thai and Vietnam rice price are mainly influenced by Indonesian price at the previous period and Indonesian import ban policies. Adjustment speed of Vietnam toward long-run equilibrium is faster than Thailand; however speed adjustment of Indonesia is not significant. Key words: Spatial Integration, Rice Market, Johansen Co-integration, VECM

3 1. INTRODUCTION Rice is the major staple food in Indonesia and thus plays an important role in economic, politic and social aspects. Indonesia is the third largest producer of rice in the world after China and India. It first achieved rice self-sufficiency in 1984, but its self-sufficiency status has fluctuated since then, meaning that in some years it imported rice to meet local demand. Rice remains the staple food of the country with Indonesians eating on average about 139 kg of rice per year among the highest in the world (IRRI, 2012). However, today Indonesia is a net importer of rice. Indonesia has been importing rice for decades as supply has not grown enough to meet unexpected jumps in demand during disasters or times of failed crops. Since the population increases every year, it has caused the increasing of food demand in domestic market. Reliance of Indonesia on imports has caused domestic market may significantly affected by the international price volatility. In 2011, Thailand was the largest rice exporter country to Indonesia market, followed by Vietnam (AFSIS, 2012). Indonesia Central Bureau of Statistics shows that both of them are the main exporters of rice to Indonesia since 2007, followed by India, Pakistan, China, and US. Since September 1998, Indonesia has liberalized its rice market by removing monopoly of BULOG which is National Logistic Agency, and allowing private importers to enter the rice market. It is expected to bring about a better functioning of markets. Supply and demand that formerly depend on domestic market, now is become integrated with world market. Meanwhile, trade liberalization prevailed by the countries, both importers and exporters will indirectly affect supply and demand of food commodity that determines the world price. Therefore, the changes in world market will affect domestic market. Degree of market integration has often been used to measure the success of market liberalization and structural adjustment policies in developing countries (Mushtaq, 2006). It may be evaluated in terms of a relationship between the prices of spatially separated markets. Therefore, such information can help the government to decide the extent to which price transmission can be considered as efficient across different region. Johansen and Jusellius (1990) derived the maximum likelihood estimation using sequential tests for determining the number of co-integrating vectors. Therefore, it is allowed to test multiple co-integrating vectors in multivariate framework. Several studies on market integration o agricultural product apply Johansen co-integration technique (Alam, 2010; Gosh, 2011; Jaramillo, 2012). This study examines spatial market integration of Thailand and Vietnam rice market in Indonesia. The result hopefully will be able to provide recommendations to Indonesian government and related parties to design policies ensuring food security in the domestic market. 2. METHODOLOGY This study attempts to analyze market integration of Thailand and Vietnam rice market in Indonesia by using Johansen co-integration method which allows us to measure the price long-run relationship among the rice prices in Thailand, Vietnam, and Indonesia. Meanwhile, Vector Error Correction Model is derived to examine the short-run price relationship and speed of adjustment toward the equilibrium. The methods are specified as follows

4 2.1 Co-integration Approach Co-integration focuses on the long-run relationships between bivariate or multivariate price series. Thus, co-integration among non-stationary prices means that a linear combination of the series is stationary and therefore prices tend to move towards the long-run equilibrium relationship. The Maximum Likelihood method of co-integration, due to Johansen (1998) and Johansen and Juselius (1990), specified the kth order VAR representation of Pt as: Pt = Pt-1 + µ +βt + εt, (t = 1,2,, T) (2-1) According to Gosh (2011), the procedure for testing co-integration is based on the error correction model (ECM) representation of Pt given by: Pt = Γi Pt-1 + µ + βt + εt (2-2) where: Pt is an (nx1) vector of I(1) prices; I = -(I - Π1 - - Πi); i = 1,2,, k-1; Π= -(I - Π1 - - Πi); each of Πi is an (n x n) matrix of parameters; εt is an identically and independently distributed n-dimensional vector of residuals with zero mean and variance matrix, Ωε; µ is a constant term and t is trend. Since Pt-k is I(1), but Pt and Pt-I variables are I(0), equation (2-2) will be balanced if ΠPt-k is I(0). Therefore, it is the Π matrix that conveys information about the long run relationship among the variables in Pt. The rank of Π, r, determines the number of co-integrating vectors, as it determines how many linear combination of Pt are stationary. If r = n, the variables are non-stationary in levels. If r = 0, no linear combination of Pt is stationary. If 0 < rank (Π) = r < n, and there are n x r matrices α and β such that Π = αβ', then it can be said that there are r co-integrating relations among the elements of Pt. The co-integrating vector β has the property that β'pt is stationary even though Pt itself is non-stationary. The matrix α measures the strength of the co-integrating vectors in ECM, as it represents the speed of adjustment parameters. Further, two likelihood ratio test-statistic are used. The null hypothesis of at most r co-integrating vectors is tested by: Trace statistic ( )= ln(1 ) (2-3) (3-3) The null of r co-integrating vector against the alternative of r + 1 is tested by: Maximum eigenvalue statistic ( )= log(1 ) (2-4) where is the estimated eigenvalues obtained from the Π matrix; and T is the number of usable observations. 2.2 Vector Error Correction Model. According to Engle and Granger (1987), if two trending, say I (1), variables are co-integrated, their relationship may be validly described by an Error Correction Model (ECM), and vice versa

5 Suppose the general equation of error correction model is: yt = β xt + α (yt-1 γ1xt-1 ) + εt (2-5) where yt-1 γ1xt-1 is known as the error correction term. Provided that yt and xt are co-integrated with co-integrating coefficient γ, then yt-1 γ1xt-1 will be I(0) even though the constituents are I(1). Indeed, it is possible to have an intercept in either the co-integrating terms or in the model. Whether a constant is included or not could be determined on the basis of financial theory, considering the arguments on the importance of a constant. Briefly, the error correction model can be interpreted as follows. y is supposed to change between t-1 and t as a result of change in the values of the explanatory variables, x, between t-1 and t, and also in part to correct for any disequilibrium that existed during the previous period. While γ defines the long-run relationship between x and y, while β describes the short-run relationship between change is x and changes in y. Broadly, α describes the speed of adjustment back t equilibrium, and its strict definition is that it measures the proportion of last period s equilibrium error that is corrected for (Brooks, 2008). An error correction model for three variables xt, wt, and yt, that were co-integrated, may be expressed as: yt = β1 xt + β2 wt + α (yt-1 γ1xt-1 γ2wt-1) + εt (2-6) 3. DATA AND EMPIRICAL FINDINGS 3.1 Data Collection The data series that is used in this study consist of monthly data rice price for the period from January 2000 to December The data for Thailand and Vietnam are FOB price of 25% white rice broken collected from FAO (Food and Agriculture organization) and were transformed to Indonesian currency (IDR) by using exchange rate from Indonesian Central Bank (BI); while Indonesian retail price of medium rice is obtained from Ministry of Agriculture Republic of Indonesia. All the data are adjusted by using consumer price indices (CPI) to obtain the real price. Eviews 7.1 software was used to analyze the price series. 3.2 Analysis of Unit Root Test As a precondition to estimate model describing the relationship among prices and its determinants, Augmented Dickey Fuller, Phillips Perron, and KPSS tests is conducted to verify stationary or presence of unit root in the individual series of the model. The result discovers that the price series stationer at their first differences I (1), which is a pre-requisite for the co-integration analysis. The result of co-integration test can be quite sensitive to lag length. An obvious procedure is estimating a vector auto regression (VAR) on the difference series to define the lag order for the data analysis. This study uses the Schwartz information criteria (SC) and Hannan-Quinnn (HQ) information criteria which obtain lag length

6 3.3 CO-INTEGRATION ANALYSIS Johansen Co-integration Test For the multivariable model, co-integration of rice price in Indonesia, Thailand, and Vietnam is tested using Johansen s maximum likelihood procedure respect to seasonal dummy (S1, S2, S3, S4, S5, S6, S7, S8, S9, S10, S11), dummy of world food crisis (SFC), dummy of import ban of Indonesia period (SIB1), and dummy of import ban of Indonesia period (SIB2). The result of the trace test and the maximum eigenvalue are presented in Table 3-1. Table 3-1. Johansen Co-integration Test of Price Variables Trace Test Max-Eigen Value Test H0 λtrace Critical Value (0.05) p-value λmax Critical Value (0.05) p-value r = *** *** r = r = *** denotes statistical significance at 1% level The value of trace and maximum eigenvalue test indicate that the rank of Π is 1 (at 1% level of significance), which are and respectively. The null hypothesis of no co-integrating vector is rejected, but the null hypothesis of at most one co-integrating vector is not rejected. It indicates that there is one long-run relationship of the three rice price variables; means there is one linear equation in the long-run. This is because Indonesia reliance its rice import to Thailand and Vietnam, so the price policy that decide by these countries will affect rice price in domestic market Co-integration and Vector Error Correction Model The results of Johansen co-integration test indicate that we can continue the analysis to vector error correction model (VECM) which means that there is a long-run equilibrium relationship among Indonesia, Thailand, and Vietnam rice prices. The estimated long-run relationship (t-statistic in parenthesis) based on co-integration test may be written as: lninat = lnTHAt lnviett (3-1) [-2.858]*** [-1.336] Note: *** significant at 1% level 337

7 The result of equation (3-1) suggest that the coefficient of co-integration of Thai rice price (β1) is found to be and significant at 1% level; means that if there is 1% change of Thai rice price, there is about 67.4% Indonesian rice price change at the domestic market in the long-run. However, it has a negative relationship which deviates from long-run equilibrium. Meanwhile, the coefficient of co-integration of Vietnamese rice price (β2) is not statistically significant. Therefore, it may not affect the price change at Indonesia rice price in the long-run. The Error Correction Model (ECM) helps to determine if the Law of One Price (LOP) of a particular good in markets located in different spaces holds, as well as to know how fast the price in one location adjusts to the price of other location (Jaramillo, 2012). Furthermore, Vector Error Correction Model (VECM) analysis describes the dynamic equilibrium relationship of short-run and long-run in a system of equations. Although there is a long-run balance among the prices, but there are deviations from the short-run equilibrium relationship. It can be conclude that VECM is a combination of short-run and long-run relationship between the prices of different markets (Anwar, 2005). VECM regress changes in the price variable lagged deviation from long-run equilibrium relationship and also lag deviation of prices in short-run period. Deviation from equilibrium, as the reflection coefficient by VECM will bring changes to the balance between these co-integrated variables. The coefficients of error correction term in the VECM are a measure of the adjustment speed toward long-run equilibrium relationship between markets (Enders, 1995). According to Anwar (2005), the adjustment speed toward long-run equilibrium (α) is shown by the absolute value of error correction, which is interpreted as disequilibrium between the actual prices with long-run equilibrium level. The larger coefficient is indicates the speed of adjustment toward long-run equilibrium and vice versa. Table 3-2 shows the result of VECM coefficient of rice market integration of Indonesia, Thailand and Vietnam in domestic rice market. The speed of adjustment toward the long-run equilibrium of Indonesian rice price is 0.005; however, it not statistically significant. Meanwhile, in short-run term, rice price changes in Indonesia are mainly influenced by seasonal dummy on the first period (S1) about 9.4% and dummy of import ban in (SIB1) about 5.9%. It can be explained by the facts of the yearly cycle. At the beginning of each year, rice price in Indonesia always experience an increasing price due to scarcity before harvest time (Ministry of Agriculture Republic of Indonesia, 2012). Therefore to meet the domestic demand, government need to import rice. It indicates the importance of long-run co-integration relationship in process of determining rice price in Indonesia. Moreover, dummy of import ban in is also significant. In January 2004, Indonesian government implied import ban in order to stabilize rice price, since it was over supply in domestic market, and thus its price was decreasing. The government ban the rice import since the domestic stocks and productions was forecasted enough to meet domestic demand. This policy was abolished in June 2005 since the rice prices begin to increase. While, Thai rice price changes are significantly influenced by Indonesian rice price at previous period (lag 1) about 48.5%, its own rice price at previous period (lag 1) about 29.8%, dummy of seasonality on the first period about 7.2 %, dummy of food crisis in (SFC) about 6.9%, Indonesia s import ban in (S1) 338

8 about 10.8%, and Indonesia s import ban in (SIB2) about 7.6%. Therefore 1% price change in each significant variables Indonesia will affect Thai rice price in the short-run as much as its short-run coefficient which are presented in the Table 3. The speed adjustment of Thai rice price toward long run equilibrium is 21.5% and significant at 1% level. This price situation can be explained by the fact that Indonesia is net importers which import Thai rice in a large amount instead of other ASEAN countries, thus rice policy and demand shock at domestic market will affect the price at its exporting country. While global food crisis that occurred in also affect Thai rice price. At that time, the Thai price was increase sharply. On the other hand, Vietnam rice price is also mainly influenced by Indonesian price at previous period (lag 1) about 50.9%. Dummy of Indonesia s import ban both in (SIB1) and (SIB2) also significantly affect rice price change in Vietnam. Recently, Vietnam has been the largest exporter of rice in Indonesia market. As a consequence, Indonesian government decided to import rice from Vietnam in a large amount because the price was always below Thailand. Hence, demand shock at domestic market will affect the price at its exporting country, as Vietnam does. Table 3-2. Estimated Vector Error Correction Model Error Correction D(LNINA) D(LNTHA) D(LNVIET) CointEq *** 0.322*** [ 0.087] [ 3.406] [ ] D(LNINA(-1)) *** *** [-0.896] [-3.507] [-3.291] D(LNTHA(-1)) * 0.604*** [ 0.651] [ 1.875] [ 3.406] D(LNVIET(-1)) [-0.524] [ 0.992] [-1.266] C [-0.135] [-0.789] [-1.082] S *** 0.072* [ 2.651] [ 1.867] [ 0.934] S [ 0.341] [ 0.587] [-0.338] S [-0.272] [-0.108] [-0.276] S [-0.892] [-0.327] [ 0.053] S [-0.467] [-0.578] [-0.123] S

9 [ 0.573] [ 0.352] [-0.139] S [-0.523] [-0.186] [-0.474] S [-0.041] [-0.738] [-0.620] S [-0.034] [ 0.602] [ 1.237] S [-0.022] [-0.677] [-0.116] S [-0.030] [ 0.068] [ 0.769] SFC ** 0.076* [ 0.489] [ 1.989] [ 1.958] SIB ** 0.108*** 0.127*** [ 2.308] [ 3.898] [ 4.105] SIB ** 0.113*** [ 0.605] [ 2.380] [ 3.152] R-squared F-statistic Log likelihood Akaike AIC Schwarz SC Moreover, it is also influenced by Thai price at previous period (lag 1) about 60.4%. Thai was the leading exporter of rice in the world, thus the rice price decided by Thailand government will affect Vietnam in determining the rice price. Dummy of food crisis (SFC) is also influence rice price change in Vietnam as much as 7.6%. At that time, since India banned its export to world market has led rice supply in world market decrease. It caused the rice price increase dramatically, and Vietnam experiences the same situation. Meanwhile, the adjustment toward the long-run equilibrium of Vietnam rice prices is 32.2% which is statistically significant at 1% level. 4. CONCLUSION AND RECOMMENDATION Based on the result, it can be concluded that there is the existence of long-run relationship among Thailand, Vietnam, and Indonesian rice in the domestic market. It is shown by the result of co-integration analysis that there is one co-integrating equation of the three variables namely Indonesian rice price, Thailand rice price and Vietnam rice price. This is because Indonesia rely its rice import to Thailand and Vietnam, so the price policy that decide by these countries will affect rice price in domestic market in the long-run. While the results of Vector Error Correction Model show that Indonesia has the highest influence in 340

10 determining rice price in its domestic market, Thailand, and also Vietnam in the short term. Indonesia s characteristic as a net importer of rice which import rice in a huge amount except when it prevail its import ban in particular time as the inventory and production estimated to be enough to satisfied domestic demand has led its domestic market as one of the biggest rice market in Asia. Therefore, government policies to determine rice demand from world market will particularly influence its exporting countries, such as Thailand and Vietnam. However, the rice price in other country do not really affect Indonesian rice price in a short term. Therefore, the government needs to determine policies aimed to improve rice market through optimizing efficiency of economy, aspect of production, harvesting, after-harvesting, processing, and also marketing; in order to alleviate its reliance to world import. Hopefully, this policy can improve both consumers and producers welfare. On the other hand, BULOG which is the National Logistic Agency should improve its performance to manage domestic rice industry. Moreover, promoting food diversification program will be a good choice for the government to support the idea of reducing rice import quantity from other countries. REFERENCES AFSIS. (2012) Asean Food Security Information System (AFSIS) Project: Report on ASEAN Agricultural Commodity Outlook. Ministry of Agriculture and Cooperatives Thailand. Bangkok. Alam, M. J., J. Buysse, Andrew M. McKenzie, Eric J. Wailes & Guido Van Huylenbroeck. (2010) Linkage between World and Domestic Prices of Rice under the regime of Agricultural Trade Liberalization in Bangladesh. Contributed Paper Prepared for Presentation at the Australian Agricultural and Resource Economics Society Conference, Adelaide, South Australia. Adelaide. Anwar, C. (2005) Prospect of Natural Rubber Indonesia in International Market Integration and Export. Doctoral Dissertation. Bogor Agricultural Institute. Bogor Brooks, C. (2008) Introductory Econometrics for Finance Second Edition. Cambridge University Press. Cambridge. Enders, W. (1995) Applied Econometric Time Series First edition. John Wiley & Sons. New York. Engle, R. F., & C. W. J. Granger. (1987) Cointegration And Error Correction: Representation, Estimation and Testing. Journal of Econometrica, 55(2): Gosh, M. (2011) Agricultural Policy Reforms and Spatial Integration of Food Grain Markets in India. Journal of Economic Development, 36(2): IRRI. (2012) Indonesian Farmers Earn More Thanks to Rice Breeding. Available at: o-rice-breeding&lang=en. Accessed in May Jaramillo, J.L., Naude, A. Y., & Cote, V.S. (2012) Spatial Integration of Mexico and United States in Grain Market: The Case of Maize, Wheat, and Sorghum. International Association of Agricultural Economist Triennial Conference. Brazil. Johansen, S. (1988) Statistical Analysis of Cointegration Vectors. Journal of Economic Dynamics and Control, 12(2):

11 Johansen, S., & K. Juselius. (1990) Maximum Likelihood Estimation and Inference on Cointegration with Application to the Demand for Money. Oxford Bulletin of Economics and Statistics Ed. 52. Mushtaq, K., Abbas F., Abedullah, & Ghafoor. A. (2006) Testing the Law of One Price Rice Market Integration in Punjab, Pakistan. Pakistan Journal of Agricultural Science, 43(3): Ministry of Agriculture Republic of Indonesia. (2012) Available at Acsessed in December