Thailand_Tuna_PriceTransmission_RevisedSubmission_1Nov12

Price Transmission Relationships in Tuna Value Chain in Thailand Umesh Bastola1,3, Madan Mohan Dey1, and Amporn Laowapong2 Aquaculture / Fisheries Center, University of Arkansas at Pine Bluff, 1200 North University Drive, Mail Slot 4912, Pine Bluff, AR 71601, USA Economic Analysis of Fishery Commodities Section, Department of Fishery, Kasetsart Campus, Chatuchak Bangkok 10900 Thailand Presently Ph D Student, School of Economic Sciences, Washington State University, Pullman WA 99164, USA Correspondence: Madan Mohan Dey E-mail: mdey@uaex.edu ABSTRACT Thailand is the largest importer of skipjack tuna as well as the largest exporter of canned tuna in the world This particular study focuses on studying the trend of tuna prices in the import (frozen skipjack) and export (canned) level, exploring causal relationships, and examining asymmetry in price transmission in the tuna value chain in Thailand The two price series were found to be cointegrated with canned tuna price Granger causing the skipjack tuna price A vector error correction model (VECM) showed that 41% of shock in the long-term equilibrium adjusted in the current month and fully adjusted in 1.4 months Test of price transmission asymmetry using von Cramon-Taubadel (1999) approach suggested symmetric relationship indicating identical response in the skipjack tuna price due to any policy intervention that either increases or decreases the canned tuna price in Thailand Key words: Price Transmission Asymmetry, Cointegration, Tuna, Thailand, Value Chain Running Head: Price transmission in tuna fishery in Thailand Introduction Tuna is one of the most valuable seafood commodities in Thailand It was the second highest seafood product to earn export value (24% of total seafood export value) after shrimp (38%) in the country in 2009, when 224,542 Million Baht1 (35 % of total food export) of seafood were exported (Laowapong, 2010) Similarly, tuna is the most dominant seafood imported in Thailand which accounted for about 52% of total fish import value (35.8 Billion Baht) in 2009 Thailand is the world’s leading importer of frozen tuna and exporter of canned tuna (GLOBEFISH, 2010) The major export of Thai tuna is in canned form (98%) in air-tight containers with a negligible proportion (2%) in frozen vacuum pouch (Figure 1) Almost 95% of the raw material used in canning factories comes from import in the form of frozen whole tuna while 50% of the domestic catch is the supply to these factories Of the total processed tuna, about 90% is exported and the remaining is consumed in the domestic market The major importing countries of Thai canned tuna in the order of decreasing volume imported in 2009 are the U.S., Egypt, Libya, Australia, Canada, and Japan (GLOBEFISH, 2010) The main market segments of the tuna value chain in Thailand are i) first-hand market from domestic supply, ii) import market of frozen tuna, iii) final domestic market, and iv) final export market of canned tuna (Figure 1) Among these, import market of frozen tuna and final export market of canned tuna are the two most important market segments both in terms of volume and value of fish marketed This study investigates trends and relationships between tuna prices in these two levels of tuna supply chain in Thailand 1 US$=34.3351 Baht in 2009, 31.7270 in 2010 Exploring relationships among prices at different levels of supply chain render valuable information for policy implication In one hand, it allows for investigating existence of potential market power in the supply chain while on the other hand, it also identifies future effects of policy intervention in one level of the chain on the other Besides, there may have existed seasonal trend in prices and marketing margins at different levels of supply chain, the knowledge of which may help to formulate policy options that benefits for all those who rely on seafood industry for their livelihood Price transmission relationships have been widely studied in agricultural sector while a few of such studies are also conducted in seafood marketing sector Meyer and Taubadel (2002) have presented a survey of the literature on asymmetry in price transmission They have also elucidated the possible causes of asymmetry in price transmission as: existence of market power, adjustment and menu costs, and other causes such as government intervention, asymmetric information, and inventory management Some of the studies that examined price transmission relationships in seafood marketing include: Jaffry (2005) in French hake; Floros (2007) in different fish species in Greece and UK; Asche, Jaffry, and Hartmann (2007) in salmon in Norway; and Shinoj et al (2008) in major fish markets in India However, no study has yet analyzed price transmission relationships in seafood markets in Thailand The main purpose of this study is to develop an improved knowledge of the dynamics of tuna value chain in Thailand The specific objectives include: i) to examine trends and seasonality in prices of tuna at two important levels of value chain (import market of frozen tuna and export market of canned tuna); ii) to explore the causal relationships among the import price of frozen tuna and the export price of canned tuna; and iii) to examine the asymmetry of price transmission relationships between the two price series Theoretical framework There may exist a causal relationship, either short-term or long term, between these two price series Identifying these relationships may have valuable implications for any policy interventions Hence, this study has tried to explore if any such relationships exists among the two price variables of tuna value chain Time series data are generally non-stationary Basically, various economic forces play role to affect variables like price over time Hence, employing ordinary least squares regression techniques to investigate causal relationships among variables in time series data may result into spurious results Therefore, testing whether data is stationary or contains unit root is the first step in any time series analysis Unit root test of price series In this study, we have used Augmented Dicky-Fuller (ADF) (1979, 1981) test to check the presence of unit root in the price series If there is no unit root, then the price series are said to be stationary in levels or integrated of order zero (denoted I(0)) If there is a unit root but differencing the series once makes it stationary, then it is said to be integrated of order one (I(1)) The null hypothesis of the ADF test is that the series contains unit root (I(1)) versus alternative hypothesis of no unit root (I(0)) The ADF tests for each series in log values were performed in levels and first differences with a constant as well as with a constant and a trend The appropriate number of lags chosen was based on Schwarz Information Criteria (1978) for each test Granger Causality Test Once the data are found to have unit root, we can test for Granger causality to see if there exists any causal relationships between variables A time series X is said to Granger-cause Y if X values provide statistically significant information about future values of Y Its mathematical formulation is based on linear regression modeling of stochastic processes (Granger 1969) Hence, one way of application of Granger causality test in marketing research may be to identify whether a price variable causes another price variable to change or not The null hypothesis is formulated such that its rejection allows us to conclude that the particular variable Granger causes the other variable Here, we have used pairwise Granger causality test to detect the causal relationships between the two price series Test of Cointegration and Vector Error Correction Model Two or more unit root series are said to be cointegrated if their linear combinations are stationary Cointegration analysis involves establishing statistically sound long-run relationships between time series data Series that are non-stationary but have a common stochastic trend may be co-integrated There are two approaches for testing cointegration, namely, Engle and Granger test (1987) and Johansen test (Johansen 1988) In this study, we followed Johansen’s cointegration test procedure This test is based on vector autoregressive (VAR) system In this framework, two asymptotically equivalent tests for cointegration, a trace test and a maximum eigen value test have been used A vector error correction model for each cointegrating equations have been developed to see how various forces react to establish a long-run equilibrium between the given cointegrated pairs of price series Test of price asymmetry Method followed by von Cramon-Taubadel (1999) outlined as below have been used in this study to study the price transmission asymmetry in the supply chain: P    P    ECT t     ECT  t    ( L)P   ( L) P  i, t j, t 2 i, t  j, t  t Where, ECT is the error correction term obtained from the cointegration equation The test of asymmetry in price transmission is the test of null and alternative hypotheses given as: H :     2 H a :     2 The rejection of null hypothesis indicates that the price transmission is asymmetric while the failure to reject the null hypothesis confirms symmetric price transmission relationship Source data Monthly time-series data obtained from the FAO GLOBEFISH (GLOBEFISH AN 11101, 11102) with time period from Jan 2005-May 2010 were used in this study Two price series namely, frozen skipjack tuna price and canned tuna price in the Thailand tuna supply chain, were analyzed The frozen skipjack price averaged 2.1 USD per kg with a standard deviation of 0.34 USD while the canned tuna price averaged 12.2 USD per carton2 with a standard deviation of 4.88 USD Both the series were transformed to natural logarithms before conducting econometric analysis Trends and seasonality in prices Figure depicts the price movement patterns of each price series over time There is no large fluctuation in prices over time Both the price variables are found to move in a similar fashion Monthly seasonal price index was calculated for each series For that, One carton contained 48 six-ounce-chunks of tuna prices at each individual month were averaged, which were then divided by the sum of such averages of all 12 months, and then multiplied the ratio by 1200 As shown in Fig 3, though canned tuna prices are relative stable across different months, the skipjack tuna prices reveal some seasonal variation The skipjack prices tend to be generally lower from January to July and comparatively higher in other months Results and Discussions The unit root tests have been performed at level data and the first difference including both: 1) a constant and 2) a constant and a linear trend The tests are applied to the variables in logarithms The results of the ADF test indicate that the null hypothesis of having unit root could not be rejected for both series at level but rejected at first difference, and hence, both are integrated of order one (I(1)) (Table 1) Table shows the result from pairwise Granger causality tests We found that the canned tuna price Granger caused skipjack tuna price, but not vice versa Hence the Granger causality was unidirectional Testing for the presence of cointegrating relationships between the two variables, Johansen’s cointegration test yielded the following results (Table 3) From Table 3, it can be seen that there exists one cointegration equation between skipjack and canned tuna prices as given by trace test and maximum eigen value test The resulting cointegrating equation is as follows: ln Skipjack Tuna = -2.9189+ 1.0006 ln Canned Tuna Thus, from the above cointegrating equations, 1% increase in canned tuna price causes about the same percent increase in skipjack price This indicates that there is a unitary elastic relationship between these two prices The error correction model for skipjack tuna resulted in the following equation:  ln stt 0.011 ln st t   0.079  ln ctt   0.413* (ln stt   2.918  1.006 ln ctt  )  0.120 R 0.2393 *  0.254  (0.106) * p  value  0.05 Similarly, the error-correction model in case of canned tuna is expressed by the following relationship:  ln ctt 0.004  0.049 ln ctt   0.119  ln stt   0.004(ln stt   2.918  1.006 ln ctt  )  0.007  0.139  0.066 (0.058) R 0.061 The models illustrate how deviations from long run equilibrium in series are restored due to adjustment processes The first ECM is significant with a R-square value of 0.2393 The error correction coefficient (or the long-run adjustment parameter) in the model is also significant at the 1% level The error correction coefficient shows that 41% of any shock in the equilibrium is corrected in the current period The value (λ = (10.413)/0.413 = 1.42} implies that for the shock in the system to be fully adjusted, it requires 1.42 months of time None of the other coefficients are statistically significant In the second ECM with canned tuna as the dependent variable, we found that none of the coefficients as well as the model R-square was statistically significant This result is in consistent with the results shown by the Granger causality test where found unidirectional causality from canned tuna to skipjack tuna price The price transmission relationship given by the von Cramon-Taubadel (1999) approach is expressed as: P  0.013  0.032P  0.049 ECT t   0.180 ECT  t  ST , t CT , t (0.043) (0.266) (0.159) (0.197)  0.082P  0.277P ST , t  CT , t  (0.142) (0.265) Here, the estimated coefficients of the positive and negative residual terms were not statistically significant The null hypothesis of equality of these two coefficients could not be rejected at the 5% level (F1, 58 = 0.150; p-value = 0.697) Hence, it indicated that the price transmission relationship in tuna supply chain was symmetric SUMMARY AND CONCLUSION The trends and relationships among prices of tuna in Thailand, have been explored in this study Imported frozen skipjack prices are found to be relatively less fluctuating over time as compared to canned tuna price Nevertheless, prices in these two series are found to move in a similar fashion The skipjack tuna price showed seasonality in prices, being lower from February to July A unidirectional Granger causality was observed where canned tuna price was found to influence the skipjack tuna price Moreover, the two price series were cointegrated depicting a long-run equilibrium relationship The statistically significant 10 terms in VECM showed that 41 % of deviations in the long run equilibrium were adjusted in the current period and gap above the equilibrium in one month was found to produce downward pressure on skipjack tuna price in the subsequent month Furthermore, the price transmission in the tuna value chain in Thailand was found to be symmetric Thus, the response of skipjack price to increase in the price of canned tuna was identical to that to the price decrease To conclude, any policy that affect the one level of tuna supply chain in Thailand would affect the other level However, any shock in the system would be adjusted fast enough and maintain the long-run equilibrium Any policy intervention that triggers price increase in one level of the supply chain will not have differential response in the other level of the chain to the price decrease, thus, providing information on potential effect of any policy intervention on tuna industry in Thailand An on-going fish supply-demand projection analysis, currently being implemented by the International Food Policy Institute/World Bank/University of Arkansas at Pine Bluff/FAO, indicates that the retail price of tuna is expected to increase over the next 20 years or so (up to 2030) Our finding of symmetric price transmission between frozen skipjack tuna and canned tuna in Thailand has a positive implication for tuna fishermen in the region The results reveal that tuna canneries in Thailand not have market power in the buyer market; increases in export price of canned tuna are expected to pass fully to the import market of frozen skipjack tuna Given that the volume of frozen import of tuna is about times higher than that of domestic catch in Thailand, the welfare of tuna fishermen in Thailand is expected to be improved with increase in price of tuna in the global market 11 REFERENCES Asche, F., S Jaffry, and J Hartmann 2007 Price transmission and market integration: vertical and horizontal price linkages for salmon Applied Economis 39:2535-2545 Dickey, D.A., and Fuller, W.A (1979) Distribution for the Estimates for Auto Regressive Time Series with a Unit Root Journal of the American Statistical Association 74:427-31 Dickey, D.A., and Fuller, W.A (1981) Likelyhood Ratio Statistics for Autoregressive Time series with a Unit Root Econometrica 49:1057-72 Engle, R.F., and Granger, C.W.J (1987) Cointegration and Error Correction Representation, Estimation and Testing Econometrica 55: 251-76 Floros, C 2007 Causality and price transmission between fish prices: new evidence from Greece and UK European Journal of Social Sciences 4(2):147-159 GLOBEFISH 2010 Globefish Market Report: Tuna-May 201, Thailand FAO Granger, C.W.J (1969) Investigating causal relations by econometric models and crossspectral methods Econometrica 37:424-438 Jaffry, S 2005 Asymmetric price transmission: A case study of French hake value chain Marine Resource Economics 19:511-523 Johansen, S (1988) Statistical analysis of cointegration vectors Journal of Economic Dynamics and Control 12:231-254 Johansen, S., and Juselius, K (1990) Maximum Likelihood Estimation and Inference on Cointegration with Applications to the Demand for Money Oxford Bulletin of Economics and Statistics 52:169-210 12 Laowapong, A 2010 Fisheries Value Chain: Kingdom of Thailand A report submitted to Food and Agriculture Organization for the project entitled A Value-chain Analysis of International Fish Trade and Food Security with an Impact Assessment of the Small-scale Sector Schwarz, G (1978) Estimating the Dimension of a Model Annals of Statistics 6:461464 Shinoj, P., B Ganesh Kumar, R Sathiadhasb, K.K Dattaa, M Menona, and S.K Singha 2008 Spatial price integration and price transmission among major fish markets in India Agricultural Economics research Review 21:327-335 von Cramon-Taubadel, S., and J.P Loy 1999 The identification of asymmetric price transmission processes with integrated time series Jahrbucher for Nationalokonomieund Statistik 218:85-106 13 Table-1: Summary of Augmented Dickey Fuller test of unit root for the two price variables at their level and first difference ADF test at level ADF test at first differences Variable name Constant Constant and trend Constant Constant and trend Skipjack Tuna -2.110 -2.029 -9.237** -9.245** Canned Tuna -1.401 -1.147 -7.361** -7.391** ** denotes rejection of null hypothesis of having unit root at the 1% significance level * denotes rejection of null hypothesis of having unit root at the 5% significance level 14 Table Results of the pairwise Granger causality test Null Hypothesis Canned tuna does not Granger cause skipjack tuna Skipjack tuna does not Granger cause canned tuna *denotes null hypothesis rejected at 5% F-Statistic 6.325** 1.880 Decision Reject H0 Accept H0 ** denotes null hypothesis rejected at 1% 15 Table Results of cointegration rank test between skipjack and canned tuna prices No of cointegration Trace Statistics Maximum Eigen Values p-value* equation 13.748 0.028 *MacKinnon-Haug-Michelis (1999) p-values Statistics p-value* 13.483 0.019 16 FIGURE CAPTIONS Figure Marketing channel of tuna in Thailand in 2009 Figure Prices of frozen skipjack tuna (c & f Thailand) and canned tuna (c & f USA) Figure Seasonal variation in prices of frozen skipjack tuna (c & f Thailand) and canned tuna (c & f USA) 17 Tuna from Thai vessels >100,000 tons 50% 50% Tuna from import >800,000 tons Factories Domestic 10% consumption 90% Export Figure 1: Market channel of tuna in Thailand in 2009 18 Figure Prices of frozen skipjack tuna (c & f Thailand) and canned tuna (c & f USA) (1 carton = 48*6 oz chunks v/f) Source (GLOBEFISH AN 11101, 11102) 19 Figure Seasonal variation in prices of frozen skipjack tuna (c & f Thailand) and canned tuna (c & f USA) 20