FISHERY BULLETIN: VOL. 87, NO. 2, 1989 



our analysis to the pot and longline, or fixed gear, 

 sector of the fleet, because these vessels produce 

 the higher quality sablefish desired by Japanese 

 consumers. 



We consider Tokyo the central wholesale market, 

 because it is playing an increasingly dominant role 

 in the handling of sablefish in Japan. Before the im- 

 position of harvest restrictions, when the Japanese 

 fishing fleet was responsible for providing most of 

 the domestically consumed sablefish, the centralized 

 market in Tokyo played a less important role in the 

 distribution of sablefish; it may have handled less 

 than 50% of that consumed. As Japan was forced 

 to rely increasingly on imports of sablefish, the clear- 

 inghouse in Tokyo has become more prominent by 

 providing a conduit for the imported product. 

 Throughout most of the 1980's, the Tokyo market 

 has generally handled 60-70% of the sablefish sold 

 in Japan. 



MARKET INTEGRATION AND 

 PRICE ANALYSIS 



This section defines market integration and pre- 

 sents the formal models used to empirically assess 

 the structure of any price integration which may 

 exist. 



Market Integration 



Markets integrated by prices are those markets 

 in which prices do not behave independently. In 

 market economies, market information to the differ- 

 ent participating economic agents is largely trans- 

 mitted by prices, so that an understanding of the 

 manner in which markets are integrated by prices 

 can contribute to the general process of policy for- 

 mulation. Markets may be integrated by prices to 

 different degrees and along some dimensions but not 

 others. Geographical links or interregional trade are 

 among the most important. Two major issues of 

 these spatial price linkages have been most frequent- 

 ly examined: whether or not markets are integrated 

 by prices, and if so, the extent and nature of this 

 integration. 



Regression analysis is the preferred method of 

 assessing market integration. This procedure im- 

 plicitly assumes that prices of commodities in spa- 

 tially dispersed markets can independently move in 

 a nonintegrated market system. Changes in weather 

 patterns or government policies affecting the quan- 

 tity and price of a commodity in a market should 

 not have an impact on prices in other markets when 

 the market system is not integrated by the price 



mechanism. Price movements across markets would 

 be fundamentally independent, and price move- 

 ments within markets would reflect local responses 

 to local conditions. 



Dynamic Spatial Price Differentials 



Ravallion (1986) recently proposed a dynamic 

 model of spatial price differentials from a central 

 market to local markets for a tradeable good. Ra- 

 vallion's model permits each local price series to 

 have its own dynamic structure, allows for any cor- 

 related local seasonality or other characteristics, and 

 provides for an interlinkage with other local mar- 

 kets. Moreover, the alternative hypotheses of inte- 

 gration of markets by price and market segmenta- 

 tion are encompassed within a more general model, 

 thereby allowing for nested statistical testing. Final- 

 ly, Ravallion' s dynamic model distinguishes between 

 the concepts of instantaneous market integration 

 and the less restrictive idea of integration as a long- 

 run target of the short-run dynamic adjustment 

 process. Thus, while short-run adjustment could be 

 statistically rejected by the data, so that trade does 

 not immediately adjust to spatial price differentials, 

 it is still possible to determine if there is any long- 

 run tendency toward market integration. 



Ravallion (1986) proposed the following 

 econometric model of a T-period series of prices for 

 A'' regions: 



+ c,X„ + ea, i = 2,...,iV, (1) 



where market 1 is the central market (here Tokyo), 

 X^ {i = 1,2 N) is a matrix of nonprice influ- 

 ences on local markets, the e,,'s are appropriate 

 error processes, J is the number of time periods to 

 be lagged, and the a's, b's, and c's are parameters 

 to be estimated. 



Several hypotheses about interregional trade and 

 market integration can be formulated as linear 

 parameter restrictions on Equation (1) and tested 

 by f'-tests (Ravallion 1986): 



Market Segmentation 



The null hypothesis of local market segmentation 

 states that changes in the central market prices will 

 have no effect, immediate or lagged, on prices in 

 the ith local market. Market i could be called seg- 



342 



