AN EMPIRICAL ANALYSIS OF LAW OF ONE PRICE IN BANGLADESH FISH MARKETS: A GENERALIZED ADDITIVE MODELING APPROACH  

The Law of One Price has a long history in economics and does provide  an important  basis for arbitrage behavior thought to characterize movements of commodity prices over time and across spatially-distinct markets. There is a vast literature that has evolved from linear regression models to classes of models that consider nonlinear price linkages investigating this phenomenon. Recently, the literature addressing the Law of One Price has concentrated on methodologies suitable for adapting structural change and mean-shifting behavior with unknown break points and variable speeds of adjustment among regimes that overcomes the problems of nonstandard test statistics and parameters. The regime switching behavior stems from unobservable transactions costs that may result in discrete trade/no trade regimes or smooth, continuous transitions between different states of the market. Among the models that integrated this mean-shifting behavior and the reasons behind these price movements are Goodwin, Holt, and Prestemon (2012), Enders &Holt, 2012; Holt & Terasvirta, 2012; Ng & Voselgang, 2002.  

The current article proposes a new class of semi-parametric model for spatial and regional price behavior that accommodates mean-shifting behavior in a vector autoregressive modeling framework to test the validity of Law of One Price for perishable fish products in Bangladesh.  These vector autoregressive models adopt the Generalized Additive Models (GAM) estimation procedures proposed by Hastie and Tibshirani (1986) and Linton (2000). In particular, the backfitting and integration algorithms developed for GAM model estimation to incorporate a non-parametric mean shift in the linkages describing individual pairs and larger groups of market prices are utilized. The empirical specification involves simple and vector error correction models that relate price differences to lagged values of prices and price differentials.

The current application is to six major regional fish markets (Khulna, Mymensingh, Rangpur, Faridpur, Barisal, and Jessore)  in Bangladesh for seven main freshwater aquaculture species ( Catla catla, Labeo rohita, Ctenopharyngodon idella, Anabas testudineus, Oreochromis niloticus, Pangasius hypophthalmus, and Hypophthalmichthys molitrix). We employ exhaustive data, collected by the WorldFish Center, consisting of weekly prices of fresh and chilled fishes at the wholesale and retail levels covering the period from October 2010 to June 2015.  

Additionally, impulse response functions are used to evaluate the dynamics of regional price adjustments to localized shocks in individual markets. Implications for regional price adjustments are also discussed in the paper.  Finally, suggestions for further extensions of the semi-parametric analysis of regime switching behavior are offered.