EXPANSION OF MERCOSUR S AGRICULTURAL EXPORTS TO THE EU: AN EMPIRICAL ASSESSMENT OF THE TRADE FLOWS. Jyrki Niemi 1. Ellen Huan-Niemi 1

Transcription

1 EXPANSION OF MERCOSUR S AGRICULTURAL EXPORTS TO THE EU: AN EMPIRICAL ASSESSMENT OF THE TRADE FLOWS Jyrki Niemi 1 Ellen Huan-Niemi 1 Oliver von Leebur 2 Petra Salamon 2 1 MTT Agrifoo Research Finlan, Luutnantintie 13, Helsinki, Finlan. yrki.niemi@mtt.fi an ellen.huan-niemi@mtt.fi 2 Institute of Market Analysis an Agricultural Trae Policy, Feeral Agricultural Research Centre (FAL), Bunesallee 50, D Braunschweig, Germany. petra.salamon@fal.e an oliver.leebur@fal.e Paper prepare for presentation at the 11th Congress of the EAAE (European Association of Agricultural Economists), The Future of Rural Europe in the Global Agri-Foo System, Copenhagen, Denmark: August 24-27, 2005 Copyright 2005 by Jyrki Niemi, Ellen Huan-Niemi, Petra Salamon, an Oliver von Leebur. All rights reserve. Reaers may make verbatim copies of this ocument for non-commercial purposes by any means, provie that this copyright notice appears on all such copies.

2 EXPANSION OF MERCOSUR S AGRICULTURAL EXPORTS TO THE EU: AN EMPIRICAL ASSESSMENT OF THE TRADE FLOWS Abstract. This paper provies new evience on income an price elasticities of eman for agricultural exports from Mercosur countries to the EU. Econometric moels are constructe for eight agricultural commoities beef, cocoa, coffee, orange uice, poultry, sugar, soya an wheat - exporte from Mercosur to the EU. A moelling approach base on the error correction mechanism is use in orer to emphasise the importance of the ynamics of trae functions. The results inicate that there is a relatively weak eman response to income an price changes in the EU. However, the results also suggest that relative-price variations affect significantly the eman for Mercosur commoity exports, implying that the exporter s market share is influence by price competitiveness. Key wors: agricultural trae, European Union, Mercosur, econometric moels, cointegration JEL classification: C22, Q17 1. Introuction Market access for agricultural goos is one of the main issues of the current EU-Mercosur Free Trae Agreement (FTA) negotiations, which began in April Both sies were cautiously anticipating a successful conclusion by the year 2004, but the agricultural knot remains a huge stumbling bloc on the roa to the final agreement. For the EU, agricultural an foo proucts are most vulnerable to competition from Mercosur countries, which are well enowe with natural resources, both lan an mineral. Agriculture is also one of the key sectors of Mercosur economies, in spite of the evient success of the manufacturing sector uring the last ecaes. The European Union (EU), which represents one of the worl s largest markets for raw materials an agricultural proucts, with imports of nearly 67 billion in 2003, is a particularly attractive an very sought-after market for Mercosur exporters. The EU accounts for 35% of all Mercosur agricultural exports to the worl. During the perio between 1990 to 2003, Mercosur agricultural exports to the EU rose from 7 billion to 11 billion, showing an average annual growth rate of 3.5%. Over the years, Mercosur countries have also manage to increase their market share in the EU quite substantially. On the other han, EU exports of agricultural proucts to Mercosur are not significant. Thus, the agricultural trae balance has clearly tilte in favour of Mercosur, with a trae surplus of 10 billion in Despite the success in penetrating to the EU market, the Mercosur countries has been concerne with the agricultural protection policy of the EU. The maor irritant in EU-Mercosur agricultural trae relations has been the variable levies/tariffs an other iscriminatory measures against Mercosur proucts, such as sugar an beef. Therefore, Mercosur countries have taken a special interest in encouraging the EU to liberalise its trae in agriculture. Mercosur countries hope that with trae liberalisation, the member countries will be able to improve their market access for agricultural proucts in the EU. The future WTO negotiations uner the Doha Development Agena will provie an important base for extening the process of trae liberalisation. This stuy attempts to increase our knowlege of the behavioural relationships unerlying agricultural trae flows between Mercosur an the EU. More specifically, the obective is to provie new estimates of short- an long-run elasticities of import an export eman for commoities exporte from Mercosur countries to the EU. The paper consists of five sections an it is organise as follows. Section two lays out the general theoretical an methoological framework employe by the stuy for moelling the ynamic relationships of commoity trae. Section three explains how the theoretical structures are implemente in ynamic econometric moels in practice. Section four presents the empirical results of the estimate moels constructe for eight agricultural commoities beef, cocoa, coffee, orange uice, poultry, sugar, soya, an wheat - exporte from Mercosur to the EU. In section five, we present a summary of our main conclusíons. Page 2 of 10

3 2. Theoretical an methoological framework Imperfect competition arising from prouct ifferentiation unerlies the moelling framework of this stuy. The estimation of import eman systems is erive from Armington s (1969) moel, where it is assume that the same goos of ifferent origins are imperfect substitutes within an importing country s commoity market. Furthermore, in orer to reuce to number of parameters to be estimate, the moel assumes a constant elasticity of substitution (CES) for each prouct pair. Following the moel, the importing ecision is split into two stages. The solution to the utility maximisation problem for the first level of ecision yiels the overall eman scheules for commoity imports M of importer, given a commoity import price P an a level of constant ollar income Y, an is expresse as M p m P = k 1 Y (1) D where k 1 is a constant with expecte sign k 1 > 0; D is the eflator; an m p is the price elasticity of import eman for goos M. The income elasticity is equal to unity, a hypothesis that will later be teste. Once the level of expenitures Y for the importe commoity M has been etermine, the solution to the utility maximisation problem of how much of the commoity to purchase from alternative suppliers - let us say an exporter of interest i an its competitors m, which refer each of the n-1 other foreign supplying countries, to market whose corresponing export prices are P i an P m - may be expresse as X i p x Pi = k 2 M (2) P where X i is the quantity of the goos exporte from country i to country, k2 is a constant; P i is the price of the goos importe from country i to country ; P is the average price of the goos importe to country ; an x p is the relative-price elasticity of export eman. The empirical analysis of the stuy is base on econometric moels which capture the ynamics unerlying trae an price formation in commoity markets, an it is conucte by means of recently evelope econometric concepts. Among these, the so-calle general to specific approach avocate by Henry (1986) is applie in the context of ata series whose (non-)stationary properties are investigate. Furthermore, the notion of cointegration (Engle an Granger, 1987) of a set of variables is analyse. The approach follows closely the moelling strategy evelope in a series of papers by Davison et al. (1978), Henry (1986), Lor (1991), Urbain (1992), Baneree et al. (1998) an Niemi (2003). Given that economic time series often exhibit non-stationary stochastic processes, the econometric specification is conucte in a framework that allows for non-stationary but potentially cointegrate variables. The approach aopte is to convert the ynamic moel into error correction formulation, an it is shown that this formulation contains information on both the short-run an longrun properties of the moel, with isequilibrium as a process of austment to the long-run moel. Equations specifie in this manner allow the relevant economic theory to enter the formulation of long-run equilibrium in levels while the short-run ynamics of the equation are etermine by growth rates. Since the valiity of the error correction specification requires the existence of a long-run relationship or cointegration between the variables concerne, the econometric analysis begins with the tests for the existence of a cointegrating vector. The first step in the analysis of cointegration is to etermine the time series properties (i.e., the orer of integration) of each variable, whether they have a unit root or not. Tests for unit roots are performe using the augmente Dickey-Fuller univariate Page 3 of 10

4 tests. Having establishe the orer of integration of each variable, tests for cointegration are unertaken an the nature of any cointegrating vectors explore. A formal test of cointegration is carrie out following the resiual-base approach propose by Engle-Granger (1987) as well as the sequential testing proceure put forwar by Perron (1988). 3. Econometric analysis of Mercosur s agricultural exports to the EU 3.1. Data The econometric analysis of the stuy is conucte with a sample of annual ata which covers Mercosur s maor commoity exports to the EU from 1976 to The analysis uses 4- an 5-igit prouct-level ata base on the Stanar International Trae Classification (SITC). For the purpose of this stuy, the prouct heaings are efine as follows: beef (SITC 0111), cocoa (SITC 072), coffee (SITC 0711), orange uice (SITC 05851), poultry (SITC 0114, 01181), sugar (SITC 0611, 0612), an soya (SITC 08131, 2222) an wheat (SITC 0411, 0412). Volume an value ata on trae flows over the perio are obtaine from UN Commoity Trae Statistics Database (UN Comtrae). Volume ata is compile in kilograms, an value ata in US ollars. The transaction value is the value at which goos were sol by the exporter, an inclues the cost of transportation an insurance, an freight to the frontier of the importing country (c.i.f. valuation). The unit prices of EU imports ( P EU ), an unit prices of exports by an iniviual Mercosur country ( P i ), are erive by iviing value by volume. These iniviual prouct unit values are subect to an error in measurement. If import eclarations are inaccurate, customs ata may be incorrect. Scobie an Johnson (1975) have shown that, if the observe value an the volume ata contain errors of measurement for actual transactions, the estimate elasticity of substitution will be biase towars zero. Moreover, unit values suffer from the traitional f.o.b./c.i.f. valuation problems. Elasticity estimates are base on c.i.f. prices, which, because they inclue changes in trae resulting from transportation an istribution costs or from tariffs, o take into account all price ifferences between suppliers to the ultimate consumer (Lor, 1991). Therefore, in this stanar formulation the observe real prices in exporting country i assume fixe transfer costs. The gross omestic prouct (GDP) inex an the consumer price inex (CPI) are use as a measure of economic activity (Y EU ) an price eflator ( D EU ) of the EU, respectively. The source of the ata is the International Financial statistics atabase of the International Monetary Fun (IMF) Import eman specification Moels such as equation (1) are usually specifie in log-linear form by assuming that stanar trae theory relates exports an imports to explanatory variables through a multiplicative form which can be erive within a cost minimisation framework (Urbain, 1992). The first-orer stochastic ifference equation as a logarithmic function of the theoretical relationship in (1) is, therefore, expresse as P P ln α + (3) M t = 0 + α1 lny t + α 2 lny, t 1 + α 3 ln + α 4 ln + α 5 ln M, t 1 v 1t D D t t 1 where the expecte signs are α 1, α 2 > 0; α 3, α 4 < 0; an 0< α 5 < 1. The lags in the moel are specifie as the maximum to be expecte in the light of the nature of import eman an the evience of previous econometric stuies. The maximum lag length for annual time-series ata is usually equal to one on the hypothetical basis that economic agents are characterise by one-year planning horizons. The results of the cointegrating regressions in Appenix A show that eman for commoity imports in the EU market ( ln M EU ) has a steay-state response to the omestic economic activity ( lny EU ), an a transient response to the constant ollar price of imports (P/D). Page 4 of 10

5 Transformation of equation (3) to incorporate an ECM riven by ln Y EU an (ln P /D ) to the equation (3), with an aitional lagge variable of ln Y EU, results in the following import eman specification: P P M ln M t = + lnyt + lny, t + ln ln D + α D + 0 α1 δ2 1 α3 δ4 δ5 Y t t 1 + v1 t (4) where δ 2 = (α 1 + α 2 + α 5-1), δ 4 = ( α 3 + α 4 ), an δ 5 = (α 5-1). The expecte signs of the coefficients are α 1 > 0, δ 2 > δ 5, -1 < δ 5 < 0, an α 3, δ 4 < 0. The fifth term of the equation, ln (M /Y ) t-1 is calle the error correction term, while δ 5 is the feeback coefficient. The error correction term captures the austment towar the long-run equilibrium. If δ 5 is statistically significant, it states what proportion of the isequilibrium in ln M t in one perio is correcte in the next perio. The long-run ynamic solution of a single-equation system generates a steay-state response in which growth occurs at a constant rate, say g, an all transient responses have isappeare (Currie, 1981, Lor, 1991). With growth rates of omestic economic activity an import eman, ln Y t = g 1 an ln M t = g 2, respectively, the long-run ynamic equilibrium solution of equation (3), in terms of the original (anti-logarithmic) values of the variable, is ( δ / δ ) ( ) t M = k Y P / D (5) 1 where k 1 = exp {[-α 0 + (1-α 1 )g 1 ]/δ 5 }. Equation (5) encompasses the static equilibrium solution when g 1 = 0. The income elasticity of import eman is expresse as m y = 1 - (δ 2 /δ 5 ). The price elasticity of import eman is m p = -δ 4 /δ Export eman specification δ / δ In terms of the general stochastic ifference specification, the export eman relationship in (2) is expresse as P P ln β + (6) i i X it = 0 + β1 ln M t + β 2 ln M, t 1 + β 3 ln + β 4 ln + β 5 ln X i, t 1 v 2t P P t t 1 where the expecte signs of the coefficients are β 1, β 2 > 0; β 3, β 4 < 0; an 0< β 5 < 1. The ynamics for the export eman relationship is assume to be of relatively small orer, an can therefore be restricte to cases where the lagge values of the variables are of one year. The Lagrange multiplier (LM) tests are again performe for omitte higher lagge variables. The results of the cointegrating regressions suggest that in most cases the eman for exports of a Mercosur country i ( ln X i ) has a steay-state response to the import eman of the EU ( ln M EU ), an a transient response to the relative price of the EU market (lnpi/lnp EU ). In the secon case, eman for exports from country i to country (X i ) has a steay-state response both to the import eman (M ), an to the relative price (P i /P ) of that market. In other wors, X i, M, P i /P are cointegrate. An ECM is obtaine from the cointegration regression of X i on M an P i /P. The following transformation of (6) incorporates an ECM riven by import eman M : P P ln β + (7) ( X i M ) v t i i X it = 0 + β1 ln M t + γ 2 ln + γ 3 ln + γ 4 ln / t 1 2 P P t t 1 where γ 2 = b 3, γ 3 = (β 3 + β 4 ), an γ 4 = (β 5-1). The expecte signs of the coefficients are β 1, γ 2 > 0, γ 3 < 0, an -1 < γ 4 < 0. The relative price term in the foregoing specification have been so transforme Page 5 of 10

6 as to nest the ifferences formulations of the variable in the levels form of the equation. The error correction term, γ 4 ln (X i /M ) t-1, measures ivergences from the long-run equilibrium an corrects for previous non-proportional responses in the long-run ynamic growth of export eman. Since in ynamic equilibrium ln M t = g 2, ln X = g 3 an ln (P i /P ) t = 0, it follows that the solution of (7), in terms of the original values of the variable, is ( i / ) X = k M P P i 2 γ 3/ γ4 where k 2 = exp {[-β 0 + (1-β 1 )g 2 ]/γ 4 }. Therefore, export eman is assume to have a unitary elasticity with respect to the level of import eman in the geographic market. The price elasticity of export eman is expresse as x p = -γ 3 /γ 4. Where the long-run response between the export eman of a country i an imports of its traing partner is not necessarily proportional, an aitional term (explanatory variable for imports of a country lagge by one perio) is introuce into the equation (6). The founations of the error correction moel (ECM) specification, use in equations (4) an (8), rest on the seminal work of Sargan (1964). The ECM specification can be erive as a simple reparameterisation of a general autoregressive istribute lag (ADL) moel. The iea of incorporating the ynamic austment to steay-state targets in the form of error-correction terms, suggeste by Sargan an evelope by Henry an Anerson (1977) an Davison et al. (1978), among others, offers the possibility of revealing information about both short-run an long-run relationships. 4. Estimation results 4.1. The import eman functions The short- an long-run responsiveness of agricultural commoity imports to changes in incomes an own-prices in the EU are summarise in Table 1. Coefficient signs an magnitues are acceptable in terms of a priori expectations. The moels also track the sizes an the irections of changes in the volume of EU agricultural imports fairly well. Consiering that the equation explains the rate of changes in the import volumes, the R 2 values ranging from 0.42 to 0.67 can be consiere quite satisfactory. The results for own-prices elasticities inicate that they are statistically ifferent from zero in six out of the eight commoities, an, of these, one is significant at the 1% level, two at the 5% level, an three at the 10% level. The estimate income elasticities have expecte positive signs an are significantly ifferent from zero at the 10% level in the equation for all commoities, excluing sugar. All the coefficients of the lagge error-correction terms appear highly significant at 1-percent level. Therefore, the eviation from the equilibrium level of the import eman ue to ranom shocks represents a significant eterminant of its short-run ynamic behaviour. The coefficient estimates on the own-price terms confirm the expectation that eman for commoity imports in the EU is relatively inelastic with respect to price. The price elasticities range from to in the short-run, an from to in the long run. The policy implication of this fact is that exchange rate policies an commercial policy intervention measures in the form of tariff an non-tariff barriers to trae woul not be very effective in changing the quantity of imports emane. Soya oil has the largest long-run price elasticity ( -0.62) an the remaining five have elasticities less than 0.5 in absolute terms. Coffee has the lowest long-run price elasticity ( -0.09). Beef an sugar i not show the expecte sign of price elasticity, though neither was statistically significant. Beef an especially sugar are highly protecte agricultural proucts in the EU internal markets. The stanar tariffs an aitional import uties for sugar are so prohibitive that almost all sugar imports are from eveloping countries that receive preferential treatment from the EU. Thus, price an income elasticity estimation for sugar may be an impossible task. (8) Page 6 of 10

7 Table 1. Short-run an long-run elasticities of import eman in the EU for selecte commoities. Commoity Price elasticity Income elasticity Short-run Long-run Short-run Long-run Beef Cocoa Coffee Orange uice Poultry Soya Sugar Wheat - -0, Note: - Not significant at the 10% level The long-run income elasticities are in a range between 0.12 for coffee an 1.17 for soya. Orange uice is foun to have a unitary elasticity with respect to income. Poultry has income elasticity close to unity, an beef, cocoa an wheat have elasticities significantly less than unity. The large ifferences in the income elasticity have implications for sales by exporters. Soya imports have been sensitive to income changes. Thus, soya exports have a consierably stronger growth potential in the EU than other commoities, because of a strong response of buyers in the EU to improvements in their real income. The high income elasticity of soya is supporte by a strong eman for its use as an ingreient in compoun animal fee. The level of eman for compoun fees epens on the livestock inustry, an the livestock inustry epens in turn on the level of eman for meat an other livestock proucts. Overall, the results suggest that agricultural commoity imports of the EU are not very sensitive to income changes an are consiere necessary goos in the sense that eman increases slower than economic activity goes up. This means a relatively weak growth potential for the selecte commoities in the EU market. At the same token, imports of these commoities are not susceptible to larger swings of eman uring business cycles, either. The trae-weighte average long-run income elasticity of import eman across commoities is relatively low ( 0.53). The finings are consistent with the earlier stuies. The low income an price elasticity of eman for primary commoities have been recore in many stuies covering a wie range of commoities. The coefficients on the error correction terms in the import eman relationships measure austment towars the long-run relationship between import volumes, economic activity, an prices. In the case of cocoa an coffee, the coefficients of the error correction terms are close to unity in absolute terms. This fact reflects the relatively quick response of EU importers to changes in income an prices, i.e. it oes not take a great eal of time for import eman to resume its long-term equilibrium growth path when a short-run isequilibrium arises between import eman an income. Nevertheless, importers of beef an poultry aust to income an price changes relatively slowly The export eman functions The elasticity estimates of export eman equations for the maor commoity exports of Mercosur to the EU are reporte in Table 2. The signs an magnitues of the estimate coefficients are broaly in line with theoretical expectations. Relative prices an error correction terms are strongly significant with an austment coefficients ranging from 0.16 to Furthermore, the moels explain the changes in the volume of Mercosur agricultural exports to the EU rather accurately. Gooness of fits is acceptable with an R 2 in a range between 0.37 an The moels also pick up quite well the turning points an rapi rises in export eman. As expecte, relative price movements affect significantly the trae flows of all commoities, implying that exporter s market share has been influence by price competitiveness. Relative prices are statistically ifferent from zero in 8 out of the 11 trae flows, an, of these, two are significant at Page 7 of 10

8 the 1% level, two at the 5% level, an four at the 10% level. The three exceptions are the export eman for poultry an sugar from Brazil, an wheat from Argentina where the relative price coefficient i not result in statistically significant estimate. Table 2. Dynamic equilibrium solutions of export eman functions for selecte commoities from Mercosur into the EU. Commoity Exporter Beef Argentina Brazil Relative price elasticity of export eman Response to changes in the level of EU imports Short-run Long-run Short-run Long-run Uruguay Cocoa Brazil Coffee Brazil Orange uice Brazil Poultry Brazil Soya Brazil Paraguay Sugar Brazil Wheat Argentina Note: - Not significant at the 10% level The sizes of relative price coefficients, of course, iffer by commoity as well as by source of supply in each commoity. The short-run relative price elasticity of export eman range from 0.37 to -1.47, an the long-run elasticity from to In other wors, there is a great eal of variation in the export performance between ifferent commoities an among iniviual Mercosur countries. Therefore, care shoul be exercise in generalisations about the price elasticities of eman for the region s commoity exports. The observe ifferences in relative-price coefficients by trae flow reflect the ynamic aspect of the EU agricultural trae, in which particular trae flows rise an fall with price competition. Among the trae flows uner examination, the export eman for Paraguayn soya is the least sensitive to relative price changes, followe by Brazilian coffee exports. In contrast, the relative-price coefficients of Brazilian beef an orange uice exports are exceptionally large, an -3.45, respectively. These finings, combine with the result of import price elasticity in Table 2, inicate that, although agricultural imports are relatively insensitive to price changes on a commoity basis, once the total amount to be spent for imports of a commoity is etermine then the EU importers seek cheaper proucts, so that price competition among suppliers is inevitable. On the other han, the sharp contrast of relative price coefficients in the same commoity ustifies the assumption that importers istinguish agricultural proucts by place of prouction, even though the proucts are calle by a common commoity name. The austment of export eman from one level of foreign import eman to another is etermine by the error correction term. The error correction terms for all the trae flow equations are strongly significant with an austment coefficient showing wie variations from 0.16 to Among the small feeback coefficients that of cocoa exports from Brazil eserve attention. The Page 8 of 10

9 estimate coefficients, -0.16, imply a very slow austment towars the estimate equilibrium state. It takes over 10 years for this trae flow to aust to 90% of its new steay-state solutions. In contrast, exports of Brazilian beef, coffee an soya plus Argentinean beef, have the coefficients of the error correction terms above 0.5 in absolute terms. This fact reflects a relatively quick response of exports to changes in the level of EU imports an relative price, i.e. it oes not take a great eal of time for export eman to resume its long-term equilibrium growth path when a shortrun isequilibrium arises between export eman an import eman. For example, it takes only three perios for Brazilian coffee exports to the EU to aust to 90% of its new steay-state solutions. The estimation results also confirm the assumption that export eman for commoities from Mercosur has, in general, more or less proportional response to changes in the level of EU import. Therefore, at given relative-price levels, any increase or ecrease in commoity imports by the EU woul be reflecte in an almost equivalent percentage change in its eman for exports from Mercosur countries. In other wors, the market share of the country oes not change unless relative prices change in homothetic eman. However, if the estimate coefficient of the import response variable is significantly greater than unity, it is a goo inication for an exporting country that its exports can expan more than others an its share increase as EU market grows. Among the selecte commoity trae flows, orange uice an beef from Brazil have clearly more than proportional response to changes in the level of EU imports. 5. Concluing remarks This paper has attempte to apply a reasonably flexible ata etermine, ynamic moel to estimate the short-run an long-run effects of changes in income an prices on agricultural commoity trae between Mercosur an the EU. Therefore, a moelling approach base on the error correction mechanism (ECM) was use in orer to emphasise the importance of ynamics of trae functions. Application of a series of iagnostic tests supporte the use of this approach. Econometric moels were constructe for eight agricultural commoities beef, cocoa, coffee, orange uice, poultry, sugar, soya an wheat - exporte from Mercosur to the EU. The results for the estimate import eman functions suggest that there is a relatively weak eman response to income changes in the EU. The results also emonstrate the inelastic nature of price responses in the EU eman for the importe commoities. The policy implication of this fact is that trae policy measures in the form of tariff an non-tariff barriers are not very significant in changing the quantity of imports emane. In the case of beef an sugar, imports woul likely increase far more than those of any other commoity if import protection is abolishe. However, ue to the high level of protection, which istorts import eman responses, we were not able to get statistically significant price elasticity estimates for beef an sugar. The coefficient estimates of the export eman functions inicate that relative-price variations affect significantly the eman for Mercosur commoity exports by the EU, implying that exporter s market share is influence by price competitiveness. Furthermore, the sharp contrast of relative price coefficients in the same commoity across countries ustifies the assumption that importers istinguish agricultural proucts by place of prouction, even though the proucts are calle by a common commoity name. The estimate moels of this stuy can be use to assess the results of trae policies on commoity trae between Mercosur an the EU. Hence, a fruitful avenue for future research woul be an analysis of how ifferent governments regulations affect the volume of trae an what kin of welfare effects such regulations might have. References Armington, P. (1969). A Theory of Deman for Proucts Distinguishe by Place of Prouction. International Monetary Fun Staff Papers 16: Baneree, A., Dolao, J. Galbraith, J. & Henry, D. (1993). Co-integration, Error Correction an the Economic Analysis of Nonstationary Data. New York: Oxfor University Press. 320 p. Page 9 of 10

10 Baneree, A., Dolao, J. & Mestre, R. (1998). Error-correction Mechanism Tests for Cointegration in a Single-Equation Framework. Journal of Time Series Analysis 19: Currie, D. (1981). Some Long-Run Features of Dynamic Time Series Moels. Economic Journal 91: Engle, R. an Granger, C. (1987). Co-integration an Error Correction: Representation, Estimation an Testing. Econometrica 55: Davison, J.E.H., Henry, D.F., Srba, F., an Yeo, S. (1978). Econometric Moelling of the Aggregate Time-Series Relationship Between Consumers Expeniture an Income in the Unite Kingom. Economic Journal 88: Dickey, D. an Fuller, W. (1981). Likelihoo Ratio Tests for Autoregressive Time Series with a Unit Root. Econometrica 49: Henry, D.F. (1986). Econometric Moelling with Cointegrate Variables: An Overview. Oxfor Bulletin of Economics an Statistics 48: Honma, M. (1991). Growth in Japan s Horticultural Trae with Developing Countries: An Economic Analysis of the Market. International Foo Policy Research Institute (IFPRI), Research Reports 89. Washington DC: IFPRI. 87 p. Lor, M. (1991). Imperfect Competition an International Commoity Trae: Theory, Dynamics an Policy Moelling. Oxfor University Press, New York. 403 p. Niemi, J. (2003). Co-Integration an Error Correction Moelling of Agricultural Commoity Trae - The Case of ASEAN Agricultural Exports to the EU. Agricultural an Foo Science in Finlan 12: Supplement No. 1. Perron, P Trens an Ranom Walks in Macroeconomic Time Series. Journal of Economic Dynamics an Control 12: Phillips, P. an Ouliaris, S. (1990). Asymptotic Properties of Resiual Base Tests for Cointegration. Econometrica 58: Scobie, G. & Johnson, P Estimation of the Elasticity of Substitution in the Presence of Errors of Measurement. Journal of Econometrics 3: Urbain, J-P. (1992). Error Correction Moels for Aggregate Imports: The case of two small an open economies. In: Dagenais, M. & Muet, P-A. (es.). International Trae Moelling. Chapman & Hall Lt., Lonon. pp Page 10 of 10