Trade Flows Impact of European Union Expansion

International Business: Research Teaching and Practice 2007 1(1) Trade Flows Impact of European Union Expansion Augustine M. Nwabuzor School of Business and Industry, Florida Agricultural and Mechanical University One SBI Plaza, Tallahassee, Florida 32307-5200 Hudson C. Nwakanma School of Business and Industry, Florida Agricultural and Mechanical University One SBI Plaza, Tallahassee, Florida 32307-5200 With the addition of new members to the EU from 2004 through 2007, expectations were high for trade expansion between new and old members. This study estimates the extent to which such an increased trade potential exists. Using Gravity Model estimates, the study compared actual to predicted export/import values for 2003 for ten new nations that were previously member states of the Soviet economic bloc or the Council for Mutual Economic Assistance. For some new members, notably, the Czech Republic, Slovenia and Romania, a robust gap between actual and potential exports were identified. There was also pronounced over-trading in nations like Estonia and Lithuania. Overall, a deepening in the process of trade liberalization within both new and old members was clearly noticeable. The results indicate that international trade of new EU members adjusts to normal gravitational forces of income and population. Additionally, the study predicts increased trade intensity between new and old members and the likelihood of success for trade expansion policies within most new members. INTRODUCTION Between 2004 and 2007, the EU increased its membership from 15 to 27 members. Apart from Cyprus and Malta, the new members were drawn from former member states of the Soviet economic bloc or the Council for Mutual Economic Assistance (Comecon). The ten new EU members that were former members of Comecon are Bulgaria, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovakia, Slovenia, and Romania. Together, these nations form the central east European trading Telephone: (850) 412-7735 Email: nwabuzor@embarqmail.com 16

International Business: Research, Teaching and Practice 2007 1(1) bloc and for them the end of Comecon meant the opening of their economies to Western markets. Trade patterns in these nations had started to change sharply by 1995 with their export-import structure showing a marked tilt towards the EU (Pravorne, Skrorohoda, Strods and Tkachevs, 2003). The new members of the EU had succeeded in re-orienting their economies in the immediate years before admission as full members. The re-orientation included concluding trade agreements with other non-eu nations to promote trade with them and to steer away from trade relations with their former socialist allies. A major concern of previous studies has been determining whether the trade potential between the EU and these new members has been exhausted. Though results have been mixed, the most recent studies (Hamilton and Winter, 2002) have tended to conclude that trade potential between the EU and these new members is either at or above potential, while other authors, (Jakab, 2001; Laaser and Schrader, 2002), have shifted their attention to the determinants of trade relations and the emergence of specialization patterns. Our objective in this paper is to examine the trade patterns of the EU and its ten new ex-comecon members and to re-explore the issue of future trade potential. This will be done using Gravity Model estimates for these nations. Hypothetical coefficients of the export and imports Gravity equations will be applied to a nation s trade data for 2003 to predict its trade flows for the year. By comparing actual to predicted values for 2003, it will be possible to draw conclusions whether the nations are over- or undertraded. The argument will be that evidence of under-trading with other EU members implies the existence of increased trade potential for the nations involved. The existence of such a potential will have obvious policy implications for each nation involved, such as export drives directed at other members of the union The technique of using coefficients of the Gravity Model to predict trade flows has been successfully used in several previous studies. Byers et al., (2000) used them in modeling trade flows for Scandinavian nations; so did Pravorne et al., (2003) in the study of trade flows for the Baltic states and Caetano and Galego (2005) used them to examine EU trade flows. Beginning with a historical perspective on the trade patterns of these nations, this study will explore the theoretical conception of the Gravity Model. A short literature review will set the stage for discussing the use of the Gravity approach to modeling trade flows. We will then estimate the export and import Gravity equations for the data on the new and old EU nations in the study. Finally, the main results of the prediction of trade flows for these countries will be presented as the empirical results of the exercise are summarized. Trade Patterns for European Nations: Before and After the Collapse of Comecon A common feature of the economies of all the European nations featured in this study is the dramatic increases in trade recorded from 1993 to 2003. Exports for the older 15 member nations practically doubled from US$ 666 billion to US$ 1,370.2 billion (See Table 1). France and Germany nearly tripled their exports while Austria, 17

Nwabuzor &Nwakanma EU Expansion and Trade Flows Finland, the Netherlands, Portugal, Spain and Sweden more than doubled theirs. Imports similarly grew significantly from US$ 624.6 billion in 1993 to US$ 1,325.9 billion in 2003. Finland, Greece, Ireland, Italy, the Netherlands, Spain and Sweden all more than doubled their imports. The ten central east European nations also experienced phenomenal increases in their volume of international trade during the same period. Their exports nearly quadrupled from almost US$ 60 billion in 1993 to more than US$ 228.5 billion in 2003. Every member of that group either doubled or tripled their exports during the period. Hungary s and Poland s exports grew from US$ 8.2 billion and US$ 13.6 billion in 1993 to US$ 43.5 and US$ 61.0 billion respectively in 2003. Imports for the group rose from US$ 71.4 billion to US$ 253.6 billion. The most dramatic increases occurred in Poland and Slovakia whose imports in 2003 stood at US$ 66.7 and US$ 22.6 billion up from US$ 13.6 and US$ 5.4 billion respectively in 1993 (see Table 1). A distinct characteristic of the economies of these nations in the early 1990s was that trade flows were intra-soviet oriented. For instance, until 1993 the Baltic States: Latvia, Estonia and Lithuania, sent more than 50 percent of their exports to Russia and other member nations of the Commonwealth of Independent States. They drew more than 60 percent of their imports from the same region (see Table 2). However this was quickly followed by a process of economic liberalization that translated to important economic transformations and changes in the external trade of these nations. EU markets soon became very important for the exports of these new members, whose products faced much stiffer competition in the wider global market. Indeed, by 1999 well over 50 percent of the exports of the Baltic nations were sent to the EU while almost 60 percent of their imports were drawn from that region (see Table 3). The pattern is the same for the other new members covered in this study. It is possible to identify some main characteristics in the trade patterns between the old and new EU members. First is the very high degree of rapid and generalized openness that the economies of the new members have experienced. In global terms and defined as the weight of the external trade to GDP, that degree of openness has been estimated at around 56 percent in 1993 and up to 80 percent by 2003. External trade has been very important to these new EU members with the increase in trade openness in Hungary, Slovakia, the Czech Republic and Estonia, exceeding 100 percent by 2004. It is also important to note that these nations suffered deteriorating structural trade deficits during the period. The deficits were considerable, accounting for up to 7.5 percent of GDP on average and up to 10 percent in the case of Bulgaria (Caetano and Galego, 2005). Second is the progressive shift of the economies of these new members towards the economies of EU members, discussed previously. By 2001, the EU was already accounting for up to 66 percent of the exports of the new members (Jakab, 2001). Germany, for example, consistently accounted for over 42 percent of the exports to and 45 percent of the imports from the new members between 1993 and 2003 (see Table 2). 18

International Business: Research, Teaching and Practice 2007 1(1) 19

Nwabuzor &Nwakanma EU Expansion and Trade Flows Comparative figures for Italy were 15 percent and 12 percent respectively. The intensification of trade between old and new members continued, although there were accompanying commercial imbalances. Trade relationships with the old EU members were responsible for the deficits of the new member nations in 2001. Such deficits reached more than 54 percent in 2003. Figures available (Pravorne et al., 2003) suggest that while Italy, France and Finland have been responsible for about 74 percent of EU trade surplus with these nations, Germany, Austria, Denmark, Greece and Portugal have maintained deficits in their trade with them. Trade intensity in three dimensional terms takes into account: 1. The evolution registered in the exports of a given country; 2. Imports in the country of destination; and 3. Weighted flows of world trade during the period under review. Based on the above definition, one can easily see that trade intensity across the countries covered in this study varies quite a bit. The Czech Republic, Hungary, and Poland are among the new members that are involved in reciprocal trade while Germany, Austria, and Finland are among the old. The intensity of bilateral trade is more pronounced in some cases such as Austria and Germany with Hungary, the Czech Republic, Slovenia and Slovakia; Greece with Bulgaria and Romania; and Finland and Sweden with the Baltic States. In contrast, the level of trade is low between the new members and Ireland, Spain and Portugal. The indication is that the intensity of trade is higher for neighboring nations that tend to have closer cultural, economic and historical ties (see Table 3). The evidence suggests that proximity of the states to one another has a significant influence on the patterns of trade. Those new members that share a common border with the old members account for a full 82 percent of the trade of the new members with the EU. In contrast, the share of the Baltic and Balkan states is 7.5 percent and 10 percent respectively. Portugal, Ireland and Greece accounted for only 2.5 percent of the trade with the new members (Pravorne et al., 2003). The weight of intra-area trade among the new members suffered a decline between 1991 and 2003. The magnitude of that decline was small (from 15.6 percent to 13.5 percent) but significant. There was an even sharper decline in the relative intensity of trade during the period, over 35 percent between 1991 and 2003. This decline in the intensity of trade was clearly noticeable in former Czechoslovakia, Estonia and Hungary after 1997. However, trade was particularly intense in the Baltic and Balkan sub-regions (see Tables 4 and 5) thus reinforcing the proposition that proximity is a decisive factor of trade intensity, a situation likely to persist in future years. 20

International Business: Research, Teaching and Practice 2007 1(1) 21

Nwabuzor &Nwakanma EU Expansion and Trade Flows Table 3. European Union 1993-2003: Intensity of Exports Index in Bilateral Trade EU 15 New Members 1993 2001 2003 1993 2001 2003 Austria 5.12 4.22 4.52 Bulgaria 0.86 1.39 1.28 Belgium 0.60 0.84 0.82 Czech. Republic 1.33 1.81 1.78 Denmark 1.14 1.20 1.25 Estonia 1.33 1.77 1.81 Finland 2.23 2.20 2.24 Hungary 1.54 1.84 1.98 France 0.73 1.14 1.12 Latvia 1.48 1.66 1.72 Germany 2.40 2.83 2.84 Lithuania 0.92 1.44 1.48 Greece 3.24 3.99 4.01 Poland 1.77 1.82 1.84 Ireland 0.21 0.34 0.33 Romania 0.84 1.56 1.58 Italy 1.63 2.19 1.86 Slovakia 1.64 1.67 1.65 Netherlands 0.77 1.02 1.01 Slovenia 1.07 1.80 1.96 Portugal 0.12 0.32 0.28 Spain 0.51 0.96 0.98 Sweden 1.10 1.57 1.61 United Kingdom 0.65 0.71 0.75 EU 1.41 1.68 1.68 New Member 1.39 1.76 1.78 Source: Constructed from IMF International Statistical Yearbook and CHELEM Database-CEP11 Table 4. Description of the Variables used in the Gravity Models Variable GDP i POP i GDP j POP j DIST ij DUM i DUM(O) Du(N) Description GDP per capita of each EU country included Population of the country GDP per capita of trading partner Population of trading partner Distance in kilometers between the capitals of the EU nations and its trading partners = 1 if the trading partner has a common border with the given EU nation = 1 if the trading partner is among the 15 old EU members = 1 if the trading partner is a new EU member 22

International Business: Research, Teaching and Practice 2007 1(1) Table 5: Exports Gravity Equation Regression Results Variable B T-value GDPi 1.991 27087 *** POPi 0.851 30658 *** GDPj 0.901 34318 *** POPj 0.717 37480 *** DISTij -0.877-15018 *** DUMi CB 0.889 5080 ** DUM2 (NEW EU) 0.283 2056 ** DUM3 (OLD EU) 0.105 780 Constraint -9.815-15535 *** Adjusted R Square =0. 845 F = 501330 Signif. F= 0.000 Number of Observations = 941 Notes: *** shows statistically significant coefficients at 1 percent level of significance ** denotes statistically significant coefficient at 5 percent level of significance # Dependent variable: log of exports from l to j Trade Index: Intensity of Europe Index (IEI) IEI = Xij k * (Xw) 2 = Xi*Mj Xij k Xi*Nj*Xw k Xw k Xw*Xw THEORETICAL FOUNDATION The analysis of the impact of trade on the economies of nations has often centered on the market model. The model suggests that the best results are obtained by nations that adopt a policy of free flow of exports and imports without stifling controls. The rapid economic development of the Asian Tigers : South Korea, Taiwan, Hong Kong and Singapore is often cited as a case in point. However, as Krueger (1990b) reminded, these Asian nations originally had restrictive trade policies, in the form of quantitative controls on imports. Trade and the Economies of Nations Linking trade to economic development, Grossman and Helpman (1990) established that nations practicing a free trade policy grew faster than those that did not. They concluded that developing nations stood to gain more from unrestricted international trade since they do not have the means (human and physical capital) to conduct the research needed for the development of new products. The Grossman- 23

Nwabuzor &Nwakanma EU Expansion and Trade Flows Helpman studies affirmed the earlier findings of Heller (1977) that the rapid growth of exports accelerates the economic growth of a country since exports are indeed part of the national output. Similarly, Balassa (1978) established that export oriented policies lead to more efficient resource allocation providing an incentive for the development of both domestic and foreign markets. Immediately following these findings, efforts were made to quantify the place of exports in economic development. Citing the dramatic successes achieved by nations pursuing export oriented policies and the equally dramatic failures of nations that practiced more restrictive trade policies, Tyler (1981) established that a 17.5 percent increase in exports tended to produce a 1 percent increase in GDP. He concluded that nations that do not develop the export sector by insisting on restrictive economic policies will settle for lower rates of economic growth. In a further attempt at quantification, Feder (1982) separated the export and nonexport components of output. He used the simple definitional equation Y=N+X; which stated that GDP (Y), was equal to N+X where N = Non-export sector and X = Export Sector. Feder found evidence that marginal factor productivity in the export sector was higher than in the non-export factor. By applying his model to the findings of earlier studies (Balassa, 1978; Michaely, 1977; Heller and Porter, 1978; Tyler, 1981), Feder established that a 10 percent increase in exports induces a 1.3 percent growth in the non-export sector. All these studies helped establish that in free trade, export promoting policies are useful in advancing economic development. The studies do not, however, address the more general question of trade flows among nations. The Gravity Model The conceptual foundation for the analysis of trade flows is the Gravity Model. It states that bilateral trade between any two countries or regions is directly related to their incomes and inversely related to the distance between them. One can, therefore state the gravity equation as follows: M ij = α k Y i β Y j γ N j,and N j ζ d ij ε U ij where M ij is the trade flow between country i and j, α k is a constant, β, γ, ζ, ε, μ are coefficients, weighted geometric averages, Y i and Y j are incomes per capita in countries i and j, respectively. N i and N j are populations in i and j, and d ij is the distance between countries i and j. The U ij is a lognormal distributed error term with E (ln U ij ) = 0. Frequently, dummy variables are also included in the model to take into account preferential trade factors between i and j. Early initial doubts about the theoretical soundness and consistency of the Gravity Model (Tinbergen, 1962; Poyhonen, 1963) have quieted following the successful application of the model by Linneman (1966) and Leamer and Stern (1970). These contributions were followed by efforts to derive gravity equations from models that 24

International Business: Research, Teaching and Practice 2007 1(1) assumed product differentiation. Anderson (1979) showed that the gravity equation can be derived from properties of expenditure systems. This was followed by Grossman (1998) who submitted that a sufficient condition for obtaining a gravity equation is the perfect product specialization assumption. The empirical success of the equation lies in the proposition that as an exporting country increases the supply of its product, the importing country will increase its consumption proportionately, thus raising the volume of trade between them. Early applications of the Gravity Model attempted to relate bilateral trade to such variables as the distance between the countries and the size of the gross domestic product. In one such application, Deme (1995) established that the rate of growth of imports between member nations of the Economic Community of West African States (ECOWAS) was larger than the rate of growth of imports from industrialized nations. As used by Deme and others, a basic tenet of the Gravity Model is to link trade flows with economic size (approximated by output). It also addressed factors that inhibit trade flows (such as tariffs) as well as factors that increase transaction costs (such as transportation bottlenecks). In that scenario, exports represent production and the ratio of an exporter s domestic market production to its foreign market production. The implication is that an increase in national production will mean an increase in export supply while an increase in the domestic market to foreign market production ratio will lead to lower export supply. The Deme (1995) logic worked well in analyzing trade flows in West Africa where markets are characterized by very high tariffs and frequent importation bans in a bid to conserve meager annual foreign exchange earnings. Severe transportation bottlenecks in the region and other trade resistance issues, such as the difference between official exchange rates and parallel market rates, also make the focus on transaction costs appropriate. European economies, in both the old and new EU members, are open and transportation and other trade impediments, while present, are certainly not severe. Therefore, the model used for studying trade flows in West Africa may not be entirely appropriate in analyzing European markets. A different, and perhaps more relevant, approach was proposed by Evenett and Keller (1998) who summarized three types of trade models that differ in the way product specialization is obtained in equilibrium: 1. Technological differences between countries in the Ricardian model of trade; 2. Factor proportions outside the vector space-of-diversification in the Hecsher-Ohlin model; and 3. Increasing returns of the firm in the Increasing Returns to Scale model. By demonstrating that each of these models is a limiting case of a model with imperfect specialization, Evenett and Keller (1998) showed that perfect product specialization is not the key assumption of the success of the Gravity Model. The studies further established that factor proportion differences are important 25

Nwabuzor &Nwakanma EU Expansion and Trade Flows determinants of trade flows within the context of imperfect specialization. Additionally, increasing returns is a cause of product specialization along the lines of trade models with both perfect and imperfect specialization of production. There have been many other uses of the Gravity Model to estimate trade flows among European nations, particularly between the EU and other parts of Europe. In many such studies, trade between the EU and the East European nations, as well as trade with the Baltic States has been modeled. In one of the latest of these studies, Angier, Gaziorrick and Laitong (2004) provided the first serious attempt at an empirical examination of the possible impact of rules of origin on patterns of trade among European nations. The methodology employed was an augmented Gravity Model where the focus was the impact within the Pan-European system of accumulation. The results suggest that rules of origin do restrict trade and that the accumulation of such rules could increase trade in the order of 50 percent with the impact greater on intermediate than on manufacturing trade. Other more important studies include Wang and Winters (1994). This study established that the potential trade between the European Free Trade Association and the Central European nations was larger than the actual trade. The study estimated trade flows from 1984 to 1986 between the regions. Baldwin (1994), van Beers and Biesen (1996), Iversen (1998), Cornet and Iversen (1998), Pravorne et al. (2003) and Caetano and Galego (2005) all probed various aspects of European trade flows. Baldwin (1993, 1994) updated the Wang-Winter model by positing that bilateral trade flows were a function of: GDP level of the countries involved; The population of the nations; and The distance between the exporting and importing nation. Baldwin used this model to estimate trade potential for Albania, Croatia, the Czech Republic, Bulgaria, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Slovakia, Russia, Ukraine, and Belarus. There were two scenarios to the Baldwin analysis: in one, the medium term scenario, it was assumed that the Central and Eastern European nations will indeed integrate fully into the world economy. In the other, the longer-term scenario, it was assumed that the per capita income in these Central and Eastern European exports will increase in the direction of the European Free Trade Association. The finding was that the Central European markets could account for up to 17 percent of the European Free Trade Association s exports to all of Europe. In a study that covered 49 nations, including six Central European countries, Van Bergeijk (1990) provides the results of the estimation of the Gravity Model based on population data and income levels for East-West and East-East trade. The estimates showed that all bilateral trade flows in 1985 between East and West were about 12.5 percent, as large as the Gravity Model prediction, based on income, population and 26

International Business: Research, Teaching and Practice 2007 1(1) distance. The model analysis further established that the actual turnover of East-East trade was 50 percent lower than its potential. Cornet and Iversen (1998) attempted an estimate of trade flows in the Baltic region taking into account trade relations between the EU and Central European nations. Two years later, Byers (2000) used hypothetical coefficients from recent trade data for the Scandinavian countries to predict trade flows in the Baltic countries of Latvia, Lithuania and Estonia. In both studies, the researchers attempted to control for different phases of integration so that they could differentiate between various forms of trade barriers associated with different forms of bilateral trade links. Both studies successfully explained the present trade relations in the Baltic region on the basis of data for income, population and proximity. METHODOLOGY In studying the bilateral trade between old EU members and the newer ones (essentially between EU and the nations of Central Eastern Europe), we estimated a Gravity Model for the period 1993 and 2003. This also enabled us to predict the trade adjustments associated with the removal of trade barriers. The analysis was limited to total exports and followed the practice of previous studies that modeled trade flows as a function of the sum of the GDP of both trading countries (GDT), the degree of similarity between them (SIM) and the economic distance between the countries (ED). This methodology is the same as was employed by Egger (2000) and Fontagne (1999). Also included are the geographic distances between the nations, the existence of a common border, and two additional dummies: EU OLD (indicating whether the country belongs to the 15 older members of the European Union) and EU NEW (which equals one if both trading partners belong to the group of new members of the EU). Several specifications may be used to estimate a Gravity Model. We began by estimating hypothetical coefficients of exports and imports gravity equations using 2003 data for the EU new members in order to predict trade flows for these nations in 2004. Such an analysis will be based on the Gravity Model with the following equations: X ij = b 0 +b 1 GDP +b 2 POP +b 3 GDP +b 4 POP +b 5 DIST ij +DUM k +u i i j j where all the variables (excluding dummies) are expressed in natural logarithms and dependent variable X ij represent the exports from the country of origin (i) to the trade partner (j) in exports equation or imports from country j to i in imports equation. Independent variables include Gross Domestic Product per capita and population in countries i and j respectively and distance between the capitals of the two countries. Dummies are used to control for different kinds of virtual distance and proximities. A full description of the variables used is given in Table 4. 27

Nwabuzor &Nwakanma EU Expansion and Trade Flows Dummies have been chosen to take into account the impact of different preferential trade agreements between some of the countries as these substantially reduce costs connected with distance. It is noted that historical ties are often an explanatory variable for current trade flows (Laaser and Schrader, 2002). Hence, a dummy for common borders (contiguity) is included since trade between neighbors is usually less impeded by transaction costs. Following the example of Caetano and Galego (2005), we also used panel data to take account for unobserved country heterogeneity. Similarly, using the proposals of Fontagne (1999), Egger and Pfaffermayr (2003), and Cheng and Wall (2005), we used a general specification based on trading pair-specific or bilateral common effects. The argument is that this type of general model tends to give better predictions because it considers common bilateral effects. There is a further assumption that there are systematic differences across pairs of countries that are captured by country-pair constants. These effects control for all time invariant factors that are specific to each of the trading pairs: Exportsijt=α 0 +δ ij +γ t +β 1 GDT ijt +β 2 SIM ijt +β 3 ED ijt +β 5 dist ij +β 6 Frontier ij +β 7 EU ij +β 8 CEFTA ij +ε ijt where δ ij represents the unobservable country pair effect, γ t the observable time effect and ε ijt is the remainder stochastic disturbance term This time dummies were included to capture the effects of any variables affecting bilateral exports that vary over time and that are constant across country pairs. It should also be noted that in these specifications we have allowed the country-pair effects (δ ij ) to differ according to the direction of the trade, which means that δ ij is not equal to δ ji. Both the dependent and explanatory variables included in the model are again logarithms with the exception of dummies. The Hausman test was used to compare the Within estimator from the fixed effects model and the random effects of GLS estimator, testing the null hypothesis of no correlation between the individual and time effects and the regressors. The test rejected the existence of no correlation. Thus, in order to obtain consistent and nonbiased estimators, a fixed effects model was estimated using the Within estimator. Besides, heteroscedastic consistent standard errors were calculated for all regressions in order to correct for heteroscedasticity problems. The relationship between trade flows and the explanatory variables was estimated by the Weighted Least Squares to overcome the problem of heteroscedasticity usually presented in Gravity Models estimated by OLS. We only considered those trading partners whose exports and imports values exceeded US$ 1 million and for which all other required data was available. Estimation of Gravity Model for New EU Members International Trade The regression results for exports show that the coefficients have the expected signs and almost all of them are highly significant (see Table 5). These coefficients are significant even at a 1 percent level. Export flows from these nations are positively 28

International Business: Research, Teaching and Practice 2007 1(1) correlated with GDP per capita and with population in both countries of origin and country of destination. Distance between the nations is a reducing factor. These findings are in keeping with the theoretical prediction that bilateral trade flows between these nations and their trading partners comply with gravitational force. As to the regression for imports, all signs are in the expected direction and all coefficients are statistically significant at a 1 percent level (see Table 6). Both GDP per capita and population show positive coefficients. However, gravitational forces of income and population are less important for imports than for exports. In both the imports and the exports models, the results are statistically important at very high significance levels. Adjusted R squares (0.850 for exports equation; 0.796 for imports equation) are comparable to other gravity regressions. They show that the independent variables explain better than 84 percent of the variation in the dependent variables in both equations. Analysis of Potential Trade Flows The models presented form the basis for predicting exports to, and imports from, other countries for new EU members. From these, ratios between actual and predicted trade flows can be calculated for 2003. If the ratio between actual and predicted values for exports (or imports) exceeds 1, then the country involved is said to be overtraded in exports (or imports) with the corresponding nation or region. If that value is less than 1 then it is under-traded, meaning that there are possibilities for further trade expansion. These ratio comparisons are presented in Tables 7, 8a and 8b. The estimates from the Gravity Model already presented were used to analyze whether the potential trade between the older 15 nations of the EU and the newer members are above or below the actual level. Predictions of trade between the old and the new EU members in 1997 and 2003 were computed and compared with the actual values. The same basic procedure has been used to analyze trade relations among the new EU members for the period 1997 2003. We then analyzed the ratio between predicted and actual values. The results of the potential versus actual exports for each of the 15 older EU members compared to the total of the new members is shown in Table 7. The table also shows the results of potential versus actual exports from each of the new members to the older 15 members. It clearly shows the deepening of the process of trade liberalization within the new and the older EU members. Differences, however, exist between the exports of the older 15 members and the exports of the new members. The results indicate that there might be a gap between actual and potential exports for some new members such as the Czech Republic, Estonia and Romania. There are other instances where, in the case of exports, and considering trade with all other countries, actual trade is less than predicted. Latvia, in the Baltic region is a case in point. The indication is that Latvia has a very robust trade potential with other non-baltic Sea nations. Under-trading is also observable in the case of exports from Slovenia. For both countries, the extent of under-trading is as much as 18 percent. 29

Nwabuzor &Nwakanma EU Expansion and Trade Flows Table 6. Gravity Equation Regression Result Variable B t-value GDPi 0.657 15469 *** POPi 0.870 26215 *** GDPj 0.787 25986 *** POPj 0.813 29583 *** DISTij -0.458 6720 *** DUMi CB 1360 6779 *** DUM2 (NEW EU) 1415 7829 *** DUM3 (OLD EU) 908 5183 *** Constraint -10105 13246 *** Adjusted R Square = 0.791 F = 320280 Signif. F = 0.000 Number of Observations = 941 Note: *** statistically significant coefficients at 1 percent level of significance A far more prevalent observation is the general over-trading that most of the new EU members exhibit in this study. Estonia, for instance, is almost 1.5 times overtraded, a reflection of the evident openness of that nation s economy. The predicted value of Estonia s exports is lower than the actual in practically all nations. Elsewhere in the Baltic, Lithuania s actual exports are more than 20 percent higher than the predicted value. Lithuania is however, under-traded with the other new EU members thus creating a potential for exports. Lithuania s over-trading with the old European members is a result of that nation s very intense trade with the United Kingdom (UK). Exports to the UK are over 3.2 times the predicted value. We can draw the same inference of considerable over-trading for other new EU members such as Poland and the Czech Republic (see Tables 8a and 8b) indicating little room for trade expansion between these nations and the older 15 member nations of the EU. The continuing permanent transformation of the economic structure of the new EU members should be noted. This transformation makes a long term prediction of their future trade potential tenuous at best. Yet, future trade expansion is certain and this will impact real incomes and market reforms in these nations as others have already noted (Fontagne, 1999; Auxilioux and Pajot, 2001). There are interesting observations to be made from the evolution of the predicted trade flows from each new EU member to the total new member s trade compared to actual values. There is clear evidence that not all these new members have the same evolution during the same time period. For some counties, the gap between potential and actual is, particularly after 1997, narrowing with time. For others the gap is getting bigger, as in the case of Slovakia, Slovenia and Estonia. As has been noted, there is a reduction in trade intensity between Estonia and Slovakia and the rest of the new invisible texxxxt 30

International Business: Research, Teaching and Practice 2007 1(1) Table 7. Potential versus Current Exports between Old and New European Union Members Origin Country (Potential/current) Origin Country (Potential/current) 1993 2001 2003 1993 2001 2003 Austria 0.7946 1.2372 1.2472 Bulgaria 1.1006 1.0051 1.1216 Benelux 1.0478 0.9633 1.0645 Czech Republic 0.9854 0.3785 0.7689 Denmark 0.8808 1.2738 1.2816 Estonia 2.7494 1.4008 1.2167 Finland 1.5153 1.9269 1.8645 Hungary 0.9817 1.0061 0.9966 France 1.0800 0.8253 1.0267 Latvia 1.0381 1.1515 1.1268 Germany 0.9214 1.0297 1.0363 Lithuania 0.9788 0.7578 0.7149 Greece 0.6769 0.8815 0.9277 Poland 0.5680 0.4784 0.4146 Ireland 1.7688 1.0481 1.1728 Romania 1.0670 1.2158 1.1617 Italy 1.0887 1.2056 1.3678 Slovakia 1.2877 0.7259 0.7647 Netherlands 0.8206 1.0783 1.0966 Slovenia 0.8222 1.0201 1.1767 Portugal 1.5321 0.7476 1.2688 Spain 1.4672 0.9393 0.9782 Sweden 1.3649 1.6173 1.7846 United Kingdom 0.9405 1.1350 1.2647 Destination Country Destination Country 1993 2001 2003 1993 2001 2003 Bulgaria 1.0885 0.9851 0.8679 Austria 0.7508 0.7593 0.9169 Czech Republic 0.7814 1.0467 1.1264 Benelux 1.1162 0.4942 0.6788 Estonia 0.9868 1.2684 1.2869 Denmark 0.7608 0.6718 0.6456 Hungary 1.8355 1.2684 1.2686 Finland 0.9161 0.8577 0.8765 Latvia 0.9673 1.0308 1.6477 France 0.8958 0.5033 0.8676 Lithuania 1.5754 0.9695 0.9874 Germany 0.7980 0.6095 0.7125 Poland 1.5205 0.8691 0.9333 Greece 0.8793 0.8688 0.8764 Romania 0.9096 1.261 1.1567 Ireland 0.8050 0.3771 0.3869 Slovakia 1.0465 0.0876 0.9223 Italy 0.9338 0.6957 0.6958 Slovenia 1.2559 0.9316 0.9567 Netherlands 0.7139 0.7712 0.788 Portugal 1.0416 0.6648 0.9430 Spain 1.2272 0.4829 0.4819 Sweden 1.1446 0.8757 0.9648 United Kingdom 0.8767 0.6659 0.7168 31

Nwabuzor &Nwakanma EU Expansion and Trade Flows Table 8a. Export Flows of 10 New Members to 15 Older European Union Members (percent) 1993 1995 1999 2001 2003 Bulgaria 4.9 10.6 27.7 39.2 46.4 Czech R 18.4 26.9 46.9 58.6 64.8 Estonia 3.7 54.0 62.7 72.5 75.1 Hungary 16.8 24.8 39.5 48.8 56.2 Latvia 39.9 39.2 48.3 62.0 61.8 Lithuania 3.0 26.4 50.1 48.1 49.2 Poland 15.6 29.8 44.6 59.7 69.2 Slovakia 11.6 19.8 24.7-38.6 Slovenia 12.6 18.9 25.3 29.4 36.5 Romania 11.7 19.6 22.7 29.8 33.9 Table 8b. Import Flows of 10 New Members to 15 Older European Union Members (percent) 1993 1995 1999 2001 2003 Bulgaria 10.3 11.6 21.3 27.4 33.5 Czech R 16.9 25.8 44.9 51.6 63.3 Estonia 6.1 66.0 57.7 56.4 58.2 Hungary 5.7 16.3 22.8 37.6 44.9 Latvia 29.5 44.6 43.9 53.2 54 Lithuania 2.9 37.2 46.5 40.6 45.6 Poland 16.4 30.6 45.8 53.7 20.6 Slovakia 13.8 19.4 26.7-36.5 Slovenia 10.5 18.3 24.6 31.6 33.9 Romania 13.4 17.3 19.4 23.8 29.6 Source: Constructed from IMF International Statistical Yearbook and Chelem Database members after 1997 (see Table 3). A widening of the gap between potential and actual trade in these nations is, therefore, not surprising. Nevertheless, the fact that actual trade in most countries seems to be approaching the potential trade is in keeping with the developments in trade intensity between these nations. CONCLUSION The central issue of this paper has been to determine which new members of the EU are over- or under-traded relative to the predicted trade flows based on Gravity Models. The focus has been to determine where unfilled trade potentials exist as a basis 32

International Business: Research, Teaching and Practice 2007 1(1) for formulating trade policies by some of the new members. Accordingly, export and import gravity equations were run using the nations trade data for 1993-2003. We were able to demonstrate that the international trade of these new members adjusts to the normal gravitational forces of income and population (larger markets). It was also established that distance had a repelling effect on transactions between trading partners. In general, the newer members have buoyant trade relations with the richer members of the old EU 15, but large distances between the newer members and their trading partners tended to reduce the trade flows between them. The liberalization of trade between the older members of the EU 15 and the newer members has helped to promote an intensification of trade among all the countries in the two groups. The major gains in trade appear to have occurred among the countries of Central Europe. A primary conclusion of this study is that, although potential trade appears exhausted for some countries in the short run, there remain strong possibilities for trade expansion. The analysis also suggests opportunities for further improvements resulting from economic development in the new member nations. The implication is that the accession of the new members to the EU will positively affect the intensity of trade. Trade impacts of enlargement are also affected by geographical and economic factors. The enlargement can be expected to trigger trade intensity that should revive old economic partnerships among neighboring nations. Depending on their technological knowledge and factor endowments, this should affect the levels of welfare in the countries involved. In the short and medium term, some countries will experience negative effects, but the longer term impacts will be positive because of the favorable environment that results from economic and monetary stability in the new members. This will in turn generate significant opportunities for trade expansion in the EU as a whole. The predictions point to increased trade intensity between the new and old EU members. However, it is fair to assume that will not happen to the same degree and at the same time to all new members. Previous experience has shown that trade flows tend to increase as income levels converge, as demand structures become more similar and as international networks expand. Thus, one can expect that, depending on the competitiveness performance of each country, changes are bound to occur in the comparative advantage patterns. What the study has clearly established is that trade expansion policies by any of the new members directed at the rest of the EU are likely to be successful. It is tempting to speculate on the impact of this trade liberalization and expansion on the economies of the old and new EU member states. Critics have often argued that continued liberalization and expansion in trade will produce devastating social and economic conditions in the new members. It is claimed that the gap in wages, which is currently five to eight times lower in the new members, will widen. Average per capita GDP, currently 24,250 euros for old EU members will continue to lag in the new 33

Nwabuzor &Nwakanma EU Expansion and Trade Flows members (where Hungary and Latvia, for instance, have GDP per capita values of 7,080 euros and 3,740 euros respectively [Salzmann, 2004]). Similarly, unemployment, said to be the main cause of poverty, is twice as high in the new members as the older ones. Unemployment levels in the Czech Republic and Slovakia were as high as 11 percent and 16.6 percent respectively by the middle of 2004. In short, the argument is that: trade expansion will increase the gulf between the rich and the poor countries of Europe under conditions where, in contrast to earlier rounds of the European expansion, no significant measures exist to compensate for such tendencies. Extremely low wages in east European countries will be employed as a lever to undermine wages and living standards in the wealthier countries in the west (Rippert, 2004). The above argument cannot be sustained. Trade expansion, particularly export expansion, can only mean the production of more goods and services (for export) in the exporting countries. Our study suggests that there is scope for such expansion in the new EU members. Such expansion will lead to increased employment, not less, and to higher wages. Besides, one needs to take into account the increased flow of foreign direct investment and employment this expansion will generate. Total German trade with the new members accounts for 40 percent of the nation s total. It is, therefore, not surprising that German companies employed 350,000 workers in Poland, the Czech Republic and Hungary by the middle of 2004. The benefits of EU membership are real to the new members and, in spite of the difficulties in agreeing on a budget in 2005 and suggested cuts in proposed aid to new members, they, the new members, continue to look for growth and prosperity from the continuing trade expansion. REFERENCES Anderson, J. (1979) A theoretical foundation for the gravity equation, American Economic Review, 69 (1): 106 116. Angier, P., Gaziorrick, M. & Laitong, C. (2004) The impact of rules of origin on trade flows, Economics Working Papers Archive: http://ideas.respec.org/p/wpa/wuwpit/0404001.html (accessed January 13, 2006). Auxiliou, V. & Pajot, M. (2001) Trade issues in European Union enlargement, Paper presented at the 75 th International Conference on Policy Modeling for European and Global Issues, Brussels, Belgium July 2001. Balassa, B. (1978) Exports and economic growth: Further evidence, Journal of Developmental Economics, 5: 181 189. Balassa, B. (1985) Exports, public choices and economic growth in developing countries after the 1973 oil shock, Journal of Developmental Economics, 18: 23-35. Baldwin R. (1994) Towards an integrated Europe, Centre for Economic Policy Research, London. Bali Indonesia Travel Portal www.indo.com/distance/index 34

International Business: Research, Teaching and Practice 2007 1(1) Byers, J et al. (2000) Capital and trade flows in Europe and the impact of enlargement, Kiel Working Paper # 1001, 2000. Caetano, J. & Galego, A. (2005) Trade flows among CEEC and EU countries: What future perspectives? University of Evora, Evora Portugal. Centre D etudes Prospectives Et D informations Internationales, CHELEM, database www.cepii.fr/anglaisgraph/bdd/chelem.htm Cheng, I-Hui & Wall, H. (2005) Controlling for heterogeneity in gravity models of trade, The Federal Reserve Bank of St. Louis, 87(1): 49-63. Cornet, S. & Iverson, L. (1998) Proper economic specification for the gravity equation, World Economy, 21(3):397-401. Deme, M. (1995) The impact of ECOWAS on intro-regional trade flows: An empirical investigation, The Review of Black Political Economy, Winter: 113-129. Egger, P. (2002) A note on the proper econometric specification of the gravity model, Economic Letters, 66(1): 25-36. Egger, P. & Pfaffermayr, M (2003) The proper econometric specification of the gravity equation: A three-way model with bilateral interaction effects, Empirical Economic, 28(3): 571-580. Evenett, S. J. & Keller, W. (1998) On the theories explaining the success of the gravity model, NBER Working Papers #6529, Cambridge, Massachusetts. Feder, G. (1982) On exports and economic growth, Journal of Development Economics 12: 59-73. Fontagne, L. (1999) Le Potetiel d Echanges entre l Union Europeenne et les PECO, Revue Economique, 50: 1139-1168. Grossman, G. (1998) The new growth theory: Trade, innovation and growth, American Economic Association Papers and Proceedings, 1998, 86-91. Grossman, G. M. & Helpman, E. (1990) The new growth theory: Trade, innovation and growth, American Economic Association Papers and Proceedings, pp. 86-91. Hamilton, C. & Winter, C. (2002) Opening up international trade with Eastern Europe, Economic Policy, 14: 77-116. Heller P. (1997) Exports and growth: An empirical reinvestigation, Journal of Development Economics 14: 191-193. Heller, P.S. & Porter, R.C. (1978) Exports and growth: An empirical reinvestigation, Journal of Developmental Economics, pp. 191-193. International Monetary Fund, (2005) International Statistical Yearbook, 2004, Washington DC. Iverson, L. (1998) On the theories explaining the success of the gravity model, NBER Working Papers #6529, Cambridge Mass., 1998. Jakab, Z. (2001) How far has trade integration advanced?: An analysis of the actual and potential trade of the three Central and Eastern European countries, Journal of Comparative Economics, 29: 276-292. Krueger, A.O. (1990a) Asia trade and growth lessons, American Economic Association Papers and Proceedings, pp. 108-112. Krueger, A. O. (1990b) Government failures in development, Journal of Economic Perspectives, 4(3):9-23. Laaser, C. F. & Schrader, K. (2002) European integration and changing trade patterns: The case of the Baltic States, Kiel Working Papers #1088, Kiel. 35

Nwabuzor &Nwakanma EU Expansion and Trade Flows Leamer E.E. & Stern, R.M. (1970) Quantitative International Economics, Allyn and Bacon Publishing: Boston, Massachusetts. Linneman, H. (1966) An Econometric Study of International Trade Flows, North Holland Publishing: Amsterdam, The Netherlands. Michaely, M. (1977) Exports and growth: An empirical investigation, Journal of Developmental Economics, 4: 49-53. Nwabuzor, A.M. & Anyamele, O.D. (2001) Improving trade among African Nations: Lessons from the West African Sub-region, Conference Proceedings, Academy of International Business SE, November 2001. Poyhonen, P. (1963) A tentative model for the volume of trade between countries, Weltwirtschaftliches Archives, 90: 93-99. Pravorne, I., Skrorohoda, N., Strods, A. & Tkachevs, O. (2003) The gravity model and prediction of trade flows between Baltic States and their trade partners, Euro-Faculty Working Papers in Economics, #17, University of Latvia. Rippert, U. (2004) The consequences of eastward expansion of the European Union, http://www.wsws.org/articles/2004/may2004/expa-m01.shtml (accessed January 13, 2006). Salzmann, M. (2004) The social consequences of European Union expansion, http://www.wsws.org/articles/2004/apr.2004/euro-a20.shtml (accessed January 13, 2006). Tinbergen, J. (1962) Shaping the World Economy: Suggestions for an International Economic Policy, The Twentieth Century Fund, New York, New York. Tyler, W. (1981) Growth and export expansion in developing countries, Journal of Development Economics, 9: 121-130 Van Beers, J. & Biesen, C. (1996) Opening up international trade with Eastern Europe, Economic Policy, 14: 77-116. Van Bergeijk, P. (1990) Diplomatic barriers to trade, de Economist, Springer, The Netherlands March 1990; 140(3): 77-96. Wang, Z. & Winters, R. (1994) Developing nations and multilateral trade negotiations, World Bank Policy Research Working Paper # 1846, Washington DC. World Bank Governance Indicators database, www.worldbank.org 36