Turkish Trade Flow and the EU : A study of a potential membership

J

I

B

S

JÖNKÖPI NG UNIVER SITY

Turkish Trade Flow and the EU

A s t u d y o f a p o t e n t i a l m e m b e r s h i p

Bachelor Thesis in Economics Author: Adela Rivero

Robert Said

Bachelor Thesis in Economics

Title: Turkish Trade Flow and the EU. A study of a potential membership. Authors: Adela Rivero and Robert Said

Tutor: Associate Professor Johan Klaesson, PhD Candidate Johanna Palm-berg, and PhD Candidate Hyunjoo Kim

Date: February 2008

Subject: Turkey and the European Union, gravity model, trade diversion, trade creation

Abstract

This bachelor thesis examines how the trade flow of Turkey would change if the coun-try becomes a member in the EU. This is done by analysing the Turkish trade flow with the EU and ROW over the period 1995 – 2005. The main question is if trade creation and trade diversion occurs?

The model used in this paper is a gravity model influenced by Soloaga and Winters (2000) gravity model. To be able to answer the purpose of this paper we used a cross-sectional regression and base our analysis on our results.

The conclusion of this paper is that Turkey’s trade flow with the EU will increase if full-membership is accomplished. This leads to trade diversion and trade creation to-wards the other members-states within EU. We see Turkey as the key to the door for Middle Eastern countries and the EU; this implies that if Turkey becomes a member -state in EU, the trade could increase between these two continents.

Kandidatuppsats inom Nationalekonomi

Titel: Turkiets handelsflöde och EU. En studie om ett mjöligt med-lemsskap.

Författare: Adela Rivero och Robert Said

Handledare: Assisterande Professor Johan Klaesson, Doktorand Johanna Palmberg och Doktorand Hyunjoo Kim

Datum: Februari 2008

Ämnesord: Turkiet och den Europeisk Unionen, gravitations modell, handelsomfördelning, handelsökning

Sammanfattning

Denna kandidatuppsats undersöker huruvida Turkiets handelsflöde kan förändras av ett eventuellt Europeiskt medlemskap. Det är gjort genom att analysera Turkiets handels-trend under perioden 1995 – 2005. Huvudfrågan är huruvida handelsfördelning och handelsökning sker?

Modellen som används i uppsatsen är en gravitations modell influerad av Soloaga och Winters (2000) gravitations modell. För att kunna besvara syftet i uppsatsen har vi an-vänt oss av en tvärsnittsregression och baserat analysen på resultaten.

Sammanfattningsvis ser man att Turkiets handelsflöde med EU ökar av ett eventuellt medlemskap i EU. Detta leder till handelsökning samt handelsfördelning gentemot nu-varande medlemmar. Vi anser även att Turkiet är nyckeln Mellanöstern samt EU, vilket innebär att fullt medlemskap ökar handeln mellan dessa två kontinenter.

Table of Contents

1

Introduction... 1

1.1 Problem... 1

1.2 Method... 1

1.3 Thesis-Outline ... 1

1.4 History of the European Union... 2

1.5 Background of the Turkish economy ... 2

2

Theoretical framework... 4

2.1 Comparative Advantage ... 4

2.2 Relative Prices ... 6

2.3 Production Possibilities ... 7

2.4 Trade Creation and Trade diversion ... 8

3

Institutional and Empirical framework ... 9

3.1 The accession of the Eastern countries... 9

3.2 Descriptive Data ... 9

4

Regression Model... 12

4.1 Gravity model ... 12

4.2 Variable formulation... 12

4.3 Model formulation ... 14

4.4 Regression Analysis and Results... 15

4.5 Results ... 16

4.6 Statistical errors ... 17

4.7 Joint Hypothesis testing... 17

5

Analysis ... 19

6

Conclusion and suggestions to further research... 22

References ... 23

Figures ...

2.2 Figure Relative prices ... 6

2.3 Figure Production possibilities ... 7

3.1 Figure Trade with EU member states ... 10

3.2 Figure Trade with non-EU members ... 11

Tables...

4.4 Table Expected Signs of the Variables in the Gravity Model ... 16

4.7 Table Joint Hypothesis Test (1995-2005) ... 18

Appendices ...

Appendix 2: Individual Variable Testing (1995-2005) ... 27

Appendix 3: White's Heteroskedasticity test ... 28

Appendix 4: Descriptive statistics ... 29

Appendix 5: Year Dummies and Turkey's GDP (1996-2005)... 30

1 Introduction

Over the last decades the international trade has increased tremendously and many cou-tries find it beneficial to sign trade agreements and join customs unions. A free trade area is a selected group of countries that have agreed to remove all barriers on trade be-tween themselves but they are free to implement different commercial policies towards third countries. On the other hand a customs union is a free trade area with a common external tariff and a common external trade policy, where the member countries can use different import quotas. The European countries have for many decades seeked for economic integration in the world and this has led to a strong international union, the European Union. The EU has become one of the major global actors in terms of trade and foreign direct investment (FDI). The benefits of a membership in this union are greate and many non-member countries are waiting to affiliate to be apart of those benefits (Nello, 2005). Tukey as one of the applicants for a full membership, draw our attention in particular.

Turkey have applied for a full membership in the EU for over four decades ago which however resulted in a customs union.. The article “Trade Policy Review” states that Turkey’s query for associated membership in the EU would later become a foundation for an Association Agreement in 1963, (the “Anakara Agreement”). The Ankara agree-ment is a three-step process between the ECC and Turkey to build up a customs union. This indicated that the EU and Turkey would build up a customs union with a deadline in 1995 (Sübidey, 2005). In 1987, twenty-four years after the “Ankara agreement”, Tur-key applied for full membership in the EU. However, due to complications in TurTur-key’s economical and political environment the process had been going so slow that it was almost not being operated (Redmond, 2007).

In this paper, we will analyze however trade creation and trade diversion occurs todwars the European Union.

1.1 Problem

The purpose of this paper is to analyze how Turkey’s trade-flow pattern will be affected by a full membership in the European Union. The main question raised is; will trade creation and trade diversion occur towards the European Union?

1.2 Method

The time period for the analysis is 1995-2005. A modified version of the Soloaga and Winters (2000) gravity model will be used. The analysis is limited to the EU 25 thereby the latest members will not be part of the analysis. The selection criteria we have is that countries must have a GDP/CAP greater than Turkey. We have this criteria to limit our analysis and we end up with 55 trading partners. The main focus will be on exports and imports of Turkey with the 55 different trading countries in the world.

and empirical framework in Chapter 3. Our intention of this order is to facilitate an un-derstanding of the issues surrounding eventual membership for the reader. In Chapter 4 we outline the gravity model which will be used in our regression itwill also present the regression analysis respectively the results of the regression. Chapter 5 presents our analysis. The final chapter presents the conclusions and suggestions to further research.

1.4 History of the European Union

The idea behind the European union is that the European countries seeked for peace and stability towards war. The initiating countries were the France, Italy, Germany and Benelux countries and their aim was to make war materially impossible. The integration process started with the first European organization, UNECE (United Nations Economic Commisission of Europe). Its initial aim was to carry out economic reconstruction and encourage cooperation among all the states of Europe. As the years went by the process continued and the enlargement of the union continued as well as the increasing organi-sation (Nello, 2005). In 1950, the European countries created the European Coal and Steel Community and later on the agreements which led to the signing of the Treaty of Rome in 1957. In 1958 these countries created the European Economic Community (EEC) for a common market. Europe enjoyed economic growth and the union became greater with other countries joining and political change within member states such as the unification of East and West Germany led to further enlargement. In 1993, the single market for the European countries took action and the `four freedoms´ of; movement of goods, services, people and money where implemented (EU). The unification of Europe and the enlargement of EU has continued and created a common curreny for many of the member countries. In 1999 the Helsinki European Council took the decision to open enlargement negotiations with Malta and to treat Turkey as a candidate. The union grew in 2004 when 10 new countries became members, today the enlargement is even greater and Turkey is one of the countries waiting to affiliate (Nello,2005).

1.5 Background of the Turkish economy

Turkey’s economic history has been turbulent with many transformations, economic re-forms and policy changes. The politicians decided to open up the economy for the inter-national trade in the 1990s after many years of closeness to the world economy. Turkey on the other hand started started as early as in the 1950´s to cooperate with the other European countries. In 1948 they became members of the OECD and the year after that they joined the Council of Europe (Nello, 2005).

Years later the poltiticans in Turkey wanted to continue the cooperation with the euro-pean countries and in 1963 Turkey signed an Association Agreement, the Ankara Treaty with the community. This agreement on the other hand set off discussions about the dual European and Muslem identities of Turkey and the implications for EU member-ship. The process for the Turkish affiliation continued and one of the rejections of this process was a participation in the European Political Co-operation. In 1980 the military took over in Turkey which implied that the relations with the EU remained suspended (Nello, 2005).

tion. With a new regime, Turkey opened up to the international market. These changes led to new trade agreements for example in March 1995 they signed up to the World Trade Organization (WTO) as an original member (Sübidey, 2005).Becoming a member of the EU became the next goal for the country to achieve.

Today, most of the European countries are members of the union which is expanding over time, Turkey being one of three countries on the list of candidate members. The country has for a long time been cooperating and been a member of a trade union with EU. Furthermore, the country has a number of barriers to become a full member of EU due to economical and political issues; these problems have delayed their application for becoming full members of the union. Turkey applied as early as in 1987 for full mem-bership but, as EU was going trough major internal changes and some other impedi-ments with the developimpedi-ments of the Eastern Bloc countries and former Soviet Union, Turkey could not become a full member. On the other hand EU offered to intensify its economic relations without necessarily rejecting full membership in the future (Hoak-man and Togan, 2005).

The process continued and on January 1, 1996 the outcome was a customs union for in-dustrial goods. This agreement implied that all inin-dustrial products except the goods of the European Coal and Steel Community that fulfil the European Community norms could flow, without restraints between Turkey and the member countries of EU (Togan, 2005). Turkey could then start their planning for a full membership due to the customs union agreement. Due to the exclusion of the agricultural sector in the trade agreements, the EU gave Turkey an opportunity to adapt their agricultural policies to the EU stan-dards for future membership. As Turkey aims for a full membership to EU, they have been placed in a program of preparation and the customs union sets out one part of the grounding. In the Helsinki program Turkey was included in the “National Program of the Adoption of the Acquis” (Hoakman and Togan 2005, page 127) which implies that Turkey has the basis of EU legislation. The program includes the important economic aspects of the legislation and it also includes a common market with the so called four freedoms (Nello, 2005).

2 Theoretical framework

Economists have been arguing about the meaning of international trade, and all of them agree that countries that trade between one another are doing so because it is beneficial for them. One country exports its most efficiently produced good or service to the sec-ond country and vice versa. (Howse and Trabilcook, 1995)

2.1 Comparative Advantage

As stated in Krugman (2006), comparative advantage (which David Ricardo developed as the main reason for trade between countries, called the Ricardian model) is an impor-tant factor when discussing international trade:

“Trade between two countries can benefit both countries if each country exports the goods in which it has a comparative advantage” (Krugman and Obstfeld 2006, page 12).

An example of comparative advantage is given in the figure 2.2.1. Quantities produced for both goods are presented, horizontally and vertically, as well as a production possi-bility frontier curve (linear) for no trade and internal and external trade. The products given are skates and electronic heaters, which we will continue to use in our example of goods in this chapter. The figure shows how the purchasing power frontier (PPF) curve shifts when the country opens for international trade.

2.1 Figure Comparative advantage

(Source: Krugman and Obstfeld, 2003)

Quantity of skates, Qs Quantity of electric heaters, Qe Quantity of electric heaters, Qe Quantity of skates, Qs No Trade Trade No Trade Trade Home Foreign

The Ricaridan model states that there is only one factor of production affected by inter-national diversities causing comparative advantage to appear; labour output. Two very famous Swedish economists did not agree with this argument, Eli Heckscher and Bertil Ohlin. Heckscher and Ohlin were against the belief that labour was the only factor of production explaining trade. Instead, they put more emphasis on the relative abundance of labour in different countries that enlightens trade appropriately. They agree that tech-nology of production is also a major factor in the model where production of different goods is strongly connected to the different factors of production. This given informa-tion regarding the model is stated in relative abundance at all times. Now, if country A has 70 million workers and 140 million acres of land, it has a one-to-two labour to land ratio. Hence if country B has 70 million workers and 70 million acres of land, it has one-to-one labour to land ratio, we easily distinguish that country B is labour abundant and country A is land abundant (Krugman and Obstfeld, 2003).

2.2 Relative Prices

In this case we can provide a figure (2.2) where trade bonds with the convergence of the relative prices. Skates and electronic heaters are used as the goods produced. If we as-sume that skates are labour intensive goods which are produced in country B, as the price of skates increases, this will in turn increase the purchasing power of labour in terms of both goods as it declines in terms of trade for land. What this means is that in-dividuals in country B, which receives wages due to skate production, gain from trade but those who receive wages due to electronic heater production becomes worse off (it is classified that it is a land intensive good). If we now look at county A, the production and trade of electronic heaters will make individuals better off, and those receiving wages due to skate production are worse off. Furthermore, looking at figure 2.2, we as-sume that country A is on point two and country B on point one before trade was intro-duced. When trade occurs then we can clearly see a convergence of the relative prices causing the prices of country B products to decline and price in country A to rise which will end up on point three.

2.2 Figure Relative prices

(Source: Krugman and Obstfeld, 2003)

Relative price of skates, Ps/ Pe Relative quantity of skates, * * Qe Qe Qs Qs + + RS* RS RD 1 3 2 RS = Relative Supply RD = Relative Demand

2.3 Production Possibilities

If we now consider two countries, A (home) and B (foreign), that can produce the same goods under a free trade agreement. Both countries specialize in producing one of these goods which they are best at, with the cost of production as significant factor. Country A is producing skates (situated in cold climate) and country B is producing electric heaters (situated in milder climate), and we assume that these products produced are the main exports of the respective countries. If on a monthly basis, country A produces 10000 skates and 0 electric heaters, and country B produces 1000 electric heaters and 0 skates, and if 1 unit of electric heater is equal to 10 units of skates in terms of produc-tion cost trade between these countries would be beneficial, making both countries bet-ter off, which is explained more specifically in figure 2.3. In addition, Country A can produce electric heaters efficiently by having skates made and traded against electric heaters, and country B can produce skates by having electric heaters made and traded against skates, which is described as gains from trade. Both countries gain and can pro-duce both goods which are demanded in the countries (Krugman and Obstfeld, 2006).

2.3 Figure Production possibilities

(Source: Bade and Parkin, 2004)

This kind of trade known as comparative advantage in this specific case occurs in un-ions/agreements such as the EU, NAFTA as well as the ASEAN Free Trade Area.

1000 10 000 9000 900 1000 100 PPF (Country A) Quantity of skates, Qs Quantity of elec-tric heaters, Qs (Country B)

2.4 Trade Creation and Trade diversion

Trade creation and trade diversion are two words that are related to trade agreements. Trade creation occurs when the member-countries in a union increase their trade and trade more within the union. Becoming members of these different unions the countries divert their trade to the union countries and away from their previous trading partners. As international trade has increased over the years the need for trade agreements and custom unions followed the pattern. Having customs unions among countries abolishes the problems of intra-union trade of commodities coming from other countries. Jacob Viner was the first to integrate the words trade creation and diversion into customs un-ions. He distinguished trade creation from the trade diversion effects of a union. (Horvat, 1999)

The attempts for trade creation started early in Western Europe. In 1958 the European Communities (EC) and in 1960 with the European Free Trade Area the consequences of this ended up in the elimination of tariffs in trade between member countries. Nowadays the inducement for intra European trade has been given an incentive trough the creation of the Single European Market in 1993 (Grimwade, 2000).

The aim for the future member countries is to increase trade when joining a customs un-ion, according to Viner (1950). This is the outcome of a customs union whereby coun-tries can increase the trade volume with one another.

The other effect of a customs union is trade diversion and according to Viner (1950) this is only beneficial for the members of the union. As the member countries divert their trade to the other members there is an economic inefficiency and an economic surplus.

3

Institutional and Empirical framework

The enlargement of the EU in the last years has been substantial; some of the latest members have been Eastern Europe countries. Next in line for inclusion is Croatia, fol-lowed by Turkey and Macedonia. Turkey applied for full membership as early as in 1987 and has in recent years become a much debated topic in the enlargement process (Hoekman and Togan, 2005). One of the main purposes for the union is to create a sin-gle market for the European countries with free mobility; this implies having a common trade policy protecting the member-countries against non member countries (Allgårdh and Norberg, 2004).

The integration process for Turkey would imply many structural changes and adopting new agricultural policies; the main transformation would be adopting an institutional framework compatible with EU´s Common Agricultural Policy (CAP). The transforma-tion and adoptransforma-tion to EU would not be an effortless task. Previous studies show that a full membership for Turkey would imply that the trade patterns will not change substan-tially but on the other hand the trade volumes would increase with EU. The pressure on Turkey from the EU concerns many themes, among them the reforms of the agriculture sector will require that the majority of the population will be confronted with severe changes (Hoekman and Togan, 2005).

3.1 The accession of the Eastern countries

The enlargement process for the EU was divided in two important dates, the first one in 1998 with some of the candidates in Europe; Cyprus, Estonia, Hungary, Poland, the Czech Republic and Slovenia. The second accession was in 2002 where Bulgaria, Lat-via, Lithuania, Malta, Romania and Slovakia began their negotiations. These countries where considered to have a good base for an eventual enlargement, the only country that was excluded was Turkey. The candidate countries had regular progress checks which were recorded in annual reports from the commission. In 2002 the Copenhagen Euro-pean Council considered that 10 of these countries had fulfilled the conditions that the EU had set up. These countries had to sign the Accession Treaty in 2003 in Athens and their full memberships were established in 2004 when the actions were approved. The enlargement has led to a greater market for these new members and even for the previ-ous members. Reforming actions has taken place for the new member countries; they have diverted their trade towards EU and also implemented new laws (EU).

3.2 Descriptive Data

In this paper we have analyzed the interval period of 1995 to 2005, the figures show the trade performance of the Turkish economy. Our data is collected from the European commission’s statistical yearbook; with this data we have created two figures, the first figures present Turkey’s trade with the EU 25 and the second presents Turkey’s trade with non-EU members

The economy has had two major recessions during this period. In the year 2000 the economy performed reasonably well but the preceding and following years are pre-sented in the figure (3.1) as recession years in terms of trade. Imports from the EU

(3.2) is presented for the imports from non-EU member states but not as drastic as the previous decrease.

Another important factor that we can conclude from the figure (3.2.1) is that imports overall has increased the following years.

3.1 Figure Trade with EU member states

(Source: European Commission, Statistical yearbook, 2006)

Figure 3.1 presents the trade pattern for Turkey with EU member states. This figure states that exports have increased constantly over time, but with two major declines in 1999 and 2001. It is clearly seen that in the years 1999 and 2001 the Turkish economy was in an economic crisis. In 1999 Turkey suffered an economic crisis due to a massive earthquake which yielded in a major financial crisis in 2001. Trade with the EU mem-bers was affected and the Turkish economy suffered from a decline in the export rates. Imports have followed the same pattern as exports but with fewer disturbances. Figure 3.1 shows a steady increase in imports from the European member states despite a small decline in the same year as the export. The data for export during 1995-1998 was not aviable in the Statistical yearbook.

0 5 10 15 20 25 30 35 40 45 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Years

Value (Mio Euro)

3.2 Figure Trade with non-EU members

(Source: European Commission, Statistical yearbook, 2006)

This figure 3.2 shows Turkey’s trade with non-EU member-states. The tendency is quite similar; there is an upward trend in both exports and imports. The main difference here is that imports from non-European member-states declined during the two crisis years, while exports actually increased from 1999 to 2001 and kept increasing after that. In summary it is clear that during the year’s 1999 and 2001 Turkey’s economy experi-enced an economic crisis which led to a decline in exports to EU 25 and an increase in imports to rest of the world (ROW).

In addition, after Turkey signed the contract in 1996, figure 3.1 show an increase in im-ports from the EU during its first years compared to the ROW. Furthermore, these two figures can be combined with the figure (5.1) which we achieve from our regression, and will be discussed in chapter 5. The data for both export and import in this figure during the years 1995-1997 was not aviable in the Statistical yearbook.

0 10 20 30 40 50 60 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Years Value (Mio Euro)

4

Regression Model

4.1 Gravity model

One of the most successful models used to explain the international trade flows is the gravity model. This model can include and explain different types of flows such as mi-gration and transports of goods (Bergstrand, 1985).

The gravity model is used to analyze the trade between two countries and can be de-scribed with this simplified equation.

Equation (1): ij j i ij A Y Y D T = × × /

The equation provides us the variables that decide the volume of the trade between two countries. The determining factors are the size of the two countries GDP and the dis-tance between the countries (Krugman and Obstfeld, 2006).

ij

T the value of the trade flow between country i and country j

A is a constant term i Y is country i´s GDP j Y is country j´s GDP ij

D is the distance between the two countries

In addition, when talking about the gravity model, we consider trade between two coun-tries to rely generally on their transaction costs and substantial size. These factors are basically the GDP, population, land area, distance, cultural similarities etc (Soloaga and Winters, 2000).

4.2 Variable formulation Trade factors

When countries trade, there are some aspects that needs to be distinguished so trade be-comes as efficient as possible. These are: distance, language, and culture. An explana-tion for these variables is given in the below secexplana-tion.

Dependent variables

Imports, ΙΜ One of the basic statements in economics is that imports are goods and ij. services that are purchased from other countries. The foreign companies provide their goods to the domestic consumers due to their demand for their good or service (Bade

Furthermore, we have removed three years (1999 – 2001) for Luxembourg since data was unavailable for these years (UN Comtrade).

Independent variables

Gross domestic product, GDPij. GDP is the total value of all final goods and services produced in a country for a given year (time period). In addition, this can be related to purchasing power since GDP corresponds to a country’s income. What purchasing power simply means is that the goods and services bought have the same relative value in monetary terms across different currencies (Bade and Parkin, 2004). The data for the GDP is collected from The United Nations Statistics Division and the rate is in current US dollars.

Gross domestic product per capita, GDPCij, where GDPC is measured in the way that GDP is divided by the population. This will denote an eye, of overall perspective for a country’s wealth, which is positively associated to trade (Bade & Parkin, 2004). Our data for our selected years is collected from Worldbank.org and shown in US dollar rate.

Distance, Dij. Distance is an important element since costs can be cut down the closer the partners are located to one other (Hacker & Johansson, 2001). Furthermore, for an industry (country) to accumulate cost, it trades with neighbour countries (Brakman, Garretsen & Marrewijk, 2001). In essence, since distance reflects transport and transac-tion costs and is therefore assumed to diminish trade as the distance between trading partners increases the greater the costs. Our findings for distance have been collected from geobytes.

Dummy variables

Language & Culture, L. When it comes to language it is clear that this has a role within trade. If two countries have the same language, trade between these countries is more likely to appear. The communication is at the same level for both countries and trade is at its best without mistakes appearing (Hacker & Johansson, 2001). Language does have a connection with culture in the sense of that language is an aspect in culture. As stated in Krugman and Obstfeld (2006), people sharing the same life standards and values. Dressing similarly, watching the same type of movies, eating the same type of food, and most of all listening to the same type of music. The importance of this is that culture has its part within trade. When looking at language and culture, both is positive to trade where language makes it easier to communicate and culture makes it easier for countries to trade in the sense of similar life standards. We have coded the countries such as 1 for countries sharing the same language as Turkey and coded 0 for countries not sharing same language. The countries which share the same language with Turkey are; Iran, Cyprus, Bulgaria, Syria

Armenia, Iraq, Azerbaijan, and Georgia. The border countries have a code 1 and the non-border countries have a code 0. However, we excluded Syria, Armenia, Iraq, Azer-baijan, and Georgia, since their GDP per capita is less than Turkey’s and we are only looking at the countries which have larger GDP per capita than Turkey.

Block Dummy, PkiPkj, the block dummy is going to represent whether or not the coun-tries are members of the European Union. Few economists have stated that free trade agreements are good for countries, which are introduced to boost trade (Bowen, Hol-lander & Viaene, 1998). Furthermore, one of the European Union’s major goals is to remove trade barriers to the member states to help them to increase their wealth as well as economic growth (Allgårdh & Nordberg, 2004). 1 and 0 is going to represent holding membership or not, respectively. The block dummy is coded as 1 if the country is a member of the EU and 0 if it is not a member.

4.3 Model formulation

In this paper we will use a gravity model which is influenced by the model that is used by Soloaga and Winters (2000). In their paper they study what effects regionalism has on trade and their conclusions are that new regionalism does not boost the intra-bloc trade they find increased trade diversion to EU and EFTA. Previous studies about Tur-key’s affiliation to EU are done by Hoekman and Togan (2005), where they use the gravity model to analyze the trade flows between Turkey and EU. The authors empha-size on the fact that Turkey has had free trade with EU since 1996, except for agricul-tural products. However, their conclusions about agriculagricul-tural products are that trade pat-terns will not change but there is a possibility of an increase in the volume of trade. This differs from our study as we look at EU-25, including year dummies and exclude vari-ables as well as the focus on trade creation and trade diversion.

Equation (2) explains the trade flow from country j to i if neither country were a mem-ber of a trade union. We have used a dummy variable to distinguish the countries being members of the European Union and those countries which are not members. This vari-able is the block dummy,PkiPkj.

Equation 2, Econometric Approach

(

)

ε

β

β

β

β

γ

β

β

∑

jdenotes trading partner with Turkey ij

ij

D is the distance (in Km) from Turkey’s capital city to the trading country’s capital city

ij

B is a dummy that takes the value 1 if the trading partner shares border with Turkey and zero otherwise

ij

L is a dummy that takes the value 1 if the trading partner shares the same language with Turkey

kj kiP

P is a dummy taking the value 1 if the trading country is a member of EU and tak-ing the value 0 if not (Soloaga and Winters, 2000)

D

Y is a year dummy taking the value 1 for each reference year

05 04 03 02 01 00 99 98 97 96 1 D D D D D D D D D D Y n i D = + + + + + + + + +

∑

In addition, when we analyze the regression results we are specifically going to look at whether the signs are positive or negative respectively.

4.4 Regression Analysis and Results

In order to analyse whether trade creation and trade diversion is accomplished for Tur-key, if they become a member-state of the EU, we use the following hypothesis. We have analyzed different variables in our regression and focused mostly on the bloc dummy variable.

The hypothesis for our paper is:

0 : 0 : 1 ≠ = i i o H H

β

β

Our null hypothesis H₀states that the β values are zero. The alternative hypothesis H₁

states that the β values can differ. In order to analyse the effect of an eventual member-ship for Turkey we have used a gravity model. We have used cross-sectional data to run a Weighted Least Squared (WLS) log linear regression testing at a 5 percent signifi-cance level and a critical t-level of 1.96 (Guajarati, 2003).

Furthermore, we made use of annual export data for 55 countries that trade with Turkey (Appendix 1) for 1995-2005 from the UN-COMTRADE database, using 14 different variables.

Table (5.1) presents the expected signs for our regression. The GDPjvariable is ex-pected to be positive because this variable reflects on the economic size of a country and the country’s absorptive capacity. The GDPCj variable is on the other hand

uncer-tion also promotes division of labour and implies the presence of economies of scale in production and therefore also the opportunities and desire to trade with a greater variety of goods. The distance variable is a measure for transport and transportation costs and is expected to have a negative sign. These measurements are related to different things; transport costs relate to distance while transportation costs reflect better to information and smaller cultural differences when countries are neighbouring. When using the grav-ity model the dummy variables often denote factors that increase trade such as language, border and the bloc dummy in our case (Nilsson, 2000).

4.4 Table Expected Signs of the Variables in the Gravity Model

4.5 Results

The first step in our regression is testing all independent variables individually against the dependent variable, which is done to confirm if they correlate. The table presents the results that we got when we ran an Ordinary Least Squared (OLS) regression for each of the variables.

The results are seen in the Appendix 2. We can conclude that the results achieved from running the OLS regression are significant at a 5 percent level of significance and a Variable Expected

Sign

Motivation

j

GDP + Economically larger countries import more

j

GDPC +/- Depends on relative changes between population

and gross domestic products

Distance - Transport and transaction costs

Language + Common language can lead to increased trade

Border + Possibility for direct land communication

Block- dummy

explanatory variables that are supposed to correlate positively and our regression values conform to our expectations. On the other hand, distance variable which is expected to be negatively correlated to the dependent variable is correct as well.

4.6 Statistical errors

To observe if there is any errors in the data we run a White’s general heteroscedasticity test. The conclusions we can make from these values are that we have a problem with heteroscedasticity (see Appendix 3). The problems that we can associate with hetero-scedasticity are that OLS is no longer efficient among linear estimators; the solution to this is to run a WLS regression instead of OLS to correct the heteroscedasticity issue (Guajarati, 2003).

In addition, by looking at the multicollinarity, we founded that two variables were inter-fering with the other variables and gave us difficulties, so we removed GDPCj and the border variables. Due to the fact that we removed two variables, we are aware of the possibility of specification errors. Guajarati states:

“specification bias arises from incorrect specification of the model used in the analysis” (Guajarati, 2003, p. 366).

Furthermore, information can be seen in appendix 4 where we present descriptive statis-tics.

4.7 Joint Hypothesis testing

The final step in our paper is to test a joint hypothesis to analyze whether or not the de-pendent variable correlates with the indede-pendent variable. In our hypothesis we exclude one independent variable GDPC_j and the border dummy.

After running our regression we find that border, language and block dummy are not significant. This does not change our results for the other variables which still are sig-nificant. Furthermore, when having a significant level at 5 percent, the critical F-value is 2.13 (df 14, ∞), and our observation F-value is 129.7194 (table 6.4). We will reject our null-hypothesis since our F-value is higher than the critical F-value. Our regression line fits the data as we have a value of R-squared equal to 76 percent. Furthermore, our criti-cal statistics is at 5 percent significant level is at 1.96 (df ∞). Since all of our t-statistics values lies in our critical region we reject the null hypothesis (Guajarati, 2003). Given the results in table (6.3.1) our conclusion is that all the variables apart from the bloc dummy are significant, regarding our expected signs. The expected sign for the bloc dummy is positive but we received a negative sign. This can most likely be due to the fact that Turkey is not a member of the EU. Comparing to the earlier studies, such as Nilsson (2000), we can see the same results of countries which are not members and be-comes members of a union.

4.7 Table Joint Hypothesis Test (1995-2005) Included observation 602

Dependent Variable: Import Value Independent

Variables Coefficient Std. Error t-Statistics Prob.

GDPj 0.988284 0.025587 38.62452 0.0000

Distance -0.000203 1.47E-05 -13.80538 0.0000

Language 0.613228 0.248102 2.471680 0.0137

Bloc dummy -0.302766 0.124802 -2.425970 0.0156

F-Value: 129.7194 R-squared: 0.755729

Furthermore it is important to point out that we included year dummy variables in the regression, but have chosen to not include them when describing the results. The year dummy variables are explained in more detail in figure 6.3.2.

5 Analysis

In this paper, we have studied how the Turkish trade-flow pattern will be affected by a full membership in the European Union. Discussions concerning this topic have been made in earlier research, chapter 3, due to the fact that Turkey, many decades ago, ap-plied for EU membership. Even though Turkey is not a full member in the EU, they have trade agreements specifically in the industrial sector. In 1996, Turkey signed a cus-toms union agreement for industrial products. This agreement was beneficial for the economy, which can be seen in figure 5.1, where we clearly see that Turkey’s GDP in-creased between the years (1996 – 2005). Furthermore, Sübidey (2004) found that the trade agreement in 1996 for the industrial sector was beneficial for the country. Due to these results, he stated that similar effects would be achieved as the other sectors be-come included in new trade agreements. However, Turkey needs to reform their politi-cal situation if they want a full membership. Our statement now becomes stronger due to the fact that we achieve similar results. Following our discussion in chapter 3, we can also state that Turkey will be better off joining the union, due to the benefits of being a member.

When analysing our results from the regression we can see that the signs are as ex-pected, apart from the block dummy variable. The block dummy variable was expected to be positively correlated with the dependent variable, due to the fact that if countries are members of a union they are expected to trade more with each other. In our case Turkey is not a member of the European Union hence, Turkey trades more with ROW than with the EU.

If we take into consideration the conclusions made by Soloaga and Winters (2000), that new regionalism does not boost the intra-bloc trade they find increased trade diversion to EU and EFTA in their regression. On the other hand, comparing this with our situa-tion we can draw the conclusions that Turkey is trading more with non-EU members due to the fact that EU excludes the non-members and provide benefits to the members. Cadsby and Woodside (1993) argue that Canada will benefit if they join the Free Trade Agreement (FTA), NAFTA. The emerging competition from the southern country Mex-ico is not beneficial for Canada’s trade with the US. The authors find that due to NAFTA, Mexico can play a larger role in trade with the US compared to Canada. Com-paring Turkey with Canada conclusion is drawn that Turkey is forced to trade with ROW due to the non-membership of the EU. However, considering Turkey becoming a member-state, the possibility for their trade to divert towards EU would be greater than trade to ROW. Continuing with arguments concerning NAFTA, Krueger (1999) finds that Mexico expanded their trade when they joined NAFTA, trade creation. The re-moval of tariffs and political hinders helped the Mexican economy and increased the share in GDP in trade. Turkey has as said before already joined a customs union in the industrial sector, but there are still other sectors needed to be included. If Turkey be-comes a member-state, the possibility of trade creation is truly. However, Krueger (1999) do not find trade diversion in her analysis due to the fact that two thirds of the trade was already with the U.S. On the other hand Turkey needs to incorporate other sectors if they become member. Hence, we believe that diversion in these sectors will occur.

We have also looked specifically on Turkey’s GDP changes during the years we have chosen to analyse and trade. The figure on next page clearly shows how there has been an upward tendency since they signed a trade agreement in 1996.

5.1 Figure Turkeys’ GDP (1996 – 2005)

Figure 5.1 explains the deviations in Turkey’s trade during the time period 1996-2005. The reference year is 1995; hence the figure shows how Turkey’s GDP has changed from one year to another compared to the reference year (appendix 5). Furthermore, the time period is set by year dummies on the Y-axis where these control the measurements for changes over time. The X-axis represents Turkey’s GDP during this time period. . As mentioned earlier, Turkey signed a trade agreement with EU which opened up the doors for a common European market. As seen in figure 5.1, after signing the agree-ment, 1996, Turkey’s economy has been increasing over time, but there are two major drops. In 1999 and 2001, economic crisis in Turkey caused their GDP to decrease. When running our regression the result achieved is similar to the Turkish GDP drop (1999 and 2001 having the lowest values). During these two recessions Turkish trade flow decreased with both EU-members and ROW.

can clearly see a relation between these figures when looking at the contract that Turkey and EU sign in 1996. The increased trade towards the EU countries and the growth for Turkey’s GDP is stronger in the beginning years taking into consideration that trade with ROW still is great after the crises. Trade increases if Turkey’s GDP increases. The calculated elasticity shows how sensitive trade is towards changes in Turkish GDP:

Υ Χ ∆Χ ∆Υ = ∆Χ ∆Υ = * % % ε Where,

ε

% is the percentage change in the coefficient

∆Χ

% is the percentage change in Turkey’s GDP ∆Υ is the change in the coefficient

∆Χ is the change in Turkey’s GDP Υ is the average of the coefficient Χ is the average of Turkey’s GDP

By looking at the statistics achieved from the figure (5.1) Turkey’s GDP being X in the formula and the year dummies (coefficient) are Y. The calculated elasticity achieved is 0.00395 which means that the supply is price inelastic (Appendix 6). This implies that Turkish trade flow would not be affected severely due to price changes, considering Turkey being a member-state (Varian, 2005).

Furthermore, in our paper, we concentrated on Turkey’s trade flow and how it will be affected for an eventual membership. We did not look into the cultural aspect, such as politics, where Turkey needs to adjust for EU-policies in the first case and accept the union’s rules. This is one of the major problems that Turkey has, but we excluded this from our paper.

To sum up the arguments we have mentioned earlier in our analysis we want to point out that if Turkey would become a member-state, they would divert their trade towards the member countries. Becoming a member they would have access to a greater market and by that it would imply trade creation, specifically in the sectors where no trade agreements are signed. We believe that it is a mutual benefit of an eventual membership and that Turkey’s trade flow would increase for those parts of the economy which are not included in any agreements. Considering this we do take into consideration that this is only a paper analysing the trade flow pattern and not including all other important factors, as political and cultural aspects.

6 Conclusion and suggestions to further research

The purpose of this paper is to analyse how Turkey’s trade-flow pattern will be affected by a full membership in the European Union. The main question raised is; will trade creation and trade diversion occur towards the European Union? The conclusions are drawn from the empirical findings and the theoretical background.

The conclusion of this paper is that the trade flow for Turkey would increase if they be-come a member-state of the EU. By looking at the trade flow pattern during 1995-2005, we can clearly see and upward trend in Turkey’s GDP by looking at the figure 5.1 as well as the significant results we achieve from our regression. From our elasticity re-sults, we conclude that as long as Turkish GDP is increasing trade will increase. We also believe that trade diversion and trade creation will occur with an eventual member-ship towards EU.

In addition, we cannot tell what the future holds but we can look in to the trend due to their trade flow increase and conclude that Turkey’s trade flow and economy will be better off as a member-state.

An interesting topic to extend this analysis is to put more emphasis on the agricultural and industrial sector. Dividing these two sectors and analyzing their influence on the Turkish trade flow with EU and ROW. Other important factors are the political social and cultural aspects. An empirical study with these aspects in consideration would broaden the analysis of an eventual membership. To divert the study towards EU and analyze whether or not is beneficial for the member countries to have Turkey as an af-filiate. The Middle Eastern countries have an important role regarding trade with Tur-key, an important and interesting topic is too see how they would be affected if Turkey becomes a member-state.

References

Allgårdh, O. and Norberg, S. (2004) EU och EG.rätten; En läro- och handbok om EU och i EG-rätt. 4e upplagan. Stockholm: Norsteds Juridik AB

Arjan, M., and Lejour, R. A. de Mooij, (2005) Turkish Delight: Does Turkey's Acces-sion to the EU Bring Economic Benefits?Kyklos, 58 (1), 87-120.

Bade, R and Parkin, M (2004) Foundation of Macroeconomics. 2nd ed. Boston: Pearson Addison-Wesley

Bade, R. and Parkin, M. (2004) Foundation of Microeconomics. 2nd ed. Boston: Pearson Addison-Wesley

Barkman, S. Garretsen, H and van Marrewijk, C. (2005) An introduction to geographi-cal economics, Cambridge: Cambridge University Press

Bergstrand, H. J. (1985) The Gravity Equation in International Trade: Some Microeco-nomic Foundations and Empirical Evidence. The review of Economics and Statistics, 67(3), 474-481

Bowen, Hollander and Viaene (1998), Applied International Trade Analysis. Palgrave Macmillan (United Kingdom)

Cadsby, C. and Woodside, K.(1993), The effects of the North American Free Trade Agreement on the Canada-United States trade relationship. Canadian Public Policy / Analyse de Politiques, 19 (4), 450-462.

Colton, N. and Neaime, S. (2003). Implications of the Introduction of the Euro for the Mediterranean Countries. Thunderbird International Business Review, 45(1), 31-49

Europa – The EU at a glance. (Updated: daily). [Online] Available from: http://europa.eu/abc/history/index_en.htm - Retrieved: 2007-09-25

Europa.eu – The 2004 enlargement: the challenge of a 25-member EU. (Updated: 2007-01-23) [Online]. Available from: http://europa.eu/scadplus/leg/en/lvb/e50017.htm Retrieved: 2007-10-10

Eurostat – Statistical yearbook 1998 – 2005 (Updated: 2006). [Online] Available from:

http://epp.eurostat.ec.europa.eu/cache/ITY_OFFPUB/KS-CV-06-002/EN/KS-CV-06-002-EN.PDF - Retrieved: 2007-10-01

Eu-upplysningen – (Updated: 2007-06-15) [Online]. Available from: http://www.eu-upplysningen.se/Kort-om-EU/EU-pa-tio-minuter/

Retrieved: 2007-09-25

Federal Research Division – The Library of Congress (Updated: 2007-11-01). [Online] Available from: http://www.loc.gov/rr/frd/ - Retrieved: 2007-25-09

Grimwade, Nigel, (2000) International Trade: New Patterns of Trade Production and Investment. 2nd ed. Routledge

Gujarati, N.D. (2003) Basic econometrics. 4

th

ed. Boston: McGraw Hill

Hacker, S. and Johansson, B. (2001) Sweden and the Baltic Sea Region: transaction costs and trade intensities. In Bröcker. J and Herrmam Hayo, ed. Spatial Change and interregional flows in the intergrating Europe: essays in hounoir of Karin Peschel. Hei-delberg: Physica-Verlag

Hoekman, Bernard M. and Sübidey Togan (2005). Turkey: Economic Reform and Ac-cession to the European Union. The World Bank

Howse, R. and Trebilcook, M. (1999) The Regulation of International Trade. 2nd ed. London: Routledge

Herndon, VA, USA: World Bank.

Horvart, Branko (1999) the theory of international trade, Palgrave Publishers

Krugman, P.R. and Obstfeld, M. (2003) International Economics Theory and Policy. 6

th

ed. Boston: Addison Wesley

Krugman, P.R. and Obstfeld, M. (2006) International Economics Theory and Policy. 7 ed. Boston: Addison Wesley

Nello, Senior Susan (2005) The European Union: economics, Policies and History,

Berkshire: McGraw-Hill Education

Nilsson, L (2000) Trade integration and the EU economic membership criteria. Euro-pean Journal of Political Economy Vol. 16 (2000) 807 – 827

Ramanathan, R. (2002) Introductory Econometrics with Applications, Fifth Edition. South- Western, Mason

Redmond, J (2007) Turkey and the European Union: troubled European or European trouble? International Affairs.83(2), 305-317

Sübidey,T (2004) Turkey: Toward EU Accession. The World Economy, 27(7), 1013-1045

Sübidey, T (2005) Turkey: Trade Policy Review. The World Economy, 28(9), 1229-1262.

Varian, R. Hal (2005) Intermediate Micro Economics: a modern approach. 7th ed. W. W. Norton & Company, New York, London

Viner, Jacob (1950) The customs union issue, New York: Carnegie endowment for In-ternational Peace

Worldbank. (Updated: 2007) Worldbank database. [Online] Available from http://www.worldbank.org/

United Nations. (Updated: 2006) UN Comtrade Database [Online]. Available from: http://comtrade.un.org/db/dqQuickQuery.aspx

Appendix 1 Trade Countries 1995 - 2005 NR Country NR Country 1 Argentina 30 Kuwait 2 Australia 31 Latvia 3 Austria 32 Libya 4 Bahamas 33 Lithuania 5 Belgium 34 Luxembourg 6 Brazil 35 Malaysia 7 Bulgaria 36 Malta 8 Canada 37 Mexico 9 Chile 38 Netherlands

10 China 39 New Zealand

11 Colombia 40 Norway

12 Costa Rica 41 Poland

13 Croatia 42 Portugal

14 Cyprus 43 Romania

15 Czech Republic 44 Russian Federation

16 Denmark 45 Saudi Arabia

17 Dominican Republic 46 Singapore

18 Estonia 47 Slovak Republic

19 Finland 48 Slovenia

20 France 49 Spain

21 Germany 50 Sweden

22 Greece 51 Switzerland

23 Hungary 52 Thailand

24 Iceland 53 United Kingdom

25 Iran 54 Uruguay

26 Ireland 55 USA

27 Israel

28 Italy

Appendix 2

Individual Variable Testing (1995-2005)

Variable Coefficient Std. Error P-value R-squared

GDPj 0.949687 0.031219 0.0000 0.606654 GDPCj 0.327810 0.079390 0.0000 0.027631 Distance -0.000157 2.09E-05 0.0000 0.085712 Language 0.071826 0.415728 0.8629 0.000050 Border 0.248444 0.363430 0.4945 0.000778 Block Dummy 0.667490 0.188185 0.0004 0.020538

Appendix 3 Method:

White’s Heteroskedasticity Test

F-statistic 2.562592 Probability 0.000000 Obs*R-squared 137.0523 Probability 0.000000

Appendix 4 Descriptive Statistics Mean Std. Deviation N Import Value 19.0036 2.31992 602 GDPj 4.6093 1.90267 602 Block_dummy .45 .498 602 Language .05 .228 602 Distance 4826.48 4329.757 602 D96 .09 .286 602 D97 .09 .286 602 D98 .09 .286 602 D99 .09 .284 602 D00 .09 .284 602 D01 .09 .284 602 D02 .09 .286 602 D03 .09 .286 602 D04 .09 .286 602 D05 .09 .286 602

Appendix 5

Year dummies and Turkeys’ GDP (1996-2005)

Years (Dum-mies) Coefficient (X) Turkeys GDP (Y) D96 -0.556509 178.061 D97 -0.474845 186.061 D98 -0.528760 197.587 D99 -0.679534 181.69 D00 -0.356975 198.23 D01 -0.763393 143.096 D02 -0.448159 182.973 D03 -0.123891 240.596 D04 0.022662 302.561 D05 0.142968 362.461

Appendix 6 Price elasticity 00395 . 0 3316 . 217 3766436 . 0 * 896525 . 3 888455 . 0 * = = Υ Χ ∆Χ ∆Υ =

ε