T he Effects of EU -membership
How distance and borders have been affected by the
EU-membership and the effects on Swedish trade with the EU14
Bachelor’s thesis within Economics
Authors: Lisa Olofsson 880115 Lisa Wassén 901122
Tutors: Charlotta Mellander (Supervisor) Mark Bagley (Deputy Supervisor)
Bachelor’s Thesis in Economics
Title: The Effects of EU-membership
Authors: Lisa Olofsson 880115
Lisa Wassén 901122
Tutors: Charlotta Mellander (Supervisor)
Mark Bagley (Deputy Supervisor)
Date: May 2013
Subject terms: Trade, EU, Sweden, Intra-industry trade, Gravity model
Abstract
This bachelor thesis investigates how distance and borders have been affected by the EU-membership and what effect it had on Swedish trade with the EU14. By comparing the time periods 1980-1994 (before the entry) and 1995-2009 (after the entry), it is possible to see changes in trade.
Three commodity groups important for Swedish trade were analysed in order to see the effects of the EU-membership. The chosen commodity groups are; (i) Chemicals and related products, nes. (ii) Machinery and transport equipment (iii) Crude materials, inedible, except fuels.
By modifying the original gravity model to fit our purpose and use it for re-gression analysis, the results are interpreted using trade theories. OLS-regressions were run, using five variables known to have an impact on trade. The Grubel-Lloyd index was used to analyse the level of intra-industry trade, for the chosen commodity groups, it was shown that the trade patterns did not change drastically after the EU-membership. The regression results also showed that some variables became less important for trade after the member-ship. Sectors were affected in different ways and it is assumed that other fac-tors became more important for trade after the EU-membership such as sharing policies and regulations.
T able of Contents
1
Introduction ... 1
1.1 Purpose ... 2 1.2 Limitations ... 2 1.3 Outline ... 22
Background ... 3
2.1 European Union ... 32.2 Sweden’s Trade with the EU14 Member States ... 3
3
Theoretical Framework ... 5
3.1 Ricardian Model ... 5
3.2 Heckscher- Ohlin Trade Theory ... 5
3.3 Inter-Industry Trade ... 6 3.4 Intra-Industry Trade ... 6 3.5 Gravity Model ... 7 3.6 Previous Studies ... 8
4
Method ... 10
4.1 Model Description ... 104.2 Explanation of commodity groups and variables ... 11
4.2.1 Expected Results ... 12
5
Empirical Findings ... 13
5.1 Descriptive Statistics ... 13 5.2 Correlation Analysis ... 14 5.3 Statistical Errors ... 16 5.4 Regression Analysis ... 17 5.5 Grubel-Lloyd Index ... 246
Conclusion ... 25
6.1 Suggestions for Further Studies ... 26
Figures
Figure 5.1………..15
Tables
Table 2.1: Swedish Trade 2012 ... 4Table 5.1: Descriptive Statistics for Export ... 13
Table 5.2: Descriptive Statistics for Import ... 14
Table 5.3: Correlation: exports in the sectors and variables, 1980-1994…...14
Table 5.4: Correlation: exports in the sectors and variables, 1995-2009…...15
Table 5.5: Correlation: imports in the sectors and variables, 1980-1994…...16
Table 5.6: Correlation: imports in the sectors and variables, 1995-2009…...16
Table 5.7: Regression results for chemical exports ... 18
Table 5.8: Regression results for machinery exports ... 19
Table 5.9: Regression results for crude material exports ... .20
Table 5.10: Regression results for chemical imports ... 21
Table 5.11: Regression results for machinery imports ... 22
Table 5.12: Regression results for crude material imports ... 23
Table 5.13: Grubel-Lloyd Index 1980-1994 ... 24
Table 5.14: Grubel-Lloyd Index 1995-2009 ... 24
Appendix
Appendix 1 ... 30Appendix 2………...….31
1
Introduction
This section gives an introduction to the thesis. The purpose and limitations of the paper are clarified and the outline is presented.
Increased trade between the European countries is one of the European Union’s (EU) most important priorities. It was one of the main arguments for a Swedish membership, and after years of discussion, Sweden joined the EU on January 1st, 1995 (Regeringen, 2013). When establishing the EU, the belief was that an economic union would benefit each member country and increase trade. A single market through liberalization of tar-iffs, with more varieties for consumers in each country, was said to increase welfare. EU’s four freedoms; freedom of movement of people, goods, services and capital, are important for the single market (EU Commission, 2013). The single market is the inter-nal market within the European Union. The open borders have helped globalization in both EU and the world as a whole.
As some theories state, trade varies with distance. A nation tends to trade more with countries surrounding it, which is consistent with Swedish trade: some of the country’s largest trading partners are the countries in Northern and Central Europe. With the use of the gravity model, trade is analysed between the years of 1980-1994 (before the EU-membership), and between the years of 1995-2009 (after the entry into the EU), we can then see if the theories leave any questions unanswered. By discussing inter- and intra-industry trade, we can see what kind of trade is most important for Sweden and domi-nant for the single market. As explained in section 3.5, trade is most likely affected pos-itively by the size of trading countries’ GDP and negatively affected by the distance be-tween countries. Being part of a customs union should also increase trade bebe-tween the members, as will be investigated in this paper.
This paper investigates the effects Sweden’s entry into the EU had on distance and shar-ing land borders and whether these factors changed in their importance for trade. Did Sweden see a change in trade patterns with the EU14 countries following the EU-membership? The EU14 countries are the 14 countries besides Sweden that were mem-bers in EU in 1995; Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain and United Kingdom (UK) (EU, 2013a). Several of the countries are some of Sweden’s largest trading partners, and have been since international trade was first developed. Have the trade patterns changed or stayed the same? Have distance and sharing land borders become less influential after Sweden’s EU-membership in explaining trade between Sweden and the EU14? We are interested in how the EU membership affected Sweden and Sweden’s trade. Sweden has for a long time been dependent on trade, even more so now, with the open borders to the rest of the EU. Some Swedish industries have benefitted from the integration, while some industries have been hurt by the increased competitiveness (Magnusson, 2010). For this reason, the paper will explore how three industries’ trade have been affected by the EU membership. The industries focused on are: Chemicals and related products; Machinery and transport equipment and Crude materials, inedible, except fuels. These commodity groups have been chosen because of their importance to Swedish trade. Magnusson (2010) discusses the industrial revolution and how iron and wood became important inputs in countries’ production. Since these were goods Sweden was well-endowed with exports increased heavily to countries such as France, UK and Germany.
With a strong infrastructure, Sweden was dependable and the trade relationship that was built has stayed strong since.
1.1
Purpose
The purpose of this thesis is to analyse how distance and borders have been affected by the EU-membership and what effect it had on Swedish trade with the EU14. By study-ing the time period between the years of 1980-2009, we can see changes in trade. Three commodity groups are studied to analyse the trade patterns between the countries; (i) Chemicals and related products, nes. (ii) Machinery and transport equipment (iii) Crude materials, inedible, except fuels.
1.2
Limitations
By restricting the time period to 15 years before the Swedish EU membership and com-pare it to the 15 years after the entry, the analysis will not be biased towards one of the time periods. Limiting the number of trade sectors to three, one can make a deeper anal-ysis of the changes in trade patterns.
1.3
O utline
Section 2 gives the background of the EU, as well as the history of trade between Swe-den and other European countries. This gives the reader a possibility of understand the analysis better.
In section 3 we state theories regarding trade, such as comparative advantage and intra-industry trade. Previous literature regarding the subject is also mentioned.
Section 4 describes the method used for our regression. An explanation of our model as well as a description of variables in the model is included.
In section 5 the empirical analysis is shown, where descriptive statistics, correlation as well as regression analysis is explained and discussed using theories mentioned in sec-tion 3.
The conclusion in section 6 will give the reader a final remark of the effects of the EU-membership on Sweden’s trade. Suggestions for further research is also discussed.
2
Background
The background of EU and Sweden’s trade with the EU14 is presented in this section in order to give a broader understanding of the topic and the construction of the union.
2.1
European U nion
Collaboration between several European countries was established in the 1950’s on both the political and economic level, in order to preserve the peace that was seen after World War II. Six countries founded the union, and as the economy grew, more coun-tries entered the common market. With the economy following a positive trend, and the member states not paying tariffs, the economy got further fuel to function well on the market (EU, 2013b). More countries entered the EU and in the end of the 1980’s, six member states had become twelve. Today, EU has 27 member states and more countries are waiting to become members. The central forces in the EU tried to solve free-trade problems by launching a six-year program that would benefit the European market. In 1993, the EU market was completed and could then guarantee what we call the four freedoms; free movement of people, goods, services and capital (EU, 2013b). These freedoms exist so as to bring down barriers and increase opportunities for citizens and businesses in all member countries.
The EU single market led to decreased prices through competition. For firms it has be-come easier to do business across borders. Removing barriers between members has in-creased trade and made it possible to exchange goods in ways that was impossible to do before. Free movement of capital allows citizens to buy shares in another country than the residential one, making financial markets and services more competitive and effi-cient. Firms are able to invest more easily in other countries and it is easier for multina-tional companies to exist and move across borders (European Commission, 2013). The service sector is becoming increasingly important but it has been noted that it does not yet work as efficiently as hoped. In 2000 there were discussions about reducing cross-border barriers to service since the existing barriers have a negative impact on the cost of services. To see a fully integrated society, these borders need to be eliminated (European Commission, 2013).
2.2
Sweden’s Trade with the EU14 Member States
Sweden joined the EU in January 1995 and the membership was seen as a further inte-gration between Sweden and the European community. Many of the agreements were already included in the EES-agreement from 1992, leading to an easy transition (Regeringen, 2013).
Sweden has ever since the Iron Age traded with other economies and trade has always had a high priority. As mentioned by Magnusson (2010) it was clear that foreign de-mand was important for how the industries developed during Sweden’s industrial trans-formation. It was especially UK’s industrial revolution that spurred demand for Swedish products, such as wood and iron. This helped Sweden’s economy and development, and ultimately made UK one of Sweden’s most important trading partners. The natural re-sources together with an efficient organization and access to good technology led to Sweden being the successful exporting country we see today (Magnusson, 2010).
60 percent of Sweden’s imports come from countries within the EU14, and the manu-facturing industry along with the forest, steel and mining industries accounts for a sig-nificant part of Sweden’s export (SCB, 2013b). In 2012, Germany was Sweden’s largest trading partner and the trade relation between the two countries has been good through-out the years. The three commodity groups chosen for this paper are all major trading products between Sweden and Germany (SCB, 2013c). Countries close to Sweden tend to be like-minded when discussing international interests as can be seen in the UK. There is a similarity of markets in the countries, with several multinational companies and high share of privatized activities (Sweden Abroad, 2013).
In Table 2.1 below, the percentage share of trade for Sweden’s different trading partners is shown. The products are exported to numerous countries around the global market, especially to the large market in the European Union (Ekonomifakta, 2013). As can be seen in below table, Sweden’s largest trading partners within the union are countries with a smaller distance to Sweden, as is explained by the Gravity model in section 3. As mentioned above, EU14 represents roughly 60 percent of Swedish imports and more than 50 percent of Swedish exports. This shows that a smaller distance and a common market increases the possibility of trade between two countries.
Table 2.1 Swedish Trade 2012
Swedish Exports 2012 Swedish Imports 2012
EU14 countries Total: 51.54 EU14 countries Total: 59.87
Germany 10.32 Germany 17.41
United Kingdom 8.05 Denmark 8.37
Finland 6.63 United Kingdom 6.80
Denmark 6.32 Netherlands 6.29 Netherlands 5.16 Finland 4.92 France 4.97 France 4.24 Belgium 4.73 Belgium 3.92 Italy 2.13 Italy 2.66 Spain 1.47 Spain 1.97 Austria 0.83 Ireland 1.34 Portugal 0.35 Austria 1.05 Ireland 0.32 Portugal 0.50 Greece 0.22 Luxembourg 0.26 Luxembourg 0.06 Greece 0.14
Rest of Europe 20.37 Rest of Europe 24.09
Asia 12.48 Asia 10.56
North America 6.55 North America 4.04 Africa 3.26 Central-/South America 0.82 Central-/South America 3.13 Africa 0.44
Oceania 1.33 Oceania 0.16
Unspecified continent 1.29 Unspecified continent 0.00
3
T heoretical Framework
This section focuses on the theoretical part of the paper. Different trade theories are mentioned in order to give the reader an understanding to why nations trade. The theo-ries discussed will be connected with the regression results in the analytical section.
3.1
Ricardian Model
Several theories discussing why trade develops between countries have been established through history. One of the first models is the Ricardian model, developed by David Ri-cardo. The model discusses how limitation of resources causes different production pos-sibilities in countries. It states that international differences in opportunity costs of pro-ducing a certain good leads to trade. A country focuses on propro-ducing and exporting that good for which it has a comparative advantage in production (Krugman & Obstfeld, 2009). Through trade both countries benefit, since more can be produced through spe-cialization of production. It was through comparative advantage Sweden first gained power in the international market, by having products that other countries had little of (Magnusson, 2010). One of the problems with the Law of Comparative Advantage is that it takes labour as the only factor of production. This is however not completely cor-rect, as can be seen through intra-industry trade.
3.2
H eckscher- O hlin T rade T heory
The Ricardian theory as well as the Heckscher-Ohlin trade theory (H-O) aims to explain the patterns of trade and specialization. The Heckscher-Ohlin trade theory, developed by economists Eli Heckscher and Bertil Ohlin, uses the concept of comparative ad-vantage like the Ricardian model. However, it does not come from total factor produc-tivity but rather from factor abundance as a reason for specialization. Differences in fac-tor endowments are the only source of trade in this model. In addition to the facfac-tor abundance a higher intensity of goods can also lead to a higher comparative advantage (Morrow, 2010).
The H-O model analyses two countries, two factors of production (capital and labour) and two commodities and by using the autarky prices it shows which of the countries is the most efficient in producing a certain good (Jones, 1957). This trade model states that the labour abundant Country A will produce and export labour intensive goods since a comparative advantage over the other country B will be present (Davis, 1995). Drawing a conclusion of the latter, a country will produce and specialize its production on the good in which they hold a comparative advantage; the good that is intensive in the fac-tor the country is abundant in. The H-O trade theory emphasizes the importance of re-sources and technology of production since the comparative advantages come from in-teractions between countries (Davis, 1995).
In the paper “The Heckscker-Ohlin Model in Theory and Practice” (1995) by Edward E. Leamer, he states that the goods traded are in fact factors of production, such as land, labour and capital. By this statement, Davis (1995) discusses that the traded goods are mainly factors of production from an economy which is abundant in a certain factor to an economy where that factor is scarce.
3.3
Inter-Industry T rade
Comparative advantage has a large role in inter-industry trade. Countries engaged in in-ter-industry trade will trade the good that is related to their factor abundance. Falvey (1981) describes how the H-O trade model has taken the fundamentals of inter-industry trade, since it assumes that each economy produces a good in which they have a com-parative advantage in producing. Inter-industry trade can be described as countries trad-ing goods from different sectors, such as export cars and import clothes (one-way trade) and not cars (Volvo) for cars (Audi) as in the case of intra-industry trade (two-way trade). Inter-industry trade is a procedure of specialization rather than trade with differ-entiated products. As stated by Davis (1995, pp. 217-218), “a fall in inter-industry trade is exactly offset by a rise in intra-industry trade” and the opposite.
Countries tend to have different amounts of factors of production as well as a difference in their factor abundance and are therefore assumed to be a part of inter-industry trade (Krugman & Obstfeldt, 2009). By having inter-industry trade, industries can benefit from clustering together and specialize the production on the good which they produce more efficiently. The knowledge spillovers and the history of producing a certain good may lead to inter-industry trade. Inter-industry trade can be seen in Sweden as well, even though Sweden does not have complete one-way trade. Sweden has a large wood and iron sector, the country will therefore export these products and have a low share of import of products related to the industries mentioned (Altunbas et.al, 2013). Inter-industry trade can in small parts exist, as in the case of wood or iron products in Swe-den. However, intra-industry trade is more common since consumers have a large taste of variety as assumed in geographical economics. The trade therefore consists of prod-ucts from the same industry but slightly differentiated (Brakman, Garretsen & van Marrewijk, 2009).
3.4
Intra-Industry T rade
Intra-industry trade is most likely between countries with similar levels of economic development and is common in trade between developed countries. Countries exchange products within the same industry, contradicting the law of comparative advantage men-tioned above. Instead, it is said that it is consumers’ love-of-variety that influences this trade (Brakman, Garretsen & van Marrewijk, 2009). Intra-industry trade is behind the huge increase in industrial trade among OECD countries that has been witnessed in re-cent years (Kadar, 1981). Balassa and Bauwens (1988) test the determinants of intra-industry trade in Europe. Their results showed that intra-intra-industry trade is positively re-lated to average per capita incomes, country size and the existence of a common border, and negatively related to distance, similar to what the gravity model states.
Intra-industry trade is dependent on economies of scale and can therefore be connected with New trade theory, which focuses on increasing returns to scale and trade between countries with similar factor endowments. New trade theory was developed by econo-mists who saw that much of world trade was not following the rule of comparative ad-vantage. Paul Krugman was one of the first economists to discuss this theory, and in his 1992 article “Does the New Trade Theory Require a New Trade Policy?” he mentions economies of scale as one of the main characteristics of New trade theory. As men-tioned by Krugman and Obstfeldt (2009), one cannot predict the pattern of intra-industry trade. What the model shows is that countries similar to each other will trade, but it does not show which country will produce which good. If two trading partners
have similar capital-labour ratios, there will be more intra-industry trade than inter-industry trade, because of the lack of comparative advantages when countries develop similar levels of technology and have a good availability of capital and skilled labour. Krugman and Obstfeldt (2009) explains intra-industry trade to create larger markets and thereby increase the benefits of trade. By producing fewer varieties within a county, production on a larger scale is possible. If product differentiation is important, intra-industry trade creates even more benefits for a country. The more differentiated goods are the larger are the gains from trade. It is also more common in manufactured goods than in raw materials (Krugman & Obstfeldt, 2009).
3.5
Gravity Model
The Gravity model of trade aims to explain how trade between two countries is posi-tively affected by the economic size of the nations, usually explained through GDP. It is negatively affected by the distance between the countries, in other words; less distance between countries makes trade between them more likely (Brakman et al., 2009). The first economist to introduce the gravity model was Tinbergen in his 1962 publication “Shaping the World Economy”. He used the law of gravity from physics to see whether it could provide a reasonable theory of trade between countries, and what variables that determine trade. Tinbergen argued that the more similar countries are in the size of GDP, the more likely is trade to occur. The economic size of a country sets the limit to how much a country can trade; if two large countries trade, more can be exchanged than if a small country traded with a large country, since small countries will demand less (Tinbergen, 1962). The basic model can be seen below:
Ti,j = A((YiYj)/Di,j)
Ti,j = Trade between countries i and j A = Constant
Yi = GDP of country i Yj = GDP of country j
Di,j = Distance between country i and j
The Gravity model allows for several different explanatory variables, and dummy vari-ables can be easily added to control for different country characteristics such as shared borders, colonial history, free trade agreements etc. It is frequently used because of its high statistical explanatory power and empirical achievements (Deardorff, 1998). Tin-bergen (1962) also praises the model and says that it explains much of trade flows be-tween countries. Other economists, such as Andersson and Wincoop (2003) consider the model to not have strong theoretical foundation. They state that bilateral trade between two regions decreases as trade barriers to several regions decrease, since trade is then spread between more regions.
Despite arguments against the model, it is still a good method for measuring trade flows, therefore we consider it useful for this paper and it is the basis for our research. Not only can the Gravity model explain the amount of trade between two nations, it can also show the impact of free trade areas. The model has been developed in several dif-ferent ways, including borders and transport costs, to better explain trade in reality (Pal-grave Dictionary, 2013). Transport costs involve such things as language barriers, tariffs and quotas. When uniting in a free trade area, these costs decrease, which should show
an increasing trade flow between the countries in the free trade area. Changing trade patterns are likely to occur when it becomes inexpensive to trade with certain countries. According to Deardorff (1998), trade between two countries should be as likely as na-tional exchanges when there is a customs union and no trade barriers. This is not con-sistent with reality but one can still see an increase in trade with a reduction in trade bar-riers.
Soloaga and Winters (2000) used the gravity model to analyse the effects of preferential trade agreements (PTA) on intra-block trade. In their paper it was investigated whether countries had faced trade creation or trade diversion after entering trading unions. They found that intra-block trade had not increased as much as assumed after entering a trade union. However, the EU witnessed some trade diversion and saw the most change of the investigated trading blocks. Even though this paper does not deal with trade creation and trade diversion, the model is still applicable, in a modified version, as can be seen in section 4.1.
3.6
Previous Studies
Several studies have been made on the topic of the European Union and its impact on member countries. There are few papers concerning Sweden’s membership and its con-sequence on trade. However, similar studies can still give a good explanation on how a country can be affected by the membership in a trade union. The basic Gravity model or extensions of it is used in roughly all papers regarding EU, its member states and the trade between them. According to the prediction of Kokko (1994), trade liberalization would lead to Sweden increasing its export and competitiveness, since lower tariffs means lower cost of trade.
In his paper, Nilsson (2000) argues that countries wanting to become members of the union have to fulfill the criteria that is set and have to, already before the entry, be inte-grated with the market in the EU. Trade patterns of a country tend to change after the entry into the EU since the trade barriers within the zone decrease. Therefore a change in specialization towards goods that have a demand in the EU will occur (Zaghini 2005). In “Evolution of Trade Patterns in the New EU Member States” by Zaghini (2005), he argues that the structural changes have led to adjustments in both the patterns of specialization as well as trade. Developing countries that are new members of the EU tend to focus on specializing production and export the goods which they have a com-parative advantage in producing as an initial step to catch up with the developed ones. With more “developing countries” entering the union with their low-tech production, the amount of low-tech production existing in Sweden will move to other EU-countries that have a comparative advantage over Sweden in producing these products.
A study conducted by Svenskt Näringsliv described how the EU-membership has af-fected Sweden in different ways. The authors show that the membership has had a posi-tive effect on the GDP-growth, and this is not only obvious in Sweden, but also in many other countries that have joined the union. The single market's purpose is to increase the welfare in Europe by improving competitiveness, however, the authors claim that there is not enough authority in EU to push these issues as far as they need to go. In many cases, countries are permitted to get around the rules. The price reduction has helped several countries, not only Sweden (Gidehag, Öhman & Larsson, 2001).
By taking the best from each member country; policies and economic structure, the goal was and still is, to make EU a stable and good market for its member states. However, it has since the entry been argued that forcing countries with different culture and thoughts together may have negative consequences. This can lead to disputes regarding how a country should be ruled; one structure cannot be applied by everyone. Zaghini (2005) does however believe that the convergence of EU-countries will reduce the risk of a declining common market.
4
Method
This section is an introduction to the empirical part, and here our hypothesis is present-ed, as well as the model description and the explanation of variables. The basic gravity model is modified to fit the purpose of this paper.
This paper tests the hypothesis whether the EU-membership has affected Swedish trade with EU14. The null hypothesis states that the EU-membership did not affect Swedish trade with EU14, against the alternative that Sweden’s trade has been affected by the membership. If the null hypothesis is not rejected, then Sweden’s trade has not been af-fected by the entry into the union. To find out what variables that are important for the regression model we ran several regressions to see whether changes in the data set would have an impact on the regression results.
4.1
Model Description
For our analysis, we will use the Gravity model, discussed in section 3.4 as well as the Grubel-Lloyd index to check for inter- and intra-industry trade.
The Gravity model has been extended in several different ways to incorporate more ef-fects on trade, for example by Flam and Nordström (2006). The model has been adjust-ed to fit the purpose of this paper. In many Gravity models, both GDP and GDP per capita are included. This can however lead to biased results, since an increase in GDP per capita tends to follow and increase in GDP. Therefore GDP per capita is not includ-ed in this paper.
Flam and Nordström (2006) used only exports as their dependent variable, to show the effects on trade between EMU members and non-members. Though we do not deal with EMU, the basic idea of comparing one-way trade flows between countries is interesting. As mentioned in section 3.5, Soloaga and Winters (2000) investigated how intra- and inter industry trade are affected by union blocks. The dependent variable is exports to a specific country, which is affected by opening up of borders (associated with a union), whether both countries are members of the union and the expected imports from non-members of the union.
By modifying Soloaga and Winters’ Gravity model, which can be found in Appendix 1, the equation used for this research is:
LnTij = β0ij + β1LnGDPi + β2LnGDPj + β3LnDij + β4Bij + β5Uij + εij
Where:
Tij = value of exports/imports from country j to country i GDPi = Gross Domestic Product of country i
GDPj = Gross Domestic Product of country j
Dij = Distance between capital cities in country i and j
Bij = Dummy for country i and country j sharing the same borders
To further be able to analyse the collected data, panel data is typically used. It is an in-formative and efficient regression model which is suitable when looking for the dynam-ics of changes, since it gives a more thorough picture than other regression models (Gu-jarati & Porter, 2009). However, an OLS (Ordinary Least Squares) gives similar results, and can be easily used in SPSS, and was therefore applied in this paper.
The Grubel-Lloyd index was used to test for levels of intra-industry trade between Swe-den and the EU14 countries. The closer the value is to one, the larger is the level of in-tra-industry trade. If the number is zero, there is complete inter-industry trade (Brak-man, Garretsen & van Marrewijk, 2009). This index is used to see whether the trading patterns have changed after the Swedish entry into EU. The equation for the Grubel-Lloyd index can be seen below:
GLi=1-[(|EXi-IMi|)/(EXi+IMi)]
Where:
EX= exports for Sweden in specific sector i IM= imports for Sweden in specific sector i
Before 1990 the GDP and trade data for Germany only considers West Germany, after the reunification of West and East Germany in 1990 the data reflects the whole country. West Germany was part of the EU whereas East Germany was not, hence we chose to only consider data for West Germany until 1990. Because of the strong trading bonds between Sweden and Germany, it is important to include Germany in the analysis. Bel-gium and Luxembourg had an economic union until 1999, called BLEU. In the econom-ic union, trade data until 1999 was aggregated between the two countries. Because of this we aggregate the two nations GDP for the years 1980-1998 so as to make it one country. From 1999 the data is separated; one may see it as one country split into two. For the distance variable between the years of 1980 to 1998 we use the distance be-tween Stockholm and Brussels, since Belgium is a larger trading partner than Luxem-bourg (SCB, 2013).
4.2
Explanation of commodity groups and variables
Commodity group description
Trade data used in this paper was collected from UN Comtrade for the years 1980-2009. By using (SITC) Rev 2 it is possible to divide trade data between different commodity groups, to see the effect on these specific sectors from Sweden’s EU-membership. Looking at statistics from SCB one can see that the selected commodity groups for this paper are three important trading sectors (SCB, 2013c). SITC Rev. 2 was chosen since it contains data for the needed commodity groups for all countries in the study.
Exports/Imports, Tij is the dependent variable. It shows the value of exports/imports in
US dollars, from country j to country i, as a function of the GDP of country j, GDP of country i, distance between the countries, with dummy variables for transport costs and the years of EU-membership. The data was collected from UN Comtrade, a well-known and reliable source, though it shows nominal values and does not take inflation into ac-count. Because of our interest in both export and import, the dependant variable does at
times represent exports, and at other times imports. This will be clearly stated in the empirical part, so as to cause no confusion.
Gross Domestic Product, GDPi,j are the first and second independent variables in the
regression model. It shows the nominal value of GDP in country i/j in current US dol-lars. GDP is the value of goods and services produced in a country during a specific pe-riod of time, in this paper, we have used yearly GDP values. Data was collected from United Nations’ statistical database.
Distance, Dij is the third independent variable, and shows the distance between the
eco-nomic centers between Sweden and its trading partners. Capital cities are usually ex-pressed as the economic centres, thus this definition is used in this paper as well. Trucks and trains are mostly used for transporting goods in the EU, the data was therefore col-lected from Google Maps since we can then see the actual distance goods travel be-tween countries rather than air distance.
Border, Bij is the first dummy variable. Countries that share borders have lower
transport costs, which lead to stronger trade relations. Countries that share borders with Sweden are Denmark and Finland, these two countries are coded 1, and countries with-out border to Sweden are coded 0.
Block, Uij is the second dummy variable. It is used to control for the EU membership
and see whether there have been trade effects of it. A customs union is said to increase trade through free trade aggrements and other eliminations of trade barriers. The dummy takes the value of 1 from the year that Sweden became a member of EU (1995-2009 in this paper), and the value of 0 for the earlier years (1980-1994 in this paper).
4.2.1 Expected Results
According to the Gravity model used for this paper, the distance variable is expected to be negative. This result comes from the fact that the longer distance there is between two countries, the less likely they are to trade since it becomes more expensive to ship. The border variable is likely to be positive, since sharing borders indicate shorter and easier transports. Countries sharing borders tend to have similar cultures which may al-so affect trade positively.
The expected sign of the variable GDPi and GDPj is positive, since the higher a coun-try’s GDP is, the more likely is trade to occur. The block variable is expected to be posi-tive, since a trade union decreases tariffs and other trade barriers. As the model predicts; the membership should increase trade within the union, but the trade with countries out-side of the union remains unknown.
5
Empirical Findings
In this section the empirical results are presented and analysed. Descriptive statistics informs the reader of the differences in the data set, and the correlation analysis measures the strength of linear association between the dependent and independent var-iables. The regression findings show the effect on the dependent variable when a change in the independent variables occurs.
5.1
Descriptive Statistics
The first step in our analysis is to check for variations in the variables. In Table 5.1 and 5.2 below, one can see the descriptive statistics for export and import of the three com-modity groups and the variables distance, GDPi and GDPj. The mean value shows the centrality of a set of observations while the standard deviation measures the spread (Aczel & Sounderpandian, 2009).
Table 5.1 below shows the descriptive data for variables involved in export. The mini-mum value of distance represents the distance between Sweden and Finland whereas the largest distance is between Sweden and Portugal. As can be seen in Table 2.1, Finland is Sweden’s third largerst export market within the EU14. Portugal on the other hand, is the country furthest away from Sweden and a small export market for Sweden. This shows that even though countries are members of trade unions, distance still matters. The minimum and maximum values should not be more than three standard deviations away from the mean for there to be a normal distribution of a variable (Aczel & Sounderpandian, 2009).
The estimated values increase over time due to economic growth. Since the values in trade and GDP are nominal, and do not take inflation into account, one needs to be care-ful when drawing conclusions. The great variations in value of exports are due to differ-ences in country size leading to differdiffer-ences in capacity of trade. By looking at the data for exports for each commodity group, one can see that export of chemicals from Swe-den to the EU14 has increased. The lowest value of exports is recognized for Luxem-bourg in 2001 (USD 1.25 million) while the same products exported to Germany in 2007 (USD 2.39 billion) show the highest value. These two countries represent the smallest and largest destinations for Swedish export for all three sectors.
Table 5.1 Descriptive Statistics for Export
N Minimum Maximum Mean Std. Deviation
Chemicalsa 401 1.25 2 394.63 290.38 355.58 Machinerya 401 8.37 5 209.66 1 140.57 1 068.12 Crude materialsb 401 145 781.00 1 915 042.26 272 970.69 305 920.15 Distance (km) 401 481.00 3601.00 1930.21 973.99 GDP EU14a 401 19 904.00 3 623 687.77 586 487.67 713 772.45 GDP Sweden a 401 97 693.15 486 158.67 248 916.05 104 825.95 Valid N (listwise) 401
Due to the Swedish recession in the beginning of the 1990’s, the trade flow of Sweden changed (Magnusson, 2010). Therefore we can see a decrease in exports during that time period and it is observed for all three commodity groups. By looking at the nomi-nal export values for the sectors, the values have increased after the entry into the EU, specifically crude materials saw a great peak in 1995-1996 (UN Comtrade, 2013). The increase in the trade flow can both be explained by the EU-membership and Sweden coming out of the recession as mentioned above.
The lowest values of import for chemicals and machinery come from Greece, whereas the lowest import value for crude materials comes from Luxembourg. The highest value on the other hand, is the import value from Germany to Sweden in all three commodity groups. It is then clear to see that Germany is Sweden’s most important trading partner within the union when looking at the data.
Table 5.2 Descriptive Statistics for Import
N Minimum Maximum Mean Std. Deviation
Chemicalsa 401 0.81 2 935.53 377 467.19 486.58 Machinerya 401 0.33 15 500.94 1.265.89 1 952.86 Crude materials b 401 32.32 795 222.26 89 315.22 117 000.11 Distance (km) 401 481.00 3601.00 1930.21 973.99 GDP EU14a 401 19 904.00 3 623 687.77 586 487.67 713 772.45 GDP Sweden a 401 97 693.15 486 158.67 248 916.05 104 825.95 Valid N (listwise) 401
Notes: aExpressed in millions of US Dollars b Expressed in thousands of US Dollars
5.2
Correlation Analysis
By conducting a correlation analysis we can identify the relationships between the vari-ables and see whether they are correlated or not, and the importance of the relationship between the variables can then be measured. The range of values are between -1 and 1, where the correlation becomes stronger as the values move further toward 1 in either di-rection (Aczel & Sounderpandian, 2009). In Appendix 2, one can find correlations that includes the block dummy; a dummy variable that takes the EU-membership into ac-count. It was observed that the block dummy is significant for all sectors for both ex-ports and imex-ports. By dividing the time period into blocks, before and after the entry in-to the EU, one can see how variables change between the time periods. In Table 5.3 and 5.4 below, the results of the correlations for exports are summarized.
Table 5.3 Correlation between exports in the three sectors and the variables, 1980-1994.
Chemicals Machinery Crude Materials Distance -0.734*** -0.657*** -0.558*** Border GDP EU14 0.392*** 0.691*** 0.264*** 0.715*** 0.125 0.796*** GDP Sweden 0.341*** 0.337*** 0.138 *** Correlation is significant at the 0.01 level (2-tailed).
Table 5.4 Correlation between exports in the three sectors and the variables, 1995-2009.
Chemicals Machinery Crude Materials Distance -0.463*** -0.455*** -0.429***
Border 0.288*** 0.236*** 0.203***
GDP EU14 0.790*** 0.748*** 0.756***
GDP Sweden 0.238*** 0.140** 0.149** *** Correlation is significant at the 0.01 level (2-tailed). ** Correlation is significant at the 0.05 level (2-tailed).
Distance is negatively correlated with exports for all commodity groups at all times. It is clear to see that the correlation analysis shows that the entry into the EU has decreased the effect of distance for Swedish exports within the three sectors.
Having a land border with a trading partner is a significant factor and is positively cor-related with exports for Sweden during these years. For crude materials in 1980-1994 this variable is not significant but becomes so in 1995-2009. The EU-membership seems to have decreased the importance of having a common land border for chemicals and machinery. This may be because of lower tariffs between countries in the union. The decreased importance of these two variables after the entry into the EU follows the theory of the Gravity model.
GDP for EU14 is positively correlated with exports and is high for all sectors. Compar-ing the results for exports before and after the EU entry shown in Table 5.3 and 5.4, one can see that the effect of GDP for the EU14 has become more important, except for crude materials. However, the correlation is still strong and positively related to exports of these products. In Figure 5.1, Sweden’s chemical exports to Denmark can be seen as a function of Denmark’s GDP, the line is steep and the fit is good, which indicates strong correlation between GDP of Denmark and their imports of Swedish chemical products. Denmark is used as an example since the trade relation between Sweden and Denmark has during the chosen time period been good. 1
Figure 5.1 Denmark’s GDP and Swedish chemical exports to Denmark
1A scatter plot was conducted for Spain as well, showing the correlation between chemical exports and
GDP of Spain. Trade between Sweden and Spain is less than between Sweden and Denmark, therefore it is interesting to investigate whether the fit would change dramatically, however, the fit was strong for this country as well.
GDP for Sweden is significant for all commodity groups except for crude materials 1980-1994, however the correlation becomes significant at the 0.05 level, in the analysis of the years 1995-2009. For machinery, the level of significance has decreased to the 0.05 level, hence the correlation is weaker than before. The correlation for chemicals has also decreased but is still significant at the 0.01 level. These observations suggest that Sweden’s GDP has become less important for exports of chemicals and machinery after the EU-membership.
Table 5.5 Correlation between imports in the three sectors and the variables, 1980-1994.
Chemicals Machinery Crude Materials Distance -0.697*** -0.664*** -0.746*** Border GDP EU14 0.218*** 0.620*** 0.243*** 0.656*** 0.410*** 0.488*** GDP Sweden 0.267*** 0.226*** 0.130 *** Correlation is significant at the 0.01 level (2-tailed).
Table 5.6 Correlation between imports in the three sectors and the variables, 1995-2009.
Chemicals Machinery Crude Materials Distance -0.597*** -0.593*** -0.489***
Border 0.223*** 0.170** 0.368***
GDP EU14 0.645*** 0.628*** 0.571***
GDP Sweden 0.172** 0.155** 0.171**
*** Correlation is significant at the 0.01 level (2-tailed). ** Correlation is significant at the 0.05 level (2-tailed).
The impact on imports of the different variables can be seen in Table 5.5 and 5.6 above. As for exports, the correlation between distance and imports is negative, and the im-portance of this variable has decreased with the EU-membership, meaning that within the union, distance has less effect than for the years before the entry. Land borders are less correlated with imports for the years 1995-2009 than for years 1980-1994, except for chemicals where the correlation has increased. This may be because of harder con-trols and regulations within the EU on how to transport dangerous goods, such as chem-icals, leading Sweden to focus on importing from neighbouring countries to decrease the distance of transport. For the two other investigated commodity groups having common borders still affect imports positively, but is not as important as before.
For the GDP EU14 variable we can see that it is significant and positively correlated. It has decreased somewhat for machinery but it is still highly correlated after the entry. The correlation between Sweden’s GDP and Swedish imports has decreased with the entry for both chemicals and machinery, and the level of significance has decreased to a 0.05 level of significance. Crude materials on the other hand, witnessed an increased correlation with imports and have become significant at the 0.05 level.
5.3
Statistical Errors
To check for errors in the data we ran a test for autocorrelation. The Durbin-Watson test statistic showed some variation between the two time periods and between commodity groups. The Durbin-Watson test statistic takes on the value of 2 when there is no auto-correlation and 0 when there is positive autoauto-correlation. The closer the test statistic is to
4, the more negative autocorrelation there is (Gujarati & Porter, 2009). With the range of 1.55 for machinery to 2.55 for crude materials in exports, it is possible to see that the sectors are affected differently by the variables. For imports it ranges between 1.92 for crude materials to 2.73 for chemicals. Even though there is some autocorrelation in cer-tain cases, the numbers are small enough to assume that the problem of autocorrelation is small.
By looking at scatter plots, one can see if there is heteroscedasticity in the data. By us-ing the log terms of the variables we found that there was little heteroscedasticity prob-lems in either dataset. Because of this we can use OLS estimation rather than a Weighted Least Squares (WLS) estimation of the regression, which could be used if heteroscedasticity was present (Gujarati & Porter, 2009).
By using the variance inflation factor (VIF) we can detect whether there is multicolline-arity in the data sets. If VIF is higher than 10, multicollinemulticolline-arity exists, and this occurs when R2 is above 0.9 (Gujarati & Porter, 2009). In exports in the first time period, high R2 values can be seen in the chemical and machinery sectors, leading us to assume mul-ticollinearity in these sectors. We have conducted regressions excluding Germany the years before 1991, and in these regressions, the R2 value decreases, as can be seen in Appendix 3, indicating that including Germany before 1991 may lead to overestimation. The reason for excluding Germany before 1991 is that we believe that the reunification may cause errors in the regression since it changes the data. When excluding Germany before 1991 some changes in significance levels can be seen, but in general the differ-ences in results are small and therefore we decided to continue using regressions includ-ing Germany for all years.
5.4
Regression Analysis
In this part of the empirical findings the regression results are presented and the hypoth-esis is tested. Here we can see how the different variables’ importance for the regression have changed with the EU-membership.
Exports
For all export regressions, the observed F-values are all higher than the critical F-value of 3.02 (df 5,∞) at the 0.01 level of significance. Thus we reject the null hypothesis, in-dicating that the EU-membership has had an effect on Swedish exports. All F-values are lower in the second time period than in the first one, which indicates that there have been changes after the entry, nevertheless they are still significant. R2 follows the same pattern, with decreased values in the second time period, even though the regression line fits the data well, the fit is not as strong as before the membership. The export variables’ values for all three commodity groups have changed, comparing before and after the en-try into the union. The regression results do not show changes in trade patterns but ra-ther how the variables have changed in their effect on Swedish exports.
Chemicals and related products
In Table 5.7 below, one can see that all variables are significant and affect the export of chemicals, even though the level of significance differs between variables. As suggested by the model, distance affects trade negatively, but after entering a trading block, such as the EU, distance becomes less influential for trading between nations within the block. The regression results of this paper are identical with this prediction. The large
change in the distance variable indicates that Sweden is more likely to export to coun-tries further away after the EU-membership. One variable that has seen an increased ef-fect for the export of chemicals is the border dummy. After the entry into the union, sharing borders became more important for Sweden, as can be seen in below table. The findings tell us that Sweden is more likely to export chemicals to Denmark and Finland; its border countries, rather than other countries within EU14. Increased competition on the EU single market may have caused Swedish comparative advantage to decrease and led to Sweden focusing on exporting chemicals to neighbouring countries. One may see this as a contradiction to the findings of the distance variable. However, it may also be seen as a general increase in Swedish export of chemicals to the EU14 countries.
The Gravity model assumes GDP to have positive effects on trade, which is consistent with the results for chemical exports. The importance of GDP for EU14 has increased with the Swedish EU membership, and it leads us to believe that the larger a country’s GDP is, the more chemicals will be exported from Sweden to that country. A reason for the increased value in the GDP EU14 variable is that chemicals has become an increas-ingly important sector for Swedish exports, making it more dependent on GDP of trad-ing partners (UN Comtrade, 2013). GDP for Sweden is assumed to be positively related to exports and this is consistent with the results from the regressions. The relevance of this variable has however decreased after the entry, suggesting that Sweden’s GDP has a lower influence over exports of chemicals to the EU14 member states. Because of better trading bonds within the common market, the effect of Swedish GDP decreases.
Table 5.7 Regression results for chemical exports
1980-1994 1995-2009 Constant -3.538** (1.681) -15.664*** (3.512) Distance -1.236*** 0.070) -0.295*** (0.093) Border 0.427*** (0.120) 1.788*** (0.158) GDP EU14 0.654*** (0.024) 1.065*** (0.030) GDPSweden 0.534*** (0.067) 0.314** (0.133) F-value 820.012 453.295 R2 0.945 R2: 0.900 N 194 205
*** Significant at the 0.01 level (2-tailed). **Significant at the 0.05 level (2-tailed). Note: Numbers in brackets are the standard errors.
Machinery and transport equipment
Table 5.8 shows the regression results for the machinery sector, where one can see that the distance variable negatively affects exports. Following the theory of the Gravity model; it affects exports less after the EU-membership. Sharing land borders negatively affected exports from Sweden before the entry into the EU. It can be seen that the level of significance is low, which may indicate that the border variable did not affect exports
as much as other variables in the model. For the years after the EU entry, sharing bor-ders positively affected exports of machinery. Products in this sector tend to be large and more difficult to transport, which may be a reason for exporting more to common border countries. When looking at the two mentioned variables, we can see that this sec-tor follows the same pattern as for the chemical secsec-tor. Even though low transport costs make it more affordable to export to countries further away, Sweden tend to export much to its border countries.
The value of the variable for GDP for EU14 has increased, which means that GDP EU14 has become a more important variable for the regression since the Swedish entry into the EU. These goods are investment goods and are dependent on GDP. A larger GDP makes it possible for a trading partner to import more of Swedish goods; the larger a trading partners’ GDP is, the larger is the probability that they will purchase goods, as predicted by the Gravity model. With the four freedoms mentioned in section 2, factors of production can easily move across borders, creating similar possibilities for member states to produce differentiated products through economies of scale. Within the EU, several countries have experienced a convergence in GDP, which in theory and reality leads to intra-industry trade. In the second time period it can be seen that Swedish GDP is no longer significant and we can therefore not draw any conclusions on the effect of Swedish GDP.
Table 5.8 Regression results for machinery exports
1980-1994 1995-2009 Constant 2.609 (2.234) 4.594 (4.720) Distance -1.274*** (0.093) -0.532*** (0.125) Border -0.314* (0.160) 1.044*** (0.213) GDP EU14 0.525*** (0.032) 0.905*** (0.041) GDP Sweden 0.580*** (0.090) -0.165 (0.179) F-value 330.842 185.005 R2 0.874 0.786 N 194 205
*** Significant at the 0.01 level (2-tailed). ** Significant at the 0.05 level (2-tailed) *Significant at the 0.1 level (2-tailed). Note: Numbers in brackets are the standard errors.
Crude materials, inedible, except fuels
In Table 5.9 one can see that as consistent with the previous commodity groups, the dis-tance variable follows the same patterns as for sectors mentioned above. Disdis-tance be-comes less important for Swedish exports in the second time period, which can be ex-plained by decreases in tariffs and quotas. Because of decreased transport costs, trade between nations is more likely to occur. The border dummy has gone from being nega-tively related to posinega-tively related to exports. EU’s common laws and regulations help explain the large change from a negative to a positive value; even though neighbouring
countries tend to have similar principles on trade, having the same laws makes trade even more likely. Connecting the results with theory, we see that some products within this sector are important for Swedish exports, such as wood and iron ore. Because of Sweden’s history of producing these goods and having abundance in it, the comparative advantage is still strong. By opening up to free trade, countries are assumed to special-ize, and produce the goods in which they have a comparative advantage in, as suggested by both the Ricardian model and the Heckscher-Ohlin model. However, the opening up of borders may affect countries comparative advantage negatively as well, as sectors that once had a strong comparative advantage now face more competition when markets open up and more countries enter the single market of the EU. Factor abundance in some sectors also create comparative advantage, which can be seen in Sweden for sev-eral industries within this sector.
GDP for EU14 positively affects exports of crude materials, even more so after Sweden joined the EU. An increase in EU14 GDP will increase exports in this sector more than in the two previous sectors, as can be seen through the high value of the variable. Crude materials are intermediate goods, marked by inter-industry trade. A change in GDP will have a higher impact on trade, leading to the high value of the variable. An increase in GDP for Sweden negatively affects exports of crude materials in the first time period, whereas after Sweden’s entry into the EU, the variable becomes insignificant and it is therefore not possible to draw conclusions on the effect of the variable.
Table 5.9 Regression results for crude materials exports
1980-1994 1995-2009 Constant 16.062*** (2.683) -2.178 (5.966) Distance -1.246*** (0.112) -0.645*** (0.158) Border -0.356* (0.192) 1.123*** (0.269) GDP EU14 0.822*** (0.038) 1.097*** (0.051) GDP Sweden -0.369** (0.108) -0.142 (0.226) F-value 326.301 167.833 R2 0.873 0.770 N 194 205
***Significant at the 0.01 level (2-tailed). ** Significant at the 0.05 level (2-tailed) *Significant at the 0.1 level (2-tailed). Note: Numbers in brackets are the standard errors.
Imports
The critical value is 3.02 (df 5,∞) at the 0.01 level of significance. Our observed F-values are all higher than the critical F-value, leading us to reject the null hypothesis and conclude that the EU-membership has affected Swedish imports from the EU14 mem-ber states. The F-values decrease in the second time period for all commodity groups. The R2 value shows a good fit for the three sectors, but also show a decreasing value in the second time period. When looking at the variables, one cannot see a specific pattern
between the three commodity groups, suggesting that they are sensitive to different changes in the variables. Import regressions show slightly different results compared to exports, indicating that import and export are affected by different factors of trade.
Chemicals and related products
In Table 5.10 the import results for chemicals is shown. The decrease in transport costs has made the distance variable less influential, but it is still a highly significant variable. Even though we can see that distance affects imports of chemicals less then before, the negative effect on imports is still high. This suggests that there is a high probability that Sweden will import from nearby countries rather than countries further away in the un-ion. This is due to the fact that many chemicals are high-risk goods, making them more costly to transport. In below table one can see that the border dummy is negative; lead-ing us to draw the conclusion that Sweden imports chemicals from EU14 countries that do not share land borders with Sweden, contradicting the results of the effect of dis-tance, since that variable indicates that Sweden is more likely to import from neighbour-ing countries. These results indicate that Sweden may import chemicals from countries that are not border countries but still rather close to Sweden. Nations other than Finland and Denmark may be larger producers of chemicals which leads Sweden to import from these countries.
One can see that GDP for EU14 is positively related to Swedish imports of the product, and the variable has increased after the EU entry in 1995. Larger economies can devel-op larger industries, and as explained through economies of scale, the cost of producing decreases when a country produces a larger amount of one variety rather than several varieties. This makes the goods less expensive and Sweden can therefore import more of the varieties produced in other countries. Swedish GDP has become insignificant in the second time period, while it was highly significant before the EU membership. One needs to take this insignificance into consideration, joining the EU, and experiencing an integration in the larger market, seems to have made Swedish GDP less central for im-ports.
Table 5.10 Regression results for chemical imports
1980-1994 1995-2009 Constant 8.850** (3.449) 3.688 (6.529) Distance -3.055*** (0.143) -2.280*** (0.172) Border -2.488*** (0.246) -0.949*** (0.294) GDP EU14 0.423*** (0.049) 0.859*** (0.056) GDP Sweden 0.830*** (0.138) 0.358 (0.248) F-value 312.030 176.841 R2 0.868 0.779 N 194 205
*** Significant at the 0.01 level (2-tailed). ** Significant at the 0.05 level (2-tailed). Note: Numbers in brackets are the standard errors.
Machinery and transport equipment
Table 5.11 shows regression results for imports of goods in the machinery sector. The distance variable has seen a small change but is still negatively related to imports of machinery. Borders show a negative impact on imports in both time periods, indicating that Sweden is more likely to import machinery from non-border EU14 countries. The change following the EU-membership is small, as is the change in the distance variable. This indicates that the variables are approximately as important for imports after the EU-membership as they were before. Economies of scale and differentiation of products lead Sweden to import varieties of products not produced at home, from other EU14 countries. Sharing land borders should have a positive impact on trade since it indicates that transport is shorter and relatively inexpensive. Common borders may also indicate similar cultures and language, making it easier to trade. A negative sign means that Sweden is less likely to trade with its border countries; in this paper, Denmark and Fin-land. The results may arise because of few neighbouring countries to Sweden being part of the EU14, a more even distribution between neighbouring and other countries in the data set may give different results. Another explanation may be that large trading part-ners such as Germany and the UK are large producers of machinery and transport equipment. This leads Sweden to import goods within this sector from Germany and UK rather than from neighbouring countries.
GDP of EU14 is positively related to imports, but the value has surprisingly decreased in the second time period, and clearly does not follow the same pattern as for the other commodity groups. The EU-membership may have caused other factors to become more important for imports; such as political influences, which can be one explanation to the decreased importance of the GDP EU14 variable. The variable for Swedish GDP in the second time period is insignificant, hence we cannot draw a conclusion on the ef-fect of Swedish GDP for imports in this sector.
Table 5.11 Regression results for machinery imports
1980-1994 1995-2009 Constant 7.880** (4.971) 11.975** (6.071) Distance -2.698*** (0.207) -2.478*** (0.160) Border -1.359*** (0.355) -1.659*** (0.274) GDP EU14 0.792*** (0.071) 0.717*** (0.052) GDP Sweden 0.426** (0.200) 0.298 (0.230) F-value 194.818 178.241 R2 0.804 0.780 N 194 205 *** Significant at the 0.01 level (2-tailed). ** Significant at the 0.05 level (2-tailed). Note: Numbers in brackets are standard errors.
Crude materials, inedible, except fuels
The regression results for crude materials can be seen in Table 5.12 below. As in previ-ous cases the distance variable becomes less influential in the second time period, ex-plained through decreases in transport costs. The border dummy has gone from being insignificant in the first time period, to highly significant in the second time period. The positive effect in the second time period indicates that even though reduced transport costs have made it possible for Sweden to import crude materials from countries further away, it is still very likely that they will trade with neighbouring countries. Goods with-in this commodity group tend to be traded with-in large quantities, causwith-ing high transport costs, hence it is easier and less costly to import from neighbouring countries. The re-sults of the distance and the border variables seem to contradict each other, however it can also be interpreted as increased Swedish imports in general. Some crude materials may be imported from Spain and Portugal after joining the EU, while other products are imported from neigbouring countries.
GDP for EU14 has a positive influence on imports of crude materials and the variable increased remarkably in the second time period. We conclude that after joining the EU, GDP for EU14 plays a larger role for imports of this sector than it did before. As in the case of chemicals, a larger GDP indicates that a country can produce more goods for a lower cost, making it possible for Sweden to import more of these products in the crude materials sector. Crude materials is used as intermediate goods in production of other goods. These goods tend to be investment goods which are dependent on GDP, leading crude materials to be affected by the same factors as investment goods. GDP for Swe-den is insignificant for both time periods and can therefore not be analyzed further.
Table 5.12 Regression results for crude materials imports
1980-1994 1995-2009 Constant 18.084*** (4.350) -14.834 (9.100) Distance -1.886*** (0.181) -0.484** (0.240) Border -0.286 (0.311) 2.402*** (0.410) GDP EU14 0.472*** (0.062) 1.086*** (0.078) GDP Sweden 0.034 (0.175) 0.244 (0.345) F-value 125.824 84.329 R2 0.726 0.627 N 194 205
*** Significant at the 0.01 level (2-tailed). ** Significant at the 0.05 level (2-tailed). Note: Numbers in brackets are the standard errors.
