University of Groningen Faculty of economics and business
Master thesis international economics and business
How will Brexit affect GDP and jobs in the UK and the EU countries?
Name student: Erick Vinkes Student number: s2773856
Student mail: e.vinkes@student.rug.nl
Date: 07-07-2017
Supervisor: de Vries, G.J.
2 Abstract
What will happen economically to the UK and the EU, once Brexit has happened. Using the world input-output database, estimations can be made about the effects of Brexit on the GDP
and jobs within the UK and the EU countries. The biggest loss in shares of GDP will be for the UK. Other countries with high GDP losses are Ireland, the Netherlands and Belgium. The
job loss will be highest in the UK, followed by Germany and France.
Keywords: GDP, Jobs, Brexit.
3
Table of contents
Table of contents ... 3
1. Introduction ... 4
2. Literature review ... 5
2.1 General trade theory ... 5
2.2 Trade organisations ... 8
2.3 Non-tariff barriers ... 11
2.4 Global value chain ... 12
2.5 Brexit ... 13
2.6 Hypotheses ... 15
3 Data ... 16
4 Methodology ... 16
4.1 Method ... 16
4.2 Tariffs ... 19
4.3 Trade and job elasticity ... 20
5 Results ... 23
5.1 Gross domestic product ... 23
5.2 Jobs ... 26
6 Concluding remarks ... 29
Reference list ... 32
Appendix A: tariff profile of the EU ... 34
Appendix B: products groups divided over WIOD sectors ... 35
4
1. Introduction
Never before in the history of the European Union (EU) has a country left its organisation.
But after a referendum in the United Kingdom (UK) the people voted to leave the EU, a so- called Brexit. Since this has never happened before media and politicians all draw their own conclusions. The ‘leavers’ are convinced that leaving the EU will strengthen the UK economy.
One of the arguments that the leavers make is that the UK can, once again, make its own trade deals. The leavers also believe that a Brexit will create jobs for lower educated workers, as it will reduce the effects of globalisation. Lastly, when the UK leaves the EU it also cuts any funding going to Brussels, this money can be invested in infrastructure and schooling which should improve economic growth. However the ‘remainers’ are convinced that a Brexit will be an economic disaster, they wish for the UK to remain a part of the EU. Being part of the EU grants access to the single market, which means the EU is one territory without any internal borders or other regulatory obstacles to the free movement of goods and services.
Remainers state that if the UK is to be cut off from the single market, the newly imposed tariffs will hurt exports to the EU and imports from the EU which will result in job losses and reduced GDP. The leavers and remainers are still hotly debating the effects of Brexit on the UK.
A rapidly expanding literature explores how Brexit may affect the UK. An influential paper is Ottaviano et al. (2014) who wrote about the costs and benefits of leaving the EU. Ottaviano et al. (2014) looked at the effects of tariff barriers and non-tariff barriers that will be imposed after Brexit. Two scenarios are explored. In the optimistic scenario there will be no tariffs, but non-tariff barriers will increase. In the pessimistic scenario, tariffs will rise to most favourable nation status, while non-tariff barriers rise even more. The authors conclude that over the course of ten years the UK’s GDP will shrink with 1.23% in the optimistic scenario and with 3.09% in the pessimistic scenario. This paper focuses on the tariffs and non-tariff barriers that arise between the EU and the UK. Its results are in line with what remainers predicted.
This thesis will follow Ottaviano et al. (2014). However, the main focus of this paper is the upcoming tariffs. There will be two scenarios: the most favourable nation (MFN) scenario and the better deal scenario. In the MFN scenario tariffs will be raise to a lowest possible tariff under regulation of the World Trade Organisation (WTO). In the better deal scenario tariffs will be lower because the UK and the EU negotiated a better deal. While Ottaviano et al.
(2014) focussed on the GDP of the UK, this paper will look at the effects of Brexit on GDP
and jobs of both the UK and the member countries of the EU. The results of this paper show a
5
big GDP loss for the UK. Depending on the tariffs the UK loses 8.3% or 4.1% of their GDP.
The job losses will be a staggering 1,078,400 or 539,200 jobs. Of the EU countries: Brexit has the biggest impact on Ireland, Malta, the Netherlands and Belgium in terms of decline in GDP. While the impact of Brexit is largest in Germany, France, the Netherlands and Poland in terms of jobs lost due to reduced international trade.
The paper will start with the literature review, which includes a discussion of general trade theory, trade organisations, global value chains, Brexit and the hypotheses. In section 3 the dataset is described. In section 4 the methodology will be explained, which consist of the method used, the tariffs and elasticity’s. After that the results are discussed, followed by the concluding remarks.
2. Literature review
For the first time in the history of the EU a country is actively planning to leave the union.
Whether the UK will leave the EU is still being debated, but in 2016 on the 23 rd of June 51.9% of the voters chose in favour of leave. The Prime Minister Cameron retired and so the new prime minister May is in charge of making the UK leave the EU. While the politics surrounding the Brexit are exciting, this paper will focus on the costs of Brexit. Specifically what will happen to the trade between the UK and the EU. And how this will affect the gross domestic product (GDP) and jobs of the UK and the EU. In this literature review the
following points will be discussed: general trade theory, the effects of joining trade
organisations, non-tariff barriers and the global value chain. Then there will be a closer look at the UK and Brexit. Lastly the hypotheses of this paper will be formulated.
2.1 General trade theory
Before the 19 th century most of the countries in the world were mercantilist. This meant that countries were trying to keep production within their own borders. This was accomplished by imposing high tariffs on imported finished products and high taxes on the export of raw products, meanwhile there were low tariffs on the import of raw products and low taxes on the export of finished products. Colonies played a role in this, being the suppliers of raw materials from the new world. Usually these colonies did not make any finished products themselves.
The colonies were also not allowed to trade with any nation except for the owner of the
colony. Mercantilism was based around maximizing export to gain as much gold as possible
(The Economist, 2013). But this mercantilist view would change.
6
In 1776 Adam Smith published the book ‘an inquiry into the nature and causes of the wealth of nations’, in which Smith tried to counter mercantilism. Smith argued that a country needs to import what it cannot produce cheaply and export what it can produce cheaply. This directly opposes the ideas of mercantilism where countries try to keep as much production within its borders. However, in the theory of absolute advantages it may occur that between countries A and B, country A has no absolute advantage over country B and country B has multiple absolute advantages over country A, if this is the case no trade will happen. In 1817 Ricardo published the book ‘on the principles of political economy and taxation’ and in this book he would prove that countries without absolute advantages can still trade. He called this theory comparative advantage. In his world famous example between Portugal and England he proves that England can still trade with Portugal, even if Portugal is more efficient at producing every available product. Countries have limited labour and should always make the product it can make most cheaply compared to the trading partner, this would result in
countries specializing in the production of one good. Specialisation would make sure the maximum number of a product would become available on the market, which could be traded between countries to increase welfare. Although countries did not truly specialise in one product, his view of free trade has been taken over by many economists. Eventually the theory would be picked up by Eli Heckscher and Bertil Ohlin, who would write the Heckscher-Ohlin model. In Ricardo’s comparative advantage theory, Ricardo looks only at the amount of labour within the trading countries, while Heckscher-Ohlin will look at labour and capital available within the trading countries. In the Heckscher-Ohlin model a country that has an abundance in capital will make capital intensive products to export, while importing labour intensive products and vice versa for countries that are abundant on labour. The theories of Smith, Ricardo, Heckscher and Ohlin are describing a world where consumption and production can be far away from one another through trade, this is called the first great unbundling (Baldwin 2006). For the most part of history, cities and villages in the world had consumption and production happening within a few miles of each other. Trade helped countries to export goods that they could make cheaply and import goods that are cheaply made elsewhere, in other words production and consumption were disentangled. The first unbundling went into overdrive once the industrial revolution started. Through steam technology, transporting goods over large distances became easier and cheaper than before.
This meant that most goods and even food to a certain extent could be produced at one side of
the country and consumed on the other side. As a result companies could locate themselves
wherever they wanted to, so they located themselves in cities where a large number of
7
labourers were living. Because of the growing demand for labour by companies in cities, many labourers left the rural areas to go to the cities where they could find jobs, this is known as urbanisation. Further improvements in transportation technology made it easier to transport goods over increasingly larger distances.
The important take away point from these theories is that free trade increases welfare. Welfare increases because free trade makes it possible to buy the relative cheapest products from other countries. Here is a fictional example to illustrate this. If there is no free trade Norwegians may buy very expensive French wines (because of high tariffs) or they are not be able to buy French wines at all. Making wine in Norway is not easy because of the climate and thus very expensive. Free trade with France would increase the welfare for the Norwegians, because they are now able to buy cheap(er) French wines. It is however important to note that an increase in welfare does not equal an increase in GDP. Frankel and Romer (1999) wrote a paper on the effects of trade in 1985 on the living standards. To calculate the effects of trade on living standards the authors did not only use trade ratio’s (the amount of trade to GDP), but also added other variables like distance to trade partners, landlocked, population and area.
They find that trade does increase living standards, an increase of one percent in the trade ratio increases standard of living with a half percent. Irwin and Terviö (2001) did this research again for a much larger time frame. The time frame they used include the pre-world war I era, the inter war era, the great depression and the post war era. They come to the same
conclusion, a country that trades more, has a higher standard of living. Baldwin, Mutti and Richardson (1980) calculated the effect of a 50% tariff reduction on trade, employment and economic welfare. In this paper the authors looked not only to gains from this tariff reduction, but focussed on adjustment costs. Adjustment costs are the costs companies (and society) make when capital and labour moves to different industries. With a tariff reduction some industries will close down, while others will grow rapidly, this means that capital and labour need to move (adjust). This results in almost no changes in employment or the absolute number of traded goods, but the welfare gains are enormous, namely 100 million dollars on a yearly basis. The adjustments costs are twenty times as small as the welfare gains. This paper shows how important free(er) trade is for GDP growth.
However, there are some academics that are critical about the strength of the relation between
lower tariffs and economic growth. Rodrik and Rodriguez (1997) wanted to examine if
countries with lower barriers to trade would grow faster after country-specific characteristics
have been controlled for. By examining multiple papers about the relation between trade
8
barriers and economic growth, they conclude that most literature is plagued with
misspecifications or suffer from use of bad measurements of openness. What the authors mean by this is that sometimes proxies for openness are used that have an independent effect on economic growth (for example property rights). Another example is that trade weighted tariff averages are biased and that there are no studies (at that time) that provide alternatives.
Then there are other problems that are not addressed in the literature, for example whether tariffs work different in low versus high income countries or small versus big countries. The last point the authors make is that more micro data needs to be used, early papers about this subject show that firm are self selecting. This means that more productive firms will export, not that export increases production. The authors end their paper by stating that they don’t think that trade restrictions will increase economic growth, but that they want to make clear that current papers may overestimate the effect of tariff reductions on economic growth. Six years later Melitz (2003) published a paper about self-selecting firms. In this paper Melitz shows that firms are heterogeneous, each with their own level of productivity. Only the most productive firms are able to export their products to other countries. Exporting products will lead to reallocation of resources from less productive firms to more productive firms. This reallocation of resources will result in a welfare gain.
2.2 Trade organisations
Now that it is clear how trade affects welfare and GDP, it is important to know how trade can
increase or decrease. One way to increase trade is by joining a trade organisation, because
trade organisations are lowering tariffs between member countries. After the Second World
War high-income countries decided to work together to lower tariffs. A large agreement was
made between these countries, this agreement is called the general agreement on tariffs and
trade (GATT) and its goal was to reduce tariffs and increase trade between signatories. This
agreement would lay the groundwork for the World Trade Organization (WTO). The goal of
the WTO is to reduce tariffs and increase trade between its members. But what are the effects
of joining a trade organisation and are the goals of trade creation met when a country joins a
trade organisation? Rivers and Tomz (2007) wrote a paper on the effects of being a (standing
or formal) member of the WTO and the effects of joining multiple trade organisations or sign
trade agreements with non-WTO members. The authors looked at two factors: institutional
standing and institutional embeddedness. Institutional standing means the degree to which a
country, territory, or nongovernmental actor has rights and obligations arising from the
institution. WTO regulations do not only count for formal members, but also for standing
9
members such as (ex)colonies. Standing members have rights and little or no obligations.
Formal members have rights and obligations, but are allowed to not accept certain obligations and rights, this makes their standing different from other formal members. Institutional standing is important when calculating the effects of trade organisations, as some formal members may have lower institutional standing (lower in this case means less rights and obligations) than other formal members or even standing members. Institutional
embeddedness is influenced by trade organisation memberships. When a country belongs to multiple trade organisations (or other international organisations such as the EU) institutional embeddedness will decrease. The authors used the gravity model to calculate the effects of being member of a trade organisation. The results show that trade increased between formal and standing members and that trade grows even faster when a member is also a participant in other trade organisations. Trade also grows between (formal and standing) members of the WTO and non-participants of the WTO. The authors conclude that being a member of the WTO will increase trade. Being a participant in multiple international organisations will complement the trade increasing effects making trade grow even faster. Breuss (2001) wrote a paper about EU enlargement at the start of this millennium. This paper predicts the effects that EU enlargement will have on the GDP of the newcomers and the fifteen members of the EU. In the early 2000’s negotiations were held to increase the EU from fifteen countries to twenty-seven countries. Such an enlargement of the EU will have consequences for all the countries involved: there are trade effects, single market effects, factor movement effects and costs of enlargement. The author used the Oxford Economic Forecasting World Macro
Economic model to calculate the effects for every country. The trade effects are asymmetrical, 70% of all exports from the newcomers went to the EU while only 4% of all the exports from the EU went to the newcomers. Through tariff reduction the EU would grow an extra 0.05%
GDP over 5 years, while each of the twelve new comers grow an extra 1% to 5% GDP. The single market effects are composed of productivity changes and price level changes. The small countries in the EU get a 0.5% growth of their GDP through productivity changes, while large EU countries gain half of that. The new comers will have an additional 1% GDP growth. The changes in prices will make the EU’s GDP grow by 0.5%, while the newcomers GDP will increase by 1%. The factor movement effects consist of FDI flows and migration.
FDI will flow from the EU to the newcomers, leading to a small GDP decrease in the EU of
0.2% and an increase of GDP of 0.75% in the countries of the newcomers. Migration flows
shows the opposite, a growth of 0.2% GDP in the EU and a decline of 0.1% in the countries
of the newcomers. Lastly, there are the costs of enlargement. Every EU member has to pay
10
0.1% of their GDP to help the newcomers prepare for EU membership. This money will be spend to improve infrastructure and reduce corruption in the newcomers. This will lead to a GDP growth of 2% for the newcomers. The paper clearly shows that joining the EU has many economic advantages. It is also important to note that while the EU can be considered a trade organisation, it is more than that. The EU is a political union between European nations, with its own leadership that has the power to overrule laws made by governments on a national level. It also helps with developing countries through financing infrastructure and fighting corruption. It even helps member states that are in financial trouble. Carrere (2004) wrote a paper about the effects of regional trade agreements (RTA). This paper used data from 130 countries over the period 1962 to 1996. The methodology was based on the gravity model with panel data and ordinary least squares (OLS). The results show that trade tends to grow significantly between signatories of the RTA. However, trade with the rest of the world seems to decrease. The signing of regional trade agreements seems to create trade diversion. Aitken (1973) wrote a paper about two different trade organisations in one continent between 1951 to 1967. In this paper Aitken (1973) explores how the European Economic Community (EEC) and the European Free Trade Association (EFTA) behave individually and together. Through use of a cross sectional trade model the author tries to calculate the trade creation and trade diversion, within the trade organisation and between the trade organisations. Both
organisations show a growth in gross trade creation (GTC), but EEC’s gross trade creation growth is higher. The EFTA shows a lot of trade diversion with regards towards the EEC, this means that countries of the EFTA choose to trade with a country within the trade organisation after the most efficient country joined the EEC and thus had to take over the EEC’s tariff policies. Interestingly, the same did not happen the other way around. Countries that joined EEC joined a custom union, which meant that tariff for the outside world were harmonized between members. This led to a reduction of tariffs and to trade creation in countries outside the EEC and within the EFTA. This effect of ‘external’ trade creation held on until 1964, after that tariff changes created trade diversion from EFTA to EEC countries.
The effects of leaving a trade organisation are not well known, as leaving a trade organisation happens rarely. However, Vandenbussche et al. (2017) wrote a paper about the effects of the trade policies Trump proposed in his race to become president of the United States (US). This paper shows the effects of protectionism on EU trading partners and while this is not the same as the US leaving a trade organisation, it does show the effects of cancelling trade agreements.
Vandenbussche et al. (2017) is looking at direct exports, but also at indirect exports towards
11
the EU. The focus is only on goods, not services, because services will not get increased tariffs under Trump. However, services will still get hit as indirect exports (some services are needed to produce goods, which can be exported). It is clear that the American and European economies are highly interconnected. There are two scenarios: a scenario in which tariffs are increased to 5% and a scenario where tariffs are increased to 15%. In the 5% tariff scenario job losses are 50,000 for the EU. The GDP of EU countries shrinks with 0.1%. Exports will decrease by 5.3%. In the 15% tariff scenario the job losses are 240,000 for the EU. The GDP of EU countries shrinks with 0.4% and exports will decrease with 24%. This paper shows that protectionism hurts the EU and while the effects of this protectionism are not quantified for the US, it can be assumed that this protectionism hurts the American economy as well.
Protectionism can also lead to so-called trade wars in which nations start increasing tariffs on each other, which can result in a loss of welfare.
2.3 Non-tariff barriers
With all the work done to reduce tariff barriers, non-tariff barriers (NTB) start to gather more attention. As discussed previously many countries and organisations in the world are actively trying to reduce tariffs and with great success. Novy (2014) shows that trade costs
(transportation costs and tariffs) are declining. Through the use of microdata and the gravity model Novy (2014) calculated that the average cost of trade declined by 40% over the time period 1970 to 2000.
Some examples of NTB are: competition laws, health and safety regulations, technical
standards and customs clearance procedures. In the paper of Dean et al. (2008) it is calculated
how NTB affects the price of consumer products. Their sample consists of 47 consumer
products, traded between 60 countries in 2001. Through use of input-output data the authors
concluded that NTB can increase the prices of consumer products by up to 54%. However, the
authors state that the variation in the sample is large. Nonetheless it is clear that NTB do have
a large impact on the prices of consumer goods. For the United Kingdom most forms of NTB
are not going to cost the citizens money, because when the UK leaves the EU they will still
have the same rules and regulations as the EU, at least for a few years after Brexit. With the
UK free to change trade policies and trade institutions, regulations and rules may change over
the course of years. Another NTB for the UK and the EU will be the establishment of a border
between the UK and the EU with custom offices. McCallum (2008) wrote a paper about the
effects of borders on trade. The paper focuses on the border between Canada and the US,
countries that are partners in a FTA and that are similar in culture, language and institutions.
12
Even so, a border still hurts the trade potential, McCallum (2008) concludes that the trade between Canada and the US should be 50% larger.
While the effects of NTB are interesting, they will not be part of this research. It is only included here to show that Brexit is more than just an increase of tariffs.
2.4 Global value chain
In the first part of the literature it is explained that countries should produce products they can produce relatively cheap for export purposes and import products it cannot produce relatively cheap. However, in this modern global economy countries make only a part of a product, this is due to the second unbundling. The second unbundling is explained thoroughly in a paper from Baldwin and Evenett (2015). Improvements in information and communication
technology (ICT) made it possible to ‘transport’ schematics, plans and knowledge cheaper and easier, which made it possible to coordinate difficult tasks from far away. Because of this production processes could be fragmented over multiple countries, each country offering whatever part of a product they can make cheapest through use of economies of scale.
Production of parts of a product happens in countries that can produce these relatively cheap.
These parts will be assembled in a country that provides assembly relatively cheap. The marketing happens in a country that can provide these services relatively cheap. Many countries can become, or already are, part of a global value chain and as a result economies are more interconnected than ever. The global value chain created a problem with measuring trade flows. When countries were still trading in finished products, export flows and import flows were informative variables. However, since every nation now makes a small part of a product, gross trade flows overestimate the economic value. For example: Country A imports three parts of a computer, each of these parts is worth $100. Country A then assembles these parts together creating a whole computer, this is then exported for $1000. Country A imported for $300, but exported for $1000, however the added value is only $700, thus gross trade flows is overestimating the worth of the traded good. A new measurement was needed, this would be how much value each nation adds to a product. Johnson and Noguera (2012) looked into the measurement of trade flows versus measurement in value added. Through use of input-output data on 94 countries, they came to the conclusion that value added exports represent about 73% of gross exports on average. Regions that are lower than average are Europe and East Asia. While North and South America, South Asia, Oceania, the Middle East and Africa are higher than average. The authors conclude that gross exports flows are
overestimating value and that value added is a better measurement. Timmer et al. (2013) also
13
delved into this subject. They calculated GVC income and GVC jobs in 27 European countries with the use of the World Input-Output Database. The authors show that countries that export a large amount of manufacturing products are usually depending on large amounts of imported intermediates. Thus gross export figures are overestimating the value creation in those exporting countries. The authors conclude that export flows cannot be used as a
measurement of competitiveness anymore. Gross export flows show that Europe has a competitive advantage in low to medium technology industries, while GVC based
measurement show an ongoing increase in high tech industries. There is also a shift from low skill workers to high skilled workers in Europe. It seems that the comparative advantages of European countries lies within high skilled production and that GVC development are magnifying this competitive advantage.
2.5 Brexit
In this literature review it is explained how trade theory works and that free(er) trade is important for welfare and GDP. By leaving the EU, tariffs will be imposed between the EU and the UK. It is still unclear how high these new tariffs will be.
However, before delving deeper into Brexit, it is important to describe the current setting. In the book ‘the UK regional-national economic problem’ written by McCann (2016) is a chapter based on the UK’s trade performance. In the post war era the UK did what most countries did, they traded with their old colonies. UK’s bigger trading partners were in south and east Africa, Australia, South Asia, Caribbean and North America. Later this changed, when the UK joined the EEC. The UK’s trade shifted from trading with far away colonies to trading within the European region, while becoming more open to trade and FDI. In the 1990’s trade and FDI was growing rapidly and wouldn’t stop until the crisis of 2008. In 2011 47.5% of all exports went to the EU, while 50.6% of all imports came from the EU. The UK is highly interconnected with the EU. The UK’s GVC participation is low, the UK ranks 26 th of all 34 OECD countries. This ranking is based on how much of the production is used as part of the GVC. This lower ranking means that the UK is linked in the GVC less than other OECD countries. However, it seems that the UK is adding more value to intermediate products then comparable countries, resulting in a value added export of about the same to comparable major European nations. Los et al. (2017) wrote an interesting paper about
dependency on the EU and voting ‘remain’ or ‘leave’. In this paper the UK is split in multiple
regions, the dependency on the EU is measured by the share of export to the EU. Regions
with more export in value added to the EU, regions that get more income from export to the
14
EU and regions that gain the most GDP from exporting to the EU tend to vote in favour of leave. This is an almost paradoxical result. The regions that had the highest percentage ‘leave’
voters will likely be hit the hardest by Brexit. The reason why people voted ‘leave’ may not be influenced by economic prognoses as much as by cultural or political reasons. This shows that Brexit is more complicated than a country leaving a trade organisation.
A rapidly expanding literature explores how Brexit may affect the UK. One of the first papers was written by Ottaviano et al. (2014). In this paper the authors looked at the effects of tariff barriers and non-tariff barriers that will be imposed after Brexit. The authors also included the gains of taxpayers’ money that was previously sent to Brussels. There were two scenarios explored, in the optimistic scenario there will be no tariffs, while non-tariff barriers will increase. In the pessimistic scenario tariffs will rise to most favourable nation status, while non-tariff barriers rise even more. In both scenarios the gains from the money that previously was sent to Brussels remains the same. The conclusion was that the UK’s GDP will shrink with 1.23% in the optimistic scenario and with 3.09% in the pessimistic scenario. Dhingra et al. (2016) wrote a paper about the Brexit as well. This paper follows Ottaviano et al. (2014) to a certain extent, but it does not look at GDP, but at standard of living. It also has the same two scenarios as the research by Ottaviano et al. (2014). The paper by Dhingra et al (2014) does show standard of living changes for the UK, the EU and the other available countries in the world input-output database. The British population will lose about 1% in the optimistic scenario and 2.3% in the pessimistic scenario. Other countries that stand to lose the most are Ireland, the Netherlands and Belgium.
While this may seem depressing for the people living in the UK, the effects of the Brexit are not certain. The papers written about Brexit calculate the effects of (non)tariffs barriers imposed between the EU and the UK. But this is only one part of the story. While it is true that tariff can and probably will be imposed between the UK and the EU, the other side is that the UK is free to make trade deals with other countries around the world. This can be a way for the UK to mitigate losses. The results of Brexit are hard to predict and there will be a lot of uncertainty of what will happen.
Like the papers of Ottaviano and Dhingra, this paper will also use two scenarios. A more
pessimistic scenario in which the EU and the UK will impose MFN tariffs and a more
optimistic scenario in which the EU and the UK will negotiate a better deal, this will be
explained in more detail later in this paper.
15
2.6 Hypotheses
Based on the literature review so far it is safe to assume that the GDP of the UK and the EU will shrink. In both the EU and the UK, some citizens will lose their job as a result of Brexit.
However, it is hard to hypothesize how much the UK and the EU will lose each. The gains of joining the EEC in 1973 can be a benchmark for this. There will be multiple hypotheses in this paper.
Hypothesis1: Ireland, the Netherlands and Belgium are losing the largest share of GDP of the EU countries.
Knowing that countries usually trade more with close neighbours, Ireland, the Netherlands and Belgium are likely to lose more than other EU countries.
Hypothesis 2: Germany, France and Poland will lose the most jobs.
Big countries tend to have a larger population and thus more jobs, the chance of them losing more jobs are thus larger. Likely Germany, France and Poland will lose the most.
Hypothesis 3: The UK’s GDP will shrink with 4% to 6% in the optimistic scenario and 2% to 3% when MFN tariffs will be applied.
It is difficult to predict how much the UK's GDP will shrink. However, the UK joined the
EEC in 1973, this can be used as a benchmark. In 1974 the GDP growth of the UK was
-2.5%. It is interesting that the GDP shrinks as a result of joining a trade union. Overman and
Winters (2011) wrote a paper about this. The UK’s economy had to switch from trading with
(ex)colonies to trading with developed countries in Europe. Joining the EEC resulted in
higher tariffs for the (ex) colonies and lower or no tariff with EEC members. This led to a
decline in trade with their big trading partners and an increase in trade with less important
trading partners. The -2.5% GDP growth is the total package, so it includes changes in tariffs
and NTB. In this paper only the effect of tariffs are calculated, so it would probably be much
lower than -2.5%. On the other hand, the tariff changes will be quite more severe than the
tariff changes in 1974, because the world is a lot more interconnected. Another benchmark
that can be used is how the GDP of the UK reacted to economic downturn. In the financial
crisis of 2009 the GDP growth of the UK was -5%, because this happen quite recently it is a
great benchmark for this paper. A good estimate would be a loss of 4% to 6% GDP in the
MFN scenario and 2% to 3% in the better deal scenario.
16
Estimating how much jobs the UK will lose will be difficult without a benchmark, but it is safe to assume that the UK will lose more jobs than any EU country.
3 Data
This research is based on data from the World Input-Output Database (WIOD), made by Timmer et al. (2015). WIOD provides a lot of data about a number of variables such as:
production, jobs, emissions, compensations and more. There are multiple countries and every country is divided in a number of sectors. All this data is in millions of dollars and provided within a large matrix. In the age of fragmentation, WIOD can be used to explain trade flows through value added instead of gross export and import values. WIOD can also be used to calculate various footprints, for example the carbon dioxide footprint. WIOD can also be used to calculate a number of shares, such as input shares or employment shares. However, to set up WIOD some assumptions had to be made to make the model work. One assumption is that every sector only produces one product. This is not a realistic assumption, because industries usually produce multiple products and some products are produced in multiple industries.
Another assumption is that production is proportional. This is also not a realistic assumption as it does not allow for substitution or for economy of scale. The last assumption is that all prices are one. While this may seem unrealistic, it can be done through units of measurement.
For example, 500 gram of apples is $0.50, so a kilogram of apples costs $1. Thus apples will be measured in kilos.
WIOD has two releases, in this paper the 2013 release will be used. In the 2013 release is data on 40 different countries and the rest of the world as a 41 st country. All EU countries are available except for Croatia. There is data from multiple years, ranging from 1995 to 2011. In this paper data from 2009 will be used, because data on jobs is not available in later years.
4 Methodology
The methodology chapter will be divided in three parts. The first part focuses on the method, the equations used are be based on methodology provided in the paper of Vandenbussche et al. (2017) and Los, Timmer, de Vries (2015). In the second part the tariffs will be discussed and the third part is about trade and job elasticity.
4.1 Method
An input-output decomposition technique is used to calculate how Brexit will affect the
17
economies of the EU and the UK. The decomposition technique, explained in the next pages, will decompose the data in value added created and jobs needed per industry per country based on the final demand of a country. This information is needed to understand how the EU and the UK are affected by the Brexit.
There are N countries, each of these countries has S industries and F production factors.
Output created in an industry in a country is using labour and capital, which are domestic production factors, and intermediate products, which are created domestically or imported from foreign countries. Final demand is consumption and investments by household and the government. Final demand can be fulfilled by domestic suppliers or imported from foreign countries. Because of the large number of exports and imports of intermediate and final products, it is important to name source and destination countries/industries. The source country will be defined as i, while a destination country will be defined as j. The source industry will be defined as s, while the destination industry will be defined as t. It is important to note that i country can be the same as j country, in that case the products are used
domestically as either an intermediate or a final product. For industries, s can be the same as t, in this case products from an industry are used as an intermediate in the same industry. If market clearing is assumed, all products created within a country will be used either
domestically or in foreign countries, as intermediates or for final demand. Furthermore, it is assumed that products have one price. This lead to a product market clearing condition:
(1)
In equation 1, is the value of the total output of industry s of country i. The term represents the products sold by industry s in country j. The term represents all the products sold by industry s for intermediate use in industry t of country j.
The next step is to calculate the intermediate inputs as a share of gross output, also known as the input coefficients. This will be defined as the A-matrix (SN x SN), which will be
calculated in the following way:
(2)
In this equation, represents the total output produced in each country-industry. The term
shows the intermediate input needed as a share of gross output (for both domestic
18
and foreign output). Each of the is placed in the A-matrix. Resulting in the following A-matrix:
With this A-matrix, equation one can be rewritten as:
In this rewritten equation 1, is a vector with the output levels from each industry-country.
The term is a vector that contains the demand for final products in country j. This rewritten equation one can be expressed in the more compact form:
(3)
This equation can be rearranged, to arrive at a fundamental input-output identity created by Leontief (1936):
The I term in the Leontief inverse is called an identity matrix. It is a (SN x SN) matrix with ones on the diagonal and zero’s everywhere else. The is better known as the Leontief inverse and it shows the gross output needed to produce one unit of final demand.
To calculate the effects of Brexit on GDP and employment, both must be linked with final demand. To do this the output per sector will be the standard function of the quantity of value added and employment per industry needed in the production process. For this the following equation is used:
(4)
The term has been used before and is the output of industry s of country i. The term
is the value added in that industry in that country. The term is the amount of
value added created per output created within industry i in country s. This will be a column
vector (SN x 1), the elements in this vector are industry and country specific. Now a new
19
matrix can be created, that shows how much value added is created within each country- industry.
(5)
This O-matrix (SN x SN) shows us how much value added is created in each industry in every country. If all the value created in each country would be summed, it would equal the GDP.
The term is a matrix with the elements of on the diagonal, the rest of the matrix is zero.
For employment, the same calculations will be made.
(6)
In this equation, is the number of workers required to produce one product of output and this is again a column vector (SN x 1). While is the number of workers per industry per country. This leads to the following equation:
(7)
The K-matrix (SN x SN) shows the number of jobs per industry per country needed for final demand in each country j.
4.2 Tariffs
With the K-matrix and the O-matrix the impact of the Brexit can be calculated. But first the
two scenarios need to be fleshed out. This is difficult because the WIOD database is based on
sectors, not on products. A 2016 report by the WTO states the average MFN tariff levied by
the EU (Appendix A). These tariffs are sorted in product groups and for MFN tariffs they are
sorted in three sections: AVG, Duty free in % and Max. AVG stands for average and is the
simple average of all the tariffs in that product group, duty free is the share of duty-free HS
six-digit subheadings in the total number of subheadings in the product group and max is the
maximum tariff paid within the product group. To get from these tariffs to tariffs based on
sectors within WIOD, the WTO product groups (with MFN tariffs) are linked to the WIOD
sectors. This can be found in appendix B. Some WIOD sectors consist of multiple product
groups, the tariffs of these will be averaged as well. In some cases, the finished product is not
available, an example of this is plastic. In such a case, the tariff for products needed for plastic
production is used instead. Setting up tariffs per sector like this is not ideal. The best way
would be to divide products over the sectors and take a weighted average per sector. Due to
20
time and data constraints that was not possible. The tariffs per sector can be found in table 1 on the next page, this shows the MFN tariffs per sector and the tariffs in case the UK and the EU negotiate a favourable deal (50% of the MFN tariffs). The service sectors are not stated in the table. That is because services cannot (easily) be exported or imported. However, tariffs will have an effect on the value creation and number of jobs in the service sectors, because services do deliver ‘intermediates’ to the sectors in table 1. Services will need a tariff for this research, so the average of table 1 will be used. The average of the tariffs is 4.6% for the MFN deal and 2.3% in case the UK and the EU negotiate a better deal.
4.3 Trade and job elasticity
The only thing left to address, is the trade and job elasticity. Trade elasticity means that when tariffs drop with a certain percentage, demand increases with a certain percentage as a result.
In the case of Brexit, tariffs go up with a percentage and thus the demand for those goods go down. Job elasticity means that when production goes up, more jobs are needed for
production. When production decreases, demand for jobs will go down too. Production in
other countries will go down as a result of Brexit, so people around Europe will lose their
jobs. The trade elasticity used in this paper will be -2. This means that when a tariff goes up
by 1%, demand for those products go down by 2%. This is the same elasticity used in the
paper of Vandenbussche et al. (2017). Setting the trade elasticity is difficult because in the
academic theory trade elasticity goes from -2 to -6, according to Simonovska and Waugh
(2014). And every product has its own trade elasticity, which makes the matter even more
complicated. For this reason it is decided to follow Vandenbussche et al. (2017) and stick with
the trade elasticity of -2. Job elasticity is also the same as the one used in Vandenbussche et
al. (2017), which is 0.4. This means that when production goes down by 1%, the amount of
jobs goes down by 0.4%.
21 Table 1: MFN tariffs and better deal
This provides the necessary information to calculate the effects of Brexit. When the O-matrix and K-matrix are calculated, they show how much value added and jobs are needed for export or domestic use. These values will be adjusted due to the increased tariffs and the trade
elasticity. Here is a fictional example to illustrate this. Imagine that the Irish rubber and plastic sector exports 100 million US dollar of value added for final demand in UK. The tariffs for this sector will be either 2.5% in case of the MFN scenario or 1.25% in case of a better deal.
With a trade elasticity of -2, the exports will go down by 5% or 2.5% respectively. So the new value added exported to the UK will be 95 million dollars (100 - 5%*100 = 95) or 97.5 million dollars. The amount of jobs needed for the export of 100 million dollar in the rubber and plastic sector is 1,000. The job elasticity is 0.4, seeing that production goes down with 5%
Sector Average MFN tariff Better deal
Agriculture, Hunting, Forestry and Fishing 15.7 7.85
Mining and Quarrying 2.0 1
Food, Beverages and Tobacco 16.3 8.15
Textiles and Textile Products 9.0 4.5
Leather, Leather and Footwear 4.1 2.05
Wood and Products of Wood and Cork 0.9 0.45
Pulp, Paper, Paper , Printing and
Publishing 0.9 0.45
Coke, Refined Petroleum and Nuclear Fuel 2.5 1.25
Chemicals and Chemical Products 4.5 2.25
Rubber and Plastics 2.5 1.25
Other Non-Metallic Mineral 2.0 1
Basic Metals and Fabricated Metal 2.0 1
Machinery, Nec 1.9 0.95
Electrical and Optical Equipment 2.7 1.35
Transport Equipment 4.3 2.15
Manufacturing, Nec; Recycling 2.6 1.3
22
or 2.5%, the amount of jobs goes down with 2% (5% * 0.4 = 2%) or 1% in the pessimistic and optimistic scenario. The number of jobs left after Brexit in this sector will therefore be 980 or 990. In conclusion, in this example, Ireland will lose 5 million dollar of value added and 20 jobs in case MFN tariffs are imposed and 2.5 million dollar of value added and 10 jobs in case a better deal is negotiated.
There is a weakness when calculating the impact of Brexit like this. The O-matrix shows all the value added that is created and where it will eventually go, directly and indirectly. To get back to the fictional example, this means that the value added export from Irelands plastic and rubber sector may have come to the UK via Germany’s machinery sector. However, in the O- matrix all the value added is shown as directly exported from EU countries to the UK (or from the UK to EU-members). Because of this the loss in value added may be over or
underestimated. Here is another fictional example to explain this. In this example only the MFN scenario is taken into account for the sake of simplicity. The Irish rubber and plastic sector exports 100 million dollar of value added to the UK, but in reality only 50% of this is exported directly to the UK, the other 50% goes indirectly to the UK via Germany's
machinery sector. The direct exports have a tariff of 2.5%, which results in 5% less demand because of the trade elasticity. So 50 million dollar of value added gets reduced by 5%.
Ireland will directly export 47.5 million dollar of value added to the UK. The indirect exports go via Germany's machinery sector. That sector has a tariff of 1.9%. So the exports of plastics and rubbers go down with 3.8% because of the -2 trade elasticity. This will result in a value added export of 48.1 million to the UK via Germany. In total 95.6 million dollar of value added will arrive in the UK, this is more than in the previous example in which only 95 million dollar of value added arrived in the UK. In the first fictional example the loss in GDP was overestimated, but the opposite could be true if the machinery sector had a higher tariff than the plastic and rubber sector.
Adequately addressing the potential bias due to indirect value added trade flows is beyond the
scope of this thesis. Future research should seek to examine the size and direction of the bias
due to my omission of these indirect trade flows in the analysis. It might be that my method
overestimates the effects for some products and underestimates for others, thus balancing out
in the aggregate and leaving aggregate results largely unaffected. But this is speculative and
will require additional research.
23
5 Results
The results will be split in two halves, in the first half the results on value added (GDP) will be explained. The second half will be about jobs.
5.1 Gross domestic product
The results will be shown in four tables. The first table of this chapter will show the effects of Brexit on the GDP (in billions of US dollars) of the UK and EU (excluding the UK). The second table of this chapter will include the UK and the North-western European countries:
Austria, Belgium, Denmark, Finland, Germany, Ireland, Luxembourg, the Netherlands, Norway and Sweden. The third table of this chapter will include the UK and the
Mediterranean countries: Cyprus, France, Greece, Italy, Malta, Portugal and Spain. The fourth and last table will include the UK and the Eastern European countries Bulgaria, Czechia, Estonia, Hungary, Lithuania, Latvia, Poland, Romania, Slovakia and Slovenia.
Table 2 shows the effects of Brexit on the EU (excluding the UK) and the UK. The table shows the GDP in 2009 in the second column, what their GDP would be in case MFN tariffs would be levied in the third column, what their GDP would be in case they could negotiate a better deal in the fourth column. The last two columns are the percentage differences 1 . Column five shows the percentage difference between column two and three, while column six shows the percentage difference between column two and four.
Country (1)
GDP 2009 (2)
GDP MFN tariffs
(3)
GDP better deal tariffs
(4)
Percentage difference column 2 and
3 (5)
Percentage difference column 2 and
4 (6)
UK 2,008.4 1,842.0 1,925.2 -8.283 -4.141
EU 12,794.4 12,774.2 12,784.3 -0.158 -0.079
Table 2: GDP of EU and UK in billions of US dollars
If MFN tariffs would be levied the EU’s GDP would shrink with 0.158%, while the UK’s GDP would shrink with 8.283%. If the EU and the UK negotiate a better deal, the EU’s GDP shrinks with 0.079% while the UK’s GDP will shrink with 4.141%. It is clear that the UK depends more on the EU than the EU depends on the UK. As a result, the UK has more to lose than the EU. The difference in losses between the two are very high, the EU loses a fraction of their GDP while the UK loses a significant amount of their GDP.
1
