Political attention for globalisation: An empirical test of the embedded liberalism thesis

(1)

10-8-2018

Political attention

for globalisation

An empirical test of the embedded

liberalism thesis

(2)

(3)

Political Attention for Globalisation

An empirical test of the embedded liberalism thesis

Master Thesis 10-8-2018

T. van de Putte

Student number: 1272837

Master Public Administration: Economics & Governance Leiden University

(4)

(5)

1 Introduction

“When you’re almost 800 Billion Dollars a year down on Trade, you can’t lose a Trade War! The U.S. has been ripped off by other countries for years on Trade, time to get smart!”

So wrote Donald Trump (2018) on his Twitter account on the 2nd_{of June, 2018. His tweet is remarkable}

in a time where international trade has flourished. Since the collapse of the Bretton Woods system around 1970, the answer to economic stagnation has been free trade (Frieden, 2006). The period from 1970 until today is known as the period of globalisation, funded on the movement of neoliberalism. Globalisation “is a process that erodes national boundaries, integrates national economies, cultures, technologies and governance, and produces complex relations of mutual interdependence” (Dreher, 2006, p. 1092). Neoliberalism underlies the economic aspect of globalisation, being a political theory which advocates for free trade and individual property rights (Harvey, 2006).

Now, the President of the United States seems to introduce protectionist measures, such as tariff barriers (Helmore, 2018). Within the period where the paradigm of neoliberalism rules the words of the President are unprecedented and certainly against the trend of freer trade (Harvey, 2006). President Trump’s statement brings up the question who benefits from curbing free trade. No one, is the answer of the Dutch Minister for Foreign Trade and Development Cooperation, Sigrid Kaag (ANP, 2018). “A trade war only knows losers.1_{” But does mister Trump have a point? Are the profits of free}

trade equitably shared? On the one hand, there are authors who claim that there is a race to the bottom of welfare state provisions due to globalisation (Starke, 2006; Allan & Scruggs, 2004; Genschel, Kemmerling, & Seils, 2011). On the other hand, there are researchers who say that the welfare state compensates for the negative effects of globalisation (Rodrik, 1998; Swank & Steinmo, 2002; Walter, 2010; Hays, Ehrlich, & Peinhardt, 2005).

Bivens & Lawrence (2015) show us in the graph below that a gap has been developed over time between the development of productivity and hourly compensation of workers. The gap is originated somewhere around 1970, exactly around the time the Bretton Woods system collapsed. Productivity continued to rise, while the hourly compensation only slightly rose. The continued rise in productivity must have led to a grow in total welfare, but the typical American worker did not see this increase in welfare in his hourly compensation. Is this just a coincidence or is there a relationship with globalisation?

(8)

Graph 1 Disconnect between Productivity and a Typical Worker's Compensation, 1948-2014 (Bivens & Lawrence, 2015)

Trade comes with benefits and losses, it creates winners and losers. Open economies tend to have larger governments, especially small countries such as the Netherlands (Katzenstein, 1985). There are two big reasons for this. Firstly, citizens of open economies want to be compensated by the external risk free trade brings. This is called the compensation hypothesis (Rodrik, 1998). Secondly, governments in favour of free trade want to maintain support of the public and thus provide them with insurances. This is called the embedded liberalism thesis (Hays, Ehrlich, & Peinhardt, 2005). With this compensation, no matter the reason for it, losers of globalisation are then less losers as they would be without compensation.

Whereas a lot of researchers investigated the compensation hypothesis and the embedded liberalism thesis, both on macro and micro level, they do not provide an in-dept analysis of the policy agenda of the government (Walter, 2010; Rodrik, 1998; Burgoon, 2001). This research tries to fill in that gap. The central question of this research is therefore:

“Is there a correlation between trade liberalisation and political attention for losers of globalisation?” With freer trade in the world, and the fact that opener economies have bigger governments, I expect that trade liberalisation leads to more political attention for losers of globalisation. Pivotal are however other variables which influence the underlying causal relationship, such as unemployment,

(9)

immigration and economic growth. The influence of those three variables is the domain of the sub-questions of this research. It is important for policy makers to have a mirror: do they forget an important topic?

The practical relevance of this research derives from the importance not to miss a major trend in the world economy. If policy makers fail to acknowledge the impact of globalisation, countries might be confronted with more than just Donald Trump twittering free trade is bad for his country. Brexit, the leave of the United Kingdom from the European Union, is another example of people worrying about globalisation. Brexit has been linked to deindustrialisation and immigration, both the effect of pro globalisation policies in the past (Jack, 2018). Brexit thus may be the result of the underestimated effects of globalisation in the past.

The Netherlands is an example of a small state which has been more exposed to free trade than the large European economies (Katzenstein, 1985). This is why I use this particular country in this research. I measure the political attention for globalisation, especially for losers, on the basis of troonredes, executive speeches given by the Dutch monarch every year. In various regression models, I regress the political attention on economic globalisation.

In order to examine this I am first providing the existing theoretical framework. This chapter falls apart in two parts: a historical background on the liberalisation of trade and theories on globalisation related to compensation. At the end of this chapter I present the hypotheses. In the next chapter I describe the research method and case selection with its rationale. Thereafter I present the results and analyse them, both in a separate chapter. I finish with my conclusions.

(10)

2 Theoretical framework

At first I will describe how and why free trade was introduced in modern history. Four periods can be distinguished. The first period was until approximately 1830 where there was little openness in the international economy. The second period is the period where economies became more open and international trade flourished, the period of classical liberalism. This endured from 1830 until 1945. The third period is the period of embedded liberalism, the era of Bretton Woods (1945 – 1970) and the fourth the period of neoliberalism, from 1970 until now (Frieden, 2006; Kindleberger, 1975). Secondly, theories on the compensation of losers and the political attention for this will be brought up.

2.1 Historical background trade liberalisation

2.1.1 The world before free trade

In the past, there have been multiple international economic regimes. According to Kindleberger (1975) it was not until the early eighteenth century that free trade, trade with no governmental limitations at the border, internationally began to manifest itself. Tuscany was ahead of its time with the implementation of free trade in 1737, when it allowed corn from an other region of what we now call Italy to be imported without restrictions. It was only one hundred years later, after the Napoleonic War, that the big steps to free trade were made. Kindleberger (1975) delineates some measures that were taken: the abolition of the prohibition of silk exports, freedom to export coal and the manor and guild system that deteriorated. Each European country had its own reasons for moving to free trade, although the European countries in this time should not be seen as different economies, but more as one entity moving to free trade, since ideological factors played their role as well.

2.1.2 Classical liberalism and the gold standard

The Great Transformation, written by Karl Polanyi (1980) describes the change in the international economic regime during the nineteenth century. The great transformation is the transformation of a society in which social relationships were inherently connected with a man’s economy into a self-regulating market society where a man’s economy is disconnected with everything except his individual interest in the possession of material goods (Polanyi, 1980, p. 46). Until 1815, several wars were fought on the European continent. It was not until 1914 that a great war broke out which threatened the global balance of power. Four defining factors contributed to these hundred years of peace: (1) the balance-of-power system between the Great Powers; (2) the international gold standard; (3) the self-regulating market; (4) the liberal state (Polanyi, 1980, p. 3). Trade and peace were being linked in the nineteenth century, with trade being dependent on an international monetary system at the end of the century: the Gold Standard.

(11)

Until the end of the nineteenth century, money was made of metals, mainly gold and silver. The percentage of metal varied across the currencies and when a lord needed more money, he lowered the percentage of metal in the coins. When merchants were trading, they agreed upon which coin they would use for the trade, taking into account the percentage of precious metal in the coins they would use. Hence, the actual prices of goods were in gold. In the late nineteenth century, people started to identify currencies with nations due to the rise of the nation state. One could argue that the world was one monetary union, with gold – and initially also silver – as the world currency. This is called the Gold Standard (Baldwin & Wyplosz, 2012, pp. 378-379). With one worldwide currency and no capital restrictions, free trade could flourish.

2.1.2.1 Hume’s price-specie mechanism

The gold standard and a country’s current account, more specifically the trade balance are related and self-correcting. When a country exports more than it imports, the money supply would increase as gold would flow into this country. The increased money supply would in turn raise inflation. With higher domestic prices and a fixed exchange rate – because the world is one monetary union with the gold standard – the domestic economy becomes less competitive, since foreigners have to pay higher prices. Consequently, this leads to a current account deficit (and thus lower exports). The opposite could also be the case. When a country has a negative current account – it imports more than it exports – the money would flow out thereby causing lower money supply, deflation, more international competitiveness and eventually a current account surplus. This mechanism is called Hume’s price-specie mechanism (Baldwin & Wyplosz, 2012, p. 380).

Around 1900, paper money entered the world, but initially changed nothing to Hume’s mechanism, because paper money rested on three principles: (1) paper money could always be converted into gold at banks; (2) central banks kept as much gold in their vaults as they printed banknotes for; (3) free trade and no capital restrictions. The first principle entails that there is a fixed exchange rate between banknotes and gold. Hence, a banknote symbolised the value of gold. The underlying monetary system had not changed (Baldwin & Wyplosz, 2012, pp. 380-381). The second principle means that every banknote is backed by a mass of gold. The third principle, however, needs to be further deepened with the principle of the impossible trinity.

2.1.2.2 The impossible trinity

The impossible trinity principle was first described by Mundell in a paper on the implications of the increased mobility of capital (Mundell, 1963). In 1963 he wrote that the world economy is becoming increasingly open and that this has effects for monetary and fiscal policies of countries with fixed or

(12)

flexible exchange rates. Baldwin & Wyplosz (2012, p. 361) describe the impossible trinity principle as follows:

“Only two of the three following features are compatible with each other:  Full capital mobility

 Fixed exchange rates

 Autonomous monetary policy.” In a figure, this would look as:

Figure 1 Adapted from Baldwin & Wyplosz (2012), p. 362

Given the circumstances a government may opt for option A, B or C. There is no best choice in all circumstances. A fixed exchange rate may be useful in order to improve the external competitiveness of firms in a country. Monetary policy autonomy may be useful in order to boost external competitiveness when domestic demand becomes lower. Depreciation then is the keyword to ease the pain of a recession. In the financial economic crisis of 2008 for example, the Greek currency could not depreciate in relation to the other European countries because they were in the same monetary union (Schinkel, 2009). I will now give an example of each of the three options, starting with the Eurozone.

Whereas the Eurozone as a whole follows approach C, the eurozone internally follows option A. Capital restrictions in the Eurozone have been gone as well as the possibility to have autonomous monetary policy. The prerogative to create monetary policy now lies with the European Central Bank, member states thereby abrogating their autonomy on monetary policy. With capital controls, and thus opting for option B, a country can have a autonomous monetary policy and a fixed exchange rate. This significantly boosts the external competitiveness of a country, but also comes with downsides. Baldwin and Wyplozs (2012, p. 365) name for example the fact that capital controls need to be administered and enforced by the government, which costs money, since people try to evade the restrictions or taxes. One example of option B has been the Bretton Woods system.

(13)

The system of the international gold standard collapsed in 1914, the start of the First World War. Due to hostilities the supply of gold became disturbed and governments therefore issued debt and printed additional money, not covered by gold. Rationing schemes kept prices stable during the war, but after the war hyperinflation made its appearance. Trouble was not gone, because The Great Depression hit the economies in the 1930s. Many countries let their currencies depreciate since this would boost exports. But the depreciation of the currency of one country is the appreciation of the currency of another, which lead to the notion that the international economy needed an international system (Baldwin & Wyplosz, 2012, pp. 381-384).

2.1.3 The era of Bretton Woods

Around the end of the Second World War the governments of influential countries, such as the United States, Canada and Western European countries, came together to discuss a new international monetary system. Option A of the impossible trinity had been tried with the gold standard, so now it was time to try option B. Capital controls were put in place and every currency became linked to the US Dollar. Exchange rates were fixed, but could be slightly adjusted and countries had full monetary autonomy (Baldwin & Wyplosz, 2012). The International Monetary Fund (IMF) was created in order to prevent countries from unilateral devaluating their currencies. The stability of the international monetary system became the shared responsibility of the IMF (International Monetary Fund, 2018). Ruggie (1982, p. 392) argued that the period after the War II is fundamentally different than before. From 1945 on, the international monetary system became subordinate to domestic social and economic policy. At the same time, welfare states were being set up. Ruggie calls this ‘embedded liberalism’. In the 1960s however the Bretton Woods system collapsed, because capital restrictions were cut. In 1971 the United Stated World Government ceased the convertibility of the US Dollar into gold as the dollar was seen as overvalued due to inflation. In 1973, currencies started to fluctuate again (International Monetary Fund, 2018; Baldwin & Wyplosz, 2012).

2.1.4 Globalisation and neoliberalism

Frieden (2006, p. 359) refers to the collapse of the Bretton Woods system as the end of an era. The period of stable growth in the developed world combined with the rise of the welfare state came to an end. The period of embedded liberalism was over. The period of globalisation started (Frieden, 2006).

Frieden (2006, pp. 105-111) brings up the question if free trade, both capital as well as physical products, could go hand in hand with fair trade (in the context of the gold standard). He writes:

(14)

“Even in an era of rapid growth, even in countries growing rapidly, even if free trade is the best possible policy for the economy as a whole, even orthodox economists accept that there are both winners and losers from freer trade.”

Kapstein (2000) published an article about these losers and winners based on four other scientific papers. He writes that the last 40 years income inequality has risen in industrial countries as well as trade. Free trade leads to more total welfare, but also has distributive effects. Imports for example changes the domestic employment. There is also a consensus that unskilled workers and the least educated are losers from globalisation and high skill workers are winners due to their competitiveness. Kapstein (2000, p. 381) also mentions that immigration is one of the factors influencing domestic labour markets, and thus creating winners and losers, although it is not the only factor influencing this. Freer trade thus brings winners and losers. The second part of the theoretical framework of this research will focus on the winners and losers from globalisation. Although Frieden’s quote is in the context of the era of the golden standard, from 1973 and on, it is again highly relevant. The period from 1973 until now has been dominated by the worldwide paradigm of neoliberalism. ‘Liberalism’ because of the emphasis on the value of freedom, and ‘neo’ because classical liberalism ruled the times of the gold standard. David Harvey (2006, p. 145) comes up with a definition of neoliberalism:

“Neoliberalism is in the first instance a theory of political economic practices which proposes that human well-being can best be advanced by the maximization of entrepreneurial freedoms within an institutional framework characterized by private property rights, individual liberty, free markets and free trade.”

With neoliberalism as the ruling paradigm and free trade as dogma the question rises who benefits from this. As said, the quote above about losers and winners of free trade is relevant again.

2.2 Theories on globalisation

2.2.1 Globalisation theory in general

Globalisation has already been discussed in this research, but what is globalisation actually? Dreher (2006, p. 1092) gives a definition: “It is a process that erodes national boundaries, integrates national economies, cultures, technologies and governance, and produces complex relations of mutual interdependence.” In his research he also defines three dimensions of globalisation: an economic one, a political one and a social one. I focus on the economic dimension in this research, entailing (freer) trade of goods and services and (freer) flow of capital.

(15)

I categorise theories on the globalisation of the last fifty years in two main categories: the first category are theories which claim that there is a so called ‘race to the bottom’, a competition between states in which they compete on social policy leading to a smaller welfare state; the second category consists of theories on the compensation of citizens for either their support for free trade or their increase of external risk due to free trade. Another distinction which can be made is the distinction between macro-level research and micro-level research. The macro-level research focuses on macro-economic statistics, at country level, as the value of imports, exports and government spending, while the micro-level research focuses on surveys with inhabitants and their attitude towards free trade (Walter, 2010). In this part, I follow the dichotomy between race to the bottom on the one hand, and the compensation on the other hand. When the micro or macro dimension is relevant, I will mention this.

2.2.2 Race to the bottom

With regard to the race to the bottom effect we first have to consider the political dimension of free trade and it counterpart protectionism. Strange (1970) and Held (1991) argue that the pace of the change in the international economic system is out of step with the change in international political system. This leads to external disturbance of domestic economic policy areas such as unemployment, interest rates and monetary reserves. The effectiveness of domestic policies will also go down and states can hurt other economies with their own policies on purpose. To solve these problems, two routes may be chosen: the first response may be cooperation and the second response is defensive. Scholte (2011) endorses the gap between the economic globalisation and political globalisation. Nation states all have their systems of accountability, which most international organisations lack. The world is globalising, but the nation states are not. Because of this, Strange (1985a; 1985b) pleas for pragmatic protectionism if there is enough competition in the domestic market and if there are social grounds to do so. She writes (1985a):

“We seem to be surrounded these days by important people who tell us that protectionism is a terrible thing, and that the creeping tide of trade restrictions must be turned back if we are ever to escape from world recession.”

One of the fields in which this has empirically been investigated is taxation. In the case of taxation, we know that in total globalisation has not led to a race to the bottom due to internal, in casu budgetary and external, in casu globalisational pressures (Swank & Steinmo, 2002). Genschel et al. (2011) define two mechanisms for tax rates in the European Union (EU): firstly, market integration and enlargement increase competitive pressure; secondly, tax coordination and jurisprudence reduce the overall tax

(16)

competition. Compared to the rest of the world, they find that the tax competition in the EU is stronger. This can be interpreted as evidence for the race to the bottom.

The analysis of Strange that the pace of economic integration is out of step with the political integration still holds. It is clear that in the case of the European Union, the path of coordination is chosen above the defensive path to solve the problems the gap between economic integration and political integration has caused.

2.2.3 Compensation and embedded liberalism

Katzenstein (1985) describes in his book that there is a difference between small en large countries in Europe. During their history smaller countries have had economies that were more open than the large industrial countries. This openness of their economy led to corporatist arrangements and to domestic compensation. In other words: the openness of an economy is positively correlated with the size of the welfare state.

The question why this is the case is answered by Rodrik (1998, p. 997). He found evidence that “government spending play a risk-reducing role in economies exposed to a significant amount of external risk”. Citizens thus ask for compensation to the risk which comes with opening up the economy. This is called the compensation hypothesis. The openness of an economy is measured as the imports plus exports divided by the gross domestic product (GDP) of a country. The size of a government is determined by the percentage of government spending related to the GDP. Burgoon (2001) further investigates the compensation hypothesis. He found that the effect of globalisation on the welfare state depends on the welfare state programme.

Next to these macro-level studies Walter (2010) conducted a study on the microfoundations of the compensation hypothesis. According to her, the causal mechanism at microlevel has been largely passed over. In her study she analyses survey data from Switzerland2_{. She finds proof for the}

compensation hypothesis as the Swiss losers from globalisation expressed feelings of economic insecurity and were proponents of welfare state intervention. Walter distinguishes two sides of the compensation hypothesis: a supply side and a demand side. She focuses on the demand side: do the people want to be compensated? The supply side entails: does the government (want to) provide compensation?

Hays et al. (2005) did research on the supply side. Governments want free trade, and in order to establish that, they have to keep their constituents happy by providing them insurances and other

2_{Although Katzenstein (1985) uses Switzerland in his book as an example of a small country, it is not clear why}

(17)

transfers to compensate the losers from globalisation. They call this the embedded liberalism thesis. Wat is different compared to the compensation hypothesis, is that the embedded liberalism thesis does not depart from the viewpoint of the citizen, but of the government, while the compensation hypothesis departs from the needs of the citizens. The authors further claim that is it important to make a distinction between the imports and exports of an economy. The variable openness is wrongly operationalised into the indicator imports plus exports divided by GDP. They argue that a rise in imports creates losers, a rise in exports winners. In the inquiry they find evidence, both microlevel and macrolevel for the support of the embedded liberalism thesis.

One of the researches in contradiction with the latter is the research done by Iversen and Cusack (2000). They claim that, and find evidence for, the fact that it is not globalisation, but deindustrialisation which contributes to the call for compensation. Their study uses macro-level data. Deindustrialisation is mainly driven by domestic factors such as increasing productivity and a change of consumer’s demand. The deindustrialisation variable is defined as “100 minus the sum of manufacturing and agricultural employment as a percentage of the working-age population.

Lastly, as mentioned in the introduction in the context of Brexit next to industrialisation, the issue of migration needs to be addressed. Burgoon (2014) looked at the effect of international migration on the compensation hypothesis and the race to the bottom effect. He has found that immigration has a strong positive effect on them both. Unemployment, extraction of benefits from the welfare state and shared values with the native population play an important role. The research uses survey data to determine the effect of migration.

2.3 Agenda setting and hypotheses

We can conclude from the theories above, that there is a race to the bottom effect when domestic economies are globalising. However, sometimes this downward pressure has been eliminated by upward pressures such as international coordination. This is at odds with the findings from the research on the compensation hypothesis and the embedded liberalism thesis, from which we should conclude that more globalisation in a country leads to more social spending and not less as we would expect from the race to the bottom theories.

As written in the introduction, this research is about political attention. According to Kingdon (1995, p. 405), the agenda setting process and the specification of alternative policy choices always precede the implementation of policies. This means that before opening up the economy, there must be political attention for globalisation and its effects.

(18)

Based on these theories, the following hypotheses can be formulated:

 The opener the economy, the more political attention for losers of globalisation;

 The government recognizes the distinction between winners and losers from globalisation;  The higher the unemployment rate, the more negatively formulated attention for

globalisation;

(19)

3 Research Method

As said in the introduction, this research is about the influence of the openness of the economy on the political attention for the compensation of losers of globalisation. I do not try to create an all-encompassing or holistic theory to clarify the reason behind the causality of this mechanism. Nonetheless, something can be said about the political attention for compensation. Where Hays et al. (2005) looked if citizens want to be compensated, what defines the people that want to be compensated and how the government can influence this, I look for evidence if the government shows any political attention for globalisation. With no political attention for globalisation and its consequences the embedded liberalism thesis does not hold. In order to compensate citizens that are losers from (economic) globalisation, the government must first show that it is aware of this problem and furthermore that the political attention for this must grow if the economy becomes opener. Two variables can thus be identified: openness of the economy as an independent variable and the political attention for globalisation as dependent variable. I will first elaborate on the independent variable openness. Thereafter, I will dwell on the dependent variable then, I will discuss the control variables and lastly the case selection and method of analysis.

3.1 Independent variable

As stated in the previous chapter, researchers differ in their approach to operationalise the concept of openness of an economy. Most researchers among whom Rodrik (1998) use the indicator ‘imports plus exports divided by the GDP’. As earlier written, Hays et al. (2005) argue that it is better to only look to the imports relative to the GDP. I will use both in different models, deriving the statistics from Statistics Netherlands, or in Dutch Centraal Bureau voor de Statistiek, often abbreviated to CBS. The statistics I used are the gross domestic product, the economic growth, volume of imports and exports, the unemployment rate, the immigration, and the number of inhabitants in The Netherlands with and without the Dutch nationality.

3.2 Dependent variable

3.2.1 Choosing documents: King’s Speeches

The dependent variable, political attention, is harder to measure. The first problem with this variable is that there are thousands of government and non-government documents available which can all be used, but never come close to an all-encompassing research. Therefore it is necessary to look for documents which include the main political discourse. Every third Tuesday of September3_{the Dutch}

(20)

main policy intentions for the coming year. Headlines in the political discourse of that year are always included, this eliminates the problem of selection bias. Breeman et al. (2009) and Mortensen et al. (2011) use speeches given by the head of state to analyse the political attention for a lot of topics, for example foreign policy, macroeconomy, environment and defence. Coding executive speeches is thus a good way to analyse the political attention. However, they do not use a category directly linked to globalisation, something which I do in this research.

3.2.2 Coding the troonredes

In order to find the political attention for globalisation, the sentences about globalisation should first be coded as such. In the coding process, a distinction can be made between (1) specific policy intentions for the coming year, (2) general statements about policy problems and (3) mere ceremonial statements (Breeman, et al., 2009, p. 5). The first category contains statements as (fictional) ‘the governments spends next year one hundred million euros on active labour market policies to curb the rise of the unemployment rate’. The second category includes statements as ‘control over government spending and a low public debt are of vital importance to maintain the welfare level we are used to’. The latter category, the ceremonial statements, are just formal, traditional sentences which are mostly being used at the beginning or at the end of a speech, called troonrede in Dutch. The ceremonial phrases which are the same in the speeches: ‘Members of Parliament’4_{and ‘in your difficult task you}

may feel supported by the realisation that many people wish you wisdom and pray with me for strength and God’s blessing’5_{. Over the years this phrase changed, but always stayed purely}

ceremonial.

Furthermore, text about globalisation needs to be identified. I divided the text into different paragraphs, following the approach of parlement.com (e.g. Parlement.com, n.d.). Mostly, I coded paragraphs as text on globalisation, but sometimes I coded sentences and even phrases as well. Coding whole paragraphs entails that introductory sentences are included. The direct link to globalisation then follows in the sentence thereafter, but would come completely out of the blue without an introductory sentence preceding it. This is mainly the case when the whole paragraph is about globalisation. Coding individual sentences or phrases occurs when the main topic of a paragraph is not about globalisation, but about something else, for example defence policy. There is then a small reference to globalisation.

4_{Originally in Dutch: ‘Leden van de Staten-Generaal’}

5_{Originally in Dutch: ‘u mag zich in uw zware taak gesteund weten door het besef dat velen u wijsheid}

(21)

Important topics coded as globalisation are:  European cooperation and coordination;  world trade;

 imports and exports of the Netherlands;

 references to an ‘open economy’ and ‘open society’.

According to Dreher (2006, pp. 1092-1094) globalisation consists of three subcategories: economic globalisation, political globalisation and social globalisation. The first one corresponds with the four topics I mentioned above. With political globalisation he looks at the number of embassies, membership in international organisations and participation in UN Security Council Missions. Social globalisation is about personal contact, information flows and cultural proximity (Dreher, 2006, p. 1094). In coding the troonredes I leave the political and social globalisation out of consideration. Moreover, statements with as main topic foreign policy are not coded as globalisation.

3.2.3 Operationalisation into indicators

After coding the text of the King’s and Queen’s speeches, I counted the total number of words of every speech. All mere ceremonial statements were excluded from this total wordcount. After that, I counted the words which were coded as something to do with globalisation. I calculated the percentage of words on globalisation of each troonrede in order to make data from different troonredes comparable. This percentage is the indicator for political attention. I call this indicator percwordg. I also created two other indicators, toneglob and loswin. Toneglob symbolises the tone of the text on globalisation. This is an ordinal indicator varying from zero to four. The loswin indicator shows if domestic (people with influence on the national policy agenda) losers, winners, both or neither one are mentioned. No abstract terms but concrete groups or people. In the table below, the indicators are specified with their values:

Table 1

Meaning of Indicators toneglob and loswin

toneglob loswin

value meaning value meaning

0 very positive 0 losers mentioned

1 positive 1 neutral (both/ none)

2 neutral 2 winners mentioned

(22)

3.3 Control variables

Other variables that matter, partly based on theories described in the previous chapter, partly based on common sense, are the unemployment rate, migration, if the year is an election year for the Tweede Kamer (Dutch lower house/ House of Commons) and deindustrialisation.

The openness of the economy has effects on the labour market of a country, especially in the European Union where there is free movement of workers. Competition of workers in export industries can lead to a loss of jobs. Therefore, the domestic (in casu Dutch) unemployment rate is included in the models (see table 4, 5 and 6).

One can also argue that in election years the government uses less specific policy intentions, because a new cabinet needs to be formed. The government then leaves new policy intentions to a new formed cabinet. The term ‘election year’ is maybe a little bit misguiding. In 1994 for example, there were elections held, but they took place early that year leading to the new formation of a cabinet in August of that year (Parlement.com, n.d.). However, there also were troonredes read by the King during an outgoing (caretaker) cabinet. Years when an outgoing cabinet prepared the troonrede were coded as ‘1’ and years with a conventional cabinet ‘0’.

Migration should also be taken into account. As written in the chapter theoretical framework, Burgoon (2014) finds an effect of migration on the compensation hypothesis and on the race-to-the-bottom-effect. I retrieved statistics on migration from Statistics Netherlands (CBS). I calculated the proportion of non-Dutch inhabitants as a percentage of the total population each year. I call this indicator percfor. Besides the percentage of foreigner that actually lives in the Netherlands, I also look at the non-Dutch immigrants as a percentage of the total population. This indicator I call percimnond. With non-Dutch people, I mean people without the Dutch nationality.

Lastly, the deindustrialisation should be taken into account, because deindustrialisation may be the real reason citizens want to be compensated, not globalisation in terms of openness (Iversen & Cusack, 2000). I follow the same calculation method as Iversen and Cusack (2000):

𝑑𝑒𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑖𝑎𝑙𝑖𝑠𝑎𝑡𝑖𝑜𝑛 = 100 −𝑚𝑎𝑛𝑢𝑓𝑎𝑐𝑡𝑒𝑟𝑖𝑛𝑔 + 𝑎𝑔𝑟𝑖𝑐𝑢𝑙𝑡𝑢𝑟𝑎𝑙 𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡 𝑤𝑜𝑟𝑘𝑖𝑛𝑔 𝑎𝑔𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛

Statistics Netherlands counts the number of jobs in different sectors, each major sector receives a letter. For the manufacturing and agricultural employment, I included sector A: agriculture, forestry and fishing; and B-E: industry (without construction industry).

(23)

3.4 Case selection and research method

Katzenstein (1985) identifies a couple of small countries that have a different approach to globalisation then bigger countries. Smaller countries are more likely to compensate for globalisation. Among these ‘small countries in world markets’ is the Netherlands, a country with a historical tradition of trade (Kindleberger, 1975, p. 25). Besides this, the Netherlands is the world most globalized country according to the KOF index of globalisation 2017 (ETH Zürich KOF Swiss Economic Institute, 2017). Thus, the Netherlands is a typical case to investigate when it comes to the openness of an economy. With regard to the temporal limitations I investigate the period 1969 until 2016. This has two main reasons. Firstly a practical reason: calculation methods of GDP, unemployment rates et cetera are being updated. Netherlands Statistics publishes in their National Historical Accounts6_{data going back}

to 1969, which makes the data per year comparable and thus internally valid. The second reason is more fundamental and theoretical. As I described earlier, around 1970, the period of embedded liberalism made room for neoliberalism. The collapse of the Bretton Woods system was a main contributor to this. After 1970 globalisation flourished and international trade was tremendously boosted. This makes the period 1970 until 2016 in particular appropriate for research on globalisation. In the research period the international financial system was not fundamentally changed. This is beneficial to the internal validity of this research.

3.5 Method of Analysis

In order to investigate the relationship, and on top of that the causal mechanism, between the openness of an economy and the political attention for globalisation I use regression models. In these models I include the control variables mentioned. I also lagged some variables. The results of this regression model are shown in the next chapter.

Possible problems with the reliability are not due to the regression analysis with the same dataset. However the collection of the data might be less reliable. The economic indicators derived from Statistics Netherlands can easily be retrieved from their database, using the GDP by final expenditure. The data collection from the troonredes however is harder to replicate, because some sentences lie in a grey area. The question if a sentence refers to globalisation is therefore harder to determine, but with the handles provided one should come close to my results. The internal validity in some models is thin, due to the amount of data available. More data would boost the internal validity. The external validity is limited. The more a country looks like the Netherlands, the more external valid a claim is with regard to the conclusion of this research.

(24)

4 Results

In this chapter I present the results of the analysis. Firstly, I am shortly showing descriptive statistics of the coding process and the variable election year. Then I will show the economic data retrieved from Statistics Netherlands and calculations I did with these data in order to create the indicators for this research. After that, I will show the results of the regression models.

So first the descriptive statistics. In the table below you will find the details of the variables year, percentage of words on globalisation, the tone of the words on globalisation, if losers or winners are mentioned and if in that particular year the troonrede has been prepared by an outgoing cabinet.

Table 2

Descriptive Statistics of Coding Process

variable observations mean std. deviation minimum maximum

year 49 1993 14.29 1969 2017

percwordg 49 8.43 4.64 1.37 23.36

toneglob 49 1.86 1.12 0 4

loswin 49 0.88 0.53 0 2

electy 49 0.18 0.39 0 1

As stated in the table, I analysed 49 troonredes, one each year. The most important variable is the main dependent variable, percwordg. The variable could theoretically vary between zero and one hundred, since the number embodies a percentage. The variable actually varies between 1.37 and 23.36 per cent. Loswin and electy are nominal variables, so we cannot give much meaning to the descriptive statistics of these variables. Of the mean of the variable toneglob we can derive that on average the attitude of the government towards globalisation is slightly positive.

The results of the descriptive statistics of the economic data is shown on the next page in table 3. What we can see here is that some statistics are missing because they were not yet available. If there are 48 observations, the data of 2017 is missing. The economic growth of 1969 is also missing, leading to 47 observations there. The variable ratimex is the ratio between imports and exports. A ratio of 1 means that the import equals the export. Impgdp is the ratio between imports and GDP. The GDP is measured in real terms at the price level of 2010 in million euro’s, as well as the presented values for imports and exports. From this data, we can easily see that the GDP of the Netherlands has triples in fifty years. On average, the Netherlands had an annual economic growth of 2,36% during 1969 and 2016.

(25)

Table 3

Descriptive Statistics of Economy Related Variables

variable observations mean std. deviation minimum maximum

gdp 48 446738.3 143374 226785 672093 econgr 47 2.36 1.91 -3.8 6.1 import 48 213839.6 138655.6 58092 506750 export 48 242981.6 160241.8 56272 583395 ratimex 48 0.89 0.04 0.84 1.04 impgdp 48 0.43 0.15 0.26 0.75 openness 48 0.92 0.35 0.50 1.62 unempr 49 5.34 1.78 1.6 9 percfor 49 3.90 1.02 1.49 5.69 percimnond 48 0.50 0.19 0.25 1.10 deindustr 48 87.58 2.80 80.82 91.23

Some variables are worth it to have a closer look on their variation. In the graph below, the development of the real GDP, imports and exports are shown.

0 100000 200000 300000 400000 500000 600000 700000 800000 va lu e in m ill io n eu ro

Development of real GDP, Imports and Exports

(26)

26 Tom van de Putte | UNIVERSITEIT LEIDEN

What is clear from this graph, is that the Netherlands real GDP has grown over the years, as well as the imports and exports value. The dip around 2009 is the great recession that hit the world economy then. The exports grew faster than the imports over the last fifty years.

In the next graph, the course over time is shown for ratios concerning the openness of the economy.

Graph 3 Development of Openness and Deindustrialisation

On the left y-axis is the outcome shown for the impgdp indicator (imports divided by GDP) and for the openness indicator (imports plus exports divided by GDP). On the right-hand side the degree of deindustrialisation is shown. What we can see is that the three indicators for the internationalisation of the economy all move in the same direction, but nevertheless differ in variation, especially in the years between 1975 and 1993, where deindustrialisation runs concave and openness and impgdp run convex. Deindustrialisation is something that is actually mentioned in the troonredes of that time. In 1978, the government says that it can nu longer uphold firms that do not have perspective in the future. Cabinet intended to shift its stimulation policy to companies that can cope with the international competition. It recognized in 1978 that the industrial developed countries are faced with the inevitable fact that they should produce goods and services of technological high quality.7

7_{In the troonrede in Dutch: “In de internationale arbeidsverdeling voltrekken zich belangrijke veranderingen.}

Het in stand houden van bedrijven die geen perspectief meer hebben zou die veranderingen miskennen. Daarom wil het kabinet geleidelijk overgaan van steun aan zwakke bedrijven naar stimulering van

bedrijfstakken die zijn opgewassen tegen de internationale concurrentie. De indrustieel ontwikkelde landen staan voor de noodzaak zich toe te leggen op technologisch hoogwaardige goederen en diensten. Allereerst ligt hier een taak voor de ondernemingen zelf.”

74,00 76,00 78,00 80,00 82,00 84,00 86,00 88,00 90,00 92,00 94,00 0,000 0,200 0,400 0,600 0,800 1,000 1,200 1,400 1,600 1,800 19 69 19 71 19 73 19 75 19 77 19 79 19 81 19 83 19 85 19 87 19 89 19 91 19 93 19 95 19 97 19 99 20 01 20 03 20 05 20 07 20 09 20 11 20 13 20 15 20 17 de in du st ria lis at io n ra tio im pg dp a nd o pe nn es s

Development of Openness and Deindustrialisation

(27)

In order to test the hypothesis, we should look whether the variation in the percentage of words explains the variation in the three indicators shown in graph 3. To test this, I use a simple ordinary least squares regression model. Since I use three different main explanatory independent variables, those three will be shown. The results of this model are shown in the next two tables.

In table 5 the dependent variable percwordg has been lagged, because there is good reason that it takes time to recognize the trend of globalisation. Politicians need to be aware of the direction the economy is heading. In addition, effects of migration take time to manifest.

What is remarkable is that with the lagged variable by two years more coefficients are significant and that the (adjusted) R-squared is higher. This means that the models in table 5 explain more variation than the models in table 4. Model 4, 8 and 12 are interesting since including the percentage of non-Dutch immigrants is statistically significant and eliminates the effect of the three openness variables. The embedded liberalism thesis entails that losers should be compensated and that governments do so in order to maintain support for free trade (Hays et al., 2005). In four troonredes the government only shows attention for losers of globalisation, whereas in ten years it shows only attention for winners. With regard to the tone of the texts the following can be said: in thirteen cases the government holds a negative attitude towards globalisation compared to 21 cases a positive attitude. In 15 cases this attitude is neutral. By reading the troonredes, it is clear that in the ‘70s agricultural entrepreneurs are frequently indicated as losers from globalisation. This has to do with the agricultural policy of the European Union or its predecessors. In 1972 for example, the Queen read: “To accompany the current structural changes in connection with decisions made by the European Economic Community in the agricultural sector, interest subsidies will selectively be applied.”8

In 1973 the government writes: “In order to prevent a disproportionate loss of income in the agricultural sector as a result of the revaluation, a temporary compensation will be granted in the context of a refund of value added tax”9_.

In table 6 some models regarding the tone and winners or losers are presented. In the next chapter, I interpret the results and compare them with the theories mentioned earlier in this research.

8_{Originally in Dutch: “Ter begeleiding van de huidige structurele veranderingen in de agrarische sector zullen,}

(28)

Table 4

Results OLS Regression Analysis

dep var t (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) openness 4.02* (1.93) 3.95* (1.97) 3.83 (2.73) 4.00 (2.67) impgdp 8.68* (4.21) 8.53 (4.28) 7.99 (5.82) 8.48 (5.79) deindustr 0.39 (0.24) 0.44 (0.28) 0.39 (0.63) 0.29 (0.29) umempr 0.10 (0.37) 0.12 (0.37) -0.16 (0.44) percfor 0.10 (0.93) 0.16 (0.91) -0.02 (1.76) perimnond 0.07 (4.73) 0.24 (4.70) 2.47 (4.25) constant 4.62* (1.88) 4.15 (2.58) 4.43 (2.64) 4.61 (2.03) 4.57* (1.93) 4.00 (2.63) 4.26 (2.65) 4.53 (2.06) -25.60 (20.68) -29.29 (23.30) -26.07 (48.53) -18.49 (24.17) (adjusted) R2 _0.09 _0.05 _0.05 _0.05 _0.08 _0.05 _0.04 _0.04 _0.06 _0.02 _0.01 _0.02 number of observations 48 48 48 48 48 48 48 48 48 48 48 48

(29)

Table 5

Results OLS Regression Analysis with Lagged Percentage of Words on Globalisation

dep var: (t – 2) (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) openness 5.80* (1.91) 5.83* (1.93) 5.65* (2.66) 2.20 (2.49) impgdp 12.54* (4.15) 12.57* (4.19) 12.01* (5.67) 4.79 (5.36) deindustr 0.52 (0.27) 0.66* (0.29) 0.14 (0.63) 0.00 (0.30) umempr -0.19 (0.39) -0.17 (0.39) -0.54 (0.44) percfor 0.08 (1.00) 0.14 (0.99) 1.17 (1.76) perimnond 9.26* (4.32) 9.28* (4.29) 11.87* (3.93) constant 3.02* (1.90) 4.03 (2.83) 2.84 (2.97) 1.65 (1.93) 2.95* (1.93) 3.86 (2.88) 2.64 (2.96) 1.60 (1.95) -37.48 (23.83) -46.77 (24.89) -9.25 (48.86) 2.75 (25.63) (adjusted) R2 _0.17 _0.14 _0.13 _0.22 _0.17 _0.14 _0.13 _0.22 _0.08 _0.07 _0.04 _0.20 number of observations 46 46 46 46 46 46 46 46 46 46 46 46

(30)

Table 6

Difference in Tone and Losers/ Winners

whole sample only losers

lag2 (t-2) lag1 (t-1) lag2toneglob is dependent variable lag1toneglob is dependent variable dependent variable is lag2percwordg unempr 0.30* (0.09) 0.86 (0.83) econgr -0.22* (0.08) 0.24 (2.58) openness 13.79* (3.66) 16.43* (4.34) percimnond -0.14 (7.18) electy -8.46 (9.48) constant 0.15 (0.51) 2.31* (0.23) -2.82 (4.06) -10.01 (10.72) R2 _0.20 _0.15 _0.67 _0.74 N 47 47 9 9

(31)

5 Analysis

At the end of the chapter ‘theoretical framework’ I formulated hypotheses, expectations for this research. Now, I will briefly discuss each hypothesis in light of the findings the results of this research provides. From this analysis, an answer can be formulated to the question whether there a correlation is between trade liberalisation and political attention for losers of globalisation.

The political attention for losers of globalisation derives from the embedded liberalism thesis: compensation by the government in order to maintain support for trade liberalisation. In order to compensate, there first must be political attention for this. The other vision derived from the literature is that the difference in pace in economic integration on the one hand and political integration on the other hand is out of step, causing a race to the bottom. In this chapter, I test which of the two visions has the upper hand. First, I start with analysing the three minor hypothesis. I end with testing the hypothesis that gives an answer to the central question in this research.

5.1.1.1 Hypothesis 1: The government recognizes the distinction between winners and losers from globalisation

Confirmed. This hypothesis is grounded in the quote from Frieden (2006, pp. 105-111) that “(…) even orthodox economists accept that there are both winners and losers from freer trade”. The governments recognizes in 4 troonredes the position of losers and in 10 troonredes winners from globalisation.

5.1.1.2 Hypothesis 2: The higher the unemployment rate, the more negatively formulated attention for globalisation

Confirmed. The foundations of this hypothesis lie partly in the research of Burgoon (2014) and partly in common sense. Burgoon names unemployment as a factor that plays an important role. It makes sense to think that a higher unemployment rate leads to negative expressions by a democratic government on globalisation. ‘It is the outer world which causes unemployment: the government did not cause this’, might be the thought expressed.

The answer to the question whether how the unemployment rate is related to the tone of political attention for globalisation can be found in table 6. For every 1% added unemployment, the tone becomes 0.30 more negative, since the toneglob variable varied from 0 to 4, with 0 being very positive and 4 being very negative. With a constant of 0.15, the tone without unemployment would be very positive. Since the correlation coefficient is significant, the hypothesis needs to be confirmed.

(32)

5.1.1.3 Hypothesis 3: The higher the economic growth, the more positive attention for globalisation.

Confirmed. This hypothesis relies on the logic that the higher the economic growth, the more euphoric the citizens (including the government) are about their economy and thus for the direction the economy is heading. With an economy opening up during the research period we can expect that the economic growth has an impact on the tone towards globalisation.

In table 6, the second model from left is the relationship showed between economic growth and the tone of globalisation. Since the troonredes are being read in September and being built on the economic prognosis of the coming year, table 6 reports a one year lagged variable for the tone. The correlation coefficient is significant and negative. Again, the higher the value for the tone, the more negative. From this we can deduce that 1% extra economic growth leads on average to a decrease in tone by 0.22, meaning that the higher the economic growth, the more positive the tone on globalisation. The hypothesis is confirmed.

5.1.1.4 Hypothesis 4: The opener the economy, the more political attention for losers of globalisation

Firstly, it is important to determine if the economy opens up at all. This thought is constructed with due consideration of the theory on neoliberalism in historical context. It takes no rocket science to determine on the basis of graph 2 and 3 that the economy opens up, the more time passes. Although there are different approaches used to describe the openness of the economy, all ways point in the same direction: the Dutch economy becomes opener during the period from 1969 until 2017.

Secondly, we should look in what direction all political attention develops and if this is correlated with trade liberalisation. The embedded liberalism thesis underlies this thought. Trade liberalisation leads to more trade which in turn leads to compensation from the government in order to maintain support for free trade. The will to compensate must then first be on the agenda of the policymakers. The variation of political attention must then be explained by the variation in openness. In table 4 and table 5 lies the answer to this. I earlier argued that using a lagged variable makes sense in this case, therefore making the results from table 5 more appropriate. Nevertheless, table 4 shows a significant effect of openness on the percentage of words on globalisation – embodying political attention for globalisation. The same goes for the import in relation to GDP. The determination coefficients are higher in table 5 than in table 4, which means that the models in table 5 have greater explanatory value.

In table 5, however, the effects are still significant, but now remain significant when controlling for other variables such as unemployment rate and the percentage of foreigners (no Dutch nationality). When we control for non-Dutch immigration though, the effect of openness on political attention is not significant any more. Therefore, we can conclude that the opener the economy, the more political

(33)

attention there is for globalisation, with the caveat that the variation in political attention is better explained by immigration.

How is then the relation between political attention for globalisation and immigration? Burgoon (2014) found that immigration has a strong effect on compensation by governments. In order to compensate, the government must first recognize the relationship between globalisation and immigration. Because of globalisation, it is easier to move to another country. It is not the trade an sich that makes governments show attention for globalisation, but it is the migration that interferes (Burgoon, 2014). With this research as basis, we can expect that immigration has its effect on political attention for globalisation.

I created two indicators for migration: the percentage of foreigners living in The Netherlands and the percentage of immigrants coming to the Netherlands each year. The percentage of foreigners is in no model significant and the openness variable remains significant. How different with the indicator percentage of non-Dutch immigrants. The incorporation of this variable into a model leads to the highest adjusted determination coefficients from all models and, most importantly, as I wrote earlier, it eliminates the formerly significant effect of the openness variables. The conclusion must be that it is not the openness in terms of trade which contributes the most to political attention for globalisation, but immigration that does. The expectation that the more immigration, the more political attention for globalisation is therefore confirmed.

Thirdly, we have to look whether the import relative to the GDP is a better explanatory factor for political attention for globalisation than the openness indicator that also includes the exports. I can be short about this: this is not the case. The thought that only the imports are a better explanatory factor is deduced from the research of Hays et al. (2005), where they argue that imports create losers and exports create winners from globalisation. Since policies try to solve problems and government creates policies, the government may have more attention for losers and thus imports are a better explanatory factor than the openness indicator which also uses the exports in the calculation.

In this research, I found that both indicators of an open economy have a significant effect on the political attention in the simple models 1, 2, 3 and 5, 6, 7. The correlation coefficient is higher with the imports only indicator, but this says nothing about the explanatory power of the models. In order to assess the explanatory value, we must look at the determination coefficient. The (adjusted) determination coefficient (R2_{) is for both of the indicators the same in each model (see table 5). Hence,}

(34)

welfare state. Deindustrialisation would be mainly the result of domestic factors. Hence I expect that deindustrialisation and political attention for globalisation are not related since globalisation and deindustrialisation are not related according to their inquiry.

In graph 3 can be seen that the deindustrialisation indicator differs in its gradient from the other two variables but the general trend is the same: upwards. If the variation of the deindustrialisation explains the variation in political attention for globalisation can be seen in table 5. From the four models with deindustrialisation, only model 10, which also incorporates the unemployment rate – is statistically significant. This model has an adjusted R2_{of 0.07. The correlation coefficient is also very low. The}

simplest model (model 9) does not give a significant effect. All together this does not indicate a relationship between deindustrialisation and political attention for globalisation.

Lastly, now that disturbance factors are investigated, I come to analyse the political attention of losers in relation to trade liberalisation. Globalisation has winners and losers (Frieden, 2006). The compensation hypothesis and the embedded liberalism thesis are also about losers. Losers must be compensated in order to maintain support for free trade, or because the people ask for it themselves. The answer to either confirm or reject this hypothesis can be extracted from table 6, the two models at the right side of the table. If only the years when losers are being mentioned are taken into account, the correlation coefficient of openness on the two year lagged percentage of words on globalisation is 13.79 and significant. The determination coefficient is 0.67, which means that in this case the model explains 67% of the variation. The number of observations is low, though. The simple model thus confirms the hypothesis.

The other model takes also control variables into account. The determination coefficient is even higher (0.74) than in the simple model. The other remarkable thing is that the openness variable is still statistically significant, while the controls are not. The hypothesis can be confirmed, but with one major caveat.

What is proven here, is that if we only take political attention for losers into account, that there is a strong relationship between the openness of the economy and this attention. It does not mean that opening up the economy leads to more annual specific attention for specific groups of losers from globalisation. This suggests that theories from authors who wrote that the race to the bottom effect is eliminated by compensation are more likely to be true. In the next chapter, I elaborate on this.

(35)

6 Conclusion

In this research I tried to give an answer to the question to what extent trade liberalisation lead to political attention for losers of globalisation. Furthermore I investigated the effect of the unemployment rate, economic growth, migration and deindustrialisation by using several regression models with political attention for globalisation as the dependent variable. The dependent variable has been operationalised into the indicator percentage of words on globalisation.

The general answer to the main question must be that an opener the economy leads to more political attention for losers, if they are mentioned. If all years are included, also the years where there is no record of losers from globalisation, the variation in openness of the economy explains the variation in political attention, except for those models that include the annual migration in percentage of the total population. It is thus not the openness of the economy that is the most important factor in explaining political attention, but migration, although migration and trade are also related to each other. The mechanism between migration and trade openness however, can be part of future research.

The effects must also be placed into perspective. In case of all years, when openness goes up by 1, the percentage of words goes up by 5.80 percent point. Openness is a variable with in the used dataset a minimum of 0.50 and a maximum of 1.62 during 48 years. If only the years with explicitly mentioned losers where the correlation coefficient is 16.43 are taken into account, then the effect is stronger, but this says nothing on if the mentioning of losers becomes more frequent when time passes. Only that when losers are mentioned, the political attention correlates with the openness.

What this research confirmed is that a high economic growth leads to a positive attitude of the government towards globalisation and that a high unemployment rate does the opposite. Another finding is that deindustrialisation and political attention for globalisation are not related. Since openness does, we can conclude that deindustrialisation is (also) domestically driven. This is consistent with the theory of Iversen and Cusack (2000).

This research also contains weak spots. I will first address the internal validity. If the database were larger, the results would be more valid. This data is unfortunately not available since we cannot look into the future. But the internal validity is thin if you take into account that the construction of the models with only the years mentioning losers used nine observations.

With regard to the reliability, there is a weak spot in coding the troonredes, the annual speeches given by the King on behalf of the whole government. Some sentences are very clear in their relationship with globalisation, but there is always a grey area. The data from Statistics Netherlands on the contrary

(36)

Overall, this research provided a useful glimpse of the correlation underneath the embedded liberalism thesis. Further research in relation to migration is desirable, because of the disturbance effect of it which popped up in this research. It puts a bomb under the thesis that it is trade of goods and services as well as capital mobility which leads to compensation by the government. Or it implicates that the government is compensating for the wrong reasons. Immigration is likely to be a key factor for support for free trade.

(37)

7 References

7.1 Books, Articles en Websites

Allan, J. P., & Scruggs, L. (2004). Political Partisanship and Welfare State Reform in Advanced Industrial Societies. American Journal of Political Science, 48(3), 496-512.

ANP. (2018, July 6). Minister Kaag: 'Handelsoorlog kent alleen maar verliezers'. Retrieved July 17, 2018, from nu.nl: https://www.nu.nl/economie/5350082/minister-kaag-handelsoorlog-kent-alleen-maar-verliezers.html

Baldwin, R., & Wyplosz, C. (2012). The Economics of European Integration (4th ed.). Maidenhead: Mc Graw Hill Education.

Bivens, J., & Lawrence, M. (2015). Understanding the Historic Divergence Between Productivity and a Typical Worker's Pay. Washington DC: Economic Policy Institute.

Breeman, G., Lowery, D., Poppelaars, C., Resodihardjo, S. L., Timmermans, A., & Vries, J. d. (2009). Political Attention in a Coalition System: Analysing Queen's Speeches in the Netherlands 1945-2007. Acta Policitca, 44(1), 1-27.

Burgoon, B. (2001). Globalization and Welfare Compensation: Disentangling the Ties that Bind. International Organization, 55(3), 509-551.

Burgoon, B. (2014). Immigration, Integration and the Support for Redistribution in Europe. World Politics, 66(3), 365-405.

Dreher, A. (2006). Does Globalization Affect Growth? Evidence from a new Index of Globalization. Applied Economics, 38(10), 1091-1110.

ETH Zürich KOF Swiss Economic Institute. (2017, April 20). KOF Index of Globalization 2017:

Netherlands are Most Globalized Country. Retrieved July 11, 2018, from KOF Swiss Economic Institute: https://www.kof.ethz.ch/en/news-and-events/media/press-releases/2017/04/kof-globalization-index-2017.html

Frieden, J. (2006). Global Capitalism: Its Fall and Rise in the Twentieth Century. New York: W.W. Norton & Company.

Genschel, P., Kemmerling, A., & Seils, E. (2011). Accelerating Downhill: How the EU Shapes Corporate Tax Competition in the Single Market. Journal of Common Market Studies, 49(3), 585-606. Harvey, D. (2006). Neo-Liberalism as Creative Destruction. Geografiska Annaler. Series B, Human

Geography, 88(2), 145-158.

Hays, J. C., Ehrlich, S. D., & Peinhardt, C. (2005). Government Spending and Public Support for Trade in the OECD: An Empirical Test of the Embedded Liberalism Thesis. International

Organisation, 59(2), 473-494.

Held, D. (1991). Democracy and Globalization. Alternatives, 16(2), 201-208.

Helmore, E. (2018, May 31). US to hit EU with steel and aluminum tariffs, report says. Retrieved July 17, 2018, from The Guardian: https://www.theguardian.com/us-news/2018/may/30/trump-administration-tariff-steel-aluminum-eu-trade-tension