The effect of the Brexit Referendum on the value of real estate investment companies in the European Union

(1)

The effect of the Brexit Referendum on the value of real

estate investment companies in the European Union

Amsterdam Business School Master Thesis

Name Ralph de Bruijne

Student number 10334367

Program Master in Finance

Specialization Real Estate Finance & Corporate Finance Number of ECTS 15

Supervisor G. Marcato Completion July 2018

Abstract

This research investigated the impact of the Brexit Referendum on the value of real estate investment companies in the UK and the EU. In order to assess whether the Brexit Referendum imposed different effects between EU member states due to potential relocations from the UK to the EU, event studies are performed on a relocating sample and a non-relocating sample. The relocating sample consists of companies from EU member states which are popular potential relocating locations. The non-relocating sample represents companies from the other EU member states. This study uses all publicly listed REITs and REOCs in the UK and in the remaining 27 EU member states. In addition, the role of foreign exchange exposure on stock returns during the Brexit Referendum outcomes is analysed. The results show that the Brexit Referendum imposed a significantly negative impact on companies incorporated in the UK. For EU incorporated companies no significant results have been found. In addition, no significant difference is observed for the potential relocating countries relative to the non-relocating countries. With respect to the role of foreign exchange exposure, the results show that geographic investment allocation and shareholder nationality significantly impacted the returns of real estate investment companies following the Brexit Referendum.

(2)

2

Statement of originality

This document is written by Ralph de Bruijne who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(4)

4

List of Figures

Figure 1 Timeline Events included in this Research Figure 2 Timeline around the Event

Figure 3 Cumulative Average Abnormal Returns per Sample Figure 4 Map of the EU and Eurozone

(5)

5

List of Tables

Table 1 Sector Composition of Real Estate Investment Companies Table 2 Sector Composition Reduced Sample

Table 3 Average Abnormal Returns for the UK Sample Table 4 Cumulative Average Abnormal Returns UK Sample

Table 5 Average Abnormal Returns for the Relocating and Non-relocating Samples

Table 6 Cumulative Average Abnormal Returns for the Relocating and Non-relocating Samples Table 7 Average Abnormal Returns for the UK Sample with a Real Estate Index

Table 8 Cumulative Average Abnormal Returns for all Samples with Real Estate Indices

Table 9 Average Abnormal Returns for the Relocating and Non-Relocating Samples with Real Estate Indices

Table 10 Cumulative Average Abnormal Returns for the Total Sample Table 11 Cumulative Average Abnormal Returns per Country

Table 12 Regression Output of Regressions Performed for the UK Sample

Table 13 Regression Output of Regressions Performed for the Relocating Sample Table 14 Regression Output of Regressions Performed for the Non-relocating Sample Table 14 Regression Output of Regressions Performed for the Total Sample

Table 15 Regression Output of Regressions Performed for the UK Sample

Table 16 Regression Output of Regressions Performed for the Relocating Sample (robustness check) Table 17 Regression Output of Regressions Performed for the Non-relocating Sample (robustness

check)

(6)

6

List of Abbreviations

Abnormal return = AR

Abnormal average return = AAR

Centre for Economic Performance = CEP European Communities = EC

European Economic Area = EEA

European Coal and Steel Community = ECSC European Union = EU

Normal return = NR Return = R

Real Estate Investment Trust = REIT Real Estate Operating Company = REOC SIC = Standard Industry Classification

GICS = Global Industry Classification Standard United Kingdom = UK

(7)

7

1. Introduction

On Thursday the 23rd_{of June 2016, a referendum was held in the United Kingdom (UK) regarding the}

position of the UK in the European Union (EU). By means of this referendum the UK decided whether or not to leave the EU. The result of the referendum was a win of 51.9 percent to 48.1 percent for leaving the EU. Resulting in a withdrawal by the UK on the 29th_{of March 2019. By submitting Article}

50 on March 29th_{2017 Theresa May notified the EU at the 29}th_{of March 2017, the UK will be leaving}

the EU. Negotiations regarding the terms and conditions under which the UK will depart started on June 19th_{2017. The first agreement regarding the separation conditions was reached on the 8}th_of

December 2017. In March 2018, the European Union negotiator Michel Barnier and the British minister for the Brexit David Davis came to an agreement with respect to the transition period. During this period, the current rules and laws regarding the Customs Union and the internal market will continue to apply for the United Kingdom. This transition period will be until the first of January 2021.

Looking at the demographics factors which led to this result NatCen Social Research (2016) find that persons with the least economic resources, being nationalistic and against immigration identified their selves as the working class dominated the leave side. The most important issues which influenced people’s votes during the Brexit Referendum were the economy, immigration and sovereignty/EU bureaucracy. NatCen Social Research (2016) found that the British inhabitants whose main concerns were education, poverty and inequality, and the economy were significantly more likely to vote for remaining in the EU. On the other hand, the main concern for the leaving party was immigration. Another factor which contributed to a win of the leave party was a higher turnout, and especially ‘new’ voters. Compared to the 2015 general elections the turnout for the Brexit Referendum was 6.1 percent higher. The majority of this new group voted for leaving the EU. In combination with the people who leant towards voting remain being more likely not to vote this pushed the outcomes in the leave direction.

The results of the Brexit Referendum induced high levels of uncertainty on European stock markets. Aristeidis & Elias (2018) found a decrease of 9.1 percent in value for the London Stock Exchange during the first ten minutes of trading after the outcomes were announced. The first trading day after the outcomes announcements let to a decrease in value of 3 percent. Besides the effect of political uncertainty an intended withdrawal from the EU led to uncertainty regarding the access to the Single Market. Being a member of the EU the UK has access to the single market and was a member of the European Union Customs Union. This Single Market is characterized by the four freedoms. The four freedoms guarantee the free movement of goods, capital, services and

(8)

8

labour. The four freedoms were introduced in order to increase economic integration. The Customs Union entails being able to trade with other members of the Customs Union without any imposed trade tariffs. The intended withdrawal of the EU created uncertainty regarding the bilateral trade agreements between the UK and the EU. Depending on the outcomes of the Brexit negotiations the Brexit Referendum is expected to impose an effect on im and export for the UK and the EU.

One of the main reasons international companies have located their European offices in the UK is due to the accessibility of Europe through the Single Market. EY (2017) states that the UK has historically been the most popular destination for foreign direct investments. EY (2017) found a record of foreign direct investments in the UK of 20.9% of the total European foreign direct investments in 2015. These investments in the UK created more than 40,000 jobs. Any change in the accessibility to the Single Market imposes implications for international companies located in the United Kingdom. Centre for Economic Performance (2016) believe foreign direct investments will decrease in the UK. Their research lists three reasons leading to the decrease in foreign direct investments. These reasons include: loss of access to the Single Market, different legislation and regulation between countries will increase the costs and the final reason is due to postponements because of Brexit induces uncertainty. EY (2017) found by means of a survey that 14 percent of international companies with UK based activities are planning to transfer their activities partially or completely elsewhere.

Economic indicators are fundamental drivers for real estate pricing. Multiple studies such as Case et al (1999), Green (1997), Gholipour et al (2014) and Kong et al (2015) found relationships between real estate pricing and economic indicators. These indicators comprise among others GDP and economic growth. The combination of changes in foreign direct investments, economic effects and possible relocations could have imposed an effect on real estate prices.

Previous studies by Ramiah et al (2016), Aristeidis & Elias (2018), Sanitas et al (2018), Crafts (2016) and Sampson (2017) investigated the expected effect of the Brexit on the economy. Their findings show a negative effect on the economy as a result of the Brexit Referendum outcomes. However, to the best of my knowledge no studies have been performed on the effects of the Brexit Referendum on real estate prices through the EU.

Therefore, this research is going to test the effects of the Brexit Referendum on real estate prices in the EU. This will be tested my means of an event study on stock prices of real estate investment trusts and real estate operating companies from EU member states. In addition, in order to test for a potential relocating effect due to the loss of access to the Single Market for UK located companies separate studies will be performed for the UK, the potential relocating countries and non-relocating countries. As a robustness check, the event study will be performed on multiple

(9)

9

event dates in the Brexit Referendum process and for three event windows in order to increase the validity of the results. On top of that the event studies will be performed with equity and real estate indices. In conclusion by means of a linear regression on the cumulative abnormal returns the role of foreign exchange exposure on prices during the Brexit Referendum will be researched. Variables included in the regression comprise of geographic investment allocation, shareholder nationality, sector specialization, market capitalization, liquidity ratios and historical beta.

This research contributes to existing literature since to the best of my knowledge no research on the effect of the Brexit for real estate prices across Europe have been conducted so far. Therefore, besides measuring the effect of political uncertainty, this research potentially provides international investors of a better understanding of the convergence of real estate prices and political instability in the EU.

This research will be structured in the following way: section 2 comprises the literature review which provides background information on the EU, the Brexit Referendum and the effect on the economy and previous research referenda. This section will conclude with the hypothesis development. Following the literature review section 3 entails the methodology used in this research. Section 4 provides information regarding the data used in this research. The results will be displayed and analysed in section 5 after which section 6 entails the robustness checks followed by the conclusion.

(10)

10

2. Literature review

This section provides background information regarding the European Union, the Brexit Referendum and previous referenda with similar characteristics. Furthermore, this section provides an overview of previous research regarding the imposed effect of the Brexit Referendum on the economy of the UK.

2.1 European Union

The European Union (EU) is an economic and political cooperation between 28 European countries. These countries authorized the EU to conduct collective policy. The EU conducts policy on different sectors such as: agriculture, fishing, consumer rights, the environment, immigration and the fight against terrorism. The EU member states cooperate on defence, foreign policy, trade, internal market and economic policy. The foundation of the EU lies in the aftermath of World War II. In the aftermath of World War II Belgium, West Germany, France, Italy, Luxembourg and the Netherlands established the European Coal and Steel Community (ECSC). The aim of the ECSC was to have a regulatory authority to regulate the production of steel and coal. Due to the success of the ECSC the same six countries decided to cooperate on more sectors such as atomic energy and the economy. This led to the establishment of the European Communities (EC). This was the beginning of the EU as we know it today. In 1973 the UK joined the EC together with Denmark and Ireland. In 2002 twelve member states decided to introduce a common currency in order to promote economic integration. This led to the introduction of the Eurozone which currently consists of 19 European Union member states which use the euro as currency.

The European Union is treated as one Single Market and a Customs Union. This Single Market is characterized by the four freedoms. The four freedoms guarantee the free movement of goods, capital, services and labour. The four freedoms are introduced in order to increase economic integration. The Customs Union entails a common external tariff on goods and services imported to the EU. According to Crafts (2016) joining the EU by the UK led to an increase in GDP compared to the situation in which the UK would have stayed in the EFTA. The increase in GDP from joining the EU has been estimated by around 10 percent. With memberships fee’s being on average 1.5 percent of GDP, the benefits outweigh the cost.

2.2 The United Kingdom European Union membership referendum

The United Kingdom European Union membership referendum commonly referred to as the Brexit Referendum is a non-binding referendum which took place on June 23 2016 in the UK and Gibraltar.

(11)

11

The aim of this referendum was to indicate the level of support among the British people for remaining a part of or leaving the EU.

Under pressure of Members of the UK parliament and the rise of Eurosceptic political party UKIP Prime Minister Cameron announced that in case of a win for his Conservative Party during the general elections in 2015 a referendum would be held regarding EU membership. On February 22 2016 Prime Minister Cameron announced to the House of Commons that a referendum would be held on the 23rd_{of June 2016. Hunt and Wheeler (2018) summarize several reasons why the leave}

campaign wanted the UK to leave the EU. The main reasons were: they felt the EU was holding the UK back due to too many rules, the membership fees were too high for what the UK received in return and instead of being forced to created legislation in consensus with other EU nations they wanted the UK to be able to impose their own laws again. Another main issue was immigration. The leave campaign wanted to decrease the number of immigrants by regaining control of the border. Due to free movement of among others labour in the EU, inhabitants from EU countries were able to live and work in the UK without restrictions. According to Hunt and Wheeler (2018) the reasons why the remain campaign wanted to stay in the EU was because they believed the UK economy received a big boost due to their membership. Import and export from and to EU countries was easier and the inflow of young immigrants increased economic growth and payed for public services. On top of that they believed the UK would be more secure when combining strength with 27 other EU members rather than operating alone.

The Brexit Referendum was held on June 23 2016 after which the outcomes were announced on June 24 2016. The results showed a win for the leave campaign of 51.9 percent to 48.1 percent (European Union committee, 2018). Becker et al (2017) found that relative to the remain side; the leave side was dominated by population with lower levels of education, higher unemployment and lower income. In addition, the results showed geographical segmentation. Urban areas showed a higher level of support for remaining in the EU opposed to rural areas. After the win for the leaving campaign Prime Minister David Cameron announced he was resigning after which Theresa May took over.

On March 29 2017 Prime Minister Theresa May notified the EU the UK will be leaving the EU by triggering Article 50. From that moment a two-year period began during which the UK and EU should come to an agreement regarding the terms and conditions of the separation referred to as the Brexit negotiations (Hunt & Wheeler 2018). On March 19 2018 UK and EU negotiators agreed upon a transition period until the 1st_{of January 2021. During this transitions period the current}

legislation regarding free movement will stay in place and the UK will get the possibility to arrange their own trade deals which would start when the transition period expires. According to Morris

(12)

12

(2017) in case a deal would not have been reached regarding a transition period there would have been consequences for almost every aspect of daily life. Without an agreement, the rules of the World Trade Organization would come in place which would lead to imposed tariffs on import and export between the UK and the EU ranging from 2 to 3 percent for industrial products up to 20 to 40 percent on agricultural products. The services sector in the UK would be hit hard as well. Without a deal, UK business will lose their passporting rights. Without passporting rights companies are not allowed to sell their services in EU member states, without having licenses in all the member states separately. Looking at regulations the UK would lose their membership to a number of EU bodies in industries such as the aviation and pharmaceutical industry.

2.3 Effect on UK economy

Previous research by Liu et al (2017) and Bachetta & van Wincoop (2012) investigated the effect of political events on asset prices. Li et al (2017) investigated the effect of the Bo Xilai scandal on the performance of Chinese stocks. They state that increased political uncertainty will lead to a drop in stock prices. According to Li et al (2017) the Bo Xilai scandal imposed a huge impact on political stability in China. Their results showed a significant negative return on stock prices. Companies with higher sensitivity levels to political changes were affected more severe. Their findings show political uncertainty being harmful for company’s value. Bachetta & van Wincoop (2012) investigated the relation between volatility and sharp drop in equity prices around the world during several events. The research comprised of the global financial crisis in the fall of 2008, the Greek sovereign debt crisis in May 2010 and the U.S. debt debate and intensifying European debt crisis in July/August 2011. They found that the uncertainty created by these events was felt all over the world. These events led to significant decreases in equity prices.

Research by Ramiah et al (2016), Schiereck et al (2016), Sanitas et al (2018), Sampson (2017) and Cambridge Econometrics (2018) researched the expected effect of the Brexit on the UK economy. Ramiah et al (2016) studied the effect of the Brexit on the stock markets per sector. Their main results show that the banking sector and travel and leisure sectors were hit the hardest by the Brexit. Ramiah et al (2016) also researched the real estate sector. Their results show a 6.05 percent negative abnormal return for real estate investment trusts and a 3.65 percent negative abnormal return for the real estate investment and services sector.

Schiereck et al (2016) found that the Brexit announcement had a significant effect for financial institutions across the EU. Schiereck et al (2017) tested the reaction of UK banks, non-UK EU banks and international banks with major EU operations in the UK. Their results show a negative significant effect on stock prices for all three types of banks. However, the reaction is significantly

(13)

13

less negative for banks which have their main EU operations located in the UK.

Samitas et al (2018) used 1000 simulations for both the pessimistic and optimistic scenario in order to measure the long- and short-term effect of the Brexit on the economies of both the EU and the UK. They find that the effect of the Brexit would be less substantial for the UK’s economy relative to the economy of the EU. However, the effects of the Brexit are expected to hit the UK much sooner than the EU.

Sampson (2017) states that due to the openness which characterizes the economy of the UK and due to the high degree of trade with the EU the Brexit will exert a negative effect on the economy of the UK. The amount of export and import within the European Union accounted for 44% and 53% of all ex and import.

Cambridge Econometrics (2018) modelled the effects of the Brexit on the UK on several parameters depending on different scenarios how the Brexit deal would look like. Scenarios are ranging from keeping the status quo with the UK remaining in the Single Market to a scenario where the UK has no membership in the European Economic Area (EEA) or Customs Union and no transition period. The modelled parameters include: export to rest of the world, import from rest of the world, population, gross value added, investment, employment and productivity. The results of Cambridge Econometrics (2018) show negative coefficients for all parameters in all scenarios relative to the status quo scenario.

Centre for Economic Performance (CEP) (2016) researched the effect the Brexit would impose on the foreign direct investments in the UK. CEP (2016) state that foreign direct investments convey direct and indirect benefits for countries. The direct benefits are due to increased productivity and because foreign firms typically pay higher ways. The indirect benefits are created because foreign direct investment increases competiveness which makes domestic companies increase their performance. On top of that foreign companies bring along new technologies and managerial knowledge which could be incorporated by domestic companies. CEP (2016) lists three causes of possible decrease in foreign direct investment in the UK after the Brexit. Due to being a part of the Single Market, international trade from the UK to the EU does not bring along large costs due to trade tariffs. Another reason why foreign direct investment in the UK could stagnate after the Brexit is due to different legislation. For multinationals with locations across Europe it would become more costly and difficult to coordinate their supply chains due to different legislation and regulation between the UK and the EU. The last reason why Brexit might lead to decrease foreign direct investment according to CEP (2016) is because of uncertainty. Since there is no agreement reached regarding the terms and condition concerning international trade, companies will be inclined to postpone or cancel foreign direct investment in the UK. Empirical results by CEP (2016) expect as a

(14)

14

result of the Brexit foreign direct investment inflow will decrease with approximately 22 percent for the UK. The cost of the decrease of foreign direct investment will have a negative effect on productivity and real incomes will decrease by 3.4 percent.

In contradiction to the beliefs of CEP (2016), Capital Economics (2016) expects that foreign direct investments will not decrease as much since they believe access to the Single Market is not the single reason companies invest in the UK. In addition, Capital Economics (2016) state that the expectations of a decrease in foreign direct investment in the UK is based on the situation where the UK does not have any trade agreements with the EU anymore. In the situation where the UK and EU sign trade agreements the UK could still function as a gateway to the EU and therefore the Single Market. On top of that other reasons why companies invest in the UK is because the UK has a good name in attaining credit, protection of minority investors and the English language (Capital Economics, 2016).

Li & Chevapatrakul (2016) investigated the interdependence between stock markets during the Brexit. Their research was focused on the UK, France, Germany and Switzerland. Looking at the period between January 2016 and September 2016 their results show that France and Germany were the net transmitters of volatilities during the Brexit while the UK and Switzerland received the transmitted volatilities. However, when looking at the day of the referendum their results shows evidence that the UK and France were the main transmitters of volatility to other countries. Their findings show that the levels of volatility spillover are significant among the research countries.

Real estate performance is related to several economic indicators. This relationship is found by studies such as: Case et al (1999), Green (1997), Gholipour et al (2014) and Kong et al (2015). Case et al (1999) researched the correlation between international real estate markets and GDP. Case et al (1999) found a significant relation between GDP and real estate returns. Green (1997) found a unidirectional relationship between housing investments and economic growth in the United States from 1959 to 1992. Research by Gholipour et al (2014) found a positive causal relationship between property price increases and economic growth in both the long and short run. Kong et al (2015) investigated the relation between real estate investments and economic growth in China. They find that on a national level for either housing investments, investments in commercial real estate and for other investments a positive significant relationship. Investments in housing showed the highest effect on economic growth. With the expectation that the Brexit will impose an effect on the economy of both the UK and EU member states and due to the relation between economic performance and real estate performance, a significant effect on real estate returns could be expected.

(15)

15

Additionally, real estate performance is affected by the portfolio type. Anderson et al (2012) found a positive relationship by portfolio composition and stock return. Their results show that diversified portfolios show higher returns compared to property type specialized portfolios. The reason behind this is because diversified portfolios are better protected against property-type specific risk and have more investment opportunities, which let them have more opportunities to the best performing properties.

CBRE (2018), PWC (2016) and Gerald Eve LLP (2017) looked at the potential consequences for the real estate sector caused by the Brexit. CBRE (2018) expects the economic impact of the Brexit to be uncertain which will lead to a decrease in confidence and economic growth and a decrease in consumer spending due to currency-induced inflation. However, the decrease in value of the Pound Sterling led to increased international real estate investments and tourism which led to increased export and demand for offices, retail, industrial and hotel real estate. With respect to the impact on the labour market CBRE (2018) expects a fall in net migration and a lower amount of hired employees. This will lead to a decrease in demand for office space.

PWC (2016) however states that the Brexit Referendum did not impose a significant impact on the real estate sector in the short term. PWC (2016) explains the absence of an effect to the long-term nature of real estate investments. In combination with the long-long-term nature of investments PWC (2016) found that transactions continued to be carried out after the Brexit Referendum, due to pre-agreed clauses concerning discounts in case of a Brexit. Therefore, there was no sudden stop in real estate transactions which contributed to an insignificant short-term effect. On top of that PWC (2016) explains the insignificant effect for real estate by the continuation of rents. Due to the longer duration of lease terms a decrease in profitability is not expected for the majority of the properties. However, for the longer-term property prices could become under pressure to the increase of yields do to the uncertainty created by the Brexit. In line with EY (2017) they believe that the relocation of companies could imply a significant effect. This effect is initiated due to changes in supply and demand.

Gerald Eve LLP (2017) claims that due to a weaker Pound Sterling which attracted overseas investment, the yields have remained at low levels, wherefore the Brexit did not impose a major effect for the UK real estate sector.

2.4 Previous researched referenda

Looking at historic referenda, two referenda are held which show similarities with the Brexit Referendum. The Swiss federal popular initiative against mass immigration; in short, the Swiss immigration referendum in 2014 and the Quebec referendum in 1995. The Swiss immigration

(16)

16

referendum was held in order to provide a decision regarding limited immigration through quotas. The Quebec referendum took place in 1995 and asked the inhabitants of Quebec whether Quebec should proclaim national sovereignty and therefore separate itself from the Canadian Federation.

Hanke (2017) investigated the effect of the Swiss immigration referendum on the value of Swiss companies. Hanke (2016) investigated the effect of sudden restrictions on immigration on firms with high shares of immigrants by the use of an event study. According to Hanke (2016) less immigration restrictions are good for business and investment environment. In 2002 an agreement between the European Union and Switzerland became in force regarding quasi free movement of EU and Swiss citizens. The bilateral agreement between the European Union and Switzerland are mutually dependent and know a ‘guillotine clause’. This means that whenever one treaty is not accepted by a party the agreement is terminated. According to Hanke (2016) new legislation with respect to immigration is usually a process with a long duration which leads to a gradual effect on value of companies. However, since this decision was taken by means of a referendum the outcome came as a surprise and capital shifted immediately. The referendum in Switzerland shows similarities with the Brexit with respect to the effect on the openness of their economies. In both events the results of the referendum would decide to either keeping their bilateral agreements with the EU or staying in the EU. Hanke’s (2016) results show a negative effect of the referendum on firms which employ more immigrants. These firms faced negative abnormal returns and their returns were lower compared to firms who employ fewer immigrants.

Beaulieu et al. (2015) studied the effects of the Quebec referendum in 1995 on the value of Quebec companies. According to Beaulieu et al. (2015) exit polls 23 days in advance of the referendum did not show a clear expectation regarding the outcome. Therefore, the Quebec Referendum has in common with the Brexit Referendum that there was not a clear expected outcome prior to the referendum. Due to which financial markets could not act on the political instability before the outcomes were released. With an outcome of 48.1 percent against 51.9 percent the Brexit Referendum did not show a clear result either till the final outcomes was announced. Another similarity between both referenda is regarding the separation of the particular state or province from the EU or the Canadian Federation respectively. Beaulieu et al’s (2015) results show a positive significant effect on stock prices of Quebec’s companies. These results imply that remaining in the Canadian federation was perceived as good news by financial markets. The effect on returns is however smaller for companies which are less exposed to political risk such as multinationals compared to domestic companies for whom political uncertainty holds a bigger impact.

(17)

17

2.5 Hypothesis development

This section provides an overview of previous research and the fundamentals which have led to the development of the tested hypotheses.

Several studies show that an effect on capital markets could be expected due to the Brexit Referendum. For example, Aristeidis & Elias (2018) state that the day after the referendum results were announced the London Stock Exchange market fell 9.1% during the first ten minutes of trading. On the first trading day, the London Stock Exchange closed with a decrease of three percent. In addition, the value of the Pound Sterling decreased to a 31-year time low level compared to the United States dollar. Burdekin et al (2017) researched the effect of the Brexit on global equity markets. Their findings show a negative effect of the referendum on global stock markets. The Eurozone suffered the hardest, with the PIIGS countries the heaviest. PIIGS countries are weaker countries with high debt levels and include Portugal, Ireland, Italy Greece and Spain. Burdekin et al (2017) found the countries suffering the least of the referendum to be fiscally sound open economies outside the EU.

Monfared & Pavlov (2017) researched the effect of the Brexit on residential real estate prices in different areas in London. Since London is the home to thousands of EU members which were left in uncertainty after the potential threat of having to leave the UK due to changes in the free movement of labour between the EU and the UK. Monfared & Pavlov (2017) researched the effect of the Brexit on real estate prices for areas in London which contain a high portion of EU members in their population. Their results show a drop in value in the researched areas ranging from 1.9 percent to 3 percent in the four months after the Brexit Referendum.

Okunev & Wilson (1997) found evidence for a non-linear relationship between the stock market and direct real estate. The pricing of direct real estate differs from the pricing in the stock market due to infrequent sales. Where stocks and bonds are traded on a continuous basis, real estate transactions occur a lot less frequent. Therefore, political events are not priced in directly in real estate prices (Geltner, 2014). The way to research the effect of events on real estate prices is by focusing on real estate investment companies. This research will be focusing on real estate investment trust (REITs) and real estate operating companies (REOCs). REITs and REOCs are companies which trade on a stock exchange and own or finances income producing real estate. This research focusses on equity REITS and equity REOCs, which only invest in properties and not in mortgages. The difference between a REIT and a REOC is that REOCs reinvest their earnings into their business opposed to REITs which distribute their earnings to their shareholders. In this research the terminology used for REITs and REOCs will be real estate investment companies. Liow and Yeo

(18)

18

(2018) found that real estate investment companies show a higher volatility compared to other financial instruments like stocks due to the combination of investment characteristics. Real estate investment companies combine the characteristics of investing in stocks and investing in direct real estate. As stated in the literature review the Brexit imposed an effect for the UK economy. Since economic circumstances are a determinant for real estate prices it can be assumed that the real estate investment markets are affected by the Brexit. However, since there are opposing views regarding the effect of the Brexit, this research test empirically effect of the Brexit on the value of real estate. In this research real estate investment companies represent the value of real estate. By combining the effect on stock markets and the relationship between the stock market and real estate due to the Brexit the following hypothesis can be derived:

H1: The Brexit Referendum has a significant negative impact on real estate investment companies in the United Kingdom

As mentioned in the literature review, losing access to the Single Market will impose an effect on the UK economy. CEP (2016) states that one of the reasons international companies located their European activities in the UK was due to the Single Market. According to CEP (2016) being located in a member of the Single Market is beneficial for companies since they do not face large cost from trade tariffs on their exports to the EU. CEP (2016) believes losing access to the Single Market will decrease foreign direct investments in the UK. CEP (2016) expects foreign direct investments will decrease after the Brexit due to the loss of access to the Single Market. On top of that international companies which are currently located in the UK are thinking of relocating their business in order to keep access to the Single Market. The Independent (2018) summarized some key announcements of companies regarding a relocation since the Brexit Referendum. These include among others: HSBC, Lloyd’s of London, Barclays, Bank of America, Moneygram, European Medicines Agency and European Banking Authority are moving their headquarters from the United Kingdom. On top of that JPMorgan and UBS moved hundreds and thousands of jobs respectively from London

Possible relocation could impose a negative effect on the UK economy. At the same time this could impose positive effects for the countries to which companies relocate since foreign direct investments increase. Foreign direct investments will increase the demand for real estate in the region surrounding the investment. In 2016 after the referendum EY (2017) conducted a survey on representatives of companies with foreign direct investments in Europe regarding where companies might relocate their operations to from the UK. The results show that the countries mentioned the most as a possible relocating location are: Germany, the Netherlands, France, Italy and Spain. Therefore, it could be expected that the effect due to the potential relocation of companies will

(19)

19

more positive in these countries compared to other European Union member states. Deloitte (2018) agrees with the potential positive effect for EU member states due to relocation. Deloitte (2018) states that multiple countries across Europe have been able to profit from the Brexit due to the relocation from the UK to EU member countries. However, this had not led to a downturn in the UK market. Although for the coming years the London real estate market will not show growth. The reason behind this is because companies are hesitating to undertake actions due to uncertainty as a result of the Brexit negotiations.

The Department for International Trade (2017) states that trade is one of the fundamentals of economic growth and activity for the UK and the world economy. For the UK international trade makes up for around 60 percent of GDP in 2016. According to the Department for International Trade (2017) international trade affects the economy of the UK positively through the increase in jobs, wages and innovation. On top of that international trade increases the availability of more products at lower costs. In 2016 the total amount of im and exports combined between the EU and UK was 553 billion Pound Sterling making the EU the main trade partner of the UK. Walker (2017) researched the main trade partners of the UK based on the im and exports of goods. In the top ten of trading partners for im and exports, the five countries which are mentioned as relocating locations are represented. Therefore, the effect of relocating could be affected by the expected negative effect of a possible loss in im and export for companies based in those countries. On top of that Burdekin et al (2017) found that the PIIGS countries suffered the most from the Brexit Referendum. Italy and Spain are both represented in the potential relocating countries and in the PIIGS countries. Bouoiyour and Selmi (2016) claim that because UK trade is heavily focused towards the EU, the Brexit imposed a significant impact on not only the UK stock market but more heavily on the Germany and the French stock markets. In order to investigate whether the Brexit caused a positive effect for real estate in relocating countries or whether due to the expected negative effects for trade the real estate markets reacted negatively compared to non-relocating countries the following hypothesis is tested:

H2: The Brexit Referendum affected real estate investment companies in relocating countries positively where non-relocating countries were not affected by the Brexit Referendum.

According to Chung et al (2015) increasing globalization of the economy led to an increase in importance of foreign exchange risk for investors and corporations. Foreign exchange risk is the risk on a change in value of an investment due to changes in the exchange rate. Foreign exchange risk is important for international investors. Since real estate investment companies invest in different countries besides their domestic market they face foreign exchange risk. Foreign exchange risk

(20)

20

depends on the level of foreign exchange rate exposure and therefore differs among real estate investment companies. Real estate investment companies face different type of foreign exchange risk. When real estate investment companies invest heavily in countries with a different currency, the received rents face exchange risk. Fluctuations in exchange rates will lead to fluctuations in value of rents received from investments in foreign currencies. Ostler (2018) found a drop of 10.4% of the Pound Sterling against the euro. In relative value this meant a drop from €1.3017 on June 23rd to €1.1663 on July 6th. Hung (1992) found a significant effect of exchange rate changes on profits of United States manufacturing companies. He found that due to an appreciation of the home currency companies’ profits are inclined to decrease. Harper and Aggarwal (2005) state that however to a less extent, even companies who have a minimum of foreign involvement are likely to suffer from exchange risk. This is because these companies are likely to have non-domestic competitors or use non-domestic suppliers. On top of that companies are affected by exchange risk through interest rates and interest rate changes. Foreign exchange risk also holds for investors investing in REITs. Unstable exchange rates lead to higher volatility in the value of investments induced by volatility in exchange rates.

Tu & Han (2011) found a significant negative effect for international shareholders on the volatility of companies. The effect was negatively significant for mutual funds and for other investors. This implies that international shareholders do impose an effect on the volatility and therefore could impose an effect on the returns of stocks. In addition, Li et al (2010) found a negative significant relationship between the volatility of stock returns and the portion of foreign ownership in emerging markets. Therefore, the performance of real estate company stocks could be affected by the nationality of the shareholders.

Ngo (2017) find REIT values to be co-dependent on exchange rate changes. For example, changing rates have an effect on tourism which affect hotel & resort REITs’ income. Retail shopping is affected by changes in exchange rates which in consequence affects the income of retail REITs. Ngo’s (2017) results show that the REITs which are classified as: Health care, industrials/office, lodging/resort, residential, retail and self-storage show a negative significant effect due to exchange rate changes. These findings show that REIT performance is affected by exchange rate changes. The combination of these theories has led to the formulation of the third hypothesis:

H3: Foreign exchange exposure affected the performance of real estate investment companies negatively during the Brexit Referendum.

(21)

21

3. Methodology

This section encompasses the methodology used in this research in order to test the formulated hypotheses. Hypotheses 1 and 2 test the effect of the Brexit on the UK, relocating and non-relocating stock markets. According to Binder (1998) the dominantly methodology used to test the effect of events on company value in empirical research is the event study. Therefore hypotheses 1 and 2 are tested by conducting an event study. In order to test whether foreign exposure had an effect on the behaviour of real estate investment companies during times of the Brexit Referendum multiple regressions will be performed and analysed. The first part of this section encompasses the event study methodology followed by the regression analysis methodology.

3.1 Event study

In efficient markets the impacts of events on the value of companies can be measured through stock prices. One methodology which can measure the effect of events on company values is by the use of an event study. The event study measures and analyses the difference in return during the investigated event and the normal returns. This difference between the measured and expected normal return is called the abnormal return. Following de Jong & de Goeij (2011), the methodology of conducting an event study is split into three steps: identification of the event, specification of a benchmark which measures the normal stock return and the final step is the calculation and analysis of the abnormal returns.

Identification of the event

The main event studied in this research is the Brexit Referendum. The Brexit Referendum is held on the 23rd_{of June in the UK with the aim of deciding whether or not the UK should leave the EU. The}

results of the referendum were announced the day following the referendum day and therefore the main event in this study is June 24 2016. However, in efficient markets, event dates might not yield meaningful results in the situation were prior to the event date events took place which could have influenced stock prices. In the situation of the Brexit Referendum announcements regarding the increase or decrease in probability of actually leaving the EU could have affected stock markets besides the outcomes announcements. Therefore, other events are included in this research which could have affected stock markets prior to and after the main event date. Figure 1 shows a timeline of the included events.

(22)

22 January 23 2013: PM Cameron states he wants to hold a referendum when his Conservative party wins the

2015 election June 9 2015: The European Union Referendum Act 2015 passed the House of Commons February 22 2016: Referendum date is announced June 24 2016: Referendum outcome is announced March 29 2017: PM May triggered Article 50 which notifies the European Union the United

Kingdom is leaving the Eruopean Union February 9 2018: Negative outcomes negotiations March 19 2018: Agreement in principle is find concerning transition period Figure 1

Timeline Events included in this Research

date the incumbent Prime Minister at the time David Cameron announced that he wanted to hold a referendum regarding the position of the UK in the EU. This referendum would take place when his conservative party would win the elections in 2015. This announcement was the first official announcement of a potential Brexit and therefore could have imposed a meaningful effect on the value of real estate investment stocks. The second event date included is June 9 2015. On this date the House of Commons passed the European Union Referendum Act 2015 on its second reading. The passing of this act made it legal to hold a non-binding referendum. On February 22 2016 the referendum date was officially announced for which this date is included in this research as the third event date. The fourth and main event date is the announcement of the referendum outcomes. On the referendum day the polls did not show a clear outcome and therefore the outcomes announcement is expected to show the biggest effect. Due to the inability of the market to adjust on the referendum day based on expected outcomes. Since the referendum was non-binding, the results of the referendum did not by certainty mean that the Brexit would take place. On the 29th_of

March 2017, the incumbent Prime Minister at the time Theresa May triggered Article 50. By triggering Article 50 the UK officially announced they would withdraw from the EU. Triggering Article 50 is followed by a two-year negotiation period in this research called the Brexit negotiations. The Brexit negotiations were concerned with the terms and conditions under which the UK would leave the EU. The sixth event date included is February 9 2018. On this day EU negotiator Michel Barnier stated in a press conference substantial disagreement remained between the United Kingdom and the European Union, therefore it was unclear whether or not negotiators would reach an agreement regarding a transition period ("Barnier: Brexit transition 'not a given'", 2018). As explained in the literature review, not reaching an agreement regarding a transition period could impose a huge

(23)

23

effect on the UK and EU economies. The final event date included in this research is March 19 2018. On this date an agreement in principle was reached concerning a transition period. During this period, the current rules and laws regarding the Customs Union and the internal market will continue to apply for the United Kingdom. This transition period will be until the first of January 2021.

Normal returns

In an event study normal returns have to be specified in order to function as a benchmark to find the abnormal returns. Abnormal returns (AR) can be found by subtracting the benchmark or normal return (NR) from the actual return (R). The following formula shows the relation between abnormal returns, normal returns and returns:

𝐴𝐴𝐴𝐴𝑖𝑖𝑖𝑖 = 𝐴𝐴𝑖𝑖𝑖𝑖− 𝑁𝑁𝐴𝐴𝑖𝑖𝑖𝑖

Prior to finding the abnormal returns, the normal returns have to be identified which will be used as a benchmark. The estimation of the normal return is performed over a period prior to or following the event date. The period over which normal returns are estimated is called the estimation window. Figure 2 shows the timeline around the event. In figure 2 T1 till T2 shows the estimation window. De Jong & de Goeij (2011) states that there are multiple models which could be used in order to estimate the normal returns. This research uses the market model of Fama et al (1969). The market model is used since the market model accounts for differences in beta while other models assume that betas of all stocks included in the research are equal to one. The market model is defined as:

𝐴𝐴𝑖𝑖𝑖𝑖 = 𝛼𝛼𝑖𝑖+ 𝛽𝛽𝑖𝑖𝐴𝐴𝑚𝑚𝑖𝑖+ 𝜀𝜀𝑖𝑖𝑖𝑖

In this model the abnormal returns are the residuals of the above model. Normal returns are therefore identified by:

𝑁𝑁𝐴𝐴𝑖𝑖𝑖𝑖 = 𝛼𝛼�𝑖𝑖𝑖𝑖+ 𝛽𝛽̂𝑖𝑖𝐴𝐴𝑚𝑚𝑖𝑖

According to de Jong and de Goeij (2011), 𝛼𝛼� and 𝛽𝛽̂ represent the ordinary leased squares estimates of the regressed coefficients.

For this research the estimation window will be 100 days prior to the event window. 100 days prior to the event window is in line with Hanke (2016). Hanke (2016) researched the effect of the referendum regarding sudden immigration restrictions in Switzerland on the value of Swiss companies. As mentioned in the literature review the Brexit Referendum and the referendum

(24)

24 Figure 2

Timeline around the event

Source: de Jong & de Goeij (2011)

studied by Hanke (2016) showed great similarities regarding the consequences on the openness of their economies as a result of the referendums.

Abnormal returns

As mentioned the abnormal returns represent the difference between returns and normal returns. Abnormal returns are identified during the event window. In figure 2 t1 to t2 represents the event window. The event window is a pre-determined number of trading days surrounding the event date. The event window should not overlap the estimation window. This research uses three event windows a shorter, a middle and a longer event window. The shorter and longer event windows will serve as a robustness check and therefore could contribute to the validity of the findings. Testing for multiple event windows is in line with Schiereck et al (2016). Schiereck et al (2016) investigated the effect of the Brexit Referendum on the stock and credit default swaps market. Schiereck et al (2016) used an event window solely comprising of the event day and an event window which added an additional day. Following Schiereck et al (2016) the shorter event window will comprise of the event

(25)

25

day and one additional day. The middle event window applied in this research contains one day prior to the event date and four trading days post the event date. The longer event window contains one day prior to the event date and six trading days post the event date. During the middle and long window markets could have processed further information and readjust in case of overreaction due to the surprising element of the event. In conclusion this means that the shorter event window is indicated as (0, +1), the estimation window is indicated as (-101, -1). The middle event window is indicated as (-1, +4) and the accompanying estimation window is indicated as (-102, -2). The longer event window is indicated as (-1, +6) and the associated estimation window (-102, -2).

De Jong & de Goeij (2011) state that in order to increase the informativeness of the analysed abnormal returns they ought to be averaged. The reason behind this is because the majority of movements in prices of stocks are related to factors unrelated to the studied event. By averaging abnormal returns the effects not caused by the particular events are cancelled out. The abnormal returns are averaged by the following formula:

𝐴𝐴𝐴𝐴𝐴𝐴𝑖𝑖 = _{𝑁𝑁 � 𝐴𝐴𝐴𝐴}1 𝑖𝑖𝑖𝑖 𝑁𝑁

𝑖𝑖=1

In order to test the effect of the event on the whole event window instead of the event date we use the cumulative abnormal return. In the cumulative abnormal return, abnormal returns found in the event window are summed by the following formula.

𝐶𝐶𝐴𝐴𝐴𝐴𝑖𝑖 = � 𝐴𝐴𝐴𝐴𝑖𝑖𝑖𝑖 𝑖𝑖2

𝑖𝑖=𝑖𝑖1

In order to test the effect over a sample of companies the cumulative average abnormal return is used. The formula for the cumulative average abnormal return is defined as:

𝐶𝐶𝐴𝐴𝐴𝐴𝐴𝐴 = _{𝑁𝑁 � 𝐶𝐶𝐴𝐴𝐴𝐴}1 𝑖𝑖 𝑁𝑁

𝑖𝑖=1

Abnormal returns are tested by the following hypothesis:

𝐻𝐻0: 𝐸𝐸(𝐴𝐴𝐴𝐴𝑖𝑖𝑖𝑖) = 0 𝐻𝐻1: 𝐸𝐸(𝐴𝐴𝐴𝐴𝑖𝑖𝑖𝑖) ≠ 0

Abnormal returns are tested to be statistically different form zero at the one, five and ten percent significance level. According to de Jong & de Goeij (2011) the most common manner of testing the null hypothesis is by a t-test. In order to perform the t-test we assume the abnormal returns, which after averaging are the average abnormal returns to be independently and identically distributed.

(26)

26

The abnormal returns are assumed to follow a normal distribution with mean zero and variance σ2_.

The null hypothesis for average abnormal returns is tested by the following test statistic and corresponding estimator of σ: 𝑇𝑇𝑇𝑇1= √𝑁𝑁𝐴𝐴𝐴𝐴𝐴𝐴_St𝑖𝑖 𝑇𝑇𝑆𝑆 = �_{𝑁𝑁 − 1 �(𝐴𝐴𝐴𝐴}1 𝑖𝑖𝑖𝑖 𝑁𝑁 𝑖𝑖=1 − 𝐴𝐴𝐴𝐴𝐴𝐴𝑖𝑖)2

For the cumulative abnormal returns the same test statistic can be applied. In this case the hypotheses, test statistic and estimator of σ are defined as:

𝐻𝐻0: 𝐸𝐸(𝐶𝐶𝐴𝐴𝐴𝐴𝑖𝑖𝑖𝑖) = 0 𝐻𝐻1: 𝐸𝐸(𝐶𝐶𝐴𝐴𝐴𝐴𝑖𝑖𝑖𝑖) ≠ 0 𝑇𝑇𝑇𝑇2= √𝑁𝑁𝐶𝐶𝐴𝐴𝐴𝐴𝐴𝐴_𝑇𝑇 𝑇𝑇 = �_{𝑁𝑁 − 1 �(𝐶𝐶𝐴𝐴𝐴𝐴}1 𝑖𝑖 𝑁𝑁 𝑖𝑖=1 − 𝐶𝐶𝐴𝐴𝐴𝐴𝐴𝐴𝑖𝑖)2

In order to test for the difference in effect for the relocating and non-relocating country in hypothesis 2 the following test statistic is used:

𝑇𝑇𝑇𝑇 𝑓𝑓𝑓𝑓𝑓𝑓 𝑆𝑆ℎ𝑒𝑒 𝑑𝑑𝑑𝑑𝑓𝑓𝑓𝑓𝑒𝑒𝑓𝑓𝑒𝑒𝑑𝑑𝑑𝑑𝑒𝑒 = 𝐶𝐶𝐴𝐴𝐴𝐴𝐴𝐴1− 𝐶𝐶𝐴𝐴𝐴𝐴𝐴𝐴_σ 2 ∆ σ∆= �_𝑁𝑁σ1 1+ σ2 𝑁𝑁2

(27)

27

3.2 Regression analysis

In order to test whether or not foreign exchange exposure plays a role in the performance of real estate investment companies in times of the Brexit, multiple regression analyses will be performed. The regression will be performed with the cumulative abnormal returns as dependent value around the time the Brexit outcomes were announced. Separate regressions will be performed by UK companies, companies incorporated or legally registered in relocating countries and companies incorporated or legally registered in non-relocating countries and on the total sample.

In equation 1 the investment allocation of the researched REITs is regressed upon the cumulative abnormal returns for each company in the sample. The allocation is divided between UK investment, investment in Eurozone and investment non-Eurozone EU members and non-EU investments. The investments outside the EU variable will be the omitted variable. Figure 4 in the appendix shows which countries belong to the Eurozone and the EU. Ostler (2018) found a 10.4 percent drop in the Pound Sterling following the Brexit. A devaluation of a currency leads to a devaluation in value of an investment in that currency. Therefore, the effect of the Brexit is expected to be more severe for companies which invested more in UK real estate. The value of the variables is in portions of the total amount invested by real estate investment companies.

Ngo (2017) show that REITs which are classified as: Health care, industrials/office, lodging/resort, residential, retail and self-storage show a negative significant effect due to exchange rate changes. Therefore, in equation 3 the sectors in which the real estate investment companies invest are included. Dummy variables are included for the sectors: industrial, hotel & resort, office, health care, residential, retail and specialized. The dummy for diversified is the omitted variable. Since not all samples contain date regarding each investment sector depending on the sample other sectors will be omitted as well. For the UK regression the hotel & resort sector will be omitted, in the relocating regression the industrial and healthcare sectors will in addition be omitted and for the non-relocating regression the hotel & resort sector will be omitted.

(1) CAR = β0 + β1 UK investment + β2 Investment in Eurozone + β Investment in EU non-Eurozone + β4 sector specific dummies + ε

Foreign exchange risk for investors is the potential volatility of their investment due to changes in exchange rate. The nationality of shareholders could also impose an effect on the cumulative abnormal returns. Shareholders investing in the same currency as they use for their other financial activities face lower foreign exchange risk compared to investors who invest in a foreign currency. As mentioned a drop in value of the Pound Sterling imposes an effect for the value of foreign

(28)

28

investments. In order to test if this assumption is true the nationality of investors is added in equation 2. Nationalities are grouped in: UK based investors, Eurozone based investors, non-Eurozone EU countries based investors and other countries based investors. The outside EU nationality variable is the omitted variable. The variables are in percentages of the total amount of investor.

(2) CAR = β0 + β1 UK shareholders + β2 EU Eurozone shareholders + β3 EU non-Eurozone shareholders + β4 sector specific dummies + ε

Equation three combines equation one and equation two in order to test for the effect of foreign exchange exposure through both indicators geographic investment allocation and shareholder nationality.

(3) CAR = β0+ β1 UK investment + β2 Investment in Eurozone + β3 Investment in EU non-Eurozone + β4 UK shareholders + β5 non-Eurozone shareholders + β6 EU non-non-Eurozone shareholders + β7 sector specific dummies + ε

In the final equation the following control variables are included: market capitalization and liquidity ratio. Market capitalization represents the total market value of the outstanding shares of a company. The natural logarithm of market capitalization is added as a variable because Barber & Lyon (1997) and Lakonishok & Shaprio (1986) found a significant effect of size on stock return, implying stock returns are related to the size of companies. Lakonishok & Shaprio (1986) found that larger companies perform worse compared to smaller companies when markets are going down. The last included variable is the liquidity ratio. Liquidity ratio represents the ability of a company to repay its debts. Liang and Wei (2012) find evidence that liquidity risk is a determinant of stock pricing on a global level. In addition, Brounen et al (2009) researched the relationship between liquidity and company value. Brounen et al (2009) found a positive significant relationship between liquidity ratio and company values.

(4) CAR = β0 + β1 UK investment + β2 Investment in Eurozone + β3 Investment in EU non-Eurozone + β4 UK shareholders + β5 Eurozone shareholders + β6 EU non-Eurozone shareholders + β7 sector specific dummies + β8 LN Market capitalization + β9 Beta + β10 liquidity ratio + ε

For the regressions on the total sample interaction dummies are included for the variables regarding geographic investment allocation and shareholder nationality. These interaction dummies are included in order to assess the effect of those variables on cumulative abnormal returns per sample.

(29)

29

4. Data

This section comprises of how the data used in this study was retrieved, and what actions have been undertaken in order to prepare the date to be included in this research.

In order to calculate the effect of the Brexit on performance of companies the daily stock prices are used to calculate returns. The data regarding stock prices are retrieved from the Compustat database. Data from the Compustat database is provided by Wharton Research Data Services (WRDS). Datastream provided the daily prices of indices for the countries included in this research. Indices were included per country and therefore in order to investigate hypothesis 3 data was required regarding foreign exchange exposure. The variables on which data was required were: geographic investment allocation, shareholders nationalities, sector specialization, liquidity ratio’s and market capitalizations. Data regarding geographic allocation is retrieved from Thomson ONE. Thomson ONE provides percentages of geographic allocation for companies. Compustat provides information regarding the sector specialization. The Amadeus database from Bureau van Dijk provided date regarding the origins of the shareholders. The Amadeus database was provided by WRDS. Data regarding company’s market capitalization and liquidity ratio is retrieved from the Orbis database.

4.1 Sample selection

Sample selection for the real estate investment companies followed the following steps. First stock prices from June 2012 until May 2018 were downloaded from WRDS for EU member states. This range covers the estimation and event windows for the 7 investigated event dates. Based on Standard Industrial Classification (SIC) codes and Global Industrial Classification Standards (GICS) codes non-real estate investment companies are excluded from the sample. After excluding non-real estate investment companies from the sample, mortgage REITs and other non-equity real estate companies are excluded from the sample. This is because this research focusses on the effect of real estate prices which is one of the underlying variables of real estate investment companies’ values. Thereafter the sample is divided by sectors. Based on the GICS the division is based on the following sectors: diversified, industrial, hotel & resort, office, health care, residential, retail and specialized. The specialized sector represents companies who do invest in specialized properties which however cannot be subdivided in the previous mentioned sectors. An example of a specialized sector is self-storage. Following EY (2017) who believes the most probable relocating locations for UK based companies to be: France, Germany, Italy, the Netherlands and Spain, a separation is made between the UK, the relocating countries, and the remaining countries. In this research the remaining group

(30)

30

will be called the non-relocating countries. The non-relocating countries are EU member states which are neither the UK nor the relocating countries. The non-relocating group represents 18 countries. Table 1 show the number and mean total assets and share of total assets of included real estate investment companies per sample and per sector. Not every sample contains companies for all sectors. The UK and the non-relocating sample are missing companies for hotel & resorts and the relocating sample is missing companies for industrial. Table 1 show the mean total assets per sector for each sample. On top of that the table shows the percentage of representation of sector in the samples. The table show that for the relocating and non-relocating samples the diversified sectors represents the majority of the companies included. The UK sample shows a more proportional distribution of included companies.

Table 1

Sector Composition of Real Estate Investment Companies

UK Relocating Non-relocating N Total assets (in m €) % of tot UK N Total assets (in m €) % of tot relo N Total assets (in m €) % of tot non-relo Diversified 41 1016.725 35.3% 115 1440.713 60% 167 865.545 90.1% Industrial 4 3016.769 10.2% 0 0 0% 2 1388.683 1.8%

Hotel & resort 0 0 0% 2 9974 7.2% 0 0 0%

Office 4 2873.772 9.7% 21 1657.571 12.6% 6 981.451 3.8% Health care 2 1520.379 2.6% 0 0 0.0% 1 1173.162 0.8% Residential 4 977.483 3.3% 1 493.000 0.2% 6 320.086 1.3% Retail 6 7377.197 37.5% 16 3393.625 19.6% 6 536.724 2.1% Specialized 2 1324.348 2.2% 2 557.000 0.4% 2 79.251 0.1% Total 63 1874.785 100% 157 1760.153 100% 191 839.485 100%

Table 1 presents the sector composition for the real estate investment companies included in this research. Sector composition is based on GICS and SIC codes. The composition is made for three samples: companies incorporated in the UK, in relocating countries and in non-relocating countries. The table provides the number of included companies, the mean value of assets in millions of euros and the share of Sector assets per sample.

4.2 Event study

In order to perform an event study data regarding the stock prices is needed. As stated this data is retrieved from WRDS for all the event dates. Since the market model is used in order to find the normal returns, daily indices prices were required. For each of the countries included in the samples the main index is used in order to find the normal returns. Where available all-share indices were used which represent a value weighted index of all the shares listed on the stock market. When an

(31)

31

all-share index was not available other major stock indices were used per country. For the robustness checks real estate indices per country have been retrieved from Datastream. Table 11 in the appendix show how many companies have been included in this research per country. Since some countries had only one real estate investment company listed on the stock exchange, it has not been possible to perform an event study for each country for each event date separately. The countries which lacked enough companies are: Lithuania, Latvia, Portugal and Romania. The companies from this country are included in the studies on the relocating and non-relocating sample. However, this research does not include a regression for these countries separately.

4.3 Regression analysis

In order to perform the regression analysis to answer hypothesis 3 some additional data was necessary. As mentioned data regarding geographic segmentation, shareholders nationality, sectors, market capitalization and liquidity ratio have been retrieved from Compustat, Amadeus, Thomson ONE and Orbis. However not all data was available for every company included in the samples. Therefore, the regressions are performed on a reduced sample. The reduced sample consists of 53 companies in the UK sample, 96 in the relocating sample and 86 in the non-relocating sample. Table 2 shows the number of observations and the mean market capitalization for each sector per sample. The table shows that the total mean market capitalization of the UK and the relocating sample is almost equal in value, while the non-relocating sample’s mean market capitalization is substantially lower.

Table 2

Sector Composition Reduced Sample

UK Relocating Non-relocating

N Mean market cap (in m €) N Mean market cap (in m €) N Mean market cap (in m €)

Diversified 32 715.216 73 1158.847 69 363.938

Industrial 4 1882.116 0 0 2 1135.918

Hotel & resort 0 0 2 6011.294 0 0

Office 3 1775.073 15 782.705 5 715.048 Health care 2 945.925 0 0 1 987.517 Residential 4 661.687 1 273.523 6 216.782 Retail 6 3447.719 10 1953.42 4 462.877 Specialized 2 1087.389 2 191.476 1 90.946 Total 53 1209.997 103 1201.357 88 399.880

Table 2 presents the sector composition for the real estate investment companies included in this research. Sector composition is based on GICS and SIC codes. The composition is made for three samples: companies incorporated in the UK, in relocating countries and in non-relocating countries. The table provides the number of companies included and the mean market capitalization in millions of euros.

The effect of the Brexit Referendum on the value of real estate investment companies in the European Union