678430

Co-Movements of UK, European

equity markets following the EU
referendum and its implications on
the global portfolio diversification
theory.

Shaun Parris

678430

BA (hons) Accounting with Finance

Word Count: 10,000

Supervisor
Lena Itangata
March 2017

1|Page
678430

Statement of Originality

This dissertation is submitted in partial fulfilment of the requirements for the degree of BA
Hons Accounting with Finance. I declare that this dissertation is my own original work.
Where I have taken ideas and/or wording from another source this is explicitly referenced in
text.

I give permission that this dissertation may be photocopied and made available through the
university library, in printed from and/or electronic form.

I provide a copy of the electronic source from which this dissertation was printed. I give my
permission for this dissertation, and electronic source, to be used in any manner considered
necessary to fulfil the requirements of the University of Portsmouth Regulations,
Procedures and Codes of Practice.

Word count: 10,000 (excluding covers page, acknowledgements, abstract, glossary,

bibliography, appendices, headings, list of tables/figures and tables in the body part).

Signed:

Date:

2|Page
678430

Acknowledgement

First of all, I would like to thank my supervisor Lena Itangata for her continuous help,
guidance and knowledge throughout the duration of this dissertation.

I would also like to thank my family and friends for their continued help, support and
encouragement during my time at the University of Portsmouth.

A special thanks to the best housemates for their help and amazing time together.

Finally, thank you for taking the time to read this dissertation.

3|Page
678430

Abstract

Stock Market co-movements have been highly researched particularly in relation to

international diversification and reducing market risk. Past studies have looked at the
implications of global shocks such as the financial crisis or 2001 terrorist attacks. This
research adapts past studies by looking into the changing Co-movements of all European
Union stock markets with the addition of Japan and US markets. Investigating the
implications of the EU referendum on stock market Co-movements both pre and post event.
To conduct this a range of empirical test have been employed including the critically
examined methods of Johansson cointergration testing, Grangers causality test and VAR and
VECM methodologies.

The findings of this research prove inconclusive as to weakening or strengthening co-

movements between the UK and its European counterparts. The findings indicate that the
EU referendum has had a series of effects on European markets both increasing and
decreasing relationships in a variety of locations and regions. By looking at the variety of
methods the studies assess the dynamic relationship the UK holds with its UK counterparts
and likewise the best and worse markets for investors to invest in when looking to maximise
diversification benefits. Additionally, it helps to access European groupings more notably
the relationship and influence Western Europe exerts on southern and Eastern Europe but
also in line with past studies the big influence that US market play on global contagion and
interdependencies.

4|Page
678430

Table of Contents

Statement of originality...................................................... 2
Acknowledgements .............................................................................. 3
Abstract................................................................................................ 4
Table of contents.......................................................................................... 5
List of tables.................................................................................................. 7
List of Appendixs8
Glossary........................................................................................................ 9
1.0 INTRODUCTION .....................................................................................10
1.1. Literature Gaps...........................................................................................12
2.0.0 Literature review....................................................................12
2.1.0. Contagion an Empirical review ............................................................ 13
2.1.1. Contagion in the Financial crisiss . 13
2.1.2. Contagion in political uncertainty ..14
2.1.3. Contagion in the European debt crisis.15
2.1.4. Channels of Contagion Volatility. 16
2.1.5. Information flows16
2.1.6. Behavioural Theories18
2.1.7. Bull and Bear Markets a comparison19
2.1.8. Review of Contagion 19
2.2.0. Interdependency as a cause of Co-movements21
2.2.1. Channels of Interdependency Fundamentals.23
2.2.2. Channels of Interdependency - Trade and Investment...24
2.2.3. Channels of Interdependency - Geographical Location.. 25
2.2.4. Channels of Interdependency -Technology..26
2.2.5. Correlation in Emerging markets...... 27
5|Page
678430

2.2.6. Review of Interdependency 28

3.0.0. Methodology.29
3.0.1. Correlation coefficient29
3.0.2 Principle component analysis 30
3.0.3 Johannsens cointergration test30
3.0.4. Grangers Causality test32
3.0.5. VAR & VECM..33
4.0 Data & results35
4.1. Pearson Correlations coefficients36
4.2. Principal component Analysis.40
4.3. Stationarity Testing ..43
4.4. Lag Length estimation..43
4.5. Grangers Causality Test..47
4.6. Johansens Cointergration Test.50
4.7. VAR or VECM..58
5.0 Summary and Conclusion..71
6.0 Bibliography.72
7.0 Appendix78

6|Page
678430

List of Tables

Tables:
Figure 1 FTSE 100 Share Price 20th 25th June 2016.11
Figure 2 The benefits of International diversification, Solnik (1995) 11
Table 1 Selected Economic data for the European Union, Japan and US .21
Table 2 Descriptive statistics (natural log differences) ..........36
Table 3 Correlation of the UK Stock Market with the Other Stock Markets37
Table 4 Most and Least Correlated Stock Markets ..38
Table 5 Average Correlation of each Stock Market with other Stock Markets 39
Table 6 PCA Pre Period 41
Table 7 PCA Post Period42
Table 8 Stationery testing results for Log Differenced data44
Table 9 Stationery Testing Results for Daily Closing Prices 45
Table 10 Lag Length Criteria 46
Table 11 Grangers Causality test results for the UK 48
Table 12 Johansens test results for the UK to entire sample (Pre period). 49
Table 13 Johansens test results for the UK to entire sample (Post)... 49
Table 14 Johansen-cointergration test results for UK to markets (Pre period) 52
Table 15 Johansen-cointergration test results for UK to markets (Post Period)54
Table 16 Johansen-cointergration test results for UK to regions (Pre Period)56
Table 17 Johansen-cointergration test results for UK to regions (Post Period) 57
Table 18 VAR and VECM Results (pre and Post)..59

7|Page
678430

List of Appendixes

Appendix 1 - Volatility for major European Stock Markets calculated to 10 basis points
standard Deviation between Jan 2006 November 2016.79
Appendix 2 - Stock Market returns between December 2010 November 2016 Indexed
weekly at 15/12/201680
Appendix 3 - Stock Market prices between Jan 2016 November 2016 Indexed daily from 4th
Jan 2016..81
Appendix 4 - Stock Prices indexed at December 2006 Jan 2009.82

Appendix 5 - Major Trading Partners for Europe and the U.S. (As a % of total).83

Appendix 6 - Trading Hours UTC....84

Appendix 7 - Stock Market Indexs used ..85

Appendix 8 - Pearson Correlation Coefficient Matrix (Pre period)86

Appendix 9 - Pearson Correlation Coefficient Matrix (Post period)..87

Appendix 10 - KMO & Bartletts test results for PCA88

Appendix 11 - PCA Pre Period Full Results .89

Appendix 12 - PCA Post Period Full Results 90

Appendix 13 - Stationery testing equations91

Appendix 14 - stationery Testing equations (2)92

Appendix 15 - Lag Length Criteria Equations93

Appendix 16 - Full Grangers Causal Linkages Results.94

Appendix 17 Markets in Each Region for Johanson and VAR/VECM Purposes95

Appendix 18 Ethics Approval form..96

8|Page
678430

Glossary
ADF Augmented Dicker Fuller
ARCH - Autoregressive Conditional Heteroscedasticity model

BRIC - Brazil, Russia, India and China

CESE - Central, Eastern, and South-eastern Europe

DAX Deutscher Aktien index

DCC - Dynamic Conditional Correlation model

DCC-GARCH - Dynamic Conditional Correlation General Autoregressive Conditional

Heteroscedasticity model

EGARCH Exponential GARCH model

EU European Union

FTSE - Financial Times Stock Exchange Index

GARCH General Autoregressive Conditional Heteroscedasticity model

HAC - heteroscedasticity and autocorrelation consistency

MGARCH Multivariate GARCH model

MIST - Mexico, Indonesia, South Korea, and Turkey

NIKKEI Japanese Stock Market Index

PCA Principal Component Analysis

PP PhillipsPerron

S&P 500 Standard & Poors 500 Index

SD Standard Deviation

VAR Vector Auro-regression

VaR Value-at-Risk model

VECM Vector Error-Correction Model

9|Page
678430

1.0. Introduction

The EU has been pivotal in the economic development of Europe since its establishment in
1991 promoting free markets, trade and innovation, yet since the 2008 Euro crisis its
existence has been questioned. The EU has attracted a great deal of attention from
investors, regulators, financial agencies, portfolio managers & policy makers. Together the
Economies of the EU account for 15% of World trade, some economies rank among the
largest and most influential of the 21st century and combined forming the second largest
global economy. Following the British public opinion, a Referendum was called in June to
decide the fate of the UKs existence in the EU.

On the 23rd June the British public voted to leave the European Union and this was
immediately reflected in a surge of volatility inside leading world markets, the FTSE 100
falling 8% (Figure 1) signifying significant market uncertainty and volatility before leaping to
its highest value since August 2015. On the Friday Global stocks lost around $2 trillion
fuelled by global uncertainty. The sterling to Dollar fell to its lowest since 1985 having
implications on global trade. Stocks tumbled in Europe. Frankfurt .GDAXI and Paris .FCHI
each fell 7 percent to 8 percent. Italian .FTMIB and Spanish .IBEX markets posted their
sharpest one-day-drops ever. Given the volatility of markets and important role the EU
markets play not just in terms of size but also global economic the UK economy needs
specific research, predominantly in terms of its movements with European markets
following the announcement on the 23rd June to aid investors in understanding the
international diversification benefits.

10 | P a g e
678430

Figure 1: FTSE 100 Share Price 20th 25th June 2016 Source LSE

Figure 2: The benefits of International diversification, Solnik (1995)

The research aim of this paper is to better understand the co-movements of UK and EU
economies following the EU referendum to better Investors investment decision in regards
to portfolio formation and global diversification theory. Sharpe (1964) empirical study
identified the relationship between risk & return under the capital asset pricing model.

11 | P a g e
678430

Global portfolio theory was initially established by Markowitz (1991) identifying the benefits
of expanding your portfolio internationally to reduce market risk (See Solnik 1995, Figure 2),
therefore the analysis of cross market co-movements is of great significance with regards to
optimal portfolio allocation and risk management (See Markowitz, H. M. 1991). Co-
movements and correlation is a critical factor investors take into consideration when
making investments decision because this feature relates to how much benefit can be
obtained from diversifying internationally. The portfolio diversification implication of the co-
movements of national stock markets has long been a prevalent research topic in finance.
Low correlation between national stock markets is evidence in supporting the benefit of
global portfolio diversification 7 therefore reduced risk (See, e.g., Levy and Sarnat (1970),
Lessard (1976) and Meric (2015).

1.1. Literature Gaps

Throughout my research I have uncovered a several literature gaps to be addressed. Firstly,
empirically much of the studies into co-movements and correlation has focused on the US
market due to its pivot role and global links. There has been a lack of insight into co-
movements and its implications on the UK following events of such nature. Research papers
have focused extensively on emerging markets particularly that of the BRIC economies but
additionally there is a lack of studies in regards to the EU referendum and also the
implications of political events on stock market co-movements.

12 | P a g e
678430

2.0. Literature Review

This section focuses on the leading channels of co-movements and correlation. Looking at
the short term (Contagion) and long term time horizons (Interdependence), this literature
review will focus on those two core elements, initially proving their existence and then
looking into their root causes and how these factors could be implied following the EU
referendum.

2.1.0. Contagion an Empirical review

One theory that can aid us in understanding the effects of the European referendum on
global market co-movements is the theory of contagion and interdependence,
understanding the differences can critically aid investors decision in international portfolio
allocation (see, Forbes and Rigobon, 2002). Empirically a lot of research has been conducted
on the subject from political to economic events and likewise a multitude of definitions for
the term. For the purposes of this paper contagion can be defined as An increase in cross-
country correlations during crisis or shock times relative to correlations during tranquil
periods.

2.1.1. Contagion in the financial crisiss

Meric and Nygren (2015) proved the theory of contagion in their study of the 2007 financial
crisis in which shocks were felt around the world and the events implication on the co-
movements of European stock markets. By using weekly stock market returns it compared
the correlation of 21 market indices for a 5-year period prior to and after the crisis. Results
show that contagion did indeed spread from the US on a global scale as correlation
increased by 41%, a significant increase from the long term interdependency in these
regions. Additionally, through PCA the research identified only 1 statistically significant
component following the financial crisis as opposed to 2 components in the stable years.
Zhang & Li (2014) identified a similar trend by comparing co-movements between the US

13 | P a g e
678430

and Chinese equity markets, finding an upwards trend in correlation by using daily spot
checks and Johannsens cointergration testing. Their findings identified no structural long
term relationship between China and the US but then contagion following the 2008 crisis
that led to an upwards structural break. Yarovaya, & Lau (2016) built on these studies by
looking at co-movements between the UK, BRIC and MIST countries before and after the
financial crisis, they too identified contagion from the shock event in the UK. By looking at
daily returns and the GARCH model as opposed to weekly they were able to consolidate on
the findings by implementing more powerful tests of cross-market movements, additionally
application of the asymmetric causality test by Hatemi-J (2012) to highlight the flow of
contagion. The events of the 2007 financial crisis showed that shock was felt not only on a
regional level but also globally significantly.

Moreover, the contagion felt through global economies since 2007 is not an anomaly,
historical research into the 1987 Stock Market crash has provided similar results under
similar methodologies. Arshanapalli & Doukas, (1993) produced undistinguishable results of
the amplified co-movements of global markets following the shock in 1987 by using
cointergration testing on pre and post event returns. Also identifying a level of
interdependency between the European countries studies, specifically that of Germany,
France, the UK as well as the US. However, Forbes & Rigobon (2002) challenged these
findings stating it was merely down to long term interdependence in the region (explored in
the second part of this literature review); studies by Gelos and Sahay (2001) uncovered
similarly results in regards to the 1997 Asian crisis, 1998 Russian default and the Czech crisis,
this study is extremely beneficial in identifying the possibility of contagion following regional
shocks, therefore its significant to this research. Additionally, Yang 2003 identified spill over
effects in regards to the Asian crisis by using the Garch model. Historical studies of financial
shocks under a variety of methodologies have proven that contagion does indeed exist,
however political shocks also demonstrate an equal view.

14 | P a g e
678430

2.1.2. Contagion in political uncertainty

The September 2001 terrorist attacks is stated to be one of the biggest political events since
the turn of the century. Hon (2004), studied the effect of the September 2001 terrorist
attacks through GARCH and concluded that in times of global uncertainty and crisis, returns
from international diversification substantially diminish in the 3-month period following the
event. Additionally, identifying that following the shock event, markets moved more closely
with that of the US following this events & increased interdependency. Meric (2007) results
aligned closely with this looking at the long term implications of the event, applying
correlations coefficient for the 5-year period before and after, the studies identified that the
event led to long term interdependencies. Moreover, it highlighted weaknesses in Meric
approach, for simplicity the study was only taken using weekly close prices and past studies
on co-movements have proven the need for daily or intraday analysis.

2.1.3. Contagion in the European debt crisis

The event that most closely aligns with the situation under study is the European debt crisis
of 20082015 as a result of the region involved and mixture of political and economic
consequences. This event has been covered extensively and can give us an insight into the
regional implication of a Eurozone shock as well as the short term spill over effects and long
term interdependency that the UK has with its European counterparts and global regions.
Wasim (2013) investigates not only contagion in the region but also on a global scale,
applying Multivariate DCC-Garch (see Engle 2002) on daily stock returns identifying
significant contagion both within the European region but also on a global scale with the US,
Japan and BRIICKs. Moreover, finding that Spain & Italy were the most contagious countries
in the EU. Additionally, the use of DCC-GARCH allows us to identify linkages in both stable
and shock events without taking volatility into account unlike that of correlation coefficient
which is positively skewed towards periods of high volatility. Syllignakis, (2011) also looked
at the implications of the Euro debt crisis on equity Market co-movements particularly that
of Germany, Central and Eastern Europe, again with identical findings to that of Wasim
(2013) through the use of the GARCH. This research has looked into the theory of contagion

15 | P a g e
678430

and no doubt empirical studies have proven its existence but now we are going to look into
the channels of contagion;

2.1.4. Channels of Contagion Volatility

Volatility has been highlighted as a signal of Market contagion as stated by Longin & Solnik
(1994) who investigated conditional correlations under the Bivariate GARCH model (see
Bollerslev 1990) and co-variances between the periods of 1960-1990. Historical events as
above prove that shock events cause downward spiralling volatility due to the public
perception of shock events and news (see Schewert, 1990) and in every event this has been
the basis point for accessing and identifying contagion. The evidence (Appendix One)
highlights increased Volatility through standard deviation in all stock markets during
September 2008 (The Lehman brothers Bankruptcy). This can be related to Appendix Three
which has been drawn up to identify market contagion for European Stock Markets in 2016.
It clearly identifies a negative shock in equity prices in late June the time of the EU
referendum and further support for the theory of contagion following the EU referendum.
Furthermore, King and Wadhwani (1994) used the GARCH model to come to the conclusion
that volatility or periods of unobservable economic variables concluded to contagion, this
was later supported by Karolyi & Stulz (1996) whos studies recognized return correlation
between markets increases with the volatility in each market. However, Ross (1986)
findings challenged this view arguing that volatility is caused by an underlying information
flow or macroeconomic announcements but also that unobservable economic variables
can be used to understand contagion. These 2 factors will now be discussed further:

2.1.5. Channels of contagion - Information flows

Connolly (2003) built on Rosss finding of Unobservable economic variables, suggesting
that foreign market returns convey information distinct from the public information flows
about economic fundamentals. Because of this, domestic traders have an incentive to use
unobservable information from previous foreign market returns and use it for their

16 | P a g e
678430

domestic trading, therefore any noise1 from one domestic market will be carried over to the
foreign market and in turn cause contagion. By looking at Intraday movement Connolly
Identified overnight spill over effects from US to Japan equity markets following volatile
stock movements, particularly on the basis of industrial production and the unemployment
rates. These studies are supported by the findings of McQueen and Roley (1993) and Karolyi
& Stulz (1996) who used intraday correlations, identifying that returns between certain
countries are normally distributed namely a large overnight shock in one market will lead to
stronger correlations the next day in other markets, further evidence of contagion. This was
compiled through the use of ARCH(1,1) & GARCH(1,1) on correlations matrix to further
support findings under periods of high volatility.

As highlighted by Ross (1986) an influence on short term co-movement dynamics is the role
of macroeconomic information. In times of market stress individuals react more to adverse
news due to risk preferences & loss aversion, therefore this flow of news can cause short
term contagion. Karolyi & Stulz (1996) identified that the release of macroeconomic news
(I.e.Unemployment, interest rates) has a small and insignificant effect on the co-movement
of US and Japanese stock markets. Furthermore, Connolly & Wang(2003) builds on these
finding in the US and Japanese equity markets. Most importantly finding that the macro
news effect is too small to account for any economically sizeable part of the return Co-
movement among the markets studies. Macroeconomic news cannot cause contagion itself
but more so the unobservable information that accompanies these new release and the
behaviour of noise traders in causing short term spill-overs effects in domestic and foreign
markets. This is important to understand for the Eurozone as it assists portfolio managers in
diversifying following a macro news event such as that the court ruling on the 4 th November
that following these findings could cause a smaller shock and therefore contagion. However,
Lux (1995) researched into contagion and the implication of unobservable information in
the form of excess volatility and finds it is as a result of herding behaviour, stating contagion
can be interpreted as an attempt to draw information from what the others do.

1
'Noise - The term used to describe an investor who makes decisions regarding buy and sell trades without
the use of fundamental data. These investors generally have poor timing, follow trends, and over-react to good
and bad news.

17 | P a g e
678430

2.1.6. Channels of Contagion - Behavioural Theories

Herding has emerged from the behavioural finance theory, Nofsinger and Sias (1999)
consider herding as a result of a group of investors trading in the same direction over
a period of time. There have been several studies into the prominence of Herding during
shock events. Christie and Huang, (1995) & Sharma (2015) identified motivations of herding
nature following times of market stress specifically that of the 2007 financial crisis. Chang
(2000) employs cross-sectional standard deviation of returns for the period to 2000 and
identifies herding occurs in developing but not developed nations as it is an action of the
irrational/noise investor. Chiang & Zheng (2010) contradicts these studies by directly linking
herding behaviour to financial contagion identifying that herding behaviour arises in the
shock country and then spreads out. On top of this the study identifies herding nature in
developed economies such as the UK, Germany as well as developing nations highlighting its
global presence. Syllignakis (2011) delves further into the understanding of the herding
effect, looking and Central and Eastern Europe between 1999-2009, through the application
of DCC-Garch and correlation coefficients, the study identifies that emerging markets are
more prone to the herding effect due to poor market efficiencies & greater noise trading.
Due to this investor usually base their decisions off the more reliable information on offer in
developed and efficient markets. Understanding the implications of herding is important for
portfolio managers as these studies identify that Herding occurs more frequently in the
developing nations and significantly reduces the benefit of global diversification.

Furthermore, the Herding effect is not the only behavioural theory that has been developed
in reference to contagion, Kodres & Pritsker (2001) & Kaminsky (2000) highlight the role of
cross-market hedging as a channel of contagion. The ideology behind this theory is that
through this channel, investors transmit idiosyncratic shocks from one market to another by
adjusting their portfolios' exposures to shared macroeconomic risks even when two markets
have no common macroeconomic variables. This research has been critical in identifying
cross market hedging as a factor. Cipriani, Gardenal, & Guarino (2013) research support the
existence of cross-market hedging, building on past studies and identifying that markets are
more susceptible to cross market hedging if A.) There is little to no information asymmetry

18 | P a g e
678430

and B.) Markets are Transparent and efficient at pricing. In turn this supports the argument
that cross-market hedging does indeed exist in the European Union due to the very nature
described above. Above all else it also highlights portfolio managers need to diversify
internationally and this itself has increased market contagion as international deposits
increases, much the same as the growth in financial services, hindering international
diversification benefits.

2.1.7. Bull and Bear Markets a comparison

Through my studies I have highlighted that contagion is prominent in times of shock, more
so this relates to negative returns or Bear Markets. For the purpose of this research we
also need to look at the prominence of contagion in Bull markets as following the EU
referendum the FTSE 100 hit a yearly high which in turn could signal contagion. Empirical
studies as discussed above have focused and proven contagion in negative shocks. Ang and
Bekaert (1999) and Longin & Solnik (1995), provide evidence for increased correlation
between international equity markets in Bear markets and decreases in Bull markets over
the period 1965-1995 both short and long term, as discussed in the second part of this
literature review (see Aityan et al.,2010)., this is supported by the view that rational
investors weight more on loses than gains. In times of losses individual are more likely to
follow the herding effect as identified by Nofsinger and Sias (1999). Yao (2002) challenged
this argument by looking at the top 5% of positive and negative returns over the period
1987-2001, there finding suggest that contagion is more prevalent in Bear than Bull Markets
in both developed and developing nations.

2.1.8. Review of Contagion

Empirical studies have proven that contagion does indeed exist in every corner of the globe,
on top of this identifying some of the leading channels of contagion, also recognising that
contagion is more prevalent in Bear markets. By using these finding we can make good
judgment as to the contagion points of the EU referendum timeline as well as the foreign
markets most likely to be hit with contagion. Additionally, Appendix four highlights high

19 | P a g e
678430

short term co-movements during the 07/2008 financial crisis and this can be closely linked
with Appendix three which looks into the 2016 period studied for the purposes of this essay,
again highlighting high volatility around the EU referendum. This is important finding for
portfolio managers in their hope of beating the market, optimal portfolio allocation and
diversification. Section2.2 of this literature review looks at the findings and causes of long
term co movements between global economies in hopes of understanding long term co-
movements within the Eurozone following the events that unfolded in June 2016.

20 | P a g e
678430

2.2.0. Interdependency as a cause of Co-movements

Table One: Selected Economic data for the European Union, Japan and US

Exports of Goods Imports of goods

Countries and services (% of and services (% of
statistics GDP Growth (%) GDP) GDP)
2000 2015 2000 2015 2000 2015
Austria 3.4 0.9 43.4 53.4 42 49
Belgium 3.6 1.4 71.9 84.4 69.2 82.7
Bulgaria 5 3 365 66.5 41.8 65
Croatia 3.8 1.6 36.5 49.4 39.6 46.6
Cyprus 5.7 1.6 70.4 * 67.9 *
Czech Republic 4.3 4.2 48.3 84.5 50.2 78
Denmark 3.7 1.2 44.9 53.3 38.2 46.9
Estonia 10.6 1.1 61.6 79.8 64.9 75.7
Finland 5.6 0.5 42.1 37.3 32.9 37
France 3.9 1.3 28.2 30 27.1 31.4
Germany 3 1.7 30.8 46.9 30.6 39.1
Greece 3.9 -0.2 23.7 30.1 34.7 30.3
Hungary 4.2 2.9 66.8 * 70.5 *
Ireland 10.2 7.8 94.5 121.4 80.7 100.6
Italy 3.7 0.8 25.7 30.2 24.8 27
Latvia 5.4 1.9 36.9 58.8 44.9 60.2
Luxembourg 8.4 4.8 147.5 213.8 121.7 177.6
Malta 6.8 * 90.6 * 98.4 *
Netherlands 4.2 2 66.5 82.8 60 71.5
Poland 4.3 3.6 27.2 49.4 33.6 46.6
Portugal 3.8 1.5 28.2 40.3 39.2 39.6
Romania 2.4 3.7 32.7 41.1 38 41.6
Slovakia 1.2 3.6 54.1 93.8 56.6 91.4
Slovenia 4.2 2.9 50 77.8 53.7 68.5
Spain 5.3 3.2 28.6 33.1 31.6 30.7
Sweden 4.7 4.1 44.1 45.2 38.2 40.9
UK 3.8 2.3 26.3 27.4 28.2 29.4
USA 4.1 2.4 10.7 12.6 14.3 15.5
Japan 2.3 0.5 10.9 17.9 9.4 18.9
World Average 4.33 2.4 26.26 29.14 25.33 28.536
*Source World Bank Data source

21 | P a g e
678430

This section reviews the causality of interdependency; as defined as the existence of cross
market linkages in a time of economic stability (Bonfiglioli & Favero ,2005). Empirical studies
have shown that interdependency has been increasing since the late 50s due to the range
of factors examined now. Appendix one looks at the long term movements of stock markets
on a 10-year period and it is clear that there is evidence of contagion during the financial
crisis, however taking an initial look returns of global equity markets seem to be very spatial
following the EU referendum. Interdependence between markets varies over time so it is
therefore a better comprehension of why national stock markets display different degrees
of interdependence over time and is critical to understanding the benefits to international
portfolio diversification for portfolio managers.

There has been much debate as to the differentiation between interdependence and
contagion as highlighted by Forbes & Rigobon (2002). Challenging findings in section one of
this literature review and identified many developed equity markets already hold high
interdependence and therefore the prominence of contagion is not so influential in these
regions. Forbes and Rigobon (2008) challenged the methodologies of many economists
especially the use of conditional correlations coefficients due to their biases towards
periods of high volatility.

Furthermore, past studies have shown that interdependence has only started becoming
relevant since the 1987 financial crash, Agmon (1972) used weekly returns and identified
some lagging interdependence between UK, USA, German and Japans equity markets
although nothing of significance. Similarly, Von Furstenberg and Jeon (1989) found that only
economic events have significant effects on the stock price changes of four major markets
studied. Claiming that national stock market interdependency may simply be as a result of
contagion events. However, following the stock market crash of 1987 markets have become
progressively more correlated, this shock event has been highlighted in many studies as the
pinpoint for increasing correlation and the decreasing benefits of international
diversification (see von Furstenberg and Jeon 1989, and Bertera and Mayer 1990). Eun and
Shim (1989) challenged this through looking at daily market returns under VAR due to its
ability to look at Multivariate variables as opposed too Bivariate under Grangers causality
test. Finding interdependency between many pairs including that of UK-US and US-Japan;

22 | P a g e
678430

although this was due to time lag between stock exchanges opening and closing, more so
that Japan and UK equity markets responding to yesterdays movements.

As proven above there has been much debate as to if interdependency was present before
the crisis in 1987 but since then studies have highlighted a significant increase in conditional
correlation over the last 3 decades in Europe (see Bracker et al 1999, Aityan et al., 2010 &
Malliaris and Urrutia ,1994). Longin & Solnik (1995) used use extreme value theory in
bivariate monthly models to identify increasing correlations between 19601990. While
Gerrits & Yuce (1999) identified long term interdependency in Europe through the ADF test
to assess if the data is non-stationery and thereby apply Johansen cointergration test to see
strong links between the Dutch, German and UK markets. likewise, Baele and Inghelbrecht,
2010; identified regional market correlations have increased substantially over the last three
decades, reflecting increasing economic and financial integration in 14 leading European
equity markets. In addition, these studies linked interdependency and contagion together,
identify that countries with high interdependence (In this case the Eurozone) will have
greater spill over effects under a contagion event. Baele and Inghelbrecht, (2010) is an
excellent study to support our understanding of spill over effects as well as long run co-
movements in the Eurozone. Studies have identified that interdependence is in fact reality
but at the heart of international diversification is the assumption that markets are not
perfectly correlated. Therefore, it is possible to reduce portfolio risk without sacrificing
expected return via international diversification to reduce or remove unsystematic risks.

2.2.1. Channels of Interdependence Fundamentals

He et al (2014) highlighted a variety of factors leading to long term interdependence of
national equity markets, firstly the studies identified the role of economic fundamentals
building on the finding of Solnik (1974). Baele & Inghelbrecht (2010) studied time varying
integration between 14 European countries over a 35-year period and recognised the
significant role of the single currency in strengthening interdependency further. In addition,
Flavin et al (2002) & Roll (1992) stated this was down to foreign exchange aversion and the
familiarity effect. Sasaki et al (1999) strengthened this finding identifying the impact of

23 | P a g e
678430

monetary and credit policies on strengthening the interrelationship between the securities
markets, this has a significant implication on the Eurozone a monetary union. Nasseh &
Strauss (2000) combined these theories on interdependence in Europe and by using the
cointergration test for the period of 19621995 and identified that Co-movements have
continually increased especially in regards to economic fundamentals such as short term
interests rates and CPIs. More so their studies identified the diversification benefits of
investing in the UK due to not being a part of the single currency. Therefore, the
implementation of a single market in the Eurozone has played a significant factor in
enhancing the interdependence in this region and furthermore significantly reduced benefit
from international diversification, although macroeconomic policies are only a small part of
global interdependence, many studies have focused on the role of trade and foreign
investment.

2.2.2. Channels of Interdependence - Trade and Investment

Kaminsky (2000) identified the role of the cross border financial interlinkages & Services,
more so the importance of the lending and credit function in rationing credit and financial
institutions susceptibility to halt investment following crisis or shock periods. On the note of
financial interlinkages rvai at al (2009) identified the presence of contagion due to
significant increases in foreign ownerships of the banking systems in CESE particularly their
reliance & exposure to Western Europe banks and financial services whereby credit from
the CESE region has grown more than 23% annually.

Arshanapalli & Doukas (1993) & Aloui et al. (2011) highlighted the role of trade and
multinationals in increasing co-movements of international equity markets in the long term.
Consequently, equity markets have become more integrated due to multiple market listings
in addition to global exposure to economic events. Eiling & Gerard (2014) discussed the role
of GDP as a catalyst for trade where growth stimulates trade, resulting in financial
interlinkages. Table one goes some way to highlighting this, identifying key economic
statistics from our sample and in turn that Economic growth has led to increased trading
links, it also recognises that trading links has increased since the turn of the century and as a
result global interdependence will have increased. Liu et al (2006), brought together a

24 | P a g e
678430

variety of studies looking at trade in global regions, by using cointergration tests the findings
established that trade isnt a significant factor in many parts of the world (Eg,Asian region)
however the European peninsular were significantly interdependent as a result of imports.
Furthermore, areas with closer trading patterns should have more predictable and
interdependent returns with one another. This is as significant finding as Appendix Five
highlights the main trading partners of the countries studied in this essay, it shows that
many Eurozone countries are heavily dependent of their Eurozone counterparts particular
as you look at the least developed of these nations such as Romania (81% reliant on
European Imports) but likewise many of the most developed nations are still heavily reliant
on European Imports Germany (70%), the most important of these nations for this study is
the UK, who are 60% reliant of Europe for both imports and exports, by using the finding we
can be assured that the UK and Europe are heavily interdependent on each other on the
basis of trade. Additionally, Roll (1992) studies identified that markets are more
independent on the basis of their industrial composition as opposed to the role of trade.
Much like Liu et al (2006), Bracker (1999) finding supported trades but also placed emphasis
on geographical distances & locations.

2.2.3. Channels of Interdependence - Geographical Location

Bracker (1999) stated that interdependence has also been driven by geographical locations
and that is still the case for many of the developing regions of the European Union. In this
case they are more susceptible to long term interdependencies & contagion, however with
the world getting smaller, geographical location is becoming less prevalent for some of the
most diverse & developed economies as identified in Appendix Five (ie US, UK that a highly
interdependent globally). Groenen and Franses (2000) & Heaney et al. (2000) support this
view identifying that stock markets cluster on a regional basis due to trade, while Flavin et al
(2002) & Portes & Rey (1999) argued that trade is only a relevant driver for goods markets
while physical location and trading costs should be less of an issue for equity market.
Therefore, challenging the ideology that trade plays a factor in interdependence. It was
found that physical location did still matter in particular that of overlapping market hours
and therefore countries of close proximities tend to have increased interdependence. This
relates to studies raised Karolyi & Stulz (1996) who identified significant overnight returns

25 | P a g e
678430

between Japan and US Markets, more so highlighting that traders base investment decision
on what else is happening in the global economy (Herding effect). Appendix Six shows the
market open times for the sample studied and shows us under Flavins research European
markets will be highly interdependent as they observe intra-day news, but also any market
that is open during a similar time period (King & Wadhwani 1990). On top of the role of
trading hours Flavin highlights the role of the familiarity effect or home bias on investors
behaviour, in short this is investors tendency to invest in local markets and has significant
implication on the developing nation interdependence as UK investors are reluctant due to a
lack to familiarity as well as lack of market credibility, regulation & additional information
costs in many regions, particularly Eastern Europe.

2.2.4. Channels of Interdependence - Technology

Flavin et al (2002) also argued that home bias is becoming less influential in investor
behaviour and its prevalence in causing interdependence. Suggesting increasing
interdependence of markets is down to decreasing information cost and increased
technology levels. Flavin states the ability for market participants to gather information and
trade simultaneously at low cost has rendered geographical location irrelevant besides the
presence of the familiarity effect & tax considerations. Similarly, Arshanapalli, (1993) states
the importance of technology in reducing the cost of cross border information and in turn
allow financial institutions and portfolio managers to invest in further reaching markets.
Conversely Portes & Rey (2005) highlights there is still an element of information asymmetry
and a lack of transparency in many markets, particularly emerging markets & this coupled
with the familiarity bias & ambiguity aversion means many of these markets give greater
diversification benefits but at a cost due to the increased risk associated with un-stable and
inefficient markets. However, the growth in information and technology is a supporting
factor for many other channels such as cross-market hedging, without transparent
information and the ease of carrying out transactions, global markets would simply be less
interdependent upon one another and there would be less of a presence of cross-border
deposits & likewise less international diversification from portfolio managers leading to

26 | P a g e
678430

cross-market hedging. Ultimately the benefits of international diversification will continue

to decrease as information inefficiency decreases and markets develop.

2.2.5. Correlation in Emerging markets

Empirical studies have highlighted that historically Emerging markets offer attractive
investment opportunities due to their low correlations. However, studies have proved
correlations significantly increase following shock events (see Hwang, J. (2014) Zhang, B., &
Li, X. (2014)). Wasim (2013) states this is as a result of their dependency on the developed
world but more so emerging economies are more susceptible to contagion during shock
events. Additionally, as well as short term contagion long term interdependency is changing
for developing markets. Sasaki et al. (1999) identified the role of market liberalisation, this is
important for the Eastern and Central European economies in the Eurozone. Empirical
studies on emerging markets particularly that of the BRICS & CESE have been studied,
identifying as economies move from central planned to free-market economies their
interdependence increases extensively. Studies by Zhang & Li (2014) and He et al (2014)
supports this by looking at coefficient of determinates, DCC-Garch and Johansens
cointergration test on daily stock returns over a 12-year period. Arguing that as markets
liberalise it allows for increased cross-market deposits, trade as well as further technology
and information efficiency which are key channels of not only interdependency but also
ofsusceptibility to contagion. Additionally, they studied China as they liberalised between
2001 and 2009 and identified that market openness led to further exposure to volatility spill
overs particularly when foreign investment was opened up in 2002. Eiling & Gerard (2014)
studied market development in Eastern Europe and it implications on interdependency the
studies found that correlation in the region has increased from 0.1 in 1990 to 0.8 in 2014,
providing empirical evidence to support the argument above. Eiling highlighted this was
down to increased cross-border deposits & cross-market hedging because of relaxed foreign
investment as well as the increased information transparency, thereby reducing portfolio
managers reluctance to invest in riskier markets. Empirical studies have shown that as
markets develop so does their exposure to shocks & as a result there is a greater presence
of global interdependence. Using this we can understand that economies in Eastern Europe

27 | P a g e
678430

will have experienced a rapid increase in interdependence and susceptibility to contagion

from the UK.

2.2.6. Review of Interdependency

The Growth of Interdependence has made countries more susceptible to contagion (See
Baele & Inghelbrecht 2010) and the factors discussed are critical to understanding long term
interdependence in the region, ultimately research suggest that the European economies
are heavily interdependent on one another and as a result susceptible to contagion. It also
helps us understand the long term relationship between the UK and Europe. Following the
EU referendum announcement on the 22nd June $65 Billion of foreign investment has been
halted and there is still uncertainty surrounding the free trade movement following the
LEAVE result. In turn these could reduce trade links, the role of financial services and
therefore the prevalence of interdependence within the Eurozone particularly with Eastern
Europe. However, the development of the UK equity markets, location and information
efficiency could mean the UK is still highly correlated with its European counterparts.

28 | P a g e
678430

3.0.0. Methodology

Past studies have used a range of methods to assess the relationships between market
returns. The most popular methodologies can be categorized into four groups (1) cross-
market correlation coefficients, (2) ARCH & GARCH model, (3) cointergration and Granger
causality analysis, and (4) VAR & VECM. This research first employs Pearson correlation
coefficient to understand the effects of the event on linear correlations. Then using PCA to
see if the event has led to any change in grouping of the 28 countries examined. Following
this we then look to access the long run relationship of the sample markets through the use
of Johansens cointergration test and the Grangers causality test to look into linear spill-over
effects before and after the event. Lastly we use unrestricted VAR on non-cointergrated
markets and VECM on cointergrated markets to access linear interdependencies of the
entire sample size before and after the event to access the markets dynamic relationship

3.0.1. Correlation coefficient

Pearson correlations coefficient has empirically been used (E.g. Meric, 2015; Connolly &
Wang, 2003; Syllignakis & Kouretas, 2011) to access the linear relationships over a given
time period, denoted by;

(EQ1)

Pearsons correlation coefficient gives an absolute result as opposed to other methods such
as Grangers causality test & Johansens cointergration test and a basic understanding that
can be further tested on the relationship between European equity markets. However,
Forbes & Rigobon (2002) argued it is biased towards periods of high volatility and skewed
heavily I favour of higher correlation due to its reliance on standard deviation and is
therefore not a suitable measure for proving contagions existence. This is why the finding of

29 | P a g e
678430

correlation coefficient is supported with other methods in the form of cointergration,

Grangers and VAR methodologies, which counter this bias towards volatility. Additionally, it
only accounts for linear relationships and fails to see that there may be more variables
influencing market returns and to solve this we later look VAR & PCA that understands these
complex relationships markets hold. Likewise, Correlation coefficients fail to identify the
direction of the spill over effects which is why Grangers causality test in employed to
evaluate this.

3.0.2 Principle component analysis:

PCA is a data reduction technique that allows us to identify groupings and to understand co-
movement patterns (see, e.g., Philippatos et al. (1983) Meric and Meric (1989))This research
performs PCA on the natural log differences using Keisers rule, whereby statistically
significant principal components with an Eigen value greater than 1 are extracted for
analysis. Additionally, we use the Varimax rotation to maximize the factor loadings of the
stock markets in each principal component with similar movement patterns as well as
allowing for non-orthogonal solutions. The PCA technique groups markets in terms of their
similarities and patterns under principle components. The stock markets that are highly
correlated would have high factor loadings in the same principal component. Past studies
have highlighted that following events there is a reduced number of significant components
so using this method can help us understand if the EU referendum has caused any increased
grouping or conversely following the correlation results any reduced grouping with that if
the UK; this can indicate interdependence groupings and areas to avoid for investors hoping
to diversify.

3.0.3 Johansens cointergration test

Cointergration analysis exists to understand whether a long run equilibrium relationship
exists between variables. There are 2 main models to test for cointergration; Arshanapalli, &
Doukas (1993) use Engle-Grangers two step cointergration test but empirically many
researchers have focused on the use of Johansens cointergration test (see, Yarovaya & Lau,
2016; Zhang & Li, 2014; Humpe & Macmillan, 2009; Liu, Lin & Lai, 2006) This is because the

30 | P a g e
678430

test permits more than one cointergration relationship so is more generally applicable than
the EngleGranger test which is based on the DickeyFuller test for unit roots in the
residuals from a single (estimated) cointergration relationship. We use Johansens
cointergration test (see Johansen 1988) to understand if there is significant
interdependencies between the sample markets and if so how the region has changed
following the EU referendum. Markets said to be cointergrated with the UK will not offer
the same diversification benefits as those without cointergration relationships, therefore if
we can spot the outliers we can identify markets that offer statistically significant
diversification benefits to investors looking for to reap the rewards of international
diversification.

Consider g I(1) variables (g 2). The VAR(k) model of these variables is:

EQ (2)

Where Y represents a g 1 vector and i a g g matrix of coefficients. To perform the

Johansen test, EQ(2) needs to be converted to the following VECM:

EQ (3)

There are two types of the Johansen cointergration tests: one is the eigenvalue trace test
and the other is the maximum eigenvalue test. The formula to calculate the trace test
statistics is:

EQ (4)

The maximum eigenvalue statistic which the null hypothesis of cointergration relations

against the alternative of cointergration relations. This test statistic is computed as:

31 | P a g e
678430

EQ (5)
for .

Five models are possible in conducting a cointegration test: For the purposes of this
research we employ Model 3 where series have a certain linear trend but the cointegration
equation only has an intercept term; as in line with empirical studies on cointergration of
international markets but also because past findings access that all in all markets conform to
some kind of probability distribution over the long term time horizon.
EQ (6)

Where yt is univariate, and Xt is a vector. If t NT (t = 1,, T), Dt() = 1; otherwise, Dt() = 0.

= TB.T
And TB represents a possible break point.
We test all values to the 5% significant level where the null is the number of cointergration
equations, when probability is more than the 5% significant value we can accept the null
equations to determine the number of cointergration equations for that sample.

Markets that exhibit no cointergration under the test are followed up under VAR while if our
sample are cointergrated we will be using the VECM to assess relationships further. Given
our sample size the results should be accurate as Johansens cointergration test suffers from
asymptotic properties. We apply a non-stationery data set established later in the research
as well as using the optimum lag function later established

32 | P a g e
678430

3.0.4. Grangers Causality test

Grangers test has been used to capture linear interdependencies between bivariate series
(see Scheicher 2001, Huang, Yang & Hu 2000). It can be used to identify the direction of
causal linkages and can be very useful to identify spill over effects of or on the UK,
something that both coefficients and cointergration tests do not identify. Additionally, we
use Grangers test to understand the region and markets that may not be subjective to
interceptions in the region and as a result could be great opportunities for diversification.
Two assumptions are made when forming grangers causality test:
(i) The future cannot cause the past. Strict causality can only occur with the past causing the
future.
(ii) A cause contains unique information about an effect that is not available elsewhere.

To determine Grangers test we first run bivariate regressions of the form:

(EQ7)

For all possible pairs of series in the group. The reported F-statistics are the Wald
statistics for the joint hypothesis:

(EQ8)

For each equation. The null hypothesis is that does not Granger-cause in the first

regression and that does not Granger-cause in the second regression (see Granger, C. W
1969). Tested at a 1-10% significant value. Therefore, those with a probability less that the
10% significant are stats to have causal linkage. One weakness in Grangers model is there is
no absolute answer, more so a simple yes or know to the implications of linear
interdependencies and no means of comparison as to if causal linkages have improved over
the time period.

33 | P a g e
678430

3.0.5. VAR & VECM

Empirically Grangers test & Pearsons Correlation Coefficient has gone out of fashion as it
only compares two variables at once which is inconvenient with a data sample of 28
markets therefore can produce misleading results if the true relationship of X & Y involves
three or more variables, likewise the standard Grangers test does not capture the impact of
positive and negative shock on causal links between variables, which is significant to our
studies during the time period that exhibit both bull and bear markets following the EU
referendum. Because of these weakness studies by Yarovaya, & Lau. (2016) have adapted
the methodology employed.

Therefore, we apply VAR (See, Beirne et al 2010; Yang et al 2006; Eun & Shim, 1989) that
can be applied to multivariate data series to test for linear interdependencies of multiple
time series. VAR can only be applied on stationery data which we test for under the ADF and
PP test that look for unit roots in data. VAR is applied to those markets that do not have
long term cointergration as established by Johansens cointergration test and is denoted by:

(EQ9)

where is a vector of endogenous variables, is a vector of exogenous variables,

and are matrices of coefficients to be estimated, and is a vector of

innovations that may be contemporaneously correlated but are uncorrelated with their own
lagged values and uncorrelated with all of the right-hand side variables.

VECM is a restricted VAR designed for use with non-stationary series that are known to be
cointergrated through Johansens cointergration test. Testing for cointergration using an
estimated VAR object, Equation object estimated using nonstationary data & understanding
of the number of cointergration relationships through Johansens cointergration testing. The
VECM model has been applied a lot in empirical studies (see Humpe & Macmillan 2009) to
test for increased linear interdependencies events studies in highly interdependent regions.

34 | P a g e
678430

4.0 Data & results

In addition to the UK stock market we study 28 other countries in the form of the entire EU
as well as the US and Japan due to their globally significant markets. The indexs used can
been found in appendix 7. Data is derived from the Bloomberg database at local currency
for the period January 4th 2016December 31st 2016 which is split into 2 data sets covering
pre event (December 30th 2015June 22nd 2016) and post event (June 23rd 201611th
January 2017). Therefore, the study looks at 263 daily closing prices for the year and any
discrepancy in days for markets is uses the past days closing prices. The daily returns are
computed as natural log differences denoted by:

LN(x)LN(x-1) where x is the daily index closing price

Table 2 demonstrates the descriptive statistics on market indexes. The majority of market
indexes were negatively skewed besides that of UK this tells us that negative returns are
more common than positive returns over this time period. Research has told us that
contagion is more prevalent in Bear markets so this can be significant in our findings of
contagion. Kurtosis highlights that many markets, significantly the likes of Spain and Ireland
experience sharp peaks in stock indices returns. The JarqueBera test indicates that returns
are not normally distributed due to rejection of the Null hypothesis for all series. Lastly the
standard deviation tells us about the volatility of the stock market and this is significant for
our calculation of Pearsons correlation coefficient and as discussed earlier its reliances on
market volatility.

35 | P a g e
678430

Table 2: Descriptive statistics (natural log differences)

N Mean Std. Deviation Skewness Kurtosis Jarque-Bera
Series Statistic Statistic Statistic Statistic Statistics P-Value
UK 261 .0005753 .0104446 .009 1.341 18.17 0.00011
Japan 261 -.0000083 .0163755 -.148 4.750 234.72 0.00000
US 261 .0003747 .0081249 -.438 2.599 77.67 0.00000
Sweden 261 .0002125 .0119061 -1.219 8.612 835.71 0.00000
Spain 261 -.0000938 .0161860 -2.173 16.304 2976.57 0.00000
Slovenia 261 .0002648 .0065599 -.437 1.575 33.38 0.00000
Slovakia 261 .0003293 .0104302 -.160 1.219 16.05 0.00033
Romania 261 .0001391 .0087606 -.878 6.456 466.17 0.00000
Portugal 261 -.0000410 .0119726 -.850 3.265 140.99 0.00000
Poland 261 .0005469 .0097901 -.533 1.935 50.48 0.00000
Netherlands 261 .0003320 .0119076 -.412 2.475 70.19 0.00000
Malta 261 .0002117 .0041588 -.268 1.635 30.28 0.00000
Luxembourg 261 .0008044 .0144370 -.294 1.827 37.76 0.00000
Latvia 261 .0008754 .0089855 1.619 10.035 1161.62 0.00000
Italy 261 -.0003787 .0181305 -1.232 8.286 779.73 0.00000
Ireland 261 -.0001355 .0141025 -2.284 15.227 2643.39 0.00000
Hungary 261 .0012226 .0101743 -.496 1.776 42.72 0.00000
Greece 261 .0002742 .0200348 -1.613 11.301 1442.9 0.00000
Germany 261 .0002677 .0130060 -.790 3.389 145.39 0.00000
France 261 .0001695 .0131354 -.999 6.373 464.87 0.00000
Finland 261 .0001161 .0120806 -1.225 7.720 684.44 0.00000
Estonia 261 .0007341 .0054211 -.079 1.864 35.73 0.00000
Denmark 261 -.0002399 .0126093 -.373 2.470 68.58 0.00000
Czech 261 -.0001131 .0105795 -.667 2.062 62.62 0.00000
Croatia 261 .0007934 .0048250 -.187 1.733 31.11 0.00000
Bulgaria 261 .0010810 .0071076 -.031 11.370 1347.29 0.00000
Belgium 261 -.0001194 .0114678 -.850 4.453 236.37 0.00000
Austria 261 .0004138 .0135785 -.899 3.941 195.3681 0.00000
Cyprus 261 .0001401 .0076364 -.137 2.654 56.79 0.00000

4.1. Pearson Correlations coefficients

We construct 841 correlations for the pre and post period samples and the full details can
be found in appendix 8 & 9; all with the exemption on Bulgaria, Latvia, Malta, Slovakia &
Slovenia are significant at the 5% critical level so we can therefore cannot reject the null
hypothesis that correlation between these markets is different to 0. Therefore, we perform
Grangers causality test to further understand the relationship between Poland and its
counterparts. Table 3 identifies the correlation coefficients between the UK and sample for
the pre and post period and clearly identifies a shock period with a significant drop in

36 | P a g e
678430

correlation of 26.37% between the 6 months before and after he EU referendum. As

expected the UK is highly correlated with developed, regional markets such as Germany,

Table 3: Correlation of the UK Stock Market with the Other Stock Markets
Correlation coefficents
Stock Market Pre Period Post period %change
Japan .284** .202* -28.87
US .617** .604** -2.11
Sweden .875** .689** -21.26
Spain .853** .662** -22.39
Slovenia .326** .099 -69.71
Slovakia .074 .004 -95.21
Romania .502** .416** -17.13
Portugal .799** .744** -6.88
Poland .568** .500** -11.97
Netherlands .930** .827** -11.08
Malta -.076 .121 -258.48
Luxembourg .657** .602** -8.37
Latvia .143 -.031 -121.49
Italy .826** .627** -24.09
Ireland .780** .658** -15.64
Hungary .506** .524** 3.56
Greece .429** .494** 15.15
Germany .834** .736** -11.75
France .919** .791** -13.93
Finland .853** .670** -21.45
Estonia .088 .254** 190.03
Denmark .775** .583** -24.77
Czech .628** .450** -28.34
Croatia .231* .212* -8.23
Bulgaria .291** .120 -58.78
Belgium .887** .761** -14.21
Austria .822** .621** -24.45
Average 0.571 0.479 -26.37

France, Netherlands and Belgium which is in line with the finding of Groenen and Franses
(2000). What is more interesting to see is the benefits of diversification in Central and
Eastern European countries, particularly that of Latvia and Malta. However as identified in
past studies these regions are associated with greater uncertainty due to their lack of
development and in turn this is why many investors veer away from such markets so will be
heavily skewed as a result of these factors (ambiguity aversion). These finding show a clear
indication of a significant drop in correlation and its European counterparts with all bar

37 | P a g e
678430

Hungary and Malta showing a decrease in correlation empirical evidence of the presence of
contagion following the EU referendum.

Table 4: Most and Least Correlated Stock Markets

Most correlated Stock Markets
Pre Period Post Period
Correlation Correlation
Country Pair Ceofficent Country Pair Ceofficent
France - Netherlands 0.964** France - Germany 0.946**
France - Belgium 0.945** France - Netherlands 0.944**
Belgium - Netherlands 0.94** Sweden - Finland 0.927**
UK - Netherlands 0.93** Netherlands - Belgium 0.923**
France - Spain 0.923** Spain - France 0.921**
UK - France 0.919** Belgium - France 0.921**
Spain - Italy 0.916** Netherlands- Germany 0.914**
Spain - Netherlands 0.907** Belgium - Germany 0.891**
Germany - France 0.904** Spain - Italy 0.881**
France -- Sweden 0.902** France - Italy 0.866**
Average 0.925 Average 0.913
Least Correlated Stock Markets
Pre Period Post Period
Malta - Estonia -0.218 US - Slovakia -0.071
Germany - Malta -0.188 Slovakia - Bulgaria -0.055
Malta - France -0.133 Slovakia - Greece -0.055
Malta - Netherlands -0.127 Slovakia - Ireland -0.044
Malta - Japan -0.124 Slovakia - Poland -0.038
Malta - Portugal -0.119 UK - Latvia -0.031
Malta - Poland -0.114 Slovakia - Belgium -0.026
Malta - Ireland -0.109 Slovakia - Portugal -0.025
Malta - Belgium -0.104 Slovakia - Italy -0.021
Malta - Finland -0.092 Slovkia - Austria -0.012
Average -0.133 Average -0.038

Tables 4 & 5 give a more accurate representation of the correlation matrix. Again showing
evidence of weakening interdependence in the region. Table 4 gives us the strongest and
weakest pairs, as a result the weakest pairs mainly forming of Malta are the most attractive
jurisdictions to invest in to optimise the diversification of a portfolio. Similarly, the most
correlated markets offer little to diversification. Additionally, Table 5 identifies the best and

38 | P a g e
678430

Table 5: Average Correlation of each Stock Market with the other Stock Markets

Average Correlation Coefficent

Stock Market Pre Period Post Period % Change
UK 0.5541 0.4621 -16.60
Japan 0.2644 0.2611 -1.27
US 0.3641 0.4146 13.87
Sweden 0.5543 0.4110 -25.84
Spain 0.5680 0.5174 -8.91
Slovenia 0.2564 0.1184 -53.81
Slovakia 0.0531 0.0050 -90.58
Romania 0.3691 0.3403 -7.80
Portugal 0.5348 0.4991 -6.67
Poland 0.3927 0.3707 -5.62
Netherlands 0.5729 0.5351 -6.60
Malta -0.0660 0.0783 -218.70
Luxembourg 0.4117 0.4128 0.26
Latvia 0.1252 0.0512 -59.09
Italy 0.5419 0.5056 -6.70
Ireland 0.5180 0.4845 -6.47
Hungary 0.4143 0.3904 -5.78
Greece 0.3721 0.4358 17.14
Germany 0.5325 0.5337 0.23
France 0.5761 0.5522 -4.15
Finland 0.5566 0.4015 -27.86
Estonia 0.1496 0.1919 28.24
Denmark 0.5220 0.3951 -24.33
Czech 0.4965 0.3795 -23.57
Croatia 0.2192 0.1783 -18.68
Bulgaria 0.2229 0.0848 -61.97
Belgium 0.5749 0.5316 -7.54
Austria 0.5409 0.4949 206.25
Cyprus 0.1528 0.2123 206.25
Average 0.405 0.366 -7.72

worst markets to invest in following the EU referendum; a high average correlation indicates
that a stock market is very well integrated with the other stock markets. Such Stock market
is not a good prospect for portfolio diversification (eg France, Belgium). Conversely Investors
can maximize the portfolio diversification benefit by investing in the stock markets with a
low average correlation coefficient with the other markets (Ie Malta, Slovakia and Latvia).
These findings are significant as they oppose the findings of rvai at al (2009) who found

39 | P a g e
678430

significant interceptions in the CESE region, albeit that is not the case as many of these
countries still offer significant benefits to investors looking to diversify their portfolios

These findings are not in line with the theory of contagion that identifies significant
decrease in correlation in Europe following the EU referendum (average of -7.72%). This
could signal the immediate impacts of the EU Referendum and the UK separation with its
European counterparts in the form of leaving the trade union as examined by Arshanapalli &
Doukas (1993) & Aloui et al. (2011) as a channel of contagion. It could also be as a result of
significant global uncertainty in the 6-months following the event. However, these findings
have serious implications for investors both in and outside of the UK looking to benefit from
international diversification as it is clear from the correlations that the UK is significantly less
correlated with the sample market post referendum which signals great benefits for UK
investors developing international portfolios.

4.2. Principle component analysis

For the purposes of this paper we analyse PCA under the Keisers rule, thereby extracting
any component with an Eigen value greater than one. We initially use KMOs and Bartletts
test to analyse is there are any significant components that can be drawn from the data,
both prove significant and results for the pre and post period can be found in Appendix 10,
Additionally the complete results of components and factor loadings can be found at
appendix 11 & 12. PCAs findings are exceptionally useful when it comes to VAR/VECM later
on to understand correlated groupings.

Table 6 highlight 5 principle components for the pre event period; bold indicates this is a
significant component while standard indicates it is still significant enough to be accounted
for in other components. More importantly it identifies 2 key components; (1) which
signified independence in the Western world & (2) that groups many of CESE. While the
following 3 components account for a much smaller size and especially component 5 which
forms of the 2 smallest markets by market capitalization. More significantly is the variance
these 5 components account for, in some markets it explains over 90% of the variance (Ie
France, Netherlands) and clearly supports empirical evidence of high interdependence in
the region, but likewise these are markets to avoid.

40 | P a g e
678430

Table 6: PCA Pre Period

Rotated Component Matrix - Pre Period

Component & Factor Loadings Total Variance
Explained in these
1 2 3 4 5 components
France .952 0.939
Netherlands
.946
0.929
Belgium .936 .209 0.927
UK .923 0.876
Sweden .916 0.872
Spain .908 .203 0.879
Italy .903 0.841
Finland .893 0.847
Germany .891 0.836
Austria .869 0.809
Ireland .837 .226 0.755
Portugal .833 .253 0.787
Denmark .816 .281 0.754
Luxembourg .699 0.524
Czech .689 .347 .210 0.664
US .623 .396 0.566
Hungary .509 .303 .359 .219 0.551
Poland .495 .373 .308 0.513
Romania .467 .354 .317 0.471
Croatia .674 0.574
Slovenia .214 .646 0.469
Japan .232 .606 .286 0.525
Greece .441 .575 .217 0.6
Latvia .764 0.615
Bulgaria .284 .479 .296 0.434
Slovakia .852 0.74
Estonia .398 .294 .553 0.603
Malta .823 0.704
Cyprus .276 .347 .251 .486 0.498
% of 44.45% 9.65% 5.86% 4.92% 4.45%
variance
Explained
Cumulative 44.45% 54.10% 59.96% 64.88% 69.33%
%

41 | P a g e
678430

Table 7: PCA Post Period

Rotated Component Matrix - Post period
Component & factor Loadings Total Variance Explained
Stock Market 1 2 3 4 5 6 7 in these components
Netherlands
.938
0.919
Belgium .921 0.907
France .919 .253 0.946
Germany .899 0.882
Ireland .861 0.787
Portugal .839 .242 0.791
UK .817 0.735
Spain .814 .399 .231 0.892
Sweden .802 .413 0.875
Finland .797 .386 0.858
Austria .796 .228 0.773
Italy .788 .341 .284 0.837
US .771 0.618
Luxembourg .752 0.609
Denmark .704 0.568
Poland .581 .332 .247 0.524
Greece .573 .368 .518 0.771
Hungary .534 .391 .283 0.571
Czech .529 .351 .272 0.514
Japan .215 .754 .109 .231 0.692
Romania .360 .584 .258 .207 0.593
Slovakia .523 .443 -.411 0.7
Croatia .723 0.598
Cypus .204 .723 0.606
Estonia .742 0.628
Bulgaria .846 0.771
Slovenia .356 .283 .408 .307 0.483
Malta .921 0.856
Latvia .980 0.966
% of 46.01% 7.52% 4.64% 4.24% 3.85% 3.49% 3.47%
variance
Explained
Cumulative 46.01% 53.52% 58.16% 62.40% 66.25% 69.74% 73.21%
%

Table 7 identifies the Post event PCA; again the principle component analysis of post event
markets identifies 2 significant components accounting for region interdependencies that in
turn should be avoided if looking to invest in volatile times. More significantly though is the
shift of countries we now see each component account for less variance in market returns
and these finding support that of our Pearson correlation coefficient in that, the 6 months
following the EU referendum European co-movements are weaker. This can be identified
from the increase in principle components and the shifts of economies particularly in

42 | P a g e
678430

smaller markets and CESE. These results provide great evidence for the shift in regional
interdependence as well as outlier markets such as that of Latvia and Malta that lack
correlation with any economies in the sample selected, these markets offer the best
opportunities for diversification for any European Investor.

4.3. Stationery testing

For the purposes of our research we must apply non-stationary data samples to the
Johansens cointergration test and stationery data sample on VAR/VECM & Grangers
causality test. First we must identify stationery and non-stationery data, to do this we
employ the ADF (see appendix 13). We double check for unit roots under the PP test (see
appendix 14) as it accounts for heteroscedasticity and autocorrelation consistency unlike
that of the Dicker Fuller test, more so it has no underlying assumptions. However similar to
Johansens cointergration test it is based on the asymptotic assumption. We first apply
these test to natural log differences and both tests highlight this data has no unit roots and
is therefore stationery (See table 8). We then compute these 2 tests on daily closing prices
identifying Unit roots for all markets at the 5% significant level therefore this data is non-
stationery (see Table 9).

4.4. Lag lengths

Lag Lengths are an important component for VAR, Cointergration and causality testing;
using the tests identified in Appendix 15. Finding an optimum lag length of 1 for non-
stationery data as expected given our research into the opening and closing times of
predominantly EU equity markets. Secondly an optimum lag length was established of 0 for
the stationery data set (Full results can be seen from table 10). Unlike empirical studies we
choose the optimum lag length by choosing the most favourable result of all tests.

43 | P a g e
678430

Table 8: Stationery testing results for Log Differenced data

ADF - Log Dif Data PP- Log Dif Data

Market T- Stat Prob* Market T- Stat Prob*
Austria -9.43679 0.0000 Austria -14.81282 0.0000
Belgium -10.00182 0.0000 Belgium -14.32698 0.0000
Bulgaria -13.17749 0.0000 Bulgaria -15.94646 0.0000
Croatia -14.44304 0.0000 Croatia -14.55867 0.0000
Cyrpus -15.52354 0.0000 Cyrpus -15.52135 0.0000
Czech -14.88762 0.0000 Czech -14.88890 0.0000
Demark -16.14619 0.0000 Demark -16.15778 0.0000
Estonia -16.60836 0.0000 Estonia -16.62096 0.0000
Finland -15.56346 0.0000 Finland -15.63800 0.0000
France -15.02645 0.0000 France -15.07380 0.0000
Germany -14.77439 0.0000 Germany -14.72341 0.0000
Greece -14.60652 0.0000 Greece -14.56189 0.0000
Hungary -15.48648 0.0000 Hungary -15.55735 0.0000
Ireland -10.14420 0.0000 Ireland -13.31648 0.0000
Italy -16.28922 0.0000 Italy -16.37239 0.0000
Japan -17.77131 0.0000 Japan -17.73643 0.0000
Latvia -18.27463 0.0000 Latvia -18.34751 0.0000
Luxembourg -15.68670 0.0000 Luxembourg 15.75464 0.0000
Malta -16.60543 0.0000 Malta -16.60449 0.0000
Netherlands -9.53018 0.0000 Netherlands -14.15168 0.0000
Poland -11.96833 0.0000 Poland -13.79018 0.0000
Portugal -9.86953 0.0000 Portugal -13.41450 0.0000
Romania -11.49895 0.0000 Romania -16.56202 0.0000
Slovakia -20.50169 0.0000 Slovakia -22.86588 0.0000
Slovenia -16.63111 0.0000 Slovenia -16.62751 0.0000
Spain -15.18100 0.0000 Spain -15.20836 0.0000
Sweden -16.68590 0.0000 Sweden -16.97226 0.0000
UK -15.15070 0.0000 UK -15.23120 0.0000
US -17.85129 0.0000 US -17.85849 0.0000
*Mackinnon (1996) One-sided P-values *Mackinnon (1996) One-sided P-values

44 | P a g e
678430

Table 9: Stationery Testing Results for Daily Closing Prices

ADF - Closing prices PP- Log Dif Data - Closing Prices

Market T- Stat Prob* Market T- Stat Prob*
Austria -0.68662 0.8469 Austria -0.64305 0.8573
Belgium -2.85138 0.0527 Belgium -3.47580 0.0094
Bulgaria 2.26551 1.0000 Bulgaria 3.09938 1.0000
Croatia 1.64047 0.9996 Croatia 1.33108 0.9988
Cyrpus -2.84653 0.0533 Cyrpus -3.06674 0.0304
Czech -2.92678 0.0437 Czech -3.21694 0.0201
Denmark -2.91959 0.0445 Denmark -3.10133 0.0277
Estonia -0.50880 0.8860 Estonia -0.45113 0.8969
Finland -1.90258 0.3309 Finland -1.99359 0.2897
France -2.12291 0.2359 France -2.30080 0.1725
Germany -1.14341 0.6990 Germany -1.35730 0.6030
Greece -1.79078 0.3846 Greece -1.90359 0.3304
Hungary 0.17862 0.9708 Hungary 0.32380 0.9792
Ireland -2.67193 0.0804 Ireland -3.12975 0.0256
Italy -3.46698 0.0096 Italy -3.47239 0.0095
Japan -2.30164 0.1723 Japan -2.30303 0.1718
Latvia -0.80802 0.8148 Latvia -0.67720 0.8492
Luxembourg -0.81846 0.8119 Luxembourg -0.62702 0.8610
Malta -2.07395 0.2555 Malta -2.21819 0.2004
Netherlands -1.96802 0.3010 Netherlands -1.68805 0.4362
Poland -0.30782 0.9205 Poland -0.08781 0.9483
Portugal -2.79311 0.0606 Portugal -3.25981 0.0178
Romania -1.16411 0.6904 Romania -1.27322 0.6425
Slovakia -3.43204 0.0107 Slovakia -3.55365 0.0074
Slovenia -1.42660 0.5691 Slovenia -1.42902 0.5679
Spain -3.40504 0.0116 Spain -3.61226 0.0610
Sweden -1.60636 0.4779 Sweden -1.43683 0.5640
UK -0.57630 0.8721 UK -0.42483 0.9016
US -0.71500 0.8398 US -0.62579 0.8612
*Mackinnon (1996) One-sided P-values *Mackinnon (1996) One-sided P-values

45 | P a g e
678430

Table 10 Lag Length Criteria

Logged Differences Data Results (Stationery)

Lag LogL LR FPE AIC SC HQ

0 25771.42 NA 1.4e-122* -201.1205* -200.7327* -200.9645*

1 26410.51 1133.375 4.5e-122 -199.9883 -188.7435 -195.4657
2 27003.25 921.5316 2.4e-121 -198.4941 -176.3922 -189.6048
3 27647.74 860.9962* 1.2e-120 -197.4042 -164.4451 -184.1482
4 28406.69 847.8919 4.3e-120 -197.2085 -153.3923 -179.5858
5 29315.24 816.2757 1.2e-119 -198.1816 -143.5083 -176.1922

* indicates lag order selected by the criterion

LR: sequential modified LR test statistic (each test at 5% level)
FPE: Final prediction error
AIC: Akaike information criterion
SC: Schwarz information criterion
HQ: Lag
VAR Hannan-Quinn information
Order Selection Criteriacriterion
Endogenous variables: AUSTRIA BELGIUM BULGARIA CROATIA CYPRUS CZECH DENMARK EST...
Exogenous variables: C
Date: 01/21/17 Time: 20:37
Daily
Sample:Closing Prices
12/30/2015 Data results (Non Stationery)
1/11/2017
Included observations: 257

Lag LogL LR FPE AIC SC HQ

0 -35337.83 NA 6.14e+83 275.2282 275.6287 275.3893

1 -28847.42 11465.55 5.11e+64* 231.2639 243.2783* 236.0955*
2 -28153.82 1068.728 1.99e+65 232.4111 256.0394 241.9132
3 -27391.34 1002.797 6.63e+65 233.0221 268.2643 247.1948
4 -26491.60 980.2592 1.47e+66 232.5650 279.4211 251.4082
5 -25321.27 1010.949* 1.19e+66 230.0021* 288.4721 253.5158

* indicates lag order selected by the criterion

LR: sequential modified LR test statistic (each test at 5% level)
FPE: Final prediction error
AIC: Akaike information criterion
SC: Schwarz information criterion
HQ: Hannan-Quinn information criterion

46 | P a g e
678430

4.5. Grangers causality testing:

As explained above we use Grangers causality testing to capture linear interdependencies
between the UK and the other sampled markets. We use an optimum lag length of 1 on the
natural log difference that as identified earlier provide a stationery data set for us to
conduct our research.

This research use Grangers causality test to identify spill over effect and how they changed
over the given time period. Full UK results can be seen in table 11 and highlighted cells
indicate where Null has been rejected and therefore we confirm there are spill-over effects.
Its significant to see how volatile causal linkages are over the 12-month period but also how
there is a lack of causal linkages with most of Europe. We can use these findings to
understand the best and worse markets for UK investors to diversify in, for instance the UK
has very strong causal linkages on Finland over the last 6 month of the year so this market
would not be beneficial. However as emphasised it only captures linear linkages and as seen
in the PCA analysis there are groupings of markets that correlate together which is what
Grangers test fails to identify. On top of this Appendix 16 highlights all causal linkages for
the entire sample size (F-stats & P-value available upon request), this can be valuable to any
European investors trying to understand how markets react to each other. For instance, the
US signifies multiple causal linkages whereby these market respond to US markets
movements the following day and in turn these markets may not be beneficial for investors
looking to diversify away from the US portfolios. The Finding in Appendix 16 again support
my findings in PCA as it signifies the complex relationships Europe markets hold with each
other but also the increased uncertainty in the last 6 month of the year during the EU
referendum. To study the implications of the EU referendum further we must look into the
multivariate interdependencies through first accessing cointergration and then applying
these results to VAR/VECM dependent upon Cointergration results.

47 | P a g e
678430

Table 11: Grangers Causality test results for the UK

Pre-period Grangers Causality testing Post-Period Grangers Causality testing

Markets Tested F-Stat Prob Markets Tested F-Stat Prob
UK Austria 1.66256 0.1998 UK Austria 2.50003 0.1162
Austria UK 0.06856 0.7393 Austria UK 0.13688 0.7172
UK Belgium 0.68374 0.4100 UK Belgium 0.09513 0.7582
Belgium UK 1.17997 0.2796 Belgium UK 3.88007 0.0509
UK Bulgaria 2.26768 0.1348 UK Bulgaria 0.00424 0.9482
Bulgaria UK 3.55049 0.0620 Bulgaria UK 0.94121 0.3337
UK Croatia 8.33545 0.0046 UK Croatia 0.40681 0.5247
Croatia UK 0.07253 0.7882 Croatia UK 1.70952 0.1932
UK Czech 6.82883 0.0102 UK Czech 1.69419 0.1952
Czech UK 0.00810 0.9285 Czech UK 2.30700 0.1311
UK Cyprus 1.27176 0.2618 UK Cyprus 0.51464 0.4744
Cyprus UK 0.00135 0.9707 Cyprus UK 0.23916 0.6256
UK Denmark 8.42650 0.0044 UK Denmark 0.76667 0.3828
Demark UK 4.59625 0.0341 Demark UK 0.05742 0.8110
UK Estonia 7.74458 0.0063 UK Estonia 5.48695 0.0206
Estonia UK 1.13822 0.2882 Estonia UK 0.11474 0.7353
UK Finland 1.09297 0.2980 UK Finland 34.81820 0.0000
Finland UK 0.65216 0.4210 Finland UK 4.75849 0.0309
UK France 0.11133 0.7392 UK France 0.04877 0.8255
France UK 0.02591 0.8724 France UK 1.15954 0.2835
UK Germany 0.71463 0.3997 UK Germany 0.23419 0.6992
Germany UK 0.32497 0.5697 Germany UK 0.39148 0.5326
UK Hungary 2.58341 0.1107 UK Hungary 0.39960 0.5284
Hungary UK 0.44128 0.5078 Hungary UK 3.17021 0.0772
UK Greece 8.91918 0.0034 UK Greece 0.01581 0.9001
Greece UK 0.44128 0.3719 Greece UK 5.22299 0.0238
UK Ireland 0.07410 0.7859 UK Ireland 0.70327 0.4031
Ireland UK 0.18444 0.6684 Ireland UK 0.88720 0.3479
UK Italy 0.10505 0.7464 UK Italy 0.02135 0.8840
Italy UK 0.07575 0.7836 Italy UK 3.47356 0.0645
UK Japan 36.30410 2.E-0.8 UK Japan 1.88792 0.1717
Japan UK 0.35617 0.5518 Japan UK 1.23281 0.2688
UK Latvia 2.31076 0.1312 UK Latvia 9.96731 0.0020
Latvia UK 2.24130 0.0137 Latvia UK 0.02847 0.8663
UK Luxembourg 1.09549 0.2974 UK Luxembourg 6.08908 0.0148
Luxembourg UK 0.13178 0.7173 Luxembourg UK 2.48932 0.1169
UK Netherlands 0.96110 0.3289 UK Netherlands 0.16910 0.6816
Netherlands UK 1.55471 0.2150 Netherlands UK 0.45200 0.5025
UK Malta 0.00958 0.9222 UK Malta 0.14075 0.7081
Malta UK 1.91427 0.1691 Malta UK 2.48932 0.2507

48 | P a g e
678430

UK Poland 5.69773 0.0186 UK Poland 1.18707 0.2778

Poland UK 0.06719 0.7959 Poland UK 0.37570 0.5409
UK Portugal 0.26574 0.0186 UK Portugal 0.43100 0.5126
Portugal UK 0.03113 0.7959 Portugal UK 2.97375 0.0869
UK Romania 1.04348 0.3091 UK Romania 2.56634 0.1115
Romania UK 1.26736 0.2626 Romania UK 5.75263 0.0178
UK Slovenia 3.22044 0.0753 UK Slovenia 1.61836 0.2055
Slovenia UK 0.09423 0.7594 Slovenia UK 1.54883 0.2154
UK Slovakia 2.01931 0.1580 UK Slovakia 0.25569 0.6139
Slovakia UK 0.32377 0.5705 Slovakia UK 1.34365 0.2484
UK Sweden 1.84327 0.1772 UK Sweden 36.19050 0.0000
Sweden UK 0.71707 0.3989 Sweden UK 10.29510 0.0017
UK Spain 0.17775 0.6741 UK Spain 0.01869 0.8915
Spain UK 0.24316 0.6229 Spain UK 6.35608 0.0128
US UK 10.18140 0.0018 US UK 5.47363 0.0208
UK US 0.26975 0.6045 UK US 2.34074 0.1283

Table 12: Johansens test results for the UK to entire sample (Pre period)
Panel Null Hypothesis Trace Test Prob Max. Eigenvalue 5% critical value
UK - Entire Sample r 17 385.9074 0.0001* 58.75736 0.6816
r 18 327.15 0.0002* 55.3892 0.5677
r 19 271.7608 0.0007* 52.1539 0.4383
r 20 219.6068 0.0024* 48.74154 0.3219
r 21 170.8653 0.0103* 37.33809 0.06614
r 22 133.5272 0.015* 34.0741 0.52
r 23 99.452 0.0271* 25.84207 0.713
r 24 73.60995 0.0241* 23.8501 0.4665
r 25 49.75985 0.0327* 18.30844 0.4696
r 26 31.45141 0.032* 16.1707 0.215
r 27 15.28063 0.0538 12.267 0.101
Results - Johansson's cointergration test indicates 26 cointergrating equations for the entire sample

Table 13: Johansens test results for the UK to entire sample (Post period)

Panel Null Hypothesis Trace Test Prob Max. Eigenvalue 5% critical value
UK - Sample r 17 337.0609 0.0415 57.92683 0.7228
r 18 279.1341 0.0861 47.34478 0.9221
r 19 231.7893 0.1029 42.82325 0.9027
r 20 188.9661 0.1205 38.5192 0.8714
r 21 150.4469 0.1401 36.22045 0.7281
r 22 114.2264 0.2001 27.2629 0.9073
r 23 89.96347 0.1726 24.98749 0.7683
r 24 61.95798 0.1797 19.39118 0.7974
r 25 42.5848 0.143 16.77803 0.5991
r 26 25.80677 0.1346 12.97067 0.4549
r 27 12.83611 0.121 10.82939 0.1635

49 | P a g e
678430

4.6 Johansens Cointergration Testing

We employ Johansens cointergration test for the pre and post period to understand
cointergration the period before and after the EU referendum. We use the daily closing
prices that were identified to be non-stationery and also an optimum lag length of 1.
Additionally, Johansson cointergration test is useful to access whether to use VECM or VAR
and progressing forward we will be using VECM for those relationships said to be
cointergrated under Johansens test and VAR for those that are not. We first test the entire
sample for the pre and post period to identify if they are all cointergrated to some extent
however tables 12 & 13 highlight the failure of the test to access all 29 markets. Therefore,
to test relationships further I look at the UK to each market and further this look at
cointergration on a regional basis.

Tables 14 & 15 highlight the results of the UK to individual countries results and clearly
identifies the UKs role in the EU before and after the referendum. My findings are very
much in line with past studies that identified cointergration with many of the developed
markets such as Western Europe both pre and post event (Arshanapalli & Doukas, 1993).
Additionally, there are several markets that have experience structural changes following
the EU referendum, most significantly the US that was no longer cointergrated with the UK
following the referendum. Significantly, markets in Czech, Poland & Greece are now
cointergrated with the UK and in turn will not be as beneficial to UK investors. These
findings can also be used in tangent with our results from Grangers causality test as
Johansens cointergration test does not assess the direction of causal spill overs. Looking
back, we can clearly see that Greece now influences the UK and this is now reflected in
Johansens cointergration test. These 2 tests in conjunctions are very useful to identify how
markets affect one another and therefore the markets to avoid, likewise its clear from both
Johansens, Grangers tests & empirical studies accessed above, that markets such as France,
Germany and the US are ones for UK investors to avoid when looking to benefit from
international diversification.

However, the more significant findings are based on the markets in which the UK is neither
cointergrated before or after the event as this signifies significant benefits for investors
looking to diversify, the results highlight many of these are Central and Eastern Europe (i.e.

50 | P a g e
678430

Denmark, Latvia, Malta) and in turn these support findings from Bracker (1999) that market
commove on a region basis. Therefore, to test cointegration of the markets further I look at
cointergration on a regional basis (Table 16/17). As expected findings clearly support my
research in the literature review that market correlate due to development and region basis
and this is reflected in results from the Western Europe & G10 samples and therefore offer
a lack of diversification benefits (See, Aloui et al2011). It also highlights the regions that can
be beneficial to investors such as Central and South Europe that have no cointegration
equations so would be the best regions to invest in.

51 | P a g e
678430

Table 14 - Johansen-cointergration test results for UK to markets (Pre period)

Panel Null Hypothesis Trace Test Prob Max. Eigenvalue 5% critical value
UK - Austria r=0 18.81502 0.0152 12.06876 0.1081*
r1 6.746262 0.0094** 6.7462623 0.0094

UK - Belgium r=0 20.97537 0.0067 16.00131 0.0263

r1 4.974059 0.0257** 4.974059 0.0257**

UK - Bulgaria r=0 12.63823 0.1288* 10.20056 0.1991*

r1 2.437665 0.1185 2.437665 0.1185

UK - Croatia r=0 11.98611 0.1576* 10.45025 0.1840*

r1 1.535865 0.2152 3.841466 0.2152

UK - Cyprus r=0 13.1461 0.1096* 9.2199 0.2683*

r1 3.926133 0.0475 3.926133 0.0475

UK - Czech r=0 14.58828 0.0681* 12.69103 0.0873*

r1 1.897254 0.1684 1.897254 0.1684

UK - Denmark r=0 11.34287 0.1913* 8.37386 0.342*

r1 2.969011 0.0849 2.969011 0.0849

UK - Estonia r=0 19.35582 0.0124 18.62237 0.0096

r1 0.733449 0.3918* 0.733449 0.3918*

UK - Finland r=0 16.67165 0.0331 14.05472 0.0539*

r1 2.616931 0.1057* 2.616931 0.1057

UK - France r=0 17.04696 0.029 14.2646 0.0691*

r1 3.686459 0.5490* 3.841466 0.0549

UK - Germany r=0 21.12663 0.0064 14.2646 0.0503*

r1 6.879944 0.0087** 3.841466 0.0087

UK - Greece r=0 15.08975 0.0575* 11.55634 0.1284*

r1 3.841466 0.0601 3.533409 0.0601

UK - Hungary r=0 20.72347 0.0074 14.2646 0.0067

r1 1.201198 0.2731* 3.841466 0.2731*

UK - Ireland r=0 16.78542 0.0318 13.61402 0.0632*

r1 3.171399 0.0749* 3.171399 0.0749

UK - Italy r=0 20.4734 0.0082 13.53916 0.0649*

r1 6.934233 0.0085** 6.934233 0.0085

52 | P a g e
678430

UK - Japan r=0 25.85748 0.001 20.90714 0.0039

r1 4.950336 0.0261** 4.950336 0.0261**

UK - Latvia r=0 14.59756 0.0679* 11.86033 0.116*

r1 2.737228 0.098 2.737228 0.098

UK - Luxembourg r=0 24.22977 0.0019 22.74395 0.0018

r1 1.485816 0.2229* 1.485816 0.2229*

UK - Malta r=0 15.87269 0.0439 12.3448 0.0984*

r1 3.527894 0.0603* 3.527894 0.0603

UK - Netherlands r=0 18.43192 0.0175 12.14015 0.1055*

r1 6.291769 0.0121** 6.291769 0.0121

UK - Poland r=0 10.62566 0.2356* 6.906249 0.5003*

r1 3.719414 0.0538 3.719414 0.0538

Uk - Portugal r=0 22.12594 0.0043 13.54827 0.0646*

r1 8.577674 0.0034** 8.577674 0.0034

UK - Romania r=0 16.83671 0.0312 11.6321 0.1252*

r1 5.294616 0.0225** 5.204616 0.0225

UK - Slovakia r=0 19.51415 0.0117 15.94035 0.0269

r1 3.573794 0.0587* 3.573794 0.0587*

UK - Slovenia r=0 12.4096 0.1383* 10.6379 0.1733*

r1 1.771616 0.1832 1.771616 0.1832

UK - Spain r=0 19.05161 0.0139 13.63307 0.0627*

r1 5.418541 0.0199** 5.418541 0.0199

UK - Sweden r=0 20.0423 0.0096 16.39107 0.0227

r1 3.65123 0.056* 3.65123 0.0560*

UK - US r=0 19.8127 0.0105 18.91645 0.0085

r1 0.896251 0.3438* 0.896251 0.3438*
* denote the number of cointergration equations
** denotes 2 cointergrating equations established for this data set

53 | P a g e
678430

Table 15 : Johansen-cointergration test results for UK to markets (Post Period)

Panel Null Hypothesis Trace Test Prob Max. Eigenvalue 5% critical value
UK - Austria r=0 17.93589 0.021 17.13192 0.0171
r1 0.803977 0.3699* 0.803977 0.3699*

UK - Belgium r=0 23.4082 0.0026 15.41297 0.0328

r1 7.995408 0.0047** 7.995408 0.0047**

UK - Bulgaria r=0 18.98533 0.0143 18.03645 0.0121

r1 0.94888 0.33* 0.94888 0.33*

UK - Croatia r=0 18.58613 0.0165 16.58299 0.0211

r1 2.00314 0.157* 2.00314 0.157*

UK - Cyprus r=0 19.12577 0.0135 15.067 0.0373

r1 4.05877 0.0439** 4.05877 0.0439**

UK - Czech r=0 19.35104 0.0125 13.92024 0.0566

r1 5.430797 0.0198** 5.430797 0.0198**

UK - Denmark r=0 13.42845 0.1* 11.22183 0.1434*

r1 2.206626 0.1374 2.206626 0.1374

UK - Estonia r=0 15.36515 0.0523* 15.29318 0.0343

r1 0.071976 0.7885 0.071976 0.7885*

UK - Finland r=0 28.63296 0.0003 24.34998 0.0009

r1 4.282961 0.0385** 4.282961 0.0385**

UK - France r=0 17.97768 0.0207 16.75147 0.0198

r1 1.226209 0.2681* 1.226209 0.2681*

UK - Germany r=0 20.05989 0.0095 19.01722 0.0082

r1 1.042672 0.3072* 1.042672 0.3072*

UK - Greece r=0 21.22156 0.0061 20.60861 0.0044

r1 0.612946 0.4337* 0.612946 0.4337*

UK - Hungary r=0 17.27311 0.0267 17.26587 0.0163

r1 0.007238 0.9318* 0.007238 0.9318*

UK - Ireland r=0 18.78828 0.0154 15.28057 0.0344

r1 3.507709 0.0611* 3.507709 0.0611*

UK - Italy r=0 23.24805 0.0028 22.72258 0.0018

r1 0.525473 0.4685* 0.525473 0.4685*

54 | P a g e
678430

UK - Japan r=0 15.93389 0.0429 15.5759 0.0309

r1 0.362306 0.5472* 0.362306 0.5472*

UK - Latvia r=0 14.21549 0.0772* 12.92046 0.0806*

r1 1.295028 0.2551 1.295028 0.2551

UK - Luxembourg r=0 15.76168 0.0456 12.46115 0.0945*

r1 3.300523 0.0693* 3.300523 0.0693

UK - Malta r=0 14.06072 0.0813* 13.84842 0.0581*

r1 0.212306 0.645 0.212306 0.645

UK - Netherlands r=0 18.57954 0.0166 15.94329 0.0269

r1 2.636243 0.1044* 2.636243 0.1044*

UK - Poland r=0 21.65955 0.0052 21.65564 0.0029

r1 0.003906 0.9489* 0.003906 0.9489*

Uk - Portugal r=0 18.20221 0.0191 15.45754 0.0322

r1 2.744673 0.0976* 2.744673 0.0976*

UK - Romania r=0 19.89134 0.0102 14.52822 0.0454

r1 5.363118 0.0206** 5.363118 0.0206**

UK - Slovakia r=0 24.67432 0.0016 15.85103 0.0278

r1 8.832295 0.003** 8.832295 0.003**

UK - Slovenia r=0 18.15028 0.0194 12.252477 0.1014*

r1 5.895512 0.0152** 5.895512 0.0152

UK - Spain r=0 21.4965 0.0055 17.09728 0.0174

r1 4.399219 0.0359** 4.399219 0.0359**

UK - Sweden r=0 21.66762 0.0052 19.50173 0.0068

r1 2.165895 0.1411* 2.165895 0.1411*

UK - US r=0 15.09303 0.0574* 13.09282 0.0759*

r1 2.00208 0.1573 2.00208 0.1573
* denote the number of cointergration equations
** denotes 2 cointergrating equations established for this data set

55 | P a g e
678430
Table 16: Johansen-cointergration test results for UK to regions (Pre Period)
Panel Null Hypothesis Trace Test Prob Max. Eigenvalue 5% critical value
UK - East Europe r=0 203.29 0.0246 53.22116 0.1493*
r1 150.0689 0.1455* 37.43577 0.6554
r2 112.6331 0.2345 34.5306 0.4901
r3 78.10246 0.4302 23.8245 0.8356
r4 54.27796 0.4496 19.04694 0.8191
r5 35.23102 0.4358 13.16244 0.8758
r6 22.068 0.2947 10.96339 0.651
r7 11.10519 0.2051 6.24578 0.5819
r8 4.85941 0.0275 4.85941 0.0275

UK - West Europe r=0 208.0323 0.0132 58.98247 0.0441

r1 149.0499 0.1610* 36.46716 0.7138*
r2 112.5827 0.2356 30.28382 0.7641
r3 82.2989 0.2919 24.92422 0.7722
r4 57.37468 0.3247 16.79061 0.9229
r5 40.40407 0.2084 14.32764 0.7979
r6 26.05643 0.127 12.66557 0.4835
r7 13.39085 0.1012 9.762898 0.228
r8 3.6279 0.0568 3.6279 0.0568

UK - North Europe r=0 55.8407 0.0074 28.53893 0.0376

r1 27.30178 0.0945* 16.33061 0.2061*
r2 10.97117 0.2133 8.479951 0.332
r3 2.49122 0.1145 2.49122 0.1145

UK - Central Europe r = 0 228.935 0.1319* 51.59723 0.4687*

r1 177.3377 0.311 43.20642 0.6347
r2 134.1313 0.5034 34.49172 0.8201
r3 99.6396 0.6131 27.04064 0.9149
r4 72.59896 0.6318 25.6877 0.7233
r5 46.91126 0.7642 18.03191 0.876
r6 28.879 0.7741 12.66074 0.9028
r7 16.21861 0.6968 8.954376 0.8361
r8 7.264232 0.5471 4.198061 0.8379
r9 3.066171 0.0799 3.066171 0.0799

UK - South Europe r=0 57.89015 0.3058* 22.78303 0.5471*

r1 35.10712 0.4424 14.53821 0.7837
r2 20.56891 0.3851 10.8824 0.659
r3 9.686512 0.3058 7.430201 0.4396
r4 2.25631 0.1331 2.25631 0.1331

UK - G10 r=0 220.4513 0.021 52.69859 0.1647*

r1 167.7527 0.0164 44.62304 0.248
r2 123.1297 0.0702* 35.85671 0.4066
r3 87.27296 0.1661 29.39357 0.4644
r4 57.87939 0.3062 18.24622 0.8649
r5 39.63317 0.2359 17.12223 0.5695
r6 22.51093 0.2709 12.05811 0.05419
r7 10.45282 0.2475 8.038597 0.3748
r8 2.414224 0.1202 2.414224 0.1202

56 | P a g e
678430

UK - G7 r=0 115.8491 0.001 40.07757 0.0443

r1 75.30447 0.0171 30.14245 0.1310*
r2 45.16202 0.0876* 21.05284 0.273
r3 24.10918 0.1959 12.08901 0.3578
r4 10.02017 0.2792 9.122546 0.2761
r5 0.897626 0.3434 0.897626 0.3434

UK - AP r=0 50.09036 0.001 30.38898 0.019

r1 19.70138 0.0109 19.31292 0.0073
r 2** 0.38846 0.5331* 0.38846 0.5331*
* denote the number of cointergration equations

Table 17: Johansen-cointergration test results for UK to regions (Pre Period)

Panel Null Hypothesis Trace Test Prob Max. Eigenvalue 5% critical value
UK - East Europe r=0 205.6686 0.0181 51.20393 0.2154*
r1 154.4647 0.0912* 39.74728 0.511
r2 114.7174 0.1902 33.6324 0.5493
r3 81.08498 0.3294 26.79499 0.6474
r4 52.29 0.4491 21.11828 0.6753
r5 33.17172 0.5473 17.71234 0.5192
r6 15.45937 0.7496 9.477367 0.7922
r7 5.982005 0.6977 4.808734 0.7658
r8 1.173271 0.2787 1.173271 0.2787

UK - West Europe r=0 207.5141 0.0142 68.035 0.0044

r1 139.4791 0.3596* 45.3077 0.2197*
r2 94.17142 0.7724 34.75961 0.4753
r3 59.41181 0.9569 17.99426 0.9923
r4 41.41755 0.9323 15.51481 0.9659
r5 25.90274 0.8924 11.56919 0.9486
r6 14.33356 0.8214 8.392007 0.8782
r7 5.94155 0.7024 5.797462 0.6394
r8 0.144089 0.7042 0.144089 0.7042

UK - North Europe r=0 59.36637 0.0029 35.28768 0.0042

r1 24.07869 0.1971* 15.55894 0.2517*
r2 8.519753 0.4116 6.817505 0.511
r3 1.702248 0.192 1.702248 0.192

UK - Central Europe r = 0 170.3136 0.4739* 40.96403 0.7605*

r1 129.3495 0.6362 32.89457 0.881
r2 96.45495 0.7096 28.45691 0.8591
r3 67.99805 0.7856 26.48579 0.669
r4 41.51225 0.9213 15.55825 0.965
r5 25.95401 0.8908 11.02435 0.9648
r6 14.92965 0.7846 10.1304 0.7323
r7 4.799255 0.8298 4.682817 0.7812
r8 0.116438 0.7329 0.116438 0.7329

57 | P a g e
678430

UK - South Europe r=0 70.87987 0.0411 30.4777 0.1207*

r1 40.40217 0.2084* 19.53574 0.3742
r2 20.86643 0.366 13.79549 0.3821
r3 7.070941 0.5694 6.996635 0.4896
r4 0.074306 0.7852 0.074306 0.7852

UK - G10 r=0 249.7687 0 69.48418 0.003

r1 180.2845 0.0022 64.1971 0.0021
r2 116.0874 0.1643* 39.2233 0.2316*
r3 76.87507 0.4743 28.96915 0.4935
r4 47.90592 0.7256 21.92091 0.6138
r5 25.98501 0.8898 9.959542 0.9855
r6 16.02547 0.7105 8.319767 0.8831
r7 7.7057 0.04974 5.941918 0,6207
r8 1.763782 0.1842 1.763782 0.1842

UK - G7 r=0 111.1028 0.0029 59.59891 0.001

r1 51.50393 0.5714* 28.8172 0.1783*
r2 22.86873 0.9658 9.587642 0.9899
r3 13.09909 0.887 7.185247 0.9463
r4 5.913843 0.7057 4.617106 0.7892
r5 1.296737 0.2548 1.296737 0.2548

UK - AP r=0 26.95921 0.1026* 20.91042 0.0536*

r1 6.048796 0.6898 14.2646 0.6866
r2 0.616839 0.4322 0.616839 0.4322
* denote the number of cointergration equations

*Markets in each sample can be found in Appendix 17

4.7. VAR & VECM

We conduct VAR on those samples with no Cointergration equations using stationery data
sets and VECM on regions with Cointergration using non-stationery closing prices using a lag
length of 1 (denoted as -1). We can therefore use the results of VAR and VECM to highlight
the dynamic relationship at play in Europe. Table 18 highlights our findings for both pre and
Post period (T stats and P Values). Some of the findings are very significant to UK investors
looking to invest in regional portfolios. For example, VECM clearly identifies Markets that
are highly correlated though 2 or less cointergration equations but additional highlighted
cells signal causal linkages upon the dependant. The findings of VECM and VAR significantly
support our finding in PCA and Grangers causality test, identifying strong linkages in regions
particularly of that of Western Europe and Northern Europe. Additionally, highlighting the
causal linkages established in Grangers causality test between Greece and the UK. Investors

58 | P a g e
678430

can use these findings to identify the regions to and not to invest in, in particular Western,
Southern and Northern Europe are all significantly Interdependent with the UK in addition
to the economic groups of the G7&G10. We would expect to see these results as they are
very much in line with studies by Wasim (2013) & Bracker (1999) that identified strong
correlation between highly developed markets. However, what is more significant to see is
the significant changes between pre and post in southern and Western Europe, there is a
surge increase in short term causal links and as with past studies this has been used to
highlight contagion. Therefore, UK investors should avoid these markets if they are looking
to benefit from international diversification. Additional the research identifies the markets
that influence the dependant variables and its no surprise that the US and Italy exert
significant influences in their retrospective regions .Lastly my findings clearly identify that
Central and Eastern Europe of the best diversification benefits to investors and in line with
past studies this is as a result of global exposure of many of these markets whose market
cap is small in comparison to Western and Southern European markets coupled with
geographical location and many of the factors discussed in the literature review.

59 | P a g e
 VAR - Central Europe - Pre Period
Austria Croatia Czech Germany Hungary Poland Romania Slovakia Slovenia UK
Austria (-1) -0.72957 0.24668 0.06296 -0.05089 0.34305 -0.28216 0.28833 1.41616 -0.73309 -0.28865
0.46580 0.80520 0.94980 0.95940 0.73160 0.77790 0.77320 0.15700 0.45370 0.77980
Croatia (-1) 0.10658 -0.68610 0.07850 -0.22560 -0.04797 2.02130 0.87197 -0.43353 0.00751 -0.08692
0.91510 0.49280 0.93740 0.82160 0.96170 0.04350 0.38340 0.66470 0.99400 0.93080
Czech (-1) -0.02399 -0.15783 -1.19829 0.46040 0.98143 -0.22908 -0.90610 -0.91413 0.39905 -0.20633
0.98090 0.87460 0.23110 0.64530 0.32660 0.81890 0.36510 0.36090 0.68990 0.83660
Germany (1) 0.08436 3.33978 0.07583 -0.13945 -0.27908 0.51766 1.22260 0.21916 -0.62919 0.19292
0.93280 0.00090 0.93960 0.88910 0.78020 0.60480 0.22170 0.82660 0.52940 0.84710
Hungary (-1) 0.67215 -0.69152 0.01561 -0.50055 -0.63461 0.23978 -0.55654 0.31568 1.36781 1.17484
0.50160 0.48490 0.98750 0.61680 0.52580 0.81050 0.57800 0.75230 0.17170 0.24030
Tables 18: VAR and VECM Results (pre and Post)

Poland (-1) 0.00063 0.10511 -0.38199 -0.12720 -0.17556 0.38910 -0.37606 0.60708 1.61632 -0.04983
0.99500 0.91630 0.70250 0.89880 0.86070 0.69730 0.70690 0.54390 0.10630 0.96030
Romania (-1) -0.79710 0.41201 0.24627 -0.99945 0.17273 -0.85606 -0.58301 -0.11504 0.51702 -1.42539
0.42560 0.68040 0.80550 0.31780 0.86290 0.39220 0.56000 0.90840 0.60520 0.15430
Slovakia (-1) -0.14143 2.19148 0.55030 -0.44810 0.72366 0.41685 0.53446 -1.77757 -0.97133 -0.88295
0.88760 0.02860 0.58220 0.65420 0.46980 0.67690 0.59310 0.07580 0.33160 0.37750
Slovenia (-1) 0.55600 -0.17487 -0.32119 0.14426 -1.50779 -0.73030 -1.91055 -1.07159 -0.61338 0.37346
0.57830 0.86120 0.74810 0.88530 0.13190 0.46540 0.05630 0.28410 0.53980 0.70890
UK (-1) 0.92200 -1.33610 1.45087 0.90580 0.70666 1.20862 0.17973 -0.28969 0.68635 0.12661
0.35670 0.18180 0.14710 0.36520 0.47990 0.22710 0.85740 0.77210 0.49260 0.89930

60 | P a g e
678430

C -0.63222 0.21920 -0.96759 -0.31326 1.01133 -0.24658 -0.62296 0.47663 -0.20095 -0.20672
0.53330 0.82650 0.33350 0.75410 0.31210 0.80530 0.53340 0.63370 0.84080 0.83630
 VAR - Central Europe - Post Period
Austria Croatia Czech Germany Hungary Poland Romania Slovakia Slovenia UK
Austria (-1) -1.40493 -0.98318 0.12121 0.920537 -1.65297 -0.43285 -0.0565 1.119885 0.337247 0.200631
0.1603 0.3257 0.9035 0.3575 0.0986 0.6652 0.955 0.263 0.736 0.841
Croatia (-1) -0.32524 2.397601 -0.5022 0.010253 -0.45571 -0.80892 -1.36993 -1.08176 -0.61074 0.573102
0.7451 0.0166 0.6156 0.9918 0.6487 0.487 0.1709 0.2796 0.5415 0.5667
Czech (-1) -0.27096 1.822342 -0.10061 -1.841307 -0.54557 -0.30237 -1.18277 0.458129 1.010002 -2.14771
0.7865 0.0686 0.9199 0.0658 0.5855 0.7624 0.2371 0.6469 0.3127 0.0158
Germany (1) 0.858803 -0.67716 2.835822 -0.74004 0.910996 1.277995 0.343614 0.352238 -1.59813 0.023468
0.3906 0.4984 0.0046 0.4594 0.3625 0.2015 0.7312 0.7247 0.1103 0.9813
Hungary (-1) 1.532539 0.033087 0.025577 1.741588 0.488509 -0.23196 1.38359 -0.5889 0.701824 1.485934
0.1256 0.9736 0.9796 0.0818 0.6253 0.8166 0.1667 0.556 0.4829 0.1375
Poland (-1) 0.828939 1.99394 -0.96021 0.412996 -0.47075 -0.44556 -0.96517 -1.22681 -0.42978 -0.05175
0.4073 0.0464 0.3371 0.6797 0.6379 0.656 0.3346 0.2201 0.6674 0.9587
Romania (-1) 2.168276 -1.19366 1.399841 2.327851 2.14764 1.66487 1.442695 1.157436 1.565029 2.18015
0.0303 0.2328 0.1618 0.0201 0.0319 0.0962 0.1493 0.2473 0.1178 0.0294
Slovakia (-1) -2.74974 -1.15238 -0.2913 -1.441131 -1.2151 -0.54468 -0.97576 -3.62583 -0.02826 -1.40367
0.006 0.2494 0.7709 0.1498 0.2245 0.5861 0.3294 0.0003 0.9775 0.1607
Slovenia (-1) -0.65845 -1.57169 0.567921 -0.387551 0.316451 0.94459 -1.70726 -0.54342 -1.52285 0.440674
0.5104 0.1163 0.5702 0.6984 0.7517 0.345 0.088 0.5869 0.128 0.6595
UK (-1) -0.05096 -0.32709 -1.35728 -0.104545 -0.47359 -0.02851 -1.17223 -0.3264 1.148393 0.600512
0.9594 0.7437 0.1749 0.9168 0.6359 0.9773 0.2413 0.7442 0.251 0.5483

61 | P a g e
678430

C 0.78026 3.11565 0.60805 0.3454 1.93298 1.23777 1.819396 0.50397 0.66326 0.5867
0.4354 0.0019 0.5433 0.7298 0.0535 0.216 0.0691 0.6144 0.5073 0.5575
678430

VAR - Southern Europe - Pre Period

Cyprus Greece Italy Malta UK
Cyprus (-1) -0.21684 -0.96798 -0.29452 -0.72345 -0.07864
0.82840 0.33350 0.76850 0.46970 0.93730

Greece (-1) 1.31853 0.37163 1.04934 -0.99552 1.15023

0.18790 0.71030 0.29450 0.31990 0.25050

Italy (-1) -0.45929 -0.05714 -1.03108 0.57514 -0.62410

0.64620 0.95450 0.30290 0.56540 0.53280

Malta (-1) -0.69947 -0.44380 -1.61475 -0.03708 -1.48828

0.48450 0.65740 0.10690 0.97040 0.13720

UK (-1) 0.63386 1.84178 0.24271 -0.21514 0.06691

0.52640 0.06600 0.80830 0.82970 0.94670

C 0.16766 0.05953 -1.02922 0.35412 -0.02530

0.86690 0.95260 0.30830 0.72340 0.97980

VECM - Southern Europe - Post Period

D(Cyprus) D(Greece) D(Italy) D(Malta) D(UK)
Coint EQ(1) -5.37849 -14.40171 -7.41986 -2.33024 -4.07722
0.00000 0.00000 0.00000 0.02010 0.00010

Cyprus (-1) -7.54914 0.38925 -0.53245 -0.19279 -0.85984

0.00000 0.69720 0.59460 0.84720 0.39020

Greece (-1) 2.58534 0.97357 5.98543 0.54132 3.71151

0.00990 0.33060 0.00000 0.58850 0.00020

Italy (-1) 3.02032 3.99875 -4.71617 1.83039 2.88874

0.00260 0.00010 0.00000 0.06760 0.00400

Malta (-1) 0.71811 1.69438 0.43765 -6.40273 0.62123

0.47290 0.09070 0.66180 0.00000 0.53470

UK (-1) -1.68210 -1.36352 -0.54776 -1.57428 -5.73094

0.09300 1.13959 0.58400 0.11590 0.00000

C 0.60975 1.13959 0.73289 0.14541 0.36486

0.54220 0.25490 0.46390 0.88440 0.71530

62 | P a g e
 VECM- Eastern Europe - Pre Period
D(Bulgaria) D(croatia) D(Czech) D(estonia) D(Hungary) D(Latvia) D(Romania) D(Slovakia) D(Slovenia) D(UK)
Coin EQ(1) -0.73079 1.98218 -0.55502 2.48891 1.26900 0.93523 4.62308 0.39723 2.03753 -0.67515
0.46510 0.04770 0.57900 0.01300 0.20470 0.34990 0.00000 0.69130 0.04180 0.49970
Bulgaria (-1) -1.74643 -1.94788 -2.09140 -2.57833 -3.63039 -1.68820 -2.55045 -1.62941 -1.92745 -1.56575
0.08100 0.05170 0.03670 0.01010 0.00030 0.09170 0.01090 0.10350 0.05420 0.11770
Croatia (-1) -0.54788 -0.73828 0.23372 -0.85820 -0.00055 -0.12082 0.61139 -0.49478 0.04059 0.09171
0.58400 0.46050 0.81520 0.39100 0.99960 0.90390 0.54110 0.62090 0.96760 0.92690
Czech (-1) 0.15729 0.52467 -0.92881 -0.38378 1.81545 -1.34249 -0.58511 -0.06169 0.54140 0.07416
0.87500 0.59990 0.35320 0.70120 0.06970 0.17970 0.55860 0.95080 0.58830 0.94090
Estonia (-1) -0.11125 0.40751 -1.13287 -0.97045 -0.01125 -0.17747 -1.31412 0.92848 0.20029 -0.54546
0.91140 0.68370 0.25750 0.33200 0.99100 0.85920 0.18910 0.35340 0.84130 0.58560
Hungary (-1_ -0.02737 1.03476 0.16829 0.49193 -0.32908 1.62431 0.92267 0.59579 1.86667 1.70372
0.97820 0.30100 0.86640 0.62290 0.74220 0.10460 0.35640 0.55140 0.06220 0.08870
Latvia (-1) -0.06074 -0.70785 0.00846 0.31729 0.04807 -1.62995 -0.44800 -0.57546 -0.63890 -1.47725
0.95160 0.47920 0.99330 0.75110 0.96170 0.10540 0.96430 0.56510 0.52300 0.13990
Romania (-1) 0.98577 0.33440 0.51654 2.06770 0.37557 0.51870 -0.54276 0.10210 0.78236 -1.34996
0.32450 0.73810 0.60560 0.03890 0.70730 0.60410 0.58740 0.91870 0.43420 0.17730
Slovakia (-1) -0.58999 0.44665 0.86887 -0.65558 0.16138 0.19339 -0.93258 -2.47661 -1.38962 -0.69935
0.55530 0.65520 0.38510 0.51220 0.87180 0.84670 0.35120 0.01340 0.16490 0.48450
Slovenia (-1) 0.31943 -0.69703 -0.01922 -1.12650 -1.96325 0.09129 -2.97128 -1.09511 -0.83536 0.65751
0.74950 0.48590 0.98470 0.26020 0.04990 0.92730 0.00300 0.27370 0.40370 0.51100
UK (-1) 0.75737 1.49284 2.37201 1.59941 1.15916 1.18927 1.25294 1.41488 0.08074 0.22096

63 | P a g e
0.44900 0.13580 0.01790 0.11000 2467.00000 0.23460 0.21050 0.15740 0.93570 0.82520
678430

C 0.12860 -0.14280 -0.86947 1.34223 1.12532 0.40743 -0.80037 0.35534 -0.09546 -0.04149
0.89770 0.88650 0.38480 0.17980 0.26070 0.68380 0.42370 0.72240 0.92400 0.96690
 VECM - Eastern Europe - Post Period
D(Bulgaria) D(croatia) D(Czech) D(estonia) D(Hungary) D(Latvia) D(Romania) D(Slovakia) D(Slovenia) D(UK)
Coin EQ(1) -0.92544 -2.40956 -1.49729 -1.05722 -5.74917 -1.57350 -8.16199 3.35568 -0.87506 -0.52310
0.35490 0.01610 0.13460 0.29060 0.00000 0.11590 0.00000 0.00080 0.38170 0.60100
Bulgaria (-1) -3.85262 -0.77460 -0.14681 0.77330 0.43849 1.52013 0.11531 0.58214 1.21481 0.72331
0.00010 0.43870 0.88330 0.43950 0.66100 0.12870 0.90820 0.56060 0.22470 0.46960
Croatia (-1) -0.02178 -3.77438 -0.58211 -0.32204 1.32431 -0.30980 0.83234 -0.42247 -0.39735 0.03822
0.98260 0.00020 0.56060 0.74750 0.18560 0.75680 0.40540 0.67280 0.69120 0.96950
Czech (-1) -0.60786 2.40629 -5.08430 -1.04975 0.12035 -0.97443 0.51512 -0.07834 0.05149 -1.48932
0.54340 0.01630 0.00000 0.29400 0.90420 0.33000 0.60660 0.93760 0.95890 0.13670
Estonia (-1) -0.32840 -0.78444 -0.78433 -5.94371 -1.92992 0.06702 1.08086 0.32230 -0.21346 -0.84971
0.74270 0.43290 0.43300 0.00000 0.05380 0.94660 0.28000 0.74730 0.83100 0.39560
Hungary (-1_ 0.41189 1.80048 0.53639 5.17604 -3.98558 0.87352 3.49687 -2.28363 0.59576 1.62142
0.68050 0.07200 0.59180 0.00000 0.00010 0.38250 0.00050 0.02260 0.55140 0.10520
Latvia (-1) -0.46947 0.66866 0.85898 0.30895 0.95340 -6.28667 1.31385 1.21956 -0.88825 0.08509
0.63880 0.50390 0.39050 0.75740 0.34060 0.00000 0.18910 0.22290 0.37460 0.93220
Romania (-1) 0.03437 0.44253 2.17619 1.10779 5.21468 0.32047 0.72157 -1.25665 2.06425 1.71472
0.97260 0.65820 0.02970 0.26820 0.00000 0.74870 0.47070 0.20910 0.03920 0.08660
Slovakia (-1) -1.24668 -1.74383 -0.06687 -1.82442 -3.20943 -0.32396 -5.17001 -5.88250 -0.08218 -0.75141
0.21270 0.08140 0.94670 0.06830 0.00140 0.74600 0.00000 0.00000 0.93450 0.45250
Slovenia (-1) -1.31442 -2.53182 0.88754 -1.12700 0.32533 -0.12680 -2.75065 -0.02795 -8.54241 0.75425
0.18890 0.01150 0.37500 0.26000 0.74940 0.89910 0.00600 0.97770 0.00000 0.45080
UK (-1) 0.97485 -2.10735 1.37329 0.61060 -0.74564 2.85779 -3.39616 0.85004 -0.10881 -3.50709

64 | P a g e
0.32980 0.03530 0.16990 0.54160 0.45600 0.00430 0.00070 0.39550 0.91340 0.00050
678430

C 0.14504 0.51125 0.33714 -0.00687 0.31945 -0.00795 0.41359 -0.06805 0.30614 0.22090
0.88470 0.60930 0.73610 0.99450 0.74940 0.99370 0.67920 0.94580 0.75950 0.82520
 VECM - Western Europe - Pre Period
D(Belgium) D(France) D(Germany) D(Ireland) D(Luxembourg) D(netherlands) D(Portugal D(Spain) D(UK)
Coin E(1) -0.15353 -0.66020 1.31454 -0.45798 0.27952 -2.12083 -0.48813 0.00600 0.47599
0.87800 0.50930 0.18900 0.64710 0.77990 0.03420 0.62560 0.99520 0.63420
Begium (-1) 1.00242 0.83909 0.94916 0.79928 1.41430 -0.45761 0.55513 0.08685 1.04292
0.31640 0.40160 0.34280 0.42430 0.15760 0.64730 0.57890 0.93080 0.29720
France (-1) -1.42167 -1.84800 -0.97593 -1.47711 -1.87813 -1.03248 -1.76611 -1.15019 -1.49454
0.15540 0.06490 0.32930 0.14000 0.06070 0.30210 0.07770 0.25030 0.13540
Germany (-1) 0.22084 0.64978 -1.16746 0.45881 0.06270 0.47036 0.75567 0.75407 0.37198
0.82530 0.51600 0.24220 0.64650 0.95000 0.63820 0.45000 0.45100 0.71000
Netherlands (-1) 1.25561 1.25628 1.69622 1.31889 1.26051 2.17061 1.82615 1.46586 1.67117
0.20960 0.20930 0.09020 0.18750 0.20780 0.03020 0.06810 0.14300 0.09500
Ireland (-2) 0.53083 0.16500 0.77428 0.61100 0.89035 -0.33911 -0.14344 -0.01788 0.30059
0.59570 0.86900 0.43900 0.54130 0.37350 0.73460 0.88600 0.98570 0.76380
Luxembourg (-1) 0.08373 0.53540 0.37990 0.20683 -0.52550 -1.09145 0.27045 0.42808 0.41479
0.93330 0.59250 0.70410 0.83620 0.59940 0.27530 0.78690 0.66870 0.67840
Portugal (-1) -0.35303 -0.33691 -1.06350 0.02764 -0.39125 1.05887 1.10589 -0.20190 -0.48031
0.72410 0.73630 0.28780 0.97800 0.69570 0.28990 0.26900 0.84000 0.63110
Spain (-1) -0.42006 0.28612 0.05023 -0.49761 -0.75621 -0.27464 -0.40945 -0.18668 -0.88500
0.67450 0.77490 0.95990 0.61890 0.44970 0.78370 0.68230 0.85200 0.37640
UK (-1) -0.59479 -0.68289 -0.76714 -0.74024 0.14854 -1.29318 -1.16863 -0.67611 -0.92634

65 | P a g e
0.55210 0.49480 0.44320 0.45930 0.88190 0.19630 0.24280 0.49910 0.35450
678430

C -0.54931 -0.41165 -0.45804 -0.05088 -0.57208 0.13888 -0.39445 -0.64008 -0.10110

0.58290 0.68070 0.64700 0.95940 0.56740 0.89340 0.69330 0.52230 0.91950
 VECM - Western Europe - Post Period
D(Belgium) D(France) D(Germany) D(Ireland) D(Luxembourg) D(netherlands) D(UK) D(Spain) D(Portugal
Coin E(1) 7.15982 9.29387 5.85601 6.79256 4.73592 6.41250 7.26981 11.74262 7.53892
0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
Begium (-1) -4.14631 -1.44713 -1.24654 -1.42133 0.19713 -0.54254 0.51293 -2.46070 -1.49712
0.00000 0.14810 0.21280 0.15550 0.84380 0.58750 0.60810 0.01400 0.13460
France (-1) 2.35514 0.85024 2.01396 2.02937 2.11546 2.19009 1.46456 2.00339 2.21013
0.01870 0.39540 0.04420 0.04260 0.03460 0.02870 0.14330 0.04540 0.02730
Germany (-1) -2.86515 -3.27760 -4.30912 -2.14388 -1.11236 -2.70559 -2.63433 -2.77376 -2.51694
0.00430 0.01100 0.04420 0.01590 0.26620 0.00690 0.00850 0.00560 0.01200
Ireland (-2) 3.37145 3.63758 2.39847 -0.43726 1.51863 3.17485 2.84327 4.62059 3.18778
0.00080 0.00030 0.01660 0.66200 0.12910 0.00150 0.00450 0.00000 0.00150
Luxembourg (-1) 1.69408 2.38031 1.09920 2.48426 -3.36814 1.93072 2.19003 3.05020 2.22763
0.09050 0.01750 0.27190 0.01310 0.00080 0.05380 0.02870 0.00230 0.02610
Netherlands (-1) -0.88601 -1.97224 -1.75640 -2.65900 -2.73343 -3.65560 -2.08945 -1.45167 -0.82607
-0.42936 0.04880 0.07930 0.00790 0.00640 0.00030 0.03690 0.14690 0.40890
UK (-1) -0.42936 0.96931 0.78010 1.87907 0.84374 0.23732 -3.50085 0.99557 -0.18606
0.66770 0.33260 0.43550 0.06050 0.39900 0.81240 0.00050 0.31970 0.85240
Spain (-1) 2.11453 3.05162 2.13874 4.37306 1.57890 1.51195 2.53291 -0.91935 1.68606
0.03470 0.00230 0.03270 0.00000 0.11460 0.13080 0.01140 0.35810 0.09200
Portugal (-1) -0.02469 -0.95300 -0.22628 -1.24657 -0.72240 -0.07072 0.90375 -0.02419 -3.92595
0.98030 0.34080 0.82100 0.21290 0.47020 0.94360 0.36630 0.98070 0.00010

66 | P a g e
678430

C 0.55071 0.72640 0.59011 0.58647 0.45739 0.52334 0.39108 0.96812 0.50911

0.58190 0.46770 0.55520 0.55770 0.64750 0.60080 0.69580 0.33230 0.61080
678430

VECM - Northern Europe - Pre Period

D(Denmark) D(Finland) D( Sweden) D(UK)
Coin EQ(1) -1.85425 -3.86144 -1.80469 -2.04952
0.06440 0.00010 0.07180 0.01640

Denmark (-1) -1.91185 -1.38930 -1.14734 -1.48784

0.05650 0.16540 0.25180 0.13570

Sweden (-1) 0.71381 -0.87568 -0.74462 -0.25776

0.47570 0.38170 0.45690 0.79670

Finland (-1) -0.09947 0.99410 0.68763 0.62495

0.92080 0.32070 0.49200 0.53230

UK (-1) 1.33892 1.06047 1.22985 0.65237

0.18130 0.28950 0.21940 0.51450

C -0.48216 -0.42768 -0.59337 -0.01456

0.62990 0.66910 0.55320 0.98840

VECM - Northern Europe - Post Period

D(Denmark) D(Finland) D( Sweden) D(UK)
Coin EQ(1) -5.95000 -8.18202 -3.72823 -5.68506
0.00000 0.00000 0.00020 0.00000

Sweden (-1) -2.00602 1.98692 -0.57488 0.99751

0.04540 0.04740 0.56560 0.31900

UK (-1) 6.82498 -0.15533 6.60256 2.74916

0.00000 0.87660 0.00000 0.00620

Finland (-1) -0.67848 -1.79519 -2.67160 -0.23718

0.49780 0.07320 0.00780 0.81260

Denmark (-1) 1.45960 0.31944 0.54732 -6.53174

0.14500 0.74950 0.58440 0.00000

C 0.04563 0.36033 0.00440 0.25697

0.96360 0.71870 0.99650 0.79730

67 | P a g e
 [ 2.64940] [ 2.70148] [2
C -7.906782 -3.484337 -2
(12.9384) (5.79790) (3
[-0.61111] [-0.60097] [-0
VECM - G10 - Pre Period
D(Germany) D(France) D(Belgium) D(Italy) D(Japan) D(Netherlands) D(Sweden) D(UK) D(US)
Coint EQ(1) -1.92361 -0.06809 0.31095 -0.57599 1.36231 -0.00682 -1.14554 0.13081 1.40645
0.05470 0.94570 0.75590 0.56480 0.17340 0.99460 0.25230 0.89600 0.15990
Coint EQ (2) 1.81574 0.03351 -0.35927 0.56597 -1.69285 -0.05643 1.05594 -0.20590 -1.66868
0.06970 0.97330 0.71950 0.57150 0.09080 0.95500 0.29130 0.83690 0.09550
Germany (-1) -0.70762 0.76749 0.31956 1.26879 -0.50237 0.48408 0.92558 0.43090 -0.60223
0.47940 0.44300 0.74940 0.20480 0.61550 0.62840 0.35490 0.66080 0.54720
France (-1) -0.42383 -1.09020 -0.64586 -0.75166 1.78491 -0.75445 -1.08682 -0.82873 -0.75529
0.67180 0.27590 0.51850 0.45240 0.07460 0.45080 0.27740 0.40750 0.45030
Belgium (-1) 1.31268 1.13069 1.42650 1.18510 0.32293 1.34939 1.48181 1.29635 1.10271
0.18960 0.25850 0.15400 0.23630 0.74680 0.17750 0.13780 0.19520 0.27040
Italy (-1) -1.06109 -0.99077 -0.57285 -1.67379 1.24682 -1.07492 -0.86712 -0.93683 -0.90559
0.28890 0.32200 0.56690 0.09450 0.21280 0.28270 0.38610 0.34910 0.36540
Japan (-1) -0.09091 -0.10850 -0.26373 -1.14496 -2.00465 -0.41479 -0.75849 -0.33692 1.99554
0.92760 0.91360 0.79200 0.25250 0.04530 0.67840 0.44830 0.71380 0.04630
Netherlands (-1) 0.62840 0.75331 0.24777 0.89354 -1.09538 0.57867 0.39492 0.54059 0.64915
0.52990 0.45140 0.80440 0.37180 0.27360 0.56290 0.69300 0.58890 0.51640
Sweden (-1) 0.22392 -0.67920 -1.32770 -0.69937 -1.25388 -0.76469 -1.18055 -0.97456 -0.44027
0.82290 0.49720 0.18460 0.48450 0.21020 0.44460 0.23810 0.33000 0.65980
UK (-1) -1.21354 -0.71945 -0.45353 -1.32603 -0.35652 2.13501 -0.31583 -0.76208 -0.78745
0.22520 0.47200 0.65030 0.18510 0.72150 0.03300 0.75220 0.44620 0.43120
US (-1) 2.64940 2.70148 2.88790 1.99773 3.27849 -0.21262 2.27043 2.53621 -0.75749

68 | P a g e
0.00820 0.00700 0.00400 0.04600 0.00110 0.83170 0.02340 0.01140 0.44890
678430

C -0.61111 -0.60097 -0.67454 -1.35525 -0.93897 -2.12615 -0.59432 -0.20528 0.17124

0.54130 0.54800 0.50010 0.17570 0.34800 0.83170 0.55240 0.83740 0.86410
 [-0.12328] [-1.20314]
C 0.000476 0.000611
(0.00085) (0.00084)
[ 0.56123] [ 0.72573]
VECM - G10 - Post period
D(Germany) D(France) D(Belgium) D(Italy) D(Japan) D(Sweden) D(UK) D(US) Netherlands
Coint EQ(1) -5.17140 -6.65561 -5.40396 -2.65481 1.05096 -1.31056 -5.67456 -1.18397 -5.54505
0.00000 0.00000 0.00000 0.00800 0.29350 0.19030 0.00000 0.23670 0.00000
Coint EQ (2) -5.15618 -6.69102 -5.40978 -2.91666 0.62673 -1.05630 -5.59810 -1.23214 -5.47676
0.00000 0.00000 0.00000 0.00360 0.53100 0.29110 0.00000 0.21820 0.00000
Germany (-1) -0.67353 2.92793 1.72534 1.53187 0.92021 2.19003 3.40541 0.69883 2.54578
0.50070 0.00350 0.08470 0.12580 0.35770 0.02870 0.00070 0.48480 0.01100
France (-1) 3.39677 2.83616 3.66498 3.11475 -0.90431 2.40423 3.31115 1.60823 3.75686
0.00070 0.00460 0.00030 0.00190 0.36600 0.01640 0.00100 0.10810 0.00200
Belgium (-1) -1.39221 -1.76578 -3.51011 -1.52946 1.73983 -1.25751 -1.36348 -1.44471 -1.47547
0.16410 0.07770 0.00050 0.12640 0.08220 0.20880 0.17300 0.14880 0.14040
Italy (-1) -1.48252 -1.87830 -1.41444 -2.78933 2.69166 -1.81345 -1.69737 0.71793 -2.40382
0.13780 0.06060 0.15750 0.00540 0.00720 0.70000 0.08990 0.47290 0.01640
Japan (-1) -0.00721 -0.53331 0.00466 2.41772 -3.05155 -1.29821 -1.77320 0.45229 -1.08937
0.99420 0.59390 0.99630 0.01580 0.00230 0.19450 0.07650 0.65110 0.27620
Sweden (-1) -4.24549 -6.34846 -4.37809 -5.76971 -3.33215 -6.38390 -4.74267 -2.83314 -3.75000
0.00000 0.00000 0.00000 0.00000 0.00090 0.00000 0.00000 0.00470 0.00020
UK (-1) -0.37867 1.15589 0.91674 0.93251 1.94698 0.19213 -3.36108 -0.14939 0.34021
0.96980 0.24800 0.35950 0.35130 0.05180 0.84770 0.00080 0.88130 0.73380
US (-1) -0.86993 -1.46052 -0.84118 0.26133 4.69910 -0.37885 -1.21784 -6.34661 -1.33935
0.38450 0.14440 0.40040 0.79390 0.00000 0.70490 0.22350 0.00000 0.18070
Netherlands (-1) -0.12328 -1.20314 -1.10911 -1.50955 -3.12587 -0.19942 -0.49820 0.32254 -2.61505

69 | P a g e
0.90190 0.22920 0.26760 0.13140 0.00180 0.84200 0.61840 0.74710 0.00900
678430

C 0.56123 0.72573 0.63497 0.75368 0.58190 0.03942 0.39354 0.42075 0.54191

0.57480 0.46820 0.52560 0.45120 0.56080 0.96860 0.69400 0.67400 0.58800
678430

VECM - G7 - Pre Period

D(Germany) D(France) D(Italy) D(Japan) D(UK) D(US)
Coint EQ(1) [2.307854] [3.173815 [2.629908] 0.90771 3.86231 3.02435
{0.0213} {0.0016} {0.0087} 0.36440 0.00010 0.00260

Coint EQ (2) [-6.679916] [-1.506677] [-0.521998] 2.08550 -1.93201 1.25493

{0.4968 {0.1324} {0.6018} 0.03740 0.05380 0.20990

Germany (-1) [-0.670048] [0.478872] [-0.053248] 0.76003 -0.56246 -0.63780

{0.5031} {0.6322} {0.9576} 0.44750 0.57400 0.52380

France (-1) [0.792441 [-0.979931] [1.012060] -0.44895 0.64818 -0.37528

{0.4284} {0.3275} {0.3119} 0.65360 -0.51710 0.70760

Italy (-1) [-0.899344] [-0.569180] [-1.639490] 1.48238 -1.08847 -0.65409

{0.3688} {0.5694} {0.1016} 0.13870 0.27680 0.51330

Japan (-1) [0.947293] [0.983171] [0.094603] -2.08550 1.20607 2.94479

{03438} {0.3259} 0.9247} 0.03740 0.22820 0.00330

UK (-1) [-0.204389] [-0.302752] -0.85199 0.91102 -0.12846 -0.07499

{0.8381} {0.7622} 0.39450 0.36260 0.89780 0.94020

US (-1) [2.787876] [2.884274] 2.18692 3.15399 2.25656 -0.92492

{0.0055} {0.0041} 0.02910 0.00170 0.02440 0.35530

C [-0.539393] [-0.564587] -1.34169 -1.03530 -0.13532 0.24438

{0.5898} {0.5725} 0.18020 0.30090 0.89240 0.80700

VECM - G7 - Post Period

D(Germany) D(France) D(Italy) D(Japan) D(UK) D(US)
Coint EQ (2) -2.83894 -3.06061 2.06282 5.52102 -4.19495 0.57216
0.00460 0.00230 0.03950 0.00000 0.00000 0.56740

Germany (-1) -3.28712 -1.32765 -1.41241 1.61305 -0.93690 -1.51433

0.00110 0.18470 0.15820 0.10710 0.34910 0.13030

France (-1) 2.34784 1.30678 1.01841 -3.51115 3.06534 1.35677

0.01910 0.19170 0.30880 0.00050 0.00220 0.17520

Italy (-1) -1.46053 -1.57832 -2.41774 2.26415 -1.15264 1.32546

0.14450 0.11490 0.01580 0.02380 0.24940 0.18540

Japan (-1) -0.24633 -0.99718 2.25136 -3.85160 -2.40922 0.89704

0.80550 0.31900 0.02460 0.00010 0.01620 0.37000

UK (-1) -0.65990 -1.01720 -1.32446 0.09190 -5.29321 -1.19265

0.50950 0.30940 0.18570 0.92680 0.00000 0.23340

US (-1) -0.59924 -1.10234 0.41820 4.25489 -0.98387 -6.02637

0.54920 0.27070 0.67590 0.00000 0.32550 0.00000

C 0.53893 0.56359 0.63090 0.52120 0.31602 0.39545

0.59010 0.57320 0.52830 0.60240 0.75210 0.69260

70 | P a g e
678430

VECM - Pacific - Pre Period VAR - Pacific - Post period

D(UK) D(US) D(Japan) (UK) (US) (Japan)
Coint EQ(1) -1.37788 2.01010 2.57049 UK(-1) 0.15755 1.33858 -0.06750
0.16920 0.04510 0.01060 0.87490 0.18150 0.94620

Coint EQ(2) 0.50515 -2.88952 -2.90529 US (-1) 2.21673 -1.73715 2.03978

0.61380 0.00410 0.00390 0.02720 0.08310 0.04200

UK(-1) -2.02597 -2.41903 1.44042 Japan (-1) 0.84822 2.28480 -2.17663

0.04360 0.01610 0.15070 0.39680 0.02280 0.00690

US (-1) 2.93943 -0.98438 3.35839 C 1.04375 0.65536 1.20140

0.00350 0.32560 0.00090 0.29720 0.51260 0.23030

Japan (-1) 0.90959 2.45093 -2.14894

0.36370 0.01480 0.03230

C 0.04793 0.44967 -1.35512

0.96180 0.65320 0.17630

71 | P a g e
678430

5.0 Conclusion and summary

Through the use of a range of methodologies we have been able to access the dynamic
relationship of European markets before and after the EU referendum. My findings identify
significant changes in Europe pre and post event with significantly lower correlation across
the board. This was supported by our findings in PCA, whereby the post period identifies
more components to explained the same sample set, this could be as a result of global
uncertainty as well as UK detracting its self from it European counterparts, unlike past
studies on correlation that focus on financial crisis I hope to develop an understanding of
the implications of political uncertainty on stock market co-movements. Although Pearsons
correlation coefficient and PCA identified a weakening relationship and further
diversification benefits in Europe, they were not supported by the findings in Johansens
cointegration testing and VAR/ VECM methodologies. These methodologies actually
identified significantly stronger relationship in many regions, very much in line with past
studies on the relevance of geographical locations & market development on
interdependency. Although the methodologies employed have provided inconclusive
evidence as to either weakening of strengthening co-movements for the 6 months following
the EU referendum they provide useful knowledge for the UK investor looking to diversify,
by looking at my findings, less developed markets and those not within a close proximity
offer significantly better diversification benefits for those investors looking to reduce market
risk, significantly those in Central and Eastern Europe, Malta and Cyrus.

Lastly to research further into this topic research would also need to be conducted on not
just a 6-month post event period but also on the long term implications of the EU
referendum over the 2-year time span until article 50 it triggered; as unlike financial shock
the effects of the EU referendum have and will be slow burning and my findings show just
how much uncertainty there is in not just European markets but also global markets.
Ultimately my findings have been inconsistent in proving Contagion in Europe but thats not
to say we can use these findings to understand the UK dynamic relationship with its EU
counterparts during 2016 and access the best markets for investors looking for optimal
portfolio allocations.

72 | P a g e
678430

6.0 Bibliography

Agmon, T. (1972). The Relations among equity markets: A Study of Share Price Co
Movements in The United States, United Kingdom, Germany and Japan. The Journal of
Finance, 27(4), 839-855.

Ahmad, W., Sehgal, S., & Bhanumurthy, N. (2013). Eurozone crisis and BRIICKS stock
markets: Contagion or market interdependence? Economic Modelling, 33, 209-225.

Aityan, S. K., Ivanov-Schitz, A. K., & Izotov, S. S. (2010). Time-shift asymmetric correlation
analysis of global stock markets. Journal of International Financial Markets, Institutions and
Money, 20(5), 590-605.

Ang, A., Bekaert, G., 2002. International asset allocation with time-varying
correlations. Review of Financial Studies 15, 11371187.

Arshanapalli, B., & Doukas, J. (1993). International stock market linkages: Evidence from the
pre- and post-October 1987 period. Journal of Banking & Finance, 17(1), 193-208.
doi:10.1016/0378-4266(93)90088-u

rvai, Z., Driessen, K., & Otker-Robe, I. (2009). Regional financial interlinkages and financial
contagion Within Europe. IMF Working Papers, 1-42.

Baele, L., & Inghelbrecht, K. (2010). Time-varying integration, interdependence and

contagion. Journal of International Money and Finance, 29(5), 791-818

Beirne, J., Caporale, G. M., Schulze-Ghattas, M., & Spagnolo, N. (2010). Global and regional
spillovers in emerging stock markets: A multivariate GARCH-in-mean analysis. Emerging
markets review, 11(3), 250-260.

Bertera, E., C. Mayer, 1990, Structure and performance: Global interdependence of stock
markets around the crash of1987, Euopean Economic Review 34, 1155-1180

Bollerslev, T. (1990). Modelling the coherence in short-run nominal exchange rates: a

multivariate generalized ARCH model. The review of economics and statistics, 498-505.

Bonfiglioli, A., & Favero, C. A. (2005). Explaining co-movements between stock markets: The
case of US and Germany. Journal of International Money and Finance, 24(8), 1299-1316.

Bracker, K., Docking, D. S., & Koch, P. D. (1999). Economic determinants of evolution in
international stock market integration. Journal of Empirical Finance, 6(1), 1-27

Calvo, S. G., & Reinhart, C. M. (1996). Capital flows to Latin America: is there evidence of
contagion effects?. World Bank Policy Research Working Paper, (1619).

73 | P a g e
678430

Chang, E. C., Cheng, J. W., & Khorana, A. (2000). An examination of herd behavior in equity
markets: An international perspective. Journal of Banking & Finance, 24(10), 1651-1679.

Cipriani, M., Gardenal, G., & Guarino, A. (2013). Financial contagion in the laboratory: The
cross-market rebalancing channel. Journal of Banking & Finance, 37(11), 4310-4326.

Connolly, R. A., & Wang, F. (2003). International equity market comovements: Economic
fundamentals or contagion? Pacific-Basin Finance Journal, 11(1), 23-43

Dickey, D. A., & Fuller, W. A. (1979). Distribution of the estimators for autoregressive time
series with a unit root. Journal of the American statistical association, 74(366a), 427-431.

Eiling, E., & Gerard, B. (2014). Emerging equity market comovements: trends and
macroeconomic fundamentals. Review of Finance, rfu036.

Engle, R. (2002). Dynamic Conditional Correlation. Journal of Business & Economic Statistics,
20(3), 339-350

Edwards, S., & Susmel, R. (2001). Volatility Dependence and Contagion in Emerging Equity
Markets.

Eun, C.S., S. Shim, 1989, International transmission of stock market movements. Journal of
Financial and Quantitative Analysis 24, 241-256

Forbes, K. J., & Rigobon, R. (2002). No Contagion, Only Interdependence: Measuring Stock
Market Comovements. The Journal of Finance, 57(5), 2223-2261.

Flavin, T. J., Hurley, M. J., & Rousseau, F. (2002). Explaining stock market correlation: A
gravity model approach. The Manchester School, 70(S1), 87-106.

Gelos, G., & Sahay, R. (2001). Financial Market Spillovers: How Different are the Transition
Economies? International Financial Contagion, 329-366.

Gerrits, R., & Yuce, A. (1999). Short- and long-term links among European and US stock
markets. Applied Financial Economics, 9(1), 1-9

Granger, C. W. (1969). Investigating causal relations by econometric models and cross-

spectral methods. Econometrica: Journal of the Econometric Society, 424-438.

Groenen, P. J., & Franses, P. H. (2000). Visualizing time-varying correlations across stock
markets. Journal of Empirical Finance, 7(2), 155-172.

Hatemi-j, A. (2012). Asymmetric causality tests with an application. Empirical Economics,

43(1), 447-456.

Heaney, R., Hooper, V. J., & Jaugietis, M. (2000). Regional integration of national stock
markets.

74 | P a g e
678430

He, H., Chen, S., Yao, S., & Ou, J. (2014). Financial liberalisation and international market
interdependence: Evidence from Chinas stock market in the post-WTO accession period.
Journal of International Financial Markets, Institutions and Money, 33, 434-444

Hon, M.T., Strauss, J., Yong, S.K., (2004), Contagion in financial markets after September 11:
Myth or reality?, Journal of Financial Research, Vol. 27, no. 1, pp. 95-114

Huang, B. N., Yang, C. W., & Hu, J. W. S. (2000). Causality and cointegration of stock markets
among the United States, Japan and the South China Growth Triangle. International Review
of Financial Analysis, 9(3), 281-297.

Humpe, A., & Macmillan, P. (2009). Can macroeconomic variables explain long-term stock
market movements? A comparison of the US and Japan. Applied Financial Economics, 19(2),
111-119.

Hwang, J. (2014). Spillover Effects of the 2008 Financial Crisis in Latin America Stock
Markets. International Advances in Economic Research, 20(3), 311-324

Johansen, S. (1988). Statistical analysis of cointegrating vectors. Journal of Economic

Dynamics and Control, 12, 231254.

Johansen, S. (1992). Determination of cointergration rank in the presence of a linear trend.

Oxford Bulletin of Economics and Statistics, 54(3), 383397.

Kaminsky, G. L., & Reinhart, C. M. (2000). On crises, contagion, and confusion. Journal of
international Economics, 51(1), 145-168.

Karolyi, G. A., & Stulz, R. M. (1996). Why Do Markets Move Together? An Investigation of
U.S.-Japan Stock Return Comovements. The Journal of Finance, 51(3), 951

King, M. A., & Wadhwani, S. (1990). Transmission of volatility between stock markets.
Review of Financial studies, 3(1), 5-33.

King, M., Sentana, E., & Wadhwani, S. (1994). Volatility and Links between National Stock
Markets. Econometrica, 62(4), 901. doi:10.2307/2951737

Kodres, L. E., & Pritsker, M. (2001). A Rational Expectations Model of Financial Contagion.
The Journal of Finance, 57(2), 769-799

Longin, F., Solnik, B., 1995. Is the correlation in international equity returns constant: 1960
1990. J. Int. Money Finance. 14 (1), 326.

Lessard, D. R. (1976). World, Country, and Industry Relationships in Equity Returns:

Implications for Risk Reduction through International Diversification. Financial Analysts
Journal, 32(1), 32-38.

75 | P a g e
678430

Levy, H., Sarnat, M., (1970), International diversification of investment portfolios, American
Economic Review, Vol. 60, no. 3, pp. 668-675.

Liu, S. Z., Lin, K. C., & Lai, S. M. (2006). Stock market interdependence and trade relations: A
correlation test for the US and its trading partners. Economics Bulletin, 7(5), 1-15.

Lux, T. (1995). Herd Behaviour, Bubbles and Crashes. The Economic Journal, 105(431), 881.
doi:10.2307/2235156

MacKinnon, J. G. (1991). Critical values for cointegration tests, Chapter 13 in Long-Run

Economic Relationships: Readings in Cointegration, ed. RF Engle and CW J. Granger.

Malliaris, A. G., & Urrutia, J. L. (1992). The international crash of October 1987: causality
tests. Journal of Financial and Quantitative Analysis, 27(03), 353-364.

Markowitz, H. M. (1991). Foundations of Portfolio Theory. The Journal of Finance, 46(2), 469.
doi:10.2307/2328831

McQueen, G., Roley, V.V., 1993. Stock prices, news, and business condition. Review of
Financial Studies 6,
683707.

Meric, I., Ratner, M., Nygren, L. M., & Meric, G. (2008). Co-Movements of Latin American
Equity Markets Before and After September 11, 2001. Latin American Business Review, 8(3),
54-74.

Meric, I., Nygren, L. M., Bentley, J. T., & Mccall, C. W. (2015). Co-Movements Of U.S. And
European Stock Markets Before And After The 2008 Global Stock Market Crash. Studies in
Business and Economics, 10(2). doi:10.1515/sbe-2015-0022

Meric, I., Eichhorn, B.H., Nygren, L.M., Meric, G., (2007), Co-movements of U.S., Japanese,
and European equity markets before and after September 11, 2001: Global portfolio
diversification implications, Journal of Management Science (Spring), pp. 1-10.

Meric, I., Meric, G., (1989), Potential gains from international portfolio diversification and
intertemporal stability and seasonality in international stock market relationships, Journal of
Banking and Finance, Vol. 13, no. 4/5, pp. 627-640.7

Meric, I., Meric, G., (1997), Co-movements of European equity markets before and after the
1987 crash, Multinational Finance Journal, Vol. 1, no. 2, pp. 137-152.

Meric, I., Meric, G., (2004), The diminishing benefit of global portfolio diversification, Global
Business and Finance Review, Vol. 9, no. 2, pp. 43-51.

Meric, I., Meric, G., Ratner, M., (2000), Emerging market crisis and the ASEAN-5 stock
markets:1997-1998, International Journal of Finance, Vol. 12, no. 3, pp. 1835-1853.

76 | P a g e
678430

Meric, I., Taga, L.S., Gishlick, H.E., Meric, G., (2015), Co-movements of the worlds stock
markets before and after the 2008 stock market crash, Journal of Strategic and
International Studies, Vol. 10, no. 1, pp. 74-81.

Mobarek, A., Muradoglu, G., Mollah, S., & Hou, A. J. (2016). Determinants of time varying co-
movements among international stock markets during crisis and non-crisis periods. Journal of
Financial Stability, 24, 1-11

Nasseh, A., & Strauss, J. (2000). Stock prices and domestic and international macroeconomic
activity: a cointegration approach. The Quarterly Review of Economics and Finance, 40(2),
229-245.

Nofsinger, J. R., & Sias, R. W. (1999). Herding and feedback trading by institutional and
individual investors. The Journal of finance, 54(6), 2263-2295.

Phillips, P. C., & Perron, P. (1988). Testing for a unit root in time series regression. Biometrika,
335-346.

Philippatos, G. C., Christofi, A., & Christofi, P. (1983). The inter-temporal stability of
international stock market relationships: Another view. Financial Management, 63-69.

Portes, R., & Rey, H. (2005). The determinants of cross-border equity flows. Journal of
international Economics, 65(2), 269-296.

Roll, R. (1992). Industrial structure and the comparative behavior of international stock
market indices. The Journal of Finance, 47(1), 3-41.

Ross, S. A. (1989). Information and Volatility: The No-Arbitrage Martingale Approach to

Timing and Resolution Irrelevancy. The Journal of Finance, 44(1), 1-17

Sasaki, H., Yamaguchi, S., & Hisada, T. (2000). The globalisation of financial markets and
monetary policy. In BIS Conference Papers (Vol. 8, pp. 57-78).

Sharma, S. S. (2015). Time-Varying Herding Behavior, Global Financial Crisis, and the Chinese
Stock Market. Review of Pacific Basin Financial Markets and Policies

Sharpe, W. F. (1964). Capital Asset Prices: A Theory of Market Equilibrium under Conditions
of Risk. The Journal of Finance, 19(3), 425

Solnik, B. H. (1974). The International Pricing of Risk: An Empirical Investigation of the World
Capital Market Structure. The Journal of Finance, 29(2), 365.

Solnik, B. H. (1995). Why Not Diversify Internationally Rather Than Domestically? Financial
Analysts Journal, 51(1), 89-94. 1864

77 | P a g e
678430

Scheicher, M. (2001). The comovements of stock markets in Hungary, Poland and the Czech
Republic. International Journal of Finance & Economics, 6(1), 27-39.
Schwert, W. (1990). Stock Market Volatility . Financial Analysts Journal , 24-34.

Syllignakis, M. N., & Kouretas, G. P. (2011). Dynamic correlation analysis of financial

contagion: Evidence from the Central and Eastern European markets. International Review of
Economics & Finance, 20(4), 717-732.

Von Furstenberg, G.M., B.N. Jeon, 1990, Growing international comovements in stock price
indexes. Quarterly Review of Economics and Business 30,15-30

William G. Christie, R. D. (1995). Following the Pied Piper: Do Individual returns herd around
the market . Financial Analysts Jounal, 31-37.

Wang, Z., Yang, J., & Bessler, D. A. (2003). Financial crisis and African stock market integration.
Applied Economics Letters, 10(9), 527-533
.
Wasim Ahmad, S. S. (2013). Eurozone crisis and BRIICKS stock markets: Contagion or market
interdependence? Economic Modelling 33, 209 - 225 .

Yang, J., Kolari, J. W., & Min, I. (2003). Stock market integration and financial crises: The case
of Asia. Applied Financial Economics, 13(7), 477-486.

Yang, J., Hsiao, C., Qi, L., Wang, Z., 2006. The emerging market crisis and stock market
linkages: further evidence. Journal of Applied Econometrics 21, 727744.

Yao, J. Y., Chan-Lau, J. A., & Mathieson, D. J. (2002). Extreme Contagion in Equity Markets.
IMF Working Papers, 02(98), 1

Yarovaya, L., & Lau, M. C. (2016). Stock market comovements around the Global Financial
Crisis: Evidence from the UK, BRICS and MIST markets. Research in International Business and
Finance, 37, 605-619.

Zhang, B., & Li, X. (2014). Has there been any change in the comovement between the Chinese
and US stock markets? International Review of Economics & Finance, 29, 525-536.
doi:10.1016/j.iref.2013.08.001

78 | P a g e
678430

7.0 Appendixs

79 | P a g e
 678430
Appendix One Volatility for major European Stock Markets calculated to 10 basis points standard Deviation between Jan 2006 November 2016

80 | P a g e
 678430
Appendix Two Stock Market returns between December 2010 November 2016 Indexed weekly at 15/12/2016

81 | P a g e
 678430
Appendix Three Stock Market prices between Jan 2016 November 2016 Indexed daily from 4th Jan 2016

130

120

110

100

90

80

70
1/4/2016 2/4/2016 3/4/2016 4/4/2016 5/4/2016 6/4/2016 7/4/2016 8/4/2016 9/4/2016 10/4/2016 11/4/2016

ATX - Austria NEY - Japan SPX - USA CRO - Croatia FTSE - UK OMX - Sweden BVLX - Portugal IBEX 35 - Spain
AEX - Holland SOFIX - Bulgaria PAX - France BEL20 - Belgium ASE - Greece Maltex - Malta HEX - Finland ITLMS - Italy

82 | P a g e
Appendix678430
Four: Stock Prices indexed at December 2006 Jan 2009

83 | P a g e
 678430
Appendix Five: Major Trading Partners for Europe and the U.S. (As a % of total)

Country Imports destination Exports destination

Austria Germany (38%), Italy (6.3%), Switzerland (5.2%), Czech republic ( 4.1%), China (5.4%), US (3.2%), UK (1.6%) - Europe ( 80.1%) Germany (28%), Italy (6.5%), US (6.4%), Switzerland (5.8%), Hungary (4.2%), France (3.8%), US (3.2%) - Europe (76.1%)
Belgium -
Luxembourg Netherlands (19%), Germany (14%), France (11%), US (7.5%), UK (4.9%), Ireland (4.1%) - Europe (72%) Germany (15%), France (15%), Netherlands (15%), UK (9.3%), Italy (5.5%), US (4.4%), Spain (2.5%) - Europe (75.1%)

Bulgaria Russia (14%), Germany (12%), Italy (6.6%), Turkey (5.7%), Spain (4.9%), Romania (6.7%), UK 1.9%) - Europe ( 79.6%) Germany (12%), Italy (10%), Turkey (9.1%), Romania (6.9%), Greece (5.9%), Belgium (4.5%), UK (2.2%) - Europe (60.7%)
Germany (15%), Italy (15%), Slovenia (9.4%), Austria (8.6%), Hungary (6.1%), Russia (5.5%), Netherlands (3.5%), UK (1.3%) -
Croatia Europe (86.6%) Italy (14%), Germany (11%), Bosnia (10%), Slovenai (9.7%), Austria (6%), Serbia (5%), Russia (3.6%) - Europe ( 85.7%)

Cyprus Greece (19%), Isreal (7.3%), Russia (6.8), Germany (6.4%), UK (5.9%), Italy (5.9%), Spain (4.8%) - Europe ( 77.3%) Greece (14%), UK (7.7%), Isreal (10%), Poland (7.2%), South Korea (4.3%), Gemrnay (2.9%), France (2.8%)- Europe (60%)
Germany (27%), China (11%), Poland (8%), Slovakia (5.7%), Italy ( 4.2%), Russia (3.7%) Netherlands (3.6%), France (3.2%), UK
Czech Republic (2.2%) - Europe (74.3%) Germany (31%), Poland (5.2%), France (5.2%), UK (5%), Austria (4.3%), Italy (3.8%), Slovakia (7.3%) - Europe (86.4%)
Denmark Germany (20%), Sweden (12%), Netherlands (7.7%), Norway (5.5%), UK (4.7%), China (7.1%) - Europe ( 78.3) Germany (14%), Sweden (11%), US (7.7%), UK (7.8%, Norway (5.6%), France (3.6%), China (4.1%) - Europe ( 70%)
Russia (15%), Germany (10%), Finland (9.7%), China (7.6%), Sweden (5.9%), Poland ( 5.9%), Lithuania (5.6%), Latvia (4.8%), Sweden (15%), Russia (15%), Finland (11%), Latvia (6.6%), Spain (5.9%), Lithuania (4.9%), Germany (4%), UK (2.1%) -
Estonia UK (3.1%) - Europe (80.3%) Europe ( 83.7%)
Germany (12%), Sweden (10%) Russia (7.4%), Netherlands (5.8%), US (6.7%), Belgium (3.3%), France (3.1%), UK (5.1%) -
Finland Germany (14%), Russia (14%), Sweden (10%), China (6.8%), Netherlands (6.5%), US (3.9%), UK (3.2%) - Europe ( 76.5%) Europe ( 68%)
France Germany (18%), Beligum (8.6%), china (8.1%), Italy (7.6%), spain (6.4%), Nehterands (4.9%), UK (4.1%) - Europe ( 66.8%) Germany (15%), Belgium (8.3%), UK (7.3%), US ( 7.1%), Italy (6.9%) , Spain (6.2%), netherlands (4%) - Europe ( 63.4%)
Germany (18%), Belgium (8.6%), China (8.1%), Italy (7.6%), Spain (6.4%), Netherlands (4.9%), US (6%), UK (4.1%) - Europe
France (66.8%) Germany (15%), Belgium (8.3%), UK (7.3%), US (7.1%), Netherlands (4%), China (3.9%) - Europe (63.7%)
Nehterlands (10%), France (7.5%), China ( 8.9%, US ( 5.4%), Italy (5.4%), Belgium (5%), Poland (4.5%), UK (4.1%) - Europe US (8.6%). France (8.5%), UK (7.1%), China (6.9%), Netherlands (7.1%), Italy (5%), Austria (4.6%), Belgium (4.5%), Poland
Germany (70%) (3.6%) - Europe (64.6%)
Germany (10%), Russia (9.3%), Italy (7.8%), Iraq (7.8%), China (5.6%), Netherland (5.1%), France (4.7%), UK (2.6%) - Europe
Greece (60.5%) Turkey (13%) Italy (9.3%), Germany (6.8%), Bulgaria (5%), UK (3.8%), US (3.1%) - Germany (53.9%)
Germany (25%), China (5.1%), Russia (7.1%), Austria (6.8%), France (4.8%), Slovakia (4.7%),Czech (4.6%), UK (1.7%) - Europe
Hungary (83.7%) Germany (27%), Romania (5.4%), Italy (5.1%), Austria (5%), France (4.4%), UK (3.9%) - Europe ( 83.4%)
Ireland UK ( 33%), US (10%), Germany (8.7%), Netherlands (5.7%), China ( 5.7%) - Europe ( 69.4%) US ( 20%), UK (14%), Belgium (13%), Germany (7.6%), France (5.6%), Switzerland (5.5%) - Europe (63.2%)
Germany (15%), France (8.5%), China (7.1%), Netherlands (5.9%), Russia (4.9%), Spain (4.7%), Belgium (4.7%), UK (2.9%) -
Italy Europe ( 66.7%) Germany (12%, France (9.8%), US (8%) , UK (5.5%), Spain (4.2%), Belgium *3.3%), Austria (2.1%) - Europe ( 61%)
Japan China (22%), US (9%), Austrailia (5.7%) - Europe (13.9%) China (18%), US (18%), South Korea (7.3%),Hong kong (5%) , Germany (3.4%) - Europe (14.6%)
Latvia Lithuania (16%), Russia (12%), Poland (10%), Germany (10%), Estonia (6.8%), Finland (5.6%) - Europe (89.5%) Lithuania (16%), Russai (9.1%), Estonia (6.6%), Belarus (6%), UK (4.9%) - Europe ( 86%)
Lithuania Russia (21%), Germany (10%), Poland (9%), Latvia (6.7%), Netherlands (4.8%), Italy (4.7%), UK (4%) - Europe ( 87.8%) Russia (15%), Belarus (10%), Latvia (8.9%), Germany (6.8%), Poland (5.3%), UK (5.3%) - Europe ( 85%)
Malta Italy (14%), Russia (7.1%), Germany (5.1%), UK (4.7%) - Europe ( 53%) Egypt (13%), South Korea (7.4%), Germany (6.8%) - Europe ( 34.7%)
Netherlands Germany (15%), Belgium (9.3%), China (9.3), Russia (7.1%), US ( 6.3%) , UK (6.2%) - Europe (61%) Germany (22%), Belgium (16%), UK (9.7%), France (6.1%), Ital (5.2%), US (3.8%), Spain (2.7%) - Europe - (78%)

Poland Germany (23%), China (10%), Russia (8.1%), Italy (5.7%), Netherlands (3.9%), France (3.9%) - Europe (74.3%) Germany (25%), France (5.6%) , Italy (4.6%), UK ( 6.3%) , Czech (5.9%), Russia (4.2%), Netherlands (4.1%) - Europe ( 84.4%)
Portugal Spain (32%), Germany (12%), France (7%), Italy (5.1%), Netherlands (5.1%), UK (3%) - Europe ( 76.3%) Spain (21%), France (11%), Germany (11%), Angola (6.7%), UK (6.1%), Netherlands (3.9%) - Europe (70%)
Romania Germany (19%), Italy (11%), Hungary (7.6%), France (5.7%), Poland (4.5%), Austria (3.6%) - Europe (81.1%) Germany (18%), Italy (11%) , France (6.1%), Hungary (4.1%), Turkey (4.9%) - Europe ( 75.7%)
Slovakia Germany (17%), Czech (15%), China (7.9%), Russia (7.6%), Poland (5.8%), Hungary (5.5%) - Europe (74.7%) Germany (22%), Czech (11%), Poland (6.2%), Hungary (6%), France (5.4%), Austria (5.4%), UK (5%) - Europe (88%)
Germany (13%), France (10%) , China (7.6%), ItaLY (6.2%), UK (4.2%), Netherlands (4.1%), Portugal (3.9%), US (3.9%) -
Spain Europe ( 58.5%) France (14%), Germany (11%), Portugal (8.4), UK (7.3%). Italy (7.2%). US (4.7%), Netherlands (3.3%) - Europe ( 67.7%)
Germany ($46.5B 9.8%), United States ($51B 11%), Netherlands ($34.2B 7.2%), Switzerland (7.1%), France (5.7%), China
United Kingdom Germany ( 15%), Netherlands (7.6%), China (9.4%), US (6.7%), France (6.3%), Belgium (5.3%), Itly (4.2%) - Europe (63%) (5.7%) - Europe (57%)
United States China (20%), Canada (15%), Mexico (13%), Japan (5.9%), Germany (5.5%) UK (2.3%) - Europe (20%) Canada (17%), Mexico (13%), China (9.2%) Germany (4.2%), Japan (4.6%), UK (3.1%), France (2.7%) - Europe 23%
Slovenia Germany (17%), Italy (15%), Austria (8.8%), Croatia (4.5%), China (5.8%) - Europe (78.1%) Germany (20%), Italy (11%), Austria (8.3%), Croatia ( 6.9%), France (5.2%), Russia (4.5%) - Europe ( 87.9%)
Sweden Germany (17%), Netherlands (8.1%), Denmark (7.2%), Norway (6.6%), UK (6%), Finland (5.1%) - Europe (82.5%) Germany (11%), UK (7.7%), Denmark (7.3%), Norway (7.2%), Netherlands (5.2%), US (6.4%) - Europe (70%)
84 | P a g e
 678430

Appendix Six: Trading Hours UTC

Trading Hours UTC

Country 00:00 01:00 02:00 03:00 04:00 05:00 06:00 07:00 08:00 09:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 23:00 00:00
Austria
Belgium
Denmark
Finland
France
Germany
Hungary
Ireland
Italy
Latvia
Luxembourg
Malta
Netherlands
Poland
Portugal
Spain
Sweden
UK
USA
Japan

85 | P a g e
 678430

Appendix 7: Stock Market Indexs used

Country Index used Abbrieviation Currency

Austria Vienna Stock Exchange Austrian Traded index ATX EUR
Belgium Belgium 20 BEL20 EUR
Bulgaria Bulgarian stock Market SOFIX BGN
Croatia Croatia Zagreb Crobex Index CRO CRO
Cyprus Cyrus Stock exchange general index CYSMMAPA EUR
Czech republic Czech Traded Index PX CZK
Denmark OMX copenhagen index KAX DKK
Estonia OMX Tallinn index TALSE EUR
Finland OMX Helsinki Index HEX EUR
France CAC All share index CAC EUR
Germany DAX Performance index DAX EUR
Greece Athex Composite Share Price ASE EUR
Hungary Budapest stock exchange index BUX HUF
Ireland Ireland overall index ISEQ EUR
Italy FTSE Italia all share index ITLMS EUR
Latvia OMX Riga Index RIGSE EUR
Luxembourg Luxembourg Luxx Index LUXXX EUR
Malta Malta stock Exchange index MALTEX EUR
Netherlands Amsterdam exchange Index AEX EUR
Poland Warsaw Stock Exchange Index WIG PLN
Portugal PSI all share index BVLX EUR
Romania Budapest Bet Index BET RON
Slovakia Slovak share index SKSM EUR
Slovenia Slovenian blue chip index SBITOP EUR
Spain Madrid Stock Exchange General Index IBEX EUR
Sweden OMX All Share index SAX SEK
UK FTSE 100 UKX GBP
USA S&P 500 Index SPX USD
Japan Nikkei 300 Index NEY JPY

86 | P a g e
 678430

Appendix 8: Pearson Correlation Coefficient Matrix (Pre period)

Correlations - Pre Period

UK Japan US Sweden Spain Slovenia Slovakia Romania Portugal Poland Netherlands Malta Luxembourg Latvia Italy Ireland Hungary Greece Germany France Finland Estonia Denmark Czech Croatia Bulgaria Belgium Austria Cyprus
UK 1 .284
**
.617
**
.875
**
.853
**
.326
**
.074 .502
**
.799
**
.568
**
.930
**
-.076 .657
**
.143 .826
**
.780
**
.506
**
.429
**
.834
**
.919
**
.853
**
.088 .775
**
.628
**
.231
*
.291
**
.887
**
.822
**
.094
Japan .284
**
1 .092 .289
**
.365
**
.357
**
.134 .201
*
.341
**
.304
**
.340
**
-.124 .261
**
.130 .269
**
.304
**
.349
**
.383
**
.316
**
.336
**
.335
**
.314
**
.365
**
.437
**
.239
**
.064 .330
**
.245
**
.144
US .617
**
.092 1 .542
**
.566
**
.091 -.040 .284
**
.522
**
.364
**
.632
**
-.058 .426
**
.205
*
.550
**
.479
**
.360
**
.237
**
.544
**
.592
**
.528
**
.105 .411
**
.399
**
.117 .266
**
.601
**
.552
**
.210
*

Sweden .875
**
.289
**
.542
**
1 .845
**
.288
**
.051 .497
**
.767
**
.595
**
.876
**
-.087 .615
**
.071 .819
**
.813
**
.526
**
.494
**
.845
**
.902
**
.888
**
.141 .801
**
.705
**
.226
*
.278
**
.888
**
.841
**
.128
Spain .853
**
.365
**
.566
**
.845
**
1 .351
**
.068 .500
**
.831
**
.507
**
.907
**
-.109 .607
**
.142 .916
**
.810
**
.535
**
.482
**
.836
**
.923
**
.824
**
.142 .779
**
.749
**
.258
**
.321
**
.898
**
.831
**
.167
Slovenia .326
**
.357
**
.091 .288
**
.351
**
1 .076 .223
*
.350
**
.360
**
.311
**
-.038 .237
**
.088 .246
**
.312
**
.256
**
.317
**
.288
**
.332
**
.332
**
.233
*
.317
**
.294
**
.314
**
.122 .309
**
.326
**
.161
Slovakia .074 .134 -.040 .051 .068 .076 1 -.020 -.001 .052 .036 .005 .116 .051 .026 .046 .177 .044 -.033 .069 .083 .286
**
.076 .060 -.026 -.058 .015 -.017 .133
Romania .502
**
.201
*
.284
**
.497
**
.500
**
.223
*
-.020 1 .504
**
.462
**
.523
**
-.018 .401
**
.121 .500
**
.421
**
.546
**
.323
**
.486
**
.495
**
.477
**
.163 .454
**
.482
**
.353
**
.278
**
.531
**
.507
**
.137
Portugal .799
**
.341
**
.522
**
.767
**
.831
**
.350
**
-.001 .504
**
1 .590
**
.835
**
-.119 .645
**
.157 .816
**
.711
**
.508
**
.480
**
.745
**
.840
**
.812
**
.106 .688
**
.687
**
.301
**
.327
**
.828
**
.782
**
.122
Poland .568
**
.304
**
.364
**
.595
**
.507
**
.360
**
.052 .462
**
.590
**
1 .530
**
-.114 .442
**
.216
*
.458
**
.432
**
.426
**
.379
**
.485
**
.536
**
.567
**
.278
**
.405
**
.514
**
.248
**
.171 .499
**
.596
**
.127
Netherlands .930
**
.340
**
.632
**
.876
**
.907
**
.311
**
.036 .523
**
.835
**
.530
**
1 -.127 .663
**
.113 .870
**
.816
**
.516
**
.479
**
.883
**
.964
**
.880
**
.117 .799
**
.671
**
.280
**
.274
**
.940
**
.841
**
.143
Malta -.076 -.124 -.058 -.087 -.109 -.038 .005 -.018 -.119 -.114 -.127 1 -.055 -.083 -.086 -.109 -.011 .053 -.188* -.133 -.092 -.218* -.078 -.040 -.001 .107 -.104 -.057 .113
Luxembourg .657** .261** .426** .615** .607** .237** .116 .401** .645** .442** .663** -.055 1 .123 .616** .507** .354** .324** .607** .641** .665** .133 .523** .509** .122 .173 .582** .638** -.004
Latvia .143 .130 .205* .071 .142 .088 .051 .121 .157 .216* .113 -.083 .123 1 .133 .121 .247** .049 .105 .102 .113 .170 .134 .184* .105 .150 .097 .159 .162
Italy .826 **
.269 **
.550 **
.819 **
.916 **
.246 **
.026 .500 **
.816 **
.458 **
.870 **
-.086 .616 **
.133 1 .788 **
.540 **
.487 **
.797 **
.884 **
.803 **
.107 .763 **
.685** .219* .301** .889** .808** .144
Ireland .780** .304** .479** .813** .810** .312** .046 .421** .711** .432** .816** -.109 .507** .121 .788** 1 .502** .524** .784** .837** .768** .122 .806** .630** .283** .289** .856** .733** .139
Hungary .506** .349** .360** .526** .535** .256** .177 .546** .508** .426** .516** -.011 .354** .247** .540** .502** 1 .424** .599** .555** .522** .242** .522** .613** .226* .267** .548** .511** .236**
Greece .429** .383** .237** .494** .482** .317** .044 .323** .480** .379** .479** .053 .324** .049 .487** .524** .424** 1 .426** .462** .469** .291** .590** .534** .337** .102 .544** .497** .258**
Germany .834 **
.316 **
.544 **
.845 **
.836 **
.288 **
-.033 .486 **
.745 **
.485 **
.883 **
-.188 *
.607 **
.105 .797 **
.784 **
.599 **
.426 **
1 .904 **
.837 **
.112 .773 **
.630 **
.232 *
.306 **
.882 **
.781 **
.093
France .919** .336** .592** .902** .923** .332** .069 .495** .840** .536** .964** -.133 .641** .102 .884** .837** .555** .462** .904** 1 .889** .113 .799** .681** .263** .282** .945** .837** .163
Finland .853** .335** .528** .888** .824** .332** .083 .477** .812** .567** .880** -.092 .665** .113 .803** .768** .522** .469** .837** .889** 1 .157 .769** .709** .240** .303** .868** .804** .183*
Estonia .088 .314** .105 .141 .142 .233* .286** .163 .106 .278** .117 -.218* .133 .170 .107 .122 .242** .291** .112 .113 .157 1 .180* .179 .129 .088 .096 .122 .196*
Denmark .775 **
.365 **
.411 **
.801 **
.779 **
.317 **
.076 .454 **
.688 **
.405 **
.799 **
-.078 .523 **
.134 .763 **
.806 **
.522 **
.590 **
.773 **
.799 **
.769 **
.180 *
1 .708 **
.261 **
.317 **
.829 **
.751 **
.101
Czech .628** .437** .399** .705** .749** .294** .060 .482** .687** .514** .671** -.040 .509** .184* .685** .630** .613** .534** .630** .681** .709** .179 .708** 1 .255** .361** .697** .712** .230*
Croatia .231* .239** .117 .226* .258** .314** -.026 .353** .301** .248** .280** -.001 .122 .105 .219* .283** .226* .337** .232* .263** .240** .129 .261** .255** 1 .151 .302** .223* .250**
Bulgaria .291** .064 .266** .278** .321** .122 -.058 .278** .327** .171 .274** .107 .173 .150 .301** .289** .267** .102 .306** .282** .303** .088 .317** .361** .151 1 .303** .287** .119
Belgium .887** .330** .601** .888** .898** .309** .015 .531** .828** .499** .940** -.104 .582** .097 .889** .856** .548** .544** .882** .945** .868** .096 .829** .697** .302** .303** 1 .859** .178
Austria .822** .245** .552** .841** .831** .326** -.017 .507** .782** .596** .841** -.057 .638** .159 .808** .733** .511** .497** .781** .837** .804** .122 .751** .712** .223* .287** .859** 1 .152
Cyprus .094 .144 .210* .128 .167 .161 .133 .137 .122 .127 .143 .113 -.004 .162 .144 .139 .236** .258** .093 .163 .183* .196* .101 .230* .250** .119 .178 .152 1

87 | P a g e
 678430

Appendix 9: Pearson Correlation Coefficient Matrix (Post period)

Correlations - Post
UK Japan US Sweden Spain Slovenia Slovakia Romania Portugal Poland Netherlands Malta Luxembourg Latvia Italy Ireland Hungary Greece Germany France Finland Estonia Denmark Czech Croatia Bulgaria Belgium Austria Cypus
UK **
0 0.202 0.604 0.689 0.662 .099 .004 0.416 0.744 0.5 0.827 .121 0.602 -.031 0.627 0.658 0.524 0.494 0.736 0.791 0.67 0.254 0.583 0.45 0.212 .120 0.761 0.621 .221
Japan *
0.202 0 0.217 .058 0.501 .102 .105 0.467 0.383 0.339 0.33 .088 .165 .069 0.473 0.242 0.374 0.498 0.356 0.413 .028 .158 0.189 0.4 0.178 0.293 0.281 0.401 .179
US *
0.604 0.217 0 0.514 0.646 .016 -.071 0.317 0.614 0.421 0.721 .078 0.55 .087 0.685 0.655 0.345 0.512 0.709 0.728 0.547 0.187 0.519 0.377 0.178 .097 0.733 0.622 .208
Sweden 0.689 .058 0.514 0 0.468 -.002 .018 0.249 0.606 0.391 0.702 .057 0.532 .129 0.486 0.773 0.4 0.317 0.662 0.644 0.927 0.346 0.597 0.451 .088 .041 0.677 0.603 .086
Spain **
0.662 0.501 0.646 0.468 0 .154 -.007 0.564 0.847 0.55 0.862 .093 0.652 .051 0.881 0.711 0.585 0.721 0.853 0.921 0.452 0.19 0.578 0.563 0.239 .109 0.873 0.769 .345
Slovenia .099 .102 .016 -.002 .154 0 .000 0.234 .115 .151 .079 .145 .104 .052 .136 .105 0.247 0.196 .135 .135 .032 .151 .040 .130 0.205 .161 .100 .156 .137
Slovakia .004 .105 -.071 .018 -.007 .000 0 .145 -.025 -.038 .006 -.011 -.004 .045 -.021 -.044 .048 -.055 .012 -.001 -.008 .102 .007 .079 .016 -.055 -.026 -.012 -.069
Romania 0.416 0.467 0.317 0.249 0.564 0.234 .145 0 0.445 0.402 0.448 .025 0.304 .032 0.502 0.396 0.46 0.521 0.482 0.529 0.209 0.261 0.379 0.39 0.32 .135 0.454 0.443 .204*
Portugal 0.744 0.383 0.614 0.606 0.847 .115 -.025 0.445 0 0.541 0.848 .068 0.594 .047 0.767 0.673 0.576 0.65 0.806 0.852 0.588 0.198 0.625 0.499 0.201 0.175 0.815 0.724 .263**
Poland 0.5 0.339 0.421 0.391 0.55 .151 -.038 0.402 0.541 0 0.59 .016 0.437 .090 0.575 0.478 0.56 0.511 0.601 0.585 0.369 .104 0.371 0.459 .126 0.17 0.556 0.523 .185*
Netherla 0.827 0.33 0.721 0.702 0.862 .079 .006 0.448 0.848 0.59 0 .054 0.698 .038 0.81 0.782 0.582 0.628 0.914 0.944 0.688 0.227 0.661 0.519 0.217 .095 0.923 0.79 .261**
nds
Malta .121 .088 .078 .057 .093 .145 -.011 .025 .068 .016 .054 0 .009 .012 .119 .105 0.182 .160 .092 .101 .059 .092 .107 .084 .030 -.013 .106 .103 .110
Luxembo 0.602 .165 0.55 0.532 0.652 .104 -.004 0.304 0.594 0.437 0.698 .009 0 .046 0.591 0.658 0.413 0.482 0.722 0.726 0.582 0.176 0.492 0.505 0.205 -.009 0.696 0.629 .210*
urg
Latvia -.031 .069 .087 .129 .051 .052 .045 .032 .047 .090 .038 .012 .046 0 .012 .100 .015 .082 .057 .045 .109 .054 .126 .041 .020 .014 .041 .084 -.030
Italy 0.627 0.473 0.685 0.486 0.881 .136 -.021 0.502 0.767 0.575 0.81 .119 0.591 .012 0 0.706 0.55 0.717 0.812 0.866 0.471 0.206 0.529 0.559 0.307 .149 0.842 0.799 .376**
Ireland 0.658 0.242 0.655 0.773 0.711 .105 -.044 0.396 0.673 0.478 0.782 .105 0.658 .100 0.706 0 0.486 0.593 0.789 0.812 0.737 0.244 0.581 0.532 0.17 .048 0.828 0.748 .294**
Hungary 0.524 0.374 0.345 0.4 0.585 0.247 .048 0.46 0.576 0.56 0.582 0.182 0.413 .015 0.55 0.486 0 0.562 0.557 0.595 0.373 .140 0.449 0.475 0.219 .142 0.538 0.535 .285**
Greece 0.494 0.498 0.512 0.317 0.721 0.196 -.055 0.521 0.65 0.511 0.628 .160 0.482 .082 0.717 0.593 0.562 0 0.664 0.709 0.325 0.224 0.397 0.481 0.41 .147 0.665 0.592 .508**
Germany 0.736 0.356 0.709 0.662 0.853 .135 .012 0.482 0.806 0.601 0.914 .092 0.722 .057 0.812 0.789 0.557 0.664 0 0.946 0.653 0.23 0.635 0.519 0.267 .070 0.891 0.773 .329**
France 0.791 0.413 0.728 0.644 0.921 .135 -.001 0.529 0.852 0.585 0.944 .101 0.726 .045 0.866 0.812 0.595 0.709 0.946 0 0.639 0.201 0.613 0.564 0.257 .102 0.921 0.823 .323**
Finland 0.67 .028 0.547 0.927 0.452 .032 -.008 0.209 0.588 0.369 0.688 .059 0.582 .109 0.471 0.737 0.373 0.325 0.653 0.639 0 0.309 0.551 0.422 .117 .032 0.665 0.608 .089
Estonia *
0.254 .158 0.187 0.346 0.19 .151 .102 0.261 0.198 .104 0.227 .092 0.176 .054 0.206 0.244 .140 0.224 0.23 0.201 0.309 0 0.275 0.337 .098 .094 0.247 0.268 .177
Denmark 0.583 0.189 0.519 0.597 0.578 .040 .007 0.379 0.625 0.371 0.661 .107 0.492 .126 0.529 0.581 0.449 0.397 0.635 0.613 0.551 0.275 0 0.399 .114 -.014 0.653 0.504 .102
Czech **
0.45 0.4 0.377 0.451 0.563 .130 .079 0.39 0.499 0.459 0.519 .084 0.505 .041 0.559 0.532 0.475 0.481 0.519 0.564 0.422 0.337 0.399 0 0.185 .045 0.563 0.597 .279
Croatia **
0.212 0.178 0.178 .088 0.239 0.205 .016 0.32 0.201 .126 0.217 .030 0.205 .020 0.307 0.17 0.219 0.41 0.267 0.257 .117 .098 .114 0.185 0 .076 0.259 0.28 .290
Bulgaria .120 0.293 .097 .041 .109 .161 -.055 .135 0.175 0.17 .095 -.013 -.009 .014 .149 .048 .142 .147 .070 .102 .032 .094 -.014 .045 .076 0 .039 .093 .016
Belgium **
0.761 0.281 0.733 0.677 0.873 .100 -.026 0.454 0.815 0.556 0.923 .106 0.696 .041 0.842 0.828 0.538 0.665 0.891 0.921 0.665 0.247 0.653 0.563 0.259 .039 0 0.783 .329
Austria **
0.621 0.401 0.622 0.603 0.769 .156 -.012 0.443 0.724 0.523 0.79 .103 0.629 .084 0.799 0.748 0.535 0.592 0.773 0.823 0.608 0.268 0.504 0.597 0.28 .093 0.783 0 .236
Cypus 0.221 0.179 0.208 .086 0.345 .137 -.069 0.204 0.263 0.185 0.261 .110 0.21 -.030 0.376 0.294 0.285 0.508 0.329 0.323 .089 0.177 .102 0.279 0.29 .016 0.329 0.236 0
*. Correlation is significant at the 0.05 level (2-tailed).
**. Correlation is significant at the 0.01 level (2-tailed).

88 | P a g e
678430

Appendix 10: KMO & Bartletts test results for PCA

Pre period

KMO and Bartlett's Test

Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .934

Bartlett's Test of Sphericity Approx. Chi-Square 3305.051

df 406

Sig. .000

Post Period

KMO and Bartlett's Test

Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .929

Bartlett's Test of Sphericity Approx. Chi-Square 3608.491

df 406

Sig. .000

89 | P a g e
 678430

Appendix 11: PCA Pre Period Full Results

Rotated Component Matrix - Pre Period

Total
Component & Factor Loadings Variance
Explained in
these
1 2 3 4 5 components
France .952 .166 .065 .032 -.015 0.939
Netherlands .946 .167 .079 .000 -.026 0.929
Belgium .936 .209 .063 -.029 .049 0.927
UK .923 .112 .102 .013 -.017 0.876
Sweden .916 .169 .046 .035 .009 0.872
Spain .908 .203 .104 .040 .016 0.879
Italy .903 .122 .099 .006 .045 0.841
Finland .893 .191 .087 .073 .008 0.847
Germany .891 .158 .097 -.041 -.082 0.836
Austria .869 .179 .145 -.031 .025 0.809
Ireland .837 .226 .024 .017 .039 0.755
Portugal .833 .253 .153 -.058 -.040 0.787
Denmark .816 .281 .010 .074 .059 0.754
Luxembourg .699 .055 .066 .125 -.112 0.524
Czech .689 .347 .210 .094 .129 0.664
US .623 -.140 .396 -.014 .038 0.566
Hungary .509 .303 .359 .219 .153 0.551
Poland .495 .373 .308 .052 -.177 0.513
Romania .467 .354 .317 -.144 .081 0.471
Croatia .121 .674 .152 -.242 .153 0.574
Slovenia .214 .646 -.005 .049 -.061 0.469
Japan .232 .606 -.007 .286 -.148 0.525
Greece .441 .575 -.075 .147 .217 0.6
Latvia .040 .034 .764 .108 -.133 0.615
Bulgaria .284 .031 .479 -.187 .296 0.434
Slovakia .043 -.037 -.038 .852 .103 0.74
Estonia .015 .398 .294 .553 -.229 0.603
Malta -.089 -.074 -.102 -.050 .823 0.704
Cyprus .042 .276 .347 .251 .486 0.498
% of variance 44.45% 9.65% 5.86% 4.92% 4.45%
Explained
Cumulative % 44.45% 54.10% 59.96% 64.88% 69.33%
Extraction Method: Principal Component Analysis.
Rotation
a. Method:
Rotation Varimax
converged with Kaiser Normalization.
in 7 iterations.

90 | P a g e
 678430

Appendix 12: PCA Post Period Full Results

Rotated Component Matrix - Post period

Total
Component & factor Loadings Variance
Explained in
these
Stock Market 1 2 3 4 5 6 7 components
Netherlands .938 .170 .088 .028 .034 -.009 -.036 0.919
Belgium .921 .128 .196 .028 -.027 .041 -.018 0.907
France .919 .253 .184 -.019 .035 .034 -.021 0.946
Germany .899 .193 .188 .024 .007 .024 .005 0.882
Ireland .861 .008 .132 .136 .005 .060 .078 0.787
Portugal .839 .242 .093 -.018 .134 .032 -.028 0.791
UK .817 .051 .041 .179 .097 .055 -.136 0.735
Spain .814 .399 .231 -.113 .037 .055 -.016 0.892
Sweden .802 -.185 -.133 .413 .023 -.009 .092 0.875
Finland .797 -.242 -.090 .386 .030 -.016 .082 0.858
Austria .796 .228 .179 .085 .065 .038 .053 0.773
Italy .788 .341 .284 -.092 .074 .057 -.037 0.837
US .771 .019 .113 -.067 .035 -.020 .066 0.618
Luxembourg .752 .029 .161 .070 -.079 -.066 .007 0.609
Denmark .704 .090 -.068 .185 -.064 .101 .107 0.568
Poland .581 .332 .035 -.087 .247 .041 .070 0.524
Greece .573 .368 .518 -.076 .107 .136 .058 0.771
Hungary .534 .391 .137 .067 .165 .283 -.049 0.571
Czech .529 .351 .171 .272 -.027 .085 -.005 0.514
Japan .215 .754 .109 -.071 .231 .057 .058 0.692
Romania .360 .584 .258 .207 .102 -.040 -.006 0.593
Slovakia -.104 .523 -.198 .443 -.411 -.107 .002 0.7
Croatia .125 .096 .723 .149 .048 -.153 .048 0.598
Cypus .204 .054 .723 -.023 -.069 .157 -.092 0.606
Estonia .183 .065 .158 .742 .100 .070 .007 0.628
Bulgaria .035 .191 -.065 .056 .846 -.094 -.029 0.771
Slovenia -.014 .091 .356 .283 .408 .307 .089 0.483
Malta .059 .018 .012 .042 -.045 .921 .001 0.856
Latvia .054 .040 -.015 .024 -.002 .006 .980 0.966
% of variance 46.01% 7.52% 4.64% 4.24% 3.85% 3.49% 3.47%
Explained
Cumulative % 46.01% 53.52% 58.16% 62.40% 66.25% 69.74% 73.21%
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.

a. Rotation converged in 15 iterations.

91 | P a g e
678430

Appendix 13: Stationery testing equations

Augmented Dicker Fuller test (ADF) - see Dickey and Fuller 1979
we use the Augmented Dicker Fuller) test denoted by:

Where and the null and alternative hypothesis can be written as

and evaluated using the conventional -ratio for =

Showing that under the null hypothesis of a unit root, this statistic does not follow the conventional
Students t-distribution, and they derive asymptotic results and simulate critical values for various
test and sample sizes. The Augmented Dickey-Fuller (ADF) test constructs a parametric correction for

higher-order correlation by assuming that the series follows an AR( process and adding lagged

difference terms of the dependent variable to the right-hand side of the test regression:

92 | P a g e
678430

Appendix 14: stationery Testing equations (2)

PhillipsPeron Test (PP) See Phillips & Perron (1988).

The PP method estimates the non-augmented DF test equation 5 and modifies the -ratio

of the coefficient so that serial correlation does not affect the asymptotic distribution of
the test statistic. The PP test is based on the statistic:

where is the estimate, and the -ratio of , is coefficient standard error, and

is the standard error of the test regression. In addition, is a consistent estimate of the

error variance in EQ 5 (calculated as , where is the number of regressors).

The remaining term, , is an estimator of the residual spectrum at frequency zero.

93 | P a g e
678430

Appendix 15: Lag Length Criteria Equations

Akaike Information Criterion:

The Akaike Information Criterion (AIC) is computed as:

where is the log likelihood denoted by:

Schwarz Criterion:

The Schwarz Criterion (SC) is an alternative to the AIC that imposes a larger penalty for
additional coefficients:

Hannan-Quinn Criterion:

The Hannan-Quinn Criterion (HQ) employs yet another penalty function:

94 | P a g e
 678430

Appendix 16: Full Grangers Causal Linkages Results

Grangers Causality test results (full F-stasts & Prob. avalible upon request)
Market Pre Period Spillover on Post Period Spillover on
Austria Czech, Estonia, Greece, Japan Czech, Finland, Estonia, Latvia, Ireland, Japan, Luxembourg , Sweden
Belgium Croatia, Czech, Italy, Denmark, Slovakia, Estonia, Finland, France, Germany, Hungary, Greece Austria, Czech, Finland, Ireland, Latvia, Luxembourg, Netherlands, Sweden, US, UK
Bulgaria Czech, Netherlands, France, Hungary, Spain , UK Japan
Croatia Poland Denmark, Finland, Ireland, Luxembourg, Sweden
Cyprus Finland, Estonia, Ireland, Japan, Luxembourg, Sweden
Czech republic Croatia, Japa, Slovenia Cyprus, Finland, Estonia, Romania, Sweden
Belgium, Greece, Croatia, Ireland, Cyprus, Italy, UK, Estonia, Japan, Netherands, Sweden, Portugal, Finland, Estonia, Slovakia, Sweden,
Denmark
Spain
Estonia
Finland Croatia, Estonia, Greece, Japan Hungary, Ireland, Japan, Romania
France Croatia, Czech, Denmark, Estonia, Greece, Japan, Slovenia Austria, Czech, Japan, Denmark, Finland,Estonia, Ireland, Latvia, Luxembourg, Sweden, US
Germany Croatia, Czech, Denmark, Estonia, Greece, Japan, Austria, Czech, Finland, Estonia, Sweden, Ireland, Japan, Luxembourg, Poland
Croatia, Denmark, Estonia, Hungary, Japan, Slovenia, Slovakia Austrai, Czech, Denmark, Finland, Estonia, France, Germany, Ireland, Italy, Japan, Lativa, Netherlands,
Greece
Poland, Porugal, Sweden, US, UK
Croatia, Japan, Latvia, Slovenia Austria, Belgium, Denmark, Finland, Estonia, France, Germany, Ireland, Italy, Luxembourg,
Hungary
Netherlands, Spain, Sweden, US, UK
Ireland Croatia, Cyprus, Japan, Denmark, Estonia, Finland, Hungary , Greece Czech, Finland, Estonia, Japan, Luxembourg, Sweden
Italy Bulgria, Czech, Croatia, Denmark, Estonia, Hungary, Greece, Japan, Slovenia Czech, Denmark, Estoni, Ireland, Luxembourg, Sweden, US, UK
Latvia Croatia, Slovakia
Luxembourg Bulgria, Hungaria, Latvia,Croatia, Czech, Estonia, Japan Estonia, Finland, France, Germany,hungary, Japan, Latvia, Malta, Portgugal, Spain, Sweden
Malta Croatia, Portugal Finland, Luxembourg, Estonia
Austria, Denmark, Croatia, Czech, Estonia, Finland, Hungary, France, Japan Germany, Poland, Austria, Czech, Finland, Estonia, Japan, Latvia, Sweden, US
Netherlands
Luxembourg, Slovakia, Sweden, Latvia, Romania, Slovenia
Austria, Belgium, Denmark, Finland, Estonia, France, Germany, Ireland, Italy, Luxembourg, Sweden,
Poland
US
Portugal Bulgaria, Estonia, Croatia, Czech, Hungary, Greece, Japan, Poland, Slovenia Austria, Czech, Denmark, Finland, Estonia, Ireland, Latvia, Luxembourg, UK
Croatia, Estonia, Slovenia, Japan Austria, Belgium, Czech, Finland, Estonia, France, Germany, Ireland , Luxembourg, Netherlands,
Romania
Poland, Slovenia, Sweden, US, UK
Slovakia Austria, Greece
Slovenia Finland, Japan, Luxembourg, Romania
Bulgaria, Estonia, Czech, Greece, Hungary, Japan, Poland Austria, Czech, Denmark, Finland, Estonia, France, Germany, Ireland, Italy, Latvia, Luxembourg,
Spain
Netherlands, Sweden, US, UK
Sweden Croatia, Denmark, Czech, Estonia, Japan, Slovenia Belgium, France, UK, Malta, Ireland, Portugal, Japan, Romania
UK Croatia, Denmark, Czech, Estonia, Greece, Japan, Poland, Slovenia Finland, Estonia, Latvia, Luxembourg, Sweden
UK, Spain, Portugal, Sweden, Slovenia, Slovakia, Romania, Luxembourg, Netherlands, Poland, Austria, UK, Japan , Czech, Latvia, Denmark, Luxembourg, Finland, Estonia, Poland, Ireland, Sweden
USA Austria, Belgium, Bulgaria, Estonia, France, Greece, Italy, Croatia, Czech, Denmark, Finland, Germany,
Hungary, Ireland, Japan
Japan Czech Belgium, Germany, Czech, Denmark , Ireland, Finland, Latvia, Estonia, Luxembourg, Sweden, US

95 | P a g e
678430

Appendix 17 Markets in Each Region for Johanson and VAR/VECM Purposes

W.Europe S.Europe C.Europe E. Europe N.Europe G10 G7 Pacific

Austria
Belgium
Bulgaria
Croatia
Cyprus
Czech republic
Denmark
Estonia
Finland
France
Germany
Greece
Hungary
Ireland
Italy
Latvia
Luxembourg
Malta
Netherlands
Poland
Portugal
Romania
Slovakia
Slovenia
Spain
Sweden
UK
USA
Japan

96 | P a g e
678430

Appendix 18: Ethics Approval Form Students

Ethics Approval Form - Students

This form should be completed by the student and passed to the supervisor prior to a review
of the possible ethical implications of the proposed dissertation or project.
No primary data collection can be undertaken before the supervisor has approved the
plan.
If, following review of this form, amendments to the proposals are agreed to be necessary, the
student should provide the supervisor with an amended version for endorsement.

The final signed and dated version of this form must be handed in with the dissertation. The form
MUST be signed and dated by both the student AND the supervisor. If the dissertation is submitted
without a fully completed, signed and dated ethics form it will be deemed to be a fail. Second
attempt assessment may be permitted by the Board of Examiners.

1. What are the objectives of the dissertation / research project?

To identify the impact of the EU referendum on stock prices and their co-movements

2. Does the research involve NHS patients, resources or staff? YES / NO (please circle).
If YES, it is likely that full ethical review must be obtained from the NHS process
before the research can start.
3. Does the research involve MoD staff? YES / NO (please circle).
If YES, then ethical review may need to be undertaken by MoD REC. Please discuss
your proposal with your Supervisor and/or Course Leader and, if necessary, include a
copy of your MoD REC application for quality review.
4. Do you intend to collect primary data from human subjects or data that are identifiable
with individuals? (This includes, for example, questionnaires and interviews.) YES /
NO (please circle)
If you do not intend to collect such primary data then please go to question 15.
If you do intend to collect such primary data then please respond to ALL the questions
5 through 14. If you feel a question does not apply then please respond with n/a (for not
applicable).
5. How will the primary data contribute to the objectives of the dissertation / research
project?

6. What is/are the survey population(s)?

97 | P a g e
678430

7. How big is the sample for each of the survey populations and how was this sample
arrived at?

8. How will respondents be selected and recruited?

9. What steps are proposed to ensure that the requirements of informed consent will be
met for those taking part in the research? If an Information Sheet for participants is to
be used, please attach it to this form. If not, please explain how you will be able to
demonstrate that informed consent has been gained from participants.

10. How will data be collected from each of the sample groups?

11. How will data be stored and what will happen to the data at the end of the research?

12. What measures will be taken to prevent unauthorised persons gaining access to the
data, and especially to data that may be attributed to identifiable individuals?

13. What steps are proposed to safeguard the anonymity of the respondents?

14. Are there any risks (physical or other, including reputational) to respondents that may
result from taking part in this research? YES / NO (please circle).
If YES, please specify and state what measures are proposed to deal with these risks.

15. Are there any risks (physical or other, including reputational) to the researcher or to the
University that may result from conducting this research? YES / NO (please circle).

98 | P a g e
678430

If YES, please specify and state what measures are proposed to manage these risks.2

16. Will any data be obtained from a company or other organisation. YES / NO (please
circle) For example, information provided by an employer or its employees.
If NO, then please go to question 19.

17. What steps are proposed to ensure that the requirements of informed consent will be
met for that organisation? How will confidentiality be assured for the organisation?

18. Does the organisation have its own ethics procedure relating to the research you intend
to carry out? YES / NO (please circle).
If YES, the University will require written evidence from the organisation that they
have approved the research.

19. Will the proposed research involve any of the following (please put a next to yes or
no; consult your supervisor if you are unsure):

Vulnerable groups (e.g. children) ? YES NO

Particularly sensitive topics ? YES NO
Access to respondents via gatekeepers ? YES NO
Use of deception ? YES NO
Access to confidential personal data ? YES NO
Psychological stress, anxiety etc ? YES NO
Intrusive interventions ? YES NO

If answers to any of the above are YES, how will the associated risks be minimised?

20. Are there any other ethical issues that may arise from the proposed research?
N/A

2
Risk evaluation should take account of the broad liberty of expression provided by the principle of
academic freedom. The universitys conduct with respect to academic freedom is set out in section 9.2 of the
Articles of Government and its commitment to academic freedom is in section 1.2 of the Strategic Plan 2004-
2008.

99 | P a g e
678430

Please print the name of: I/We grant Ethical Approval

student Shaun Parris supervisor Lena Itangata

Signed: Signed
(student) (supervisor)
Date 02/03/2017 Date 02/03/2017

AMENDMENTS
If you need to make changes please ensure you have permission before the primary data
collection. If there are major changes, fill in a new form if that will make it easier for
everyone. If there are minor changes then fill in the amendments (next page) and get them
signed before the primary data collection begins.

100 | P a g e