Search

Cycles in the behavior of stock markets have been widely documented. There is an increasing body of literature on whether stock markets anticipate business cycles or its turning points. Several recent studies assert that financial integration impacts positively on business cycle comovements of economies. We consider three Western equity markets, represented by their respective stock indices: DJIA (USA), FTSE 100 (UK), and Euro Stoxx 50 (euro area). Connecting these three markets together via vector autoregressive processes in index returns, we construct "propagation values" to measure and trace, on a daily basis, the relative importance of a market as a volatility creator within the network, where volatility is due to a return shock in a market. A cross-wavelet analysis reveals the joint frequency structure of pairs of the propagation value series, in particular whether or not two series tend to move in the same direction at a given frequency. Our main findings are: (i) From 2001 onwards, the daily propagation values of markets have been fluctuating much less than before, and high frequencies have become less pronounced; (ii) the European markets are in phase at business cycle frequency, while the US market is not in phase with either European market; (iii) in 2008, the euro area has taken over the leading role. This approach not only provides new insight into the time-dependent interplay of equity markets, but it can also replicate certain findings of traditional business cycle research, and it has the advantage of using only readily available stock market data.

WOS:000390952900001 The goal of the this paper is to investigate the shock spillover characteristics of the Russian stock market during different rounds of sanctions imposed as a reaction to Russia's alleged role in the Ukrainian crisis. We consider six stock markets, represented by their respective stock indices, namely the US (DJIA), the UK (FTSE), the euro area (Euro Stoxx 50), Japan (Nikkei 225), China (SSE Composite) and Russia (RTS). Linking these markets together in a network on the basis of vector autoregressive processes, we can measure, among other things: (i) direct daily return and volatility spillovers from RTS to other market indices, (ii) daily propagation values quantifying the relative importance of the Russian stock market as a return or volatility shock propagator, and (iii) the amount of network repercussions after a shock. The last two are methodological innovations in this context. It turns out that distinct spillover patterns exist in different rounds of sanctions. Large-scale sanctions, beginning in July 2014, rendered the consequences of shocks from Russia less predictable. While these sanctions reduced the importance of the Russian stock market as a propagator of return shocks, they also increased its importance as a propagator of volatility shocks, thus making the network more vulnerable with respect to volatility shocks from the Russian stock market. This is a form of backlash that the sanctioning economies have been facing.

The degree of connectedness of equity markets on a given day can be assessed by decomposing the forecast error variance resulting from a vector autoregressive model, applied to daily returns on stock indices. This well-known procedure leads, for each day, to a spillover matrix which can be readily interpreted as a network structure and summarized into the so-called spillover index, to which much recent research work has been devoted. Taking a sequence of spillover matrices as starting point, we show how the scope of this concept can be broadened in several ways. Firstly, we develop a concept to quantify a market's potential to spread information, which is related to the eigenvector structure of spillover matrices. Secondly, a Markov chain approach allows the definition of relative market entropy, quantifying the amount of information gained from day to day. A further entropy concept can be related to the speed of shock digestion and network stability. As an empirical example, we analyze a system of five markets represented by stock indices Dow Jones Industrial Average (USA), FTSE (UK), Euro Stoxx 50 (euro area), Nikkei 225 (Japan), and SSE Composite (China). It is demonstrated that increasing trends in the spillover index as well as in speed of information digestion are an empirical fact but no logical necessity theoretical examples show that there can be opposite trends in these series.

The household income distribution (i e. how many households had which income) gives important information for social research We investigate the concentration and distribution of the income earned in 1994 by households in 19 Turkish province centers. The empirical bas is is a recent publication of DİE (Devlet İstatistik Enstitüsü). An analysis of the rank correlation between average income and income concentration reveals interesting features of Turkey., The household income distribution (i e. how many households had which income) gives important information for social research. We investigate the concentration and distribution of the income earned in 1994 by households in 19 Turkish province centers. The empirical basis is a recent publication of DİE (Devlet İstatistik Enstitüsü). An analysis of the rank correlation between average income and income concentration reveals interesting features of Turkey.

Bu tezde vektör otoregresif modelden hesaplanan öngörü hatası varyans ayrıştırması ile oluşturulmuş özel bir bağlanmışlık çerçevesi incelenmiştir. İlgili model günlük borsa endeks getirileri kullanılarak regresyon tahmini yöntemi ile hesaplanmış ve taşma tablosu oluşturulmuştur. Bir ağ yapısını gösteren bu tablo, taşma endeksi adı verilen bir ölçü ile özetlenmiştir.Bir dizi taşma tablosu ile yola çıkılarak, bu çerçeve muhtelif doğrultularda geliştirilmiştir. İlk olarak bir pazarın bilgi yayma potansiyeli hesaplanmıştır. İkinci olarak günden güne üretilen bilgi miktarı ölçülmüştür. Bunun ötesinde ağın bilgi işleme hızını ve kararlılığını ölçen bir entropi kavramı da oluşturulmuştur.Yukarıda ölçüler kullanılarak hali hazırda kullanılan model belirleme yöntemlerinin genellikle yanlış tahminlerde bulunduğu gösterilmiştir.Bu tezde uygun bir yapısal vektör otoregresif model belirleme yöntemi seçeneği olarak `ayrı varyans kullanımı ile belirleme` yöntemi önerilmektedir.Bu yöntemde incelenen veri kümesindeki yapısal şokların varyanslarındaki kırılmalar belirleme işleminde kullanılmaktadır. Ampirik bir uygulama olarak beş pazardan (`Sistemik Beşli`) oluşan bir sistem çözümlenmiştir. Bu sistemi Dow Jones Sanayi, FTSE, EuroStoxx-50, Nikkei-225 ve Şangay Bileşik endeksleri oluşturmaktadır. Ayrı varyans kullanımı ile belirleme yöntemi, pazarların müreffeh ve kararlı olduğu dönemlerde doğudaki pazarların batıdaki pazarlara göre bağıl bilgi yayma özelliğinin oldukça arttığını göstermiştir. Benzer şekilde meşakkatli zamanlarda ise bu etkinin tersine döndüğü gözlenmiştir.Ağ kararlılığı bağlamında, sistem kararlılığının taşma endeksi ile doğrusal olarak arttığı gözlemlenmektedir. Daha yüksek taşma endeksi seviyelerinde ise sistem kararlılığı önce belli miktarda azalmakta daha sonra ise tekrar doğrusal olarak artmaktadır.Sonuç olarak; ağ bağlanmışlık çerçevesi kapsamında ayrı varyans kullanımı ile belirleme yönteminin daha geniş aralıktaki ağ hareket kümelerinin incelenebilmesi için uygun bir seçenek olduğu düşünülmektedir. Our focus in this thesis is a particular network connectedness framework developed through decomposing the forecast error variance that results from a vector auto regressive (VAR) model. The model is estimated using daily returns on stock indices, yielding a spillover table which can be interpreted as a network structure and summarized in a measure so-called spillover index.Departing from a sequence of spillover tables, this framework has been enhanced in several directions with concepts to quantify: (i) a market's potential to spread information, (ii) the amount of information gained from day to day and (iii) the speed of information processing and the degree of network stability. We evaluate the performance of the extant identification approaches, and show that they are arbitrarily misestimating with respect to the above-mentioned network measures. As an alternative, we propose the use of identification through heteroscedasticity (ITH) which utilizes heteroscedasticity in the data to identify structural VARs.As an empirical example, we analyze `Systemic Five` system of markets representing stock indices DJIA, FTSE, EuroStoxx-50, Nikkei-225 and SSEC.ITH uncovers that at prosperous and stable times, eastern markets' relative importance with respect to the western ones' increase significantly, and during times of trouble their importance decrease drastically.In terms of network stability, we observe that the system stability increases linearly with increasing spillovers followed by temporary decreases and then linear increases again. We conclude that ITH could be proposed as a viable identification approach candidate in the network connectedness framework to investigate richer sets of network dynamics. 143

52 page In this study the connectedness of global stock markets are discussed. After the recent credit crisis the correlation between the financial assets are discussed more than before. In this study the connectedness cycles of stock markets’ returns are investigated by applying Forecast Error Decomposition (FEVD) from Vector Auto-Regression (VAR) models. Spillover, directional spillover and net directional spillover methods are used (Diebold & Yilmaz, 2012). The study aims to find out how “The Credit Rating Upgrades”, which was one of the most popular themes for Turkey in 2012, is affecting the change in the connectedness or not. Obtained results show that the connectedness of the stock markets of the countries on the way to the announcement of the investment grade is decreasing until the event, but after the announcement, it starts to increase. On the other hand, the stock markets of such countries become relatively the affected ones instead of being the affecting ones, but they cannot protect this capability for a longer period. Bu çalışmada küresel hisse piyasalarının birbirleriyle olan bağlantıları ele alınmaktadır. Son yaşanan kredi krizinin ardından finansal varlıklar arasındaki benzer hareketler eskisinden daha çok tartışılmaktadır. Bu çalışmada 11 uluslararası hisse endeksinin getirilerindeki olası etkileşim döngüleri Vector Autoregression (VAR) ve Forecast Error Variance Decomposition (FEVD) modelleri ile incelenmektedir. Yayılım, yönlü yayılım ve net yönlü yayılım metodları kullanılmaktadır (Diebold & Yilmaz, 2012). Etkileşimdeki değişimlerde, 2012 yılında Türkiye için hayli popüler bir konu olan “Kredi Notu Artışları”ndan ne ölçüde etkilendiği sorgulanmıştır. Elde edilen sonuçlar notu yatırım yapılabilir seviyeye yükseltilen ülkelerin hisse piyasalarının bu süreçte etkileşimlerinin azaldığını, ancak bir süre sonra yeniden arttığını göstermektedir. Öte yandan notu yükseltilen ülkelerin hisse piyasaları aynı süreçte göreceli olarak diğer piyasalardan etkilenen değil, diğer piyasaları etkileyen haline gelmiş, ancak bu özelliklerini kayda değer bir sürede koruyamamışlardır. CONTENTS 1. Introduction 6 2. Data 12 3. Methodology 15 3.1. Total Spillover 15 3.2. Directional Spillover & Net Directional Spillover 17 3.3. Spillover Plots 18 3.4. Smoothing – Local Regression 19 4. Results 21 4.1. Spillover Measures 22 4.2. Hypothesis 1 27 4.3. Hypothesis 2 32 4.4. Turkey’s Case 37 5. Robustness 38 6. Conclusion 41 References 42 Appendix A RELATIVE NOMINAL VALUES & LOG RETURNS (as %) OF THE DATA SERIES 45 Appendix B SMOOTHED RELATIVE DIRECTIONAL SPILLOVER PLOTS & SMOOTHED NET DIRECTIONAL SPILLOVER PLOTS OF THE STOCK MARKETS 49 Appendix C SMOOTHED RELATIVE DIRECTIONAL SPILLOVER PLOTS & SMOOTHED NET DIRECTIONAL SPILLOVER PLOTS OF TURKEY AS MARKED WITH CRITICAL MACRO-EVENTS 52

36 pages This thesis investigates the effect of the fraud in Societe Generale Bank in the global financial connectedness in crisis times. This fraud was exposed and released to the public towards the end of January 2008, in the middle of the sub-prime mortgage crisis. Furthermore, the research focuses on two years' time period and divides this period into three sub periods according to literature. These three sub periods contain three different market structures. This division is used for assessing the effects of this fraud with information gain theory and financial connectedness. Two hypotheses are tested in this thesis. First one is the impact of the fraud depending on these three periods. That is tested with KLIC values, which measures the difference between today's news balance and yesterday's news balance, in order to denote how much information is created today. The second hypothesis is the propagation values of selected data would decline if this fraud happened in another week after the crisis. That is tested with propagation values. The propagation values measure the strength ofrepercussions in markets when a hypothetical shock comes from these markets. As results of these two tests, empirical findings support the two hypotheses. This fraud had different impacts in different market structures. The amount of the information that one event can create, differs in different market structures. Bu tez, kriz zamanlarmdaki finansal bağlılığı, bir Fransız bankası olan Societe Generale'de yaşanan usulsüzlük ve dolandırıcılık üzerinden incelemektedir. Bu olay 2008'in Ocak ayının sonuna doğru ortaya çıkmıştır. Amerika' da başlayan konut kriziyle aynı zamana denk gelmektedir. Bu araştırma da kabaca, krizin ilk iki yılını kapsamaktadır. Bu süreç literatürde de tartışıldığı üzere, finansal piyasalarda oluşan yapısal değişiklikler temel alınarak üç alt döneme ayrılmıştır. Bu bölünme, piyasaya gelen yeni bilgi akışının ve finansal bağlılığın degerlendirilmesine imkan tanımıştır. Bu konu ile alakalı iki hipotez bu çalışmada test edilmiştir. İlki; Societe Generale bankasında yapllan dolandırıcılığın etkisi üç alt dönemde birbirinden farklı olacaktı. İkincisi; eğer bu olay ilerideki kriz sonrası bir dönemde olursa, bu dolandırıcılığın yayılma etkisi daha az olacaktı. İlk hipotez KLIC değerleri, denilen ve piyasayı etkileyecek bilgilerin miktarıdaki değişimin gün be gün ölçülmesine imkan sağlayan yöntem ile test edilmektedir. İkinci hipotez ise yayılma değerleri ile ölçülmektedir. Testlerin sonucunda çıkan değerler iki hipotezi de destekler mahiyettedir. l.Introduction 1.1 Summary of the Subprime Crisis 2 2. Literature Review 4 2.1 Contagion 4 2.2Transmission Channels 5 2.3 Three Periods 7 3. Societe Generale Scandal 11 3.1 Societe Generale 11 3.2 Jerome Kerviel 12 3.3 The Fraud 13 3.4 The Effect of Fraud 15 4. Data 16 5. Methodology 17 6. Empirical Findings 18 6.1 KLIC Values 18 6.2 Propagation Values 22 7. Summary and Conclusions 24 References 26

106 pages