This thesis studies realized volatility (RV), implied volatility (IV) and their applications in stochastic volatility models. The first essay uses both daytime and overnight high-frequency price data for equity index futures to estimate the RV of the S\&P500 and NASDAQ 100 indexes. Empirical results reveal strong inter-correlation between the regular-trading-time and after-hour RVs, as well as a significant predictive power of overnight RV on daytime RV and vice versa. We propose a new day-night realized stochastic volatility (DN-SV-RV) model, where the daytime and overnight returns are jointly modeled with their RVs, and their latent volatilities are correlated. The newly proposed DN-SV-RV model has the best out-of-sample return distribution forecasts among the models considered. The second essay extends the realized stochastic volatility model by jointly estimating return, RV and IV. We examine how RV and IV enhance the estimation of the latent volatility process for both the S\&P500 index and individual stocks. The third essay re-examines asymmetric stochastic volatility (ASV) models with different return-volatility correlation structures given RV and IV. We show by simulation that estimating the ASV models with return series alone may infer erroneous estimations of the correlation coefficients. The incorporation of volatility measures helps identify the true return-volatility correlation within the ASV framework. Empirical evidence on global equity market indices verifies that ASV models with additional volatility measures not only obtain significantly different estimations of the correlations compared to the benchmark ASV models, but also improve out-of-sample return forecasts. / Thesis / Doctor of Philosophy (PhD)

This diploma thesis focuses on modeling volatility in financial time series. The main approach to modelling volatility is using GARCH models which can capture the variability of conditional volatility of time series. For modelling a conditional mean value in time series are used ARMA models. In the series there are usually not fulfilled the assumption of earnings normality, therefore, are the earnings in most cased characterized by the leptokurtic shape of distribution. The thesis introduces some more distribution types, which can be more easily used for the earnings distribution - above all the Students t distribution. The aim of the thesis in the first part is to present the topic of financial time series and description of the GARCH models including their further modification. There are used e.g. IGARCH or other models capturing asymmetric impact of shocks such as GJR-GARCH. The second part deals with generated data, where are more in detail explored the volatility models and their behavior in corresponding financial time series. The third part focuses on the volatility estimation and forecasting for the financial time series. Firstly this concerns development of stock index MICEX secondly currency pair Russian Ruble to Czech Crown and eventually price development of the Brent crude oil. The goal of the third part is to present the impacts on volatility of chosen time series applied on the example of economic sanctions against Russia after annexation of the Crimea peninsula which happened in the first quarter 2014.

Thesis (MBA (Business Management))--University of Stellenbosch, 2009. / ENGLISH ABSTRACT: Econophysics is a relatively new branch of physics. It entails the use of models in physics applied to economics. The distributions of financial time series are the aspect most intensely studied by physicists. This study is based on a study by Kleinert and Chen who applied the Boltzmann distribution to stock exchange data to define a market temperature that may be used by investors to indicate an impending stock market crash. Most econophysicists’ analysed the tail regions of the distributions as the tails represent risk in financial data. This study’s focus of analysis, on the other hand is the characterisation of the central portion of the probability distribution. The Boltzmann distribution, a cornerstone in statistical physics, yields an exponential distribution. The objective of this study is to investigate the suitability of using a market volatility forecasting method from econophysics, namely the Boltzmann/market temperature method. As econometric benchmark the ARCH/GARCH method is used. Stock market indices are known to be non-normally (non-Gaussian) distributed. The distribution pattern of a stock market index of reasonable high sampling frequency (typically interday or intraday) is leptokurtic with heavy tails. Mesoscopic (interday) distributions of financial time series have been found to be exponential distributions. If the empirical exponential distribution is therefore interpreted as a Boltzmann distribution, then a market temperature can be calculated from the exponential distribution. Empirical data for this study is in the form of daily closing values of the Johannesburg Stock Exchange (JSE) All Share Index (ALSI) and the Standard & Poor 500 (S & P 500) index for the period 1995 through to 2008. The Kleinert and Chen study made use of intraday data obtained from established markets. This study differs from the Kleinert and Chen study in that interday data obtained from an emerging market, namely the South African stock market is used. Neither of the aforementioned two differences had a significant influence on the results of this study. The JSE ALSI log-return data displays non-Gaussian properties and the Laplace (double exponential) distribution fit the data well. A plot of the market temperature provided a clear indication of when stock market crashes occurred. Results of the econophysical (Boltzmann/market temperature) method compared well to results of the econometric (ARCH/GARCH) method and subject to certain improvements can be utilised successfully. A leptokurtic, non-Gaussian nature was established for daily log-returns of the JSE ALSI and the S & P 500 index. The Laplace (double exponential) distribution fit the annual logreturns of the JSE ALSI and S & P 500 index well. As a result of the good Laplace fit, annual market temperatures could be calculated for the JSE ALSI and the S & P 500 index. The market temperature method was effective in identifying market crashes for both indices, but a limitation of the method is that only annual market temperatures can be determined. The availability of intraday stock index data should improve the interval for which market temperature can be determined. / AFRIKAANSE OPSOMMING: Ekonofisika is ‘n relatiewe nuwe studieveld. Dit behels die toepassing van fisiese modelle op finansiële data. Die waarskynlikheidsversdelings van finansiële tydreekse is die aspek wat meeste deur fisisie bestudeer word. Hierdie studie is gebaseer op ‘n studie deur Kleinert en Chen. Hulle het die Boltzmann-verspreiding op ‘n aandele-indeks toegepas en ‘n mark-temperatuur bepaal. Hierdie mark-temperatuur kan deur ontleders gebruik word as waarskuwingsmeganisme teen moontlike aandelebeurs ineenstortings. Die meeste fisisie het die uiterste areas van die verspreidingskurwes geanaliseer omdat hierdie uiterste area risiko in finansiële data verteenwoordig. Die analitiese fokus van hierdie studie, aan die ander kant, is die karakterisering van die die sentrale areas van die waarskeinlikheidsverdeling. Die Boltzmann verspreiding, die hoeksteen van Statistiese Fisika lewer ‘n eksponensiële waarskynlikheidsverdeling. Die doel van hierdie studie is om ‘n ondersoek te doen na die geskiktheid van die gebruik van ‘n ekonofisiese, vooruitskattingsmetode, naamlik die Boltzmann/mark-temperatuur model. As ekonometriese verwysing is die “ARCH/GARCH” metode toegepas. Aandelemark indekse is bekend vir die nie-Gaussiese verspreiding daarvan. Die verspreidingspatroon van ‘n aandelemark indeks met‘n redelike hoë steekproef frekwensie (in die orde van ‘n dag of minder) is leptokurties met breë stert-dele. Mesoskopiese (interdag) verspreidings van finansiële tydreekse is getipeer as eksponensieël. Indien die empiriese eksponensiële-verspreiding as ‘n Boltzmann-verspreiding geinterpreteer word, kan ‘n mark-temperatuur daarvoor bereken word. Empiriese data vir die gebruik in hierdie studie is in die vorm van daaglikse sluitingswaardes van die Johannesburgse Effektebeurs (JSE) se Alle Aandele Indeks (ALSI) en die Standard en Poor 500 (S & P 500) indeks vir die periode 1995 tot en met 2008. Die Kleinert en Chen studie het van intradag data vanuit ‘n ontwikkelde mark gebruik gemaak. Hierdie studie verskil egter van die Kleinert en Chen studie deurdat van interdag data vanuit ‘n opkomende mark, naamlik die Suid-Afrikaanse aandelemark, gebruik is. Nie een van die twee voorafgaande verskille het ‘n beduidende invloed op die resultate van hierdie studie gehad nie. Die JSE ALSI se logaritmiese opbrengs data vertoon nie-Gaussiese eienskappe en die Laplace (dubbeleksponensiële) verspreiding beskryf die data goed. ‘n Grafiek van die mark-temperatuur vertoon duidelik wanneer aandelemarkineenstortings plaasgevind het. Resultate van die ekonofisiese (Boltzmann/mark-temperatuur) metode vergelyk goed met resultate van die ekonometriese (“ARCH/GARCH”) metode en onderhewig aan sekere verbeteringe kan dit met sukses toegepas word. ‘n Leptokurtiese, nie-Gaussiese aard is vir daaglike opbrengswaardes vir die JSE ALSI en die S & P 500 indeks vasgestel. ‘n Laplace (dubbel-eksponensiële) verspreiding kan goed op die jaarlikse logaritmiese opbrengste van die JSE ALSI en die S & P 500 indeks toegepas word. As gevolg van die goeie aanwending van die Laplace-verspreiding kan ‘n jaarlikse mark-temperatuur vir die JSE ALSI en die S & P 500 indeks bereken word. Die mark-temperatuur metode is effektief in die identifisering van aandelemarkineenstorings vir beide indekse, hoewel daar ‘n beperking is op die aantal mark-temperature wat bereken kan word. Die beskikbaarheid van intradag aandele indekswaardes behoort die interval waarvoor mark-temperature bereken kan word te verbeter.

Submitted by Marcia Bacha (marcia.bacha@fgv.br) on 2012-03-07T12:45:08Z No. of bitstreams: 1 20120306084421880.pdf: 1716342 bytes, checksum: e7f9f7df4b67ff4e12f57770620942d8 (MD5) / Approved for entry into archive by Gisele Isaura Hannickel (gisele.hannickel@fgv.br) on 2012-03-07T12:50:42Z (GMT) No. of bitstreams: 1 20120306084421880.pdf: 1716342 bytes, checksum: e7f9f7df4b67ff4e12f57770620942d8 (MD5) / Made available in DSpace on 2012-03-07T12:51:26Z (GMT). No. of bitstreams: 1 20120306084421880.pdf: 1716342 bytes, checksum: e7f9f7df4b67ff4e12f57770620942d8 (MD5) / Com origem no setor imobiliário americano, a crise de crédito de 2008 gerou grandes perdas nos mercados ao redor do mundo. O mês de outubro do mesmo ano concentrou a maior parte da turbulência, apresentando também uma explosão na volatilidade. Em meados de 2006 e 2007, o VIX, um índice de volatilidade implícita das opções do S&P500, registrou uma elevação de patamar, sinalizando o possível desequilíbrio existente no mercado americano. Esta dissertação analisa se o consenso de que a volatilidade implícita é a melhor previsora da volatilidade futura permanece durante o período de crise. Os resultados indicam que o VIX perde poder explicativo ao se passar do período sem crise para o de crise, sendo ultrapassado pela volatilidade realizada. / Started in the U.S. housing sector, the credit crisis of 2008 caused great damage in markets around the world. The effects were concentrated in October of the same year, which also showed an explosion in volatility. In mid-2006 and mid-2007, the VIX, an index of implied volatility of options on the S&P500, recorded a rise in level signaling the possible imbalance in the U.S. market. This dissertation examines whether the consensus that implied volatility is the best predictor of future volatility remains during the crisis. The results indicate that the VIX loses explanatory power to move from a period of economic stability for a period of crisis, been surpassed by the realized volatility.

This research investigates to what extent the ARCH model and the GARCH model forecasts one-day-ahead out-of-sample daily volatility (conditional variance) in European AT1 CoCo bonds compared to the Random Walk model. The research also investigates how different orders of ARCH and GARCH models affect the forecasting accuracy. Specifically, the models investigated are the Random Walk model, ARCH(1), ARCH(2), ARCH(3), GARCH(1,1), GARCH(1,2), GARCH(2,1), and the GARCH(2,2)model. The data set used in this report is 47 European AT1 CoCo bonds from 20 different issuers.The results show that 42 out of 47 CoCo bonds have daily log returns that are conditional heteroscedastic. Five CoCo bonds with homoscedastic daily log returns were CoCo bonds with significant low liquidity. The results show that the GARCH model outperforms both the Random Walk model and the ARCH model, under the assumption that the innovations follow a normal distribution. The results also show that a higherorder of ARCH or GARCH does not necessarily lead to more accurate forecasts. The GARCH(1,1) model provided the most accurate predictions. The conclusion is that the GARCH models provide accurate volatility forecasts in CoCo bonds compared to the ARCH-model, and the Random Walk model. However, the ARCH model and the GARCH model fail to forecast the daily volatility in CoCo bondswith insufficient liquidity. Furthermore, a higher order of ARCH or GARCH models does not necessarily lead to better forecast results. / Denna uppsats undersöker till vilken utsträckning som ARCH och GARCH-modeller kan prediktera daglig volatilitet i AT1 CoCo-obligationer (eng. Additional Tier-1 Contingent Convertible Bonds), jämfört med Random Walk-modellen. Uppsatsen undersöker även hur olika parametrar I ARCH och GARCH-modeller påverkar resultatet i prediktionerna. De modeller som undersöks är Random Walk-modellen, ARCH(1), ARCH(2), ARCH(3), GARCH(1,1), GARCH(1,2), GARCH(2,1), och GARCH(2,2)-modellen. Datasetet som har använts i denna forskning består av 47 Europeiska AT1 CoCo obligationer från 20 olika emittenter. Resultatet visar att 42 av 47 CoCo-obligationer har betingat heteroskedastisk daglig avkastningsdata. Fem CoCo-obligationer med homoskedastisk avkastningsdata är obligationer med signifikant låg likviditet. Vidare visar resultatet visar att GARCH modellen överpresterar jämfört med både Random Walk-modellen och ARCH-modellen, under antagandet att innovationstermen följer en normal distribution. Resultatet visar även att en högre ordning av ARCH eller GARCH inte nödvändigtvis leder till ett bättre resultat i prediktonerna. GARCH(1,1)-modellen är modellen som predikterar den dagliga volatiliten i CoCo-obligationerna med bäst resultat. Slutsatsen är att GARCH-modellen predikterar volatiliteten i CoCo-obligationer bättre jämfört med ARCH-modellen och Random Walk-modellen. Däremot kan inte ARCH-modellen eller GARCH-modellen modellera CoCo-obligationer med signifikant låg likviditet. Vidare så medför en högre ordning i ARCH eller GARCH-modellen inte nödvändigtvis till bättre prediktioner.

本文提出一個動態隱含波動度函數模型，以改善一般隱含波動度函數難以隨時間的經過而調整波動度曲線且無法描述資料的時間序列特性等缺點。本文模型為兩階段隱含波動度函數模型，分別配適隱含波動度函數的時間穩定(time-invariant)部分與時間不穩定(time-variant)部分。 本文模型在波動度的時間不穩定部分配適非對稱GARCH(1,1)過程，以描述隱含波動度的時間序列特性。本文使用的非對稱GARCH(1,1)過程將標的資產的正報酬與負報酬對價平隱含波動度的影響分別估計，並將蘊含於歷史價平隱含波動度中的訊息及標的資產報酬率與波動度之間的關連性藉由價平隱含波動度過程納入隱含波動度函數中，使隱含波動度函數能納入波動度的時間序列特性及資產報酬與波動度的相關性，藉此納入最近期的市場資訊，以增加隱含波動度模型的解釋及預測能力。時間穩定部分則根據Pena et al.(1999)的研究結果，取不對稱二次函數形式以配適實證上發現的笑狀波幅現象。時間穩定部分並導入相對價內外程度做為變數，以之描述價內外程度、距到期時間、及價平隱含波動度三者的交互關係；並以相對隱含波動度作為被解釋變數，使隱含波動度函數模型除理論上包含了比先前文獻提出的模型更多的訊息及彈性外，還能描繪「隱含波動度函數隨波動度的高低水準而變動」、「越接近到期日，隱含波動度對價內外程度的曲線越彎曲」、「隱含波動度函數為非對稱的曲線」、「波動度和資產價格有很高的相關性」等實證上常發現的現象。 本文以統計測度及交易策略之獲利能力檢定模型的解釋能力及預測能力是否具有統計與經濟上的顯著性。本文歸納之前文獻提出的不同隱含波動度函數模型，並以之與本文提出的模型做比較。本文以台指選擇權五分鐘交易頻率的成交價作為實證標的，以2003年1月1日~2006年12月31日作為樣本期間，並將模型解釋力及AIC作為模型樣本內配適能力之比較標準，我們發現本文提出的模型具有最佳的資料解釋能力。本文以2006年7月1日~2006年12月31日作為隱含波動度模型預測期間，以統計誤差及delta投資策略檢定模型的預測能力是否具有統計及經濟上的顯著性。實證結果指出，本文提出的模型對於預測下一期的隱含波動度及下一期的選擇權價格，皆有相當良好的表現。關於統計顯著性方面，我們發現本文提出的動態隱含波動度函數模型對於未來的隱含波動度及選擇權價格的預測偏誤約為其他隱含波動度函數模型的五分之一，而預測方向正確頻率亦高於預測錯誤的頻率且超過50%。關於經濟顯著性方面，本文使用delta投資組合進行經濟顯著性檢定，結果發現在不考慮交易成本下，本文提出的模型具有顯著的獲利能力。顯示去除標的資產價格變動對選擇權造成的影響後，選擇權波動度的預測準確性確實能經由delta投資組合捕捉；在考慮交易成本後，各模型皆無法獲得超額報酬。最後，本文提出的動態隱含波動度函數模型在考量非同步交易問題、30分鐘及60分鐘等不同的資料頻率、不同的投資組合交易策略後，整體的結論依然不變。 / This paper proposes a new implied volatility function to facilitate implied volatility forecasting and option pricing. This function specifically takes the time variation in the option implied volatility into account. Our model considers the time-variant part and fits it with an asymmetric GARCH(1,1) model, so that our model contains the information in the returns of spot asset and contains the relationship of the returns and the volatility of spot asset. This function also takes the time invariant in the option implied volatility into account. Our model fits the time invariant part with an asymmetric quadratic functional form to model the smile on the volatility. Our model describes the phenomena often found in the literature, such as the implied volatility level increases as time to maturity decreases, the curvature of the dependence of implied volatility on moneyness increases as options near maturity, the implied volatility curve changes as the volatility level changes, and the implied volatility function is an asymmetric curve. For the empirical results, we used a sample of 5 minutes transaction prices for Taiwan stock index options. For the in-sample period January 1, 2003–June 30, 2006, our model has the highest adjusted- and lowest AIC. For the out-of-sample period July 1, 2006–December 31, 2006, the statistical significance shows that our model substantially improves the forecasting ability and reduces the out-of-sample valuation errors in comparison with previous implied volatility functions. We conjecture that such good performance may be due to the ability of the GARCH model to simultaneously capture the correlation of volatility with spot returns and the path dependence in volatility. To test the economic significance of our model, we examine the profitability of the delta-hedged trading strategy based on various volatility models. We find that although these strategies are able to generate profits without transaction costs, their profits disappear quickly when the transaction costs are taken into consideration. Our conclusions were unchanged when we considered the non-synchronization problem or when we test various data frequency and different strategies.

This thesis aims to provide new insights into the importance of decomposing aggregate time series data using the Maximum Overlap Discrete Wavelet Transform. In particular, the analysis throughout this thesis involves decomposing aggregate financial time series data at hand into approximation (low-frequency) and detail (high-frequency) components. Following this, information and hidden relations can be extracted for different investment horizons, as matched with the detail components. The first study examines the ability of different GARCH models to forecast stock return volatility in eight international stock markets. The results demonstrate that de-noising the returns improves the accuracy of volatility forecasts regardless of the statistical test employed. After de-noising, the asymmetric GARCH approach tends to be preferred, although that result is not universal. Furthermore, wavelet de-noising is found to be more important at the key 99% Value-at-Risk level compared to the 95% level. The second study examines the impact of fourteen macroeconomic news announcements on the stock and bond return dynamic correlation in the U.S. from the day of the announcement up to sixteen days afterwards. Results conducted over the full sample offer very little evidence that macroeconomic news announcements affect the stock-bond return dynamic correlation. However, after controlling for the financial crisis of 2007-2008 several announcements become significant both on the announcement day and afterwards. Furthermore, the study observes that news released early in the day, i.e. before 12 pm, and in the first half of the month, exhibit a slower effect on the dynamic correlation than those released later in the month or later in the day. While several announcements exhibit significance in the 2008 crisis period, only CPI and Housing Starts show significant and consistent effects on the correlation outside the 2001, 2008 and 2011 crises periods. The final study investigates whether recent returns and the time-scaled return can predict the subsequent trading in ten stock markets. The study finds little evidence that recent returns do predict the subsequent trading, though this predictability is observed more over the long-run horizon. The study also finds a statistical relation between trading and return over the long-time investment horizons of [8-16] and [16-32] day periods. Yet, this relation is mostly a negative one, only being positive for developing countries. It also tends to be economically stronger during bull-periods.

本研究探討VIX 期貨價格所隱含的資訊對於S&P 500 指數波動度預測的解釋力。過去許多文獻主要運用線性預測模型探討歷史波動度、隱含波動度和風險中立偏態對於波動度預測的資訊內涵。然而過去研究顯示，波動度具有長期記憶與非線性的特性，因此本文主要研究非線性預測模型對於波動度預測的有效性。本篇論文特別著重在不同市場狀態下(高波動與低波動)的實現波動度及隱含波動度異質自我迴歸模型(HAR-RV-IV model)。因此，本研究以考慮馬可夫狀態轉化下的異質自我迴歸模型(MRS-HAR model)進行實證分析。 本研究主要目的有以下三點: (1) 以VIX期貨價格所隱含的資訊提升S&P 500波動度預測的準確性。(2) 結合風險中立偏態與VIX期貨的資訊內涵，進一步提升S&P 500 波動度預測的準確性。(3) 考慮狀態轉換後的波動度預測模型是否優於過去文獻的線性迴歸模型。 本研究實證結果發現: (1) 相對於過去的實現波動度及隱含波動度，VIX 期貨可以提供對於預測未來波動度的額外資訊。 (2) 與其他模型比較，加入風險中立偏態和VIX 期貨萃取出的隱含波動度之波動度預測模型，只顯著提高預測未來一天波動度的準確性。 (3) 考慮狀態轉換後的波動度預測模型優於線性迴歸模型。 / This paper explores whether the information implied from VIX futures prices has incremental explanatory power for future volatility in the S&P 500 index. Most of prior studies adopt linear forecasting models to investigate the usefulness of historical volatility, implied volatility and risk-neutral skewness for volatility forecasting. However, previous literatures find out the long-memory and nonlinear property in volatility. Therefore, this study focuses on the nonlinear forecasting models to examine the effectiveness for volatility forecasting. In particular, we concentrate on Heterogeneous Autoregressive model of Realized Volatility and Implied Volatility (HAR-RV-IV) under different market conditions (i.e., high and low volatility state). This study has three main goals: First, to investigate whether the information extracted from VIX futures prices could improve the accuracy for future volatility forecasting. Second, combining the information content of risk-neutral skewness and VIX futures to enhance the predictive power for future volatility forecasting. Last, to explore whether the nonlinear models are superior to the linear models. This study finds that VIX futures prices contain additional information for future volatility, relative to past realized volatilities and implied volatility. Out-of-sample analysis confirms that VIX futures improves significantly the accuracy for future volatility forecasting. However, the improvement in the accuracy of volatility forecasts is significant only at daily forecast horizon after incorporating the information of risk-neutral skewness and VIX futures prices into the volatility forecasting model. Last, the volatility forecasting models are superior after taking the regime-switching into account.

The thesis focuses on modelling and forecasting the exchange rate time series volatility. The basic approach used for the conditional variance modelling are class (G)ARCH models and their variations. Modelling of the conditional mean is based on the use of AR autoregressive models. Due to the breach of one of the basic assumption of the models (normality assumption), an important part of the work is a detailed analysis of unconditional distribution of returns enabling the selection of a suitable distributional assumption of error terms of (G)ARCH models. The use of leptokurtic distribution assumption leads to a major improvement of volatility forecasting compared to normal distribution. In regard to this fact, the often applied GED and the Student's t distributions represent the key-stones of this work. In addition, the less known distributions are applied in the work, e.g. the Johnson's SU and the normal Inverse Gaussian Distribution. To model volatility, a great number of linear and non-linear models have been tested. Linear models are represented by ARCH, GARCH, GARCH in mean, integrated GARCH, fractionally integrated GARCH and HYGARCH. In the event of the presence of the leverage effect, non-linear EGARCH, GJR-GARCH, APARCH and FIEGARCH models are applied. Using suitable models according to the selected criteria, volatility forecasts are made with different long-term and short-term forecasting horizons. Outcomes of traditional approaches using parametric models (G)ARCH are compared with semi-parametric neural networks based concepts that are widely applicable in clustering and also in time series prediction problems. In conclusion, a description is given of the coincident and different properties of the analyzed exchange rate time series. The author further summarized the models that provide the best forecasts of volatility behaviour of the selected time series, including recommendations for their modelling. Such models can be further used to measure market risk rate by the Value at Risk method or in future price estimating where future volatility is inevitable prerequisite for the interval forecasts.

This dissertation explores aspects of financial stability from three different perspectives. In the first essay, we empirically analyze to which extent popular global systemic risk measures (SRMs) yield comparable results with respect to the systemic importance of a financial institution and, in particular, from which determinants the degree of consistency of the classification by the various SRMs depends. It turns out that rank correlations, in general, are more sensitive towards macroeconomic factors such as the unemployment rate, and to a minor degree towards factors that can be interpreted in a broader sense as proxies for the stability of a bank such as the market-to-book ratio and the loans-to-deposits ratio. Further analyses reveal the inconsistency of systemic risk ranks and the difficulty to detect specific explanatory factors across several different settings. In the second essay, we assess the potential of activity on Twitter for improving forecasts of daily and intra-daily stock and index return volatilities. For this purpose, a unique high-frequency dataset of a comprehensive sample of more than 150 stocks of large international companies, systemically important banks, as well as several leading international stock indices is constructed. Our results show that there is no clear advantage of adding Twitter information by assessing the forecast performance of a plethora of different model specifications. We also reveal the necessity to consider different set-ups since they partly deliver opposing results. However, even though Twitter information is sometimes valuable, we find that forecast improvements in general remain marginal. In the third essay, we characterizes the formation of self-enforcing international financial regulation agreements. Our analysis allows evaluating the desirability and feasibility of cooperative solutions and explains the challenges associated with the process of cooperation. We model the cooperation of national financial regulators in a game-theoretical framework that considers financial stability to be an impure public good. Joint national supervisory effort is supposed to increase aggregate welfare in terms of a more stable financial system both on a global and on a local level by simultaneously generating incentives to free-ride. In our basic version of the model, we show that partial cooperation of two or three countries is stable and improves the welfare of all countries relative to the non-cooperative Nash equilibrium. Further analyses highlight the role of additional club benefits. When signatory countries of a coalition gain benefits over and above the joint welfare maximization, stable coalitions of any size become feasible.