Filters
Availability
Publication Year
Resource Type
Faculty
School/Institute
Publication Search Results
Sort
Results per page
1 - 9 of 9
-
(2010)ThesisThis dissertation is primarily concerned with mixture models for high-dimensional financial data. New flexible mixture models are introduced and implemented with fast and effective optimization routines. The stochastic gradient approach uses random gradients to update the parameters of the mixture model improving the chance of the iterates converging to a higher mode. Chapter 2 provides the details of the stochastic gradient optimization routines used. Chapter 3 suggests two new multivariate density estimators, namely the marginal adaptation mixture of normals and the mixture of normals copula. Their performances are compared with a few recent popular models such as the skewed-t model. Chapter 4 discusses covariance estimation for high dimensional data. The aim of the chapter is to improve the estimation of covariance matrices by using mixture shrinkage priors. This chapter also shows how to apply the priors to the simultaneous estimation of several covariance matrices such as in the case of mixture of normals models. Chapters 5 and 6 consider the estimation or fitting of models to time series data, when the models may experience a small number of structural breaks. Chapter 5 looks at univariate data and Chapter 6 considers multivariate data. In particular, Chapter 6 shows how to estimate a Gaussian vector autoregressive model subject to occasional structural breaks using a mixture of experts framework.
-
(2010)ThesisThis thesis investigates the dynamics of trading behaviour and provides a comprehensive assessment of the overall performance of all Finnish retail and institutional investors over the years 1995 through 2004 using detailed transaction data on daily trades and investor identities. In contrast to prior published studies for Finland, this study reveals a significantly superior performance achieved by households over institutions in the longer run, either before or after the trading expenses. Despite much lower transaction costs, institutions trade several times as much as the most active household age group that reduces their net return as much as for households. I also find institutional trading is the key contributor to the volatility of Nokia share price. This study further examines the trading performance of individual investors across gender and age bands and documents both males and females in their thirties are the best performing traders among their respective gender groups on either a gross or net basis. In term of gender, male groups show significantly higher turnover while females hold higher-risk stocks. Frequent trading reduces men’s net return more so than do women in all age groups. The life cycle hypothesis does not seem to apply to Finnish working males nearing retirement after Nokia price crashed in 2000. This age group of 60 to 69 were net buyers of Nokia shares, and hence does not indicate a movement out of a high-risk investment into cash or bonds. I find strong evidence of correlated trading among household groups with young and middle-aged males sharing the highest cross correlations between their contemporaneous buy intensity. The level of correlated trading is surprisingly persistent and increases year by year. This finding suggests that profitable herding behaviour encourages more information sharing among households over time.
-
(2010)ThesisThis thesis studies four related topics in financial economics; realized volatility modelling and forecasting in the presence of model instability, forecasting stock return realized volatility at the quarterly frequency, quarterly realized beta measurement and beta neutrality evaluation under a popular long short strategy. Recent advances in financial econometrics have allowed for the construction of efficient post measures of daily volatility. The first topic investigates the importance of instability in models of realized volatility and their corresponding forecasts. Testing for model instability is conducted with a subsampling method. We show that removing structurally unstable data of a short duration has a negligible impact on the accuracy of conditional mean forecasts of volatility. In contrast, it does provide a substantial improvement in a model's forecast density of volatility. In addition, the forecasting performance improves, often dramatically, when we evaluate models on structurally stable data. The second topic is on forecasting stock return volatility at quarterly level. The last decade has seen substantial advances in the measurement, modeling and forecasting of volatility which has entered around the realized volatility literature. To date, most of the focus has been on the daily and monthly frequency, with little attention on longer horizons such as the quarterly frequency. In finance applications, forecasts of volatility at horizons such as quarterly are of fundamental importance to asset pricing and risk management. In this chapter we evaluate models for stock return volatility forecasting at the quarterly frequency. We find that an autoregressive model with one lag of quarterly realized volatility produces the most accurate forecasts, and dominates other approaches, such as the recently proposed mixed-data sampling (MIDAS) approach. Chen and Reeves (2009) introduced a new beta measurement technique via the Hodrick-Prescott filter and found it substantially reduced measurement error and produced much better performance than Fama-MacBeth measurement approach at the monthly frequency. The third chapter extends this technique to quarterly beta measurement. The finding in Chen and Reeves (2009) is also confirmed at the quarterly frequency. Hodrick-Prescott filtered beta contains the most relevant information and follows closely the true underlying beta. This result is also used in the final chapter to construct the proxy for the true underlying quarterly beta time series. The final topic is to investigate the economic value of realized beta. Market neutral funds are commonly advertised as alternative investments offering returns which are uncorrelated with the broad market. Utilizing recent advances in financial econometrics we demonstrate that constructing market neutral funds from monthly return data can be widely inaccurate. Given the monthly frequency is the most common for return measurement in the hedge fund industry, our findings highlight the need for higher frequency return data to be more commonly utilized. We demonstrate the use of daily returns to achieve a more market neutral portfolio, relative to the case of only using monthly returns.
-
(2010)ThesisWe examine three aspects of investors’ divergence of opinion in the context of takeovers. First we examine the effect of divergence of opinion and short sale constraints on merger announcement returns. To our knowledge, no study has analysed this combined effect for acquirers and targets. Using a sample of 565 Australian acquirers and 442 targets we find that relative spread and order imbalance (proxies for divergence of opinion), significantly affect acquirer returns. This provides information to investors wanting to assess the impact on bidder returns of an impending merger. The evidence of an impact on target returns is weak. Neither of our proxies for short sale constraints (level of institutional ownership or short selling ban), present a significant relationship with acquirer returns. The relationship with target returns is very weak. Second, we examine the effect of the market reaction to a takeover announcement in conjunction with divergence of opinion on deal success. To our knowledge, no study has analysed this joint effect. We apply logistic regressions using a sample of 774 Australian acquirers and 584 targets and find some evidence that managers in takeover negotiations listen to the acquirer market reaction to the announcement when deciding whether to complete a deal or withdraw from it. We do not find evidence of managers listening to the target market reaction. Additionally, we find that deal success has a significant positive relationship with relative spread and a significant negative relationship with order imbalance, variables which proxy for divergence of opinion. Third, we examine the impact of divergence of opinion on merger arbitrage investment strategies. To our knowledge, no study has analysed this mpact. We find that by using divergence of opinion to predict takeover success, we can build a merger arbitrage portfolio that earns higher Fama and French (1993) adjusted excess returns than a typical portfolio where arbitrageurs invest in all announced deals. However, the portfolio is small and its performance sensitive to the window used to calculate the returns and the divergence of opinion proxies. The models that provide the highest merger arbitrage returns use relative spread or order imbalance to proxy for divergence of opinion.
-
(2010)ThesisThis thesis consists of three related studies on the relationship between public information arrivals, stock price formation, and market efficiency in the Australian Securities Exchange (ASX). The common theme of these three studies lies in the investigation of a distinctive proxy of public information arrivals at the firm level, which is the intensity of time-stamped announcements of companies listed in the ASX. The first study examines how public information arrivals affect stock return volatility of individual firms. Due to data limitation associated with a flow of time-stamped news at the firm level, prior studies often focus on documenting this relationship at the daily frequency or with an event study. Taking advantage of the distinctive proxy for public information made available by the ASX, this study documents that a positive relationship exists between public information arrivals and intraday volatility of stock returns at the firm level. Moreover, it finds that the effects of public information arrivals on stock return volatility become weaker if the sampling frequencies change from the intraday interval to the daily interval, or if a stock is smaller or less liquid. Overall, this study highlights the importance role of the flow of time-stamped announcements of companies listed in the ASX as a source of public information that generates significant intraday stock return volatility at the firm level. The second study analyses how the arrivals of new public information influence the process by which trades and quotes move stock prices during the normal trading hours. Prior studies show that information contained in the limit order book is related to stock price changes. This study finds that public information arrivals increase the informativeness of the limit order book on short-term stock returns, and that public information arrivals also strengthen the relationship between trading volume and stock return volatility, and between the slope of the limit order book and stock return volatility. In sum, this study suggests that public information arrivals improve the informational content of the limit order book. The third study focuses on the pre-opening period in the ASX and examines the impact of order placement during this period on the final opening price under different information arrival conditions. The existing literature in this area has documented that despite there being no or little trading, the pre-opening period still facilitates price discovery. This study adds to the existing literature by documenting that public information arriving before or during the pre-opening period increases the efficiency of the indicative opening price. Moreover, it finds that market depth, order size, stock return volatility and company announcements are the determinants of order aggressiveness during the pre-opening period, highlighting the important role of public information in facilitating price discovery during this period.
-
(2010)ThesisHedge funds’ (HF) assets under management (AUM) expanded at more than 20% per annum between 2000 and the mid-2008 peak of USD1.85 trillion. The 2008 Global Financial Crisis (GFC) led to a drastic industry contraction with an estimated 2,000 fund liquidations. Nonetheless, the industry’s AUM is expected to reach new highs by 2011. Clearly, this asset class continues to grow at a remarkable pace, despite being marked by both high attrition rates and — as the industry’s recent experience made clear — dramatic changes over time. Given the continued growth and the important role of hedge funds, tools that facilitate the understanding of their survival probabilities are of general interest and would specifically benefit investors, service providers, managers and regulators. We use data on individual hedge funds and a range of models to analyse the association between fund specific and market variables with two types of database attrition, liquidation and non-liquidation exits. First, we explore the drivers of not only HF liquidation but also non-liquidation exits. Understanding the latter exit type is of interest to institutional investors who prefer to regularly allocate capital across their HF portfolio. Secondly, we use flexible econometric techniques that can, and indeed do, identify inflection points and turning points between the two database exit types and both fund specific variables and market ones. Liquidation risk is shown to increase immediately and drastically as HF returns move to negative territory and surprisingly, liquidation risk increases at the right tail of the return distribution and decreases at the far left. Finally, we explore the dynamics of two time scales, the lifetime scale and the calendar time scale. Importantly, we find the lifetime analysis requires the flexibility of fractional polynomial models as the time interactions of many variables were non-monotonic. For example, while positive returns remain negatively related to liquidation risk, they have the least impact during the time period of peak liquidation risk, roughly at the 5.5 year age mark. Similarly, we show that the impact of variables are dynamic over the 14 year sample time period, and in particular during a number of the structural breaks, the relationship between specific market related variables and HF liquidation risk reverses.
-
(2010)Thesis“Time incongruency” occurs when there is a mismatch between the return period used to assess investment choices and an investor’s true investment horizon. Using both theoretical and empirical techniques, I demonstrate that time incongruency has a substantial impact on both the assessment of non-systematic and systematic risk, which can cause investors to make sub-optimal investment choices. I derive methods to mitigate this impact. With non-systematic risk, the impact can emerge via the use of methods of analysis that are inappropriate for the investment horizon. This can cause investors to reach inappropriate conclusions about risk diversification opportunities, affecting their portfolio choices. For example, a long term investor may employ short run risk assessments in the form of pairwise correlation analysis, although long run risk assessments, including cointegration analysis, would be more appropriate. Using data from 12 Asia Pacific equity market indices drawn from a 23-year interval, I find many instances when such an investor would draw different conclusions about risk diversification opportunities under each method, suggesting that she could make sub-optimal investment choices. In the case of systematic risk, time incongruency introduces a bias into the beta calculation. I show that this beta bias is a non-symmetrical function of the time congruent beta and the degree of time incongruency. As a result, this bias cannot be systematically diversified away across a portfolio. I derive, test and refine a solution, the “Adjusted Time Incongruency Beta Bias Adjustment (TIBBA) Model”, to mitigate this bias. I evaluate the benefits of this strategy using data drawn from four Asia Pacific markets over a 15-year interval. Although most investors will not always have certainty around their true investment horizon, most will have a fairly clear expectation about their investment timeframe. By using this expected investment horizon to choose the most appropriate method of analysing non-systematic risk (correlation for the short run, or cointegration for the long run), they can reduce the impact of time incongruency on their risk diversification choices. Similarly, by using their expected investment horizon and the Adjusted TIBBA Model, they can choose appropriate asset allocation strategies to achieve their desired level of systematic risk across their portfolio.
-
(2010)ThesisThis dissertation examines the identity of new controlling shareholders in partial corporate control transactions and its influence on firm performance and corporate policies in an international context. In a transfer of partial corporate control, the identity of controlling shareholder changs, thereby facilitating an event study of corporate changes resulted from controlling shareholder turnover. The dissertation comprises of three empirical research projects to address two questions: Firstly, does the identity of a new controlling shareholder in a partial control transfer matter to firm value? Secondly, how does new controlling shareholder identity explain the differences of firm performance and corporate policies in the long run? The equity block transactions in listed firms from around the world announced between 1996 and 2005 are filtered to arrive at the final sample of 215 corporate control transfers through private negotiation. This hand-collected dataset allows the dissertation to contribute to the existing literature on ownership concentration by introducing the identity of controlling shareholder into the theoretical framework and investigating its significance in an international context. It is argued that firm value, long-term performance and corporate policies can be influenced by the identity of new controlling shareholders because different controlling shareholders have distinct incentives, skills and styles. This dissertation finds that individual investor controlled firms outperform those controlled by corporate investors in both short-term abnormal returns and long-term performance after a control transfer; and that the sample firms controlled by individual and corporate investors adopt different policies of investment and financial leverage. The evidence presented here shows that individual controlling shareholders are better motivated to monitor managers and improve operational efficiency. Partial corporate control activities have important governance effects and controlling shareholder heterogeneity is a significant contributing factor to firm performance and decision making.
-
(2010)ThesisFinancial markets worldwide have grown rapidly over the last few decades and so have the number of modelling approaches to analyse and price financial assets. With the emergence of more complex models, emphasis has also been placed on calibrating the models to market data. This thesis consists of three distinct components which investigate the estimation of financial models and the pricing of derivatives. The focus of these studies is stochastic volatility models and the crude oil futures market. The first component investigates discrete-time stochastic volatility models by comparing the estimation performance of three maximum-likelihood procedures. The analysis is conducted empirically on the fixed income and crude oil markets and also tests the validity of different stochastic volatility model specifications in-sample and out-of-sample. The study finds that the choice of estimation procedure is important if conditional volatility estimates are required. Also, a traditional AR(1) specification for the log-variance is sufficient in the fixed income and crude oil markets. The second component introduces a three-factor short/long factor commodity model which allows for mean-reversion in spot prices, expected increases in long-term prices and a time-varying market price of risk. The model is able to accurately capture the term structure of futures prices in the crude oil futures market with evidence suggesting that risk premiums are time-varying. Using the cross-section of futures prices we estimate a time-series of the market price of risk implied by the model. We find that the risk premiums in the crude oil market are driven by the same risk factors as equity and bond markets. In the final component, the short/long factor model is extended to incorporate both jumps and stochastic volatility. Semi-analytical solutions of futures and European option prices are derived for the model. The futures and option pricing performance is compared with nested specifications. The empirical results demonstrate that although introducing jumps or stochastic volatility does not impact futures pricing much, they are required for option pricing applications. When fitting the implied volatility surface of crude oil futures options, stochastic volatility is required to fit implied volatility over the maturity and moneyness dimensions but jumps are required when fitting the steep volatility smiles exhibited by short-term options.