This thesis examines how stock returns are determined by different ex ante risk factors implied from options; these ex ante risk factors include option-implied betas, the variance, skew and kurtosis risk premia. I first compare different option-implied beta measures in future stock return prediction on the basis of Buss and Vilkov (2012). The option-implied beta proposed by Buss and Vilkov (2012) (BV) is found to outperform other beta approaches included in the research. I also propose the implied downside betas and find that the BV implied downside beta performs best and offers an improvement over the BV implied beta. However, the relationship between option-implied or implied downside betas and stock returns is not robust to firm-level variables such as firm size, book-to-market ratio or option-implied moments. These variables are correlated with option-implied betas and implied downside betas, which may obscure the beta-return relationship. Next, I investigate comprehensively whether the moment risk premia are able to predict the cross-section of stock returns. Cross-sectionally, I find that the variance, skew and kurtosis risk premia are determined differently by firm-level and risk factors. I also find that the moment risk premia have different effects on stock returns. For ex post realised stock returns, there is a negative relationship with both the variance and skew risk premia. However, the kurtosis risk premium has a noisy and insignificant relationship with realised stock returns. The price target expected return (PTER) and the implied cost of capital (ICC) are adopted as proxies for ex ante expected stock returns. I demonstrate that there is a significantly negative relationship between the variance and skew risk premia and expected stock returns, while there is a significantly positive relationship between the kurtosis risk premium and expected stock returns. The results are robust to firm-level and risk factors, sub-periods and different maturities. 3 Finally, I investigate whether the moment risk premia are able to explain future index returns at the aggregate stock market level; they are found to have different impacts on index returns depending on the return measure. Both the variance and skew risk premia are inversely related to subsequent realised S&P 500 index returns; however, the variance risk premium has a stronger relationship than the skew risk premium. The kurtosis risk premium has no effect on realised index returns. For the index price target expected return (PTER), neither the variance risk premium nor the skew risk premium has explanatory power with the PTER, while the kurtosis risk premium has a robust and positive relationship with the PTER. For the index implied cost of capital (ICC), both the variance and skew risk premia are significantly and positively related to the ICC, while the kurtosis risk premium has a significantly negative relationship with the ICC. However, the relationships between the moment risk premia and the ICC are not robust to macroeconomic variables. I also find that both the PTER and the ICC can be explained by macroeconomic factors.

No description available.

The purpose of this thesis is to investigate trading strategies based on futures contracts. The first chapter demonstrates and analyzes the exceptional performance of both carry and momentum strategies in future markets across asset classes (commodities, bonds, equities, and currencies). Individual carry and momentum returns have low correlation, generating a significant diversification benefit in the combined portfolio and a Sharpe ratio of 1.4. Individually and combined, carry and momentum strategies have significant returns not explained by the CAPM or risk factor models. However, carry returns disappear after adjusting for lagged macroeconomic variables, suggesting performance is related to business cycle risk. Expected momentum returns are only weakly related to macroeconomic variables, but co-vary significantly with hedge fund capital flow - indicating returns are related to limits to arbitrage constraints of hedge funds. The second chapter establishes the economic significance of carry and momentum trading signals. We use a model incorporating a time varying investment opportunity set into a parametric portfolio framework and derive optimal portfolio parameters. Without any ex-ante imposed relation, in-sample portfolio parameters are found to be consistent with the results of the first chapter. Furthermore, out-of-sample returns are found to be highly significant, robust to transaction costs and not compensation for traditional risk exposure, time-varying risk due to macroeconomic cycles, or funding liquidity. Out-of-sample returns are significantly related to pro-cyclical hedge fund capital flows, suggesting expected returns decrease with speculative capital. The third chapter applies our parametric portfolio framework to assess the economic significance of predictors important in commodity markets since 2001. The studied predictors are widened to include hedging pressure and three market wide predictors found in the literature to forecast returns. In contrast to our results for the whole futures market, we find little evidence for economically significant commodity strategy returns for either individual or combined predictors.

This thesis is a theoretical and empirical analysis of asset price movement including during periods characterised by financial bubbles. It can be argued that financial bubbles occur due to excessive optimism on the part of speculative investors. The positive expectations of investors encourage increases in both price and trading volume. When prices subsequently falter exodus from the market ensues resulting in both a price and trading volume crash. A key question is why do bubbles emerge and grow and subsequently burst? One answer to this question may be found through an analysis of how beliefs are formulated. In the theoretical component of this thesis (Chapter 2) by applying the feedback modelling approach with the coordination game of Ozcenoren and Yuan (2008) we model how investor beliefs are formulated. In Chapter 3 we use transaction-level data to investigate market illiquidity on the London Stock Exchange over the period 1996-2009. The time period under investigation encompasses the Internet (Dot-com) bubble (1997-2000) and house price bubble (2007-2008). Our dataset covers 1,600 stocks and more than 528 million trades. Chapter 4 present the second empirical investigation and considers whether spreads on the London Stock Exchange have become increasingly right skewed.

This thesis empirically studies asset pricing in the Chinese stock market. It consists of four separate but closely related essays in Chapters 2-5, respectively. The first essay tests the validity of the standard Fama-French model in the Chinese Main-board stock market within a relatively new sample period since quarterly financial reports became widely available. The empirical results confirm the good performance of the Fama-French model. By splitting the sample into two sub-periods representing two different market tendencies, the Fama-French model performs better but is less stable in the downward trend than the fluctuating market. The potential explanation is also provided. The complexity of traditional methods to construct the Fama-French factors limits the model's popularity. The second essay examines a simplified method to construct the factors using the Chinese style indices. The results show that the style index-formed factors lead the Fama-French model to better explaining equity returns and producing less pricing error. The finding is meaningful in the sense that it dramatically lowers the barrier to examining and applying the Fama-French model in both academic research and financial practice. However, the theoretical foundation behind the Fama-French model, i.e., which macroeconomic variables fundamentally determine the model's success, is not clear. The rest two essays make efforts to find the answer. The third essay examines the extent to which the market, size, value and momentum factors can be linked to fundamental macroeconomic variables. We find that the momentum factor consistently captures the current economic risk as well as predict future GDP growth in China. In contrast, there is no consistent relationship between the market, size and value factors and macroeconomic growth. This finding contradicts Liew and Vassalou's (2000) conclusion that the Fama-French factors can generally predict future real GDP growth in ten developed countries. The fourth essay examines whether the ability of the above four factors to explain the cross-sectional variation in equity returns, if any, stems from the fact that they are proxies for fundamental macroeconomic risk associated with future GDP growth. The results show that the market, size and value factors together contain some overlapping innovative information about future real GDP growth, and thus, reconcile previous findings in the US (Vassalou, 2003) and Australian (Nguyen et al., 2009) markets. Some possible explanations are also discussed.

The thesis extends the research in the area of momentum strategies by investigating the short-term continuation for stocks listed in the United Arab Emirates (U.A.E.) Stock Market over the period from January 2001 to June 2006. The evidence shows that winner portfolios tend to outperform loser portfolios of stocks over pre- and post-formation periods of three months to twelve months. The most successful zero-cost trading strategy selects stocks based on their returns over the previous six months and then holds the portfolio for eight months. This strategy yields abnormal returns of 1.10 percent per month, which is very close to the profits reported by Jegadeesh and Titman (1993) in the US market. The thesis continues by looking at possible explanations of momentum profits by investigating whether they can be explained by the firm size effect or book-to-market effect. The empirical results provide evidence that small-stocks exhibit a greater return than big-stocks over various holding periods, but that the difference between high B/M-stocks and low B/M-stocks is not as effective in producing abnormal returns. In order to achieve a deeper understanding of the linkages between these variables and momentum profits, I propose a multiple model of risk valuation that extends both the CAPM and the Fama and French (1992, 1993) models, by introducing a new model that assumes that momentum is explained by the sensitivity of stock returns to four-factors; market beta, firm size and book-to-market, and in addition to the oil price factor. The evidence suggests that the relationship between momentum return and the market risk and book-to-market factors is insignificant, which means that these factors are unable to explain the performance of the momentum returns, while it is positively correlated with the firm size factor and the changes in the oil price factor. These findings motivate taking a closer look at the causes behind the momentum returns. A survey questionnaire is carried out to acquire more knowledge of the momentum effect, and to identify further possible explanations of momentum in stock returns. The results from the survey questionnaire reveal that investors’ decision-making appears to be influenced by a number of factors other than fundamental factors, such as recent price movements in a stock, market rumors and friends/family opinions. The questionnaire results lead to additional insights into the causes of the momentum phenomenon.

In recent years there has been a surge in interest in two branches of research. The first is the analysis of cross‐market linkages, arising from portfolio diversification analysis, in order to measure integration between countries. The second larger branch of research is the oil price effect, which has its roots in the effect of oil prices on economic activity. Specifically the impact of oil prices on stock market behaviour which, in turn, is regarded as being representative of economic activity. This thesis attempts to measure the effect of the oil prices on cross‐market linkages between stock markets. The analysis is performed utilising the Cheung and Ng (1996) causality‐in‐mean/variance test where the effect of the oil price on cross‐market linkages is estimated as the percentage change of the magnitude of the CCF coefficient after adjustment for the oil price. Daily data nonsynchronisation correction and the application of the rolling window method as a stability test contribute to the reliability of the estimation outcomes. Recognising the significance of oil prices for the development of the world economy, the research is focused on BRIC and mature market economies providing a mix of countries regarding their economic profile(emerging/developed) and oil‐status (net oil imported/self‐sufficient/exporter). A complementary effect of oil prices on cross‐market linkages was found for the pairs of countries with similar oil status and economic profile. However, the outcome for mixed pairs does not demonstrate a prominent trend. In general, the effect of the oil price on cross‐market linkages is found to be smaller in contemporaneous terms, while higher lag dependencies are described by a higher percentage change in the magnitude of cross‐market dependencies. The research demonstrates that the application of oil price adjustment, whilst exploring crossmarket linkages, may be an appropriate technique for determining the level of integration between pairs of countries where the sign of the percentage change is confirmed as stable.

This thesis consists of an introductory chapter, three main chapters, and a concluding chapter. In Chapter 2, which was nominated for an EFMA 2014 Best Paper Award, I provide a novel risk-based explanation for the profitability of global momentum strategies. I show that the performance of past winners and losers is asymmetric in states of the global market upturns and downturns. Winners have higher downside market betas and lower upside market betas than losers, and hence their risks are more asymmetric. The winner-minus-loser (WML) momentum portfolios are exposed to the downside market risk, but serve as a hedge against the upside market risk. The high returns of the WML portfolios compensate investors for their high risk asymmetry. After controlling for this risk asymmetry, the momentum portfolios do not yield significant abnormal returns, and the momentum factor becomes insignificant in the cross-section. The two-beta CAPM with downside risk explains the cross-section of returns to global momentum portfolios well. In the third chapter, published in the Review of Finance and the winner of EFMA 2013 John Doukas Best Paper Award, I propose a new factor – the global downside market factor – to explain high returns to carry trades. I show that carry trades have high downside market risk, i.e. they crash systematically in the worst states of the world when the global stock market plunges or when a disaster occurs. The downside market factor explains the returns to currency portfolios sorted by the forward discount better than other factors previously proposed in the literature. GMM estimates of the downside beta premium are similar in the currency and stock markets, statistically significant and close to their theoretical value. I show that the high returns to carry trades are fair compensation for their high downside market risk. In the fourth chapter, I study whether or not countries‟ macroeconomic characteristics are systematically related to the downside market risk of their currencies. I find that the downside risk is strongly associated with the local inflation rate, real interest rate and net foreign asset 5 position. Currencies of countries with higher inflation and real interest rates and lower (negative) net foreign asset position (debtor countries) are more exposed to the downside risk whereas currencies of countries with low inflation and real interest rates and positive net foreign asset position (creditor countries) exhibit „safe haven‟ properties. Since inflation and real interest rates determine nominal interest rates which determinecurrency returns which, in turn, determine capital flows and net foreign asset positions, these macroeconomic variables are related. But the local real interest rate has the highest explanatory power in accounting for the cross-section of currency exposure to the downside risk. This suggests that the direction of currency trading is the reason why some currencies are exposed to the downside risk more than others. High currency downside risk is a consequence of investments in high-yield risky currencies and flight from them in „hard times‟. Currencies of low-yield creditor countries, on the contrary, provide a hedge in „hard times‟ because capital flies back to them. Currency exposure to the downside market risk has increased significantly in the 2000s when the volume of currency trading by institutional investors increased.

The Chinese government has implemented a series of financial liberalisation policies in stock market after China’s accession into the WTO in 2001. However, it remains somewhat unclear whether and to what extent China’s financial liberalisation has influenced its stock market in the post-WTO accession period. This thesis examines this issue from the perspectives of international market integration and market interdependence. This thesis mainly includes three empirical studies: 1) gauging the degree of market integration by the weak-form measure, 2) measuring the degree of market interdependence by the multi-factor R-squared measure, as well as 3) examining the relationship between these two elements by the MWALD causality test and cointegration analysis. The first empirical study finds that China’s financial liberalisation might have largely decreased the stock market segmentation between China and the world in the period from July 2003 to June 2007, but this process has been interrupted by the global financial crisis in 2008. Meanwhile, only some of China’s financial liberalisation policies might be regarded as effective, such as the QFII programme, the first round of exchange rate reform and the QDII programme. The second empirical study regards that the stock market interdependence between China and the world has experienced three stages. These stages are: 1) the stationary stage from July 2001 to June 2003; 2) the increasing and steady stage from July 2003 to June 2006, which might be thanks to China’s financial liberalisation; as well as 3) an inverted U-shaped stage from July 2006 to December 2009, which might be attributed to China’s economy overheating in 2007 and to its economic stimulus plan in 2008. This third empirical study finds that there has been a unidirectional causal relationship from market integration to market interdependence, and a cointegration relationship between them. Market integration and market interdependence are found to be highly connected but different issues. Market integration is one, but not the only, determinant of market interdependence.

This thesis investigated stock market disequilibrium focusing on two topics: the impact of multiple market makers on the market disequilibrium at the market microstructure level, and the detection of the long-run market disequilibrium in the context of bubbles and the changes in transition probabilities. The multiple market makers increased the resilience of price rather than improving its efficiency when a multiple market maker system (the NASDAQ) was compared with a single market maker system (the NYSE) in terms of lowering non-stationarity and raising predictability. On the other hand, the volatility modelling of intraday data showed that market maker’s under-estimation (higher-than-estimated size of return) increased volatility while over-estimation decreased it. Also, intraday seasonality in mean and volatility was confirmed, but leverage effects were denied in the GJR-GARCH-type models. The evidence of price bubbles in the Indian markets (1987-2008) and positive duration dependence in negative runs in the Korean market (1990-2008) were revealed using the duration dependence tests. The unconditional transition probabilities that a positive or negative run continues or ends were mostly significantly different from 0.5. On the other hand, the structural break based duration dependence test was devised to detect the changes in the transition probabilities between the market (dis)equilibrium. The NASDAQ and the Indian market showed positive duration dependence in positive runs and the Korean market displayed it in negative runs. In other words, the transition probability in those markets increases as a price run between structural breaks lengthened.