Portfolio Analysis Of Post-financial Crisis Era

In September, 2008, the global financial crisis broke out in the United States with the bankruptcy of Lehman Brothers. About half year period, the global economy was in a terrified atmosphere and the global economy was straight down sharply. As most countries launched the bailout policies, the global economy gradually shifts into a stable situation. However, due to not solving the inherent crisis or impossible completely solving the crisis in a short period of time, there were still a lot of uncertainty and instability. This period was so called as the Post-crisis Era.The global economy was in turbulence by the attack of the financial crisis. Most institutional investors were confronted with the dilemmas of serious shrinking of assets or bankruptcy. Under this complicated international financial environment, each investor was concerned with suffering risk (loss), especially for the institutional investors. The investors pay more attention to the gains and the asset value in the future when confronted with the financial crisis or unexpected adverse events. They adjust the portfolio positions according to the change of the investment environment to avoid shrinking dramatically by diversification of assets. In this paper, we focus on how to establish an effective portfolio selection model to deal with the outburst adverse event and retain the asset value as well as prevent high loss.In the preface, we first introduce the development of modern portfolio theory, and then propose the contents, emphases, and innovations in this paper.In Chapter 1, we investigate some portfolio selection theories: mean-risk portfolio theory, safety-first portfolio theory and expect utility theory, and then compare the advantages and disadvantages of some famous models, such as: mean-variance model, mean-absolute deviation model, mean-lower-semi-variance model, mean-target model, value-at-risk model and condition value-at-risk model.In chapter 2, in view of this risky era, we propose a portfolio selection model based on Conditional Value-at-Risk, which considers the investor's concern about the future asset return. Numerical experiments are implemented to test the model's rationality and validity. It is verified that our model outperforms other portfolio models based on the scenarios sampled from multi-variable normal distribution.In chapter 3, we first introduce some traditional scenario generating methods, and then propose a new method based on the principal component analysis. Regardless of the probability distribution, this method will fully use the historical data to generate scenarios, and can guarantee that the generated scenarios has the same mean and variance with the original data. Also, this method retains most of the information with fewer principal components, which provides a valuable tool to generate scenarios when there are lots of risky assets.In chapter 4, we extend the portfolio model proposed in Chapter 2 considering the investment restrictions in capital market. Numerical experiments are implemented by two scenario generating methods, the principal component analysis and multivariable normal distribution. The results show that the method based on principal components analysis has a better description of the distribution characteristics of the asset return and the optimal portfolio generated by the extended portfolio model has a dominant efficient frontier than other models. Meanwhile, the out-of-samples test show that this extended portfolio model effectively avoids high risk and realizes the increment of asset value. And we also check that the method based on principal components analysis can ensure no arbitrage opportunities by empirical test.In chapter 5, we conclude our contributions and point out some limitations of the portfolio model proposed in the paper. Finally we give an outlook for further research.