风险管理与收益预测
Risk Management and Return Prediction
摘要：
随着金融业的良好发展，市场开始吸引人们的眼球，不仅是从债券、股票到期货、期权的多元化投资选择，而且是普遍的“高风险、高回报”的心态促使人们将资金投入金融市场。人们对在给定的回报水平上降低风险感兴趣，因为没有办法既有高回报又有低风险。许多研究者一直在研究这个问题，其中最具开拓性的是1952年哈里·马科维茨（Harry Markowitz）提出的现代投资组合理论，它是投资组合管理的基石，旨在“在给定的风险下实现收益最大化”。50年后，E. Robert Fernholz的随机投资组合理论，与作为早期现代投资组合理论基础的规范性假设相反，与实际投资组合和市场的可观察特征相一致。本文在介绍了Markowitz的MPT和Fernholz的SPT的基本理论之后，进一步从应用的角度出发，尝试在Markowitz有效前沿的四个基本模型（Markowitz模型、常数相关模型、单指数模型和多因子模型）下，选择哪些投资组合，以及这些投资组合在现实世界中的表现如何。这里我们也涉及到通用的投资组合算法。封面选择投资组合作为比较。此外，将评估每个投资组合的风险价值、预期损失和相应的风险管理自举置信区间。最后，利用因子分析与时间序列模型，预测四种模型未来的绩效。
英文摘要：
With the good development in the financial industry, the market starts to catch people's eyes, not only by the diversified investing choices ranging from bonds and stocks to futures and options but also by the general "high-risk, high-reward" mindset prompting people to put money in the financial market. People are interested in reducing risk at a given level of return since there is no way of having both high returns and low risk. Many researchers have been studying this issue, and the most pioneering one is Harry Markowitz's Modern Portfolio Theory developed in 1952, which is the cornerstone of investment portfolio management and aims at "maximum the return at the given risk". In contrast to that, fifty years later, E. Robert Fernholz's Stochastic Portfolio Theory, as opposed to the normative assumption served as the basis of earlier modern portfolio theory, is consistent with the observable characteristics of actual portfolios and markets. In this paper, after introducing some basic theories of Markowitz's MPT and Fernholz's SPT, then we step across to the application side, trying to figure out under four basic models based on Markowitz Efficient Frontier, including Markowitz Model, Constant Correlation Model, Single Index Model, and Multi-Factor Model, which portfolios will be selected and how do these portfolios perform in the real world. Here we also involve universal Portfolio Algorithmby Thomas M. Cover to select portfolios as a comparison. Besides, each portfolio value at Risk, Expected Shortfall, and corresponding bootstrap confidence interval for risk management will be evaluated. Finally, by utilizing factor analysis and time series models, we could predict the future performance of our four models.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
