Estimation of Jump-Tails: Theory and Applications

最近金融市场的动荡突显了更好地了解罕见灾难或所谓的“尾部”事件的可能发生及其在不同资产和市场中的传播的重要性。拟议研究活动的主要主题涉及基于高频日内金融数据对此类事件的可靠计量经济学估计。智力优势：高频日内资产价格的可获得性促使大量快速增长的文献关注这一潜在丰富的新数据来源的分析。该项目旨在通过发展新的计量经济学程序及其经验应用，进一步扩大我们从这些数据中提取有关重要经济现象的有用信息的能力。具体地说，结合了传统极值理论(EVT)的见解，该理论是在罕见事件独立发生的假设下在精算科学的背景下发展起来的，与所谓的已实现变化度量相结合，旨在解释金融市场中经验上现实的波动聚集，作为提案的一部分，我们寻求开发的新计量经济学估计器有望为极端“尾部”事件提供比目前任何可用的方法更准确的估计。广泛影响：拟议的研究计划对经济和金融数据的理论和实证分析具有许多重要的宽泛定义的影响，应该会对理论导向型计量经济学家感兴趣。以及对“尾部”事件的估计感兴趣的应用宏观经济学家、金融研究人员和监管者。特别是，最重要、最难管理的金融市场风险总是与罕见事件有关。因此，更准确地衡量并可能预测“尾部”的能力，意味着改进风险管理程序的前景，这些程序更适合于控制较大的风险，撇开较小的“连续”价格波动不谈。通过加强我们对导致金融资产价格大幅波动或“跳跃”的经济“新闻”类型的理解，新程序的实证实施也将有助于揭示资产市场和实体经济之间的根本联系。投资者缺乏信心和对“尾部”事件的恐惧，往往被认为是2008年秋季金融危机爆发时市值大幅缩水的主要罪魁祸首之一，而罕见的灾难可能有助于解释明显的错误定价的想法，促使最近出现了一篇迅速增长的文献。该文献中提出的关键论点通常取决于严重事件的概率超过样本中具体化的事件的概率，或校准以反映异常广泛的资产和/或国家的概率。相反，我们建议的计量经济学方法有望基于实际观察到的高频数据，可靠地估计给定市场可能发生的“尾部”事件，而不需要求助于“比索”式的解释。研究结果将在研讨会和会议上广泛传播。该项目还将寻求通过让研究生和本科生参与拟议的研究活动来整合研究和教育。所有研究成果，包括新的计算机程序和数据库，都将很容易地通过网络提供。

The recent turmoil in financial markets has highlighted the importance of obtaining a better understanding of the likely occurrence of rare disasters, or so-called "tail" events, and their transmission across different assets and markets. The main theme of the proposed research activities relates to the reliable econometric estimation of such events based high-frequency intraday financial data.Intellectual Merit: The availability of high-frequency intraday asset prices has spurred a large and rapidly growing literature concerned with the analysis of this potentially rich new data source. This project aims to further expand on our ability to extract useful information about important economic phenomena from such data through the development of new econometric procedures and empirical applications thereof. Specifically, combining the insights from traditional Extreme Value Theory (EVT), developed in the context of actuary science under the assumption of independent occurrence of rare events, with so-called realized variation measures, designed to account for empirically realistic volatility clustering in financial markets, the new econometric estimators that we seek to develop as part of the proposal hold the promise of delivering much more accurate estimates for extreme "tail" events than any currently available procedures.Broader Impact: The proposed research program has a number of important broadly defined implications for the theoretical and empirical analyzes of economic and financial data, and should be of interests to theoretically oriented econometricians, as well as applied macroeconomists and financial researchers and regulators alike with an interest in the estimation of "tail" events. In particular, the most important and difficult to manage financial market risks are invariably associated with rare events. Hence, the ability to more accurately measure and possibly forecast the "tails," holds the promise of improved risk management procedures that are better geared toward controlling large risks, leaving aside the smaller "continuous" price moves. By enhancing our understanding of the types of economic "news" that induce large price moves, or "jumps," in financial asset prices, the empirical implementations of the new procedures will also help shed new light on the fundamental linkages between asset markets and the real economy. The lack of investor confidence and fear of "tail" events are often singled out as one of the main culprits behind the massive losses in market values in the advent of the Fall 2008 financial crises, and the idea that rare disasters may help explain apparent mis-pricing has spurred a rapidly growing recent literature. The key arguments put forth in that literature typically hinge on probabilities of severe events that exceed those materialized in-sample, or probabilities calibrated to reflect an unusually broad set of assets and/or countries. Instead, our proposed econometric procedures hold the promise of reliable estimating the likely occurrence of "tail" events in a given market based on actually observed high-frequency data, without resorting to "peso" type explanations.The research results will be disseminated broadly at seminars and conferences. The project will also seek to integrate research and education by involving both graduate and undergraduate students in the proposed research activities. All of the research results, including new computer programs and databases, will be made easily available through the web.