Uncovering Long-Run Economic Relationships in High-Frequency Financial Data

揭示高频金融数据中的长期经济关系

• 批准号: 9730440
• 负责人: Tim Bollerslev
• 金额: $ 24.21万
• 依托单位: National Bureau of Economic Research Inc
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1998
• 资助国家: 美国
• 起止时间: 1998-05-01 至 2003-04-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9730440&HistoricalAwards=false
• 关键词: Uncovering; Long; Run; Economic; Relationships

9730440 Bollerslev This project enhances the ability of researchers to extract useful information about important economic phenomena from high-frequency financial data. The increased availability of intraday financial data has spurred a tremendous growth in empirical research on the cross-sectional and intertemporal dependencies across financial instruments and markets. However, only recently has it become clear that the analysis of such high-frequency data necessitates the introduction of novel modeling paradigms relative to the standard methodologies employed at the daily and lower frequencies. The reliance on conventional time series techniques invariably results in the intraday-specific features dominating the inferred statistical properties of the observed high-frequency data, whereas any evidence concerning longer-run dependencies tend to be distorted or disappear altogether. These facts have effectively divorced the empirical studies of high-frequency data from the remainder the literature, and a near dichotomy has arisen between so-called microstructure studies and the more mainstream empirical literature in asset pricing finance. This project bridges this gap by developing new econometric time series models and tools and using these to improve our understanding of the type of information that induces dramatic price responses, and the forces shaping the intensity of price movements and their relation to underlying market microstructure. The empirical research is based on a long time span of intradaily exchange and interest rates. It provides a direct window to the market expectations and the importance of macroeconomic announcements across different policy regimes. The project also develops new and more accurate risk measurements based on high-frequency data and these should have immediate practical implications for the management and monitoring of financial market risks. The general results of the project should be of relevant to applied macroeconomists, time series econo metricians, financial researchers, regulators, and practitioners alike. ??

9730440 Bollerslev这个项目提高了研究者从高频金融数据中提取重要经济现象有用信息的能力。日内金融数据可用性的增加刺激了对金融工具和市场的横断面和跨期依赖性的实证研究的巨大增长。然而，直到最近才清楚地认识到，对这种高频数据的分析需要引入相对于日常和较低频率所采用的标准方法的新的建模范例。对传统时间序列技术的依赖必然导致当天特定特征主导了观测到的高频数据的推断统计特性，而任何有关长期依赖关系的证据往往被扭曲或完全消失。这些事实有效地将高频数据的实证研究与其他文献分离开来，在所谓的微观结构研究和资产定价金融领域更主流的实证文献之间出现了一种近乎二分的现象。本项目通过开发新的计量经济学时间序列模型和工具来弥合这一差距，并利用这些模型和工具来提高我们对引起剧烈价格反应的信息类型的理解，以及塑造价格运动强度的力量及其与潜在市场微观结构的关系。实证研究基于较长时间跨度的日内汇率和利率。它为了解市场预期和不同政策体制下宏观经济公告的重要性提供了一个直接窗口。该项目还开发了基于高频数据的新的、更准确的风险衡量方法，这些方法将对金融市场风险的管理和监测产生直接的实际影响。该项目的一般结果应与应用宏观经济学家、时间序列经济学家、金融研究人员、监管机构和从业人员等相关。？?