Arch, Cointegration and Common Features: Theory and Application

这是一个基于成就的项目更新，开发ARCH和相关的模型和协整测试。ARCH和协整模型现在几乎在经济学的每个领域都使用时间序列数据来检验不同经济变量之间的经济关系。这些研究通常使用单变量时间序列模型或最多包含两个或三个不同时间序列的多变量模型。这是大多数经济应用程序的主要缺点。这项资助将允许研究者继续在先前的资助下开始的关于多元ARCH和协整分析的研究。此外，该项目开发并应用一种称为共同特征的新统计程序来解决经济和金融问题。1992年4月将在加州拉霍亚举行一次会议，讨论波动模型的新发展和对金融的应用。经济时间序列具有许多显著的特征。一般来说，它们表现出一系列的相关性、趋势、季节性、异方差、偏态、峰度和各种其他特征。为了检测数据集中的每个这些特征，可以使用各种测试，其中每个测试都将所讨论的特定特征作为数据中不存在特征的零假设的替代方案。在之前的NSF资助下，研究者开发了一种新的统计程序，称为共同特征，允许分析人员确定两个或多个数据集是否具有相同的显著特征。这一程序是用来表明，主要工业国家有共同的国际商业周期的经验证据。这一程序将得到扩展和推广。该指数将被用来检验美国各部门的产出，以观察各部门在商业周期中是否会同步变动。该程序将用于确定美国境内的地区是否会一起移动。将分析资本市场的国际数据，以确定几个国家的股票市场的共同特征。这项研究应该为国际资本市场波动的本质提供新的见解。

This is an accomplishment based renewal of a project that developed ARCH and related models and tests for cointegration. Models of ARCH and cointegration are now used in almost every area of economics to test for economic relationships among different economic variables using time series data. These studies typically use univariate time series models or multivariate models with at most two or three different time series. This is a major drawback for most economic applications. This grant will permit the investigator to continue the research started under the previous grant on multivariate ARCH and cointegration analysis. In addition the project develops and applies a new statistical procedure called common features to problems in economics and finance. A conference will be held in April, 1992 in La Jolla, California on new developments in volatility models and applications to finance. Economic time series have many distinctive characteristics. Generally, they exhibit serial correlation, trends, seasonality, often heteroskedasticity, skewness, kurtosis, and various other features. In order to detect each of these features in a data set, a variety of tests are available each of which takes the particular feature in question as the alternative to the null hypothesis that the feature is not present in the data. Under its previous NSF grant, the investigator developed a new statistical procedure called common features that permits the analyst to determine if two or more data sets share the same distinctive characteristics. This procedure was used to show that there was empirical evidence of a common international business cycle for the major industrial countries. The procedure will be extended and generalized. It will be used to examine sectoral output in the U.S. to see whether sectors move together over the business cycle. The procedure will be used to determine whether regions within the U.S. move together. International data on capital markets will be analyzed to determine what features are common to the equity markets for several blocks of countries. This research should provide new insights into the nature of the volatility in international capital markets.