CAJ | 학술논문

金融数据时间序列具有丛集性和方差波动性特点，传统经典计量模型对此的解释能力不足。ARCH模型引入观测数据方差自相关假设，有力地刻划和解释了金融数据的丛集性和厚尾尖锋特性。目前，国内用此模型对股指收益率、非对称性、市场有效性、量价关系、风险管理、保证金水平等问题研究取得了多项成果。但是，在研究中也存在样本容量小，容易导致实证结果不稳定、不可靠，对杠杆效应、周内效应、羊群效应形成原因和机理研究不足等问题。后续研究应该注重对超高频数据分析、波动的持续性、无条件分布的厚尾性及高维系统的分析。在参数估计方面，应针对具体市场和样本数据，检验各种模型和参数估计方法的能力。
금융수거시간서렬구유총집성화방차파동성특점，전통경전계량모형대차적해석능력불족。ARCH모형인입관측수거방차자상관가설，유력지각화화해석료금융수거적총집성화후미첨봉특성。목전，국내용차모형대고지수익솔、비대칭성、시장유효성、량개관계、풍험관리、보증금수평등문제연구취득료다항성과。단시，재연구중야존재양본용량소，용역도치실증결과불은정、불가고，대강간효응、주내효응、양군효응형성원인화궤리연구불족등문제。후속연구응해주중대초고빈수거분석、파동적지속성、무조건분포적후미성급고유계통적분석。재삼수고계방면，응침대구체시장화양본수거，검험각충모형화삼수고계방법적능력。
The time series of financial data have the characteristics of unevenness and variance volatility, which was inadequately explained by traditional classical econometric model. The ARCH model with its introduction of autocorrelation hypothesis of observation data describes and explains convincingly the unevenness and fat-tail distribution of financial data. At present, research achievements have been made at home in the application of this model in such issues as stock index returns, asymmetry, market efficiency, price volume relation, risk management, and margin level. However, the sample volumes in these studies are small, which leads to the instability and unreliability of the empirical results, and studies are not adequate concerning the formation causes and mechanisms of leverage effect, weekday effect and herding effect. Future research should focus on the analysis of ultra-high frequency data, volatility persistence, un- conditional fat-tail distribution, and high dimensional system. With regard to parameter estimation, the capacity of different models and parameter estimation methods should be tested against the specific market and sample data.