基于区间数据的数字货币市场风险测算新方法

An Innovative Approach for Risk Measurement in Digital Currency Market Using Interval-valued Data

在数字经济时代,数字货币的出现吸引了诸多投资者与研究者,但其高波动性的价格特征,为投资决策和风险评估提出了新的挑战.为更准确地刻画这种特征,本文提出基于指数衰减加权自举法的区间变量置信域构建方法,进一步以此置信域的覆盖面积与尾部分位数作为评估数字货币市场波动率与尾部风险的新指标.以比特币为例的实证结果表明,首先,相比于传统点值模型如指数加权移动平均模型,区间变量置信域的覆盖面积能同时有效度量比特币价格的水平与极差的不确定性,这增加了对日内价格波动的度量.其次,在分析尾部风险预测效果时,相比于历史模拟法和指数加权移动平均模型预测的在险价值,区间变量置信域生成的尾部分位数在条件覆盖率与非条件覆盖率检验上的表现更优.此外,本文提出的基于指数衰减加权的自举法更有效地刻画市场的非正态分布与时变性的特征.本研究不仅为数字货币的波动分析贡献了一种新的统计工具,而且为金融市场的尾部风险管理提供了新方法和新视角.

In the digital economy,the emergence of digital currencies has attracted considerable attention from both investors and researchers.However,their high volatility characteristics present new challenges in investment decision-making and risk assessment.To capture the characteristics comprehensively,this paper proposes a novel approach for constructing confidence regions for interval-valued variables based on the exponentially decay weighted bootstrap.The coverage area of the confidence regions and tail quantiles provide new indicators for assessing the volatility and tail risks in the market.Empirical results using Bitcoin as a case study demonstrate the proposed approach outperforms other traditional point-based methods such as exponential weighted moving average in measuring the uncertainty and intraday price volatility.Furthermore,the derived tail quantiles exhibit superior predictive performance for tail risk compared to Value-at-risk methods and the exponential weighted moving average,as evidenced by various tests.The proposed methodology not only contributes a new statistical tool for analyzing digital currency volatility but also provides novel perspectives for extreme risk management in financial markets.