发帖
雷达卡
因子强度的度量：理论与实践
Measurement of FactorStrength: Theory and Practice
作者：
纳塔利娅·贝利（Natalia Bailey）
乔治·卡佩塔尼奥斯（George Kapetanios）
哈希姆·佩萨兰（M. Hashem Pesaran）
This paper proposes an estimator of factor strength and establishes its consistency andasymptotic distribution. The proposed estimator is based on the number of statisticallysignificant factor loadings, taking account of the multiple testing problem. We focus on the casewhere the factors are observed which is of primary interest in many applications inmacroeconomics and finance. We also consider using cross section averages as a proxy in thecase of unobserved common factors. We face a fundamental factor identification issue whenthere are more than one unobserved common factors. We investigate the small sample propertiesof the proposed estimator by means of Monte Carlo experiments under a variety of scenarios. Ingeneral, we find that the estimator, and the associated inference, perform well. The test isconservative under the null hypothesis, but, nevertheless, has excellent power properties,especially when the factor strength is sufficiently high. Application of the proposed estimationstrategy to factor models of asset returns shows that out of 146 factors recently considered in thefinance literature, only the market factor is truly strong, while all other factors are at best semistrong, with their strength varying considerably over time. Similarly, we only find evidence ofsemi-strong factors in an updated version of the Stock and Watson (2012) macroeconomicdataset.
本文提出了因子强度的估计量，并建立了其强度和渐近分布。拟议的估算器是基于统计上重要的因素负荷的数量，并考虑了多重测试问题。我们关注于观察到在宏观经济学和金融学的许多应用中最重要的因素的情况。在未观察到的公共因素的情况下，我们还考虑使用横截面平均值作为代理。当存在多个不为人知的共同因素时，我们将面临一个基本的因素识别问题。我们通过蒙特卡罗实验在各种情况下调查提议的估计量的小样本属性。总的来说，我们发现估计器和相关的推论表现良好。该测试在原假设下是保守的，但尽管如此，它仍具有出色的功效，尤其是在因子强度足够高的情况下。将拟议的估计策略应用于资产收益的因素模型表明，在金融文献中最近考虑的146个因素中，只有市场因素才是真正强大的，而所有其他因素至多是半强的，它们的强度在整个时间。同样，我们仅在Stock and Watson（2012）宏观经济数据集的更新版本中找到半强因素的证据。将拟议的估计策略应用于资产收益的因素模型表明，在金融文献中最近考虑的146个因素中，只有市场因素才是真正强大的，而所有其他因素至多是半强的，它们的强度在整个时间。同样，我们仅在Stock and Watson（2012）宏观经济数据集的更新版本中找到半强因素的证据。将拟议的估计策略应用于资产收益的因素模型表明，在金融文献中最近考虑的146个因素中，只有市场因素才是真正强大的，而所有其他因素至多是半强的，它们的强度在整个时间。同样，我们仅在Stock and Watson（2012）宏观经济数据集的更新版本中找到半强因素的证据。
京公网安备 11010802022788号
