流動性共變在台灣股市之現象、成因與定價

Abstract

流動性共變泛指個別股票之流動性與全體市場流動性相關之現象，流動性存 在共變現象意味著流動性並不只是個別股票的獨特風險，也可能是市場共同系統性 風險的一種，且會降低分散投資的效益，對法人的投資策略影響甚鉅。如果投資人 無法透過分散投資而沖銷此風險，則有可能會因持有流動性較差的股票而要求額外 的風險貼水。 此三年期計畫旨在提供對臺灣股市流動性共變之深度分析。第一年計畫研究 個股流動性是否有共同成分，以及此共同成分對個股流動性的影響程度。本研究將 使用多個流動性指標衡量個股流動性，分析個股流動性和全體市場流動性的關係， 並呈現不同族群個股的橫斷面和時間序列流動性特徵，包括大型股與小型股的流動 性共變差異、指數成分股與非成分股的流動性共變差異、多頭市場與空頭市場的流 動性共變差異、大幅波動和小幅波動市場的流動性共變差異等。 第二年計畫研究個股流動性共變的決定因素，試圖利用公司特性(包括公司規 模、法人持股比例、外資持股比例、媒體覆蓋度、報酬波動性等)詮釋個股對市場流 動性的敏感度，並討論各族群股票之平均流動性敏感度隨時間而演變的趨勢。本計 畫提供流動性敏感度的橫斷面與時間序列分析，深入探尋流動性共變的成因，具有 銜接第一年計畫和第三年計畫的功能。 第三年計畫探討流動性風險是否應納入定價模型中，成為投資人要求報酬的 系統性風險因子之一。首先引用上兩年度計畫的研究結果，選取適當的流動性代理 變數為風險因子的基礎，再使用Fama and French (1993)的投資組合方法建構風險因子，最後則比較CAPM 和Fama-French 三因子模型在納入流動性因子前後的表現， 觀察原本模型的異常報酬(anomalies)是否因為加入流動性因子而降低。

Liquidity commonality refers to the phenomenon whereby the individual firm's liquidity is at least partly determined by market-wide factors. Common liquidity movements imply that liquidity risk is more than idiosyncratic shocks that affect individual stock one at a time. It may be market-wide and can not be reduced by diversification. If investors have to bear the systematic component of liquidity risk, they will therefore demand compensation from holding illiquid assets. Whether a group of stocks respond significantly to their common liquidity movement will significantly influence choices made by portfolio managers’ decision that attempt to minimize the liquidity impact of asset positions. This three-year proposal intends to provide a comprehensive investigation to the liquidity commonality in Taiwan stock market by three studies, one in a year. In the year-one project, we examine whether stock liquidity show common movements and the extent to which the commonality in liquidity explain the liquidity of individual stocks. Using a variety of liquidity proxies, the study will explore the degree to which individual stock liquidity is correlated with market-wide measures of liquidity. The study also analyzes some important cross-sectional and time-series liquidity characteristics, for the overall market, including large versus small firms, index versus non-index stocks, bull versus bear markets, and calm versus turmoil markets. In the year-two project, we will explore the factors that determine the cross-sectional variation in liquidity commonality. The study relates each firm’s sensitivity to the common liquidity to firm characteristics that potentially explain the cross-sectional divergence of liquidity commonality. The explanatory variables include firm size, institutional ownership, index inclusion, foreign investor accessibility, media coverage, return volatility, degree of order imbalance, etc. In addition, the paper will study the evolution of liquidity sensitivity of Taiwan stocks over the most recent ten years. The year-two project seeks to provide knowledge about the time-series pattern of commonality and explanation about the cross-sectional divergence of the stock sensitivity to the common liquidity factor. The study bridges the gap between the year-one project that finds a common liquidity factor and year-three project that evaluates the impact of the factor in the pricing model. The third year project completes the study by analyzing whether the liquidity factor is priced and whether investors demand liquidity premium for holding less liquid assets. A liquidity measure will be selected based on the results in the previous two projects and upon which an aggregate market liquidity risk factor will be calculated. The study then tests the liquidity-included capital asset pricing model against the conventional CAPM and Fama-French three factor model. The study will verify the model by examining the time-series pattern of the liquidity premium and the cross-sectional characteristics of the risk loading across portfolios sorted by other proxies of liquidity.