추종거래자의 존재가 기관투자가 간의 쏠림에 미치는 영향

Since the development of traditional Sharpe-Linter CAPM, it has been widely used in explaining the variation of stock returns. However, researchers identified some predictable errors in the model, such as size effect and value effect. In response to the findings, studies focused on the cause and cure of the problems. This paper suggests that those effects are the result of mis-specified assumption on investors' preference. CAPM holds when investors have a mean- variance optimization preference. When investors' preference does not fully represented by mean-variance, for example, when the return has asymmetric distribution, CAPM anomalies happen. The implied option of limited liability embedded in stocks makes the stock return's distribution different from symmetry. We used the macroeconomic bond credit spread data to control for the changing undiversifiable option risk. The estimation results using U.S.'s long-term stock return and bond yield data show that the revised model does not have those systematic errors. It is also shown that the sensitivity to the macroeconomic credit spread depends on the size and B/M ratio. While the revised model resembles Conditional CAPM but it suggests different logic in explaining size and value effect.