| Summary: | 碩士 === 國立中山大學 === 財務管理學系研究所 === 103 === Market reflexivity refers to Soros(1987) describe a phenomenon of feedback-loop interaction between investors’ action and investment environments. It has been used in criticizing efficient market hypothesis and providing another explanation of the boom and burst of stock market. However, there are few empirical studies due to the difficulty of modeling the concept of market reflexivity. In this study, we quantify the level of endogeneity of price formation process and as the “market reflexivity” of the TXF index futures of Taiwan Futures Exchange, using the methodology of Hawkes conditional Poisson model proposed in Filimonov and Sornette(2012). We first examine the price movement before and after the exceeding of a critical 99 percentile of the market reflexivity measure. To test the validity of market reflexivity as a leading indicator of market fragility, the study also observes the interaction between market reflexivity and the liquidity, volatility and efficiency of the market. Finally, to decipher the factors contributing to reflexivity, we construct the trading network of TXF and invoke the concept of social network analysis such as centrality, clustering and centralization to describe the behavior of investors and the environment of market.
Our main findings are threefold: (1) There is significant volatility before and after the exceeding of the critical threshold of market reflexivity. Specifically, further drawdown is significant in bearish market when the threshold is reached. (2) Intraday market reflexivity leads the liquidity and efficiency variables by 5 minutes. (3) Herding, high frequency trading and the clustering of market significantly positively related with the market reflexivity. Our study shows an application of market reflexivity as an indicator of market fragility which reflects the impact of herding and high frequency trading.
|
|---|
