Exploring the Dynamic Model of the Returns from Value Stocks and Growth Stocks Using Time Series Mining

[[abstract]]This study considered that value stocks and growth stocks are 2-dimensional concepts. We defined the book-to-market ratio as the value factor and the return on equity as the growth factor. We used these 2 factors to divide stocks into 4 types: high-value, low-value, high-growth, and low-growth stocks. Furthermore, we explored the change in stock prices and stock returns for these 4 categories before and after the formation of investment portfolios. We also established a dynamic model showing the returns from value stocks and growth stocks, called the exponential decay model. Finally, we used Taiwan Stock Exchange data to examine effectiveness of the model during the period from 1995-2009. The results are as follows: First, high-value stocks and low-value stocks exhibit a significantly over-reacting phenomenon. Second, high-growth stocks and low-growth stocks exhibit an obviously under-reacting phenomenon. Third, in each current quarter, high-value stocks exhibit the lowest returns; however, in the subsequent quarter, they have the highest returns, and then demonstrate a slow declining trend in the following quarters. These results showed that the stock market can exhibit a dramatic response to extraordinary information and proved that the stock market requires considerable time to correct themselves from an excessive reaction, thus high-value stocks exhibited a higher return. Fourth, in each current quarter, high-growth stocks had the highest return, followed by a rapidly decreasing trend in the following quarters. The t+3 quarter returns were lower than those of low-growth stocks. This result demonstrated that the stock market does not exhibit an adequate reaction, but still remains rather efficient for routine financial information. Finally, regardless of value stocks or growth stocks, exponential decay models could accurately match with the data.[[notice]]補正完