This paper examines the predictability of stock returns at the individual firm’s level in the Tokyo Stock Exchange by quantifying the tone of the news articles during the 12-year from 2003 to 2014. We compared predictive powers of the body and the headline of news articles. As a result of examinations, we show that the tone of the news article has high predictive powers for stock returns of one trading day later. Especially in our analysis methods the headline tone which is more predictive than the body tone. The headline tone still has predictive powers even after taking the transaction cost into account. Furthermore there are low predictive powers for stock returns of two or three trading days later.

This study examines the chilling effects of the cross-side network effects that should be the one of the core concepts of the two-sided platforms theory. Firstly, the streams of academic literature such as the empirical research for the video games industry, the theoretical research for the chilling effects of the direct network effects, and the simulations for the two-sided markets are reviewed. Secondary, using the system dynamics approach, the two-sided markets model including the cross-side network effects for the video game industry is developed and validated. By the simulated results of three cases of video game consoles, the classifying in patterns for the chilling effects of the cross-side networks are identified, which is the combination of types of time-dependent forms and the variance of two sides such as user side and business side. Finally, the hypotheses concerning the management policy of the platform providers are proposed, through making the correspondence between the real event and the simulated result.