| Summary: | 碩士 === 輔仁大學 === 應用統計學研究所 === 95 === Taiwan has successfully transformed from the traditional labor-intensive to high tech Electronic industry in the past decade. This high tech industry has doubled its manufacturing output value since 1999 and the future trend of M&A in the strategic alliance and globalization strategy can be heavily addressed to gain its further competence. Among the globe OEM business segments of the Electronic productions, two top most competitive companies are UMC and TSMC in Taiwan. Since the high tech industry is greatly influenced by the market macro-economic indicators, the dynamic qualities of electronics which lead to the fast moving product prices directly result the significant stock price fluctuation. For example, in H1 of Y2000 TSMC and UMC each pursued large scale M&A. The outcome is that their stock price both climbed into its historically high level. The internet bubble; however, adversely counters their efforts in H2 and their market value drops vertically. Thus, we try to apply the value-at-risk (VaR) analysis in the stock price variation due to M&A and target to the TSMC and UMC so that we can make use of the four different kinds of risk value evaluation models for the estimations of different significant levels and moving average days. We want to choose a better forecast model to analyze the risk generated from the M&A and use the estimation of risk value to understand more about the market risk exposures in the high tech industry.
We concluded that the Monte Carlo method simulation behaved better when the moving average days is set to 50, roughly a half-quarter periods, since any forecast periods exceeding a quarter will result in poor accuracy by any extremely high penetrative rates. It also indicated that most of the estimated VaR after M&A are smaller than those before M&A, which shows a declining trend of market risk by pursuing M&A strategy. Among 4 methods, the GARCH Model could in general get more accurate estimation of stock price volatility and thus perform better in the VaR analysis.
|
|---|
