剩餘盈餘評價模型應用在台灣前五大IC設計公司之個案研究

Abstract

本研究係探討臺灣股票市場資訊不對稱與評價誤差之關聯性，以上市、櫃前五大IC設計公司為樣本，並以全球金融風暴後逐步恢復市場機制之2010年為研究期間。

本論文採用剩餘盈餘評價模型作為衡量企業實際價值(Intrinsic Value)之基準，將折現因子按不同風險溢酬分為3種可能方案，並將模型中2013年後之剩餘盈餘固定永續成長率再細分為5種可能方案，總計共15種可能方案加以探討不同因素變化下企業實際價值之變化幅度，同時藉由比較企業實際價值與市場價格之差異，衡量該公司評價誤差幅度。而資訊不對稱方面，則採4家機構法人在相近時間點，針對相同IC設計公司預估之目標價，以極端值差異幅度作為資訊不對稱幅度之代表。

研究結果發現，(一) 評價誤差與資訊不對稱呈現正相關：在15種評價誤差可能方案中，高達10種方案其評價誤差與資訊不對稱幅度呈現高度正相關；(二) 資訊不對稱方面：實證發現，前五大公司資訊不對稱幅度大小依序為聯發科、群聯、瑞昱、聯詠、凌陽；(三) 評價誤差方面：評價誤差幅度按15種可能方案而有不同結果，惟整體而言評價誤差幅度較大者包括聯發科、群聯，幅度較小者為凌陽。

The thesis aims to find the correlation between valuation error and asymmetric information in Taiwan Stock Market. The dataset employed in this essay is acquired from the top 5, including listed and OTC, Taiwan IC design companies. The period under analysis is focused on the year 2010, which is a recovering year after the global financial crisis.

The study takes Residual Earnings Model as a valuation way to reflect the intrinsic value of the IC Design companies, and evaluates the range of intrinsic values combining with 3 scenarios of different risk premiums and 5 scenarios of different residual earning growth rates. The difference between the market price and the intrinsic value determines the scale of valuation error, and the difference between the 12-month target prices forecasted by 4 different institutional investors determines the scale of asymmetric information.

The conclusions are drawn as follows. (1) Valuation error is positively related to asymmetric information：There are 10 scenarios showing obviously positive correlation between the scale of valuation error and the scale of asymmetric information within 15 kinds of scenarios in total. (2) In terms of asymmetric information：Asymmetric information in descending order is MediaTek、Phison、Realtek、Novatek、Sunplus that the empirical results show. (3) In terms of valuation error：Descending order of valuation error depends on different kinds of scenarios. Overall, MediaTek and Phison have much severer valuation error and Sunplus has slighter valuation error.