各項選股指標於台灣股票市場的實用性探討--應用對象與應用時機之研究

Abstract

本研究參考Fama and French(1992)所使用的研究方法，研究樣本包括台灣上市櫃276家公司，期間為1998年1月到2006年10月，探討目前投資業界較常使用的三個選股指標：本益比、股價淨值比與盈餘成長率，加上公司規模與貝他值(β值)等共五個選股指標在台灣股票市場對於股票預期報酬率的平均解釋能力如何。本研究的自變數與因變數的估算假設，採用最即時與最符合現實情況的資料分析方法，並針對不同投資區間與不同產業特性類股作相同的檢測，期望讓實證結果可獲得更高的實用價值，加強投資人了解該使用那些選股指標、應用對象與使用時機的理論參考與信心度。 實證結果發現，就長期而言，台灣股市存在本益比效果，也就是低本益比的公司享有較高的報酬率，盈餘成長率也是長期適用的指標，盈餘成長率較高的公司享有較高的報酬率，顯示公司的盈餘表現仍為股價表現的最佳解釋指標。就短期而言，若大盤處於下跌趨勢期間，選股指標可能會暫時失靈，但若大盤重回上升趨勢期，長期具解釋力的選股指標則可重現良好的解釋能力。 對景氣循環類股來說，不管長短期，公司規模都是很好的選股指標，而股價淨值比與本益比雖是長期適用的選股指標，但可能在短期之內失靈。景氣循環股中規模越大的公司，其報酬率愈高，低股價淨值比或低本益比的公司也有較高的報酬率。 對電子類股而言，長期並無適用的選股指標，但若大盤處於上升趨勢期，則β值可能是一項實用的選股指標，β值愈低的公司反而有較高的報酬率，與理論上變動方向相反，可能是輪漲或補漲的效應所造成，電子類股適用的選股指標異於其他類型公司。

By applying the research methodologies of Fama and French (1992) to the historical data of the 276 publicly listed companies in Taiwan between the dates of January 1998 to October 2006, we investigate the five most commonly used stock selection indicators amongst investors: price to earnings per share ratio(P/E), price to book-value(P/B), earnings growth rate, market capitalization and beta. Using these five leading indicators we are looking to determine its effectiveness in explaining and predicting stock subsequent performance. Variables that are chosen in this research is based on the most practical data analysis methodologies adopted to similar tests for different investment horizons and industries with the goal to achieve better applicability. These should help investors better understand its effectiveness in timing the usage of these indicators to increase investor confidence levels. Our empirical results show that in the long term, companies with low P/E’s enjoy higher expected returns, while earnings growth is a more appropriate long term indicator as data suggests earnings performance is still the best indicator for stock performance. However these indicators do not explain short term share price fluctuations especially when the market is in a downward trend cycle. In terms of cyclical sectors, market capitalization is both a good long/short term stock selection indicator while P/B and P/E are more appropriate for long term indicators; though ineffective in the short term. For cyclical names, both large capitalization companies and low P/B and P/E companies also provide better returns. With regards to the technology sector, there appear to be no appropriate long term stock selection indicator; however in an upward trend market, beta could be a practical indicator. Low beta companies giving better returns contradict the financial theoretical understandings. It is possible that these results explain the “catch up effect,” suggesting the technology stock selection indicator differs from other sectors.