本文旨在研究如何以台灣經濟新報之長期追蹤調查資料(panel data)估計臺灣個體廠商的Tobin's Q。由於臺經新報的資料庫起始年份並不久遠,使得國外常用以估計資產重置成本的永續存貨法(perpetual inventory method)喪失使用的先決條件。此外,臺灣資料雖有「資產重估增值」乙項,但由於法令的限制,使得該重估值亦非理想的重置成本估計值。本文利用臺灣資料的特性,重新建構廠商在樣本前的資產狀況,再配合文獻中新近提出、理論較完整的年份結構法(vintage structure method),估計臺灣廠商的Tobin's Q。結果顯示,本文建構之Q比例變數,統計上顯著地相異於直接以財務報表據估計之Q變數,或以永續存貨法衡量之Q變數,且其亦具有較理想的統計性質。本文以臺灣製造業廠商資本投資的實證分析為例,說明相對於其它二者,本文建議之Q比例變數確實有較強的模型解釋能力。此外,本文亦發現若使用不同的Q變數,將有可能對某些實證命題產生相異的結論。
This paper proposes a method to measure the Tobin’s Q for Taiwanese manufacturing firms based on the Taiwan Economic Journal (TEJ) Data Bank. The focus is particularly on the measure of the replacement cost of capital. Because TEJ only collects data from 1981, and many of the firms do not even appear in the dataset until 1988, the relatively short panel history makes the popular perpetual inventory approach inapplicable. The method proposed in this paper is based on the vintage structure method, which is further modified to take into account the uniqueness of Taiwanese accounting rules and the phenomenon of the rapid increase in Taiwan’s land prices. We empirically compare our Q variable with the ones constructed from the book values and from the perpetual inventory method. The results show that our Q variable has a better explanatory power for an investment model, and that using different Q measures may lead to opposite conclusions on some empirical issue.