我國證管會法令規定公司出售不動產應於三日內公告之，此為該事件首次公告，其後該季之季盈餘中亦將包括出售不動產損益，此為再次公告。投資人如果只注重底線數字，而不深究其數字背後的含意，則該資訊再次公告時，投資人仍然會對包含在底線數字中的出售不動產損益產生反應，此為功能固著假說之現象。本文探討我國證券市場對出售不動產事件是否存在功能固著之現象，實證結果發現：1.出售不動產事件之首次公告，確實會對股價產生影響。2.再次公告出售不動產損益時，仍會再次對股價產生影響，證明投資人對出售不動產事件存有功能固著情形。3.投資人的精明程度並不會影響資訊再宣告時的股價反應，因此無法證明我國證券市場符合擴大功能固著假說。4.公司規模會影響資訊再宣告時股價反應的程度。 / The Securities and Futures Commission in Taiwan requires that the sale of real estate by a firm should be announced within two days of the transaction. The sales revenue of real estate transaction will be publicized again when the quarterly earnings are released. In an efficient market, the price will simultaneously and fully reflects the transaction information. However, if investors react once again to the sales revenue of real estate transaction at quarterly earnings announcement, the market has functional fixation phenomenon. This study examines the existence of the functional fixation phenomenon through real estate transaction. 　　The empirical results illuminate: (1) Market price reacts to the announcement of real estate transaction. (2) The functional fixation hypothesis cannot be rejected based on the reaction of the market prices to the sales revenue of real estate transaction at quarterly earnings announcement. (3) The extent of sophistication of individual investors does not affect the reaction of stock prices to the sales revenue of real estate transaction at quarterly earnings announcement and therefore the extended functional fixation hypothesis is rejected in Taiwan security market. (4) Firm size affects the extent of market reaction to the sales revenue of real estate transaction at quarterly earnings announcement.

退休金的會計處理原本即具複雜性，加上最近退休金制度的變革，形成退休金會計獨特且豐富的情境背景，提供了許多的研究機會。本論文包含三個與退休金會計相關之議題；其中第一個議題是「勞工退休金新制之股價效應」；第二個議題是「退休金精算假設選用之誘因」；第三個議題是「退休金精算假設選用之價值攸關性」。三個議題均以效率市場假說為基礎，針對勞工退休金新制之實施與退休金精算假設之選用，分別深入探討其與股價之關聯性。 首先，本論文第一個議題探討勞退新制之股價效應。事件日股票報酬與公司特質的關聯分析，雖然部分的實證結果不顯著，然而部分的證據顯示，勞退新制公布實施時，公司股票報酬與退休金提撥不足程度、員工平均服務年資、退休金提撥率、退休金成率等公司特質具有關聯性，隱含資本市場在某種程度上，似乎可以依據勞退新制對公司不同之衝擊程度，反映於公司股票報酬。 其次，本論文第二個議題探討退休金精算假設選用之誘因。實證的結果顯示，我國上市上櫃公司精算假設選用的橫斷面差異，可能源於債務契約、提撥不足之成本以及損益平穩化等誘因。最後，本論文第三個議題探討退休金精算假設選用之價值攸關性。實證結果顯示，退休金精算假設選用具價值攸關性，投資人對公司選用精算假設保守程度有所評價，對於裁量性 (隱藏的) 預計給付義務亦有所反映，似乎適當地根據精算假設選用對於預計給付義務之影響幅度來調整對公司之評價，隱含資本市場並未功能性固著於報導的退休金義務。 / There are three essays in this dissertation. The first essay examines the impact on equity prices of the Promulgation of Labor Pension Act. Evidence reported in the paper shows that negative abnormal returns are most pronounced for firms with large underfunded level of pension plan, firms with little average years of employment, firms with low contribution ratio, and firms with low pension cost ratio. The second essay explains the cross-sectional variation in firms’ selected actuarial assumptions (rate of increase in compensation and discount rate) used to measure the projected benefit obligation (PBO). Evidence shows that firms with relatively larger debt ratio, larger underfunded level of pension plan, and larger decline in earnings tend to select more aggressive (obligation-reducing) estimation parameters. The last essay studies the association between actuarial assumptions and firm value. The results indicate that firm value is lower, conditional on the reported PBO, for firms that use higher discount rates and lower rate of increase in compensation. This is consistent with investors seeing through managers’ opportunistic choices of obligation reducing assumptions. The evidence suggests that the stock market does not fixate on reported PBO numbers, but properly appreciates the value relevance of the discretionary component of the PBO.