本論文目的在於探討「異質期望」(heterogeneous expectations)、「資訊密度」(information arrival intensity)、以及「跳躍風險」(jump risk) 這些因素對於期貨價格的影響，並且透由「跨期模型」(intertemporal models) 的建立，推導出具有封閉解形式的期貨價格理論公式。 誠如 Harrison and Kreps (1978) 所言：除非所有市場參與者的行為方式完全相同、而且他們都打算抱著股票直到永遠，否則「投機交易」(speculation transactions) 與「異質期望」就不可能自市場當中滅絕。有鑑於此，本論文在第二章中討論「異質期望」對於期貨價格的影響；同時為了反映交易者看法可能會隨時間演進而發生改變的可能性，「調整效果」(adjustment effects) 是本章另一個討論重點；第三、為了區別期貨契約與遠期契約的基本差異，「利率」這個隨機因子也被納入模型當中。由「部分均衡」(partial equilibrium) 觀點下具有封閉解形式的期貨價格公式來觀察，這三個重要因素以及彼此間存在著的複雜交互作用，可以協助解釋一些實證現象與重要變數之間的關係。 第三章主要是借用Clark (1973) 與Chang et al. (1988) 「資訊時間」(information time) 的概念，取代一般模型所使用的「日曆時間」(calendar time) 設定方法，並且額外納入「利率」與「便利所得」(convenience yield) 這兩個廣為一般期貨定價文獻所認定的重要隨機因素，推導出「部分均衡」觀點下的期貨價格封閉解。根據1998/7/21 至 2003/12/31 台灣期交所「台灣證券交易所總加權股價指數期貨」的實證結果來看，本章模型的定價績效不僅勝過「持有成本模型」(the cost of carry model)，也比同時考慮「利率」與「便利所得」兩個隨機因子的「日曆時間」模型要來的好。 第四章則是嘗試結合Hemler and Longstaff (1991) 的「無偏好模型」(preference-free model) 以及Merton (1976) 的「跳躍」(jumps) 設定，重新推導「一般均衡」(general equilibrium) 模型下、考慮「跳躍風險」(jump risk) 後的期貨價格封閉解。根據本章各種比較靜態與模擬分析的結果顯示，整個經濟體系或是「狀態變數」(state variables) 的安定程度，決定了市場變數間的關係；另一方面，這些關聯會因為「跳躍風險」規模的遞增 — 不管是肇因於「發生機率」(occurring probability) 或是「衝擊效果」(impulse effect) — 而變的更加不可預測。 / The dissertation contains three essays on intertemporal futures pricing models allowing for heterogeneous expectations, information-time based setting, and jump risk. As Harrison and Kreps (1978) have noted, unless traders are all identical and obliged to hold a stock forever, speculation would not extinguish in market, and heterogeneity in expectations yields whereby. The first essay develops intertemporal futures pricing formulas accounting for such reality, adjustment effect, and stochastic interest rate in a partial-equilibrium sense. The closed-form solutions show that the three factors complicated with each others can help to explain some existing empirics on relationships between futures prices and other important market variables such as indeterminate converging pattern. The second essay extends Chang et al. (1988) option pricing model to derive futures prices with information-time based processes. Stochastic interest rate and convenience yield are also taken into account to derive closed-form formulas. According to empirical results of transaction data of TAIEX index and its corresponding TFETX futures contract through 1998/7/21 to 2003/12/31, the analytic solution performs better than the cost of carry model and its calendar-time based counterpart, especially when information arrival intensity estimates become larger. The last essay combines Hemler and Longstaff’s (1991) preference-free model and Merton’s (1976) jump setting to measure effects from jump risk and a futures pricing formula is derived in its closed-form as well. According to miscellaneous comparative static and simulation results, the bounded degrees of state variables, or economy, affect co-varying extents among variables, while the increasing jump risk, including the size of occurring probability and its corresponding impulse effect, makes them un-tractable.