根據Arrow (1962) 率先提出的取代效果，既存獨佔廠商從事製程創新的動機較潛在進入廠商來得低 (不論創新之幅度)。根據Gilbert and Newbery (1982) 率先提出的效率效果，前者從事創新的動機較後者來得高 (非大幅度之創新) 或一樣高 (大幅度之創新)。本文採用Hotelling線性城市模型來描述新舊產品之差異 (水平和垂直)，以探討既存獨佔廠商和潛在進入廠商從事產品創新之相對動機大小，且探討所獲致的結論是否不同於Arrow (1962) 和Gilbert and Newbery (1982) 所獲致的結論。數值分析結果顯示，此問題的答案為「是，且就性質上來說，是大幅度的不同」。關鍵原因有二。一與區分非大幅度創新和大幅度創新之標準有關，另一與獲取創新的既存獨佔廠商之選擇有關。

The Arrow (1962) replacement effect shows that an incumbent monopolist has less incentive to pursue process innovations than a potential entrant, no matter what the innovation size. In contrast, the Gilbert and Newbery (1982) efficiency effect shows that the incumbent’s incentive to pursue a non-drastic innovation is higher and that both firms have equal incentives to pursue a drastic innovation. Hotelling’s linear city model is used in this study to describe horizontal and vertical differentiation between old and new products. The purpose is to examine the relative incentives of the incumbent monopolist vs a potential entrant to pursue product innovations and whether the results differ from those obtained by Arrow and those obtained by Gilbert and Newbery. The results of the numerical analysis show large qualitative difference. There are two key reasons for this; one is related to the criterion used to classify innovations as non-drastic or drastic, and the other is related to the innovating incumbent’s choice.