停讓行人政策對交通事故之因果效應：2023年《道路交通管理處罰條例》修正案之案例研究

Abstract

本文旨在評估2023年《道路交通管理處罰條例》修正案（以下簡稱「停讓行人政策」）對於行人交通事故之影響。該法案自2023年6月30日施行後，擴大未停讓行人之適用範圍至未劃設行人穿越道之路口區域，並提高罰鍰上限、新增因未停讓而致人傷亡之罰則，象徵臺灣在強化行人路權的重要里程碑。為檢驗此政策之因果效應，本文利用內政部警政署「傷亡道路交通事故資料」，結合中央氣象署氣象觀測與監理所登記之汽機車數等資料，建構涵蓋2020、2022與2023年之鄉鎮月度面板資料，並採用雙重差分法（Difference-in-Differences, DID）與動態雙重差分法（Dynamic DID）進行實證分析。 實證結果顯示，政策實施後臺灣整體行人車禍件數及人數顯著下降約15%，且於政策實施後第六個月降幅進一步擴大至23%至26%，顯示政策在提升行人道路安全上逐步取得顯著成效。穩健性檢驗結果支持模型假設，並與主要結果一致，提升估計結果之可信度。異質性分析顯示，政策效果在不同交通情境與車禍參與者間呈現差異性影響，其中於工作時間（10:00至16:00）、路口區域及涉及年長者（65歲以上行人）之行人車禍事故降幅尤為顯著。 整體而言，本研究結果顯示停讓行人政策的實施有助於提升臺灣的行人道路安全，並於非路口區域產生外溢效果，反映臺灣整體行人用路環境之改善。然而，道路工程設計與交通安全教育仍有待持續精進，以形塑良好的駕駛人社會規範，確保政策成效得以長期維持，並作為我國推動「行人優先」交通文化與「零死亡願景（Vision Zero）」之重要政策參考依據。

This study evaluates the impact of the 2023 amendment to Taiwan’s Road Traffic Management and Penalty Act, hereafter referred to as the “Yield-to-Pedestrians Policy,” on pedestrian traffic accidents. Implemented on June 30, 2023, the reform broadened the scope of yield-to-pedestrian requirements to include intersection areas without marked pedestrian crosswalks, increased the statutory maximum fines, and introduced additional penalties for violations resulting in pedestrian injuries or fatalities. This reform represents an important milestone in strengthening pedestrian right-of-way protection in Taiwan. To identify the causal effects of the policy, this study utilizes road traffic accident injury and fatality data from the National Police Agency, combined with meteorological observations from the Central Weather Administration and registered motor vehicle data from vehicle registration authorities, to construct a township-level monthly panel dataset covering the years 2020, 2022, and 2023. The empirical analysis employs a Difference-in-Differences approach and a Dynamic Difference-in-Differences framework. The results indicate that the policy led to a statistically significant reduction of approximately 15% in both the number of pedestrian-related accidents and the number of pedestrians involved nationwide. Moreover, the magnitude of the effect increased over time, reaching reductions of 23%–26% six months after implementation, suggesting a strengthening policy impact. A series of robustness checks support the identifying assumptions underlying the empirical strategy, indicating that the estimated effects are reliable. Heterogeneity analyses further show that the policy effects vary across traffic contexts and participant characteristics, with particularly pronounced reductions observed during working hours (10:00–16:00), at intersections, and in accidents involving elderly pedestrians aged 65 and above. Overall, the findings suggest that the Yield-to-Pedestrians Policy has contributed to improvements in pedestrian road safety in Taiwan and has generated spillover effects in non-intersection areas. Nevertheless, sustaining these gains likely requires complementary investments in road engineering design and traffic safety education to reinforce pro-pedestrian social norms and support the long-term advancement of a pedestrian-priority traffic culture consistent with the “Vision Zero initiative”.