全國與城市層級汽機車總體持有與使用模式之建構

Abstract

近數十年來，經濟的快速發展與高速公路持續興建導致汽機車持有率與使用量的持續攀升。此一趨勢不僅造成都市與城際交通嚴重壅塞，亦造成污染廢氣之過度排放與大量之能源消耗。為達永續運輸之目的，提出有效降低汽機車持有率與使用量對策即成為重要之課題。然而策劃有效之私人運具管理策略仍需仰賴對於影響汽機車持有與使用之重要因素的了解。 基於上述原因，本研究自數個資料庫中蒐集汽機車持有與使用之相關解釋變數資料。全國層級資料主要來自於國際道路組織整理之世界道路統計彙編，該資料庫包含全球188個國家自1963年至2008年之資料。而城市層級資料主要來自於世界大眾運輸協會整理之千禧年城市永續運輸資料庫，該資料庫蒐集2001年度全球100個城市之相關資料。根據上述之資料，將分別針對全國層級與城市層級各建構兩種模式，其一為汽機車持有長期預測模式，其二為汽機車聯立模式。汽機車持有長期預測模式主要目的在於透過估計之羅吉斯非線性迴歸模式分別預測全國與台北都會區之汽機車持有長期趨勢。其中，羅吉斯非線性迴歸模式之汽機車飽和率，將透過判別函數設定高、中、低三個水準之飽和率值。而汽機車聯立模式將汽車持有、汽車使用、機車持有、機車使用等高度相關且相互影響之變數視為系統之內生變數進行模式估計。 汽機車持有長期預測模式結果顯示，全國之汽機車飽和率分別為每千人472輛汽車和760輛機車；台北都會區汽機車飽和率分別為每千人364輛汽車和760輛機車。根據模式結果，全國與台北都會區之汽車持有率將分別在平均國民所得三萬三千美元與三萬八千美元時達到飽和；全國與台北都會區之機車持有率將分別在平均國民所得三萬美元與三萬五千美元時達到飽和。 根據全國層級汽機車聯立模式估計結果，汽車持有率與平均國民所得、公路密度呈正向變動；與人口密度呈負向變動。汽車使用與機車使用、公路密度呈正向變動；與汽車使用成本呈負向變動。機車使用與汽車使用、道路交通負荷、汽車使用成本呈正向變動；與人口密度呈負向變動。另一方面，根據城市層級汽機車聯立模式估計結果，汽車持有率與汽車使用、平均國民所得、高速公路密度（每千人長度）、平均每人每日旅次數、平均旅次長度呈正向變動；與機車持有率、職業密度、大眾運輸路網保守密度（每千人長度）、平均每公里油價呈負向變動。機車持有率與職業密度呈正向變動；與機車使用、大眾運輸服務水準（每公頃延車公里）呈負向變動。汽車使用與汽車持有率、高速公路密度、平均每人每日旅次數、平均旅次長度、大眾運輸使用成本比例呈正向變動；與機車持有率、大眾運輸路網保守密度（每千人長度）、平均私人運輸成本呈負向變動。機車使用與汽車持有率、機車持有率、平均國民所得、平均私人運輸成本呈負向變動。並基於以上結果，提出相關對策之建議。

In recent decades, the rapid economic growth and continuous construction of highways has inevitably led to greater ownership and usage of private motor vehicles. This trend has not only created severe congestion on urban roadways and intercity highways, but also excessive emissions and energy consumption. Towards sustainable transportation, it is crucial to propose countermeasures capable of effectively curtailing ownership and usage of private vehicles. However, to devise management strategies which can effectively relieve dependency on private vehicles depends on the thoroughly understanding of the key factors affecting private vehicles ownership and usage. Based on this, this study collects data of car/motorcycle ownership and usage along with their related explanatory variables from several databases. The national-level data mainly come from the database of “World Road Statistics Compilation” prepared by International Road Federation, including related information of a total of 188 countries ranging from 1963 to 2008. While the city-level data are adopted from the database of “Millennium Cities Database for Sustainable Transport” compiled by International Association of Public Transport, which includes related information of a total of 100 cities in 2001. Based on such datasets, two types of models are respectively developed for both national- and city-levels. The first model is a car/motorcycle ownership growth model aiming to forecast the long-term growth of the number of private vehicles in Taiwan and Taipei by employing logistic-curve nonlinear regression model. To determine the saturation level of car and motorcycle ownership for facilitating the model building, discrimination functions of high, medium and low saturation levels are derived from the abovementioned datasets. The second model is a simultaneous equations model to take ownership and usage of cars and motorcycles as four endogenous variables from a whole system point of view due to the strong correlational relationship among them. The results of first type of models show that the saturation levels of car and motorcycle ownership of Taiwan are 472 and 760 vehicles per thousand people respectively, and those of Taipei are 364 and 760 vehicles per thousand people respectively. The ownership of cars in Taiwan and Taipei will reach their saturation level at 33 and 38 thousands US dollars, respectively; while the motorcycle ownership will attain their saturation level at 30 and 35 thousands US dollars, respectively. As to the estimation results of the national-level simultaneous equation model, income, road density and population density are identified as the key variables affecting car ownership. Motorcycle usage, road density and car usage cost are the key variables for car usage. Car usage, road traffic load, car usage cost and population density are the key variables for motorcycle usage. These findings confirm the existence of close relationship among car/motorcycle ownership and usage. A biased result will be obtained if these endogenous variables are separately estimated. On the other hand, according to estimation results of the city-level simultaneous equation model, car usage, income, freeway density, average number of daily trips, average trip length, motorcycle ownership, occupation density, network density of public transportation and fuel price are the key variables affecting car ownership. Occupation density, motorcycle usage and level of service of public transportation are the key variables for motorcycle ownership. Car ownership, freeway density, average number of daily trips, average trip length, the ratio of usage cost of public to private transportation, motorcycle ownership, network density of public transportation and average private transportation cost are the key variables for car usage. Car ownership, motorcycle ownership, income and average private transportation cost are the key variables for motorcycle usage. Based on the results, corresponding countermeasures are then suggested.