臺灣地區家戶汽機車相互持有與使用間的關係──Ordered Bivariate Probit與SURE模式之應用

Relationship between Household Car/Motorbike Ownership and Usage in Taiwan—Applications of the Ordered Bivariate Probit and SURE Models

本文從汽、機車二種自有運輸工具之間的持有與其使用之間的替代／互補關係切入，先透過Ordered Bivariate Probit模式之應用，建立家戶汽機車共同持有的聯合機率模式，來探討臺灣地區家戶汽機車共同持有的替代／互補關係；其次再利用相似無相關迴歸（Seemingly Unrelated Regression，SURE）模式的建構與估計，來討論家戶在汽機車使用上相互間關係。研究結果顯示，透過對條件共變異矩陣的檢定，除了彰化縣與高雄市外，臺灣各縣市汽、機車持有間及使用間，具有明顯的替代關係；由於此一顯著的替代關係，使得臺灣地區家戶的汽機車持有比率較其他國家為低；而由於影響各地汽機車持有與使用的因素各有所不同，因此建議汽機車相關的交通政策應具有地區特殊性。 We examined the related policy measures with two transportation modes, motorbike and car, simultaneously by investigating the substitute/ complementary relationship between these two modes at a household level. From the estimation results of the Ordered Bivariate Probit Model, we found out that there exists a significant substitution relationship between the household car/motorbike ownership, except for Changhwa County and Kaoshiung City. Furthermore, we used the Seemingly Uncorrelated Regression (SURE) Model to investigate the household car/motorbike usage relationship. The estimation results also showed a strong substitution relationship between the household car/motorbike usage, with the exception of Changhwa County and Kaohsiung City. These consistent results suggested that the substitution relationship between car and motorbike should be taken into consideration when the government or agents formulate any policies related to these two modes. The different factors affecting the relationship between the household car/motorbike ownership and usage in different areas of Taiwan also suggested the necessity of localizing the transportation policies.

We examined the related policy measures with two transportation modes, motorbike and car, simultaneously by investigating the substitute/ complementary relationship between these two modes at a household level. From the estimation results of the Ordered Bivariate Probit Model, we found out that there exists a significant substitution relationship between the household car/motorbike ownership, except for Changhwa County and Kaoshiung City. Furthermore, we used the Seemingly Uncorrelated Regression (SURE) Model to investigate the household car/motorbike usage relationship. The estimation results also showed a strong substitution relationship between the household car/motorbike usage, with the exception of Changhwa County and Kaohsiung City. These consistent results suggested that the substitution relationship between car and motorbike should be taken into consideration when the government or agents formulate any policies related to these two modes. The different factors affecting the relationship between the household car/motorbike ownership and usage in different areas of Taiwan also suggested the necessity of localizing the transportation policies.