The asymmetric income effect on household vehicle ownership in Taiwan: A threshold cointegration approach Rong-Chang Jou a , Wen-Hsiu Huang b,⇑ , Yuan-Chan Wu a , Ming-Che Chao a a Department of Civil Engineering, National Chi Nan University, No. 1, University Rd., Puli, Nantou County 54561, Taiwan b Department of Public Finance, Ling Tung University, No. 1, Lin Tung Rd., Taichung 40852, Taiwan article info Article history: Received 6 May 2011 Received in revised form 7 December 2011 Accepted 5 January 2012 Keywords: Household vehicle ownership Asymmetric threshold cointegration test Asymmetric error correction model abstract This paper uses the asymmetric threshold cointegration test to examine the asymmetric relationship between household income and vehicle ownership in Taiwan, presenting esti- mated asymmetric error correction models. The empirical data include information on household income, car ownership and motorcycle ownership in different regions from 1974 to 2009. The results show that, first, motorcycle ownership is asymmetrically cointe- grated with household income in each region, and car ownership is asymmetrically coin- tegrated with household income in all regions except Taipei city. Second, both car and motorcycle ownership levels increase faster than they decrease in the asymmetric adjust- ment of their long-run relationship. Third, sensitivity tests for the period 1987–2009 show that the cointegration relationship of the car ownership equations vanished. Finally, we find evidence on the effects of household income on motorcycle ownership, and the effects of income variables on car and motorcycle ownership are dissimilar. This study exhibits different results across regions. These findings may be related to the development of public transit system in each region. Ó 2012 Elsevier Ltd. All rights reserved. 1. Introduction In the developing world, designing policies that can effectively control the number of vehicles has become a major chal- lenge for governments. Vehicle ownership also creates problems such as traffic congestion, inefficient travel, air pollution and excessive energy consumption. The existing literature on vehicle ownership models can be divided into two parts in terms of its approach. One body of studies has used macroeconomic analysis, including aggregated and time series data. In this context, the primary type of analysis model is the regression model (Jansson, 1989; Button et al., 1993; Mannering, 1983; Said, 1992; Stanovnik, 1990). However, approaches based on aggregated data assume that individuals are homoge- neous and do not examine the behavioral factors influencing different individuals. The other body of studies has employed microeconomic analysis, including individual and cross-section data (Giuliano and Dargay, 2006; Matas and Raymond, 2008; Dissanayake and Morikawa, 2010). However, cross-sectional surveys cannot capture the impact of changes in variables over time. It was for this reason that Nolan (2010) used pool data in his dynamic analysis of car ownership. Many previous studies have highlighted the nature of relationships between socio-economic characteristics and car own- ership. A number of studies have been undertaken that have examined the effects of income on the ownership of cars, motor- cycles or both (Dargay and Gately, 1999; Giuliano and Dargay, 2006; Dargay, 2002; Jou and Sun, 2007; Van Acker and Witlox, 2010; Potoglou and Kanaroglou, 2008; Clark, 2007; Whelan, 2007). Some studies have used household expenditure instead 0965-8564/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.tra.2012.01.001 ⇑ Corresponding author. Tel.: +886 04 2389 2088x3622. E-mail address: michelle@teamail.ltu.edu.tw (W.-H. Huang). Transportation Research Part A 46 (2012) 696–706 Contents lists available at SciVerse ScienceDirect Transportation Research Part A journal homepage: www.elsevier.com/locate/tra