Previous studies have reported that urban spatial structure and the spatial relationship between jobs and housing are strongly correlated with commuting patterns. Nevertheless, a number of studies have also supported the concept that the spatial relationship between workplaces and residences cannot be the only reason behind observed commuting behaviour. Some ‘soft’ factors also affect people's commuting patterns. When considering these ‘soft’ factors, urban land development patterns such as urban structure and the spatial relationship between jobs and housing must not be overlooked. In this paper, we first conduct a brief review of the debate about how urban spatial structure and the jobs–housing relationship affect commuting patterns. Then we mainly focus on the ‘soft’ factors related to commuting behaviour. We divided the ‘soft factors’ into city-level factors and individual- and household-level factors. These perspectives will offer insights to identify a set of key factors that could affect the patterns of commuting.

Income status is an important variable that is strongly associated with certain commuting behaviours of workers. This paper presents new evidence on how polycentric development impacts on workers’ commuting behaviour among various income groups in Beijing, China. This study suggests that three key influencing factors—the public transport network, the location of affordable housing projects and the process of employment decentralisation—have played significant roles in affecting workers’ commuting behaviour. The results of regression analysis indicate that subway and bus transport significantly and negatively influenced the commuting times of low- and middle-income workers, but the two transport modes did not have a significant influence on the commuting times of high-income workers. The findings from this research suggest that policies for promoting employment decentralisation during polycentric development have the potential to reduce workers’ commuting times through promoting jobs-housing balance in the sub-centres. The results of this study indicate that a balanced jobs-housing relationship can be achieved through adjustment of affordable housing locations, and this can be effective in shortening low-income workers’ commuting times in the sub-centres of Beijing.

Institutional and economic reforms in China have not only delivered major improvements in people’s living standards but also changed the spatial structure of Chinese cities. In addition, substantially increased commuting times have become a serious problem influencing quality of life in the mega-cities of China. Accordingly, research into the dynamics of urban restructuring in rapidly growing Chinese cities and the impacts on workers’ commuting patterns would provide important evidence in this research field. The results of this case study suggest that polycentric urban development has a significant influence on employees’ commuting times. Promoting local jobs-housing balance in sub-centres would shorten commuting times for workers. The findings also suggest that the locations and types of employment centres are significantly associated with individuals’ choice of housing and workplace locations, and accordingly their commuting behaviour. The effects of decentralization of employment on workers’ commuting times differ depending on the extent of suburbanization of different economic sectors. The findings in this research also suggest that workers’ commuting behaviour is related to individual socio-economic characteristics, such as income and education.

Substantial increases in commuting times have become a serious problem affecting the quality of life in Chinese megacities. Accordingly, research into the dynamics of urban spatial restructuring and socio-economic transition in rapidly growing Chinese cities and the impacts on workers’ travel behaviour can provide important evidence in this research field. Using data from a recent survey of employees at their workplaces in sub-centres of Beijing, a multiple linear regression analysis was employed to measure the effect of certain variables on workers’ commuting time, by identifying relationships between the various independent variables (local jobs–housing spatial relationships, individual socio-economic characteristics and choice of transport mode) and the dependent variable (workers’ commuting times). The results of the regression analysis show that jobs–housing balance has a more significant impact on workers’ commuting times than socio-economic characteristics. The results also suggest that workers’ commuting behaviour is strongly related to socio-economic characteristics such as income, age, gender and education, as well as institutional factors such as economic and land-use reforms.