Beijing’s rapid economic growth in recent years has induced enormous transportation demand. Nowadays, congestion and air pollution caused by increasing use of cars are perceived as some of the most pressing problems in Beijing.
To cope with these problems, the “low-fares of public transport” policy was adopted in 2007, by which passengers get 60 - 80% off the regular price; in the same year, subway fare decreases from 3-5 yuan to 2 yuan; from 2007 to 2009, 4 new subway lines and 1 airport express were constructed; and another 10 subway lines are under construction.
Besides these policies and measures in effect, some other measures are under discussion, such as parking lot fees, congestion charges. To evaluate and learn from the existing programs and to design the most effective and desirable programs in the future, it is critical for the policy makers to acquire as much information about travelers' behavior as possible.
In this research project, we analyze travelers’ mode choice behaviors by using revealed preference survey data in Beijing for three years from 2007. Several versions of a discrete choice model are specified and estimated. A mixed logit model, which allows for random taste variation and unrestricted substitution patterns, is preferred. Based on the estimated model, value of time savings and sample aggregate elasticities of choice probabilities are calculated.
In the second part, we estimate two transport measures’ effects on mode choices and compare their welfare impacts. One is the “low-fares of public transport” policy, in effect since January 2007; and the other is congestion charges, hotly debated but not in effect yet. By simulation, we predict mode choices that travelers most likely have made if there was no bus fare reduction and if there were congestion charges. By comparing the predicted mode choices with and without “low-fares of public transport” policy, the effect of the policy on mode shares and its distributional welfare impact are measured. Similarly, the potential effects of congestion charges are predicted by simulation.
This project will contribute to our understanding of how travelers respond to price changes, given the unique transportation conditions in Beijing that the buses are incredibly highly packed and the bus fare is very low. Another contribution of this paper is that it includes Walk, Bike, and Motorcycle in the choice set, besides other common transport modes. Not to put them into the choice set will lead to an overestimate of the effect of the reduction of public transport prices on decreasing car driving, and hence possibly make misleading policy effectiveness conclusions, given that Walk, Bike, and Motorcycle are popular modes for low income and middle income groups in Beijing. Including both work and nonwork trips in the analysis is another contribution of this paper. Most previous research focused on the morning commute to work, which is not sufficient for policy analysis, as behavior for work and nonwork trips differ and will be affected by the bus fare reduction and congestion charges differently.
