Effects of Travel Time Reliability and Commodity Characteristics on Hinterland Leg Transportation of Export Containers

Abstract

Increasing the mode share of railway in hinterland leg containers transportation requires a better understanding about the effects of critical factors on shippers’ mode choices. This paper focuses on the effects of travel time reliability (TTR) and commodity characteristics on freight mode choice. A two-stage survey is conducted in the Yiwu-Port of Ningbo corridor, China, to collect shippers’ preference data. Five model specifications are estimated using these data. Estimation results of generic parameters indicate that significant interaction effects between commodity characteristics and travel time exist. The value of generic reliability was then calculated and the effects of commodity characteristics was quantified. In addition, mode-specific values of reliability are estimated. Remarkable differences are found in the mode-specific value of reliability for different modes. Also, the effects of mode-specific value of reliability on the demand forecasting were investigated. Results imply that the mode share of railway will be underestimated if the mode-specific value of reliability is neglected, especially when travel time of railway transportation is reliable. Therefore, it is recommended that the mode-specific willingness-to-pay should be considered in railway demand forecasting and project appraisals.

Accounting for Heterogeneity of Travel Time Reliability Valuation in Shippers’ Freight Mode Choice

“…To accommodate preference heterogeneity in model estimation, several model structures have been explored ( 24 ). These models include the mixed logit (ML) model ( 10 , 13 , 19 ), the latent class model ( 25 – 29 ), and the heteroskedastic multinomial logit model ( 30 ). However, the majority of the literature has focused on the mean value of VOR, so as to obtain a better fit of the data ( 31 ).…”

“…This may be a result of its explicit meaning, making it easier for respondents to understand and make trade-offs between attributes. Nevertheless, more recent studies preferred to use the mean-variance based approach, which use SD to measure TTR ( 8 , 10 , 13 , 18 , 19 ). The reason is that VOR derived from this approach is compatible with the existing freight demand models and the cost–benefit analysis framework ( 10 , 20 ).…”

“…Heterogeneity in VOR has been found in plenty of empirical studies ( 10 – 15 ). To accommodate preference heterogeneity in model estimation, several model structures have been explored ( 24 ).…”

“…They attributed these variations to geographic characteristics, the regional economy, and others. However, the effects of observed characteristics, such as company characteristics and shipment characteristics, on VOR heterogeneity are not included in studies ( 13 ). This limits the effective segmentation of the market demand based on shippers’ preference heterogeneity, which is important for railway operation companies to design effective strategies targeted at specific shipper groups.…”

“…Many empirical studies have been conducted to quantify shippers’ willingness to pay (WTP) for TTR improvements (namely the value of reliability, VOR) ( 6 – 10 ). As for hinterland leg transportation, several papers have even indicated that shippers’ VOR (using standard deviation to measure TTR) is higher than the value of time (VOT, defined as shippers’ WTP for the travel time savings) ( 11 – 14 ). This implies that it is necessary to incorporate TTR in mode choice behavior analysis for hinterland leg transportation.…”