Influence of Parking on Location and Mode Choice

The survey was conducted with a sample of more than 1’200 pre-recruited respondents in Switzerland. Several characteristics of the respondents’ regular travel were recorded during a recruitment interview, which the stated choice experiments were based on. The overall response rate was over 80 percent, and more than 14’000 observations were used for the estimation of the choice models presented here. Two different modeling approaches were tested: one with a deterministic model of taste heterogeneity (using interaction terms for trip distance, duration and income, as has been the practice in earlier studies), and one applying a random parameters approach. Both models were estimated in willingness-to-pay space. The model results indicate that parking characteristics have a significant effect on mode and location choice. The sample distributions for the relevant willingness-to-pay measures and the sample average demand elasticities inferred by the two models, as well as the overall model fit statistics, are slightly different. The deterministic assessment of taste heterogeneity is preferable for application in forecasting models, as it allows the direct computation of the expected effects of a policy on market shares.本调研面向瑞士预先招募的1200余名受访者展开。招募访谈阶段已记录受访者日常出行的多项特征，以此作为后续陈述选择实验（stated choice experiments）的设计依据。本次调研的整体响应率逾80%，最终共获取14000余条有效观测数据，用于本文所呈现的选择模型估计工作。本次研究测试了两种不同的建模范式：其一为偏好异质性确定性模型（通过引入出行距离、出行时长与收入的交互项，沿用了早期同类研究的通用做法），其二为随机参数法。两类模型均在支付意愿（willingness-to-pay）空间内完成估计。模型结果显示，停车特征对出行方式与选址选择存在显著影响。两类模型推导得到的相关支付意愿指标样本分布、样本平均需求弹性，以及整体模型拟合统计量均存在细微差异。其中，偏好异质性确定性模型更适用于预测建模场景，因其可直接计算政策调整对市场份额的预期影响。