Transportation Policy and Project Evaluation

This chapter focuses on the analysis of cross-sectional regression, which conceptualizes travel behavior as influenced by both socioeconomic and constructed environment elements. Using observed cross-sectional data, such regression analysis assesses the association between independent variables and travel behavior. Complementary methods are crucial for obtaining unbiased estimators, and both instrumental variable and propensity score matching methods have been widely utilized in transportation research.

This chapter mainly introduces various forecasting models and explains them in relation to related research and applications. The first method discussed is the mainstream four-step travel demand model, which conceptualizes travel decision-making. The second method is the direct demand model, where cross-sectional regression analysis estimates travel demand. This is followed by a discussion of post-processing models that account for impacts not explicitly considered in the main models. Next, the abstract mode demand model is introduced, which assumes an abstract representation based on features that affect mode choice, such as cost, travel time, and frequency. The incremental demand model is subsequently utilized to examine the effects of variations in travel time, cost, and other transportation attributes. This analysis typically involves the use of elasticity analysis and pivot-point modeling. The final model discussed is the sketch planning model, used to analyze traffic and land use scenarios at a broad resolution level, providing general magnitude estimates. Additionally, the chapter covers the corridor model and the marginal demand model. The corridor model is characterized by simplicity, considering only the entry and exit points as the origin and destination. The marginal demand model focuses on travel demand, most likely influenced by specific projects and policies.

This chapter introduces causal models for project evaluation, including methods used to estimate counterfactual effects. Firstly, it discusses the regression discontinuity (RD) method, including fuzzy and sharp RD, and its application in transportation. Then, it covers the difference-in-differences methods, noting that panel data helps estimate causality and that difference-in-differences is widely used in transportation research. Finally, the chapter examines the synthetic control method, which creates a “synthetic” control unit for comparison with the treated unit. Synthetic control methods can also construct counterfactual outcome trajectories. However, the synthetic control method has seen limited application in transportation.

This chapter examines pertinent cases and practical applications using the models and methodologies expounded upon in the preceding three chapters. The inaugural case study concerns the UrbanFootprint transportation analysis, which furnishes a sturdy framework for land use planning, modeling, and data structuring. This enables practitioners, public agencies, and additional stakeholders to engage in more enlightened planning endeavors. The subsequent case scrutinizes using land use and transportation models in policy analysis. It involves evaluating transportation and land use policies, comparing scenario analyses under different policy conditions, and discussing their impacts on urban development. The third case focuses on the transportation planning evaluation of the metro’s regional transportation planning. Lastly, this chapter offers a temporal examination of congestion and public transportation within Urbanized Areas (UZAs) in the United States through time-series analysis. This analysis explores the temporal relationship between public transportation and congestion, providing insights into their interdependencies.