Transport and Spatial Economic Models

We all love and promote interdisciplinary research, but most of us prefer to begin with material written in the terminology of our own area of expertise. Continue reading the relevant section below depending on whether you are transport professional interested in understanding our spatial models; a transport researcher and/or a transport economist, or a spatial or urban economist. Otherwise, I encourage you to shift perspective now and then, and read the full text if time allows. 

Guide for Transport Professionals

Cities and their transport networks operate in close symbiosis. Our travel needs are shaped by the structure of cities, while urban form is strongly influenced by the cost of moving through space. This interdependence is driven by the behaviour of people and firms, especially in their location choices. The existence of this relationship is well established in theory, and in recent years we have seen a rise in related qualitative research at the intersection of urban planning, geography and transport.

However, quantifying the extent of transport-induced urban transformation remains a major challenge, both in academic research and in applied policy evaluation.

The spatial models we develop treat the transport network much like a traditional traffic simulator. What sets them apart is that we integrate the transport network with local labour, housing and production markets. We do this at a high level of spatial granularity, distinguishing thousands of locations or zones across large cities or regions.

These models are not merely predictive tools. They are grounded in microeconomic theory. Each household and firm in the model is assumed to make decisions based partly on rational calculation and partly on individual preferences. These decisions relate to where people live and work, and how they travel.

As a result, our models are compatible with cost-benefit analysis. A CBA conducted within our framework can produce location-specific benefit-cost ratios. This allows us to estimate the spatial distribution of impacts from transport policies and, in turn, to assess their likely popularity and political or social acceptance.

Our long-term vision is that these highly granular spatial models will become a complement to mainstream cost-benefit analysis, and ultimately form part of official CBA guidelines within the European Union and, potentially, other parts of the world.

Guide for Transport Researchers

Many transport researchers associate location choice and the prediction of urban form with land-use transport interaction (LUTI) or spatial computable general equilibrium (SCGE) models. Our work builds heavily on these branches of the literature. Below, we position our research in the context of three key steps in the evolution of spatial models for transport.

Step 1: LUTI refers to a broad category of models in which transport supply interacts with measures of urban form. The primary purpose of LUTI models is prediction, typically estimating the redistribution of households, who then become travellers, across geographical space. LUTI models are sometimes integrated with traditional transport models, creating combined frameworks that account for both travel demand and land-use responses.

Step 2: SCGE models are a specific type of LUTI model. Their distinguishing features include:

• Modelling the spatial distribution of housing prices, wages and, in some cases, firm productivity, based on rational (that is, profit-maximising) firm behaviour and a defined market structure such as perfect or monopolistic competition. This allows floorspace prices and wage responses to be modelled explicitly.

• A foundation in economic theory, from individual behaviour to aggregate market outcomes. This enables the derivation of coherent measures of economic performance, making SCGE models suitable for economic appraisal, also known as cost-benefit analysis, of transport policies.

SCGE models require significant investment in both data and expertise. As a result, their use in transport appraisal is still much less common than the mainstream partial equilibrium approach, which is usually based on travel time savings.

Step 3: In the past decade, a technique known as quantitative spatial modelling (QSM) has gained considerable attention in economics. QSMs are a subset of SCGE models that aim to improve the speed and accuracy of parameter calibration, while reducing sensitivity to confounding effects in the data that standard methods struggle to isolate.

Our main line of research focuses on developing QSMs specifically for transport analysis. We believe this approach will soon become as widely used among transport researchers as it already is in leading economics departments. Our long-term goal is to design a QSM framework that becomes part of the official cost-benefit analysis methodology for transport appraisal in Europe and beyond.

Guide for Spatial and Urban Economists

I develop quantitative spatial models for transport economic applications. Quantitative spatial economics (QSE) has had a major impact on spatial and urban economics. In many ways, the rise of quantitative spatial modelling (QSM) is driving a shift in academic writing as well. There has been a surge in recent publications, particularly from early-career researchers entering the job market, that combine QSMs with historical context, reduced-form empirical analysis, counterfactual policy simulations and welfare evaluations. While the intellectual value of this line of literature is hardly questionable, it also creates a barrier to the wider use of QSMs in applied contexts such as transport cost-benefit analysis. From this perspective, our research aims to make spatial economics more accessible to the global transport research community.

QSE has considerable potential to influence the transport sector. Transport involves one-off decisions on major infrastructure and regulatory changes that have lasting effects on the spatial distribution of economic activity. Transport already features prominently in many of the most cited QSM papers, including:

• Individual car use and urban form in Berlin (Ahlfeldt et al., 2015)

• Bus rapid transit and its effects on Bogotá (Tsivanidis, 2019)

• The suburban rail network in London (Heblich et al., 2020)

• High-speed rail and regional economies in Japan (Koster et al., 2024) and California (Fajgelbaum et al., 2024)

Despite transport being a popular subject, most of the QSE literature has developed separately from the transport research community. The authors of leading QSM papers rarely attend transport conferences or publish in transport journals. As a result, QSMs have remained largely unknown within the field over the past decade.

This disconnect also reflects a difference in modelling traditions. For example, the assumption of iceberg (ad valorem) transport costs, where travel disutility is modelled as a fraction of utility being lost with distance or time, is common in trade economics but rarely used in transport economics. While such assumptions are not difficult to adjust, adapting them to more realistic formulations requires research effort that has yet to be made.

This is exactly the research gap my work in spatial modelling addresses. I consider the following features essential for a transport-oriented QSM:

• The money, time and discomfort costs of travel should be treated separately. The valuation of non-monetary disutility should be grounded in the traveller’s utility maximisation problem.

• Most QSMs focus only on commuting, capturing its effect on residential and workplace location choice. Yet commuting accounts for less than half of all trips in developed cities. It is important to also include leisure and business travel.

• Congestion in road use and crowding in public transport are central concerns in traditional transport economics, but are still poorly represented in QSE. Allen and Arkolakis (2022) made progress on modelling road congestion within a general equilibrium setting. However, more work is needed to build on this and incorporate findings from decades of research on congestion and crowding.

• Public transport supply, including network layout, frequency and pricing, is typically treated as fixed in current QSMs. Future models should be able to explain observed supply patterns and anticipate how these might change in response to transformative policy.

Our ongoing research follows this direction. The goal is to make quantitative spatial modelling a practical and widely used tool for transport policy professionals.

You can find more details about our spatial modelling research here.

References