Strategic Transport Models are relevant but need to be better understood

There has been a lot of commentary recently on LinkedIn on the relevance or otherwise of Strategic Transport Models (STMs). It is appropriate yet again to revisit some of the well-trodden arguments in defence of the value of STMs, but with the maturity to recognise their limitations and hence to use them judiciously in contributing to our understanding of how tested initiatives impact the key performance outputs being investigated. Importantly, STMs are not just about forecasting but increasingly on an understanding of behavioural relationships today that can be used to assist in forming a position of what future settings might look like. We often describe this as ‘Vision (or decide) and Validate’ in contrast to ‘Predict and Provide’. Carefully crafted STMs have the ability to satisfy both approaches.

Strategic Transport Models: Powerful Tools or Dangerous Weapons in the Wrong Hands?

Prediction is very difficult, especially if it's about the future! - Niels Bohr

Strategic models have long been a cornerstone of transport planning, but they are increasingly scrutinised for their perceived inability to capture complex, dynamic realities. Numerous recent works have challenged the utility of these models, pointing to inherent limitations such as their reliance on historical data and assumptions of static behaviour. While these critiques are not without merit, they risk undermining the significant value that transport models bring to planning processes when used appropriately.

What is the role of transport models?

Transportation is a derived demand shaped by complex human behaviour, making it inherently difficult to model with precision. Transportation models are not just tools for forecasting demand; they also offer critical behavioural insights into the transportation system. For instance, by analysing elasticities (such as responsiveness to changes in fares, travel times, crowding, and urban design) and quantifying willingness to pay (e.g., for time savings or reliability), models provide a nuanced understanding of how people adapt to varying conditions. More importantly, models can address "what-if" questions, providing planners with a perspective on the initiatives that warrant detailed consideration [1].

Why are models being criticised?

Traditional transport modelling has relied on the "predict and provide" approach that passively forecasts travel demands by extrapolating historical trends and behaviours into the future. This approach risks keeping us trapped in a vicious cycle of ever-expanding road capacities, ultimately reinforcing the status quo. This mindset is now being replaced by a ‘Vision and Validate’ approach, which is a proactive, outcome-focused strategy, supported by new planning tools to shape demand rather than merely reacting to it. Vision and Validate overcomes the traditional limitations by encouraging communities to envision their future (after all, the best way to predict the future is to create it together!) and test policies against desired outcomes [2].

Transport models face several key criticisms that highlight their limitations in aligning with the ‘Vision and Validate’ approach. Traditional models often rely heavily on historical data and static assumptions (falling victim to the ‘ceteris paribus’ fallacy—the assumption that all other factors remain constant), failing to account for dynamic societal, technological, and behavioural changes, such as induced (or reduced) demand and real-time responses, leading to flawed projections. In many cases, behavioural nuances such as willingness to pay and time-of-day variations are frequently oversimplified or omitted. Also, the effectiveness of any model can be constrained by skill gaps among analysts, which can lead to misinterpretation and poor decision-making.

What can be done?

The good news is that, while the above criticisms are valid for older models (dare we suggest the very conventional Vanilla flavour 4-step models), they do not reflect the state of modern transport models. Advances in modelling techniques now allow for greater flexibility and responsiveness to changing conditions. For example, scenario-based modelling can account for potential shifts in travel behaviour, such as increased remote working, the adoption of electric vehicles, or the impacts of pricing mechanisms like congestion tolls. Dynamic feedback loops—such as those that capture induced (or reduced) demand or mode shifts—are increasingly incorporated into contemporary models, reducing reliance on static assumptions [3].

However, it is true that the quality of a model’s output depends on the quality of its inputs. It is also worth noting that the effectiveness of models depends as much on the clarity of the questions posed as on the inclusion of behavioural responses such as feedback effects, and the validity of assumptions [1].

Critiques of transport models often reveal systemic issues, such as errors in land-use projections or population forecasts. Improving their accuracy and relevance requires using diverse, up-to-date data sources, including real-time traffic data, GPS insights, and behavioural studies, to better inform assumptions. Models should be embedded within broader decision-support systems that integrate qualitative insights from community and stakeholder engagement to align decisions with societal objectives. Context-specific calibration is critical, as high-level generalisations often lead to misleading results and poor decisions. Transparency is equally important, with clear documentation of assumptions and limitations to enable critical interpretation by practitioners and the public. Ultimately, transport models must complement, not replace, human judgment, integrating quantitative analysis, qualitative interpretation, and stakeholder input to address the complexities of transport planning effectively.

Final remarks

Transport models are not infallible, but their structured approach to analysing transport systems and testing policy interventions makes them indispensable in enabling effective planning and allocation of public funds. For instance, transport models provide a necessary input into cost-benefit analyses (CBA). Hence, transport models cannot be avoided as a standardised tool to compare current and future scenarios that require some kind of benchmark.

Criticisms of outdated practices should inspire innovation, not rejection. By evolving transport models to align closely with the Vision and Validate framework, and maintaining transparency, practitioners can ensure these tools remain central to creating effective and sustainable transport systems. Therefore, the solution to the current shortcomings is to improve practices and build better models to guide our decisions (after all, an educated guess is better than stabbing in the dark), as well as train analysts and decision-makers to better understand the role of STMs.  Otherwise, we risk throwing the baby out with the bath water!

References

[1] D. Hensher, "What is the value of strategic transport modelling?," presented at the NSW Transport Planning Network (Planning Institute of Australia) webinar, Oct. 12, 2020. 

[2] P. Jones, "Visions require validation: new approaches for deciding and testing which policies will deliver," Tapas. network, 2023. [Online]. Available: https://tapas.network/41/jones.php. [Accessed: December, 2024].

[3] S. P. Shepherd, "A review of system dynamics models applied in transportation," Transportmetrica B: Transport Dynamics, vol. 2, no. 2, pp. 83–105, 2014.