Does COVID-19 change our view of models?

Models are still based on a number of assumptions that we seldom challenge, and some of this will be turned upside down by the pandemic

It is interesting to note that the COVID-19 pandemic has forced authorities to rely on models to support their decisions to curtail personal movement, and that the reliability of these models has been put centre stage; see for example the article in The Guardian.

This rarely happens with our own transport models.

Indeed, the models used to forecast the future of the pandemic have all good provenance: Imperial College and Oxford, but both suffer from heavy reliance on assumptions that cannot be substantiated at the moment, simply because not enough is known about the virus and its transmission effectiveness. Assumptions are made and different assumptions lead to different conclusions.

This has not stopped governments making decisions based on these models that have enormous cost and quality of life implications. A shock of this scale speeds up decision-making, as the costs of delay can be very significant.

Are our transport models any better than just a set of assumptions dressed with some advanced mathematics? I would say no. Our models have been developed over many, many years and they have sufficient emphasis on calibration to fit with real, observed data. They are still not perfect, there is no such a thing as a perfect model. But it can be claimed that their grounding on real data and rigorous procedures for validation make them considerably more reliable.

But, is this really true? Does COVID-19 change this view? Our models are still based on a number of assumptions that we seldom challenge and some of this will be turned upside down by the pandemic. Consider the following questions from my LinkedIn blog:

• As more activities moved to the internet (virtual lectures, work, meetings etc.) when the pandemic is over, will they all return to physical presence? Some would have been found to be equally or more efficiently done virtually; at least some of the time. Trip/tour generation models will have to change. 

• I suspect pre-2020 leisure and business trends for international and domestic travel will never fully recover. Trends will have to be adjusted. 

• Will people change their mode preferences? Will we decide to avoid crowded trains and prefer buses; or return to the individually owned car as I know how much effort I put keeping it clean with soap and water? 

• Is this the end of the road for ride-sharing in a small vehicle? 

• It has proved more difficult to reduce contagion in large cities like London; would people re-discover their love for smaller and more local urban areas (given that remote work is possible)?

I believe it is inevitable that some parameters in our utility or generalised functions will have to change (including Alternative Specific Constants). Of course, it is too early to say, but not too early to start collecting data during and after the pandemic to get a better grasp on this very significant shock to our standard and consistent way of life.

If you have any thoughts or suggestions, please contact Luis on luis.willumsen@kineo-analytics.com