Planning changes we need to make, and approaches to forecasting that we need to consider: What’s at stake

I’m a planner.  Like many other planners, I’m looking at the longer view and systemically, what is happening.  Our training could be better in terms of planning for shifts though and analyzing what is happening.  That’s why I was particularly grateful to connect up with Jamie Arbib and Tony Seba of RethinkX.  Below is an adaptation of a background appendix we were working on for planners, policymakers, investors, businesses and other decision makers.  Jamie wrote the first draft of this. Thank you both for your great work, that is educating so many and prompting better planning and preparation for the public sector (rather than just assuming things will continue — often assuming trendlines — as they always have).  We are discussing and thinking about some of this among the Strategic Management Committee and others at the National Academies Transportation Research Board – MV

General mistakes forecasters make

Exponential projection is widely understood.  Examples include growth, emissions, deforestation or pollution, but planners and those involved in projection in for the financial sector and media often miss the exponential nature of solutions. The solutions – technologies – are adopted in S-curves. These are driven by feedback loops, network effects and tipping points, which act as powerful accelerators to adoption and drive non-linear pathways. Furthermore, technology convergence leads to disruption points where suddenly new business models or products become possible – and change happens fast. Just as the smart phone was enabled by a number of technologies reaching a point in price and performance, so too — and soon— autonomous vehicles will be feasible as artificial intelligence, sensors, communications and computer processing come together to enable it. This will create a disruption point where the costs of mobility plummet, Mobility/Transport-as-a-Service becomes viable, and economic realities make rapid adoption inevitable, even if the timing of this inflection point is uncertain.  With technology disruption playing a much larger role than when planners were trained, this is a new factor that must be incorporated now.

Misunderstanding the systemic and interrelated nature of change

The second mistake most forecasters make is to assume that technology improves in cost and utility at an exponential rate, but that “all else remains equal.” In fact, everything is dynamic and interdependent.  This mistake means forecasters see factors as constants and not variables, or ignore the systemic implications of this dynamism across society. For example, they view the public’s skepticism of autonomous vehicles (based on surveys conducted today) as a constant brake on adoption. This fails to recognize that over time, through increased exposure to AVs and to information about them and their ability to save time chauffeuring for kids or parents, public opinion might come to view human drivers as reckless and unsafe in comparison. If every fatal accident involving a human driver comes to be seen as an unnecessary and avoidable tragedy, political pressure to accelerate adoption of AVs is more likely to grow, and change will accelerate.

Similarly, mainstream forecasters fail to recognize how business incentives and metrics change when you begin changing variables, such as moving to a new ownership model (MaaS/TaaS). Reducing travel cost per mile (especially at larger percentages of household income) becomes the new economic driver, and car manufacturing moves from planned obsolescence to longevity and a circular model. This can drive further cost savings, and create new business opportunities built on a  low-cost transport infrastructure — just as the internet created untold business opportunities based on low-cost communications infrastructure, like Amazon, Google, and Facebook. It might be that the winners are those who can develop new revenue streams to subsidize the costs of mobility and profit from them (for example, on-board advertising, entertainment, grid services, data monetization and product and services sales). Early exposure to this new system is critical for future participants, given the network effects (and hence early mover advantage) of many of these new models. Failure to lead this transition will mean that the Fords, GMs, Exxons and Googles of the future are created elsewhere.

Assuming an energy transition and not a technology disruption

Some claim that this is an energy transition, and that past analysis of energy transitions “prove” that they take 30 years to get to 1% of the energy mix, and decades more to reach materiality. There are many problems with this analysis, including the fact that it looks at figures on a global scale when most previous transitions have happened locally and then gradually spread out. But most importantly, it misdiagnoses the issue. Energy transitions are slow because they require the associated infrastructure to be built. To replace horses with cars, for example, you needed roads, gas stations, oil wells, refineries and pipelines as well as the production capacity and supply chains for cars. For the rollout of electricity, you needed power stations, distribution networks and cables to every home. But for the coming TaaS transition, the roads are already built, electricity already serves every part of the country and every household, and the vehicle-production supply chains already exist. Fewer car-size vehicles are likely to be produced, not more.  Less road space will be required, if planning and supportive requirements for data-sharing and ridesharing occur. Rather than resembling a traditional, gradual energy transition, the coming changes in transportation will represent a technology disruption, and will therefore move along an S-curve (as discussed above) far faster and further than a mere change in fuels.

Mistakes forecasters make that are specific to AVs

Our forecasts differ dramatically from mainstream thinking. We foresee that 10 years from the disruption point — that is, from the widespread regulatory approval of AVs — 95% of US passenger miles will be travelled in autonomous shared electric vehicles in a model we call Transport as-a-Service (TaaS). It is important to understand why analysis of intersecting technology disruptions and consumer uptake of technology produces different results from ordinary/older planning and analysis methodologies.

Modelling the wrong disruption

Most forecasts model only the disruption of the gasoline vehicle by the electric vehicle (EV) in a like-for-like, one-for-one substitution. These forecasters assume the cost savings and utility improvement of EVs over internal combustion engine (ICE) vehicles are minor, so there is only very slow adoption, as ICE cars are only replaced gradually in new car sales over the course of decades. However, a more meaningful disruption is likely to arrive before this first disruption is anything like complete.  That business and ownership model disruption – with Mobility/Transport-as-a-Service (TaaS) replacing car ownership (of both ICE AND EVs). With TaaS, the cost savings are dramatic, and undercut both new car sales AND the cost of keeping the existing fleet on the road. TaaS travel will be at least twice as cheap as operating a new car; used cars, even if given away, won’t be competitive on price. Existing vehicles will be increasingly less used and ultimately abandoned, totally transforming the fleet on the road.

Under-estimating cost savings of TaaS

The scale of the cost savings that will drive fast adoption of TaaS are not broadly appreciated. These cost savings come from a reduction in vehicle degradation which leads to lower maintenance costs and longer lifetimes. We conservatively base our cost assumptions for maintenance and vehicle lifetimes on what is possible now. Because an EV has only 20 moving parts compared to 2,000 for a car with an internal combustion engine, there is less friction, and there is far less to go wrong. There is also less heat and vibration that cause degradation. This means that electric vehicles can last for 500,000 miles vs. 150,000 miles for an ICE vehicle, and cost 80 percent less to maintain. Spreading the initial vehicle purchase cost over more miles driven leads to a dramatically lower cost per mile. However, we would expect that the change of business model which delivers a new economic metric — where cost per mile is the number that matters, rather than upfront cost —  will create new business incentives that deliver even greater savings than we model. We also believe EV lifetimes will improve to one million miles, while the cost of maintenance will hover at about 10% of what it costs to keep a gasoline-powered vehicle in good shape.

Mainstream forecasts are generally obsessed with the cross-over point in EV vs. ICE upfront costs or lifetime costs, which they see happening in the 2020s. However, this misses the point entirely. Cost per mile is the new key metric; upfront cost is only one factor in its calculation. Of far more importance is vehicle lifetime, and EVs offer a dramatic improvement (to 500,000 miles immediately, increasing to one million miles as incentives move from planned obsolescence to low cost per mile). Further savings come from reduced insurance costs (fewer accidents and no theft), reduced finance costs (greater capital efficiency as 10x more miles are covered per vehicle per year), and lower fuel costs (electric motors are far more efficient and electricity is cheaper than gasoline per joule). Economics is the main driver of this disruption, and there are few barriers in terms of supply or demand. The speed of adoption will be driven both by the huge cost savings (a large proportional increase in disposable income) and the systems dynamics that kick in as adoption begins.

Ignoring Systemic and dynamic factors – all else is NOT equal.

Once adoption unfolds, systemic dynamic factors drive it ever faster. But most analysts fail to account for the feedback loops, tipping points, network effects and other dynamics that drive the non-linear S-curve shape of adoption.  Everything is dynamic and interdependent. The factors that influence key stakeholders, including consumers, policymakers and businesses, are constantly changing and are inter-related. Changes to any variable ripple through the system, impacting all others.

Another example of a feedback loops is the potential death spiral of the gasoline car. Once an early tipping point in M/TaaS availability is reached and people in cities realize that a car is always available when needed, we can expect people to begin to sell their used vehicles. As used-car supply increases (and demand drops), used car prices will drop. It will become cheaper for potential buyers of new gasoline cars to buy used instead (if they don’t move to TaaS). This will cause new car sales to drop. Economies of scale in car manufacturing will begin to unwind, increasing the cost of ICE vehicles and reducing sales further. R&D in ICE will stop as manufacturers focus on TaaS vehicles (ICE drivetrain designers are already seeing lack of demand for new models), meaning ICE cars won’t develop and will become less attractive. Eventually the supply chain will break down, and ICE sales will cease entirely.

Later in the cycle of adoption, as another tipping point is reached, gas stations will begin to close, spare parts will dry up, insurance premiums will rise (as human drivers are perceived as reckless) and using an ICE car will become ever more expensive and difficult. This might be accelerated by policy as human drivers are confined to special lanes or restricted to certain times of day, or even banned completely. ICE cars left in the fleet might also be banned or restricted as the number of ICE defenders diminishes and the political space to ban ICE vehicles opens up (as the UK and France recently set dates for, in confidence it will actually happen much earlier).  The number of countries signalling these types of restrictions will increase, and the dates proposed will move up.

The importance of proper forecasts, and the risks of being wrong

Self-fulfilling nature of forecasts

Forecasts can become self-fulfilling. This happens for a number of reasons. Decisions made without sight of a better alternative can lock in a high-cost, uncompetitive infrastructure that makes a change of course harder at a later date. Short-sighted decisions can also use up money and political capital that could have been used to accelerate the transition to M/TaaS, rather than encouraging investments that create barriers to change. Furthermore, regulation is critical to both the implementation and the development of autonomous vehicles, which are the critical enabler of M/TaaS. In terms of implementation, regulation is needed to approve the use of AVs and clear barriers to adoption, and potentially to further accelerate and regulate M/TaaS to ensure monopolies don’t abuse power and take most of the cost savings. Agencies also must ensure universal access. Regulation is also key to development, because AVs learn by doing – the more cars on the road and the more miles covered, the faster they learn and the faster they can reach the critical safety threshold. This learning requires the ability to practice, and hence the need for real-world pilots. The more and the broader pilot programs are, the faster the necessary learning will occur. Public opinion and expectations can also affect the pace of adoption. Addressing concerns and communicating a bigger vision can help create the political space needed to accelerate adoption.

Forecasts from Shell, BP, the IEA and other mainstream consultants and analysts tend to presume that transportation disruption will happen gradually as the EV replaces the gasoline car over many decades. This misses the more important disruption, which is Mobility/Transport-as-a-Service based on autonomous electric vehicles (robo-taxis) replacing car ownership entirely. This is a business model disruption and can happen far faster – within a decade – with far greater benefits for the economy and society than is generally realized. It is very easy for consultants and businesses to hide behind this common viewpoint and not challenge the common wisdom; in fact, their business incentives dictate that it is safer to be close to what others are saying. The danger for policymakers is that they may take these forecasts as correct and allow them to become self-fulfilling, thus missing out on the opportunity to lead, implement, and see key benefits in the transition. As Warren Buffett says, “You pay a high price for a cozy consensus.”

So how do policymakers balance the need to invest in the current system, until at some point in the future, when a new infrastructure and system is required? That’s the central challenge. As the speed of progress heats up, this balancing act becomes more difficult. Some find the safest course, career-wise, is to base decisions on the overwhelming consensus. As the saying goes, nobody ever got fired for choosing IBM; however, there is real danger of getting left behind and there is a real and urgent need for courage and vision.

RethinkX is established as a not-for-profit, and has no intention to profit from our analysis. Our analysis is provided as a public good, and we hope decision makers can use it to make better choices. We are very happy to clarify any issues raised here, or to engage with policymakers free of charge (subject to time constraints!).

About RethinkX

RethinkX is an independent think tank that analyzes and forecasts the speed and scale of technology-driven disruption and its implications across society. We produce compelling, impartial data-driven analyses that identify pivotal choices to be made by agencies, investors, businesses, policymakers and civic leaders.

RethinkX provides evidence-driven systems analysis that helps decisionmakers who might otherwise have to rely purely on mainstream analysis. Decisions made based on the latter risk locking in investments and infrastructure that are sub-optimal, and that make societies poorer by locking them into expensive, obsolete, uncompetitive assets, technologies and skill sets.

We focus on understanding the dynamics and the systemic nature of disruption across key market sectors, according to a highly evidence-based approach. This approach is designed to facilitate decisions that maximize the benefits and minimize the costs – economic, social, and environmental – of technology disruption.

Rethinking Transportation is the first in a series of reports that analyzes the impacts of technology-driven disruption, sector by sector, across the economy. A copy of the report is available for download free of charge on our website.

We invite you to learn more at