Why Good Alternatives Still Can't Get People Out of Their Cars
Key Takeaways
- The transport desirability line provides a simple, visual way to explain why mode shift happens and why it doesn’t.
- People choose the transport mode that feels most desirable to them, not necessarily the one that others might consider objectively the best.
- Investing in a new public transport option only produces mode shift from cars if it clears the desirability bar set by the car.
- Riyadh's metro succeeded because driving was already so unpleasant that the bar was low, not because the metro was uniquely brilliant.
- Most passengers on new metro lines in London and Sydney came from other public transport, not from cars, because the metro was not more desirable to many people than the car.
- Free fares increase ridership, but the new trips come overwhelmingly from walkers, cyclists and existing public transport users, not car drivers.
- Building more road capacity raises the desirability of driving, induces new demand, and typically returns congestion to its previous level within a few years.
- Congestion charges and parking charges work because they lower the desirability of driving rather than just raising the desirability of the alternative.
- Even when public transport genuinely improves, habitual car users may not notice or believe it, which is why behaviour change programmes matter, but only once the product is good enough.
- Metro lines that appear to produce modest mode shift at opening may prove transformative over decades, but only if land use policy encourages dense, transit-oriented development around stations.
What next?
How are you thinking about mode shift, and how are you explaining it to decision makers?
Introduction
Mode shift away from cars sits at the heart of almost every major transport goal: reducing congestion, cutting emissions, improving safety, and building more liveable cities. And yet, despite decades of investment in public transport, shifting people out of their cars has proven stubbornly difficult. The evidence, at first glance, looks contradictory.
Take metros. Anyone who has spent time in transport planning knows that a well-run metro is about as good as public transport gets: fast, frequent, reliable, and accessible. Exactly the kind of service that advocates argue will persuade people to leave the car at home.
In Riyadh, that is precisely what happened. When the city opened its new metro, the cynics, who had argued that Saudis would never give up their cars, were proven decisively wrong. Patronage has been remarkable, and large numbers of people have genuinely switched away from driving. A vindication for metros everywhere, surely?
Not so fast. The Elizabeth Line in London and the new metro lines in Sydney have both been celebrated as outstanding successes, and by many measures they are, with patronage comfortably exceeding forecasts. But look more closely at where those passengers have come from, and a less flattering picture emerges. The vast majority have switched from other public transport, buses, other rail lines, rather than from cars. For all the billions of pounds and dollars invested, the actual impact on driving levels has been modest.
Free fares are also a puzzle. The theory is intuitive: if cost is a barrier, remove it, and people will switch. But research from across the world mostly shows that when fares are abolished, the new trips generated come overwhelmingly from people who were previously walking, cycling, or already using public transport. Car users, by and large, stay in their cars.
So what is actually going on? Why does the same type of intervention produce a transformative mode shift in one city and barely register in another? And what does it actually take to get people out of their cars?
This blog sets out a simple framework for thinking about these questions. The aim is to provide a clearer picture of how mode shift works that can be easily communicated to non-experts.
How Engineers Think About Mode Shift
Before setting out a simpler framework, it is worth briefly understanding how transport engineers formally approach the problem.
When engineers need to forecast how many people will use a new bus route, an upgraded metro line, or a new cycling lane, they turn to a mathematical tool called the Multinomial Logit Model. The model assigns a "utility" score to each available transport mode, a numerical representation of how attractive that option is for a given journey. It then calculates the probability that a traveller will choose each mode based on the relative utility scores across all available options.
For all its sophistication, the model has real limitations when it comes to communicating with the people who ultimately make transport decisions. The outputs are probabilities and forecast volumes, not intuitive explanations of human behaviour.
The result is that transport engineers often find themselves in a difficult position. They have powerful tools for modelling mode shift, but they are poorly equipped to explain it to decision-makers.
What is needed is a way of thinking about mode shift that is honest about the underlying logic, captures what the models are really telling us, and is expressed in terms anyone can engage with. That is what the transport desirability line is designed to do.
How To Think About Mode Shift Without The Maths
The transport desirability line is a simple idea. For any given journey, every person has a perceived most-preferred transport option, the mode they would choose if they were making a clear-headed decision based on everything they know and feel about their available options. That perceived preferred option is their desirability line.
What goes into that preference? A wide range of factors, and they vary considerably from person to person. Journey time matters, but so does reliability. Safety is a factor, both personal safety on public transport and street and road safety for drivers. So is comfort, cost, frequency, and the ease of the journey at either end. Whether you are travelling with children, need to make stops along the way, or have luggage, all of these shift the desirability line. So do less tangible things: the status associated with different modes, the culture of the community you live in, and the simple force of habit.
Let’s, for argument’s sake, say that the most desirable mode for a journey is a car, as it is in many cases in our car-dominated world. We can then add desirability lines for other modal options:
In this case, the perception of our individual is that the non-car options are less desirable
A key insight is that the desirability line is about perception. A journey that is measurably faster by public transport may still feel less desirable to someone who perceives driving as more convenient, more comfortable, or more consistent with how people like them get around. This distinction between actual and perceived desirability is an important and overlooked aspect of mode shift, which I will cover a bit more below.
Once you accept that people choose based on perceived desirability, the logic of mode shift becomes much clearer. Getting someone to switch modes requires one of two things: either you raise the perceived desirability of the alternative above that of the car, or you lower the perceived desirability of driving below that of the alternative. Investments that do neither, however well-designed or well-intentioned, will not move the needle.
This is the framework that the rest of this blog applies to metros, free fares, road widening, congestion charges and habits. The graphs that follow are not precise forecasts; they are illustrations of the underlying logic, designed to make the key dynamics visible.
How this applies to our metro lines
With the desirability line framework in mind, the contrast between Riyadh, London, and Sydney becomes much easier to explain.
In London and Sydney, the new metro lines are genuinely excellent products. They are fast, frequent, and reliable, and passengers have noticed, turning up in large numbers. But for most of those passengers, the new metro was not competing with the car. It was competing with the bus or with an older, slower rail line. For those travellers, the new metro easily cleared the desirability bar. For car users, it largely did not. The new metro lines are better than what came before, but they are not better enough to displace the car as the most desirable option for many people.
Riyadh tells a different story, and the reason has less to do with the metro's quality than with the car's starting point. Driving in Riyadh is an unpleasant experience. Traffic congestion is severe, road safety is poor, and the city's layout makes many journeys long and stressful. The car was not highly desirable in absolute terms; it was simply the least bad option in the absence of any meaningful alternative. When the metro arrived, it did not need to clear a high bar. It just needed to be better than the driving experience, and it was, comfortably. The result was a rapid, large-scale mode shift.
The lesson is not that metros work in some cities and not others, or that Riyadh's experience cannot be replicated. The lesson is that the impact of any new service depends critically on where the desirability of driving starts. In cities where driving is genuinely difficult, expensive, or unpleasant, a high-quality alternative can tip the balance on its own. In cities where driving remains convenient and deeply ingrained, it almost certainly cannot, at least not without additional measures that address the other side of the equation.
This has practical implications for how we evaluate metro proposals and set expectations for what they will deliver. A metro that fails to produce significant mode shift away from cars is not necessarily a failure; it may be doing exactly what the desirability framework would predict. The question worth asking before the ribbon is cut is not just "is this a good product?" but "is it good enough to beat the car for the people we are trying to reach, given how desirable driving currently is for them?"
Now, let’s look at some examples of how this applies to various transport policies.
What About Free Fares?
Free fares have an intuitive appeal that makes them perennially attractive to politicians and advocates alike. If cost is a barrier to using public transport, the logic runs, then removing that barrier should unlock a wave of new riders, including, crucially, people who would otherwise drive. It is a clean, simple theory. The evidence, unfortunately, is less encouraging.
Research from cities that have experimented with free fares tells a common story. Removing fares does increase public transport use. But the new trips generated often come overwhelmingly from people who were previously walking, cycling, or already travelling by public transport. The car users, by and large, stay in their cars.
The desirability line framework explains why. For most people in most cities, cost is not the primary reason they choose to drive. Journey time, convenience, reliability, comfort, the ability to carry things or travel with children, the absence of a need to plan, these factors typically weigh far more heavily in the calculation than the price of a ticket. Lowering fares raises the perceived desirability of public transport, but usually not enough to clear the bar set by the car. The gap between the two lines narrows, but it does not close.
There is a further complication. When fares are removed, and more people use the network, services can become more crowded. Crowding reduces comfort, which in turn reduces desirability, partially offsetting the benefit of the fare reduction. In some cases, perceptions of safety decline as well. The intervention that was meant to make public transport more attractive ends up making it less so in other respects.
None of this means that fares policy is irrelevant. For people on low incomes, for whom cost is genuinely a significant barrier, targeted fare reductions can make a real difference to access and social equity, which are important goals in their own right. But if the objective is mode shift away from cars, free fares are an expensive tool that the evidence suggests will largely miss its target. The money spent abolishing fares for everyone, including existing passengers who would travel regardless, might achieve far more mode shift if directed toward measures that improve public transport in ways that raise its desirability above that of the car, for example, by expanding services where they are currently poor or non-existent.
Induced Demand
If the case for building new public transport is sometimes oversold as a solution to car dependency, the case for building new road capacity is arguably undersold as a cause of it. The phenomenon known as induced demand is one of the most persistently ignored in transport policy.
The basic observation is straightforward: when you build more road capacity, traffic expands to fill it. New lanes attract new trips. Journeys that were previously suppressed because the road felt too congested become worthwhile again. People who previously travelled at different times, by different routes, or by different modes, shift back to driving. Within a few years, sometimes sooner, the new capacity is absorbed and congestion returns, only now with more lanes.
The desirability line framework makes part of the mechanism visible. Before the road is widened, driving and public transport may sit at similar desirability levels for some travellers; congestion has made the car less attractive, and public transport has become a reasonable alternative for at least a portion of trips. When the road widens, the desirability of driving jumps sharply upward. For people who were on the margin, using public transport reluctantly, or considering switching, the calculation shifts, and they choose the car.
Over time, as induced demand fills the new lanes and congestion rebuilds, the desirability of driving gradually declines from its post-widening peak. The road is wider, more cars are using it, and the baseline traffic level is higher than before. The net result is more driving, not less and public transport, starved of the passengers it briefly attracted when congestion was at its worst, is left in a weaker position than it was before the road was built.
Transport Habits
Of all the factors that shape mode choice, habit may be the most underappreciated. People are not, in practice, rational calculators. Most of us do not reassess our travel options each morning with fresh eyes, weigh up the available alternatives, and choose the most desirable one. We do what we did yesterday. And the day before. Travel behaviour, once established, tends to be remarkably sticky.
This creates a problem that the desirability line framework helps to illustrate. Even when the objective desirability of an alternative rises, because a new service has launched, journey times have improved, or a new route has opened, habitual travellers may not notice, or may not believe it. Their perception of the alternative is anchored to how it used to be, not how it is now. The desirability line for public transport may have genuinely moved, but in the minds of habitual car users, it has not. They are, in effect, making decisions based on outdated information.
This gap between actual and perceived desirability is where behaviour change interventions come in. Personalised travel planning, targeted marketing, free trial periods, and employer travel schemes are all designed to do essentially the same thing: to coax people into trying an alternative they have been ignoring, in the hope that direct experience will update their perception and shift their desirability line for good. And the evidence suggests that when these interventions are well-targeted, they can work, producing mode shift that persists long after the intervention ends.
But there is an important caveat. Behaviour change only works if the underlying product is genuinely good enough. If someone tries public transport because of a well-designed marketing campaign but has a poor experience, you will not achieve mode shift; you have confirmed their existing perception. The habit may even become harder to break the second time. Behaviour change tools are effective at closing the gap between actual and perceived desirability. They cannot manufacture desirability that does not exist.
The practical implication is that behaviour change and infrastructure investment need to work together, and in the right order. Get the product right first. Then invest in helping people discover it.
What about sticks?
So far, the interventions we have examined have largely focused on making alternatives to the car more attractive. These are the carrots of transport policy. They are politically popular and easy to communicate. They are also, as we have seen, frequently insufficient on their own.
The politically unpalatable truth that the desirability line framework makes plain is that in cities where driving is deeply entrenched, supported by decades of car-friendly infrastructure, abundant free parking, and cultural norms built around the car, no amount of carrots is likely to close the gap. The car's desirability is simply too high, and the gap between it and the alternatives too wide, for investment in those alternatives alone to tip the balance for most people. To achieve meaningful mode shift in these cities, you need to reduce the desirability of driving. You need sticks.
Sticks come in various forms. Congestion charges, parking charges, and reduced parking spaces all make driving less convenient, more expensive, or both. In the language of the desirability line, they push the car's line downward. And when that line drops far enough, ideally to a point where a decent alternative is already available, people switch:
The evidence for this is strong. London's congestion charge, introduced in 2003, produced an immediate and sustained reduction in traffic entering the charging zone, with a significant proportion of displaced car trips transferring to public transport. New York's congestion pricing scheme, though more recent, has shown similar early results. In both cases, the critical ingredient was not just the charge itself, but the combination of the charge with an existing public transport network that was good enough to absorb the displaced demand. The stick worked because there was already a viable alternative for people to switch to.
In Riyadh's case, the metro was so far above the car's desirability that no stick was needed at all, but Riyadh's starting point was unusual. In most cities with entrenched car dependency, the honest assessment is that sticks will be necessary, and that pretending otherwise will not work.
Secondary Impacts
It would be unfair to leave the assessment of London and Sydney's metros where the earlier analysis left them. The observation that most of their passengers have come from other public transport rather than from cars is accurate, and it matters for how we set expectations and evaluate outcomes. But it is not the whole story.
New metro lines do not just serve existing travel patterns; they reshape them. And the most significant way they do this is not by changing how people travel today, but by changing where people choose to live, work, and invest over the decades that follow. A metro line that opens today will influence the location decisions of households, employers, and developers for a generation or more. Housing built along a well-served corridor attracts residents who are more likely to use the metro for their daily commute than if they had settled elsewhere. Offices and retail that cluster around stations generate trips that, by virtue of where they originate and terminate, are far more likely to be made by public transport than by car. Over time, the cumulative effect of these location shifts can be substantial, not because existing car users have switched, but because the city has gradually reorganised itself around the network.
In this sense, the mode shift case for a new metro line is less about the snapshot at opening day and more about the long arc of urban development it enables and encourages. A line that appears to produce modest mode shift in its first few years may, over decades, prove transformative, provided it is accompanied by the land use policies and development patterns needed to realise that potential. Stations surrounded by car parks and low-density development will not deliver the same outcomes as stations surrounded by dense, mixed-use neighbourhoods with good pedestrian and cycling access. The infrastructure is necessary, but it is the development that follows, and the planning decisions that shape it that determine whether the long-term mode shift case is realised.
So the honest assessment of London and Sydney's metros is not that they have failed in getting people out of their cars; it is that their most important impacts have not yet been fully felt.
Conclusion
The transport desirability line is a simple idea, but it cuts through much of the confusion about why mode shift is so hard to achieve.
To shift modes, you need one of three things, or ideally all three. You need to raise the desirability of the alternative above that of the car. You need to lower the desirability of driving. Or you need to change people's perception so that they actually recognise the better option in front of them (if it exists).
Building a great metro is necessary, but not sufficient. Fares policy matters, but it is rarely the binding constraint. Sticks, such as congestion charges, parking charges, and reduced parking, are uncomfortable politically, but the evidence shows they work. And habits and culture mean that even when the numbers stack up, you may still need to help people make the switch.
For decision makers, the practical takeaway is this: before committing to a major transport investment, ask whether it will actually clear the desirability bar for the people you are trying to reach. If it won't, because driving is simply too convenient, too cheap, or too culturally ingrained, then the investment alone will not deliver the mode shift you are hoping for. You need to think about the whole equation, not just one side of it.