Most air travel forecasts predict a long-term rise in demand, with limited consideration of any limits to growth. However for any given population there will be those who have not flown recently, as well as those who never have flown. For the UK, about half the population respond to travel surveys that they did not fly in the previous 12 months. We call these the ‘infrequent flyers’.

Little is known about this group, including  whether they are likely to fly in the future. Anne Graham, of the University of Westminster, and I recently published findings of an analysis of the characteristics of this group and the reasons for their travel habits, using a survey commissioned by the UK Civil Aviation Authority. We found that infrequent flyers make up a heterogeneous consumer group whose non-flying is influenced more by budget constraints and personal circumstances than specific aviation factors such as fear of flying.

The proportion of infrequent flyers in the UK population has remained stable over time. Our findings do not suggest that this is likely to change in the future, so the infrequent flyers are unlikely to be a source of future demand for air travel on account of their increased propensity to fly.

Our paper: Graham&Metz JATM Infreq flyers published

I have a new book published on 1 September, one in a series of short books on policy and economics topics described as ‘essays on big ideas by leading writers’. My contribution is a critique of the inconsistencies of transport policy in recent decades, which I attribute to the shortcomings of conventional transport economic appraisal in identifying the benefits that arise from investment. Readers of this web-magazine will recognise many of the arguments, now brought together in a single volume at a modest price.

The article below appeared in Local Transport Today 699, 10 June 2016. It was prompted by discussions at a workshop event organised by colleagues at the Transport Institute of University College London, who are carrying out a study for the Department for Transport of social and behavioural impacts of autonomous vehicle.

There is much interest in the possibilities for autonomous vehicles, in particular driverless cars. Focus is mainly on technological feasibility, role of the driver, risks and insurance. What has not yet been sufficiently considered is the implications for traffic. How much difference would autonomous vehicles make?

There are two broad routes to driverless cars. Mainstream auto manufacturers are equipping vehicles with devices that assist the driver. Adaptive Cruise Control automatically adjusts the vehicle speed to keep a safe distance from the vehicle ahead. Lane Keeping systems alert the driver if the car is drifting out of its lane and assist in steering back. Self-Parking systems allow a vehicle to park hands-free. Such devices are contributing to a reduced role for the driver, which ultimately could lead to driverless vehicles. The crucial transition is from high automation to full automation. Because many manufacturers, BMW for instance, market their cars on performance, they are likely to encourage hands-off-the-wheel only in situations where there is little challenge to the keen driver – such as long motorway trips or slow-moving urban traffic. Otherwise, driving is to be enjoyed.

Google’s pods, lacking a steering wheel, exemplify the other route – the great leap forward to full driverless. While these electric vehicles could be privately owned, they seem particularly suitable for shared ownership, given that they are, in effect, taxis with robot drivers. Taxis are popular, and we would make more use of them if they were cheaper, which they might be if robots replaced humans. This could increase demand, adding to traffic congestion in urban areas. But possibly the technology might allow the safe distance between moving vehicles to be reduced, packing more into the available carriageway.

The main impact on traffic of shared driverless cars is likely to be via parking. Privately-owned cars are generally parked for 95% of the time, seemingly an inefficient use of resources. Sharing would allow more time in use and so fewer parked cars. But the main impact on road space would be in the suburbs and car parks, not city centre streets where congestion is most acute and where parking is limited to avoid impeding traffic.

Driverless vehicles would contribute to congestion when they are on the move empty, as do black cabs plying for business. Programming your personal driverless car to cruise round the block empty while you transact business in a shop – in effect ‘parking’ on the move – would need to be regulated, possibly banned, in city centres (although this could lessen the attractions of driverless vehicles). A two-car family might economise with one driverless car, taking the breadwinner to work, then returning for use by the house wife/husband and children, before collecting the worker at the end of the day. But this would double the number of work trips, adding to traffic.

Altogether, it seems likely that the overall impact of driverless cars would be to increase urban traffic. It would be desirable model traffic flows under a variety of driverless scenarios to understand better the implications, since there may be conflicting policy objectives.

The UK Government is keen on driverless cars. The ministerial introduction to the Department for Transport’s 2015 action plan, The Pathway to Driverless Cars, starts: ‘Driverless vehicle technology has the potential to be a real game changer on the UK’s roads, altering the face of motoring in the most fundamental of ways and delivering major benefits for road safety, social inclusion, emissions and congestion.’ The Chancellor of the Exchequer, in his 2016 Budget, made a point of announcing trials of driverless cars on the Strategic Road Network by the end of 2017.

It could turn out, however, that benefits of autonomous vehicles on inter-urban roads could be offset by increased traffic on urban roads. One way of mitigating such traffic would be to increase vehicle occupancy significantly. This may be possible though what might be termed the ‘shared-squared-driverless’ mode, involving both shared ownership and shared use.

So rather than one or two occupants, the aim would be to fill the vehicle at peak times with passengers travelling in the same direction. This would reduce urban traffic congestion through high occupancy requiring fewer vehicles, with one study suggesting that this could remove 9 out of 10 cars in a mid-sized European city. Uber has introduced uberPool, a shared taxi service with lower fares, and uberHOP, which facilitates sharing along commuter routes at peak times. Their success will depend on the ability to match enough passengers going in the same direction, and also on the willingness of people to share.

If priority were given to shared-squared-driverless vehicles through road pricing or similar demand control measures, it might be possible to avoid urban traffic congestion while offering speedy and reliable door-to-door travel. This would be facilitated by some central oversight of such vehicles to minimise conflicts and maximise efficient use of the road network (analogous to air traffic control). The outcome could allow the car to compete with rail in urban areas, in terms of speed and reliability, and could help cities without rail infrastructure better to meet the mobility needs of their citizens. However, the technological, institutional and commercial challenges to the shared-squared-driverless concept are substantial, and practical feasibility is unclear.

Colin Buchanan’s seminal report, Traffic in Towns, was published 50 years ago, decades before the possibility of driverless cars. How much difference would autonomous vehicles make to urban traffic congestion? In the medium term, congestion could worsen, unless action were taken to regulate the movement of vehicles without occupants. In the longer term, the possibility of higher vehicle occupancy offers the prospect of mitigating urban traffic congestion.




Until fairly recently, the main driver of growth of travel demand in developed economies was growth of personal incomes. This is no longer the case, a key change that is an important contributor to the Peak Car phenomenon. For the future, demographic change is the main driver of demand growth. There are four aspects: population growth, population ageing, the young deferring maturity, and the spatial consequences of these developments.

I have a chapter summarising changing demographics in a new Handbook on Transport and Urban Planning in the Developed World. A copy of the manuscript is available Changing Demographics chap 23-1-15

Transport accounts for over 60 per cent of global oil consumption and about a quarter of energy-related carbon emissions. Typical forecasts of future world vehicle ownership project substantial increases, particularly in the developing economies. The transport sector relies largely on oil for motive power and has been seen as more problematic than other parts of the economy when it comes to reducing greenhouse gas emissions.

The problem of transport greenhouse gases may be less than generally supposed, however. There is emerging evidence that individual car use, as measured by the average annual distance travelled, has ceased to grow in most of the developed economies, starting well before the recent recession, and it may be declining in some countries – a phenomenon known as ‘Peak Car’. A number of explanations have been proposed, which are not mutually exclusive and include a decline in younger people holding drivers licences, changes to company car taxation, saturation of demand for daily travel, technological constraints on faster travel, and a shift away from car use in urban areas.


The shift away from car use in cities is particularly important in a world in which future population growth will be mainly urban and the economic attractions of population density are increasingly recognised – agglomeration economics. London illustrates these developments. Over the past twenty years the population has been growing and incomes rising, but car use has held steady at about 10m trips a day. This is mainly because the city has not increased road capacity but instead has invested in public transport, particularly rail which offers speedy and reliable travel for work journeys, compared with the car on congested roads.

A growing population but no growth of car use has resulted in a marked decline in the share of journeys by car in London, from 50 per cent of all trips in 1990 to 37 per cent currently. With continued population growth projected and more investment in rail planned, the share of trips by car could fall to 27 per cent by mid-century. There is every reason to suppose that London will continue to thrive as car use declines – perhaps because car use declines.

This decline in car use from 1990 was preceded by a 40 year period of growth from 1950, the result of growing incomes, growing car ownership and at the same time a falling population as people left an overcrowded damaged city for new towns, garden cities and greener surroundings. So we see a marked peak on car use around 1990, the time when the population of London was at a minimum, when attitudes to city living began to changes.

Peak Car in the Big City

This phenomenon of ‘Peak Car in the Big City’ is not unique to London although this is the city for which we have the best data. There is evidence for something similar happening in Birmingham, Manchester and other British cities as well as those in other developed countries. The shift in economies from manufacturing to services is an important driver, as is the growth of higher education located in city centres, attracting young people for whom the car is not part of the life style.

The Peak Car phenomenon is helpful for mitigating transport greenhouse gas emissions – both the cessation of per capita car use nationally and the decline in the share of trips by car in cities. I have estimated that these changes in behavior, taken together with expected developments on low- and zero-emission technologies, could reduce UK surface transport greenhouse gas emissions in 2050 by 60 per cent compared with a 1990 baseline. This fall short of the overall target of 80 per cent reduction, but is a good deal better than conventional projections.

Peak Car is not just an emerging phenomenon to be investigated. It is a helpful trend to be encouraged to achieve both successful, sustainable cities and national reduction of transport greenhouse gas emissions.


This article is based on a recent paper of mine published in the journal Case Studies on Transport Policy, also available as a final draft Metz CaseStudies 1-5-15 pdf.

The Department for Transport has published its Road Investment Strategy with details of £15bn expenditure on the Strategic Road Network. I have a critique of this programme published in Local Transport Today. I argue that, as ever, you can’t build your way out of congestion, but you can use the digital technologies to help drivers use the network more efficiently under congested conditions. The Highways Agency places far too much emphasis on the civil engineering technologies, and too little of the digital technologies.

Metz LTT Roads pdf 18-12-14


NNNPS stands for National Networks National Policy Statement. The UK government has issued a draft for consultation. This covers national perspectives for future road and rail developments and is intended to be approved by Parliament and provide a settled input to local planning enquiries into specific schemes.

There is quite a lot wrong with the consultation draft. I have submitted a response  Metz NNNPS response 26-2-14. My main concerns are the following.

There are good reasons to suppose that the National Transport Model substantially overestimates future demand on the Strategic Road Network. The traffic projections do not therefore provide a sound basis for the scale of investment planned by the Government. The Model needs to be made public and available for independent audit.

A major shortcoming of the NTM is its inability to model the consequences of a strategic choice between greenfield and urban housing to accommodate the growing population.

There is a good case for investment in the network where this can make land accessible for development, consistent with planning policy. However, constructing additional carriageway is not an effective means of reducing congestion.

The main problem with congestion is the uncertainty of journey time. This is tackled most cost-effectively by provision of reliable predictive journey time information.

A strategic issue for national Government is where to accommodate the UK’s growing population since this will determine the kind of investment needed in the transport system. In London, transport investment can be planned to respond to population growth. Nationally, there is no such coherence.

I presented a paper (Metz ETC 15-9-13) at the recent European Transport Conference held in Frankfurt. This shows how car use in London rose from about 5% of all journeys in 1950 to 50% by 1990 – no surprise there. But then it started falling, to 38% currently, and projected to fall further to 30% before 2040. This is surprising since car use has always risen with growing incomes. However, in London we have not built new road capacity, which has constrained car use to a steady 10m trips a day for each of the past 20 years. But because the population has been growing, these car trips represent a declining share of all travel. This is the clearest illustration of the phenomenon of Peak Car.

I also suggest that growing cities in developing countries, where car use is still low, may be able to avoid this peak and move directly to a more sustainable level of around 30% for a city of 10m. The key policies to achieve this are to provide rail travel for work journeys since this can get business and professional people out of their cars, and limit parking in the city centre to keep buses, taxis, good delivery and emergency vehicles moving.

I have a new paper published in the Journal of Transport and Land Use, title ‘Mobility, access and choice: a new source of evidence’. The new source is the UK Department for Transport’s accessibility statistics, which relate where people live (census data) to where are located services and facilities such as doctors, hospitals, schools etc. My analysis allows estimates of how much choice people have, dependent on their mode of travel. I find that for those with use of a car or good public transport, levels of choice are quite high, consistent with the proposition that demand for travel has ceased to grow because we have enough choice to meet our needs.

This Journal is unusual in that it is available free of charge online without charging authors for the costs of publication. Commendable.