Rising Population, Falling Traffic: Why Has Car Ownership Fallen While London Has Prospered?
James Clowes, Transport for London
The impact on car ownership of increasing population density is estimated econometrically alongside other variables that are conventionally used. Increasing density is considered to reduce car ownership, counteracting the effect of rising incomes.
Since 1999 London’s population has grown by 15 per cent from 7.2 million to 8.4 million. Over the same period, the annual volume of traffic in London fell by the same proportion – 15 per cent. In July 2014 Transport for London (TfL) published a report titled “Drivers of Demand for Travel in London: A review of travel trends and their causes”. This study examined the reasons why London experienced modal shift from private motorised transport to public transport, walking and cycling over this period.
The reasons London experienced such substantial modal shift were found to include not only the already well understood improvements to its public transport network, but also wider social and economic changes that have taken place in London over the past two decades. While London’s public transport network was greatly improved over this period, other factors also played a substantial role: highway capacity was reduced, fuel prices increased while public transport fares decreased, average incomes in suburban areas of London did not increase for many years, population growth was skewed toward already dense areas, and the structure of London’s economy changed. These and other factors all contributed to the reduction in traffic volume seen in London from 1999 onward.
TfL is now undertaking a programme of work to explore some of the specific factors found to have contributed to the trend for declining traffic in greater depth. One strand of this work has analysed car ownership in London over time – both in terms of car ownership rates and the total number of cars owned. Between 1994 and 2005 the total number of cars owned in London grew by an average of 1 per cent per year, but since 2005 this growth has ceased. Given that since 2005 London’s population increased by 12 per cent, this lack of growth meant the car ownership rate in London decreased from 0.34 cars per person to 0.30 cars per person during the eight years to 2013.
Car ownership modelling conventions link car ownership rates to amongst other factors GDP, the mix of household types and the rate of licence holding. A saturation level is generally assumed as a maximum bound for the ownership rate, with lower saturation levels in dense urban areas and higher saturation levels in rural areas.
TfL has previously carried out work modelling car ownership, including a model presented at the 2010 European Transport Conference. The model presented in 2010 forecast an 11 per cent increase in the number of cars owned in London between 2008 and 2016, which has now been observed not to have occurred. The approach to modelling the rate of car ownership employed in 2010 did not take into account, however, the effect of population growth on car ownership rates over the modelled time period.
Population growth means the density of the area increases over time. With constant levels of car ownership, this would entail greater levels of car travel and consequently congestion. The feedback from population growth to congestion would be expected to reduce the utility of car ownership, therefore reducing the rate of car ownership – an effect not captured in the standard approach. The absence of this effect is an explanation of why TfL’s 2010 model forecast an increase that did not occur, since the modelled car ownership rate was multiplied by the forecast population size to give a total number of cars owned, treating car ownership rates as independent from population size. Building on its previous work, TfL has developed its modelling approach to allow for population growth over time to influence car ownership rates.
This project employs an econometric approach to quantify the impact on car ownership of increasing population density, alongside other variables that are conventionally used such as GDP, demographics and licence holding. The conclusion drawn is that in prosperous urban areas such as inner London, the expected effect that rising incomes would lead to increasing car ownership rates can be counteracted by the effect of increasing population densities driving down car ownership rates due to the reduced utility of car ownership and use.
Association for European Transport