The Fuse

Car-Sharing and Ride-Sharing to End Personal Vehicle Ownership? Not Yet

by Hart Schwartz | April 07, 2017

The advent of car-sharing has led to speculation about whether we will soon see a major reduction in personal vehicle ownership in urban areas. A 2016 report from Fuels Institute, entitled Shared Travel: Revolution or Evolution? An In-Depth Fleet Analysis of Shared Urban Mobility, performed a literature review of several articles and studies on car-sharing that suggested this reduction has already taken place. For instance, an oft-cited 2014 survey from consulting firm AlixPartners suggests that for every car-sharing vehicle, 32 vehicle purchases are avoided, leading to hundreds of thousands of avoided vehicle purchases.

However, there is one simple problem, at least to date, with this line of argument (revealed later in the Revolution or Evolution study): The suggested dramatic reduction in urban vehicle purchases is not corroborated by U.S. Census data on household vehicle ownership. Every year since 2005, in its annual American Community Survey, the U.S. Census has tracked the number of vehicles available per household in U.S. places, cities, urbanized areas, and metropolitan areas. A review of this data shows an entirely flat trend pattern—there has not been a marked reduction in vehicles per household, or even any reduction at all. Moreover, some metro areas and their central cities have actually seen a slight increase in vehicles per household.

The following chart shows the level of household vehicle availability in all U.S. metropolitan areas, annually, for the period from 2005 through 2015, reflecting how ownership has remained steady (2016 data was not available at publication).

There are a few qualifications to this data. Even if there has not been a general trend for all U.S. metro areas, have at least certain places seen a reduction of household vehicle ownership? The answer is no, not at all. The Fuels Institute’s Revolution or Evolution study identified metro areas with high levels of car-sharing, based on a software inventory download from the online booking systems of car-sharing companies. But the numbers haven’t moved, even for these high-volume car-sharing areas. The chart below shows a flat trend-line for selected metro areas from 2005 through 2015.


In fact, tabulating the data from the above chart shows that some of the selected metro areas even had an increase in vehicles per household in the most recent year tracked, from 2014 to 2015. The table below shows the increases in highlighted cells.


The vast majority (approximately 90%) of car-sharing fleet vehicles are located in the central cities of metro areas, and not in the outlying suburban regions.

Certainly there has been no overwhelming reduction, in these metro areas, in the level of household vehicle ownership. However, what if one narrows the lens even further, to focus only on the dense central city for these metro areas? According to U.S. Census definitions, metropolitan areas contain at least one principal city of 50,000 or more residents, and according to the Fuels Institute’s Revolution or Evolution study, the vast majority (approximately 90%) of car-sharing fleet vehicles are located in the central cities of metro areas, and not in the outlying suburban regions of these metro areas.

Therefore, might there be a reduction in vehicles per household, in recent years, in those central cities that have seen the greatest amount of car-sharing? The following chart narrows the scope from the 14 selected metro areas in the preceding chart and table, to only the central city of these metro areas.

Again, the trend remains flat, even when looking at the urban places that have the greatest density and intensity of car-sharing fleets. As with the metro areas as a whole, many central cities have even seen vehicles per household increase between 2014 and 2015, as shown in the tabulation below (in the highlighted cells for 2014 and 2015).


Why has the trend in household vehicle ownership been flat in metro areas and central cities that do experience significant levels of car-sharing? The answers are obviously open to speculation, but a few reasons come to mind. First, there may be a dual, or parallel, phenomenon, whereby vehicle ownership has become reduced for some households that use car-sharing very heavily, but at the same time vehicle ownership has increased for other households that have their own vehicles. If so, this could explain the fact that certain metro areas have both high levels of car-sharing and high and increasing levels of household vehicle ownership.

The car-sharing market share may be narrowly targeted among the small but consistent percentage of households that have never in fact had their own vehicle.

Second, the car-sharing market share may be narrowly targeted among the small but consistent percentage of households that have never in fact had their own vehicle. If this is the case, then it would only reinforce the dual trend speculated above, whereby a small but steady proportion of households use car-sharing but the rest of households have their own vehicle.

Duality: substitute or complement?

The two reasons given above strongly suggest a duality among urban residents: Some use car-sharing to reduce their vehicle ownership, or they never owned a vehicle to begin with, whereas many other urban residents continue to rely on personal vehicles. But considering this duality scenario brings up the possibility that some urban residents may even switch back and forth between car-sharing and personal vehicle ownership, either in the short or long term. If so, this would imply that some amount of substitution is occurring between car-sharing and car ownership. But how much substitution exactly, and what is the elasticity of substitution? Is car-sharing a substitute, or a complement, for personal vehicle ownership? Answering this key question—whether it is a substitute or complement—will be the topic of a future article in The Fuse.