Search the site

AuthorLavan Mahadeva

Research Director View profile

By 2035 there could be nearly half a billion cars on the road in China, creating unprecedented levels of congestion and pollution.

One solution could be widespread ride hailing through electric vehicles (“EVs”) coupled with further development in car platforms, congestion pricing, automated driving and sensors and cloud storage. Suppliers of metal into cars would need to know whether ride hailing on this scale would make sense and how many cars would be produced as a result. In this insight, CRU’s Research Director, Lavan Mahadeva, uses CRU models to estimate the feasibility and impact of ride hailing in the world’s largest car market.

For ride hailing to be successful, it must save time and money
Ride hailing will only become popular if car users themselves get a better deal. Of course, some of the benefits of ride hailing would go beyond their private advantages. Fewer cars means less pollution, for one thing. But to keep things simple, we focus exclusively on the private costs that cars users will have to pay other than those implicit in travel times. These costs can be split just as a taxi fare, into a per mile fee, a time fee and a flat per journey element:

  • The costs for the mileage travelled by each car user include fuel or power and wear and tear. Mileage costs depend on energy prices, and tend to be lower for EVs compared to internal combustion engines (“ICEs"). But because they can be charged per user mile, are not especially sensitive as to whether they are covered through ride hailing or single ownership.
  • Then there is a time fee element. Ride hailers are paying someone else to drive and manage a car rather than doing it themselves. Drivers’ salaries must be paid for, but their own productive time is freed up, and that has value. We assume that in balance there would no difference between ride hailing rather than driving in one’s own car.
  • Finally, there are costs that depend on how intensely each car is being used, or the average mileage of each car in each day. Under intense use, a car fleet will physically degrade more quickly, and the higher will be the cost of keeping it in working order or replacing it . These costs cannot easily be charged at a per mile rate because drivers making short journeys in a short time would not be fully compensated; hence the ubiquitous flat fee. By using fewer cars more intensely means that each user gets to pay a smaller share of this flat fee, but also that the size of the fee is higher. Therefore, ride hailing has two opposite effects on these intensity costs.
  • Proponents argue that an EV has fewer parts than an ICE; that those parts are falling in price and that EVs are modular. Perhaps, ride-hailing EVs can be loaded with sensors such that the fixed costs of charging, replacement and maintenance could be better monitored and shared. For these reasons, EVs may be cheaper to use intensively than ICEs. But, it is also true that batteries are a large share of the fixed costs of an EV, and, yet, there is no firm consensus on how long it will take for a battery to degenerate under intense usage.
  • The intensity of usage is also critical for metal demand. Having fewer cars in the pool means lower metal demand but more intensive use means that they will be replaced more often, increasing metal demand. Our calculations of the combined effect on car user costs and production determine both whether ride hailing is feasible and what it will mean for suppliers of metal into cars.

Exploring the impact on China 2035
The results below are based on a scenario where EV cars are much more economically resistant to usage. Our estimates of costs are from CRU’s EV model and projected into 2035.

We present the results in the form of a single chart. On the x-axis is the production of cars in China in 2035. Production follows from sales and both are calculated in CRU’s own Chinese car fleet demand and production model. In this model, the sales of cars depend on new demand for cars from new entrants into the middle class and the demand for replacement from existing car owners. Both will be affected by ride hailing, but in opposite directions, with new demand falling while replacement rises.

On the y-axis is the average annual cost of car use in 2017 dollars. We would need to see large falls in the cost of car use as the result of ride hailing if we think that ride hailing will take off.

The size of the bubbles represents the extent of ride hailing, which is measures by the number of cars on the road. In our base case, without a massive extension in ride hailing, we expect there to be about 500 million cars in China in 2035. In an extreme ride hailing scenario, there would be 200 million cars. There are estimated to be over 150 million cars in China currently, so even a pool of 200 million cars would mean 30% more cars in circulation, but being replaced more frequently.

EVs can foster ride hailing in China, but only if they are resistant to intense use

Source CRU

The chart shows that in the base case without extended ride hailing, where about 500 million cars will circulate in China in 2035. The total annual costs of using EVs at 2,200 2017 $ are slightly cheaper than those of ICEs at 2,500 2017 $. Currently it is cheaper to replace an ICE than a EV, but we expect that falling trends in battery technology will have overtaken those in ICE costs well before 2035. This could be especially true in China if the government continues to promote EV production. About 40 million cars of all types are expected to be produced annually in this base case, which is a just less than double current levels of production.

Based on our assumptions, the combination of EVs and ride hailing can lower user costs. With widespread ride hailing keeping the car stock to 200 million, the annual cost of EV car use would be 1300 2017 $, a fall of 40% below the business as usual case. The costs of using ICEs under ride hailing would be like under the base case.

This is entirely explained by the assumption that EVs are more economically resistant to usage than ICEs. If instead all cars, be they ICE or and EV, need full replacement after say, 150,000 miles no matter how intensely they are used, and if at that point, the replacement cost is not dramatically lower for an EV, then we see no advantage to EVs in ride hailing in terms of the private cost to the user. Some change in technology, such as mass adoption of autonomous driving, or a significant policy subsidy would be needed to promote ride hailing in these circumstances.

Metal demand will be affected if ride hailing takes off
We carried out further simulations of CRU’s car demand and production model to show that extended ride hailing will dramatically lower car production relative to the base case. With widespread ride hailing, car production falls by two thirds of the base case. Greater replacement cushions the blow on production somewhat. But it is swamped by the simple effect of less cars. The effect of ride hailing should be seen in the context of a base case for 2035 which envisages just less than a doubling of car production from current levels. It is also worth remembering that different metals will be demanded if a greater share of produced cars will be EVs.

To summarise, extended ride hailing is a viable possibility for China in 2035 but only if EVs turn out to be especially resistant to intense use. We will learn more in coming years about EVs when they are used intensely.

Could ride hailing take off elsewhere?
The implications of ride hailing, or indeed any other car sharing scheme, can be estimated by applying the same models for other countries and at other moments in time. One reason to expect different outcomes is that the costs of fuel relative to EV power vary over time and location. The patterns of car use can also vary tremendously across countries and cultures. The typical distance traveled by car can be expected to be relatively short in China in 2035, about 6000 miles per person per year or even less (because it can be combined with longer spells on public transport). Chinese car users are expected to have an average car mileage well below that of a US citizen even by 2035. With such short journeys, it makes less economic sense to own your own car and ride hailing becomes a better economic proposition.

The Technology Metals and Energy Consulting desk at CRU needs to be able to better understand future demand for new energy vehicles, and in particular, the rate at which consumers will change to the new vehicles on offer, in order to forecast the demand for the commodities they contain. We would appreciate five minutes of your time to complete the survey.

Take part in the CRU Survey here

Explore this topic with CRU

Technology Metals team

  • Lavan Mahadeva

    Research Director London
  • Rebecca Gordon

    Head of Technology Metals and Energy, Consulting London
  • Michael Finch

    Research Analyst London
  • George Heppel

    Consultant London
  • Ian Hiscock

    Head of South East Asia, Consulting Hong Kong