Are driverless cars the death knell of the motor biz?

Looking at the numbers

I don't pretend that the multiplexed car will eliminate car ownership entirely. Some people will still need to own cars, and some people will want to no matter what. But if the price is right, an awful lot of people will think “why take on debt to buy something if I don't need to?” – not to mention the running costs, insurance, registration and the rest.

The Royal Auto Club of Victoria (RACV) puts the cost of owning one of Australia's most popular cars, the Holden Commodore, at more than $AUD200 a week.

In that scenario, and excluding other trips (which will, after all, be spread across the second car that most car-owning households maintain), the daily commute comes at a price of $20 per hour.

So let's work to a price: the autonomous car comes at $10 per hour (not per kilometre, because as any taxi stuck in a traffic jam will attest, that's an unprofitable way to run a business), but instead of operating two hours per day, it can be in profitable service for six hours a day.

Ignoring times when there are multiple passengers, that's a daily take of $60: nearly $110,000 over five years. If you can put the car on the road for $40,000, the remaining $70,000 should be enough to operate it with a profit margin available, and the customers are getting transport cheaper than if they bought their own car.

And on shared trips (with a discount for each individual), extra over $10 per hour is cream.

The same exercise is harder to calculate for America, either because of the data the Census Bureau publishes, or because I'm not so adept at finding US census data. While it's easy to find out that the average commute across America's 50 largest cities is 23 minutes, the Census table I found din't state the average drive commute time (“Travel to Work Characteristics for the 50 Largest Metropolitan Areas in the United States by Population”).

However: since 73 percent of Americans in the sample drove alone, their drive time can't be too far away from the average. Every other sample is too small to make a significant difference.

According to the AAA, the annual cost of owning a car is $US9,000, and the quicker commute means a longer idle time – which means, on the same basis as I used for Australia, the cost of owning a car for commuting is $US32 per commute-hour.

So on a like-for-like basis, America is a better place to try autonomous-car-as-a-service than Australia. It's also got a much bigger population, making it easier to justify investing in a decent size fleet to begin with.

In either case: it's at least feasible that once the (considerable) startup costs are covered, a business model exists under which people will use autonomous cars, but never own one.

That's a deadly threat to the motor industry: like all businesses, they need growth to survive. A contraction in the market following the 2007 financial crisis was enough to send car-makers to governments seeking bailouts.

If a driverless fleet cut new car sales by just five percent, it would be a crisis for the industry. A contraction of ten percent would be catastrophic. ®