We have a momentous event coming up in my household: my son will turn 16 at the end of the month and will — if the DMV gods are agreeable — get his drivers license. This has sparked a lot of debates in my family about what driving will look like over the next 10-20 years.
In short, my son HATES the idea of driverless cars. Imagine – the club he’s been pining to join – drivers – is now threatened with extinction. Perhaps with wishful thinking, he has come up with a lot of theories about why self-driving cars will never take off.
I disagree with him, though I may be indulging in a bit of wishful thinking myself. I find few things more stressful than sitting in the passenger seat with my son at the wheel. His behind-the-wheel instructor says he’s a good driver (I wish she wouldn’t tell him that…), but I have never been quite so focused on everything that could possibly go wrong, and I would rather trust a computer to make the right decision if something does.
Also, I’ve spent enough time in Bay Area traffic jams – where one distracted driver who brakes a little too hard can slow down a whole lane of traffic – to relish the idea of smoothly flowing computer-driven cars. Research seems to back me up — simulations suggest that automated vehicles will likely reduce fuel consumption, and part of that reduction will come from fewer slowdowns due to accidents.
Here’s my son’s theory, which draws on network economics even if he doesn’t use that phrase: as long as there are enough people like him on the road, who actually want to be behind the wheel, driverless cars won’t do much to improve congestion. In the extreme, a mixture of robot-driven and person-driven cars could be worse for congestion than all person-driven.
Imagine if Silicon Valley technocrats could send for their favorite Los Angeles sushi and have it delivered by a driverless, and passenger-less, car, thereby adding cars that wouldn’t have been there. Then put those vehicles on the road with the remaining 16-year-old boy drivers, and others with an inner 16-year-old boy, some of whom get a kick out of messing with the automated cars’ sensors to make them brake quickly.
His theory was borne out by the story of the Google car getting stuck at the four-way stop as it waited for other cars to come to a complete stop. But, that doesn’t seem like an unsolvable problem to me – someone just needs to update the algorithm and stress test it versus thrill-seeking drivers.
My son also points out that his online driver’s ed course warned that no one leaves the house thinking they will get in a car accident. So, he thinks people won’t be drawn to driverless cars to protect their own safety.
Consistent with this, surveys suggest that most of us live in a Lake Wobegon world and think we’re better than the average driver. This could mean that we all want other people – particularly the drunks, texters and overly aggressive lane-changers – to be in driverless cars, but want control over our own on-road destiny. Given that we buy cars for ourselves and not others, this doesn’t lead to many autonomous car sales.
I try to explain to my son (without using the phrases “opportunity cost” or “consumer surplus”…) that driverless cars will both give us more time and make driving a lot cheaper, so teenagers will eventually find another way to mark the transition to adulthood.
On the “more time” point, think of all the things we can do instead of sitting behind the wheel of the car. With more of us able to be productive remotely, time in the car could be quite valuable.
In terms of the cost of driving, it’s hugely inefficient to have so many of us own a $20,000-plus piece of capital that we use, on average, 46 minutes per day. The capital depreciates even when we don’t use it because technological change makes newer cars more desirable.
If you could order up an autonomous car only when you needed it, the cost of the capital would be spread over many more people and rides, driving down the cost per ride. So, I explain to my son, you’ll have to really, really like driving to pass up the much cheaper alternative of renting one from the next incarnation of Uber or Lyft. In fact, GM and Lyft recently announced that they will begin testing self-driving taxis on actual roads within a year.
Cars themselves are also likely to get cheaper if they’re automated, leaving aside the cost of the automation itself. In economics, cars are the canonical empirical example of a differentiated product. Remember back to basic microeconomics, where the perfectly competitive market model works for a purely homogeneous good and market forces drive prices to marginal costs?
The converse of this is that the more differentiated products are, the higher the markups above marginal cost are likely to be (which roughly means higher company profits). In fact, economists have written dozens of papers trying to model consumer demand for cars, accounting for our demand for brands, horsepower, leather seats, etc.
My guess is that with driverless cars, consumer demand for differentiation will be much lower. Who even knows what the brand of the last bus you rode was? And, as long as my Uber driver’s car is clean and gets me where I’m going, I don’t really care what he’s driving – no self-identity there.
In a rejoinder that warms his economist mother’s heart – the boy understands incentives! – my son points out that this is another reason why driverless cars are doomed. The auto companies will figure out that they spell lower profits for them, and will use their (considerable) economic and political power to derail them.
We will see. In a battle between Google and Ford – Silicon Valley and Detroit – I might put my money on Google. At least I hope I’m right….
What do you think? For those of you with 6-year-olds, will the drivers test be the same rite of passage in another 10 years?