Third Order Effects of Autonomous Cars

The automobile allowed workers to travel to work, safely, predictable and much further than ever before. This eliminated many of the benefits of walking to work, which in turn lead to people living cities in droves, for the benefits of space that the suburbs provided. Once in the suburbs people relied on their cars to go shopping, this lead to the rise of big box stores such as Wal-Mart, and Costco.

The cheap prices of Walmart can only be achieved by the volume that they’re able to sustain because of the large store, abundant parking, and cheap land. Costco sells goods in such a quantity that they you need to drive to the store to be able to haul the goods you bought back home.

I’m going to categorize the first order effect of the automobile as enabling people to travel much further, faster. The second order effects would be the mass migration to the suburbs. The third order effect is the rise of Walmart and Costco.

Autonomous cars, much like automobiles, will have obvious, first order, effects. Harder to see is the second and third order effects that will reshape cities and the economy. Let’s first consider the current state of the industry.

A traditional Boston taxi spends only 25% of it's time occupied, averaging less than two trips per hour and taking only 1-2 people on most trips [1]. The price of a taxi medallion in Boston, at its peak, was $700,000 [4]. This lack of utilization combined with an insanely high barrier to entry has driven the prices to a level that only benefits those at the very top and not the drivers. Medallion owners are the top of the food chain, they rent out medallions to drivers for an average of $500 dollars a week [1]. Those drivers in turn have to work 12-hour shifts, seven days a week, to afford to rent the Medallion, more on top of that to rent a cab.

It's clear that normal taxis are highly under-utilized and as a result cost more than they should. This hurts the drivers, the passengers, and the economy, as fewer people can afford to travel to their destination.

This is starting to change with the rise of ride sharing companies such as Uber and Lyft. They are more efficiently routing drivers and passengers, sometimes matching multiple parties going in the same direction, as is the case with uberPool and Lyft Line. Drivers can keep a majority of their income, paying a percentage of earned wages (typically around 28%), rather than a flat fee. They own the cars that they drive and can drive the hours that they want.

This higher utilization of the driver's time and fewer fees is pushing prices down to a level where it makes more sense to take an Uber or Lyft than public transportation. Thus the ridesharing transportation market is growing doubling every six months according to Uber [2].

The next step to this progression is level-five autonomy [3]. Fully self-driving cars can be scheduled to arrive to your door when you're ready to go to class, bring you to the door of your destination, picking up people along the way.

The first order effects will include, faster, more reliable, transportation. But, as we optimize the cars, and treat them as inputs to an optimization problem, we can see benefits that exceed what is possible with ridesharing apps and humans today.

Imagine, instead of having 1-2 people in the car, you have up to 5 (since there’s no driver). Each person is going in similar direction but each person has a different start and end location. The cost of the fare will simply be a multiple of the hourly cost of the car, which is very little without a human in it, divided by the number of people in that car.

Let’s imagine you’re headed to the airport and you could pay $25 for a Taxi, that price includes the overhead of renting the medallion, renting the car, paying the radio operator fees, and paying the credit cards fees. Assuming that the driver is 100% busy, you’re only paying for the portion of time that you use. But, as we know, the driver is only 25% busy on average so you’re only paying a 4X premium. If the Medallion’s cost is $500 / week [1], the cost to rent the car is $700 / week [1], and the driver typically makes $1,000 / week [1]. That means that your fare is broken down into $5.68 to lease the medallion, $7.95 to lease the car and $11.36 for the driver’s wages.

In a self-driving car, you’re not paying for a medallion and you’re not paying for driver’s time. Thus the trip that cost $25, is now $7.95. If you fill four seats, you’re only paying $1.98 for the trip. Now that’s optimistic because of the high R&D costs that make self-driving cars really expensive right now. But after the technology has matured you should see costs similar to that.

With a ride available at the push of a button for less than the price of public transportation or owning your own car, people will stop driving and taking the T. It will reshape the urban landscape as parking lots are turned into parks and roadside parking is turned into bike paths and outdoor seating.

With transportation costing less than it ever has, new markets will open. Ordering food to your door has traditionally been expensive and time consuming. With a fleet of autonomous cars making deliveries for a few dollars, it will change the way people eat. Waiting for two-day shipping will be a thing of the past as you'll be able to order goods online and have them show up at your door an hour later. Grocery shopping will be done for you based on how much food you have in your fridge, with groceries showing up to your door when you get home.

Traffic will be a thing of the past. Cars tasked to take people into work in the morning will deliver goods in the day before taking people home in the evenings. Cars will serve a larger role than just transporting people, they will be the fabric of the new urban landscape. Despite their increased role, we'll have far fewer cars than ever before, resulting in decreased pollution, less traffic and faster time to destination.


Comments ()