In this post I will explore the economic realities of the coming autonomous car revolution for the auto industry. Recently there has been a great buzz around the imminent arrival of self-driving cars. There has also been much discussion about what this means for society and the mass movement of people. Companies like Uber have raised billions in order to maneuver themselves into a position to take advantage of the coming disruption in the auto industry and others such as Google, Tesla, BMW, Mercedes and Apple (if reports are to be believed) are in the midst developing the requisite technology that will innovate auto-tech to the point of being fully autonomous. The general mood has been one of excitement tempered with the realization that jobs in logistics such as taxi driving, truck driving and delivery driving will become a thing of the past however little has been spoken of about the effect that this technological disruption will have on auto manufacturers themselves. Naturally each entrenched manufacturer is showing a mix of confidence and bravado however when digging a little deeper into the figures and the potential economic environment that will accompany the emergence of this technology a picture of instability and almost certain difficulty for auto manufacturers is unveiled.
Mobility is a way of life
Looking at the Federal Highway Administration (FHWA) figures for 2014 we can start to build an understanding of what the USA’s vehicle usage patterns look like. In total 3.04T miles were driven of which 2.08T miles were classified as urban meaning roughly ⅔ of all miles driven were in an urban environment. These miles are significant. Uber, Lyft and similar companies operate almost exclusively in urban environments where high density populations and on average shorter individual journey times make their service attractive from a price competitive and convenience perspective. We will come back to the significance of this figure later.
This figure shows an 18.25% growth in the total number of miles driven since 2000 which when projected to 2027 amounts to almost 2.5T urban miles driven in total in the USA. Looking at the trend over the course of the past 15 years we can see that total miles driven decreased during the recession but rebounded and as we can expect another recession by 2027 it is fair to assume that this long term trend will remain mostly accurate.
Delving deeper we can analyse the total number of cars on the road, the age of cars on the road and also the number of new cars registered each year. In the year 2000 there were 225.8M cars on the road in the USA. By 2007 just before The Great Recession there were 254.4M cars on the road and by comparing the number of new vehicles registered (see below for more), the average age of cars on the road and factoring recessionary behavioural patterns we can extrapolate that there were roughly 255.5M cars on the road in 2014. This represents a sizable growth of 13.15% of road registered vehicles since 2000 which again when projected to 2027 with current usage patterns means there could be 289.1M cars on the road by then. This is highly unlikely for reasons which we will discover.
The Great Recession hollowed out new car sales
Before we continue hypothesising let’s quickly also take a look at the number of new cars registered each year since 2000. That year there were 17.8M new registrations (the most successful year to date for auto dealers on record), 16.5M in 2007, 10.6M in 2008 (the recession was clearly in full swing by then), with figures only just rebounding to 16.5M again in 2014. If there was any illusion of how impactful the recession was the fact that it took 7 years to rebound to 2007 figures in car sales should say it all. This point is also a very important part of the hypothesis presented in this post so please bare it in mind. Also of note is that about 11.22M of these were for use in mainly urban environment (guesstimate).
The average age of registered vehicles is increasing
With this information we have a considerable picture of what the auto industry looks like for the USA in the recent past but we can get an even clearer picture if we explore behaviour a little more in detail. Consider for a moment that in 2001 the number of cars on the road over 7 years old was 60.8% with the trend moving towards this figure increasing as new cars became more and more unaffordable for the average American. In fact the average price of car was $32k in 2013, a monthly payment that out of the whole country only the median income families in DC can afford.That is especially noteworthy when considering that the car in the most under utilized utility. A car is on average probably only used around 5% — 7% of the time with the rest of the time being spent idle in a parking lot somewhere waiting for it’s owner to finish work. That means, in 2014, America spent $528B on a utility it only uses at 5% of its potential efficiency. Just let that sink in for a moment.
What does all this mean for the future of the auto-industry?
Ok we are done with the hard facts. I know they’re not exciting but they provide a good foundation for the main hypothesis of the post. I have taken the liberty of making educated guesses on some of the figures from here on in but they are very conservative and solid guesses grounded in reason which I have tried to explain as best as possible. If you have better figures based on more solid scientific projections that I haven’t had time to research fully I’d love to hear them but based on some other papers I have read the figures I have arrived at are definitely near to reality.
Most established car companies are going to go bust and there is nothing they can do about it.
There is a trend in the world towards sharing resources. I live in a city. I bike, walk or sometimes take the train to get around. It’s true I also drive from time to time but when I drive I drive a Car2Go. I swipe in and pay by the minute. I’ve never taken an Uber, Lyft or Wundercar but I imagine I will do very soon. In fact I think I will stop using Car2Go self drive services and only take chauffeur style Uber services very soon. When is very soon? As soon as there are no more human drivers. True this is purely anecdotal but let me explain.
We know that cars are used 5% of time (ok it’s a guess based on each car being used about 2 hours a day). In any case we know they are very underutilised. When we implement a network of fully autonomous vehicles which benefit from an advanced technological routing and distribution efficiency management systems, car usage efficiency could get as high as 60% in peak hours and 80% in off peak in an urban environment. Add to this the change in culture with respect to more flexible working hours we can see that rush hour patterns will be negated slightly, again improving network efficiency. That said autonomous car service providers may prefer to optimize for customer satisfaction and have bigger fleets to cope with demand during peak hours. 100% efficiency in peak hours would mean maybe 50% efficiency or less during off peak. Night time hours efficiency could be expected to be at around 5–20%. Regardless all these figures are either light years ahead or at least still game changing compared to our current efficiency levels.
A 60% efficiency means that for every 12 cars currently on the road in an urban environment we would only need 1 car to replace it. Let’s revisit our figures from earlier to see what this really means. We had a car market of $528B in 2014. A rough calculation based on miles driven would point towards $359B of this being for urban based vehicles. With this in mind we can estimate the urban car market 2027 (1 car for every 12) will require a total value of $29.9B to serve the same amount of passenger/miles at current buying levels however we can expect that number to be two or three times as high due to customers demand for better experience meaning that fleets are updated more regularly than family cars are currently. Remember this is retail price so the price for the manufacturer will be much less. Let’s say cars are marked up 100% from manufacturer to retail (I couldn’t find accurate figures so if anyone knows a better figure please let me know and I will adjust the figures here) the cost of adding new cars to the fleet to be recouped is accurate give or take.
Assuming a network efficiency of 60% will be enough to achieve customer satisfaction of 95% then we can assume that that actual vehicle sales for those would don’t use autonomous taxis but instead continue to opt to buy will be at around $18B per year.
Car sales are going to crash
This leaves us with an urban car market of $47.87B in 2027 compared to $359B in 2014. That is 13.3% of the market value and is a huge hit for the industry to take even when considering that services and logistics related opportunities will generate extra revenue for providers. Backing up my back of napkin economic projections a study in Portugal showed that autonomous cars could eliminate 90% of cars on the road in turn freeing up acres of parking space (since I first drafted post this in April Barclays went for a more optimistic 40% drop in sales). Many companies are going to have to adapt their business models, grow new markets or risk becoming obsolete. My guess is that the companies with large and inflexible infrastructures, financial commitments and traditionally lacking in innovation will be the first to go (Detroit I’m looking at you). Next will be the companies that manage to successfully innovate their technologies but not their business models and perhaps more importantly fail to deleverage their cost structures into more manageable and flexible micro units. The companies that are best positioned to benefit are the startups who have everything to gain and nothing to lose. They won’t have to worry about protecting old revenue streams or transitioning business models and will therefore be best positioned to take advantage of the coming carpocalypse. The new mobility startups will also be perfectly placed to take advantage of new revenue streams that will open up with autonomous mobility. There is a potential total market of $520B in urban mobility at 25¢ per mile but let’s not forget that this will destroy an existing industry of $1.2T based on current cost per mile driven. No matter how the maths is done it looks like trouble ahead.
But I love my car!?
Wait a moment. You may be asking what if people want to still own their vehicle? This is a good question and much of American popular culture is based around the status, power and freedom of car ownership. We can look at the trends to see if this is a legitimate observation and the trends point towards sharing and cost efficiencies being the drivers of the 21st century economics in the USA. Demand for used cars was up in 2014 pushing up used car prices and pointing towards a consumer reluctance towards buying a new vehicle. On average each car in the USA drove 11,891 miles in 2014. Currently each mile driven costs 59.2¢ yet estimates point towards autonomous cost per mile being anything between 25¢-40¢ (NYC taxis cost $4-$6 per mile currently). Cyclic projections of economic recessions point toward there being another significant recession in the early to mid 2020’s. Elon Musk and Tesla aim to have their fully autonomous vehicles on the road by 2023. The stars align perfectly. Through a convergence in timing of technological innovations coming to market, economic pressures forcing people to reduce their spending and increased availability and affordability of autonomous personal transport networks the soil will be fertile for a huge and significant cultural and behavioral change in peoples transport usage patterns.
Convenience, cost, recession and a shift in perspective
Currently New York City is the only locality in the country where more than half of all households do not own a car (the figure is even higher in Manhattan, over 75%; nationally, the rate is 8%) but we can expect to see this shift drastically in the coming 15 years. Driving is becoming more and more inconvenient with every passing year. Pump prices are rising (even if oil is momentarily down), parking is difficult to find and costly and sitting in traffic is taxing on time and mood. Add to this the cost of insurance, road tax and vehicle maintenance and suddenly a personal vehicle can seem more like a burden than a luxury. Public transport has its pitfalls too and it is primarily the lack of privacy, inability to carry awkward loads and inconvenience of not getting dropped on your doorstep that keeps people driving. Public autonomous vehicle networks will not only solve the major issues of public transport but also solve the major issues of car ownership all in one fell swoop. The list of pros will so conclusively outweigh the list of cons that with a little help from recessionary economic pressures people will switch their mobility patterns and once they have experienced the power of autonomous mobility they will never switch back. As many internet start-ups have already proven despite the naysayers convenience really is king.