One of Google Inc.'s (GOOG) moonshots that is fast becoming reality is the self-driving car. GOOG has continued to research and develop the technology, logging more than 300,000 miles. Furthermore, states are starting to review the implications of these vehicles.
Both California and Nevada have legalized self-driving cars. On June 16, 2011, Nevada's Governor Brian Sandoval signed into law AB 511, which required the Nevada Department of Motor Vehicles to establish regulations governing the safety and performance of autonomous vehicles as well as establish testing areas. AB 511 also required the establishment of a driver's licence endorsement for operating autonomous vehicles.
On September 25, 2012, California's Governor Jerry Brown signed into law SB 1298, which directs California's DMV to establish necessary regulations. SB 1298, unlike AB 511, contemplates the notion of truly autonomous vehicles be defining an operator instead of driver and acknowledging that the operator may not be in the driver's seat, but rather "causes the autonomous technology to engage." This alludes to the revolutionary innovation that there will eventually be vehicles without drivers.
This article will look at two sources of value that could be derived from self-driving cars. The first is simply productivity gains from freeing drivers and the second is capital productivity savings from reducing the number of vehicles used.
Productivity increase could be enormous
Any time spent driving a car is time that cannot be devoted to other, perhaps more valuable, activities. Some sources put the time spent driving a car at around 520 to 600 hours a year. My analysis suggests a somewhat lower number around 441 hours per year. This is calculated by looking at the average number of miles a vehicle is driven and the average speed. The following table shows the distribution of miles driven by type for passenger cars (light duty vehicle (LDV) with short wheel bases) :
Passenger Car VMT by Type
|Type||LDV short wheel base (Billions of VMT)||% of Total|
|Other Arterial Rural||230||11.3%|
Source: Department of Transportation, results for 2010.
Of the over 2 trillion vehicle miles, over half are driven on urban non-interstate roads. Given stop and go traffic, this is probably not at a very high rate. The following table shows a calculation with an estimate for the time spent driving, which should include congestion traffic, but exclude time spent pumping gas or basic maintenance.
Driving Time in Light Duty Vehicles
|Type||LDV short wheel base(miles per year)||MPH||Hours|
|Other Arterial Rural||1,208||30||40|
Source: Department of Transportation, results for 2010, Author estimates and calculations.
If the speeds were a little higher, the hours driven would be a little lower and hence productivity gains would be higher. However, looking at 441 hours across 192 million vehicles is around the equivalent of 42 million man years - roughly equivalent to the working force of Germany. There would be additional gains from other vehicles as well.
However, this assumes first that driving time is completely unproductive, which is not true since you can still do certain things while driving, albeit not at the same level of focus as if you were not driving. It also assumes 100% participation, which is unlikely even in the near future. Furthermore, there is some small portion of this driving that is viewed as recreational - quite possibly many the miles driven on motorcycles would qualify and hence there would be no desire to have a self-driving vehicle.
The key assumptions for financial gain are assuming a participation rate (% of driving freed by self driving), an effectiveness rate (relative value of time freed - you're still sitting in the car so it is not complete boost) and value of time saved ($ amount per hour or year saved). Using these figures, the financial gain would be perhaps 25% participation x 50% effectiveness x 42 million x $20,000 value per year = $105 billion.
Self-driving cars represent a huge potential boost to capital productivity
In addition to freeing driver time, self-driving cars could represent a significant improvement in vehicle utilization and consequently capital productivity. For example, families could potentially reduce from two or more vehicles down to just one. The car would transport an adult to work and potentially then return home to transport other family members elsewhere. The following table shows the estimate for the capital tied up in U.S. vehicles.
U.S. Vehicle Capital
|Vehicles Type||Vehicles Registered||Average Value ($)||Total Capital ($ Billions)|
|Light duty vehicle, short wheel base||192,513,278||4,000||770|
|Light duty vehicle, long wheel base||41,328,144||6,000||248|
|Truck, single-unit 2-axle 6-tire or more||7,819,055||8,000||63|
Source: Department of Transportation - Bureau of Transportation Statistics, Department of Transportation - Federal Highway Administration, Kelly Blue Book, author estimates and calculations.
The following table repeats the analysis completed earlier for each vehicle class to determine the average MPH, time spent traveling, and hence average vehicle utilization.
U.S. Vehicle Utilization Rates
|Vehicle Type||Average VMT||Average Speed (MPH)||Hours Traveled||Utilization|
|Light duty vehicle, short wheel base||10,650||24||441||5.0%|
|Light duty vehicle, long wheel base||15,474||24||641||7.3%|
|Truck, single-unit 2-axle 6-tire or more||13,476||24||558||6.4%|
The key observation in this chart is the high amount capital tied up in light duty vehicles with a correspondingly low utilization. Trucks and Buses have much higher levels of utilization. If the speeds were a little higher, the hours driven would be a little lower and hence the utilization would also be lower. This would strengthen the case that there was additional upside in better utilizing the assets. Utilization does not consider loading or unloading times nor does it include maintenance times (including pumping gas).
Limits to Utilization Improvements
In glancing at the single digit utilization rates for the vehicles classes above, it is important to recognize that these figures will never approach 100% and perhaps not even 50%. While I looked to calculate utilizations based on distance traveled to include congestion and traffic jams, it does not include maintenance. Nor do these utilization rates typically account for truck and vehicle loading and unloading times. Furthermore, the one car family scenario considered earlier creates additional inefficient utilization. In many cases the miles traveled by the vehicle may double as many trips may become round trips in which a vehicle travels to one destination and then autonomously returns home to take passengers somewhere else. Furthermore, vehicle utilizations will be much higher during the day than at night as one would expect.
Measuring the Potential Financial Gain
The potential gains can be approximated using target utilization rates and an estimated return for the redeployment of that capital. For this analysis, I will only using light duty vehicles (LDV) since these have very low utilization rates and represent the bulk of the capital assets. There might be some meaningful gains for long haul trucking as well.
The other major consideration is that vehicle utilization will also result in additional deterioration of the vehicle, raising maintenance costs.
|Vehicle Type||Current||Round-trip||Target||Capital Savings ($ B)|
|LDV - short wheel base||5.0%||10.1%||20.1%||385|
|LDV - long wheel base||7.3%||14.6%||29.3%||124|
Source: Author calculations
This above table reflects a lot of simplifying assumptions. The first set suggest the number might be higher and include assuming full participation and assuming the cost of a self-driving car is equivalent to a current car. There would also be some drag from higher maintenance costs and additional fuel to manage return trips.
The target utilization rates might also be too high. However, if all LDVs were replaced with self-driving cars, average driving speeds could increase and utilizations would decline. Furthermore, there could potentially be other benefits from fewer automobile accidents.
So the financial gain would be the redeployment of that capital of $509 billion. If you assume a 5% return, the result would be about $25 billion. This figure is more than twice GOOG's 2012 income from continuing operations figure of $10.8 billion.
Through its self-driving cars initiatives, GOOG is addressing an area where there is a large amount of capital that is inefficiently deployed and significant upside from labor savings as well. The above analyses suggest that the labor savings are still substantially more than capital redeployment benefits.
While there are still significant technological advancements and legislative changes required, the outlook is promising. This represents one more aspect of optionality embedded in GOOG stock that is not yet represented by positive free cash flow. While GOOG would probably capture only a portion of the value created, it could still be a substantial number. Furthermore, self-driving cars would also enable other disruptions to existing business models (e.g., long haul trucking, taxis, car sharing) that could provide additional benefits.
However, I still think it will be several years before you can send you car to fill itself up and get washed while you are watching the NFL on television or reading articles on the internet.
Disclaimer: This article is for informational and educational purposes only and shall not be construed to constitute investment advice. Nothing contained herein shall constitute a solicitation, recommendation or endorsement to buy or sell any security.