Viewed from early 2016, the year 2040 seems rather distant. When pondering the idea of what that time might look like in Southern California, there is much to consider. Such considerations should inarguably include vehicles, as they have shaped and will continue to impact our region in a significant and singular way.
The term between now and 2040 will be enough to add an entire, new generation of people to our planet. From a vehicle perspective, the same 25 years will see four to five full generations of cars and trucks brought to us by current carmakers, extant startups, and likely some new high-profile automotive entrants. With this many coming model generations, competitors and potential innovators, the automobile is bound to change. The question is how, at what rate, and in which proportions these changes will impact our future vehicle landscape.
As with all things, the automobile business’ complexity increases with true examination, a fact learned perennially, and painfully, by nascent carmakers. While many reasons are given for this complexity, it is perhaps best just to recall that it is a large system, and one very highly and mutually adjusted to human society. Such systems change constantly in detail, but slowly in total. Within this context, cars have been evolving for more than a century and will continue to do so for many years.
Importantly, on the private car side, adoption of any new vehicle technology is hampered by lugubrious and slowing fleet turnover. Thanks to steadily increasing durability, the average U.S. vehicle is now more than ten years old and climbing. Many vehicles on our roads today, and nearly all of those being sold as this is written, are likely to be on the road more than a dozen years from now. Add to this the reality that when an attractive new vehicle technology is invented, it takes many years to debut and even then is typically first offered only as a premium feature or on premium cars. Many new cars sold in the 10-15 years following the entry of new technology still don’t feature it, and these cars then persist in the fleet for another 10-15 years. Understanding these realities of fleet refresh can be difficult, which may help explain why pundits from outside the industry consistently over-forecast the penetration of innovations into the automotive mix. If we want to plan realistically for Southern California’s 2040 vehicular population, however, it is incumbent on us to do so.
Automation is probably foremost among new vehicle technology in the public’s attention. The physical car itself is already mostly adapted to automated operation, with key systems having moved to by-wire, rather than mechanical, actuation years ago. Vehicles are also increasingly “connected” via communication links carried by occupants and/or the vehicle itself. Such links started simply with navigation and entertainment, but are rapidly growing into torrents of data on everything from vehicle operating systems to passengers themselves. Of course, Google and many others around the globe have recently added sensors and guidance systems to this ready platform to create stunning demonstrations of driverless cars.
Outside of privacy concerns, consumers are not really resisting any of this progress, and mostly look forward to the day when they will not need to operate their vehicles at all. This combination of willing buyers and available technology means that progress towards a connected, automated fleet will continue. Despite liability questions and a patchwork quilt of enabling legislation, cars that are fully self-driving will become available to some consumers in the near future and progress into the wider market thereafter. An interim step, however, has already been occurring.
Between now and 2040, millions of vehicles will be sold in California that are partially automated, offering automated operation under limited circumstances. This is a logical step towards full autonomy, but one many failed to forecast in 2008 when the idea of autonomous operation resurfaced in the public eye. Some partially automated vehicles are already sold today in the form of premium models that can navigate into a parking spot, or keep themselves in a highway or arterial lane, for instance. Inherent in such cars is a “hand-off” and a “hand back” from driver to car, and back again, of operational control. Currently, drivers are trained and practically attuned to full-time vehicle operation, and therefore small periods in which the partially autonomous car takes over are mostly overseen by an alert operator. As automation increases, however, and such periods of automation grow in frequency and duration, drivers’ attention will be more focused on other activities far removed from the task of driving.
Robust consumer research continues to enforce the fact that drivers are eager to spend non-driving time in the automated car on distractions such as social media, reading, gaming, work and even sleep. How to manage operator attention will be a matter requiring adjustment and coordination by carmakers, regulators, educators, law enforcement and insurers over the next several decades, as sales of partially automated vehicles increase. The complexity of this coming second vehicular transition will only be compounded as partially automated cars are later joined by fully automated ones on Southern California roads. The strain on societal systems resulting from such fundamental changes in our roadway mix will in many ways resemble that from the chaos of the dawn of the motor age in the early 20th century, which took decades to work through.
Today, progressive metro areas are attempting to change the balance between automobiles and other land uses and to limit private car presence, objectives automated transport will affect. On the upside, truly smart automated vehicles should allow recapture of some real estate currently devoted to separating manually driven and fallible autos from other roadway users such as pedestrians and cyclists. Many urban planners correctly imagine that such self-driving cars, especially if they “talk” with other users, will allow a reinvented and much more vital and integrated cityscape than our streets currently represent. Such enhanced dynamic cooperation might also come with a reduction or even elimination of curbside parking in such scenes, empty automated vehicles being capable of continuing on their own to less obtrusive parking nearby.
On the downside, cars that can drive their occupants places are also capable of dropping them off and returning unoccupied over much greater distances to home or at least to where parking is free, turning automated vehicles into two way chauffeurs and thwarting attempts to limit car use by rationing parking. Such usage behavior could quite easily increase vehicle miles travelled (VMT) for many cars and, to the extent parking fees are avoided as a result, actually reduce cost of ownership. Calculations using nearly any historical fuel prices show this would today be the case for a hypothetical driverless car in many Southern California areas. Further, at current fuel and many garage prices, automated cars could simply drive around a city unladen for hours more cheaply than they could park there.
Automation may also tend to naturally increase commute distances, travel times and the appeal of the car relative to public transportation simply by making the auto a nicer place to be. An owner who can work on a laptop, or sleep, to and from the office might find commuting from Santa Barbara or Crestline to Los Angeles acceptable on a daily basis. In the same way, many drivers who today inadvertently “peak shave” usage of our highway system by avoiding rush hours might find less motivation to do so when the car does the driving. The effortless travel afforded by automation also becomes a more formidable competitor to most forms of public transportation—especially rail, whose chief current advantages include the promise of allowing passengers to do other things while distance is covered. This propensity for drivers to spend more, not less, time inside autonomous vehicles will also lead toward change in the physical vehicle itself.
Users’ actions in other modes of transport show they tend to demand more interior space, not less, as mere passengers than they do as drivers. Non-driving time is effectively disposable time, and a more relaxing interior befits this. Studies of consumers by automakers continues to reinforce this truth, and several automated concept cars, including one by Mercedes Benz that drove itself from California to the 2015 Consumer Electronics Show in Las Vegas, have demonstrated it. The rolling lounges consumers want for their self-driving cars bear little resemblance to the tiny self-driving pods many futurists have prophesied will be borne of automation.
Trucks and other delivery vehicles, conversely, will continue to be configured more by economic than other factors, and are less likely to grow physically larger with automation. It is hard to imagine a scenario in which most of long haul, and much of shorter distance, goods and service delivery is not highly automated by 2040. To the extent such vehicles can reduce costs relative to manually operated forbears, the fleets they comprise will turn over more quickly than will the private car fleet and accelerate automation in our commercial vehicle mix by 2040. Whether the resulting reduction in operational costs also here brings more VMT, and vehicles, to our streets is yet to actually be determined but seems likely. At least the possibility for some of these automated commercial vehicles to shrink in size truly does exist, given less emphasis on costly human drivers.
Commercial vehicles aside, calculations of autonomy’s rate of availability and incorporation into the mix of our private car fleet are sobering. Absent regulatory intervention, fully automated cars will comprise around a quarter of the private regional fleet by 2035. The balance will be legacy vehicles with either limited or no self-driving capability; one among many factors leading us to a truly heterogeneous vehicle fleet in 2040.
Other key questions around automated vehicles concern so called car sharing. As already discussed, the real impact self-driving cars will have on public transit is far from well understood. Just as uncertain is how much true car sharing will be advanced by automation. Certainly, ride hailing services such as today’s Uber and Lyft will one day benefit from the elimination of drivers and their attendant costs and constraints. Beyond that, there is so far little evidence most households in Southern California or the nation are eager to be emancipated from car ownership, even after automation, through the sharing of vehicles. Conversely, certain demographic groups and households that currently don’t own cars may more often travel in them as true car sharing reduces barriers to light vehicle use. This could increase VMT, and perhaps even operating vehicle populations, especially in high density areas.
In terms of vehicle fuel type, we are also likely to see greater variety by 2040. Electric propulsion will grow, but in terms of natural market demand, mostly in conjunction with internal combustion engines (ICE) power in the form of hybrids. Cars with such hybrid ICE/electric powertrains – from full plug-in to simple start-stop – are very clearly favored by our predominant Southern California duty cycles, and the continued technological development of such vehicles means their popularity with consumers will almost certainly continue. The variety of forms such mixed powertrains take, and vehicle segments they appear in, will increase commensurately. Hybrids are already a plurality in our regional car fleet.
This leaves us, uncomfortably perhaps, still dependent in 2040 on the ICE and on gasoline as a predominant fuel source. While current low gas prices have again shifted consumers toward thirstier vehicles, the actual efficiency of those ICE cars and trucks in terms of MPG relative to mass and power output is unprecedented. Gasoline powertrains have advanced very rapidly since 2000, spurred by tightening CAFE standards. This progress has come mostly via induction, fuel system and transmission efficiency measures, and will likely continue through the next several decades. Alone, and combined in myriad ways with small batteries as hybrids, those modern ICEs’ natural majority share of the 2040 new vehicle mix is assured, at least in a fossil fuel market with any resemblance to that of the last hundred years.
Alongside those gasoline and gasoline/electric hybrid vehicles will be two alternative fuel stalwarts already firmly imbedded in our regional and national fleet today: natural gas and diesel. Compressed natural gas (CNG), with only a tiny share of the private market, continues to have promise for private vehicles, but will likely remain dogged by lack of ideal refueling infrastructure, pricing tenaciously correlated to gasoline’s, and lack of attention among alt fuel promoters. Liquefied natural gas (LNG) is currently favored as a future fuel for the long distance movement of goods, and will likely exist in Southern California by 2040 as a minority in the heavy truck mix. Diesel engines, although slightly more fuel efficient than gasoline, will continue to be market constrained by high purchase costs and emissions of criteria pollutants. Both diesel and natural gas are highly developed, dense power forms, well understood and adeptly utilized by the market for specific uses; but absent sea changes in the two fuels’ prices relative to gasoline, there is little to indicate they will represent very different percentages of the market in 2040 than they do today.
Pure battery electric vehicles (BEV) are currently and will remain a small minority, serving certain natural or created niches. Electric cars are hamstrung by chemical batteries that have progressed little from their historical origins and continue to suffer power densities and recharge efficiencies that are lower by an order of magnitude than other forms of vehicle energy storage and decline with repeated use. Despite massive regulatory, incentive, and OEM efforts over the last decade to overcome this limitation, battery improvements have been outstripped by those of ICEs in the same period. Battery cost makes BEVs unprofitable for OEMs, who subsidize their sales to meet California credit targets. While much is made of projected cost efficiencies from greater battery production volumes, these will almost surely be offset by the costs to manufacturers of increasing mainstream EV range from the currently common 70-90 miles to at least 150 by doubling battery size, something most are now contemplating as a way to remove purchase barriers for many buyers.
Workplace charging is rightly identified as a way to overcome at least BEV lack of range for many commuters, but will remain too costly to be supported commercially in most of the market. Other forms of field charging, though sought enthusiastically by many advocates, are costly and offer little promise to make electric vehicles truly more effective for general use. With regulatory attention and key credit schema now shifting to other fuels BEV will be challenged to naturally exceed five percent of the 2040 fleet in Southern California, with nearly all of that penetration subsidized in some form.
More uncertain is the future of hydrogen, which does not suffer batteries’ lack of power density or glacial refuel time. Another key way in which fuel cell vehicles (FCV) differ from BEV is technological maturity, as the former really are “new” among alt fuels and therefore truly on the steep part of the advancement curve. Several major carmakers are currently strategizing and developing new FCV as a more cost effective way than BEV to meet tightening CARB credit hurdles. Unfortunately, there is almost no retail hydrogen distribution system or, perhaps more importantly for the long term, clear means of sustainable and efficient hydrogen production. While FCV may indeed help some automakers earn ZEV credit counts that look increasingly unattainable via pure BEV sales, significant market incursion is unlikely in the next ten years. By that time, even if hydrogen becomes a truly competitive alt fuel it will be too late for it to be a significant regional player in the next 25 years, thanks again to fleet turnover realities already discussed.
In summary, we might think of our 2040 Southern California car fleet as a sort of automotive soup. Partially automated cars will almost certainly by then be the rule, running alongside a smaller legacy fleet absent automation, and joined by an onslaught of fully connected and automated cars dominating sales (but only a minority of fleet mix) by that time. Those truly driverless, private cars are likely to challenge our system by tending toward being larger, and used more, than their manually driven forebears. Congruently, a mostly automated 2040 commercial fleet, while likely including some downsized vehicles, threatens to continue the trend of more and more goods of all types and sizes being delivered to homes and businesses, adding more VMT and perhaps even more vehicles to our landscape. Car sharing will by then have long been separated from car hailing services, the former making car use more common among carless households, with the latter reaping the rewards of eliminating paid drivers. Powering the majority of these vehicles will be gasoline and gasoline/electric hybrids, followed by minorities of diesel, pure electric and NG powertrains. The wild card of FCV will also be a small minority, assuming some true progress with hydrogen’s creation and distribution is achieved. While the challenges of exactly anticipating this conglomeration of vehicle types are substantial, those of optimally accommodating it will surely be greater still.