What is the total passenger capacity of all street-legal cars in the U.S.?

Factual Questions
1

IOW, if you could put a person in every seat of every street-legal car in the U.S., how many persons would that be?

This would be excluding vehicles designed for mass transit (buses and the like) and commercial trucks, but including all personal and family vehicles other than motorcycles (but including pickup trucks, SUVs, and minivans in addition to your typical 4-5 passenger cars), and including similar vehicles used for commercial purposes (e.g. taxis, rental cars).

[The motivation for asking this question, if you’re wondering, has to do with how things would change if self-driving cars become a reality. Once you think about it, it’s pretty obvious that we have a tremendous overcapacity here with respect to cars - that the cars we have can carry far more people than we have. (E.g. my wife and I have 2 cars, total capacity 9 people, but there’s only 3 people in our household.)

My expectation is that the easy access to extra car capacity resulting from self-driving cars would reduce the need for individuals to own excess car capacity. But that made me wonder: just how much excess capacity do we have?]

2

Based on what I see driving in ATL, there is a great over abundance of capacity.

Most cars on the road are single driver cars.

Where I work, there are roughly 3,000 employees at that location, we have parking capacity for somewhere around 3,000-3,500 cars, and most slots are full daily.

So if you took a typical car capacity of 4 (SUVs and Vans more but just in general), that would be somewhere on the neighborhood of 12,000 people capacity, for 3,000 people (very few carpool or ride share, or take the bus or other forms of transit).

Throwing in the proliferation of SUVs and Vans, I would give it more of an average of 5 per vehicle, so then you are at about 15,000.

Every area is of course different. Cars are very common in ATL, probably more common than in say NYC. We could potentially cut down dramatically on needed capacity, but then it will be difficult to get people to give up that ultimate freedom.

3

The other aspect of this is the limited amount of time a person uses their car. For example, I drove to work today, by myself in a four-seater. Then the car sits there for eight or nine hours until I drive home. If, instead, I took a self-driving shared/public car to work, even by myself, the car could then be used by others.

This is the sort of thing that shared car services like Zipcar are exploiting.

4

There’s also a huge amount of nonsense in some of these breathless ideas of reduced traffic.

The roads are crowded at 8am because everybody is going to work at once. That won’t change just because the ownership of the cars changes.

If shared fleets are to be widely used, they have to be sized for *peak *demand, not the 24-hour average demand. At least until we can get lots of workers of all kinds onto staggered shifts spread evenly around the clock. Fat chance of that.

Said another way, if by magic we could replace the entire US fleet overnight with shared zipcars, traffic would be utterly UNchanged.

The thing shared fleet ownership *might *enable is a *small *reduction in the need for parking at workplaces. Some fraction of the shared cars will be needed during the day to meet the demand at 9am, 10am, 11am, 12noon, 2pm, etc. Until at 4pm the whole fleet again is simultaneously deployed to move everybody home at once. My WAG is that 1/3rd of the vehicles will be needed during the day and 2/3rds will simply be stored between commuting peaks. In the evening 10% will be needed and 90% will be stored. Overnight it’s 5%/95%. All the unused cars have to park. Someplace.
If all these shared vehicles were all magically self-driving, maybe we could cram more cars into the roads and have less congestion- and accident-related slowdowns. But that would be equally true if all the self-driving cars were individually owned versus fleet-owned.
The other idea is that shared cars, whether human or computer driven, could accommodate more than one passenger with different pickup points and different destinations. It’s a good bet that if we could put every person’s typical commute into a database and match them up in pairs we could take 20% of the cars off the highways at the expense of increased driving at both ends to pick up and drop off the second person.

Anyone who’s ever used a shared van shuttle service to/from an airport is familiar with the process. And with it adding about 50% to the duration of most journeys. It’s not obvious to me that’s a product that’ll sell. How much would you need to save on total car costs to be willing to add 20% to your commute time? 50%? Remember this extra 20%/50% is every single workday, not just that one time you’re riding to the airport to go on a weeklong trip.
Traffic density is essentially proportional to vehicle miles driven per area per time. Every moment a vehicle is shuttling between pickup/dropoff points is added mileage versus the solo trip case. IMO the net effect of shared ride vehicles, human- or computer-driven, will be to significantly increase traffic in suburbs and downtowns and commercial industrial parks while slightly reducing traffic on main highways / freeways. Remember the tragedy of the commons applies here: Your ride is faster if everybody else doubles up. But your ride also is faster if you don’t.
For autonomous shared vehicles, every mile the car travels with no person on board is a deadweight cost to the operating company and a deadweight increase in traffic for the rest of us.

In downtown Big City during the weekday the typical deadhead distance might be a couple miles of milling around waiting for a call; not too bad. To carry 50% of commuters that way the deadhead mileage might be 40 miles per passenger; big difference. Uber / Lyft’s experience to date with human driven taxis is much closer to the former case. By and large people don’t use those services today for routine mass market commuting. But that’s exactly what the end-game use case is for fleet-owned cars, human or autonomous. As such, current experience data will not scale directly.

I know Uber / Lyft know this. But lots of media commentators and regular Joe’s like us don’t seem to grasp the difference.

5

This table from the Bureau of Transportation Statistics tells you how many of each type of vehicle there were on U.S. roads as of 2014. http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/publications/national_transportation_statistics/html/table_01_11.html

The table shows a total of 260,350,938 highway registered vehicles of all types plus another 62,449 motor buses. Making guesses about how many people each vehicle could hold if filled to the gills, I come up with a total capacity of 1.072 billion people. If I can figure out how to post a table, I’ll show you how I came up with the estimate.

6

My assumption (which I’d rather not debate in GQ) is that if self-driving cars become a reality, cars will evolve to reflect the trips we largely take: single-occupancy cars. No back seat, no ‘passenger side’ of the car either. Half the width = twice the lanes, without having to build any more roads.

7

I was thinking on the self driving stuff after reading an article the other day.

For me, it seems that putting more efforts into technology and changing mindsets to allow for remote working situations would seem to have the largest impact.

I have worked at the same location for 22 yrs and have been driving from the same starting point to that location for the last 15 yrs.

In that time, the distance of 23 miles has of course not changed, but my commute has gone from 45 mins to 1.5 hrs.

That has happened due to the increase in cars on the road due to housing changes and so forth.

I can easily do my job daily from home. I have the technology, my company has the infrastructure, but due to “appearances” they do not want us to work from home more than 1-2 days per week.

I spend 3 hrs a day being unproductive driving to and from home to work, to do the exact same thing as I can do at home.

If that mindset changed, we could easily reduce more than 1,000 cars per day off the ATL roadways from my company alone.

I know many companies are forward thinking, but lots are still caught up in the past.

Sure in a completely driverless car situation, I could hotspot and work while riding, but I get car sick, so it is still mostly an unproductive 3 hrs in the car getting to and from work.

8

It’s unlikely that the self-driving car-sharing service vehicles would be filled to capacity all the time. They’ll likely have a capacity of 2 or more passengers, but most of the time they’d be used by just 1 person.

I think it’s more interesting to look at the amount of space used for storing (parking) our cars when they are not in use. Look at all the parking space necessary for every office complex, campus, big-box store, etc, not to mention the driveways and residential streets wide enough to park on. Of course self-driving cars will need to spend some amount of time at charging stations, but they don’t have to be located in city centers.

9

Well, 2 days a week is a 40% drop in commuting… nothing to completely sneeze at.

10

For what it’s worth, here are the numbers I used in generating my estimate above. Note that this assumes that every motorcycle and commercial truck is carrying 2 people. This seems unlikely.


						     individual  total 	
Vehicle type				number	       capacity  capacity
----------------------------------------------------------------------------------------------------
Light duty vehicle, short wheel base	187,554,928	      4	 750,219,713
Light duty vehicle, long wheel base 	52,600,309	      5	 263,001,547
Two axles and six or more tires 	8,328,759	      2	 16,657,517
Truck, combination  	  	  	2,469,094	      2	 4,938,188
motorcycles 	 	  	  	2,577,197	      2	 5,154,395
buses	  	  	  	  	872,027	             35	 30,520,956
Motor Bus	  	  	  	62,449	             35	 2,185,715
11

You are entirely correct. And I let my manger know how much I appreciate that.

My main point was, there is no actual reason to limit to 3 days in and 2 out. They trust me to work at home two days a week, but 3 I would be goofing off? But more people teleworking really does change the amount of traffic, much more than the hypothetical driverless cars would.

12

Many jobs are a mix of solitary work and collaborative work. The mechanical engineer on our team usually works from home 2 days a week, and that’s perfectly fine with me, but we still benefit from having face-to-face discussions a couple times a week.

13

I definitely get that.

But my job is not.

Most days I am in the office I do not even see my boss unless he physically walks by my cube and then it is just walking by.

I rarely collaborate with anyone.

Most of the people I work with are not even in my work complex.

And I know policies are not about me, but most of my group is like that.

I have actually be told by my boss that the main reason we have the arbitrary limit is because the HR at the location will not set a policy and are leaving it up to the managers, but they are afraid that someone will notice we are not at our desk the majority of the time and will tattle on us and then it could cause someone to decide to reel it back in.

14

You have two choices: store the excess cars near where they’re going to be used later, or have them contribute to traffic as they shuttle from their last load to some remote storage area.

All those cars that hauled all the people into the city center in the morning will be needed again to haul them all home in the late afternoon. So we either store them in the city center as we do now, or we have 4 traffic jams a day:

  1. Bring people from 'burbs to city center in the early morning.
  2. Ferry cars in late morning to remote parking / charging in the 'burbs.
  3. Ferry cars in early afternoon from remote parking / charging in the 'burbs back to city center.
  4. Bring people from city center to home in the 'burbs in the late afternoon.

Obviously we can do some of each. But the idea that the cars just disappear after you step out of them, or that they can all (or even most) be productively used during the off-peak times is just not true.

15

Self driving cars that do not belong to any particular individual can be parked in a much more compact way than traditional cars, so the parking problem would be reduced significantly (50-66% less space per car of the same size) because of that.

I would also imagine that many of these vehicles would be 2 passenger and usually only have 1 person in them. They could be the size of smart cars and with computers, they can drive more safely close together, and reduce traffic problems.

I could also imagine that the daily end of work routine of getting into cars would make people familiar with who lives near them, and ad-hoc carpools would form. There might even be social pressure to not hog up a whole car to yourself, if/when carpooling reduces the wait for coworkers in line behind you.

Even in cases where two people hop on the same car because they are going in the same general direction, and it would be out of the way to for each to go to the other’s destination first, there is a good solution. If there are park’n’rides near freeways, then the car can stop at the point where the two routes part ways, one rider gets directly into a second car, and both are back on their way within a minute or two. If there is a computer with good knowledge of passenger needs, including the two endpoints, it could become very efficient at creating ad-hoc car pools. That does have the promise of reducing total traffic.

I have noticed that there is a lot of traffic caused by a driver dropping off a non-driver, and then going back home, and repeating the trip again the same day. This is double the driving that is needed, and is even sometimes rewarded by letting the vehicle use the carpool lanes. This is the same issue as the empty self driving cars returning to a holding area, and since a certain amount of no-passenger driving is already effectively happening, that cost will be partially self balanced.

16

Then what happens when the family wants to go to the beach on the weekend? We take four cars, including the 2-year-old in their own car??

17

That’s sounds right, certainly in the ballpark based on more or less common knowledge of 250 million or so ‘passenger vehicles’ in the US with average capacity probably at least 4 (some non-crew cab pickups and sports cars at 2, but a greater number of larger cars/SUV’s at >4).

I wonder how much driverless cars per se would change demand though. Maybe indirectly and in conjunction with the growth of car sharing methods, also fueled by technology. But if you start with the model of everyone drives their car to work (say) alone and the cars stop needing you to drive, that doesn’t itself, seems to me, call for either smaller or fewer cars. I can see how eg. Uber might work with smaller cars if the Uber driver isn’t taking up space for the driven party, though it still depends on the distribution of number in party. A Smart Car would work for the presumably common party of 2 w/o driver but not with a driver, but obviously not for a party of 3 where a 4 seat car works either way.

In general it would seem car sharing methods would be the main thing making people abandon having their own car, assuming constant costs. But as in general with the issue of less/smaller cars for less consumption/waste the direct way is to make it more expensive to own/run cars. That definitely works, whereas ways which make it easier/cheaper to use cars tend to ‘backfire’, in that respect at least, by encouraging more travel. That’s true of Uber IME living in a dense urban area: people now Uber distances they used to walk which aren’t convenient to bus/subway let alone finding a parking space at each end.

I’m leaving out, obviously, the critical issue of whether a particular public agrees it’s other people’s right via govt to deliberately make it more expensive for them to drive. In the US with the means up to now that’s proved a total political non-starter. But in various other countries govts already do a lot to deliberately raise driving costs in order to encourage smaller cars and less use.