Ridesharing as an Alternative to Transportation Capacity Increases

» To what degree can we rely on people getting into strangers’ cars to reduce the congestion on our highway networks?

Outside of the biggest, densest cities, transit generally underperforms; with smaller populations, less significant destinations, more diffuse congestion, and far more available parking, there is often little motivation for people to abandon their cars in favor of jumping on the bus or the train. As a result, the work commuting mode share for public transportation in most metropolitan areas in the less than 5% (only five regions have shares above 10%).

Carpooling, on the other hand, attracts more than 7% of work trips in all major metropolitan areas. In many places, where public transportation options just are not particularly appealing, sharing an automobile with another person can be an excellent commuting alternative, especially for people who cannot afford to own their own vehicles.

But how useful can carpooling be in answering a city’s transportation demand? Could its use be expanded significantly to encourage more people to get out of their single-passenger automobiles — without having to set down the millions of dollars required to build and maintain transit systems?

Last weekend at the Congress for the New Urbanism’s Madison conference, Paul Minett of New Zealand’s Trip Convergence made a compelling case. Most people drive their cars to and from work, but leave them sitting unused 95% of the time, he argued — we should make their use more efficient. In many cities like his own Auckland, 94% of road space is used by single-occupancy automobiles. And transit’s frequently minuscule mode share means that even if it doubled its capacity over the next ten years, it would still account for few overall commuters.

Mr. Minett advanced the notion that ridesharing — a system by which people get out of their own cars and essentially share a vehicle with other commuters taking similar routes to work — could triple or quadruple the capacity of existing roads, all without any added physical investment, the stuff no one can afford anymore. Unlike traditional carpooling, which requires people who know each other to share itineraries, ridesharing allows trusting strangers to save the cost of commuting.

From an urban environment perspective the benefits of encouraging more people to choose to ride together rather than drive separately are indeed significant: Just like with transit, reducing the sheer number of cars entering the central business district at rush hour can play an important role in cutting the number of lane-miles required on the roads downtown and throughout the city, and it opens up more developable land for buildings rather than parking. And the reduction in overall transportation infrastructure necessary to keep a city functioning means fewer expenditures by local governments.

Problems that have made ridesharing difficult for many people to adopt in the past have revolved around the issue of coordination. Finding people who share similar routes is not always easy, and the result can be frustrating waits on the side of a busy road, hoping someone with a sign advertising a similar destination shows up. But new technologies that virtually everyone can take advantage of through devices like smart phones solve many of those problems. They allow people to find others who are traveling the same routes as they are, thus reducing the difficulty of aligning origins and destinations.

In some cities like Washington, D.C., where ridesharing is referred to as “slugging,” as a mode it has made a successful dent into congestion and reduced the individual cost of toll roads. In the San Francisco metropolitan area, 10,000 people “casual carpool” everyday. If a lack of tax funds for capital investments in transportation is a problem, thus encouraging the charging of people to use highways, making ridesharing work well in places with few transit alternatives could reduce opposition to tolling.

Cheaper “transit” options, like downtown circulators and bike and car sharing, could ensure that in central city areas people are able to get around during the day without automobiles. As part of any city’s overall effort to encourage transportation alternatives, ridesharing should clearly be considered.

In his speech, Mr. Minett admitted that there remain significant obstacles to the wide-scale introduction of ridesharing. The ease of jumping in a car or — when it’s available — catching the next bus or train is partially based on the independence of others those modes afford.

The effort to expand ridesharing’s use has other structural difficulties, namely the fact that any corridor that works well for ridesharing would probably work better for a good transit line. Fundamentally, a place where ridesharing is appealing is one where either significant congestion exists and thus where a city or state implements carpool-only lanes or where traffic control measures like tolling are implemented. In the first case, people want to share a ride because it reduces their trip times (because they can travel in the free-flowing carpool lanes) or because it reduces their costs (because they can split the toll with the others in their car).

In other words, cities that suffer from few highway backups (and which therefore have no carpool lanes or tolls) provide their citizens few incentives to share rides (or use transit).

On the other hand, in most corridors where there is significant congestion, good transit is arguably in demand and necessary to attract people out of their cars. Is it a surprise that the cities with the lowest carpooling mode shares in the U.S. are those with the highest transit mode shares? As long as it is appropriately designed, with high frequencies of service and good accessibility, transit is the more popular option.

Finding the right spot for a ridesharing program, then, is likely no simple matter.

Image above: Carpools only sign, from Flickr user Richard Drdul (cc)

30 replies on “Ridesharing as an Alternative to Transportation Capacity Increases”

It seems like rising gas prices (in addition or instead of tolls and carpool lanes) are also now providing enough of an incentive for people on tight budgets commuting long distances. Given that gas prices are increasing everywhere, carpooling might become an increasingly attractive option for a large number of people in a variety of areas, potentially helping to reduce congestion in areas that could not otherwise be served by cost effective public transportation.

It’s important to note that carpooling’s share has declined significantly since the Census first started tracking it in 1980, from 20% nationally back then to an estimated 10% in 2009. From the available data, the biggest drop occurred between 1980 and 1990 (20% down to 13%), so it’s hardly a new phenomenon. It could very well have spiked during and after the ’70s gas crises, though I suspect that’s only part of it.

The fleet of cars out there is so huge that carpooling could regain popularity pretty quickly if people felt the need for it. Continued sprawl limits the ability to rebound to historic levels, though.

Yes, conventional carpooling is undermined by sprawl outer suburban development the same way and for the same reasons that bus ridership is.

Ridesharing is an effort to toss in cellphones and ride matching in as close to real time as possible to boost carpooling share, but given a secular improvement due to those technologies, sprawl of residences and sprawl of employment centers still undermines mode share.

With carpooling focuses on a common workplace, park and rideshare would supplement common carrier public transport, while rideshare park&ride would be more of a complement for common residential area and distinct final destinations all accessible by public transport. Park and rideshare could be a means of increasing access to high occupancy lanes.

Furthermore, sprawl hurts schemes like Zipcar and Ithaca Carshare.

“Slugging” is always going to have minimal attractiveness due to security concerns. :-P

Is it too early to start considering the ramifications of self-driven cars? Although Google’s work seems shockingly successful and far along, I’d be stunned if we saw anything commercially available within a decade. But, suppose they became widely commercially available within twenty. Fleets of timeshare cars, managed by a central reservation and allocation system, could revolutionize ride sharing and flatten the need for business-district parking.

I think the ability to reduce parking is probably what they would be best at, and automated vehicles would likely replace taxis to start out with. The biggest change would be a reduction in the need for parking, which we’ve already got too much of in most places. Do any taxi services actively try to carry strangers on the same or similar trips? I’m not sure how well ridesharing can apply to automated cars — someone should probably try it with taxis first.

Ideally, automated cars will be like personal rapid transit systems, but without the need to build dedicated guideways.

These cars are still going to need parking spaces and I really don’t view this as something that is going to be a major player in the future in that people don’t like riding with other people.

I really don’t think these new types of cars are going to do anything in that most people don’t like the idea of sharing their privet cars with groups of other strangers. Such as I could picture some really funny things happening with this idea. This really is most likely going to effect taxis first and make a lot of taxi drivers lose their jobs. Also what would the legal aspect of this be if you are driving one of these cars and get pulled over and the dude before you deiced to smoke some rocks while driving.

The scheme is that you don’t need parking “downtown” the self driving car would drop you off at your destination and then drive itself to a cheap parking lot outside of downtown. Nevermind that the scheme has all the same problems of owner driven cars and some of the bugs of PRT….

I suspect that these self driving cars would be a easy targot for car thevies to steal in that they are running and there are no people in them to stop them from stealing it.

If it has computers in it that are smart enough to drive itself it’s going to be smart enough to keep itself locked unless it “knows” who it is who is trying to get into it. And smart enough to not run if someone who isn’t authorized to use it is trying to use it. And smart enough to go “home” when if it does get stolen.
…. LoJack on steriods…

Self-driving cars are not going to take off unless they are much, MUCH safer than regular cars. Self-driving trains only took off because, when used without grade crossings on a well-designed system, they are close to 100% safe.

People, as the result of an innate cognitive bias, hold automatic devices to a higher standard than human drivers. Self-driving cars could have half the accident rate of regular cars and they would still not be accepted.

What allowed slugging to take off in the DC area was the proliferation of HOV lanes in Northern Virginia; HOT lanes will diminish their utility. One blog characterized them as BRT with cars, which might be a good way to go about encouraging slugging elsewhere.

Intentionally constructing such a scheme would function like so: construct a BRT line and disallow any vehicle with fewer than 3 occupants to use it. Provision a minimum of four lanes, with two on either side for stops, two in the middle for travel. If the lanes get too crowded, implement rush-hour tolls on top of the carpool requirement.

Octaviuslll, it all boils down to efficient use of lanes.

First of all, it is a waste do build a 4-lane BRT out of the sake of some passing lanes. It would be better to build stop side bays.

In regard of HOT, it all depends on whether is there spare capacity on the lane or not. If there is, it is totally justifiable to allow toll-paying cars to enjoy increased speed up to the point the lane reaches capacity. You can control that tweaking the price of tolls. They are the monetary equivalent of ramp meters.

It’s not just the HOV facilities, it’s the fact both the origin of the trip (park and ride lots) and the destinations for the trips were several highly concentrated job centers (Downtown DC, the Pentagon, etc). The system would not work without that density at the destination, as well as the natural choke point along the route and the advantages for carpools.

I think such schemes might work if technology comes to place to allow some kind of vouching or pre-screening + smartphone app to locate someone in the vicinity wanting to share a ride + automatic payment collection.

There is potential, though limited. Many people, me included, are not keen to share a car with strangers. I’m on the high tail of driving (and I like transit projects that clear roadway for me) because I enjoy not having to interact with stranger while concentrating for my workday and not having my personal space invaded. However, many people would approve and the idea might be worthwhile to explore.

Last couple years many free or paid home accommodation sites gained some degree of robust success on their niche. You have for instance CouchSurfing, allowing people to visit other cities and stay on other’s couch for free, which has an intricate system of vouching and ID verification (like matching login data with a credit card + postcard with code sent to your home). I wouldn’t receive some stranger in my house, but apparently many people would, and the site organizes the whole process.

If it worked up to a point to get people taking strangers into their houses, it might well work to get people taking strangers into their cars. I think of a system that, for instance, requires registration + ID verification + credit card, so that people using it can be at least sure about the identity of whomever is using. Then, via smartphones/internet, people could easily enter their intended route to travel and the application could match those with cars and without cars, handling some sort of payment without direct money exchange (e.g., withdraw $$ from your card, transfers to driver’s account).

As for reasons for why carpool reduced its share on transportation since the 1970s, I’d add a tectonic shift in workplace organization that took us away, for the good and for the evil, from the model of centralized corporate compounds of large factories and large, in-house staffed only, office facilities attached.

When you had an cities like Detroit, Cleveland and many others with a handful of 40.000-workforce locations, mostly working on fixed and rigid schedules, it was way too easy to carpool, as chances were in any major neighborhood you’d have people going for the same 5-10 big employers’ workplaces.

Seems to me that the “campus” model of Google, Apple, Cisco et al is very similar to the old factories which is why Google for example has their own fleet of employee shuttles many picking up riders at rapid transit stations. The difference is that each of these firms elected to build at low density miles from any neighborhood the workers would choose to live in as opposed to the older model of living near the plant. (CTA no longer provides bus service to a large Ford assembly plant one assumes because those workers no longer live in Chicago but in surrounding suburban areas thus no concentration of rider origin) The company shuttle also bridges the stranger gap by filling the bus with similarly upscale badge wearers some of whom might even be from one’s own assembly line.

I know of two similar corporate bus schemes run by Microsoft and Boeing in the Seattle area. They are surely not a comprehensive solution but an important contribution.

As a matter of fact, the concept of corporate campus, where you have all sorts of amenities like gyms, private park and small golf course, tennis court, some ‘relaxation plaza’ with threes and so is all but impossible to implement in a high-density CBD, even if parking space were not an issue. Certain types of corporations (high-tech, bio-tech etc.) are keen to build lavish working environments such that employees feel like they are part of a family. You can have that if the cafe your employees patronize is some independent store two blocks away with no immediate association with the corporation you work in (as it happens with most Manhattan business that are not big enough to have the financial chest to dedicate prime real estate floor space for relaxation, indoor gyms etc.).

As for place of residence chosen by employees, we must account for the fact few of highly skilled employees of firms like Google will live in single-income households. As a matter of fact, I think the issue of dual-income household if often ignored and overlooked on the whole transit discussion: it was feasible for a family in 1954, when only the “breadwinner” was expected to work, to live as close as possible to the workplace, since everyone else would plan accordingly. Nowadays, especially on high-middle class-but-not-rich families, you have two adults working in different careers on different schedules and, unless one of them is willing to give up career choice for the sake of procuring employment only on the neighborhood (not happen if you are a lawyer, MD, architect, engineer), the couple will have to live somewhere distant from both workplaces, or close to one and far from the other.

“Paul Minett of New Zealand’s Trip Convergence made a compelling case” for how he would like the world to be. But in the world as it actually is, the verdict is in on carpooling: people do not like it. We’ve tried a vast array of incentives to increase carpooling (free/preferential parking; HOV lanes; free matching services; advertising campaigns) and the result is that carpooling is on the decline.

There will always be a few special cases (like the SF Bay Bridge corridor) where carpooling can attract a substantial number of trips. But there have been exactly zero successful, pro-active efforts to substantially expand carpooling. Even carpool lanes have not been show to create many new carpools; primarily they reward travelers who would be carpooling anyways.

For the past 10 years I’ve been hearing from entrepreneurs who believe that they have a technological solution that will unlock a vast market for real-time ridesharing. None of these efforts has succeeded, because the problem isn’t technology. They problem is that people tend to not like to carpool (with the social anxiety, safety fears, and schedule limitations that it presents).

breaking the scheduling limitations requires, first, as close to real time matching as possible, and two, either critical mass or an assured fallback in the form of realtime dial-a-ride or jitney cabs.

Thanks Bruce; good points. The close-to-real-time-ridematching it technically trivial. But critical mass raises a difficult transitional problem — you can’t get to critical mass overnight. So, at least initially, you need fallback of real-time dial-a-ride or jitney cabs. If you’re dealing with a fairly large geography (a precondition to carpooling making any sense at all), then you’re talking about a very substantial public subsidy cost for this fallback dial-a-ride/jitney system. Suddenly “free” carpooling has become “expensive” transit.

So, yes, the schedule limitations can theoretically be overcome, though at great expense. That leaves the barriers of social anxiety and safety. Overcome those, and we’ve taken care of the major kisses-of-death. Instead of “hell no”, people will say “yeah, I suppose I could, but why should I?” So lastly you’d need a positive monetary or time savings incentive. Smells like a losing proposition to me.

This is another benefit of the slugging in Northern Virginia – almost all of the slug pick-up spots are co-located with commuter bus park and rides, giving slug riders an additional option to get back if they don’t want to slug or can’t find a ride.

If slugging is simply taking people off a bus/train and putting them in a carpool, there’s no a societal benefit unless the rail/bus is at capacity and the sluggers are helping avoid the need to add rail/bus capacity.

I don’t know the specific circumstances in Northern Virginia. Are these sluggers people who would otherwise be driving, or otherwise be riding transit? Is the transit service operating at full capacity?

A modeling exercise for those who might want to pursue real-time ridematching: What is the necessary critical mass? As a user, I’d want to know that I can get a ride, within 15 minutes of my desired time, that requires less than 10 minutes of extra travel time on either end (for either me or the driver). (Pick your own parameters here.) There are only so many travelers on the road at any time, and only a few are making trips that meet my needs. What share of travelers must be participating in real-time ridematching in order for me to know that I’ll get a match? (Allow some share of trips (say 5%) to be unmatchable and have to use the fallback system; and account somehow for different travel patterns by time of day and location).

I suspect the answer is that even 100% participation is insufficient, with the possible exception of densely traveled areas and corridors, during heavy travel times of day.

Hi Mike. People don’t like paying taxes, yet do. People don’t like having to work for their money, yet do. Obviously both actions are required for a functioning society, and so people get on and do them regardless.

So people don’t like carpooling? I’d first argue that people don’t like commuting, regardless of how. And I’d secondly argue that people’s like for things is irrelevant from a policy standpoint other than for the purpose of working out the balance of incentives and tolls required to make commuters choose to switch.

I think we’re reaching the point with burgeoning population and burgeoning costs of fuel that some sort of methods to increase roading efficiency will be required for things to function, whether people like doing them or not. Use of higher occupancy vehicles seem like a pretty straight forward way to improve transport capacity; and at a much lower cost than building more roads.

Where I live (San Gabriel Valley, NE of downtown LA), we have carpool/HOV lanes on the 210 Freeway. It’s not uncommon to see rush hour traffic in the HOV lane moving no faster than the regular lanes.
A while back, a California state bureau had a media campaign with various celebrities’ mothers extolling the virtues of sharing (as applied to automobiles); I doubt if it did much good. One of the attractions of driving one’s own car is the freedom to listen to your favorite talk show or music without concern for anyone else’s opinion. One can stop on the way home from work without a passenger wondering what’s taking so long inside the store. And the point about large-scale factory jobs is well taken; so many jobs now require weekend and late evening shifts. The sites of three of the factories where I worked in the 1960’s are now occupied by retail stores with extended hours.

Many big employers promote carpooling schemes. Around 2003-4 many universities pushed such schemes, as they have more limited resources.

This is an easier task as everybody has a common terminus on their journey.

On the most simple and common form or carpooling sponsored scheme, employers grant some rewards privileges (premium parking, cash back, forwarded tax credits some cities give for carpooling) for carpoolers that organize and register by themselves, with some matching help from the company/institution.

Then, the employer will pay for free cab rides for each car-pooler during a period (month, semester, year) should they lose their ride back home due to someone leaving earlier than expected or having to stay later than the normal working hours. Of course, this work best with employers with more-or-less fixed working schedules, like unionized places, government agencies or 24/7 factories with fixed turns.

School drop-off also offer a great opportunity for carpools, and many already exist without formal organization: parents team up to take turns in driving kids to the school and taking them back.

While we’re waiting for better transit systems, improving carpooling options seems like a worthwhile endeavor. I think the sea of parking lots at the typical big box store could easily be converted to park and ride lots (during the day, when few shoppers are around), and those lots could serve as origin points for trips into the city/employment center. If those lots were well-established, we’d see private transit options spring up too, so someone could be dropped off at the Wal-Mart park-and-ride lot in the morning and not worry that they won’t be able to get home that night.

Eventually, we may see more mixed-use development at the big box store lots, as land values increase, and establishing these lots as integral parts of a transit system would encourage them to also be transit-oriented developments. While I don’t know if we’ll ever retrofit the exurbs for rail transit, rapid bus transit is a real possibility.

Regarding the point made above about trust, I think the density of connections developed through Facebook, etc., in recent years will have long-term consequences. Even if you don’t know the driver of the car you’re in, you could have access to a wealth of information about him/her. Plus, cell phones will make it much easier for people to pay someone directly to ride in their car, opening up opportunities for private individuals to start van/car-sharing services which would be even more regulated.

However, I agree with those who say that driving itself has to have a much higher cost (either in terms of time or money) in order to increase carpooling rates. For whatever reason, people are very tolerant of road congestion, so with a few exceptions that alone won’t be enough to get them to stop driving solo.

If I’m driving down the highway and a ridematch connection becomes available – there’s an opportunity to pick someone up and head into the HOV lane – Is the diversion more costly in time than what I would be saving? and most importantly: How am I notified? Is this going to cause accidents when people get buzzed on the phone to tell them they have a match. Lastly, how do you get past the dating problem (more guys than gals or, in this case, more riders needing to go somewhere than drivers willing to take them)?

See above ~ how you eliminate the “dating” problem is to have dedicated jitney cabs as part of the system.

Indeed, it seems potentially far more useful as a means of expanding the capacity of a jitney cab system during peak commute than as a stand-alone transit system. You’d have the rideshare cars as a special restricted class of jitney cab ~ eg, to make more than one trip in the same direction in a day, you need the jitney cab license.

Leave a Reply