Felix Salmon

When I wrote about New York’s expensive bikeshare scheme last week, I got a lot of pushback from people saying that I was missing the point. These bikes are designed for short trips around town, at a marginal price of zero: the large sums you pay if you keep them checked out for an hour or more are a deliberate attempt to discourage that behavior.

OK, fair enough — but in that case, if you’re charging high variable costs, the converse should be that you should charge low fixed costs. Most bike schemes work in much the same way: you pay a certain amount up front for membership, be it for a day or a week or a month or a year, and then the variable per-hour costs on top of that. If New York’s bikeshare scheme is indeed quite cheap, then one would expect the variable costs to be low.

But they’re not.

Ben Walsh put together some numbers for me, for various cities around the world with bikeshare schemes. Not all cities work in exactly the same way, and some have no direct comparison points with New York at all: in Chicago, for instance, the membership options are 1 month, 2 months, and 3 months. In any case, here’s what we managed to find, for New York’s three options:

It’s pretty clear that New York is at the top end of the range, here: only Frankfurt rivals it in price. To have access to New York’s bikes for one day, you need to spend $9.95 — that’s more than six times the £1 one-day membership in London, and it’s significantly more, too, than you’d pay in Washington or Toronto or Paris.

The first reaction of New Yorkers, when they hear this, is surprisingly positive: they think of it as a tax on tourists, and everybody likes taxes on tourists. Let the tourists pay through the nose for their one-day memberships, and we locals will save loads of money, through an implicit cross-subsidy, on our one-year memberships.

But that doesn’t seem to be the case, either. Look at the cost of one-year memberships, and New York is still top of the league table, at $95. That’s the same as Toronto, and more than seven times what Romans pay. (Indeed, the one-time membership fee in Rome, at €10, is barely more than a daily membership in New York.)

New Yorkers, then, are being asked to spend much more money per year than bikeshare users in just about any other major city — even if they never take out a bike for more than 45 minutes. And what worries me is the deterrent effect that these prices will have.

The first trip you take, on one of the new New York bikes, will cost you at least $10, and possibly as much as $95. Cab rides don’t cost much more than that, and you can fit four people in a cab. Experienced urban cyclists like me will definitely cough up the $95, even if that hurts a little, because we know how convenient it can be to be able to take one-way bike trips in Manhattan, especially if it’s going to rain later, or if you don’t like biking back in the dark, or if you got in to work on the subway but then just need to go a mile or so to your lunch meeting.

But the great promise of the bikeshare scheme is that it will get people onto bikes who have never biked before — people who are generally very nervous about biking at all on busy urban streets. Those people are going to want to try before they buy, and the $10 cost of a trial one-day membership is high enough to give them a good excuse not to bother.

The East River Ferry is a great example of the benefits of low entry costs and the opportunity the city’s bikeshare program is missing. When the ferry was reintroduced last June, it was free for the first 12 days. Passengers flocked and even once full fares kicked in, it remained far more popular than planners projected.

That said, from a user’s perspective there are two costs worth considering before you opt into one of these schemes, and the dollar cost is only one of them. The other one is convenience — and on that front, New York’s 10,000 bikes look as though they’re going to be everywhere you might want one, so long as you stay south of 60th Street; there are even a few docking stations in Queens! In that respect, getting on a bike is (fingers crossed) going to be much easier in New York than in most other cities — and I can definitely see how that convenience might be worth paying for.

I don’t think the pricing for these bikes is going to cause the plan to fail: New Yorkers are used to paying lots of money for convenience. I just wish there were some cheaper way of getting New Yorkers to try these things out. Because $95 is enough money — roughly the same as an unlimited monthly transit pass — that a lot of people will simply not bother.

New York’s new bike-share program, sponsored by Citibank to the tune of $41 million (plus $6.5 million from MasterCard), will go live at the end of July, and the prices are public already. Transportation commissioner Janette Sadik-Khan called them “the best deal in town short of the Staten Island Ferry.” Which, not really.

The Staten Island Ferry is free, of course, but that aside, New York transportation has a very simple pricing scheme. To a first approximation, all rides, whether on the subway or the bus or some combination of the two, are $2.50, no matter how long they are.

And the bikes cost a lot more than that.

As with all bike schemes, there’s a base price to enter the scheme — $10 per day, $25 per week, or $95 per year. Then the first half-hour of bike riding is free (45 minutes if you’re an annual member); after that, you pay on a per-ride basis as well, starting at $4 when you bike for more than half an hour.

The $10-per-day cost is already a significant expense: that’s four subway rides right there. And then the hourly charges really start to rack up if you keep the bike for some length of time. If you take the bike around Governor’s Island, for instance, and stay there for a couple of hours, you’re likely going to end up in the 3-hour time bracket, which is $49. On top of your $10 daily rental. As Garth Johnston puts it, for any real let’s-bike-around-the-city plans, you’re definitely going to be better off just buying your own bike.

What’s more, New York is significantly more expensive than similar schemes in rival cities like Washington and London. Here’s a chart of the cost of one trip, based on a 24-hour membership:

London’s 24-hour membership is just £1, or $1.61, and the cost for the second half-hour is only another £1. Which means that anybody can get on a bike, ride it for an hour, and pay just £2 — less than the cost of a journey on the Tube. In New York, by contrast, getting on the bike costs $10, and then the second half-hour costs another $4, for a total of $14. That’s more than four times the cost of the London bike. By the time you’re on the bike for 90 minutes, the New York cost goes up to $23; you’d need to be biking twice as long to pay that much in London.

Here’s the full chart, going out to the maximum charge for 24 hours:

As you can see, none of these schemes are exactly friendly towards someone just taking a bike and using it to bike around for, say, six hours. But if you do that in London, you’ll “only” pay $58: in Washington, it’s $85, and in New York, it’s $131.

I can’t think of any other area where London is so much cheaper than New York: it’s just weird to me that New York would set the prices for this scheme so high. Maybe the problem is that they haven’t found a lot of places to put the docking stations, so they’re having to set the price high to keep the demand in check. All we’ve been told so far is that the plans for the docking stations will be available “soon”; it’ll be fascinating to see how many of them there are in the first instance.

But one thing’s sure: the price difference between renting a bike and hailing a cab is very small in New York, while it’s very large in London. Which probably makes cabbies very happy, while doing very little to reduce congestion.

Everybody’s favorite transportation geek, Charles Komanoff, has a fascinating new paper out on the economics of New York’s new Tappan Zee Bridge. The old bridge is decrepit, and needs to be replaced — everybody agrees on that. And the replacement is now in the works, at a cost of $5.2 billion. But does it need to cost that much? Komanoff makes a strong case that it doesn’t.

I won’t try to summarize Komanoff’s paper here. Instead, I’ll just point to one fact which is buried there. The new bridge comes with a combined bike/pedestrian lane, 12 feet wide. And the cost of building that lane — the amount that the cost of the bridge would decrease if you simply built it without that lane — is an astonishing $400 million.

To put that number in perspective, Komanoff tells me it would cost roughly $40 million, in the same 2015 dollars, to build two bike/pedestrian lanes on the Verrazano Narrows bridge — lanes which would get vastly more traffic than the one lane on the new Tappan Zee.

As for the cost of the first three years of New York City’s ambitious bike program under transportation commissioner Janette Sadik-Khan, that was just $8.8 million, 80% of which was paid by the federal government.

In other words, for the $400 million which governor Andrew Cuomo is planning to spend on a white-elephant bike lane almost nobody is going to use, you could utterly transform the bicycling infrastructure for millions of New Yorkers in all five boroughs.

Oh, and I almost forgot — it looks as if the old Tappan Zee bridge is going to be converted into a bike/pedestrian walkway anyway, making such a facility on the new bridge even more superfluous.

But this is how big projects always work: it’s weirdly easier to raise billions for something huge than it is to add millions to an annual budget somewhere. “Gridlock” Sam Schwartz, for instance, in his clever new congestion-pricing plan, is proposing three new massive bike/pedestrian bridges: one from Jersey City and Hoboken, in New Jersey, would span the Hudson River and land just north of Chelsea Piers. A second would go from Long Island City and Hunter’s Point, in Queens, and would cross the East River to midtown Manhattan. And the third, and most ambitious, would start in Red Hook, in Brooklyn, head over to Governor’s Island, and then continue on to the Financial District.

These are utterly wonderful ideas. If beautiful new pedestrian bridges can be built by Santiago Calatrava in Venice or by Norman Foster in London, there’s no reason New York can’t follow suit. Still, it’s a bit depressing that we don’t seem to have the mechanisms to take the billions available for vanity projects, and use some small fraction of that money for things which would make a huge difference to the daily lives of millions of New Yorkers.

This phenomenon isn’t confined to government, of course: anybody working in a big corporation has seen some huge acquisition made, using money which was never available for smaller projects from existing teams which had much clearer benefits. And there are hundreds of museums around the world which never have money for important things like conservation, but which somehow manage to find enormous sums for glossy new starchitectural projects. Basically, people want to be able to see where their money is going, in the form of something large and grand and headline-grabbing. Even if there are much more sensible uses for it elsewhere.

Many thanks to Mark Bergen for finding me this data; I asked him for it because I thought that maybe we could learn something from the way in which China has managed to keep employment growing steadily through some extremely turbulent economic times.

What you’re looking at here is total Chinese employment from the All China database. Primary industry is commodities, basically, including agriculture; secondary industry is manufacturing; tertiary industry is services.

It comes as little surprise to see that agricultural employment has been falling steadily for 20 years. But it is surprising to see that if you take out the services sector, total Chinese employment has been going nowhere, and basically falling, for the same amount of time.

Caroline Baum, using a different data source, says that China lost 15 million manufacturing jobs between 1995 and 2002; according to these figures, employment in “secondary industry” was flat in those years, going from 156.6 million to 156.8 million before starting to rise again and reaching 218.4 million in 2010. (It’s worth pausing here to appreciate the sheer scale of this chart: each horizontal line is another 100 million workers.)

Meanwhile, the services industry — tertiary industry — has been on fire: it now employs 263 million people, more than are employed in secondary industry, and has doubled since 1992. All this, remember, in a country with more or less flat population growth, thanks to the one-child policy.

Of course it’s hard to find work in the services industry if you’re a rural peasant: tertiary industry is a fundamentally urban thing, which brings me to my second chart.

It comes as no surprise to see that urban employment is growing incredibly fast — 13.7 million urban jobs were created in China in 2010 alone. What does come as a surprise is to see that urban jobs are still in the minority in China — which means that there’s a lot of room for growth going forwards.

In the U.S., we had a huge construction boom in the aughts, which was concentrated on building bigger suburban and exurban residential houses. That’s good for homebuilders and makers of granite countertops, but it doesn’t really boost the economy more broadly. The Chinese construction boom, by contrast, is building cities and roads and crucial infrastructure, which allows the service economy to keep on growing at a torrid place.

Realistically, there is very little chance that global manufacturing employment is going to increase in future at a rate which will provide jobs for a growing global population. If we’re going to find jobs in the U.S. and the rest of the world, they’re going to have to be found in exactly the area where China is finding them — tertiary industry, or services.

How do you create service-industry jobs? By investing in cities and inter-city infrastructure like smart grids and high-speed rail. Services flourish where people are close together and can interact easily with the maximum number of people. If we want to create jobs in America, we should look to services, rather than the manufacturing sector. And while it’s hard to create those jobs directly, you can definitely try to do it indirectly, by building the platforms on which those jobs are built. They’re called cities. And America is, sadly, very bad at keeping its cities modern and flourishing. 1950s-era suburbia won’t cut it any more. But who in government is going to embrace our urban future?

Remember the sharp rise in taxi medallion prices over the past few years? I thought that the price was pretty justifiable back then, in October, although I did have my concerns:

Any time you see a chart like the ones above, you have to worry that there’s a bubble. Plus, there’s political risk: the mayor can print new medallions, making the existing ones worth a little less (but not a lot less, given that the income from medallions is largely fixed).

Since then, however, two things have happened. First, New York City agreed to print 2,000 new medallions — that’s a very large increase. And secondly, Charles Komanoff — you remember him — has done the hard math of what this means for congestion and taxi incomes.

The first thing to understand is that while 2,000 cars might not seem very much in the context of a city which sees 800,000 cars per day, in fact it’s huge. Taxis spend 40 times as much time driving in congested areas as private cars do, so 2,000 medallions is the equivalent of 80,000 private cars. And when you impact the amount of traffic that much in an area which is already highly congested, the effects can be enormous:

The bottom line, here, is not just significantly more congestion, with travel speeds dropping on average from 9.5 mph to 8.4 mph. That inconveniences everybody, of course, not least the cab drivers themselves, who will take a whole extra minute, on average, to get to where their passenger wants to go. That decreases the number of fares they can pick up per day, and hurts their income.

And at the same time, the number of people wanting to take a taxi is likely to go down, when traffic speeds fall, rather than up. If you have fewer people hailing a greater number of cabs, then it’s simple math that the number of fares per cab shift is likely to fall. According to Charles’s calculations, it will fall in total a good 19%.

If taxi fares stay constant, that means a 19% drop in taxi-fare income, per cab. Fares won’t rise enough to cover that fall. And of course the income for the driver is going to fall more than 19%, because the medallion owners are going to be very reluctant to drop the amount they charge the drivers per shift.

All of which implies to me that if there’s a medallion-price bubble, then the introduction of 2,000 new medallions is likely to prick that bubble. And conversely, if medallions keep on changing hands for a million dollars a pop even after those 2,000 new cars are on the road, then we can be pretty sure that there’s no bubble here at all.

Good news over at the Capital Bikeshare website, which has now been updated to make it perfectly clear that you can, after all, use your debit card to pay for a Bikeshare membership. The FAQ,which used to say that memberships require a credit card, now says “credit or debit card”; the signup page, which used to ask for your credit card details, now says “credit/debit card” at the top of that section. All of which means that although it was always technically possible to sign up for a membership with a debit card, now many more people are likely to actually do so.

On top of that, Bikeshare manager Josh Moskowitz tells me that some kind of installment plan should be in place “in the early half” of this year. It’s unclear whether that option will be open only to Bank on DC members, or whether it will be open to everyone.*

I still think that, in terms of getting the unbanked on bikes, the best approaches are always going to be ones which just get the unbanked on bikes, rather than ones which try to get the unbanked to open bank accounts which in turn will allow them to get onto bikes. But this is a move in the right direction — and a move which is probably going to cost some small amount of money for Bikeshare.

Remember that there’s a $1,000 fee charged to your card if your bike is lost or stolen. Now, what’s more likely: that your credit card has $1,000 of spare capacity on it before it’s maxed out, or that your checking account contains $1,000 in cash? I’d say the former, by a substantial margin. So the chances of Bikeshare having to chase down an individual when the $1,000 charge doesn’t go through are surely higher if that person signed up with a debit card than they are if they used a credit card.

So well done to Bikeshare for fixing its site and taking the risk that it might have to do more work chasing down the money for lost and stolen bikes. My gut feeling is that the marginal difference here is small. But government organizations like Capital Bikeshare are always overcautious and risk-averse, and I genuinely thought when I wrote my initial post on this that Bikeshare wouldn’t allow debit cards to be used for memberships at all. Now, let’s keep our fingers crossed in the hope that they come up with some way of being able to serve unbanked people without debit cards at all.

*Update: Moskowitz confirms that the installment plan will be available to everyone.

Remember the importance of counting intersections? Density alone is good, but not sufficient for a pleasant, walkable urban experience: you also need to be able to get from one place to another in a reasonably straightforward, noncircuitous manner. Cities did this naturally before the 1930s, but then urban planners started building cul-de-sacs and other ways of maximizing the effective distance between any two points.

Now, Michael Lewyn reports on some extremely unwalkable street design in a city which most emphatically predates 1930: Jerusalem, of all places. He was staying in a 32-storey residential building called Holyland Tower, whose official rendering shows lots of people on foot and just one car. But in reality, this is not a neighborhood for pedestrians:

To find the area, go to Google Maps and go to a street called Avraham Perrera. You will note that the street is in a section of looped streets that make the typical American cul-de-sac seem like a masterpiece of clarity. As a result, very little of interest is within walking distance, and what is within walking distance is hard to find unless you know the area really, really, really well.

One look at the map and you can tell this is not a walkable neighborhood. Yes, Jerusalem is hilly, but there are lots of walkable hilly cities: San Francisco and Lisbon spring to mind. This area, to the west of the city, is relatively new; it was clearly built with the idea that people would get around first and foremost using their own personal cars.

What’s more, the Holyland development seems to be targeted at Americans, who are used to the suburban lifestyle, like it a lot, and are attracted by developments which can claim to be “surrounded by 15 acres of green park”. Residential towers can be fine things, but they become very bad neighbors when they’re surrounded by nothing.

I suspect that what’s going on here is a classic case of Nimbyism: Jerusalem has a growing population, it needs a lot more residential square footage, but the locals in Jerusalem proper refuse to allow developers to build up. So those developers retreat to the hills, where, attempting to make a virtue out of necessity, they create luxury towers as removed as possible from the bustle of urban life.

I’d be interested to hear about growing cities which are getting this right, rather than wrong. There are lots of cities which predate 1930 and are very walkable. And there are lots of cities and suburbs which postdate 1930 and aren’t. But it’s now been 50 years since Jane Jacobs published The Death and Life of Great American Cities. Which new or growing cities have taken her lessons to heart, anywhere in the world?

American Banker’s Andy Peters has a jolly story about how West Virginia’s United Bank is teaming up with Washington’s bike-sharing program, to help the formerly unbanked have access to this handy form of transportation.

To check out a bike from one of Capital Bikeshare’s 110 solar-powered stations, users must first swipe a debit card or credit card.

Such a system locks out plenty of low-income commuters who use Washington’s Metro trains and buses but lack a checking account or a credit card…

Bank on DC is offering a $25 discount on yearly Capital Bikeshare memberships to those who open an account at either United Bank, a unit of United Bankshares, or District Government Employees Federal Credit Union. A Capital Bikeshare annual membership normally runs $75.

“These are not necessarily high-balance accounts, but a lot of the customers are using their accounts very prudently,” says Craige L. Smith, the chief operating officer of United Bank’s Virginia division. “We think there is real value in establishing those relationships.”

The fact is, however, that there’s a lot not to like here, most of which is elided by Peters. This scheme is not going to get the unbanked onto Capital Bikeshare in any remotely significant numbers, for a lot of reasons.

Firstly, the $25 discount comes only on the most expensive form of membership — the annual membership which the poor and unbanked are least likely to be able to afford. If you want to help bring down the cost of accessing these bikes, then charging $50 just to get started is not a great way of doing that. In many cases, that’s $50 desperately needed for food or rent: buying bike access in eleven months’ time simply isn’t on the list of priorities, no matter how good a deal it might be.

Secondly, the unbanked tend to be unbanked for many, many reasons. Some are good, some are bad. But it’s ridiculous to imagine that getting a $25 discount on a bikeshare membership is going to be enough to persuade anybody to open a bank account. Millions of dollars have been spent on all manner of imaginative approaches towards trying to get the millions of Americans without bank accounts to open one. Few if any of those approaches actually work. This one won’t work either.

Thirdly — and this is key — opening a bank account isn’t enough to get you that $25 discount. A bank account will come with a debit card, and a debit card will get you a bike for either 24 hours or three days. But as the bikeshare website clearly says, “all Capital Bikeshare memberships require a credit card”. If you want to get that $25 discount, you’re going to need not only a bank account but also a credit card.*

At the same time, even if you don’t take advantage of the $25 discount, it’s still a bad idea for the newly-banked to rent a bike even for just one day, using their United Bank debit card. Capital Bikeshare spells this out quite explicitly:

When you join Capital Bikeshare with a 24-hour or 3-day membership, a preauthorization hold of $101 per bike is placed on your card account. This is a not a charge against your account. It serves as a security deposit and will be returned to you when the hold expires. Holds may last up to 10 days, depending on the credit card company. We recommend using a credit card and not a debit or check card when becoming a 24-hour or 3-day member. Using a debit card may result in overdrafts if you don’t have sufficient funds in your account to cover the hold.

It’s easy to imagine someone opening their first-ever bank account with United Bank, using their debit card to pay $7 for one day’s biking, and then immediately getting hit by some whopping overdraft fee because of that $101 hold.

Then again, the possible charges if you rent a bike with your credit card are substantially higher. This is hidden away in the small print when you sign up for a membership:

If Member maintains possession of the Capital Bikeshare bicycle beyond the Permitted Period of Continuous Use, then the Capital Bikeshare bicycle is deemed lost or stolen, Member’s credit card will be charged a fee of $1,000, and a police report may be filed with local authorities.

(The Permitted Period of Continuous Use, incidentally, is 24 hours.)

In order to get that $25 discount, then, an unbanked person in Washington has to first open a bank account; secondly get a credit card; and thirdly sign a contract under which they agree to pay $1,000 should their bike be lost or stolen or taken out for more than 24 hours. It’s not even clear that United Bank is promising to give a credit card to anybody who opens up a bank account under this scheme, but assume that they do: then they’re immediately putting the newly-banked individual into $50 of debt, with no real idea as to whether or when that debt will get paid off. Add in late fees and the like, and the $25 savings starts looking even less desirable.

Jim Surowiecki has a column on layaway this week; United Bank should take a leaf out of that book and offer their customers the opportunity to pay the $50 membership fee at a rate of say $4.25 per month, plus whatever usage fees are run up on the Bikeshare program.

Who, under that kind of continuous-layaway scheme, would take on the responsibility of paying $1,000 if a bike were to be lost or stolen? Bank on DC is the obvious organization to do such a thing. They should post a $10,000 bond to cover the first ten times this happens, see whether the scheme is any kind of success, and then take it from there. If there are lots of stolen bikes and the $10,000 disappears quickly, then the scheme would be an interesting failure — but organizations trying new ideas should be open to failure. And there’s a good chance that the $10,000 would not be touched at all.

What’s really needed, in other words, to get the unbanked onto bikeshare schemes is not bank accounts at all — it’s a way of finding institutions which will accept the responsibility of paying $1,000 should the bike be lost or stolen. If you do that, then memberships can be given out to individuals without bank accounts at all, and they can use any old prepaid card to pay the modest usage fees they run up, without worrying about the $101 hold.

Which institutions would do such a thing? Well, there are a lot of non-profit organizations which work with the poor and try to get them mobility — they’re a good place to start. But there’s another set of institutions which might be interested as well: churches, of which there are very many in the DC area. Churches know their flocks, after all, and might well be interested in giving out memberships to those who need them, and taking on a contingent liability in the process. All you’d need is a single credit card belonging to the church, which could then deal in its own way with any congregant who ran up that $1,000 charge.

The Capital Bikeshare scheme has been built in a very cautious manner, carefully constructed so that everybody with a membership needs to have a credit card associated with that membership. That in turn allows Capital Bikeshare to be sure that it can collect $1,000 every time a member loses their bike for whatever reason. This system was set up ex ante, with no indication of how often such a fine would turn out to be necessary — and it has essentially excluded the unbanked from Bikeshare.

I would much have preferred to see an optimistic scheme at first, with the restrictions coming in later if the cost of lost or stolen bikes turned out to be substantial. But even now it’s possible to imagine ways around these problems, if some well-intentioned group has faith in its members and in the Bikeshare scheme. The Bank on DC promotion, however, is not such a way.

*Update: It seems that the website is wrong about this: you can use a debit card to pay for a 30-day or 1-year membership, and when you do so no hold is put on your account. On top of that, Bank on DC also has a scheme to help defray the $1,000 cost if your bike is lost or stolen. More details as I get them!

Update 2: Details, from Bikeshare:

• You can pay for 30-day and annual membership with a debit card.
• There is no hold on an individual’s account if they purchase a 30-day or annual membership with a debit card. The reason for this is that Bikeshare has contact information (name, address, etc.) when an individual signs up online for one of these types of memberships — but not for 24-hour or 3-day members.
• If the bike is stolen and no police report is filed, Bikeshare would charge the debit card $1,000. If the individual who is responsible for the bike does not have $1,000 in their account, Bikeshare would charge the amount available in the account and work with the financial institution to ensure no subsequent charges can be made to the account until there is a full reimbursement for the cost of the bike.
• Approximately 85 percent of annual members do not incur any usage fees when riding. The average usage fee the other 15 percent incur is between $6-$7 per individual per month.

What’s behind the rise in bus travel in recent years? It certainly seems very impressive, according to the latest research from DePaul University.

Here’s how Bloomberg’s Jeff Plungis characterizes it:

Megabus.com and BoltBus led U.S. curbside bus companies that boosted trips by 32 percent this year as travelers opted to leave their cars behind and surf the Internet while traveling.

And here’s Matt Yglesias, with a slightly different take:

Like Duncan Black, I’m far from certain that the right way to understand this is actually as intercity bus trips substituting for intercity car rides. The way I would primarily interpret it is as these services leading to additional trips that wouldn’t otherwise have been taken. Instead of riding Amtrak to New York once a year, you ride the bus three times instead.

If you look at the data, Yglesias seems closer to the mark than Plungis. Could the massive 30% rise in curbside bus ridership be accounted for by the 1% fall in private autos? Possibly. But it’s more likely that something else is going on.

Both Plungis and Yglesias, I think, miss the elephant in the room, and the obvious reason why the DePaul measurements for bus ridership have been growing at such a startling rate. Here’s how the paper puts it:

The analysis we provide also excludes all “Chinatown operators,” which have significant different qualities than mainstream operators. As a general rule, those carriers listed on the GotoBus.com web site are considered for purposes of our study to be Chinatown operators. Many of these carriers do not invest in a brand identifiable by the paint scheme or insignia on their buses.

Indeed, DePaul specifically excluded the dramatic growth of California’s USAsia Bus Lines, just because they determined that it counted as a Chinatown operator.

The obvious theory, then, is that big operators like Megabus and Bolt Bus saw the huge success of the Chintaown bus market and saw an opportunity there. They brought in branding and professional marketing and wifi and much higher safety standards, and succeeded in taking a huge amount of market share from the Chinatown operators who were never part of the DePaul survey in the first place.

That theory is borne out by my own anecdotal experience: when my friends took the bus from New York to DC or Boston ten years ago, it was normally a Chinatown bus. Today, it’s more likely to be a Bolt Bus, or even a higher-end product like the Limoliner.

In other words, the DePaul data is consistent with total bus ridership actually staying constant, with the recognized curbside buses simply taking ridership share from unrecognized Chinatown operators. In reality, I suspect that bus ridership is growing. Just not nearly as fast as the DePaul paper would have you believe.

As for the much-vaunted wifi on these buses, it’s basically the same as the wifi on Amtrak, or from Gogo in-flight: in a word, crap. If you’re working on a laptop and can download emails or web pages in the background while reading or writing something else, then it’s fine. But it’s pretty much useless for people on iPads, where the lack of multitasking means you can’t read one thing and download something else at the same time.

It seems to me that the travel industry in general has done a very bad job of adjusting to the fact that most wifi-enabled devices these days are not laptops. I even stayed at one pretty high-end hotel in England, recently, which thought that providing an ethernet cable was a perfectly good alternative to providing wifi, and which didn’t have any kind of Airport Express devices or similar that it could lend out to guests who didn’t have ethernet ports on their computers or tablets.

So far, no one’s really cracked the problem of the mobile web — we’re still in a world where connecting to the internet when on the move is far too difficult, and needs to be configured (and often paid for) on a device-by-device basis. Companies like Lightsquared want to change that, but for the time being they’re vaporware, and I’m not holding my breath for them to arrive. Which means that for the time being it’s a bit of a stretch to say — as Plungis, for one, does — that the mobile web is actually changing the way we travel from city to city.

This is a chart of the number of bike commuters in New York. It’s known as the NYC Commuter Cycling Indicator, and it comes from surveys taken ten times per year at predetermined points around the city. It doesn’t give a good count of the number of bike commuters in New York, but it gives an excellent idea of the trends: bike commuting has essentially quadrupled in the past decade, and has doubled over the past four years. Which just happen to be the four years during which Janette Sadik-Khan has run the Department of Transportation.

This is important because it shows just how effective strong leadership can be, when combined with a dedication to creating good infrastructure. And if you delve a bit into the numbers behind the indicator, this comes out even more clearly. For instance: in 2007, the Queensboro Bridge saw an average of 1,292 cyclists per day, about 80% of the 1,626 cyclists per day on the Brooklyn Bridge. By 2011, the Queensboro number had shot up to 2,904 bikers per day — 25% more than the 2,322 cyclists crossing the Brooklyn Bridge. That’s entirely a function of the fact that the Brooklyn Bridge is unpleasant for cyclists, despite the fact that by dint of its location it should be one of the busiest bike corridors in the city.

The lesson of this chart, then, is that if you build bike lanes, cyclists will appear to fill them. That’s fantastic news, since cities with lots of cyclists are always the most pleasant cities to live and work in — even for people who don’t bike themselves. New York City has a long way to go before it can be considered genuinely bike-friendly. But it’s moving in the right direction, and the bike-sharing scheme to be launched next year will provide a massive boost. Let’s hope that now Sadik-Khan has provided the necessary momentum, her successors embrace and extend what she has started.

Traffic congestion is one of those problems to which there is no single silver-bullet solutions. If a city has too much congestion, the main thing it needs to do is build out much better public transportation infrastructure — something which takes many years and many billions of dollars. But failing that, or in the mean time, is there anything else?

Balaji Prabhakar reckons he can try to give prizes to people who drive at non-peak hours; it’s a cute idea, but I don’t think it scales. And then there’s congestion pricing, of course. But right now, there’s surprisingly little evidence that it actually works, and some evidence that it doesn’t, at least in the medium term: in London, congestion went up after it went down, and much of the small decline in congestion can be attributed to better public transportation rather than the congestion charge itself.

The people at GPS maker TomTom, however, have another idea about how to reduce congestion, at least at the margin. TomTom has a lot of Live devices, which are constantly transmitting their location; on top of that, every time the old-fashioned GPS devices get plugged in to a computer and upgraded, they also upload details of where they’ve been and when. And TomTom also has data from cellphone companies about how fast various phones are moving on certain streets at certain times of day.

Put it all together, and you get a very rich database of where to find congestion, both in terms of geography and in terms of time of the day and week.

This information can reduce congestion in one of two ways. Firstly, the information can directly change driver behavior. TomTom claims that it can reduce travel times by up to 15% — which obviously reduces overall congestion in and of itself a little bit, since that car is on the road for less time. And if we ever reach a point at which 10% of drivers are using that technology, then you see further benefits: enough cars get rerouted from busy to less-busy streets that the busy streets clear up a bit and move faster. Overall, TomTom reckons that congestion gets reduced for everybody by 5%, even if they’re not using any GPS device at all.

And then there’s bigger changes which can be made at the municipal level. TomTom has a whole business which sells congestion information to cities, which can then use that information to inform minor decisions like traffic-signal timings, and major ones like rebuilding roads. Historically, cities have needed clunky hardware like cameras and in-street loops to detect how much congestion there is; the TomTom data is cheaper than that, it’s much more precise, and it covers all roads rather than just the main arteries.

This is great news, right? Yesterday, for instance, we got detailed information on pedestrian congestion in NYC. The “pedestrian volume index” has now reached 113.2 from 100 in 2007, and we know about individual blocks with great specificity: on West 14th Street between Hudson and Eighth, for instance, there were 11,166 pedestrians per day between 4pm and 7pm in September, compared with 8,911 in May and 7,055 the previous September.

That’s very useful information — but it pales in usefulness when compared to equally detailed information on automotive congestion. Finally, we can settle once and for all whether the pedestrianization of Broadway has resulted in higher or lower traffic speeds. We can find out what happens to congestion when bike lanes are installed. And we can measure congestion overall in the city, with an eye to working out how much the deadweight cost of congestion is, and whether some kind of congestion charge might make sense. Most obviously, the real-time TomTom data is good enough that it can and should be used to adjust traffic signals in real time, as and when traffic jams appear.

So, is this happening? Turns out, that question is very difficult to answer. I spoke to a couple of TomTom executives three weeks ago, and they told me that they had done a little study, just for their own edification, on the Herald Square area, to see what happened to traffic speeds there when Broadway got pedestrianized. I asked if they would send it to me, and they said that they’d be happy to — they just wanted to check with the city of New York first.

And then, last week, I was told that a New York Department of Transportation Deputy Commissioner asked TomTom not to share its data with me.

TomTom is still looking for some data, somewhere in the world, that it can share with me — but I don’t have anything yet. And this data is very hard to find, for anybody outside TomTom: the company informs me that even TomTom’s own non-profit, The Traffic Foundation, which says that it will be “tapping into TomTom’s community of 50 million users”, has no access at all to this wonderful and enormous traffic database, which now has more than 4 trillion datapoints.

If this data were a little more public, researchers could plug away at it to see what the effects of infrastructure changes are on congestion, for instance, and to help create a set of best practices for what kind of interventions get the best congestion-reduction bang for the buck. But that doesn’t seem to be happening. And TomTom is very quiet indeed about which municipalities have bought its data, and what those municipalities are doing with it.

There’s an amazing opportunity, here, to use advanced statistical and quantitative techniques to painlessly improve city life: it’s exactly the kind of thing that the TED people got so excited about when they decided to award their TED Prize to The City 2.0. But I worry that TomTom is keeping this data far too close to its chest, and that cities like New York are similarly unhappy about the idea that it might become public in any way.

Of course, TomTom is a public company, with an obligation to its shareholders to maximize the amount of money that it can generate from its database. And it doesn’t want to annoy potential clients, especially not when its share price has been bumping along at very low levels for the past couple of years. Sooner or later this information is bound to start getting used widely — can’t Google just buy TomTom, already, and make it much more public? The entire company, after all, has a market capitalization of less than a billion dollars. The profit motive, here, shouldn’t get in the way of the massive public good that can come from this database.

Emily Badger has found a fantastic paper (not online, sadly) from the University of Connecticut. Authors Chris McCahill and Norman Garrick took aerial photographs of New Haven, Hartford, and Cambridge, and started counting up the number of off-street parking spaces over time.

There’s a general tendency, in American cities, for the number of parking spots to rise inexorably — and that’s exactly what happened in New Haven and Hartford. As McCahill and Garrick write:

If places were to grow, it was assumed that most of the growth would be served by automobile, so new development would require supplemental parking facilities. In the city of Hartford, Connecticut, city officials stated in 1972 that, “the most critical improvement to [neighborhood shopping districts] which could be made at this time is the provision of off-street parking facilities”. In 1982, responding to the claim that his city had more parking than any other Connecticut city, New Haven Mayor Biagio DiLieto stated that he was “strongly committed to maintaining and improving parking facilities for workers, shoppers, and visitors in the downtown area”.

The authors add, waspishly:

Applying this thinking to U.S. cities, without knowing any other information, one would expect that the cities with the greatest increases in parking over the past fifty years have also experienced the greatest growth of development and activities. And conversely, the cities in which parking has not increased substantially might be struggling to achieve growth.

Of course, that’s not what happened at all. New Haven had 21,690 parking spots in 1951, and 106,410 in 2009; Hartford went from 47,000 to 141,000 in the same time period. But both were shrinking, rather than growing, in population.

Meanwhile, Cambridge took a different tack, and decided in 1985 to essentially ban the creation of any new parking spots. That decision marked the beginning of a reversal in its population trend: it started growing quite impressively. Here’s the chart:

Parking lots are — with only a handful of exceptions — the best possible way of destroying a city’s soul. They’re gruesome, lifeless places, and I’m constantly astonished by the way in which governments and developers are convinced that they’re a great idea. Instead, local government should act as a brake on private developers’ desires to build out new parking: while that might (or might not) be good for an individual commercial operation, it can at the same time be bad for the city as a whole. Cambridge is living proof that this can be done: other cities, including New Haven and Hartford, should follow its lead.

Shelly Banjo’s article about the multi-year waiting lists for parking spots at Connecticut train stations is going somewhat viral, for good reason:

The waiting list for a Fairfield Parking Authority permit has 4,200 people and stretches past six years…

“It’s like season tickets to the Giants—even when you’re dead they get passed down to your children,” said Jim Cameron, head of the Connecticut Rail Commuter Council…

Connecticut’s parking crunch is, in large part, a problem of supply and demand: More than 60,000 commuters head toward Manhattan on Metro-North’s New Haven train line on weekdays, but transportation officials say stations have public parking for nearly 20,000…

John Eck, a former television executive from Fairfield, kept his permit after he left his job last spring—”just in case” he needed to start commuting again.

“You hear horror stories of people missing the renewal deadline and losing the permit in other towns,” Mr. Eck said. “I wouldn’t give it up for anything.”

Eugene Colonese, the transportation department’s rail administrator, said the task force “came to a certain point and well, stopped its work for a little while.”

He said the department is still “looking for the best way to get commuters to stations, a balance we think will be between building more transit-oriented development, looking at shuttles and other public transportation, as well as parking improvements.”

The parking lot at Fairfield train station is big enough for 1,053 cars; the station sees 2,942 people, on average, ride in to NYC, and the waiting list now has 4,278 names on it. These are all big numbers. The price for a spot, however, is low: just $340 per year. Obviously, that’s well below market, and causing all manner of problems. But there’s another number that’s lower still:

We recently spoke to Director of the Fairfield Parking Authority, Cynthia Placko…

Placko told us there isn’t room for many more than the 24 bike lockers that are already there, and those are totally filled.

My guess is that it really isn’t all that hard to take the space given over to 1,053 parking spots and use it effectively to house transportation for 2,942 people. Unless, that is, those people are all taking up the space of some enormous SUV.

In a place like Fairfield, it’s hard to raise the price of parking so much that you start to incentivize car-sharing directly. So here’s my proposal: rip out a bunch of car spaces, and replace them with covered, secure parking for bicycles and scooters. Maybe motorbikes, too. Surely that’s an obviously better way of getting commuters to stations than giving them each a couple of hundred square feet of massively underpriced prime Connecticut real estate, and then acting shocked when they flock to the opportunity.

Update: Fairfield could even buy back parking slots for more than they were sold for, and convert them to two-wheeled parking. Continue to do that until there’s one empty two-wheeled parking space. And then auction off the rest to the highest bidders.

What on earth is going on with the price of taxi medallions in NYC? Two of them just sold for $1 million apiece — that’s a 42% increase just since August, when conventional wisdom had it that $705,000 was a top tick and that medallions would soon plunge in value.

Derek Thompson reckons that this is a sign of how great New York’s economy is doing:

It’s all about supply and demand. The tailwind behind medallion inflation is a cap on taxi cab licenses. Even as the economy of New York City grew at a furious pace across three decades, the number of taxi plates stayed basically constant, despite wage growth and population growth and rising demands for cross-town transportation. As a result, their value rose tremendously.

One problem with this theory is that if you look at the chart of medallion prices, it doesn’t seem to bear much relation to the strength of New York’s economy. The majority of the rise in price has happened over the past ten years, which haven’t been particularly great, economy-wise. And certainly there hasn’t been any citywide boom in the past two and a half months.

Jacob Goldstein has a similar theory.

Every cab in New York City has to have a medallion. And the city strictly limits the number — currently just over 13,000. So as New York has prospered over the past few decades, the price of medallions has gone through the roof.

The economics of this, however, don’t make a lot of sense. Cabs might get a little bit busier when economic activity picks up, but most of that extra income is taken home by drivers, rather than the owners of corporate medallions. (Corporate medallions are the ones where the owners don’t drive the car; they’re the ones being charted here.) The income from a corporate medallion is pretty steady, and was spelled out in the comments to this post:

The maximum amount a cab leased out to drivers can earn over a year is $82,524. This assumes that the cab had a driver every day and every night and that it never breaks down. The day shift maximum charge is $105. The maximum night shift charge is $129 and that is only for Thursday-Saturday. Sunday-Tuesday is $115 and Wednesday is $120.

You’ve also left out of your calcultions the cost of the car. Let’s call it $27,000 and you are required by law to buy a new one every 3 years.

This is the real reason why medallions are so expensive: good old-fashioned interest-rate calculations.

We’re basically talking about a real income stream, here, of about $75,000 per year. (Let’s assume, for the sake of argument, that the income from a taxi medallion rises at the same rate as inflation.) That’s a real yield of 7.5% on a $1 million investment — which isn’t half bad at today’s interest rates.

Put it this way: how much would a bond paying a real yield of $75,000 a year cost? At the most recent auction, the 29-year TIPS cleared at an interest rate of 0.999%. At a 1% real yield, an income stream of $75,000 a year would cost you $7.5 million.

Now you don’t actually get $75,000 a year if you own a medallion. You have to pay for maintenance, insurance, and workers comp; you also have to pay someone to manage your drivers. But even if you bring the income down to $50,000 a year, that’s still a pleasant 5% yield on your money, and what’s more it’s a yield which behaves much more like a real yield than a nominal yield. Paying $1 million for such a thing doesn’t seem silly to me, especially when there’s a lot of room for capital gains as well.

Of course, there’s risk here too. Any time you see a chart like the ones above, you have to worry that there’s a bubble. Plus, there’s political risk: the mayor can print new medallions, making the existing ones worth a little less (but not a lot less, given that the income from medallions is largely fixed).

That said, medallion owners have a lot of political clout, and historically they’ve been good at making sure that their income is maximized, rather than suffering anything which might reduce it. And when Goldstein starts dreaming of taxi deregulation, he quickly enters cloud-cuckoo land:

The medallions create a textbook example of what economists call rent-seeking behavior: Basically, gaining extra profits without providing extra benefits. If the number of taxis were allowed to increase (and if cab fares were unregulated), the number of taxis would increase and the price of a cab ride would fall.

No! if the number of taxis were allowed to increase, then the number of taxis would increase. Of course. That’s just a tautology. But the price of a cab ride would not fall, because that’s a separate piece of legislation — the price of a cab rate is set at a predetermined rate, and indeed has to be set at a predetermined rate. If you deregulated cab fares, utter chaos would result — New Yorkers would basically have to haggle over the cost of a fare every time they got into a cab.

Goldstein’s not the first person to have this cockamamie idea: Jim Surowiecki said the same thing in 1999. But in order to have a market where prices are set by supply and demand, people need to be able to choose how much they’re willing to pay to take a cab. And you can’t do that when you’re standing on the sidewalk (not in the bike lane, please!) sticking your arm out and trying to hail the first cab to turn up. In order to make that transaction work, the fare schedule has to be set, in advance, by the municipal government.

But I do worry that the way fares are set, too much money ends up going to medallion owners. If fares were brought down, the amount that medallion owners could charge drivers would also come down, and medallion prices would — finally — start to fall. Why does NYC ever raise taxi fares, when the income from those fares ends up going overwhelmingly to a handful of millionaire medallion owners? These medallions, right now, are licenses to print money. That’s why they’re getting extremely expensive. But it doesn’t need to be that way.

Traffic expands to fill the space available. This is known as Down’s Law of Peak-Hour Traffic Congestion, and has been known since 1962; new research shows that it’s true even more generally than previously suspected.

Increasing lane kilometers for one type of road diverts little traffic from other types of road. We find no evidence that the provision of public transportation affects VKT. We conclude that increased provision of roads or public transit is unlikely to relieve congestion.

Eric Jaffe draws a simple conclusion from all this:

Whenever a driver shifts onto public transportation, another one quickly grabs the open lane. That leaves just one solution to the traffic problem plaguing American cities: congestion pricing.

“We cannot think of any other solution,” says Gilles Duranton, the paper’s co-author. “As soon as you manage to create space on the road, by whatever means, people are going to use that space. Except when people have to pay for it, of course.”

I’m a fan of congestion pricing. But I’m also realistic about it, and the fact is that for all Jaffe’s enthusiasm, congestion pricing has its own Down’s-like characteristics. Jaffe raves about the “success” of congestion pricing in London and Stockholm, but the only chart he provides gives numbers for the trial period in Stockholm. If you go looking for recent data on traffic in London, Stockholm, or other cities with congestion charges, it can prove surprisingly difficult to find.

Here’s Jaffe, again, quoting Duranton:

“My feeling is, yes, people tend to be against it before they see it at work,” says Duranton. “They think it’s going to cost them more money, which directly it will, but they’re all very unclear about the benefits; i.e. traffic is way more fluid, way faster, and pollution is going down.”

There’s another way to look at this phenomenon, though. When congestion pricing is first introduced, people recoil against it — they expend quite a lot of effort to avoid the charge, and traffic goes down. Over time, however, it becomes just another part of the cost of driving, along with gas and insurance and parking tickets. As that happens, traffic goes back up again. Congestion-charge revenues go up too, of course, and those can be reinvested into public transport.

But traffic is like water — it wants to find its own level, which tends, in cities, to be maximum capacity. If you want to implement a system which keeps traffic below maximum capacity, then you need to apply significant pressure on drivers to keep them away from the roads. And that means not just implementing a congestion charge, but also regularly increasing the amount of the charge over time.

This is how the Singaporean congestion-charging system works. Think of a shutter-priority camera: you set the shutter speed, and then dial the aperture so that the exposure is correct. In Singapore, they set the amount of traffic they want, and then dial up the congestion charge until they get it. It’s much the same idea as the one behind SFPark: you set the number of empty parking spaces you want, and then dial up the parking-meter pricing until you get there.

But the point in all of these cases is that the charge has to be variable over time — specifically, it has to increase over time. Without those steady increases, drivers become inured to the congestion charge, and traffic will go back up to its former level.

As a result, drivers are pretty much never happy with congestion pricing. Either it’s painfully expensive and going up in price — expensive enough to keep them from driving — or else it doesn’t have much effect.

That doesn’t mean that congestion pricing isn’t good public policy. It is. But it’s always going to be unpopular with a powerful constituency. (Drivers, in nearly any city you care to mention, tend to have a disproportional amount of political clout.) Local politicians looking for a popular platform will run on reducing or abolishing the charge, or at the very least not increasing it. And so the old fight keeps on being fought over and over again: while increasing a charge isn’t as politically difficult as introducing one, it’s still tough.

This is something worth remembering when urbanists start waxing utopian on the subject of congestion pricing: once it’s introduced, the fight isn’t over. It’s never over. And if you leave a system long enough without increasing its price, its efficacy starts declining dramatically.