How the taxi-medallion bubble might burst

By Felix Salmon
January 20, 2012
sharp rise in taxi medallion prices over the past few years?  

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

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:

traxi.tiff

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.

7 comments

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/

Felix: “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.”

This statement seems contradictory because the value of a medallion depends on the amount charged to drivers per shift, not the driver’s income. So if the drivers’ income drops by more than 19%, it would seem the value of the medallion would drop by less than that.

But more likely, the income of the drivers would drop by less than the income of the medallion owners as there is a lower bound to what they are willing to work for and we are probably close to that bound already. (Otherwise, what prevents the owners from raising the charge to the drivers now?) And this does not take into account costs such as fuel and maintenance that are also significant and that will drop by less than 19%.

So overall, even a slight decrease in trips could result in the value of a medallion crashing by a much larger amount.

Posted by TSTS | Report as abusive

Wow. This might make cab drivers angry, causing them to start beeping, yelling, and driving reckless all the time.

Posted by bryanX | Report as abusive

If taxis spend 40 times as much time in congested areas as private cars do, why don’t we get rid of all the NYC taxis and have everyone drive in their private cars! There, congestion problem solved.

Oh wait a second, cities which push their residents to use cabs (e.g. Hong Kong) have much less congestion than cities which don’t (e.g. Moscow). So maybe we need to think a little bit harder.

Posted by johnhhaskell | Report as abusive

@JohnHaskell – Felix isn’t suggesting that everyone use private cars instead of taxis, he’s merely explaining that the congestion effect from adding 2,000 taxis (in Manhattan/B/Q and driving 24-7) is similar to adding 80,000 private cars (driving in, parking for 9 hours, driving out).

If you were to replace X number of taxis with exactly X number of private cars, then yes, street congestion would go down. And then taxi fares would go up (or demand would drastically outstrip supply resulting in long waits, which has the same economic effect if time = money), black cars and gypsy cabs would flourish, parking garage rates would rise, subway ridership would go up and current patterns of mobility would change.

Note that you could only limit private cars in order to enforce the congestion decrease with strict governmental laws/regulations. You’d need a parallel medallion system (“right to enter Manhattan”) for private cars. Otherwise your second paragraph would come true (fewer taxis, many more privates, increased congestion).

Posted by SteveHamlin | Report as abusive

Maybe I’m missing something here. If fare rates do not drop due to increased competition, demand for taxis should remain unchanged. A few people unsuccessful in hailing a cab before might now “get lucky”, or a few more in wheel chairs may become new taxi customers.

Days ago on another Reuters commentary on this subject I responded (in part): “The number of “hired trips” is determined by the number of riders, not cabs; so there should be NO effect on traffic from the same number of “fares”.

Your data bears out what I said. You show taxi trips per day (riders/fares) of 472,200 increasing to 472,000, or only 0.02%.

We need to know taxi miles driven per shift. If a given taxi is one of those that gets in the taxi queue at the airport until their”turn, goes out, and returns to do the same thing the queue is longer, “dead time” longer, and time “in traffic” less. If a given taxi is “on the prowl” in a fare-rich area, they are “in traffic” their whole shift even though they get fewer riders/fares in a shift. What will be the “mix” of these “new” 2,000 medallions?

You show a journey time reduction from 19.2 to 19.1 minutes. No “added congestion” evident here. But you also claim a purported drop in travel speed of 12.1 % (from 9.5 MPH to 8.4 MPH). These are obviously figures at odds with each other. What am I missing?

Posted by OneOfTheSheep | Report as abusive

@OneOfTheSheep — I composed the analysis that motivated and underlies Felix’s post.

What I’m calling journey time is the sum of in-cab travel time and hail time. With more taxis, the former goes up while the latter goes down by roughly the same amount (a coincidence). Journey time thus goes down almost imperceptibly, causing demand for taxis to rise, also almost imperceptibly (the 0.2% you cited, or intended to).

Medallion taxi fares are 100% regulated, so you’re right that they won’t drop despite the increased competition. I also agree with your sense that wheelchair accessibility will add little to demand.

What *will* change is the percent of taxi time spent cruising (fareless), from 35% now, to 40%.

You can see this and lots more in my spreadsheet model. Please download my report (via the “hard math” link in Felix’s post). The link to the model is under References, in the back.

Posted by Komanoff | Report as abusive

@Komanoff,

Your explanation makes sense. Thank you for your detailed response!

Posted by OneOfTheSheep | Report as abusive