One of the best economic journalists in the United States is Megan McArdle of the Washington Post. That makes her error in a recent WaPo article all the more striking. Don Boudreaux at Cafe Hayek has pointed out the error. But I want to do my own analysis because it’s a more general error that I see people make and the first time I saw a friend make was when I was 20 and really starting to “get” economics. Coincidentally, the error my friend made was when we were talking about taxicab economics and he was a cab driver.
In her analysis of the market for Uber and Lyft, here’s the key paragraph in which McArdle veers into bad economics:
The companies’ [Uber’s and Lyft’s] problems essentially boil down to this: The barriers to entry into the driving-people-around business are functionally nil. Whenever the profits in the market rise above a subsistence wage, more drivers will enter, thus competing those profits away.
Notice the last sentence, which is absolutely true as long as we specify what’s being held constant. Whenever the wages rise about the wages drivers could make in other uses of their time, then, if the other components of the job (driving versus other uses) are the same, that will cause more drivers to enter.
But that’s not a problem for Uber and Lyft. That’s good for them. All else equal, Uber and Lyft do better when drivers enter because it keeps down labor costs and gives the two companies more of a chance to make money. Uber and Lyft are not in the driving business; they’re in the transportation business, and drivers’ services are inputs into the production of transportation.
McArdle’s next paragraph proceeds along this erroneous path:
That was the problem taxi medallions had been designed to solve. Drivers still didn’t make much money, because all you needed to get started in the business was a driver’s license. But the people who owned the right to drive were able to make a tidy living, with almost no downside risk. That’s why New York City taxi medallions got so valuable: Two decades of low-interest rates made them an attractive alternative to bonds offering yields in the low single digits.
Her thinking here is that restricting the number of cabs helped cab drivers. This is precisely wrong. Restricting the number of cabs hurt cab drivers but helped owners of medallions; the restrictions caused the medallions to be an artificially scarce resource.
To see the error more clearly, imagine–perish the thought–that the government handed out a certain number of “newspaper medallions” and that that restricted the number of newspapers. How would we know that it restricted them? By noticing that the price of a newspaper medallion was greater than zero. By McArdle’s reasoning, this should help newspaper columnists like her. But it wouldn’t. The restriction would reduce the demand for newspaper columnists and thus reduce the amount newspaper columnists are paid.
Back in 1970, when I was starting to get economics, a cab driver friend of mine in Winnipeg said that if the restriction on cabs were lifted, cab drivers would make less money. I made the same argument then that I’m making here. He didn’t have one of the scarce licenses (they didn’t call them medallions in Winnipeg) and so he would have lost no wealth once the restriction was lifted but would have gained as a driver. Of course, I’m assuming that the supply of labor was not and, in the current case, is not, perfectly elastic (horizontal.) So an increase in the demand for cab driver labor would drive up the price (wage) somewhat.
I pointed out this reasoning, by the way, in my biography of David Ricardo in The Concise Encyclopedia of Economics. See the last paragraph of the bio.
The application to farmers versus owners of farm land was a point that Thomas Hazlett made very clearly in the 1980s in an article or book review either in Reason or in the Wall Street Journal. I’ve forgotten which.
READER COMMENTS
Bob Murphy
May 26 2019 at 3:07pm
Hi David,
I’m more sympathetic to McArdle and (possibly) your friend than you see to be, on this one. But first I need more information: How was your friend able to operate a cab, if he didn’t have the license? Can you explain that situation a little more?
David Henderson
May 26 2019 at 3:31pm
The owner had the license.
Bob Murphy
May 27 2019 at 12:26am
Oh, duh. For some reason I thought McArdle was assuming most taxi drivers were owner-operators. So when she wrote this:
…I thought she was making your point. I.e. I thought she was saying the people driving the NYC cabs weren’t making more money *in their capacity as drivers*, but they benefited from their ownership of the medallions.
But yeah, looking at her whole piece, I realize that’s not what she was saying. I agree with you and Don.
David Paul
May 26 2019 at 5:57pm
Hello David,
I think one can interpret this differently. Look more at the second quote you pulled out. The drivers are still not making much money, but the owners used to do that.
In this analogy, Uber/Lyft are now the owners. Currently they are not earning any money (in fact they are losing money each year), but they also can’t raise their prices later to make it back, because they broke up the system that used to guarantee those higher prices.
Their current evaluations are pretty high – so high that it would only be justified if they earned high profits in the future. But since the old rationing scheme is gone, there is no way to earn those profits in the future. Any time they try to raise their prices in the future, some competitor will come into the market since the start-up costs are very low.
Benjamin Cole
May 26 2019 at 7:43pm
I think David Henderson is correct in his analysis. Whether you are driving for Uber, or for a medallion-owner, you will not make much money if labor is abundant and cheap.
Let us hope for a couple generations of very tight labor markets in the United States. Otherwise, we may see AOC in the White House.
Side note on Uber: as smartphones proliferate and apps improve, how long until people—riders and drivers—learn to connect with each other without Uber as a middleman?
David Henderson
May 26 2019 at 9:23pm
You wrote:
Good question. I think it will be at least 5 years, but we’ll see.
Matthias Görgens
May 27 2019 at 12:30am
The problem is not so much the connection as handling payment, dispute resolution, reputation etc.
My bet is that we won’t see direct connections on a large scale. (Happy to make that more precise and into an actual bet with someone.)
But: if technology or society develops in a direction to make these direct connections more feasible, the margins that Uber and Lyft can get away with will shrink first.
Benjamin Cole
May 30 2019 at 1:39am
I think inertia is the big strength for Uber. Consumers are trained to consume, not think and work a little.
But surely a local app could do Yelp-type scores on drivers, and payments could be in old-fashioned cash. We are talking about a 40% Uber mark-up here. (Cash also benefits drivers who will not report income).
An etiquette of driver and rider uploading photos of each other before ride would be useful. Riders who did not pay would have their photos put into a “They Did Not pay” file. Sheesh, my brother was a cabbie, and some people ran away even in the old days.
I’m told people meet up and expose themselves to others on Tinder, Grinder and who knows what (I am too old for dating).
Cab rides should be easy.
robc
May 30 2019 at 9:17am
Blockchain could help with these, although it would be possible to fake reputation, but not if people only based reputation scores on raters they could certify somehow.
Mark Z
May 26 2019 at 11:11pm
It would be amusing if Uber became the first fully automated corporation. One reason though why the ease of app creation may not create the competition necessary to drive this eventuality is that Uber’s (and Lyft’s) big advantage isn’t really the app, but the network effects. It wouldn’t be difficult for almost any programmer to make another Uber or Facebook; it’s the fact that everyone already uses the existing ones that gives them a competitive advantage.
Matthias Görgens
May 27 2019 at 12:33am
Yes and No.
It’s easy to code up a Facebook or Uber clone that does what most people want (not all of their myriad features), but only at small scale.
Even with that limited feature set, scaling your infrastructure to Facebook’s and Uber’s size is a hard technical problem.
A small operator doesn’t have too worry too much about eg denial of service attacks, or large scale fraud.
ChrisA
May 30 2019 at 5:21am
Brands are even more important online than offline. Look at Amazon’s business – it is entirely dependent on brand – literally everyone else can set up their own web based selling business and undercut Amazons, on price at least. But we all use Amazon because we all know their brand.
So if I arrive in a small city I have never been to before I am not going to do a bunch of research to find a good cab company before I go, I am going to use my already existing Uber app. As a result of their brand having value, I know that Uber will ensure that most likely my service will be good.
So I really doubt Uber brand will be beaten by low cost local suppliers.
Floccina
May 28 2019 at 4:00pm
I don’t know why we aren’t there already. What do we need Uber for? They do give you a little more screening of crazies but rating can do that to some extent.
Jon Murphy
May 26 2019 at 8:20pm
Good stuff. I think I’m going to assign this in my intro micro class in the Fall
David Henderson
May 26 2019 at 9:23pm
Thanks.
Mark Z
May 26 2019 at 10:57pm
Is she not also wrong in her assertion that profits determine wages? Competition between firms drives down profits, not wages. In fact, competition between firms for labor would drive up wages. What keeps wages “at subsistence level” is competition between workers for jobs. And it doesn’t drive wages down to subsistence, but to marginal productivity.
Lastly, even though she is sort of correct that taxi medallions may benefit some cab drivers (the few that have medallions; and this is assuming the cartel’s surplus doesn’t just end up accruing to the companies that sell the medallions), it’s at the expense of all the other drivers or potential drivers excluded from the market. Even from a strictly ‘distributional’ standpoint, this is bad: dis-employing half of all cab drivers to cause the other half’s wages to double would still cause more harm than good, because of diminishing marginal utility (the disutility of losing your paycheck entirely exceeds in magnitude the utility of your paycheck doubling).
Basically, McCardle is ignoring the well-being of potential (or disemployed former) cab drivers (as well as customers) in her analysis. Classic case of seen vs. unseen.
Matthias Görgens
May 27 2019 at 12:43am
The value of the medaillon accrues to whoever is owning a medaillon when tightening measure are announced.
If a city government would introduce a medaillon system from scratch, exactly publish all the details, and auction off the medaillons, they would get the whole value. (Mostly at the cost of drivers and passengers and lots of deadweight cost.)
In practice, medaillon holders lobbied for additional restrictions over a long time. (And I don’t know if any city auctioned off the medaillons?)
A saner medaillon system would auction them off for at maximum a few years at a time only. So that subsequent developments will be reflected in the price.
That’s what Singapore does with its Certificates of Entitlemen (CoE), basically a license you need to operate a car at all in the space constrained city state.
There are about one million CoEs, valid for ten years at a time. Every month about 1% expire and are replaced by auctioning off a fresh supply.
And there’s no confusion like with taxi medallions: no body thinks that the CoE system is there to make wages for people who drive higher.
It’s too limit congestion in an economically efficient way, and raise some revenue while they are at it.
It’s an interesting system, because the auction captures the customer surplus from owning a car. So the Singaporean government could in theory scrap their other tax around cars, and preserve the same revenue on net.
Tom DeMeo
May 27 2019 at 9:34am
Its always more complicated than this.
Your medallion analogy probably wouldn’t play out that way. Consider the NBA, for instance. If the newspaper market ultimately plays out like that, someone of McArdle’s stature would do quite well, while thousands of lesser known columnists would essentially become amateurs or semi pro-columnists.
Ultimately, Uber and Lyft will have problems because they are attempting to build an enormous corporate system by predominantly tapping into marginal part time labor. This is going to be increasingly rejected as politically unacceptable.
David Henderson
May 27 2019 at 3:33pm
You wrote:
That doesn’t contradict what I said. McArdle is very good and so she would survive and do well. What I said is that less competition for her services would hurt her. Those are not contradictory. So instead of making, say, $120K a year, she would make $100K a year. (I’m making up numbers here.) Either way, she does well.
Tom DeMeo
May 27 2019 at 4:46pm
When I claimed she’d do quite well, that wasn’t what I meant.
I analogized the NBA as an example of what might happen if you constrain the number of “medallions” that can issue news content. It probably would create a high marginal return for superstar columnists, even as it restricts total demand for the quantity of working columnists. McArdle might very well be paid far more money in such a circumstance.
That isn’t a great analogy for what would happen in the case of ride service e-commerce apps. They were first, and captured huge profits, but their enormous valuations will be their downfall. They don’t actually provide much value. E-commerce networks are not that expensive to build, and competitors without the investor baggage will clean their clock in time.
Phil H
May 27 2019 at 10:44am
I have to admit, I don’t really understand the medallion argument, so I might be missing something here.
But combining two different ideas, I think McArdle’s got a point.
First she notes that Uber and Lyft have created more perfect markets, which should mean less rent. That’s the point of a market, right? That shouldn’t really be controversial.
Second, she suggests that with less rent, there will be less profit, which sounds about right. I’m thinking here of Thiel’s Zero to One, in which he recommends creating monopolies by doing better than everyone else. Uber and Lyft are past that point, they’re both approximately as good as each other. They are unlikely to be able to pull ahead with new innovations that the other can’t catch.
So I can see her point, at least. I think Don Boudreaux’s point is correct:
“there is no reason to suppose – as Ms. McArdle apparently supposes – that fares will fall to levels below those that enable competently run ride-sharing firms to earn enough profits to remain in business.”
There’s no reason to think U&L will go out of business. But there is reason to think that competition between them will mean that they will never be massively profitable.
mike davis
May 27 2019 at 12:29pm
Megan is clearly (and uncharacteristically!) confused. But David is also wrong—or at least not sufficiently clear—when he says “ Uber and Lyft are not in the driving business; they’re in the transportation business, and drivers’ services are inputs into the production of transportation”. It is much clearer to say that Uber and Lyft are in the business of building and maintaining a network that substantially reduces transactions costs between those who are in the transportation business and their customers.
Like most networks, the ride sharing network has to overcome two problems.
First, there are some level of network externalities. This doesn’t seem like a big problem because it is relatively easy for drivers and riders to belong to multiple networks (I suspect most people who use Uber also use Lyft). Also, while having multiple networks may not allow complete capture of network externalities, there are other benefits to competition between networks (e.g., Uber and Lyft may have different policies for surge pricing.)
The second and more serious problem arises from the fact that the business of providing a ride-sharing network has high fixed costs and very low marginal costs. That means the firms need to charge a price greater than marginal cost, which can be hard to do without attracting competition. This is, of course, not a problem that’s unique to ride sharing networks and it certainly doesn’t mean that governments need to impose regulations on prices and entry. Many businesses figure out a pricing scheme that solves the problem —“ruinous competition” is shibboleth popular with anti-market ideologues and crony capitalists, not economists who look at how real markets work. But so far, at least, Uber and Lyft haven’t figured out how to make that market work. That’s what’s really bugging Megan.
BTW, I want to get my prediction in writing and on the record: I predict that Uber, Lyft or their successors will adopt some sort of two-part pricing, club model. Riders and/or drivers will pay a monthly or annual fee for the right to use the network. I don’t know enough to predict what that fee will be, but something in the neighborhood of a Costco or Amazon Prime membership doesn’t seem unreasonable. I can also imagine tiered club pricing—for example, a frequent business traveler or late night reveler might pay an extra $50 for priority pick up service.
David Henderson
May 27 2019 at 3:35pm
You’re right that I should have said network business, not transportation business.
The tiered pricing with club membership prediction is interesting. Could be.
MarkW
May 27 2019 at 5:30pm
Here’s an argument in McArdle’s favor. When there are barriers to entry in a market, excess profits are earned by business owners. Does this help their workers? It may if they are organized. Think of the old days when commercial air traffic was much more tightly regulated and the Civil Aeronautics Board limited the routes and gates allocated to airlines. Unionized pilots did VERY well under this system. Fares were high and so there was much lower passenger volume and fewer pilot jobs than there would have been under an open system, but those few who were lucky enough to land an airline pilot gig earned a lot. Now, there are vastly more airlines and flights and fares are much lower and there are many more pilot jobs, but those jobs pay much less.
How much of this is also true of cab drivers? One difference is that medallions traded on the open market, so the profits of medallion-owners were limited by the need to service the loans for the medallions. This factor may have resulted in taxi drivers being no better paid than Uber/Lyft drivers. What do the data say? How do the pre-Uber earnings of NYC cab drivers compare to the average earnings of Uber/Lyft drivers now?
David
Jun 2 2019 at 2:12am
So this is more of a behavioral economics issue than pure economics. The drivers do actually (think) they care more about the utilization, how often they have a paying customer in their car, as this is how they earn money. They don’t see how the market will adjust to keep their earnings at the cost of their labor.
For lyft/uber: less cars on the road means they are getting paid more per hour, as they have more fares. As long as the amount they are paid to carry a fare doesn’t change, this stays true. Of course, as we have seen more cars will join, until the earning per driver drops to just above their costs. Lyft/Uber doesn’t particularly care one way or the other where this lands, as long as enough cars are available to provide satisfactory service to their clients. The total amount running through their platform is the same, and they are getting a slice of the total revenue. I think we all understand this pretty well.
The other case of medallions is a bit harder to see precisely because the inefficient market messes up the incentives. For a non owner driver, they do have an interest in the restricted market, as a limited number of cabs available does guarantee them a more or less reliable income per shift. Of course what they don’t see is that if the supply of cabs increased, the owners wouldn’t be able to charge the drivers the same level to rent the cabs, so the drivers earning would go down, but so would their costs. Approximately. But I do think due to the inficencies they may have been correct in not wanting more medallions.
Comments are closed.