Uber Increases Productivity

Uber drivers carry more passengers per mile driven or hour worked than do taxi drivers. In other words, the Uber system is more productive than the taxi system. That’s the big finding from a new paper by Judd Cramer and Alan B. Krueger.

On average, the capacity utilization rate is 30 percent higher for UberX drivers than taxi drivers when measured by time, and 50 percent higher when measured by miles, although taxi data are not available to calculate both measures for the same set of cities.

Four factors likely contribute to the higher utilization rate of UberX drivers: 1) Uber’s more efficient driver-passenger matching technology; 2) Uber’s larger scale, which supports faster matches; 3) inefficient taxi regulations; and 4) Uber’s flexible labor supply model and surge pricing, which more closely match supply with demand throughout the day.

Krueger co-authored an earlier paper on Uber drivers commissioned by Uber but this paper was not commissioned.


Alan Krueger's earlier paper referenced by Cowen included some questionable claims. http://www.huffingtonpost.com/2015/01/22/uber-drivers-pay-study_n_6527470.html. I can't access this new paper and it may be more objective since Krueger wasn't paid by Uber this time, according to Cowen. I suppose credibility has a price. I am paid to be an advocate (I'm a lawyer), so there's no doubt whose interests I promote. Is it acceptable among economists for an economist to be paid to advocate a client's interests? And if it is, how does one know if the economist is being objective or promoting the interests of a (current or former) client, which may or may not hire her again for another research paper?

Tut-tut, counselor: "Uber this time, according to Cowen [AlexT]". You would not make a good proofreader. And there's lots of billable hours in proofreading, as any associate doing discovery can tell you.

Of course, Tabarrok (sorry, not Cowen) mentions that Krueger was paid by Uber for the first paper but not this paper, which suggests that Tabarrok believes the two papers deserve different consideration. My view, the realist's view, is that nobody is totally objective about much of anything (including the weather) and that everything said or written has a slant. Indeed, even the questions one asks reveals a slant. On the other hand, does my slant go up or down based on how high or low the fee paid by my client? Probably not. An advocate is an advocate is an advocate. That's why lawyers can do such a good job even when doing the work pro bono publico (for the public good). But they (me) would prefer to be paid.

Economists, as well as scientists in general, ideally only make claims that can be objectively verified (loosely speaking), in the sense that their validity is independent of the person of the speaker. Hence every reader, ideally, should be able to conclude from the internal logic of the paper whether the argument is convincing or not, which means that who funded the research is irrelevant.

Yes, but dismissing a claim on its own terms and the terms of its supporting data takes work. Much easier to avoid updating your priors by just attributing all results that don't comport with your worldview to bad faith.

So we should ignore everything we read until we've had the opportunity to independently replicate (or disprove) it? Seems inefficient.

As long as the sponsor of the research is clearly labelled, it should be fine, no? We should be able to evaluate the research on the basis of its merits.

Large payments to politicized think tanks seem questionable though. I recently learned that the Koch brothers have given $4.3 million to the Fraser Institute in Canada, which helps to explain why they are so keep on promoting AGW skepticism and adopt such pro-oil stances in general (Koch's are heavily invested in Alberta tar sands).

If you can evaluate the research on the basis of its merits, why do you need the sponsor to be named?

Also, would you yourself promote causes that conflict with your beliefs if you were paid money to do so? If not, why do you think other people are likely to?

If the sponsor is named, you open your eye a little wider.

Also, it reduces the temptation to engage in outright dishonest work under the guise of imperfection and plausible deniability. Because there are always some journalists and politicians who will quote whatever suits the purpose of their narrative. When the sponsor is named, the credibility of both the researcher and sponsor are more highly at stake, and implausibly "stupid" research eventually serves to discredit both enormously.

No,l I don't think many people are being paid to promote causes that conflict with their beliefs. Rather, money seeks out the most credible possible voice to promote their dubious outlooks/perspectives.

Funny thing is, the former Canadian government passed laws to have special scrutiny targeting environmental NGOs because they receive support from citizens of other countries (usually large numbers of small donations), and this was painted as a threat to sovereignty, yet $4.3 million from one family alone was never something that suggested any such similar threat to sovereignty.

Anyways, Conservatives in Canada over the last decade were always very much a "do as I say, not as I do" sort of group, and they do not show much signs of learning after their recent electoral defeat.

One additional factor which may be driving Uber efficiency is payment speed, for short haul rides uber is probably saving the driver/rider 1-3 minutes in payment processing time. For a lot of urben transit where the ride can be <10m that adds up pretty quicjly, could probably account for a large portion of the overall efficiency.

As any regular user of urban cabs can attest, you are subjected to a poisson shaped distribution at the end of your ride as you try to get the credit card machine to work, driver fumbles for change etc., I'm increasingly guilty of almost walking out on my cab drivers when I use the service instead of uber, becoming so used to just stepping out and having payment automated. Not sure what the process is from the uber drivers side, but it seems much more efficient.

We would need data to (in)validate this claim. But given the difference in the capacity utilization mention, a difference of 1-3 minutes (2 minutes on average??) would require a turnover rate on the order of 9 minutes for the taxi driver.

Seems dubious.

'but this paper was not commissioned'

Any adherents of the Virginia School wishing to bet whether his 3rd paper will be sponsored/commissioned by Uber?

No, but he's retiring to work as a consultant for Uber, salary undisclosed...

I don't have access to the paper.

(1) Do they factor in the notion that is is nearly impossible for an Uber driver to cheat and take cash under the table, whereas at least in my experience, that is common with cabs?

(2) Do they factor in Uber's effect on the productivity of cabs? That is, Uber has been undercutting cabs with a nonprofitable business model for a while now, which has to have some effect on their productivity.

1. That's not a bug, it's a feature. We don't really want taxi cabs cheating their employers and the tax man, do we?

2. Wouldn't pressure from Uber encourage traditional taxi cab companies to become *more* competitive, not less?

1. That may or may not be a good thing, but it's not productivity. This paper isn't talking about thwarting under-the-table transportation dealings. It's discussing productivity. If you want to write a paper about Uber increasing ethical behavior, feel free. But it's not this discussion.

2. Taxis are often subject to price fixing. Operators cannot compete on price. Therefore, it stands to reason, when a loss-leader enters your market, and it's illegal to degrease prices to compete with them, your company will suffer a loss of business and productivity.

Don't you mean it's hard to compete when it's illegal to cut costs by discriminating against classes of customers like Uber and Uber small business contractors intentionally do to maximize revenue?

Or do you think price is unrelated to cost?

Which classes of customers do you allege Uber is discriminating against?

Follow up: do you allege those classes of customers are currently being meaningfully served by traditional taxicabs?

Actually it is cabs that discriminate against certain classes of customers, mostly poorer and blacker customers, by refusing to take fares. Uber drivers are not allowed to do that and are fired for refusing fares.

Yes, taxis sure must *hate* how they are vigorously forced to pick up black passengers, which results in them diligently adhering to that law! /s

In reality, it's the other way around; minorities (and people with bad destinations) have much better success in getting a ride via Uber/Lyft than with taxis.

Plus, taxis flout the taxman with impunity, while it's impossible with Uber. Why aren't we sympathizing with the unfair competition that Uber drivers face from taxis? Think about it.

While I am a big supporter of Uber, I must note that this paper doesn't do the calculation correctly. It doesn't take into account any differences in terms of service quality and externalities (positive or negative) of Uber vs. Taxis. Wasn't there some kind of criminal molester who used Uber? And I understand there are people in many countries who openly use Uber as a means of hitting on dates. Wouldn't this be a negative compared to a successfully regulated taxi system?

Of course in the end Uber may prove to be vastly superior while still taking this into account. But I do consider this calculation to be virtually special pleading.

It's like -- dare I say it -- measuring the impact of immigration while ignoring its effects on crime and the costly integration of the descendants of immigrants into the citizen community.

Taxi drivers are way worse than uber. With Uber you know exactly who the driver was, where they went, etc. With a taxi driver, unless you get the cab number you have no idea and the cab company will just circle the wagons. Lots of criminal taxi drivers out there.

Crimes committed by taxi drivers typically don't make the papers. Crimes committed by Uber drivers are front page news.

There's a selection bias effect here. Uber is the new shiny object, journalists are looking for ways to attack them because it sells papers.

A lot of violent acts occurring in taxis are against the driver. In the US, it's more dangerous to be a taxi driver than a cop. Perhaps it's a similar situation around the world. https://www.osha.gov/Publications/taxi-driver-violence-factsheet.pdf

The successfully regulated taxi system is asymmetric in data handling: it makes passengers feel safer while the driver never knows if the next passenger is the last one. It would be interesting to know if the passenger "feeling observed" (data stored in Uber servers) yields a lower driver homicide rate compared to conventional taxis.

Dalton told officers that when you "plug into" the Uber app, "you can actually feel the presence on you."


Uber drivers and taxi drivers have different supply curves. Just measuring miles driven doesn't fully capture differences in productivity.

And different demand curves.

How many Uber customers are there who are too poor to have a smart enough phone and an electronic payment system accepted by Uber are demanding services from Uber?

Regulated taxi services are required to accept cash, vouchers from government sanctioned aid agencies, etc, ie, taxis are required to meet the demand from customers Uber intentionally excludes from its demand.

Further, Uber refuses to allow demand from most of the land area of the US, and while taxis do not cover the entire US land area, they service demand from many times the land area of the US.

I live in a fairly dense suburban area that is by public policy devoid of public transit for all intents, leading to all residents needing private personal transportation when moving into the area. Once invested in the area, many residents become unable to provide their own transportation, and must rely on ad hoc arrangements with limited taxis service, most taxi service provided by airport limo/shuttle services being the best bad option if family and friends and neighbors can't provide transit.

Uber seems to be able to attract about as many transportation small business contractors as there are taxi and limo small businesses listed in the yellow pages and taxi exchanges.

Unknown are the taxi small businesses that have regular customers with regular schedules. Some Uber small business contractors make their money by finding regular customers they handle outside Uber's rent seeking system. But cards on laundromat bulletin boards and occasional ads in free newspapers and word of mouth are probably better at finding customers for such small businesses than Uber.

Bottom line, the Uber demand curve is far lower than the demand curve for taxi service by intention on the part of the Uber founders who want the highest prices to maximize the Uber rent extraction.

The reason Uber was able to get its start was because regulated taxi services weren't meeting demand. My parents lived 6 miles from the downtown area and taxis would take an hour to respond to a call, apparently because they would rather work the downtown area hoping for a lucrative airport run.

"Uber drivers...."

According to Uber, there are no Uber drivers. To get a contract with Uber, you must be a capitalist. You must own the minimum capital required by Uber to get contracts from Uber. You must pay all the costs on the capital required by Uber, as well as returns to capital and operating costs of capital, from the payment Uber gives you which is the price paid by customers to Uber minus the Uber rent.

Until the capital asset can drive itself, the capitalist small business contractors with Uber provides an operator, which is the capitalist working without any wage, or benefit. I bet most Uber drivers have no taxable income because the allowed capital related expenses are equal to the revenue they get from Uber.

The big test will be a pending class-action suit.


Which side is Uber taking?

Which side is the anti-regulation, anti-living wage, economic conservatives taking?

Aren't they all taking the position that the individuals are capitalist small businesses, not workers, and by not being workers, individuals, if any, involved in providing the services provided under contract through Uber are not due wages, unemployment benefits, union rights, workers comp, big government welfare like social security and Medicare benefits, just like IBM, Google, Microsoft do not qualify for welfare benefits?

What I find so interesting is the idea that the path to growth is with small business startups, but when Uber makes business startup trivial for millions of people, the advocates of small business startups suddenly declare small businesses to be workers.

In reality, what the advocates of small businesses want are millions of startups like Bill Gates and the Woz which then get IPO'd to enable massive profits by the Wall Street rent seekers who suck off the money that should be paid to workers building tangible capital assets. For Wall Street, they love the idea of a million losers who start businesses and go bust after laboring at extremely low wages because one in a million times these guys get lucky and a rent seeker buys in and reaps all the profits.

The small business startups that Uber, Lyft, Airbnb, et al are creating will never provide the rent seeking opportunity for Wall Street because they are captives from the start to rent seeker Uber.

Uber's business model is based on converting the equity in the Uber partner's (drivers) cars into Uber's bottom line.

With their latest round of price cuts Detroit the UberX rate is down to 30 cents a mile. (That's 22.5 cents) after Uber takes its 25% cut. It is literally impossible to make money at that rate. Uber is relying on it's "partners" to not realize that the wear and tear on the cars is an actual expense.

If you hang out on any of the Uber forums you'll see that the veteran drivers who have stuck with Uber have learned this lesson. They'll only drive an already depreciated vehicles during surges. But to fill in the gaps Uber needs a continual supply of unaware partners (i.e. low IQ immigrants) to keep the system going.

I agree that a lot of people out there act as though gas were the only expense of driving a car. However, I can also see how people would be happy to smooth their finances by "borrowing" from the equity built up in their car. Or by converting their car loan into ready cash.

Additionally, while the depreciation rate may be more or less constant across cars from an accounting perspective, it most certainly is not from a maintenance costs perspective. When considering the durability of certain models with the frequency of surge pricing, it's hardly implausible for drivers to be able to turn a small profit on their time even at such low base rates.

Do they lower their rates when supply of drivers is high and would they raise them if supply were lower?

Finally, someone nails it.

Uber is a rip off for most drivers. Not only does Uber take the "safe ride fee" off the top, then they take their percentage. (So Uber makes more than a dollar per ride off the top.)

And price reductions are rarely announced more than 24 hours in advance, so drivers have little time to plan for that. Oh, and the minimum fee can be lowered with no notice, too.

The amount of hostility here toward cab drivers (and increasingly toward Uber drivers), also is depressing.

I guess I don't understand how Uber works. Apparently, Uber sets the price which makes no sense. Drivers and passengers should agree on a price, maybe through an auction like on eBay. It might be worth more to me to get a ride from someone with 100% feedback and a nice Mercedes. Or I might prefer to accept the lowest bidder and his 20-year-old Civic.

If nobody offers a "ride sharing" service like that, someone should. If nobody wants to drive me for $20, how about $25? Okay, $30. Ah, two drivers -- hmmm, I'll take the one with the 2005 Scion who can be here in 15 minutes. Better be on time if you want good feedback!

Lol Detroit. Your example was Detroit.

A post apocalyptic city is really, really, relevant to how uber operates in functioning areas of the world.

Maybe, just maybe, Uber set that price since that is what customers in the Detroit area demanded given their limited means.m

Looking at the data that is exactly the suitation. But don't let the facts get in the way of you diatribe! Uber is capitalism, capitalism is oppression. Workers of ten world unite! Smash the looms!

I am not some OWS socialist. I've been a subscriber to Reason Magazine for 27 years and donated money to Ron Paul's Presidential campaign. And I am certainly not advocating Uber to be outlawed or banned. I believe people should be free to make their own decisions in life and if they decide they'd like to convert the equity in their car to give to Travis while earning a sub-minimum wage, that's fine by me.

That doesn't mean that Uber isn't a sleazy organization whose continued existence depends suckering in new drivers as their existing partners wise-up to the scam.

Yes, it is possible to earn a decent hourly rate, if you drive a depreciated car, work only during surges, learn the ins and outs of your particular area so as to minimize the dead miles. But Uber can't survive on those drivers alone.

Yes Detroit is an extreme example, but the price cuts in Tampa, Cincy, San Antonio, etc. are having the same effect. When taxes and depreciation are taken into account, it's not possible to make a minimum wage working 40 hours a week.

Uber's one hope is a never-ending stream of low-skilled, low IQ immigrants who will live 6 to a 2 bedroom apartment and are so glad to be out of whatever hell-hole they were living in, that they'll glady take Uber wages. (And of course the externalities will fall on the rest of us.)

I won't even get into the lease a car and drive for Uber flim-flams which are just a step above pay-day loans.

The set of consumers is different between Uber and traditional cabs. How does the author manage to factor that in his productivity computation?

Productivity per hour is also helped by the rating system that uber uses (at least here in Australia). Passengers are usually ready when the car arrives lest their tardiness leads to the driver marking them down.

"Krueger co-authored an earlier paper on Uber drivers commissioned by Uber but this paper was not commissioned."

The naivety that underlies this sentence made me smile.

Comments for this post are closed