The Uber Tipping Equilibrium

What is the effect of tipping on the take-home pay of Uber drivers? Economic theory offers a clear answer. Tipping has no effect on take home pay. The supply of Uber driver-hours is very elastic. Drivers can easily work more hours when the payment per ride increases and since every person with a decent car is a potential Uber driver it’s also easy for the number of drivers to expand when payments increase. As a good approximation, we can think of the supply of driver-hours as being perfectly elastic at a fixed market wage. What this means is that take home pay must stay constant even when tipping increases.

But how is the equilibrium maintained? One possibility is that as riders tip more, Uber can reduce fares so that the net hourly wage remains constant. Since take home pay doesn’t change we will have just as many drivers as before tipping. Under the tipping equilibrium the only change will be that instead of the riders paying Uber and then Uber paying the drivers, the riders will also pay something to the drivers directly and Uber will pay the drivers a little bit less. The drivers end up with the same take home pay.

But suppose that Uber doesn’t want to reduce fares or is somehow constrained from doing so. Does the model break down? Sorry, but the laws and supply and demand cannot be so easily ignored. If Uber holds fares constant, the higher net wage (tips plus fares) will attract more drivers but as the number of drivers increases their probability of finding a rider will fall. The drivers will earn more when driving but spend less time driving and more time idling. In other words, tipping will increase the “driving wage,” but reduce paid driving-time until the net hourly wage is pushed back down to the market wage.

At this point many readers will object that I am a horrible person and this is all theory using unrealistic “Econ 101” assumptions of perfectly competitive markets, rationality, full information etc etc. To which my response is that the first claim is plausible but irrelevant while the second is false. A new paper, Labor Market Equilibration: Evidence from Uber, from John Horton at NYU-Stern and Jonathan Hall and Daniel Knoepfle, two economists at Uber, looks at what happens when Uber increases base fares:

We find that when Uber raises the base fare
in a city, the driver hourly earnings rate rises immediately, but then begins
to decline shortly thereafter. After about 8 weeks, there is no detectable
difference in the average hourly earnings rate compared to before the fare
increase. With a higher fare, drivers earn more when driving passengers, and
so how do drivers make the same amount per hour? The main reason is that
driver utilization falls; drivers spend a smaller fraction of their working hours
on trips with paying passengers when fares are higher.

My conclusion is that increases in Uber fares are a very bad idea. Why? Increases in Uber fares–i.e. increases beyond those required to have enough drivers so that pick-up times are reasonably short–have two negative effects. First, and most obviously, higher fares increase the price to riders. Second, higher fares don’t result in higher driver earnings but do result in drivers wasting time.

The situation is very similar to the inefficient market for realtors. When realtors earn a fixed percentage of a home’s sales price, higher home prices encourage more entry into the realtor market. But we don’t need more realtors just because home prices have increased! When home prices are high, a realtor can earn enough selling a handful of homes a year to make it worthwhile to stay in the industry even though most of the realtor’s time is spent unproductively finding customers rather than actually helping customers to buy and sell homes. It would be better if commission rates fell when home prices rose but even after many years of online entry that typically doesn’t happen which is the mystery of realtor rent-seeking.

Uber is a great service for riders and it’s also great for people who need a source of flexible earnings. The fact that Uber drivers earn less than some people think is appropriate is a function of the wider job market and not of Uber policy. Indeed, Uber can’t increase take-home pay by raising fares and if we require them to do so we will simply hurt consumers and waste resources without improving the welfare of drivers.


"two economists at Uber, looks at what happens when Uber increases base fares"

I am not that gullible.

In his zeal to support theories or Econ 101, and to discount behavioral explanations, Alex chooses to ignore the support for behavioral economics research by Camerer that was confirmed in this study that workers, when they got a higher income, cut back on hours...which is exactly what behavioral economics provided.

Read the paper, and read Camerer's paper as well.

"Camerer that was confirmed in this study that workers, when they got a higher income, cut back on hours…which is exactly what behavioral economics provided."

And also what Econ 101 expects, if the income-effect is higher than the substitution effect (to come ate that conclusion, you don't need behavioral economics for nothing).

No, Miguel. Go read the article and see how it departs from Econ 101. If you read the article Alex is citing, they try to discount Camerer's findings.

Here is the conclusion of Camerer's paper.

"Our interpretation of these ndings is that cabdrivers (at least inexperi- enced ones): (i) make labor supply decisions “one day at a time” instead of inter- temporally substituting labor and leisure across multiple days, and (ii) set a loose daily income target and quit working once they reach that target."

Here is the link:

Go read it. It is not classical Econ 101.

It is also largely discredited.

Yeah, the co-author Richard Thaler just got a Nobel prize in economics.

Offer your evidence. Your word means nothing. You could be a Russian troll. Cite some evidence. Dare you.

Here's one paper that attempts to discredit the One Day at a Time paper, probably what Jason was talking about:

There's been a lot of papers written on this topic, not just the original Camerer et al.

Miguel, By the way, here is the classical model against which you should look at Camerer's finding:
"Dynamic models of labor supply predict that work hours should respond positively to transitory positive wage changes, as workers intertemporally substitute labor and leisure, working more when wages are high and consuming more leisure when its price—the forgone wage—is low (e.g., Lucas and Rapping [1969])."

Individual workers may cut back on hours, but the higher fares will attract more drivers into the pool. So overall driver hours can increase even if individual drivers reduce their hours.

Yes, that's true.

This seems wrong even on simple economic terms.

Tipping makes pay more performance-based, resulting in a better quality service and therefore more value delivered to customers and more pay.

I think you could probably demonstrate this by comparing pay in fields with tipping to similar entry-level service jobs where there is no tipping. You could also add a cross-regional comparison and look at countries that don't have the same tipping tradition and see if the same wage premium occurs there.

For example, in the U.S. wait staff make significantly more than cashiers. Is that also true in countries where they don't tip, or don't tip as much?

There's also a bit of an "ability to pay" redistribution going on I think. I can't think of a more effective way of "spreading it around" than to directly give an extra $10 to someone hustling to provide a pleasant experience.

Using Germany as an example, and comparing full time cashiers and waiters, the answer, very roughly, is that each have health insurance and a legally mandated 25 days vacation. The cashier is a bit more likely to have paid legal holidays, however. Wages are considerably more variable for wait staff, but the sort of high paying cash earnings of highly tipped waiters might, at best, exist in a few German localities with a high number of American tourists (think Heidelberg, for a concrete example which is not that far from where I live, and until the facilities were shut down a couple of years, had a fairly notable American military population too).

In other words, not much real difference.

Uber already has a rating system that deals with performance. So that is covered.

And I'm not sure tipping really increases performance that much. I usually just give a 20% restaurant tip every time unless the service was awful (1% of the time). The rating system essentially removes the 1%.

"I usually just give a 20% restaurant tip every time unless the service was awful."

Not everyone tips that way, but even if they did it would still improve quality by weeding out terrible servers.

Having dined in a number of countries, I definitely think tipping improves service in the U.S.

The effect on Uber specifically I don't know, but I took Alex's point to be a general one on tipping.

The point of tipping is to make the customer feel superior to the servant. Of course, one has to tip because the employees are forced to rely on them, but we should collectively find alternatives to tipping.

The point of tipping is to make the customer feel superior to the servant.

No, the point of tipping is a token of appreciation. This isn't that difficult if you're not attempting to manufacture an excuse for being a jerk.

A token of appreciation. Yeah, keep telling yourself that. I said that it was an obligation to tip so long as we have this abusive system, but I think it would better to substitute wages for tips.

Furthermore, why should we have to subsidize "jerks" who don't tip? That also doesn't seem fair or efficient.

The point of tipping is to ruin my meal because all of the arguments made above swirl in my head when the bill arrives, making me ill.

You haven't taken cabs in NYC, I'm guessing. The cab is filthy, the driver talks on his phone all the time and you have to wrangle your bags out of the trunk. And a 20 percent tip is de rigeur. Some service.

What increases quality service with Uber is the mutual driver-passenger rating system. If you get too many fours (on the 1-5 scale), you have a harder time getting a ride. My personal view is that the tipping protocols in this country are ridiculous -- here's 20 percent for not being mean to me -- and I would be happy to pay more for rides instead. (In NYC, Uber has made things worse by dropping ride rates to increase market share, drivers be damned.)

The Realtor percent-of-sale-price fee has flooded our town with housewives who don't know much about real estate, add no value and are satisfied to take a fee for one home sale a year. The Realtors' vaunted MLS is now irrelevant for years, and those "professionals" do little more than schedule showings for people who've already done their initial research online. My husband and I never would let a Realtor negotiate a contract, but we might pay a flat fee for the job if we had no relevant experience. Hard to see why the whole industry wasn't disintermediated 10 or 20 years ago.

Josie, nobody makes you tip 20% for bad cab service in New York. Why do you do it?

Uber seems more like Tupperware than anything else. Some people make more money, some people make less.

Does Lyft allow tipping through their app? That would be an interesting comparison.

They do. And, what is interesting is that Alex ignores two party market analysis> That Uber is competing not only for customers, but also for drivers. Uber reacted to Lyft's model, which permitted tipping.

Not only do they, but they've always had it.

The moderator squanders time dreaming up excuses for being a graceless cheapskate.

Art Deco is right on time. "At this point many people will object that I am a horrible person..."

No, by objection is that he's a twit. I'll keep 'horrible person' in reserve for the point in time he's discovered to be a participant in corruption or abuse of power at GMU.

LOL that's what you get for trying to agree with Art. He's an ornery cuss.

One other crazy statement in Alex's post: That the wage increase results in drivers "wasting time"

Again, what Alex totally misses is that what Uber was trying to address was CUSTOMERS having longer wait times. if the driver is "wasting time" waiting for a customer, this DECREASES wait time for the customer. If Uber has higher wait times it loses customers to Lyft or to on call or hail taxi services.

What is with the quality of this blog this week?

What, you mean it has recently improved?

I recall a study earlier this year or last that concluded Uber drivers set a goal for earnings for the day and, when the goal is reached, quit. The study was in the context of peak pricing (it doesn't necessarily add to supply of Uber drivers during peak demand because the drivers reach their goal quicker and quit). Tabarrok is not a "horrible person", but I think he may not understand Uber drivers.

Ie, earn enough for two dime bags?

The bigger picture would be the multiple frauds going on:

1. Investors are told it will all be cool when fares cover actual cost, plus profit.

2. Drivers are told they'll make a high return net of expenses.

Plus tax filing confusion:

"This means that the income reported on your 1099-K includes commissions and fees that never hit your bank account. These fees are deductible as a business expense, but about half of rideshare drivers aren’t deducting them."

I think they have ot do that because the credit card companies make a report to the IRS so they want them to match.

Amazon does the same for sellers. they pay you net, but the tax information is gross.

The flawed assumption of this analysis is that there is a linear correlation between Uber wages and number of Uber drivers, which is a perfectly reasonable classical assumption but completely ignores insights from behavioral economics. Not everyone would want to be an Uber dirver even if it paid better than their current job. There is stickiness in the uptake of "better" jobs, meaning that at least for a while an increase in tipping would deliver real gains to Uber workers. Considering that for many drivers Uber is a temporary thing, the period of time for which they receive increased earnings could be a large portion of their total time working for Uber.

"which is a perfectly reasonable classical assumption but completely ignores insights from behavioral economics."

Other with the obsession with "behavioral economics"; again, you don't need "behavioral economics" for nothing - the idea that "not everyone would want to be an Uber dirver even if it paid better than their current job" is perfectly mainstream with neoclassical economics: it is simply a question of assuming that jobs are different and that the disutility of some job is differente from the disutility of other job. An analogy - nobody says that the neoclassical predicts that, when the price of apples goes down, everybody will buy apples instead of oranges.

Millions of people would drive for Uber, except Uber never employs drivers, only vehicles.

If you are not a capitalists, you can not contract with Uber.

Uber is merely a simple way of starting a small business if you have the required capital. You are a pass thru business for tax purposes, with gross business income reported on a 1099 with lots of business expenses to deduct, requiring added bookkeeping and paperwork.

Ie, you need to record all business miles, which are not the fare miles, in order to deduct 57 cents per mile if that option can and is used. That 57 cents covers normal cleaning and detailing, but the cost of cleaning up barf is exceptional, but needs to be documented to be deductible.

I'm not sure how the cost of a phone is handled. Does its cost need to be prorated between business vs personal use?

But if you have a more expensive vehicle, you now need to record all costs, all uses, calculate depreciation of capital assets, allocate costs between business and personal in a way that is acceptable to the IRS. Buying a new $50k vehicle to replace your used $10k beater you use to drive to work at half the miles driven increases your personal commuting costs drastically. But your beater would get you many if any Uber contracts.

The problem with Alex's analysis is he assumes that magic or the government gives every person a vehicle suitable for getting Uber contracts. Instead he should be focusing on the barriers to entry to contracting with Uber as a business providing transportation services.

Personally, I have a barrier of entry to merely getting a ride by using Uber. I don't have the required capital: a smart cell phone. And even if I had a phone, my vehicle, a trike, does not meet Uber's business capital minimums to providing rides.

Alex needs to sign up and provide ride services for Uber before writing about the economics of Uber, a minimum of say 500 rides, just 10 per week for a year. 30 minutes per for 5 hours a week max. And then do all the tax filings to report all the after tax net earnings of ride service contractors for Uber.

Economists seem to have no clue about business economics. Listening to economists, and politicians, talk about business taxes, you might believe profit is the amount paid to workers, and the money owners and shareholders get is business costs.

What is interesting is why there is all this media focus on the supposed plight of the Uber driver and what forces are behind the efforts to upset the economics of a model that seems to work for both rider and driver.

I suspect that most of the people complaining are genuinely unaware of how the typical taxi driver in most cities is a non-owner, generally "independent contractor" who rents a medallion or car for a weekly rate and makes a very low net salary with no benefits, similar to that of Uber drivers. (There's a difference in that an "all you can drive" weekly rate incentivizes working extremely long hours and weeks, whereas Uber's surge pricing encourages irregular hours. Some people like a regular work week, even with overtime; some people like irregular hours.) Sure, there may be medallion owners and owner-operators pushing the stories, but most of the focus is out of misconception.

An independent contractor may or may not have a good deal:

'full information'

Being unfamiliar with Uber in America, don't people still tip with cash? And if so, how the hell would anyone know that, apart from the two people involved in the cash transaction?

Both Uber and Lyft offer the option to tip with your credit card through the app once your ride is complete (at least here in the U.S.).

No, no cash tips, those are forbidden by the terms of service.

BUT, the price of an Uber ride is fixed by Uber but the tip is optional and variable. Therefore even if it all averages out (some tip more, some less, some not at all) wouldn't drivers prefer a certain payment to an uncertain one?

And what about Uber itself? Uber takes a cut of the ride price, but the driver gets 100% of the tip: driver revenue = 0.8*fixed_price + tip. Therefore the equilibrium price of a ride should be higher if driver income is mostly from the fixed price (because the driver gets less of the customer's total spend). Or it will be if uncertainty doesn't reduce the perceived value of tip income.

And then there's Uber. If the total ride price (fixed price + tip) is constant, Uber's revenue will go down if it reduces the fixed price unless doing so somehow encourages more people to use the service. But if the total price of a ride (fixed price + tip) remains the same, why would reducing the fixed price increase demand?

In contrast to Uber, grad student admissions would seem to be tightly controlled, so there's a better argument there that taxing tuition waivers may lead to better outcomes for grad students (such as being classified as employees and getting benefits) or at least having the university bear most of the costs, or have the universities reduce enrollment of grad students and hire other employees. There isn't the free entry, and the universities do seem to pay at close to immiserating rates.

Having more drivers on the road means less waiting time for riders, so while the prices are higher, the service is a bit better. This is especially true for riders in less well served areas, as the large glut of drivers will cause some to avoid the main areas.

It's not clear whether the price increase is compensated by the service increase, but it probably is on the edges of the city, and folks in the center of the city are usually more price tolerant.

Yep. That's what the paper shows, that there was less wait time, which is why Uber "generously" raised rates. Also, the pricing increase changes resulted in fewer surge pricing episodes. The paper doesn't cover whether the diameter of the market increased, which is notable, because Uber, in some cities like Chicago, set an initial service area, and later expanded it while at the same time raising prices.

I think the failure of the paper is to ignore service area changes and also, while acknowledging wait time changes, failing to explain that Uber wasn't changing rates because it was "nice" but because it had to, in the face of competition for both passengers (less surge) and driver (who were defecting to Lyft).

Fewer trips means lower costs for the car (gas and wear and tear) for the same amount of money. Waiting for a fare may just be that, waiting and not driving around. Most people don't take factor in depreciation of a capital asset or future expenditures for maintenance when trying to buy that week's food or make the rent, if it does result in less trips but more money for each trip they will still net out positive in the long term.

The real wage of uber drivers is close to zero. It's primarily a way to convert auto equity into cash.

Really? You could make $38,434.42 a year (the median Uber wage) by selling your car and investing it in some sort of investment portfolio?

Tipping is annoying now because every payment system now tries to nudge you into large tips for absolutely everything.

Relevant, but NSFW language:

Two weeks ago at a Celtics game I tipped the food and beverage guy 11 bucks (15%) because it was too hard to figure out how to give him $1 on the machine.

Tipping is a way to price discriminate against customers, thus increasing net revenue for the firm/driver to split up.

The ax getting gored is the consumer. They stole some surplus from him.

There is no such thing as 'price discrimination' when the discretion is at the hands of the customer.

The discretion to buy, our not, is always the customer's, so that makes no sense.

Tipping in this case seems quite different than restaurant tipping. The driver is already getting a base fare for the ride and the customer has less pressure to tip. The customer can easily proceed with $0 tip, and often does. In most cases when tips (especially large ones) are left it's a direct result of the customer perceiving the driver going the figurative extra mile, delivering good customer service, or just forming a rapport. If customers have this sort of voluntary reaction to a good experience, why shouldn't the driver benefit from their voluntary generosity?

The main reason is that driver utilization falls; drivers spend a smaller fraction of their working hours on trips with paying passengers when fares are higher.

This could use some clarification. What are a driver's "working hours," to start with, and what do they do during these hours when they have no passenger? Go to a coffee shop? Go home? Run personal errands?

The concept seems amorphous to me, and I would like a better explanation of how it all works. As I understand it, possibly wrongly, drivers do not have to accept all passengers. So the idea that they are working during some period when they don't have a passenger needs some definition.

Your understanding is wrong, they do have to accept nearly-all passengers. You clock in and clock out in the driver app, and must respond to all ride requests quickly (seconds) while you are clocked in.

The equilibrium price could be called the "tipping point", but that would just be confusing.

> My conclusion is that increases in Uber fares are a very bad idea. Why? Increases in Uber fares–i.e. increases beyond those required to have enough drivers so that pick-up times are reasonably short–have two negative effects. First, and most obviously, higher fares increase the price to riders. Second, higher fares don’t result in higher driver earnings but do result in drivers wasting time.

1000% correct. Not only is increasing fares wrong, fares and drivers wages should be decreased until no more riders ever have to wait. I can imagine a fare so low that riders won't even have to use the app to request, the free market will place the driver on the riders door step milliseconds prior to the rider realizing he wanted that uber in the first place.

This will result in no drivers ever wasting time, riders always getting the ride they want, and earnings staying the same.

If you're not a selfish little yuppie, take a god damned taxi and tip well.

It seems you are assuming quantity of Uber rides demanded remains fixed despite prices increasing.

This does seem like a classic "Econ 101" argument, not in how it assumes perfect market, but in how it is utterly devoid of any empirical data or observations. The closest is comes is presenting an analysis from Uber itself about what happens when base fares are raised, nothing about tips.

As a side note, that conclusion drawn by Alex- that it will never make sense for Uber to raise fares- seems weird.
Because right now Uber is unprofitable- its base fares can't cover expenses.
So does this mean that Uber can never become profitable? Or is there some scheme to become profitable by reducing expenses?

Yes, reduce expenses by eliminating the driver and just using self-driving cars.

What's left out is that consumers may be (very probably I am sure) pay a lot more for Uber's services than than Uber currently charges. So allowing tipping might not cost Uber a single fare (not many). Uber is involved in a classic Rockefeller/monopoly move: under pricing taxi fleets to try to put them out of business -- hoping to out wait taxis while Uber burns through billions of new investor cash. Not sure how all that fits into econ 101. ???

Of course if Uber had to contract with a labor union for drivers the price of a ride would be set by all that drivers could squeeze from consumers -- instead of the other way around, the American way.

"My conclusion is that increases in Uber fares are a very bad idea. Why? Increases in Uber fares–i.e. increases beyond those required to have enough drivers so that pick-up times are reasonably short–have two negative effects. First, and most obviously, higher fares increase the price to riders. Second, higher fares don’t result in higher driver earnings but do result in drivers wasting time."

On the surface, Uber's and the driver's interests are aligned here as Uber takes a flat proportion of base fares. This alignment is somewhat mitigated with tips as drivers keep 100% of tips, but as long as tips represent a small fraction of total revenue, its impact is irrelevant. The reality, however, is much more complicated.

In the debate for the most optimum ride-sharing market with respect to prices, it'd be irresponsible to not consider a myriad of other independently simple factors. Let's consider the impact of low and high fares to Uber, the drivers, the consumer, and the public.

While Uber's economists may conclude that Uber's profits are not influenced by high or low fares, the reality is more nuanced. A simple Econ 101 analysis fails to take into account network effects by having a larger number of cheaper rides. An increasingly larger proportion of rides are given through the UberPool service, which requires a certain critical mass in order to aggregate multiple rider parties into one car. Additionally, more rides above that certain critical mass makes UberPool more attractive as Uber is able to offer shared rides at a lower cost to consumers.

To drivers, even if higher fares are overall revenue neutral to them, fewer rides at a higher price are preferable to the contrary. Ignoring car wear and tear and assuming the new prices take these into account, drivers work highly flexible schedules dependent on the rides market at that particular moment. Because of the highly cyclical nature of ride sharing demand (peaks during commute hours, weekend evenings, etc.), the consolidation of rides means many drivers can work fewer hours at a higher utilization ratio, which frees drivers to pursue other pursuits. Even if drivers are forced to idle and wait for their next ride, they can do so in parking lots and can get other work done in that time period. Having more free time is obviously preferable to drivers.

To riders, it's obvious that a lower fare system is preferable. Having their interests perfectly opposite to those of the drivers creates a classic buyer/supplier market. While Uber may be bias towards the consumer and sets the prices, it's still beholden to the supply/demand of rides, and can't dictate pricing without taking market forces into effect.

To the public, higher fares are preferred as it moves marginal riders to seek other forms of transport (walking, train, carpooling, etc.) and takes cars off the road, especially during commute hours. Moreover, drivers that can work less and use that time towards other pursuits may directly and indirectly benefit the community through the availability of additional services (drivers may be tutors, volunteers, etc.).

Thus, while Uber may slightly favor lower fares, the impact of higher or lower fares is mixed to different groups. You have to consider all of these impacts to determine which system is preferable.

Comments for this post are closed