A widely reported figure on ride-hail pay was corrected by its own lead author soon after it made headlines. A 2018 working paper from MIT’s Center for Energy and Environmental Policy Research reported that the median Uber or Lyft driver in its survey cleared $3.37 an hour after vehicle costs.¹ After it drew press attention, Uber’s chief economist argued that the authors had mishandled drivers’ answers to two survey questions about income, and the lead author, Stephen Zoepf, said the criticism was valid. His initial recalculation put the median at $8.55 to $10 an hour, depending on method.²

So how much do gig workers make after costs? The correction shows why the answer is slippery: it depends on which hours you count and which costs you subtract, and small choices move the result a long way. New York City’s 2022 study of app-based restaurant delivery workers, built on the apps’ own records, found they earned about $14 per hour worked in late 2021, tips included and waiting time counted, and kept about $11 after expenses.³ A 2024 UC Berkeley Labor Center study of drivers in five other US metro areas found typical pay after costs, tips included, in a similar range, with wide spread around it.⁴

For the wider picture of extra work, see what a side hustle involves, from typical hours to tax. This page covers pay per hour after costs for delivery and rideshare work, mostly in the US.

How much do gig workers make? First ask which of three numbers you see

A gig pay figure can be the fare the customer paid, the payout after the platform’s share, or what is left after the worker’s own costs, and each can be divided by paid trip time or by every hour logged in. A 2018 study of Uber’s records by Cody Cook and colleagues shows how far apart those layers sit.

Using Uber’s records for more than a million US drivers from 2015 to early 2017, the study found average earnings of about $21 per logged-in hour, which the authors state is before Uber’s service fee, fuel and depreciation. In an appendix they sketch the next layers: Uber’s average share took about a quarter, and at the running cost per mile they assumed, roughly a third of what remained would go on the car. That sketch rests on assumptions, not measured costs, and two of the five authors were Uber employees with equity in the company, as the paper discloses.⁵

Picture two neighbors comparing notes after the same Friday evening. One quotes the hourly rate the app showed for her trips; the other has taken off fuel and counted the whole evening from logging in to logging out. His figure is lower, and it is the fairer match for a regular job’s wage.

1. What the customer pays: the fare or delivery fee, plus any tip
2. Minus the platform’s share: what the app deposits is already smaller than the fare
3. Minus running costs: fuel or charging, insurance, repairs, depreciation, phone and data, tickets
4. Divided by every working hour: waiting for offers and driving to pickups count, not only time with a passenger or order

Bar heights are illustrative, not to scale; the gap between them differs by city, vehicle and year.

The lesson: before comparing any gig pay figure with an hourly job, ask what it has already subtracted and per hour of what.

Waiting and driving to pickups can move the hourly figure a lot

Logged-in hours include waiting for an offer and driving to a pickup, and US platforms pay mainly for time spent on trips. A May 2024 report from the UC Berkeley Labor Center found that published US estimates of driver pay differ in how they treat that waiting time and how they price vehicle costs.⁴

The Berkeley authors split each shift into three parts: waiting for a request, driving to the pickup, and carrying the passenger or order. In the metros studied, pay in California was based on the second and third parts; elsewhere, only on the third. Their example of how much method matters comes from Seattle, where two studies of the same market reached very different net figures, largely because one assumed 19 cents a mile in vehicle costs and the other used an earlier IRS rate of 55 cents.⁴

Their own data came from drivers who uploaded trips to a tracking app in January 2022, in five metro areas. Counting whole shifts and costs, a majority of drivers in each area earned less than the local minimum wage. The report was funded by the Ford Foundation and argues that drivers would be better paid as employees; the authors also note that their sample may earn more than the average driver.⁴

The argument has not settled. A November 2025 study that Uber commissioned from the consulting firm HR&A Advisors put its drivers’ average 2024 pay after expenses, tips included, at about $21 to $23 an hour in Chicago, Philadelphia and Portland, Oregon.⁶ A January 2026 brief by James Parrott and Michael Reich, one of the Berkeley authors, redid those sums with more waiting time counted, the IRS mileage rate and tips left out, and got about $13.⁷ Same data, same year: the gap comes from method.

In practice, a quiet weeknight can look fine trip by trip and poor across the shift, because long gaps between offers and unpaid drives to pickups stretch the hours the pay must cover. Noting log-in and log-out times, not only trips, gives a truer figure.

The IRS mileage rate is a benchmark for costs, not a bill

The US Internal Revenue Service’s business standard mileage rate is 76 cents a mile for July to December 2026, up from the rate set for the first half of the year.⁸ The IRS says the business rate rests on an annual study of the fixed and variable costs of operating a car, which makes it an average across vehicles and a tax option, not a measure of what one driver’s car costs.⁹

The MIT working paper shows how far real costs can sit from that benchmark. For its drivers, whose cars the authors describe as a used fleet, the estimated cost of fuel, insurance, maintenance, repairs and depreciation came to about 30 cents a mile for the median driver, well under the 2016 IRS rate of 54 cents, and roughly half of the median driver’s revenue per mile.¹ These figures come from the original version, still listed as under revision, so treat them as rough.

The same paper offers one possible reason some drivers underestimate costs. Repairs arrive as large, irregular bills, and depreciation on a paid-off car is not felt until it is sold or scrapped. The authors suggest that drivers who skip these costs are, in effect, borrowing against the value of their car.¹ The drivers’ incomes and hours in that paper were self-reportself-report: A measure in which people describe their own behavior, feelings or circumstances, usually by answering a questionnaire. When the same person supplies both of the things being compared, shared habits of answering can make the link between them look stronger than it is.Full entry in the glossary in an online survey with a low response rate, drawn mostly from subscribers to a driver website.¹

What to take from this: counting fuel alone makes any hourly figure too high. The IRS rate gives a full-cost reference for an average car; a particular car can cost more or less.

What New York City’s delivery data show after costs

New York City’s Department of Consumer and Worker Protection measured delivery pay from the apps’ own records and priced workers’ costs with a survey, and found that tips made up about half of earnings and that costs weighed more on car drivers than on e-bike riders. The study covered Uber Eats, DoorDash, Grubhub and Relay.³

The study

Moderate evidence

Every logged-in hour counted: New York City's delivery pay study (2022)

Counting every hour logged in, including time waiting for offers, app pay without tips averaged $7.09 an hour and tips added about the same again. Hourly expenses averaged $2.70 for e-bike riders and $4.86 for car drivers. Batteries were the largest cost for e-bikes; for cars it was depreciation, followed by traffic and parking tickets.³

The strength of this study is its denominator: it counts on-call time as work, the way a worker experiences a shift. Its limits are scope and date. It describes one dense city where e-bikes carried most deliveries, in one quarter before the city’s minimum pay rule, and it estimated expenses largely from IRS deduction rules, surveys and price quotes.³

The vehicle sets the cost floor. In the study, e-bike riders had higher net earnings than car drivers, because they made more deliveries per hour and had lower expenses.³ For anyone weighing delivery work, the vehicle belongs in the calculation before the app does.

City pay floors change the math where they apply

A minimum pay rule sets a floor under hourly pay, and New York City’s is the clearest example for delivery work: as of September 2026, the city’s Department of Consumer and Worker Protection says apps doing restaurant or grocery delivery must pay at least $22.13 an hour, not including tips, for time spent preparing or making deliveries, and the minimum rises each April.¹⁰ Grocery apps were brought under the rule in January 2026, when city law also began requiring restaurant and grocery apps to offer a tipping option at checkout.¹¹ The 2024 Berkeley report listed New York State, Seattle, Washington State and California as also having pay floor policies for some gig drivers; such rules change, so check your own city or state.⁴

The city’s quarterly app data show how the floor reshaped the work. Between late 2023 and late 2024, app pay per hour rose from about $7 to about $20, tips per hour roughly halved, and on-call hours fell by more than four-fifths. The department says some apps changed their screens in ways that discouraged tipping, and that its worker counts do not remove people using several apps at once.¹²

The city’s reasoning was simple: once an app must pay for time, idle time costs the app rather than the worker, so the app has reason to cut it. The city’s 2022 study predicted as much, expecting apps to use workers’ time more efficiently and raise deliveries per hour.³ For workers, a higher rate per hour can come with fewer idle hours on offer; the published data do not show how easy it became to find work at the times people wanted.

Pay per driver has not held steady over time

Platform pay figures age quickly. The JPMorgan Chase Institute tracked platform payments into millions of anonymized Chase checking accounts from October 2012 to March 2018, and found average monthly earnings from transportation platforms fell by about half between 2013 and 2017 as the number of drivers grew.¹³ The data show earnings, not hours, so the report cannot say whether hourly pay fell or people drove less. The institute is part of the bank, its report was not peer-reviewed, and it sees only Chase customers.¹³

Outside the US, the picture changes again. The International Labour Organization’s 2021 report, based on surveys of about 12,000 workers worldwide, found that most app-based taxi and delivery workers it surveyed received bonuses and called them an important part of income.¹⁴ A bonus-heavy pay model means one month’s rate may not predict the next.

WiserHours

Questions to ask of any gig pay figure

• Is it before or after the platform takes its share?
• Does it subtract fuel or charging, insurance, repairs, depreciation, phone and tickets?
• Is it per hour on trips, or per hour logged in, waiting included?
• Are tips included, and are they counted toward any pay floor?
• Which city, which year and which vehicle does it describe?
• Who produced it: the platform, independent researchers, or a survey of volunteers?

Find this checklist online:
https://wiserhours.com/side-hustles/how-much-gig-workers-make/#checklist-how-much-gig-workers-make-questions

Where to get advice on pay, tax and the rules

Which adviser fits depends on urgency: a labor agency if an app may have underpaid you, a tax professional before you file, and a regulated financial adviser before a large spending decision.

If you deliver in New York City and an app has paid less than the city minimum, the Department of Consumer and Worker Protection takes complaints from delivery workers regardless of immigration status.¹⁰ Elsewhere in the US, ask your city or state labor department; in other countries, your national labor inspectorate.

In the US, the IRS says its mileage rate is optional and taxpayers may claim actual vehicle costs instead.⁹ A qualified tax professional, or your own country’s tax authority outside the US, can say which suits your return. For choices such as financing a car to drive for work, a regulated or fee-only financial adviser can weigh your whole situation.

The bottom line

What a gig driver earns after costs depends less on the headline rate than on two quiet choices: which hours count and which costs come off. Official New York City data, the best-measured case, show a large gap between app pay and take-home pay once waiting time and vehicle costs are counted, and a pay floor that raised hourly pay while cutting idle time. Treat any single figure, including the ones here, as a description of one place and year, and check it against the questions above before comparing it with a job.