The Effects of Proposition 22 on Driver Earnings: Response to a Lyft-Funded Report by Dr. Christopher Thornberg

Ken Jacobsand Michael Reich

Press Coverage

Jacobs and Reich (2019) estimated that besides exempting Uber and Lyft from AB5, the promises in what is now called Proposition 22 were worth much less than they appeared. [1] The proposition promises to pay drivers 120 percent of the minimum wage, reimburse them for some of their expenses, and provide some health benefits. Jacobs and Reich pointed out that Uber and Lyft would not compensate drivers for their waiting time and expenses and that their promised reimbursements of expenses are well below the costs faced by the drivers who provide most of the rides. As a result, Jacobs and Reich estimated that the promises in the proposition were worth the equivalent of only $5.64 per hour.

As we discuss below, independent studies consistently find that Uber and Lyft drivers are paid less than the minimum wage. This is not surprising when only two companies hire all the workers in an industry. Uber and Lyft possess considerable power to set pay at below-competitive levels. Meanwhile, the drivers who provide most of the rides are locked in to the industry through their long-term financial commitments in their vehicles. Strong evidence for these “monopsony” arguments comes from Uber’s own experiments with driver pay in Boston and Houston.[2]

A Lyft-funded recent report by Dr. Christopher Thornberg and his colleagues at the UC Riverside Center for Economic Forecasting and Development (CEFD) critiques our claims.[3] We argue here that Thornberg’s report rests on evidence that is not relevant for California and that he curiously ignores a study funded by Lyft that finds lower driver pay in California.

Thornberg overestimates driver gross earnings (before expenses) based on data that is not representative of drivers in California. He also underestimates driver costs. In doing so, he significantly overstates what drivers earn on net now, and would earn under Proposition 22.

Moreover, wage and hour laws are not concerned with the average worker earnings, but what each individual driver earns. Our earlier research found that under Proposition 22, drivers could still earn as little as $5.64 an hour. Thornberg repeats the same errors and misleading statements about driver earnings under the ballot initiative that we critiqued in our earlier report. We stand by our results.

1. Gross Driver Earnings

Thornberg takes as the evidential base of his analysis estimates from an Uber/Lyft-funded study of driver compensation in Seattle, conducted by researchers at Cornell.[4] That study claims that Seattle drivers’ gross earnings are in excess of $36 per hour. But the Cornell report is an overestimate of driver earnings in California and is highly misleading for the issue at hand here.

First, earnings of Seattle Transportation Network Company (TNC) drivers are higher on average than those of TNC drivers in California as a whole. Seattle drivers’ earnings appear to be comparable to those of drivers in San Francisco. However, a study by Uber economist Jonathan Hall and Alan Krueger found that Uber drivers in Los Angeles earned 22 percent less than drivers in San Francisco.[5] But pay levels and driving conditions in Los Angeles are more representative of the state as a whole than are pay and conditions in San Francisco. Average TNC driver pay in California is therefore substantially lower than in Seattle.

Second, Parrott and Reich’s independent study of driver earnings in Seattle found average net earnings (after expenses) of $9.73 an hour, well below the minimum wage.[6] Other independent analyses of driver earnings similarly find that a majority of drivers earn less than the California minimum wage.[7]

Third, Uber and Lyft’s own estimates suggest much lower average driver earnings in California than those claimed by Thornberg. The companies presented these lower estimates as evidence in response to the California Attorney General’s recent lawsuit calling for the companies to adhere to AB5. For example, cost increases projected by Uber as a result of converting to employee status suggest average gross driver earnings (before expenses) of close to $22 an hour, much less than Thornberg’s assumed $36 .[8]

A Lyft-sponsored study of Lyft data by Dr. Catherine Tucker of MIT similarly presents a lower estimate of driver earnings for California than the Cornell study found for Seattle.[9] Since Lyft already has funded Tucker’s study of California driver earnings, it is extremely odd that another Lyft-funded report relies on a separate study from an outside area that is not representative of the state of California.

Finally, Thornberg misrepresents the findings in a national study of Uber driver pay by two Uber economists and two Stanford economists.[10] This study does not find that the drivers are already paid well above the California minimum wage. Thornberg ignores a well-known re-analysis of this study’s data that finds the drivers are paid less than the wage equivalent of $10 per hour.[11] Thornberg also asserts, without offering any evidence, that driver pay in California is higher than in the U.S. as a whole and that it has been rising since 2015. Both of these assertions are contradicted by the evidence in Hall and Krueger (2018).

2. Hourly Earnings Assumptions

Thornberg critiques the Berkeley study for basing our analysis on a state minimum wage of $13, rather than $14. In 2021 the California minimum wage for small businesses will be $13 an hour. Proposition 22 classifies drivers as small businesses. The proposition does not contain any language stating that it would apply the 2021 $14 minimum wage that will apply only to firms with more than 25 employees.

3. Driver Costs

Waiting Time and Miles

Waiting time is a necessary part of drivers’ work. Under Proposition 22, the ride hailing companies have no incentive to minimize waiting time, placing all of the risk on drivers. That is time spent returning to an area with a greater chance of getting a pick-up, or, in urban areas, circling until another pick-up is available. Under California law that is time “engaged to wait,” which is compensable work time.

Thornberg argues that we overestimate the miles driven during waiting time by relying on estimates in a report from the consulting company Fehr and Peers. Fehr and Peers use data provided by Uber and Lyft for Los Angeles and San Francisco. The estimates from Fehr and Peers are similar to research from the California Air Resources Board, which comes from Uber and Lyft data for the state as a whole.[12] CARB finds the average TNC driver in California drives 22 miles each hour. Thornberg uses data from Cornell’s study of Seattle TNC drivers. However, average vehicle speed in a core metro area like Seattle will be much lower than for California as a whole. Again, Lyft could have supplied this data to Thornberg—rather than Thornberg relying on unrepresentative data from Seattle.

Compensation for Expenses

The CEFD report claims that our cost estimates per mile are too high, since we use the IRS tax deduction rate of 57.5 cents a mile (the same the authors receive as reimbursement when they drive for work purposes on their own vehicles). They state that the IRS figure “stands in stark contrast to estimates of driving costs by AAA” which they cite as 17-23 cents an hour. The AAA pamphlet they cite, however, has costs per mile ranging from 40-87 cents, depending on size of vehicle and miles driven. Thornberg’s lower estimate excludes most of the major costs associated with owning a vehicle (depreciation, financing, license, registration, taxes, and insurance). Moreover, the California Supreme Court has ruled that the IRS rate is presumptive for reimbursing employee vehicle costs.

Further, Thornberg excludes driving costs incurred while waiting between passengers from his expense calculations.

4. Benefits

Thornberg’s estimates for the value of benefits are highly misleading. They are based on the value per hour of engaged time only, excluding waiting time. Factoring in the time drivers wait between rides reduce his estimates by one-third. They also fail to note that the hours in Proposition 22 that count toward receiving the health benefit is only based on engaged time. We estimate the value of Proposition 22’s health benefit for a 30-hour per week worker (including waiting time) at $1.20 an hour. This is well below the value of benefits mandated for employees under state and federal law.

5. Average Wage versus Minimum Wage

California minimum wage law sets a standard for each individual driver during a pay period, not the average for all drivers. Parrott and Reich’s independent study of driver earnings in Seattle, which was based on data provided by Uber and on a driver survey, found that gross hourly earnings each month varied by 25 percent over the course of a single year. The Cornell study of Seattle notes a wide variation in earnings among individual drivers. In fact, the Cornell study (Hyman p. 121) finds that 26 percent of full-time drivers earned below Seattle’s $16.39 minimum wage. How much drivers are paid on average does not tell us what share of drivers earn less than the minimum wage, and how much drivers would benefit from employee status.


We stand by the conclusion of our original report that, should Proposition 22 pass, Uber and Lyft drivers could earn as little as $5.64 an hour, well below California’s minimum wage.


[1] Jacobs, Ken and Michael Reich. 2019. The Uber/Lyft Ballot Initiative Guarantees only $5.64 an hour, UC Berkeley Labor Center and UC Berkeley Center on Wage and Employment Dynamics.

[2] Caldwell, Sydnee and Emily Oehlsen. 2018. “Monopsony and the Gender Wage Gap: Experimental Evidence from the Gig Economy.”

[3] Thornberg, Christopher, et al. 2020. ”Proposition 22: Analyzing the Impact on App-Based Drivers’ Earnings.” UC Riverside Center for Economic Forecasting and Development.

[4]Hyman, Louis et al. 2020, “Platform Driving in Seattle.” July 6, 2020.

[5] Hall, Jonathan, and Alan Krueger. 2018. “An Analysis of the Labor Market for Uber Driver-Partners in the United States.” ILR Review 71, 3: 705-32.

[6] Parrott, James, and Michael Reich. 2020. A Minimum Compensation Standard for Seattle TNC Drivers. UC Berkeley Center on Wage and Employment Dynamics. July 2020.

[7] Mishel, Lawrence 2018. Uber and the Labor Market. EPI Report May 15, 2018,; Henao, Alejandro 2017. “Impacts of Ridesourcing—Lyft and Uber—on Transportation.” Ph.D. Thesis, Department of Civil Engineering, University of Colorado,; James Parrot and Michael Reich 2020.

[8] Authors’ analysis of data in Stein, Alison 2020. “Analysis on Impacts of Driver Reclassification.” Medium May 28, 2020.

[9] Tucker, Catherine 2020. “Declaration of Dr. Catherine Tucker and Appendix A-B in support of LYFT Inc.’s Opposition to Plaintiff’s Motion for Preliminary Injunction.” Superior Court of the State of California in and for the County of San Francisco. Case No. CGC-20-584402. Tucker finds an average of $20 net earnings in the state. Her expense accounting contains many of the flaws that we discuss below for Thornberg.

[10] Cook, Cody, Rebecca Diamond, Jonathan V. Hall, John A. List, and Paul Oyer. 2020 May. The Gender Earnings Gap in the Gig Economy:Evidence from over a Million Rideshare Drivers.

[11] Mishel 2018.

[12] California Air Resources Board 2019. “SB 1014—2018 Base-year Emissions Inventory Report.” December 2019.