Gig Economy’s Highway toward Increasing Surplus Labor


gig economy

This article situates the gig economy as a tool for shoring up a falling rate of profit within the U.S. economic system. The main feature of this “gig economy” is that its businesses employ what are called private contractors instead of employees. This change allows a company to eliminate business expenses inherent to the typical economy and to shift these costs to other companies or the public sector, or to the workers themselves. Thus, a new high water mark for the gig companies’ rate of profit may be achieved. As time goes by, this high water mark will not be enough— given a financial world which demands that the rate of profit continue going up from there (and forever). Attempted moves to automation and repeated work speedups are to be expected.

My first-hand experience in the gig economy has to do with ride-sharing companies. Since I have in the past worked in conventional public transportation (passenger buses) I have had a view of both worlds.

Physical Changes

Let me first describe my former bus depot: coming in from the road you have an employee parking lot for several hundred employees, then a street-facing building with a flag pole. One half of this building houses a dispatch office, employee bathrooms, a locker area, a mailbox area, and employee break rooms with sinks, microwaves, toaster ovens, and refrigerators. The other half of the building houses management offices.

Behind that is a bus lot with a maintenance garage to one side. The garage is much bigger than the first building. It contains about a dozen large bays, lifts, and grease pits for working on buses and storage areas for tools, extra parts, and stacks of tires, along with more offices, bathrooms, and break rooms.

Further back is a large bus wash with rotating scrubbers, air dryer, etc. Behind that are natural gas storage tanks and fueling area with another office and bathroom. As you get to the back fence you see that the property extends around to one side with two more large parking lots for the buses. If every one of the hundreds of buses in the fleet were back in the lot at one time they probably wouldn’t fit. Indeed, the dispatch has to keep track of which buses are parked in the through lanes (between the spaces) overnight, to make sure they assign those buses first in the morning.

In contrast, Lyft is a gig company like Uber and a myriad of other companies that claim to use driver contractors to deliver passengers, food, or other things. Though combined it is likely they count over a million employees, there is not a single employee parking lot; the workers find parking on their own. There’s no human dispatcher, it’s an algorithm. If an employee has non-emergency questions, there’s no one who can answer—figure it out yourself: sink or swim, then get stuck with the blame if it goes wrong.

No manager can hear a real-time appeal if the algorithm records a situation wrong. Someone may check emails once every 24 hours and give a canned response. Every communication back and forth can take days.

There’s no maintenance garage, the worker pays for maintenance and insurance. If Lyft is renting the car to the driver, the company pays but the average cost is still factored in to the rental fee. Fueling and washing one’s vehicle inside and out is a necessary task done by the workers and their own expense, and with the time unpaid at that.

Finally, there are no employee bathrooms… Usually a driver has to log off just to look for a bathroom. This is unpaid time when they could be missing fares. Often the retail sector or public facilities such as libraries are left to assume the costs of providing bathrooms to gig sector employees.


A true private contractor would set their own rate, however the reliance of employees and potential customers on a particular cellphone app allows the gig economy to operate on a take-it-or-leave-it approach.

Companies like Lyft claim no authority over private contractor employees, yet retain the privilege of setting the rate of compensation on their platform. A true private contractor would set their own rate, however the reliance of employees and potential customers on a particular cellphone app allows the gig economy to operate on a take-it-or-leave-it approach. A freely associated individual worker would be hard pressed to develop their own app and market it to the public, and by that time would price themselves out of the car transportation market. In this way the principle of “socially necessary labor power” works to favor the gig company that exploits workers as a crucial aspect of the business. The power of the computer lets them do so on an unprecedented scale.

Gig companies like Lyft use a pay system similar to a taxi farebox, though the rates are quite reduced. They pay a fixed rate (50-80 cents) per mile plus a different fixed rate (10-15 cents) per minute. On top of this they offer bonuses of one type or another that can be arbitrarily given and taken away. By changing the bonus amount each week, the company is able to moderate the number of drivers on the road and adjust the real wage of these drivers according to what the companies perceive the rideshare market will bear. Since this business might fluctuate wildly over certain holidays or school breaks and the in-between times, the variable bonuses are a valuable tool for buttressing the companies’ bottom line.

The computerized analysis of profit vs. on-the-clock workforce is not unique to the gig economy. Neither is the preference for giving employees non-permanent gifts/perks instead of an extra nickel to their wage. However, the impersonal nature of the platforms and the computerization of every other aspect of “management” allows the gig companies to use these tactics on a hyperactive level. They will soon cause similar changes within the entire market, as brick-and-mortar taxi and car service companies evolve to stay competitive.

Transfer of Risk

Any employee who earns tips is in effect a stakeholder in the daily success of the business they work for. This is by design. If business is slow, tips decline, and if business is booming, tips increase. Without tips, the business would be paying their workers some higher fixed rate and therefore carry all the risk/reward of boom/bust cycles. Often tipped employees such as restaurant workers are allowed to be paid less than minimum wage, so long as tips make up the whole difference.

For transportation-based gig workers who work using a motor vehicle, another risk spectrum is situated between slow areas and busy areas. If the worker gets stuck in a slow area at the destination point of a passenger, they have a choice between 2 bad options: wait in place for a customer who might take a long time (maybe hours) to appear, or drive in the direction of a busier area while empty. Either choice involves unpaid labor time which causes their effective hourly wage to fall far below minimum wage.

Further transfers of risk include the refusal of gig companies to pay into the unemployment insurance system, the refusal to pay the employer portion of FICA/payroll taxes, and the refusal to offer group health insurance. The flimsy legality of these final 3 refusals stems from an insistence that the employees are not employees but private contractors. I return to this topic later.

Artificially Intelligent Dispatch (AI)

While a lot of what I have described about app-based car transportation could equally apply to conventional taxis, the two industries are about to diverge drastically.

Dispatch by algorithm is a technology that holds promise if used with humanistic aims. It potentially saves gasoline and vehicle depreciation for drivers who would otherwise have to roam the streets scanning for passengers with their hand up. Instead of a sytem where a passenger flags down a passing car, rideshare apps have the passenger fill out information about their destination and time they will be ready. The app automaticaly records their current position and then uses an algorithm to assign their itinerary to a specific driver.

This is sometimes (but not always) the closest driver who is signed in as ready. The exceptions can be benign, like when a driver has specified they only want to pick up rides that take them in a certain direction, or perhaps another passenger has been waiting longer in the same area.

Other times, it is conceivable that the gig company’s algorithm disadvantages the driver and/or the passenger, seemingly without their knowledge, and presumably to advance the profit interests of the company.

A word of caution: companies like Uber and Lyft zealously guard their algorithms as proprietary and secret. While they may be justifiably worried about people copying their work, another and perhaps bigger reason for secrecy is that the algorithm allows them many opportunities to willfully exploit consumers and workers, if only the embarrassment of verifiable public scrutiny can be avoided.

Even if we don’t know what is exactly in the algorithms or how often they are changed, we can still come to reasonable conclusions about what the companies potentially could accomplish with algorithms that are able to instantly cross-reference stores of metadata on drivers and passengers.

From now on, I will refer to these algorithms as “AI.” This term is meant to simultaneously include both the possibilty that every quantitative variable is coded in by a human worker in the programming department as well as the possibility that the algorithm may be coded to “learn” correlations between data sets and apply these variables automatically to its calculations.

Pricing Each Worker

In the first few weeks of my experience as a rideshare driver, I was nearing the completion of a weekly bonus, which was a set number of rides that had to be completed by 4:59 a.m. on a Monday morning. I had made a rookie mistake of overestimating the passenger demand on Sunday nights. By 10 p.m. I was within 3 rides of the goal, but agitated with endless waiting I had come up with an idea: my friend could request a superfluous, yet short ride. It would be cheap for him and get me closer to achieving the bonus. When he made the ride request, Lyft connected him not to me, but to another driver much farther away. It felt like I had a price tag relative to the other driver, and the AI dispatcher knew it.

Other drivers anecdotally report that they’ve noticed this effect. When you get closer to achieving a bonus, your probability of getting assigned to a passenger falls.

Employee Retention

Following the end of a full-time summer job I logged back into Lyft after a 3-month hiatus. Tips were much more easy to come by than I remembered. Whereas in other months maybe 1 in 20 passengers leave a tip, I experienced several days where about 1 in 3 passengers tipped me (and were noticeably nicer). However, as the days went on, my rides seemed to revert back to the earlier averages.

In Lyft and many other apps, customers are asked to give workers a numerical rating, and vice versa. This theoretically allows the categorization of pleasant and unpleasant drivers, pleasant and unpleasant passengers.

Could it be that the AI was assigning me to likely-pleasant rides with the impression that they could hook me in as a regular driver again? There’s no reason I deserved a pleasant passenger more than another hypothetical driver who had loyally worked all the way through summer. And yet it seemed to me this was the case, that the AI is able to dole out rewards to drivers who seem discontent.

The logical conclusions are more putrid when you realize the opposite possibilty also exists. For example, there is nothing holding the AI back from closely observing the approximate limit of exploitation a particular worker will put up with at a certain time, and make decisions accordingly. It can even punish drivers whom the companies hope to get rid of, all with plausible deniability.

As a Lyft driver, when you logout of your app for the day, a menu always pops up asking how you would rate your experience on a 1 to 5 scale. There is no way to leave any substantial feedback, and no one has ever followed up no matter what I entered. After my experience with the days of increased tips I was convinced that this survey is merely a method for the free collection of pure quantitative contentment data on drivers. Although it’s impossible to close the menu without choosing a number, it is possible to swipe up and force the app closed, and that is what I started doing.

Customer Retention

When you acknowledge the possibilities for what gig company’s AI could do to workers, it is also conceivable that it happens on the passengers’ end. As a passenger, does the AI notice how long you are willing to wait? It wouldn’t be fair to prioritize impatient passengers but it is likely to be financially beneficial and could avoid Uber and Lyft from losing the passenger to the other company. So will the companies do it if they are confident the patient rider won’t be able to see they were skipped?

There are few limits on the flexibility of the gig companies when we consider the AI dispatch program has instantaneous access to so much information and can also double and triple as a staffmember from multiple managerial departments including human relations to customer service to expenditure management. Conventional companies may try to take similar actions with a full staff but still not match the speed and interconnectedness of the high-tech gig economy.

Worker Actions

I’ve written several articles regarding strikes and protests carried out by workers in the gig economy. Other drivers have also taken up a pen. Here’s a tidbit from another (anonymous) author:

What gets me is companies like Uber and Lyft have made billions off the “gig” economy or what we called piece rate. The workers took all the risks and the owners made all the profit. Time to shove all the risk on their billionaire owners and the profit to the workers. Amazon without warehouse workers would fold, same for all the delivery and livery drivers for other companies.

Indeed, workers have made various demands over the years for leveling, to even out the risk and the profit.

Pressure on Municipalities

Many of the earliest protests were of taxi drivers trying to keep Uber out of their cities. This has had some early successes. (And recent ones: last summer London declined to renew Uber’s contract, denying them the ability to operate within that city.)

As the apps grew in popularity, however, the companies schmoozed the cities and negotiated on the terms that would need to be followed in order to be allowed in. Five to 10 years ago, U.S. cities began to see taxis with Uber and Lyft stickers, meaning the drivers had begun splitting their time between the apps and their company meter. There are still labor gains for taxi drivers every so often, such as LAX airport moving all rideshare services to an external lot, with only taxis allowed to pick up curbside.

In New York City there is an inflated minimum wage of $27.86 for Lyft and Uber drivers that was won through strikes and protests. The Los Angeles city council is still “studying” a proposal for a $30 minimum wage. These specific wage controls are a tacit acknowledgement that workers unfairly shoulder the responsibility for so many business expenses that in a conventional business model would normally fall on the gig companies. This is reflected in the L.A.-based Mobile Workers Alliance slogan: “$15 for our work, $15 for expenses.” The grassroots, driver-led “union” Rideshare Drivers United has held at least 3 strikes in the past year and a half. These were not successful in getting Uber and Lyft to reverse recent pay cuts but they were instrumental in raising public attention enough to the point that the California Assembly enacted a state law (AB5) to classify the drivers as employees of their apps.


The classification of “employee” is an important legal distinction because it potentially opens the door to certain benefits, including but not limited to:

  • minimum wage (hourly)
  • overtime pay
  • employer contributions to social security and Medicare payroll taxes (without an employer contributing half, the worker must pay both halves)
  • employer contributions into the unemployment insurance system (can typically be paid only by the employer)
  • workers’ compensation
  • the right to form a registered union, hold a union election under NLRB rules
  • group health insurance (or the employer is liable to penalties under ACA)
  • reimbursement for fuel, depreciation, insurance, and other expenses

When AB5 was first introduced, the platforms tried to text drivers to encourage them to call their state legislators and urge these legislators to vote no. The companies said the Assembly was trying to take away flexible hours. This push wasn’t successful.

Neither Lyft or Uber have complied with AB5 since it came into effect on Jan. 1, 2020. Drivers have not been asked to fill out any type of payroll form such as a W-4. One of Uber’s top executives told a mainstream news outlet in late 2019 that his plan was to simply ignore the law and not follow it, then leave it to company lawyers to work out.

In the meantime, a threshold for signature gatherers was reached in late May 2020 to allow the voters a chance to repeal AB5 in November. The paid signature gatherers had posted themselves outside of progressive gatherings like the Womens’ March and had said the initiative was about giving drivers the choice to be employees or contractors. In truth, there will be no such choice, since the companies will simply not hire the applicants who make the wrong one. However, many marchgoers likely found it difficult to make sense of over a dozen different proposed initiatives they were being asked to sign off on while they were walking by. There is little doubt voters will be bombarded this fall with numerous advertizements paid for by Uber and Lyft, given the companies’ plan to spend $110 million over this issue.

Acute Implications amidst Covid-19

In the current pandemic, Uber and Lyft drivers who have not worked for a W-2 employer in the past 18 months may have applied for unemployment and been approved for a weekly benefit of $0. Without at least $1 per week, they will not receive the extra $600 per week supplemental payment issued by Congress. In this way Lyft and Uber have bilked their drivers out of potentially thousands of dollars in stimulus simply because of an ideological unwillingness to pay a maximum of $7.00 of unemplyment tax per employee per year (0.1% of the first $7,000 in earnings).

Instead of suing Lyft and Uber for back taxes like New Jersey did earlier this year, California has opened up a parallel program for gig workers who can’t get any money out of UI. It is inferior to the regular UI program considering it is temporary and  lasts for a shorter time.

Future Protests

If there is an eventual waning of coronavirus, it is possible that passengers will reappear and so will the drivers. I anticipate more actions having to do with demanding unemployment compensation, healthcare, sick leave. Just before the stay-at-home order statewide, drivers were having socially-distanced protests over sick leave in front of emptied out corporate offices. Also I expect to see educational actions to help defeat the repeal AB5 initiative.

I hope to find unions and rank-and-file workers from non-gig sectors coming into this fight. Bus drivers are an example of a class that is put at enormous risk by the gig companies’ drive for market share and self-driving vehicles. A University of Louisville study has found that increased rideshare use led to a decrease in the use of buses even while it increased train ridership. The apps have already convinced some towns (ex: Monrovia) and campuses (ex: USC) to replace shuttle service with subsidized ridesharing.

A Short Note on Automation

Years ago, Uber and Lyft used the excuse that they are “tech” companies to weasel out of earlier state court cases that would have required employee protections for anyone who performs transportation functions for a transportation company. Still, it is worth monitoring the companies’ vast investments in self-driving vehicle tech.

Earlier this year, it had looked like Uber was on track to miss their 2015 goal of self-driving cars on the road by 2020. However, the company does have cars test driving around Pittsburgh, even if they aren’t carrying passengers yet due to safety concerns. Furthermore, technological advancement generally moves forward on an exponential curve, meaning that when the cars are improved, we can expect them to improve at an ever faster rate and be on the streets before we know it. Self-driving tech is Uber’s and Lyft’s end-all-be-all, most of the reason for their inflated stock prices, and they won’t give up on it.


The gig economy is a microcosm of the economy as a whole and therefore worth the attention of all workers. Its pursuit of a higher rate of profit through all kinds of shenanigans only makes absolutely clear how the basis for that rate of profit is in paying the workers less and less for their labor, and in having the authority to dispose of more and more of their labor power through unpaid time on the job.

As we move forward, the imperative is to see the positive in the negative. Through opposing the trends we see in the gig economy, we are saying something about what a more fair economy would look like. As we proceed to effect changes, we are bound to notice other contradictions and change those as well.

Buddy Bell

Buddy Bell is a labor activist and a part-time driver for Lyft.