How should we provide benefits to gig workers?

Shutterstock / Daisy Daisy

The past decade has seen an explosion in the market for “gig” workers who provide a wide variety of services for “platform companies” that contract for their services but do not count them as employees. Most notable is the growth in work at Transportation Network Companies (TNC) for which individuals drive customers from place to place (such as Uber and Lyft) and Delivery Network Companies (DNC) for which individuals drive goods from place to place (such as DoorDash and InstaCart). But the gig economy has exploded along other dimensions as well, such as freelancers on sites like Upwork and even crafts creation on Etsy.

The growth has been astonishing, although data on gig work is challenging to measure. Garin et al. (2023) use tax data to measure those earning income from platform companies. Since the nascent stages of such platform work in 2012, the number of gig workers has grown by 5 million, and by 2021 comprised approximately 3% of the U.S. workforce—about 30% of the self-employed workforce in the U.S. This was driven primarily by growth during the pandemic, where the number of platform workers more than doubled. Meanwhile, a survey of workers from the Pew Foundation in 2021 found that 9% of the workforce reported working for an online platform in the past 12 months; the difference across these sources partly reflects minimum earnings requirements for reporting gig earnings for tax purposes. The number of drivers for Uber and DoorDash is likely over 1 million each. Revenues of Uber and DoorDash are up roughly five times since 2017.

This shift has a wide variety of implications for the labor force and for society more generally. One of the most important relates to the provision of employee benefits. For most Americans, employer-sponsored benefits provide their primary source of protection against the costs of illness, short-term income losses due to injury, sickness, or unemployment, and income loss in retirement.

The employer-provided model doesn’t work well for gig workers, however. In the traditional labor market, employers provide these benefits to their full-time workers after some period of attachment to the firm. Gig workers value the flexibility of being able to vary their hours across platforms and over time. This makes it difficult to apply the traditional benefits model, raising questions such as who is sufficiently attached to the firm to qualify for benefits, and how to provide benefits when workers work across multiple platforms.

There are two solutions to this problem. The first is to reclassify gig workers into traditional employment roles, as has been proposed in recent legislation in some states, such as California’s AB 5. But such an approach is problematic for two reasons. The traditional employer model is eroding, and putting gig workers into a declining system does not appear to be an effective option. Moreover, the traditional employment relationship may constrain the flexibility that gig workers are looking for in these arrangements. As I show below, most gig workers are using this work to supplement their income, not as a main source of income, and the flexibility that these positions bring would be constrained by the rules of engagement for traditional employer-employee relationships.

The second is to introduce a new model that can address the problems that gig workers face in providing for their income and health security. I propose such a new model in this paper. This model is based on the lessons learned from a study I carried out surveying Uber drivers about their benefits preferences. In this paper, I begin by reviewing that study and the key lessons learned. I then turn to recommending a model for a new benefits platform that would provide security for gig workers.


Employer provision has been the dominant source of benefits for those in the U.S. for more than 70 years, after the political inability to provide a universal government-provided system interacted with the wage and price (but not employee benefits) controls imposed during World War II. This system revolved around lifetime employment with an employer who would provide a generous package of wages and benefits. But benefits have become much less generous over time, most notably with employer health insurance coverage of workers falling from 66% in 2001 to roughly 60% in 2021. And coverage remains very partial for employer-based savings plans, for example with only 56% of workers participating in an employer-based retirement plan.

To date, gig workers in the U.S. have been classified as independent contractors, and as such they have only been provided benefits in fairly limited situations. A lack of benefits protection raises a number of significant risks for these workers. For example, retirement is not always in the control of employees, with forced retirement contributing to the timing of job leaving for over half of retirees. One-fourth of non-retirees have no retirement savings, and less than 40% of workers felt that their retirement savings are “on track.” Among nontraditional workers in particular, 30% had saved for retirement in only some months, rarely, or never in the past 12 months.

Health insurance coverage options are available to all through the Affordable Care Act (ACA) exchanges, but premiums can be as high as 8% of income and plans can also lead to significant out-of-pocket spending risk for enrollees (although for the lowest-income workers, this is mitigated by income-related cost-sharing subsidies available through the ACA exchanges). And workers face a host of short-term income and spending risks from factors out of their control, ranging from sickness of themselves or a family member to a shortfall of work opportunities. Partly as a result, 16% of adults are not able to pay all of their current month’s bills in full, and more than one-third of adults say that they could not cover an unexpected expense of $400 completely using cash or a credit card that they could pay off.

Given these shortfalls, there is a strong argument for strengthening benefits for gig workers. But doing so through traditional employer-financing is a challenge. Benefits are tied to specific firms, which is incompatible with a sector where work is spread across multiple firms, both within and outside of the gig economy, and where workers don’t work steadily. Using internal Uber data, I find that among the sample of workers who were active on Uber at the start of a quarter, “only 22% work 20+ hours for four consecutive weeks; only 13% work 30+ hours for four consecutive weeks. Moreover, less than 5% of such workers work 20+ hours every week in a quarter, and only 2% work 30+ hours every week in a quarter.”

Reclassifying gig workers as traditional employees would not generally solve this problem because benefits are typically only available to employees with a long period of continuous work for a firm, such as the 2-month average waiting period for health insurance. Only 40% of part-time employees have access to retirement benefits. Even in places where employers are required to offer paid sick leave, there is typically a 90-day waiting period before using accrued leave. Most strikingly, of those firms that offer health insurance to their workers, only 14% offer coverage to part-time employees.

What do gig workers want?

The natural starting point for assessing the proper structure of benefits for gig workers is to understand their working circumstances and their preferences for benefits. Without a rigorous mandate on platforms to provide benefits or for workers to have them, it may be difficult to expand benefit coverage among gig workers. It is therefore important to understand what aspects of benefits provision are most attractive to them.

To provide this background, I worked with Uber, Inc. to do a survey of their drivers in the summer of 2021. The survey universe was all active drivers outside of California, and drivers were offered a $10 Amazon e-gift card as an incentive. There was a 0.34% response rate, yielding 1,063 responses. This is a very low response rate so the results should be not interpreted too precisely but rather as a convenience sample to suggest gig worker interest in benefits availability. The respondents were then weighted to match the characteristics of the overall driver population. Compared to the total U.S. workforce, the sample was more male, more non-white, and had both fewer low-educated (high school dropout) and highly educated (graduate degree) individuals.

We surveyed drivers on a wide variety of topics including their background characteristics, their sources of work and earnings, and their preferences for benefits. We considered four types of benefits: contributions towards health insurance premiums; contributions to a retirement account; contributions to an account that could be used for sick leave; and contributions to an account that could be used flexibly for short-term savings needs. To measure the intensity of preferences, we used an “unfolding” survey method to sequentially measure the value that drivers place on a dollar of benefits versus a dollar of cash.

The findings of this study were quite informative. Eight findings are particularly relevant for thinking about benefits design for gig workers (although once again the findings represent the views of the small share of drivers who responded):

  1. Drivers spread their time widely across apps. Seventy-five percent of Uber workers have worked with at least one other platform company, and more than 40% have worked with two or more.
  2. This sample of drivers largely uses the work to complement other jobs. Almost half of the drivers earn less than $5,000 per year from platform work, yet more than three-quarters of our sample has family income above $20,000 per year. Only 20% of the sample earns more than half their income from driving work, while almost one-third earn less than 5% of their income from this source.
  3. Drivers typically—but far from universally—have benefits provided from another source. Seventy-seven percent have health insurance coverage, with just over half of that coming from their own or spousal employment. Roughly 47% have a retirement savings account, almost all from employment outside of gig work.
  4. As is the case for the broader population, many drivers do not have large financial resources available for emergency spending. Fewer than half of those surveyed have $1,000 in liquid assets available to cover current expenditures, and more than one-third said that they had less than $400 available.
  5. Drivers value benefits highly. Almost 30% of drivers value an additional contribution to their retirement account or flexible savings account at more than $1, and another roughly 15% would value this contribution at $1. After accounting for the tax advantage of benefits, around half of drivers would rather have a dollar in retirement benefits than a dollar in cash!
  6. Drivers value retirement or flexible savings accounts more highly than other benefits, including health insurance coverage. This is partly because most drivers have health insurance from another source; among those without health insurance, health insurance is valued similarly to retirement coverage (but still less than flexible savings accounts)
  7. Drivers value commitments to save. Only one-quarter of drivers would prefer a savings account that lets them flexibly remove the money; almost one-quarter would prefer not to have that choice, and fully half are indifferent. There is even substantial enthusiasm for a “save more tomorrow” option that allows workers to devote a share of earnings from high-earning weeks towards a retirement savings account; if such contributions were matched, 80% of drivers would like to use such a mechanism.
  8. Drivers have widely varying preferences for benefits. Forty percent of drivers value retirement benefits most highly, but more than 20% vary health benefits most highly, and more than 35% value cash for sick leave most highly. Ranking all combinations of sick leave, health insurance, and retirement benefits, every single combination is preferred by at least 8% of the sample, and no single combination is preferred by more than 25% of the sample.

Policy design

The desire of gig workers for work flexibility, matched with their desire for benefits protection, suggests a new model for providing benefits. This model could be designed as follows.

Government-sponsored gig worker benefits platforms

Each state would be mandated to have a new Gig Worker Benefits Platform (GWBP) that can host benefits for such workers. Following the model of the ACA exchanges, these platforms could be run by the states or the federal government could set up a default model that is available to each state. The GWBP would provide a user-friendly platform on which gig platforms and workers could contribute to benefits accounts.

The state would define a set of gig employers to be included in this platform, for example those gig employers who meet a certain minimum level of payroll or sales (or both). The government may want to also consider the number of interactions that the worker has with the platform, e.g., the worker has to have at least some minimum number of interactions over some period of time (to include, for example, regular Uber drivers but exclude those who only once or twice have had something for sale on Etsy).

An interesting question is what to do about the majority of self-employed workers who don’t work for one of the platform employers. In principle, to the extent that we are allowing individuals to make their own contributions to these platforms, such workers could be included as well. But this sets up a different set of challenges that are best addressed in version 2.0 of the GWBP.

Regardless, one important lesson from the ACA implementation is clear: There must be sufficient funding and software engineering skills brought to bear in optimizing the functioning of these exchanges. Initial failure, as with the ACA exchanges, can have large spillover effects in terms of delegitimizing the effort.

Mandatory minimum contribution from gig payers

Platforms would have to make a minimum contribution to their state GWBP in respect of gig workers earning above some minimum income level. This minimum income level would be sufficient to cover the expected costs of the new mandated injury benefits described below.

Additional voluntary contributions

Any qualifying platform would be able to make additional contributions on behalf of the workers who use that platform to the GWBP. States could choose to mandate a minimum level, or they could opt to set a default level of contributions above the amount required to fund injury insurance.

Worker contribution

Gig workers would also be allowed to make contributions to the GWBP. Platforms would be mandated to facilitate these contributions. Platforms would have the option to match employer contributions at various rates and levels. But specific platform companies would not set allocations across benefits—either strictly or through defaults—since workers will often be pooling contributions from multiple platforms.

Injury benefits

All gig workers who earn above workers’ compensation (WC) minimum amounts in the state would be eligible for medical and indemnity benefits, which would parallel state WC programs. This would include a cash indemnity benefit that is a fraction of the average weekly earnings on the platform over the previous quarter, similar or identical to WC. Qualification for this benefit would follow the same determination process as WC. States could choose how best to fold this into existing WC programs for traditional employers.

Pooling on platform and worker contributions

Contributions from all platforms would be pooled, along with worker contributions, for the allocation by workers.

GWBP Options

Enrollees in the GWBP would be able to freely allocate their contributions to one of three options:

  1. Contributions towards insurance purchase on state health insurance exchange: The GWBP would be fully integrated with state or federal insurance exchanges. Individuals would be able to assess the cost of insurance and use that to decide on their contributions to an insurance account. These contributions could then be used to partially or fully offset the cost of insurance. Participants with gig income that is sufficiently low to qualify them for free Medicaid coverage or zero-dollar premium coverage would be notified and could explore whether these options are still available after including other income.
  2. Contributions toward retirement savings: The GWBP would facilitate the establishment of Individual Retirement Accounts (IRAs) for gig workers. Individuals who have their own IRA should be able to directly link these new IRAs with their existing accounts. For others, this would be a new retirement savings option.
  3. Contributions toward an emergency savings account: The GWBP would facilitate the establishment of an emergency savings account (ESA) for gig workers. Individuals would get access to a financial account in which they could deposit funds that could be withdrawn for emergencies. There are a number of options states could consider to encourage the use of these accounts for emergencies and not just deferred consumption, such as:
    • Ordeals to withdrawal: When individuals try to withdraw their funds, various barriers could be established to remind them that this is supposed to be just for emergencies.
    • Withdrawals restricted over time: For example, funds would “vest” over five years, i.e., of money contributed this year, individuals could only get 20% next year, 40% over the next two years, up to 100% over the next five years.
    • States could set up limits on balances in emergency savings accounts, beyond which funds are rolled over into retirement accounts.

Financial provider partners

The state would need to partner with a financial provider for both the IRA and short-term savings options. The state should hold a reverse auction among qualified financial providers based on the size of fees that would be charged to enrollees. States should consider allowing multiple winners with default allocation of funds based inversely on fees charged by the institutions.

State rules or defaults

States would have the option of setting allocation rules or defaults across benefits categories. For example, states could allocate contributions in particular proportions or could default contributions across categories.

Save More Tomorrow options

States would offer workers the option to participate in Save More Tomorrow (SMT) programs which would allow the worker to specify a target earnings and a share of income above that earnings that they would default to one of their savings accounts.


An attractive feature of benefits in traditional employment is their tax preference. The ability to tax prefer benefits through the GWBP is therefore an important issue to resolve and may require legal and legislative action. In particular, given that gig workers are not classified as traditional employees, many of the traditional tax rules will not apply without further legislation. There are a few options to consider:

  • Individual retirement account contributions from individuals could clearly be tax-favored through existing rules. Employer contributions would be taxable, but that would be offset by the tax deduction as the individual contributes to their IRA.
  • Emergency savings accounts are unlikely to be tax-favored without new legislation.
  • Contributions towards health insurance from platforms could be in the form of tax-favored Individual Coverage Health Reimbursement Arrangements (ICHRA) or Qualified Small Employer Health Reimbursement Arrangements (QSEHRA). It is unclear under what rules workers would qualify for these plans. For workers, any contributions that they make to health insurance would be deductible from income taxes under the self-employed health insurance deduction. Once again, the treatment of gig workers is unclear, and legislation could be helpful in resolving this uncertainty.

While some tax issues need to be resolved, it is important to note that for the vast majority of gig workers, the income tax burden they face is quite low, so this is not the primary driver of the attractiveness of these benefits.

Decision support

The GWBP would be required to have state-of-the-art decision support to assist gig workers in making their benefits decisions. This would include:

  • Tools that easily allow enrollees to track their contributions and balances.
  • Interactive and live support to consider the tradeoffs in attributing money to different aspects of the GWBP.


The rapid and continuing growth of gig workers in the U.S. suggests that the time is now to address the important issue of how these workers will access the workplace-based benefits that are essential to standard employees in the U.S. The different nature of this work and the diversity of engagement of gig workers suggest that the traditional employee model may not work well for them. In this paper, I suggest an alternative informed by data collected on the preferences of gig workers for benefits. Those preferences suggest that a flexible benefits platform is best but that there are a number of features of the architecture of such a platform that can make it both more appealing and more socially valuable. In this paper, I lay out one such structure in the hope of contributing to ongoing conversations among policymakers on this issue.

  • Acknowledgements and disclosures

    I am grateful to the Brookings Institution for funding and to Miriam Chaum, Richard Frank, Sherry Glied, Mark Iwry, Libby Mishkin and Wendell Primus for helpful comments.

    The study described in this report was funded by Uber Technologies and published by NBER. Uber Technologies had the right to review that research for factual accuracy but did not have control of the editorial content. The author did not receive financial support from any firm or person for this article or, other than the aforementioned, from any firm or person with a financial or political interest in this article. The author is not currently an officer, director, or board member of any organization with a financial or political interest in this article.

    The Brookings Institution is financed through the support of a diverse array of foundations, corporations, governments, individuals, as well as an endowment. A list of donors can be found in our annual reports published online here. The findings, interpretations, and conclusions in this report are solely those of its author(s) and are not influenced by any donation.

  • Footnotes
    1. Garin, Andrew, Emilie Jackson, Dmitri Koustas, and Alicia Miller. 2023. “The Evolution of Platform Gig Work, 2012-2021.” w31273. Cambridge, MA: National Bureau of Economic Research.
    2. This study was funded by Uber, Inc. Results are shown in Gruber, Jonathan. 2022. “Designing Benefits for Platform Workers.” w29736. Cambridge, MA: National Bureau of Economic Research.
    3. For example, Rideshare drivers in New York City receive workers’ compensation, death benefits, vision benefits, dental insurance, telemedicine, and legal assistance and accident support through the Black Car Fund, platform workers in California receive occupational accident insurance protection and healthcare stipends under Prop 22, and rideshare drivers in Washington state receive workers’ compensation, paid family and medical leave, unemployment insurance, and sick pay under HB 2076.
    4. For example, workers are first asked about their preference for a dollar of cash or a dollar of (for example) contributions to a retirement account. If they say that they prefer a dollar of cash, they are then asked if they prefer 80 cents of cash or a dollar in a retirement account. If they say that they prefer retirement, this says that they value dollars in retirement accounts at 80-100% of dollars of cash. If they say they prefer cash, the cash discount is gradually decreased until they prefer retirement contributions, and bounds on value are established.