Skip to main content Accessibility help


  • Access
  • Open access
  • Cited by 2


      • Send article to Kindle

        To send this article to your Kindle, first ensure is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about sending to your Kindle. Find out more about sending to your Kindle.

        Note you can select to send to either the or variations. ‘’ emails are free but can only be sent to your device when it is connected to wi-fi. ‘’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

        Find out more about the Kindle Personal Document Service.

        Retirement plans for contingent workers: issues and options
        Available formats

        Send article to Dropbox

        To send this article to your Dropbox account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Dropbox.

        Retirement plans for contingent workers: issues and options
        Available formats

        Send article to Google Drive

        To send this article to your Google Drive account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Google Drive.

        Retirement plans for contingent workers: issues and options
        Available formats
Export citation


This paper examines retirement saving policy for independent – or contingent – workers, a growing segment of the workforce. Because few of these workers are covered by employer-sponsored retirement plans, they often do not benefit from payroll deduction, employer matching contributions, automatic enrollment, and other provisions that encourage retirement saving. Better use of fintech, judicious changes to tax policy, and expanded Automatic IRAs would help independent workers save for retirement. In addition, we propose the creation of retirement saving accounts that attach to the worker as a supplement to, and possible replacement for, the current system of employer-sponsored accounts.

In the traditional employer–employee relationship, workers earn a salary or hourly wage and receive fringe benefits. Today, however, a wide variety of workers has alternative work arrangements. These workers may be full-time, part-time, or seasonal. They range from independent contractors and white-collar consultants to blue-collars workers such as security guards and maintenance professionals. The rise of the so-called gig economy – which allows people to hail cabs through Lyft, find household help through TaskRabbit, and book lodging through Airbnb, among other applications – has brought these workers into the spotlight. But similar issues have been present for the last several decades, as employers have increasingly substituted independent contractors or firms employing them for traditional employees. Contingent workers now account for a significant and, by most estimates, growing share of the US workforce, and their role in labor markets is expected to continue to rise.

This paper examines retirement saving policy for contingent, or independent, workers, terms we will use interchangeably. It also covers gig workers, a subset of this group who work usually temporary, flexible jobs for an online entity. Contingent workers face many challenges in preparing for retirement. On average, they earn less than traditional employees and have more volatile income, giving them less disposable income available to save and making it harder to set aside regular saving amounts.

The current employer-focused private retirement system was not designed with these workers in mind. As a result, contingent workers have less access to and coverage under work-provided retirement plans than traditional workers. Thus, they also tend to miss out on provisions that encourage retirement wealth accumulation, such as payroll-deduction contributions, automatic enrollment, and employer matching contributions. In addition, under the current system, job changes create an unnecessarily high likelihood of pension leakage. However well or poorly it serves the needs of traditional workers, the current retirement system does not meet the needs of contingent workers, who plausibly have less job stability than traditional workers, or for those people who switch back and forth between contingent and traditional jobs.

Of course, anyone with earnings can contribute to an Individual Retirement Account (IRA), but only a small percentage of people do so on a regular basis (Copeland, 2018).1 Moreover, IRAs have lower annual contribution limits than 401(k) plans and cannot accept employer matches, limiting the amount that can be accumulated. For the most part, automatic payroll deductions into an IRA only currently exist in a handful of states that have implemented Automatic IRAs.

To help contingent workers save for retirement, we propose several changes, ranging from using fintech more effectively, to expanding and developing automatic payroll-deduction IRAs, to judiciously expanding tax incentives for retirement saving.

We also present a more comprehensive solution – which we call employer-facilitated accounts – that would decouple retirement saving plans from the employer, so that individuals would have retirement accounts – much like their Social Security accounts – that follow them across employers and across various work arrangements. This new retirement saving vehicle would provide workers with access to payroll deductions and employer matches and would have contribution limits consistent with other existing plans. It would accommodate the diversity of the contingent workforce and the resulting need for new forms of retirement saving.

The reform would provide stable retirement saving arrangements for people who move back and forth over the course of their careers between traditional jobs and contingent positions. If this phenomenon becomes more common in the future, as we believe it will, ensuring that retirement saving plans function well in the presence of multiple job changes will be essential. Reforms targeted only at contingent workers would not meet this criterion. Although we propose the plan as a way to improve retirement saving for contingent workers, once employer-facilitated plans are established, both employers and employees in traditional jobs may find that this new vehicle is preferable to existing retirement accounts. For example, our proposal would reduce leakage from retirement accounts upon job changes and it would eliminate the dependence of workers on employer choices to offer retirement plans.

While this reform focuses on supply, we recognize that there is also a significant question about contingent workers’ demand for retirement savings. It is natural for those with irregular earnings to prioritize covering immediate needs over meeting long-term retirement goals. Failure to save for retirement has real consequences, however, so even contingent workers with irregular earnings should take advantage of retirement saving options.2 Making retirement contributions a regular part of the hiring process should encourage more contingent workers to save, especially if our more radical proposal is adopted. In addition to making saving easier, having a single retirement account that goes with the worker from job to job it would make it easier to keep track of past saving and encourage additional contributions.

Sections II and III provide background on the contingent workforce and the characteristics of contingent workers. Section IV explores several options for improving independent workers’ retirement security. Section V concludes.

II. The contingent workforce

In traditional employment situations, workers are salaried, or they work specific hours or shifts for an hourly wage. In contrast, contingent workers work on an ad hoc basis and are paid by the service or good they provide. In theory, contingent workers have more flexibility than their traditionally employed peers and may be able to choose when and from where they work. For example, a Lyft driver need not commit to driving 8 hours per day or at what times during the day those hours occur. However, many other contingent workers are at the mercy of their employers and earn income only when they are needed. Some have regular hours, but others are on call and get work sporadically. These workers could arrive at their job to find that the employer's needs have changed and they will not be able to work or receive pay for that day.

Many contingent workers occupy a nebulous space between ‘employee’ and ‘independent contractor.’ The Internal Revenue Service offers three criteria that employers must consider when determining a worker's status: (1) The extent to which the employer controls how a worker does his or her work; (2) The extent to which the business aspects of the work are controlled by the employer; and (3) The existence of written contracts and/or benefits, as well as the nature of the employer–worker relationship and the importance of the work to the employer (IRS 2015a). The U.S. Department of Labor (2015) issued guidelines saying that workers who rely on a single stream of income should be formally considered employees, though the U.S. Department of Labor (2017) later withdrew those guidelines.

Some recent cases at the state level have been more in line with the guidelines initially issued by the U.S. Department of Labor (2015) in classifying gig workers as employees. The Supreme Court of California (2018) set a precedent that employers should classify independent contractors and employees according to the ‘‘ABC’ test.’ Using this standard, a worker should be classified as an independent contractor ‘only if the hiring entity establishes: (A) that the worker is free from the control and direction of the hirer in connection with the performance of the work, both under the contract for the performance of such work and in fact; (B) that the worker performs work that is outside the usual course of the hiring entity's business; and (C) that the worker is customarily engaged in an independently established trade, occupation, or business of the same nature as the work performed for the hiring entity.’ Other states, including New Jersey, Illinois, and Massachusetts, use similar tests (Wiessner, 2018).

Some contingent workers, however, have characteristics of an employee and a contractor; this is especially true for gig workers. For example, a dog walker who uses Rover as an intermediary can choose when to work and what to charge (like an independent contractor) but must still provide walks of a certain basic length and quality (like an employee). Harris and Krueger (2015) propose to address these issues by creating a new category of worker – ‘independent workers’ – for people who are neither fully employees nor fully contractors.

The narrowest measure of the contingent workforce includes only workers with temporary employment arrangements and excludes self-employed workers. The U.S. Bureau of Labor Statistics (BLS, 2018) estimates that about 1.3% of the working labor force fell into this category in 2017. The broadest definition includes all workers with nonstandard work arrangements, from day laborers and agency temps to consultants and independent contractors. Just over 40% of all workers – including seasonal and part-time workers – fell into this category in 2010 (U.S. Government Accountability Office 2015).

Between those two extremes lies a definition that includes those who have either a contingent job or an alternative work arrangement as their primary job. The BLS (2018) found that 15.5 million people, representing 10.1% of the workforce, had an alternative work arrangement in 2017. About 5.9 million people, accounting for 3.8% of the workforce, had contingent jobs. In total, after adjusting for people who had both a contingent job and an alternative work arrangement, BLS (2018) found 19.8 million people (12.8% of the workforce) people who were independent workers in their primary job.

These figures are slightly down from the BLS (2005), which surveyed labor market conditions in 2005. Most other estimates, however, such as Katz and Krueger (2016), JPMorgan Chase and Co. (2016b), and QuickBooks (2016), show an increase in the number of independent workers over this period. The BLS surveys count people who have independent work as their primary job and thus does not capture what appears to be vast growth in independent work as an individual's second (or third) job.

While it receives a great deal of publicity, the gig economy accounts for only a small portion of contingent work. For example, JPMorgan Chase and Co. (2016a) estimate that only around 1% of adults earn money through the online on-demand economy.3

III. Characteristics of contingent workers

There are two broad views of contingent workers and contingent work. Some people see contingent work as an opportunity for people who have full-time jobs to generate supplemental income or for people to have work that is flexible in terms of time and location. In this view, the increase in contingent work is supply-driven; more people have become contingent workers because they want to be. An alternative view is that contingent work represents the further erosion of basic standards and protections that workers have had in the past. Under this view, the increase in contingent work is demand-driven, as employers find new ways to cut costs. Given the diversity of contingent workforce arrangements, there is clearly room for both interpretations to contain elements of truth.

According to the Bureau of Labor Statistics (2018), relative to traditional workers, contingent workers were more likely to be young (under 25), more likely to have less than a high school education and slightly more likely to hold a bachelor's degree. Just under half of traditional and contingent workers are female. About 22% of contingent workers were Hispanic, compared with 16% of non-contingent workers. Contingent workers were concentrated in professional and related occupations and were more likely than traditional workers to be in the construction, extraction, education, and health industries. They were less likely to work in retail trade or manufacturing.

Contingent workers have lower earnings than traditional workers. BLS (2018) reports that, among full-time workers, median weekly earnings for non-contingent workers were about 30% higher than for contingent workers ($886 versus $685, respectively). The difference reflects a combination of age, experience, industry, and occupation effects, as well as any inherent differences in compensation between contingent and traditional work. Katz and Krueger (2016) find significant variation in wages across types of contingent workers, with independent contractors most likely to fall into the highest-earning quintile and temporary workers most likely to fall into the lowest-earning quintile. They also find that contingent workers are more likely to work part-time, tend to work fewer hours per week, and are more likely to have multiple jobs than traditional workers. Contingent workers are also more likely to live in poverty and to receive public assistance (GAO 2015).

Differences in benefits across traditional and contingent jobs parallel the differences in wages and income. Coverage rates for employer-provided health insurance plans are twice as high for traditional workers as contingent workers (BLS 2018). Most importantly for this paper, employer-provided retirement plans were offered to only half as many contingent workers (23%) as traditional workers (48%). Participation rates were even more skewed – 18% for contingent workers compared to 43% for traditional workers. As with wages, the differences in retirement plan coverage and access between contingent and traditional workers represents the net effects of many factors. Beyond this basic information on retirement accounts, however, there is little specific evidence about the retirement security preparedness of contingent workers.

Non-employer-sponsored retirement solutions are available, but Internal Revenue Service (2018) data show that less than 3 million people (1.8% of all tax filers) made a tax-deductible contribution to an IRA in 2015. Even if contingent workers contributed to IRAs at the same rate as the general population – which seems unlikely, given their lower-than-average income – relatively few would have any significant balances in them.

IV. Mechanisms for improving contingent workers’ retirement security

Independent workers almost always lack access to an employer-sponsored retirement account that makes saving easy through mechanisms such as payroll deduction, employer contributions, and automatic enrollment and automatic escalation of contributions.4 Without this access – or other alternatives – this population may face retirement with little more than Social Security.5 To improve contingent workers’ retirement security, we consider several different approaches that may have some promise.

A. Incremental changes

1. Saving through new technology and automatic transfer

Some contingent workers are not paid through a corporate payroll system or on a regular schedule. For example, Lyft drivers have the option to ‘cash out’ and receive their earnings immediately, instead of waiting for their weekly paycheck (Lyft, 2018). In other cases, a worker is paid after a task via a payment platform such as Stripe (Fitchard, 2013). Independent contractors may work a different number of hours each week, which results in an income stream that may considerably vary over time. None of these situations make it easy for the worker to use a traditional payroll deduction retirement savings model.

Recent technological advances and savings apps may provide an answer for those with irregular earnings patterns. Lyft drivers are currently being offered payroll deduction IRAs through the fintech firm Honest Dollar. Other fintech firms offer similar savings opportunities. For instance, Uber has partnered with Betterment to provide its drivers with access to an IRA and financial counseling, with the first year free for its drivers and a discount on subsequent years (Marte, 2016). To compensate for irregular payment amounts and schedules, savers can choose to save a pre-set amount each month, save a percentage of each payment, or to only save when a payment is over a certain amount. Another option leaves it to the saver to decide when and how much to save but sends text messages to encourage saving and reminders when deposits have not been made on schedule.

In addition, Digit and other similar fintech companies offer smartphone applications that analyze a participant's bank account and, when it detects that the account contains money that is not needed for immediate expenses, automatically sweeps it into a savings account. Finally, some employers offer workers the option to split their pay between a checking account and retirement savings account with either a set amount or a percentage of income being saved. Honest Dollar has a patent application that would allow individuals to save for several individually chosen goals and receive regular updates on their progress towards meeting the goal.

These and other automated systems could allow contingent workers, including those with irregular earnings, to save routinely without having to take any specific action. A key feature is the ability to choose the system that best meets their circumstances and income stream. These benefits would be intensified if the workers were also automatically enrolled in the system upon hire. As online employment platforms for contingent workers continue to develop, they could include either benefits or the opportunity to have them as a feature of their systems.

These models do require the contingent worker to have a retirement account. Workers with pre-existing IRAs or other accounts can simply connect to the account in the same way they direct payments to their checking accounts. Workers without such an account would have to set up a new account or have their employers or the fintech provider do so. Fortunately, technology makes this an easy task. Honest Dollar, for example, advertises that it can quickly sign up gig and contingent workers with an IRA that offers a variety of investment choices (Honest Dollar, 2018). In general, an increase in fintech companies and users can be expected to continue to increase the availability of these accounts and to drive down the costs of these services.

2. State-sponsored retirement saving plans and other group plans

Contingent workers will be able to take advantage of group retirement plans such as the new state-sponsored proposals being developed by a growing number of states for workers who do not have an employer-sponsored retirement savings or pension plan. These accounts could take the form of Automatic IRAs or Multiple Employer Plans (MEPs). Initially, state-sponsored plans have been offered through employers, but some states are also making them available to contingent workers as well.

As of this writing, California, Illinois, Oregon, Maryland, New York and Connecticut are setting up plans based on the Automatic IRA (Gale and John, 2017). OregonSaves is already in operation and will make accounts available to contingent workers in the fall of 2018, while both California and Illinois are beginning to test their systems (Bradford, 2018). In addition, Washington State's Small Business Retirement Marketplace offers pre-screened plans through private providers.6 It offers simple, low-cost accounts that are available to contingent, part-time and gig workers.

The Automatic IRA model allows workers without access to a 401(k) or other employer-sponsored plan to nonetheless benefit from automatic enrollment, payroll deduction, and sensible default investments. New hires automatically start making contributions to an IRA unless they actively opt out. These contributions are usually made via direct deposit and sent to a single IRA provider, thus easing the administrative burden for the employer.

Vermont is implementing a MEP, but under current law, it will only be available to traditional employees. MEPs are a type of group 401(k) plan where a number of employers join together to offer a plan with more features and consumer protections than a typical Automatic IRA. MEPs have lower administrative costs and a simpler regulatory structure than a 401(k) and could be offered to independent workers as well as traditional employees if Congress and regulators approve. Currently, MEPs can only be offered through state plans, but Congress could also allow them to be offered by private providers. If it does so, it could also open them to independent workers (Bradford, 2018).

3. Improved tax benefits

While the tax system may have a limited effect on households’ overall level of saving, two simple changes to tax benefits would enable independent workers to increase their retirement balances. The most direct way would be to modify the Saver's Credit, which helps low- and moderate-income workers build retirement savings.

Currently, the Saver's Credit is little known and largely ineffective. Only about three out of every 10 American workers know about the saver's credit (Transamerica Center for Retirement Studies, 2015). In 2015, individuals with earnings of $30,500 or less, and married couples with incomes of $61,000 or less, were eligible for up to $1,000 or $2,000, respectively, if they made contributions to a retirement account (IRS 2014). In 2012, the mean saver's credit that was received was only $127 for individuals and $215 for married couples (IRS 2014). This is largely due to the credit's non-refundability; it can be used to reduce a filer's tax liability to zero, but any ‘extra’ is lost to the filer.

Making the credit refundable and having it deposited directly into the taxpayer's retirement account would make it much more salient for many taxpayers (Brown and John, 2017). Another reform, described in Gruber, Orszag, and Gale (2006) proposed in both legislation and early Obama Administration budgets, would change the credit to a government match; instead of receiving the credit as part of their refund where it is likely to be spent, individuals would have their contributions matched (up to a certain threshold) directly into their retirement account. The rest of their refund would be treated just as it is currently. In addition, the Saver's Credit should be able to be claimed on all tax forms. Currently, it is only available on longer tax forms that are not widely used by its target population. The restructured Saver's Credit would encourage retirement saving and help people who qualify for it to build their retirement balances faster than they could otherwise.

A second tax change that could benefit independent workers would be to ensure that spouses who are employed could contribute the full amount to a retirement account regardless of marital status. Currently, individuals may contribute up to $5,500 per year to an IRA ($6,500 per year for individuals aged 50 and over) (IRS 2015b). Traditional IRA contributions – but not Roth IRA contributions – are deductible, but if an individual or his/her spouse has an employer-sponsored retirement plan, these deductions may be limited (IRS 2015c).

The net effect of tax incentives is subject to debate. Chetty et al. (2013), using data on households from Denmark, find that reducing the tax benefits in one tax-preferred saving account caused high-income households to substitute saving into a different tax-preferred saving account rather than reduce their overall saving. Research that focuses on low-income households, however, generally finds larger impacts of saving incentives on net saving.7 In addition, if the saver's credit is reformed so that it is refundable directly into the account, its value would be as much to improve savings balances for eligible taxpayers as to incentivize participation.

B. A comprehensive approach: employer-facilitated retirement accounts

We now turn to a more far-reaching idea, restructuring retirement accounts so that they follow workers from job-to-job. We refer to these as employer-facilitated accounts, as opposed to the current system, which features employer-sponsored accounts. The reform is designed with an eye towards the needs of contingent workers, but traditional employees and employers may also find it more attractive than the present system.

This idea is motivated by several factors. First, the current retirement system depends critically on employer-sponsored plans, and that dependence can create problems. For example, most payroll deduction retirement savings plans are only offered to workers through their employers. This effectively leaves out independent contractors, gig workers, most part-time employees, and many others. If the employer chooses not to offer such a plan, even the firm's full-time workers are denied access to this effective retirement savings method. They have the option to open an IRA, but few will contribute regularly in the absence of payroll deduction.

Second, under the current system, job changes can create problems for retirement saving, and contingent workers are likely to have even more job changes than traditional employees. In many cases, when an employee leaves a job with a retirement savings account, the funds are either rolled into an IRA, which may have significantly higher fees than the 401(k) plan the worker just left, or the funds are cashed out.8 Research suggests that there are significant amounts of pre-retirement withdrawals (Argento et al., 2015). In an increasingly mobile and diverse workforce, these results are likely to be sub-optimal for workers.9

Third, it is likely that in the future an increasing number of individuals will spend time as both an employee and as an independent/contingent worker, and their retirement arrangements should function effectively for both types of employment status. For this reason, an effective reform must cover both contingent workers and traditional employees.

A sweeping way to address all of these issues would be to attach the retirement account to the worker rather than depending on the employer to offer a retirement benefit. In an employer-facilitated model – as opposed to the current employer-sponsored model – every person would have a retirement savings account that travels with him or her from job to job. Employers, including those that utilize contingent workers, would be required to provide workers the ability to make payroll deduction contributions to their retirement accounts (and to have state and federal tax withheld). This would apply to full- and part-time employees as well as contingent workers of all types. Employers would not be required to offer their own retirement account, although they could decide to do so to attract and keep the kind of workforce they desire just as they do now.10 Under the proposal, an employer that offered a retirement plan with matching contributions to its permanent employees would have to offer the same arrangement to any contingent workers it also employs.

This idea is a departure from the current system, which relies on employer-sponsored accounts. At the same time, the idea of an account that travels with the worker from job to job, and which is able to receive employer contributions, should be quite familiar, since Social Security has those features.11 Others have proposed similar models, such as the Universal Retirement Security Account (URSA) as a way to reduce costs and simplify the retirement savings paradigm (Friedman, 2015). In addition, some, including the Aspen Institute's Future of Work Initiative (2016), have called for a universal, portable, and pro-rated system that works with – rather than against – the frequent job and careers changes workers make today.12

Under an employer-facilitated account, at the time of hire, the worker would supply a Social Security number for tax purposes, a bank account number for direct deposit of the paycheck, and a retirement account number. The worker would also provide the new employer with a form indicating how large the desired ‘employee’ contribution would be with the default choice of keeping the contribution at the same level as it was in the past.13

Workers with volatile incomes may not want to commit to a retirement contribution amount in advance. To remedy this concern, these workers could be offered the option to connect to the retirement account through a savings app such as offered by Digit or a similar provider, as discussed earlier. Those with irregular earnings could have the option to save for retirement by committing to contribute a set amount each month, an amount conditional on earnings, or on a discretionary basis rather than through automatic payroll deduction.14

An open question is to what extent the account should change to conform to the plan in place at a worker's firm.

Under one approach (the ‘fixed approach’), the employee's account would remain fixed in form and contribution limit as he or she changes jobs. In a second approach (the ‘adjustable approach’), the account could be adjusted so that when the worker is employed by a firm with a 401(k) plan, the account would be subject to all 401(k) rules, with the same contribution limit, withdrawal rules, etc. that apply to that type of account along with whatever employer match that the company offers. As with the first approach, when the employee moves to a company that does not offer a 401(k), the account would become a payroll deduction IRA with all of the rules currently attached to that type of account.

Under a third approach (the ‘intermediate approach’), the employee's account would adjust the contribution limits and ability to receive an employer match to conform with the plan the employer offers, but otherwise, all employee accounts would have uniform rules for distributions, withdrawal options, etc. In this approach, the account would be adjusted so that when the worker is employed by a firm with a 401(k) plan, it would be subject to 401(k) contribution limits along with whatever employer match that company offers.15 When a firm that does not offer another type of retirement account employs the worker, the retirement account would be a payroll deduction IRA with the same contribution limits that apply to that type of account.

The fixed approach would be the simplest option for workers, but it could be disruptive to the existing retirement savings system as it would create one uniform contribution limit and enable these accounts to receive an employer contribution even when the account is not part of a 401(k)-type plan. This approach may also discourage some employers from deciding to offer a 401(k) plan that offers additional features to a payroll deduction IRA. Indeed, it may encourage some companies that now offer a 401(k) to drop that plan in favor of the simpler employer-facilitated alternative.

Relative to the fixed approach, the adjustable approach is less disruptive to the existing retirement savings system and might provide more encouragement to employers to offer a 401(k)-type plan with employer matching contributions. Most employees contribute less than the $5,500 annual IRA contribution limits, but the higher contribution limits for 401(k) plans make them more attractive to senior management. The higher limits for a 401(k)-type account would encourage employers to start such a plan. In order to take advantage of those higher limits, 401(k) anti-discrimination tests require that the plan is offered and utilized by lower paid employees also.16

The intermediate approach could be both less disruptive to the current retirement savings system than the first approach and easier for both employers and employees to administer than the second approach. We expect that this version would be most acceptable to both employers and employees.

Under any approach, contingent workers and regular employees who prefer to have a 401(k) plan would continue to be able to seek jobs in companies that offer them but would be able to avoid a break in employee contributions if they moved to a job without an employer plan. In addition, all types of workers would be able to avoid the hassle of combining different accounts of potentially different types from a number of employers over the course of a career. Workers will have the benefit of having one account throughout their careers without accidental leakage or periods without contributions when firms without a plan employ them.

Employers would have the flexibility to offer whatever type of retirement plan they believe is necessary to attract and retain the type of employee they need to be competitive. Each approach also retains the private sector nature of the retirement savings system. Private account management firms would handle employees’ accounts, and private entities would continue to be able to offer different types of plans to employers. However, the two functions would be separate with account management firms only serving individual savers, while plan managers would serve only employers. This would enhance consumer protections by avoiding potential conflicts of interest. While a change from the current system, the new structure is also likely to eliminate certain hidden charges and subsidies that are often present in today's combined system.17 Otherwise, transitioning to the new system would be fairly simple.18

Workers could either choose an investment management firm when they start their career or more likely, be assigned a provider whose accounts meet certain specifications at random through an automatic enrollment arrangement using either a national or state mechanism.19 In order to ensure that savers are not charged high fees or placed in inappropriate investments, the initial investment choice would be required to meet certain criteria such as low overall fees and being invested in low-cost index funds, probably as part of a target date fund.20 The worker would always be free to change investments or investment management firms, but in order to promote safe decision-making, they would be encouraged to stay in low-cost, buy-and-hold strategies. Enhanced consumer protections would help keep workers’ nest eggs safe for the duration of their careers. The Consumer Financial Protection Bureau and US Treasury Department would be natural candidates along with the Department of Labor to serve as regulators, ensuring respectively that savers have a reasonable menu of portfolios to choose from and that account management carries out its fiduciary responsibility (Friedman, 2015).

We believe that an employer-facilitated account structure has both the flexibility to serve both traditional employees and contingent workers and would provide them with better retirement security than the existing fragmented structure where a saver has a different account from each employer. Ultimately, what concerns a retiree is how much money he or she has available for retirement rather than the structure of the account where it is located. A single account that travels with the worker from job to job regardless of the employment relationship reduces that chance that accounts and the money they contain will be lost or cashed out when jobs end. It also eliminates the complex and confusing rules that govern rollovers. If that account can accept contributions from both 401(k) plans and payroll deduction IRAs while retaining the same underlying investments, it is likely that the total will grow faster than if the money goes into individual, smaller accounts at each employer.

V. Conclusion

As the American workforce continues to age, it will be even more important that everyone has the opportunity to build a retirement nest egg to supplement Social Security benefits. Currently, less than half of Americans are enrolled in workplace retirement savings or pension plan. There are a number of efforts to expand this proportion, but almost all of them focus solely on workers who are classified as employees. Efforts to help other types of workers save for retirement are not as common.

As the number of contingent workers continues to grow, making retirement saving easier and more financially rewarding for these workers, who are largely left out of the current retirement system, becomes ever more compelling.

Our examination of the problems in the current system and the characteristics of growing contingent employment provide an impetus to consider decoupling retirement savings plans from employers. The reform we discuss would enable every worker to have one retirement plan that they could carry from employer to employer, just like their Social Security account. It would still enable employers to select the type of retirement plan they need to attract and retain good quality employees, would reduce leakage when workers change jobs and would preserve the private sector nature of our retirement system. This reform has the added advantage of helping traditional employees to preserve their retirement savings. As noted, we believe that, in the future, many people will be contingent workers and traditional employees at different points during their working lives. Thus an effective reform must assist both roles as well as the transition between such employment states.

As more researchers focus on these issues and with continuing rapid advances in technology, we anticipate that additional solutions will be developed. For example, Harris and Krueger's (2015) proposal to create a new category of ‘independent workers’ could pave the way towards improved retirement security for certain contingent workers. Similarly, Choitz and Conway (2015) offer a series of policies that include using state-sponsored retirement savings plans to improve contingent and lower-income workers’ retirement outcomes. In any case, as the number of contingent workers grows and the average American lifespan extends, it is essential to begin to address these issues quickly.

1 Available data on the use of IRAs do not differentiate between the self-employed and employees of firms that do not offer a retirement plan.

2 For more discussion of issues facing gig workers, see Warner (2015). In addition to our proposals, policymakers should consider a comprehensive financial literacy campaign about the need to save for retirement regardless of one's employment status, encourage wider use of automatic enrollment and automatic escalation of contributions, and consider raising Social Security benefits for lower-income workers who are genuinely unable to save.

3 JPMorgan Chase & Co. Institute (2016b) estimates that the gig economy contained 2.5 million workers in September 2015. This is probably an underestimate because the calculation is based only on Chase bank account holders who earned income from among 30 different gig platforms. Chase does not offer a truly free checking account and so may miss lower income households, and the list of 30 platforms is likely to be non-exhaustive. QuickBooks (2016) estimates there were 3.2 million gig economy workers in 2015. This is probably an overestimate, since it is based on an on-line survey, and both respondents and gig economy workers are likely to be more tech-savvy than average. Manyika et al. (2015) and Katz and Krueger (2016) obtain smaller numbers than the other studies, but they count only labor platform workers, not capital platform workers.

4 Madrian and Shea (2001); Benartzi and Thaler (2007).

5 And unfortunately, even those benefits may be lower for contingent workers than they are for traditional employees due to low tax compliance and a variety of other factors. Schreur and Veghte (2018):

6 In order to be listed on Washington State's Marketplace, both a plan and its provider must meet several standards including having fees below 100 basis points.

7 Benjamin (2003), Engen and Gale (2000).

8 The Department of Labor's now defunct fiduciary rule may have had some effect on the fees charged to IRA owners, but since its demise, it is uncertain if those changes will remain. In addition, rollovers from a 401(k) plan are likely to remain higher than those charged by the plan. An important distinction is that a payroll deduction IRA will not have the same fee structure as a retail IRA as this is a group product with a simple structure and index fund investments.

9 According to Munnell and Webb (2015), a leakage occurs when funds are permanently withdrawn from a 401(k) or IRA before retirement.

10 Implementing this proposal would require that the IRS definition of ‘employee’ change so that providing access to a retirement savings account, regardless of type, would not in itself make a contingent worker become an employee. The effect of these changes could be significant, but are outside of the scope of this paper.

11 Of course, there are also many differences between Social Security, a public, defined-benefit, pay-as-you-go system, and the idea considered here, which is for private, defined-contribution, fully funded accounts.

12 The concept of pro-rating in this context is simple: It means that the amount workers contribute to their retirement is determined by the number of hours they work, dollars they earn, or some combination thereof (Foster et al. 2016). See also Harkin (2014).

13 This form from the past employer would include both the routing number and the level of past contributions. That way the level of contributions would be at least the same from employer to employer. We believe that employee contributions should be significantly greater than the 3% level often found in many of today's automatic enrollment plans and encourage employers with a 401(k)-type plan to use automatic escalation to increase the initial employee contribution level to a higher amount. At the same time, we recognize that circumstances may require the saver to choose a lower contribution rate or even to stop saving. However, this should be the exception, and the notice should be structured accordingly. We use the term ‘employee contributions’ here to mean contributions from any worker's pay.

14 In such a case, the employee's retirement account would be directly connected to his or her checking account, and this employer would not be forwarding retirement saving. However, if the employee later has a job that does have a steady income stream, the contributions would come through the employer's payroll system.

15 Legally, a 401(k)-type plan would continue to be defined as an employer-sponsored plan since it is being offered at the employer's discretion. As is the case today, an employer is not required to offer such a plan. A key difference would be that employers would no longer be responsible for investment choices since they would be chosen by the employee and would apply to his or her account regardless of where they are employed. However, anti-discrimination tests would still be applied and would be expanded to cover both of those who qualify as employees and those who are contingent workers.

16 However, those accounts will always have the same type of tax treatment as the worker moves from employer to employer. The individual would have the opportunity to choose which type of tax treatment when they are first employed or be automatically enrolled in one type of account of the other. If the employee is in a traditional account where contributions are deducted from taxable income, he or she could move to a Roth by paying back the tax preference just as they can today.

17 This change has the advantage that the account manager would be working directly for the saver – the ultimate customer – rather than indirectly through the saver's employer as in the current system. In the process, the responsibilities and administrative burdens for employers that offer a 401(k)-type account would be sharply reduced. Savers could be protected against predatory fees and inappropriate investment choices through safe harbors and appropriate default funds.

18 In the initial stages of the transition, employers that do not offer another retirement savings or pension plan would automatically enroll their existing employees into a payroll deduction IRA, while neither employers that do offer such a plan nor their employees would be affected. However, all employers would also enroll contingent employees, and we would encourage them to re-enroll existing employees who are not participating in the plan. Once all accounts have an individual routing number so that they could accept direct transfers of deposits, the new system could go fully into effect.

19 New savers could be placed in state-sponsored retirement savings plans available in the state in which they reside or a national mechanism that places savers randomly with a provider who meets certain specific standards as to fees and investment choices. New Zealand's KiwiSaver plan uses such a system. Of course, savers would have the ability to choose their own provider or to change providers at any time.

20 Of course, upper income and high-balance retirement savers could always choose to have a more sophisticated or actively managed investment choice. Our reform would allow investment firms that prefer these savers to serve them only rather than having to accept a mixture of balances as they do through the current system. However, they would have to meet suitability standards to ensure that higher cost investments are not sold to savers who will not benefit from them. We anticipate that investment options would be commodity-like standardized funds that have all the same basic elements. One option to keep costs low would be to use professionally managed pooled funds. Savers would retain the option to have more actively managed investments.


Retirement Security Project, Brookings Institution. We thank Alec Camhi for research assistance and Debbie Chalfie, Derek Dorn, Jason Fichtner, John Friedman, Hilary Gelfond, Joshua Gotbaum, Catherine Harvey, Victoria Johnson, Aaron Krupkin, Robert Toth, and Christian Weller for helpful comments. We thank the Laura and John Arnold Foundation for funding this work. The findings and conclusions are the responsibility of the authors and do not necessarily reflect positions or policies of the Brookings Institution, its funders, or any other organization.


Argento, R, Bryant, VL and Sabelhaus, J (2015) Early withdrawals from retirement accounts during the great recession. Contemporary Economic Policy 33, 116.
Benartzi, S and Thaler, RH (2007) Heuristics and biases in retirement savings behavior. Journal of Economic Perspectives 21, 81104.
Benjamin, DJ (2003) Does 401(k) eligibility increase saving? Evidence from propensity score subclassification. Journal of Public Economics 87, 12591290.
Bradford, H (2018) Efforts to help gig economy workers save for retirement gaining ground. Pensions & Investments (July 9).
Brown, JE and John, DC (2017) Improving the Saver's Credit for Low- and Moderate-Income Workers, AARP (September).
Chetty, R, Friedman, JN, Leth-Petersen, S, Nielsen, TH and Olson, T (2013) Active v. Passive Decisions and Crowd-Out in Retirement Savings Accounts: Evidence from Denmark, (December). Available at (Accessed 14 August 2016).
Choitz, V and Conway, M (2015) The Future of Work for Low-Income Workers and Families. The Working Poor Families Project (Fall/Winter). Available at (Accessed 29 June 2016).
Copeland, C (2018) Individual Retirement Account Balances, Contributions, Withdrawals, and Asset Allocation Longitudinal Results 2010–2015: The EBRI IRA Database, Employee Benefit Research Institute (January 10).
Engen EM and Gale WG (2000) The Effects of 401(k) Plans on Household Wealth: Differences Across Earnings Groups. National Bureau of Economic Research Working Paper 8032.
Fitchard, K (2013) Braintree launches Marketplace, a payments system for collaborative consumption apps, GigaOM (August 14).
Foster, N, Nelson, G and Reder, L (2016) Portable Benefits Resource Guide, The Aspen Institute (July 13).
Friedman, JN (2015) Building on What Works: A Proposal to Modernize Retirement Savings, The Hamilton Project (June 23).
Future of Work Initiative (2016) Toward a New Capitalism: A Policy Agenda to Restore the Promise of Work, Aspen Institute.
Gale, WG and John, DC (2017) State Sponsored Retirement Savings Plans: New Approaches to Boost Retirement Plan Coverage, PRC WP2017, Pension Research Council Working Paper. Available at (Accessed 13 July 2018).
Gruber, J, Orszag, PR and Gale, WG (2006) Improving Opportunities and Incentives for Saving by Middle- and Low-Income Households, Brookings (April 1).
Harkin, T (2014) HELP Chairman Tom Harkin Introduces the USA Retirement Funds Act – Bill Summary. United States Senate: Health, Education, Labor and Pensions Committee.
Harris, SD and Krueger, AB (2015) A Proposal for Modernizing Labor Laws for Twenty-First Century Work: The ‘Independent Worker’, the Hamilton Project, the Brookings Institution (Dec 7).
Honest Dollar (2018) How Do I Get Started with Honest Dollar? Available at on July 13, 2018.
Internal Revenue Service (2014) Plan Now to Get Full Benefit of Saver's Credit; Tax Credit Helps Low- and Moderate-Income Workers Save for Retirement (Nov 3).
Internal Revenue Service (2015 a) Retirement Topics – IRA Contribution Limits. Available at,-Employee/Retirement-Topics-IRA-Contribution-Limits (Accessed 1 April 2016).
Internal Revenue Service (2015 b) IRA Deduction Limits. Available at (Accessed 1 April 2016).
Internal Revenue Service (2015c) SEP Plan FAQs – Contributions. Available at (Accessed 1 April 2016).
Internal Revenue Service (2018) Table 2. Individual Income and Tax Data, by State and Size of Adjusted Gross Income, Tax Year 2015. Available at (Accessed 11 July 2018).
JPMorgan Chase & Co. Institute (2016 a) Paychecks, Paydays, and the Online Platform Economy (November).
JPMorgan Chase & Co. Institute (2016 b) Paychecks, Paydays, and the Online Platform Economy (February).
Katz, LF and Krueger, AB (2016) The Rise and Nature of Alternative Work Arrangements in the United States, 1995–2015. Available at (Accessed 28 June 2016).
Lacy, L (2016) Micro Businesses Need a Voice of their Own this National Small Business Week, Invoice2go blog, 2 May 2016. Available at (Accessed 6 July 2016).
Lyft (2018) Express Pay – Lyft Help. Available at (Accessed 13 July 2018).
Madrian, BC and Shea, DF (2001) The power of suggestion: inertia in 401(k) participation and savings behavior. The Quarterly Journal of Economics (November).
Manyika, J, Lund, S, Robinson, K, Valentino, J and Dobbs, R (2015) A Labor Market That Works: Connecting Talent with Opportunity in the Digital Age, McKinsey Global Institute (June).
Marcus, S (2016) Christie signs law creating small biz retirement marketplace, (January 19).
Marte, J (2016) Uber is Tackling One of the Biggest Challenges for Workers in the Gig Economy. The Washington Post. (August 24).
Munnell, AH and Webb, A (2015) The Impact of Leakages of 401(k)/IRA Assets, Center for Retirement Research at Boston College, Number 15–2. Available at (Accessed 29 June 2016).
Perez, ST (2016) Innovation and the Contingent Workforce, U.S. Department of Labor Blog. Accessed via on June 11, 2016.
QuickBooks (2016) Dispatches from the New Economy: The On-Demand Workforce. Available at (Accessed 24 March 2016).
Schreur, E and Veghte, BW (2018) Improving Social Security Coverage and Retirement Benefits for Independent Contractors, National Academy of Social Insurance (June).
Storey, D, Steadman, T and Davis, C (2016) Is the gig economy a fleeting fad, or an enduring legacy? Ernst and Young (June).
Supreme Court of California (2018) Dynamex Operations West, Inc., Petitioner, v. The Superior Court of Los Angeles County, Respondent; Charles Lee et al. , Real Parties in Interest. Available at (Accessed 11 July 2018).
Transamerica Center for Retirement Studies (2015) 16th Annual Transamerica Retirement Survey: A Compendium of Findings About American Workers. Available at (Accessed 29 March 2016).
U.S. Bureau of Labor Statistics (2005) Contingent and Alternative Employment Arrangements, February 2005. Available at (Accessed 11 July 2018).
U.S. Bureau of Labor Statistics (2018) Contingent and Alternative Employment Arrangements – May 2017. Available at (Accessed 11 July 2018).
U.S. Department of Labor (2015) Administrator's Interpretation No. 2015-1, (July 15). Available at (Accessed 1 March 2016).
U.S. Department of Labor (2017) US Secretary of Labor Withdraws Joint Employment, Independent Contractor Informal Guidance, (June 7). Available at (Accessed 11 July 2018).
U.S. Government Accountability Office (2015) Contingent Workforce: Size, Characteristics, Earnings, and Benefits, (April 20). Available at (Accessed 1 March 2016).
Warner, MR (2015) Asking the tough questions about the gig economy, Washington Post (June 18).
Washington State, Department of Financial Institutions (2015) Small Business Retirement Marketplace. Available at (Accessed 29 March 2016).
Wiessner, D (2018) GrubHub case could be barometer for new rules on independent contractors, Reuters (May 6).
Zumbrun, J and Sussman, AL (2015) Proof of a ‘Gig Economy’ Revolution is Hard to Find, Wall Street Journal (July 26).