September 9th, 2021 42 Minute Read Issue Brief by Allison Schrager

Wage Stagnation and Its Discontents: Rethinking the Safety Net to Encourage a More Dynamic Economy

No matter how you measure it, income is not rising as fast as it once did for many Americans, and many people feel that their economic situation is precarious and risky. Understandably, policymakers want to find ways to make work pay more and be more predictable. They have proposed various means of doing so, such as a higher minimum wage, more unionization, a jobs guarantee, industrial policy, or universal basic income (UBI). All of these proposals seek, in one way or another, to lessen the amount of economic risk that Americans face.

But upon closer inspection of the data, the problem is not new and excessive economic risk— but rather, the opposite. Many households are falling behind because they aren’t able to take risks necessary to get a piece of the upside that an economy in transition offers. Americans are switching jobs less often than in the past, and fewer new businesses are being created, which contributes to wage stagnation.

At the same time, fundamental changes in the American economy make this an ideal time for a revamp of the safety net.

A more flexible labor force, in which people change jobs and move more, as well as pursue more entrepreneurship or gig work, could increase wages by exposing workers to more upside risk. The best way to make work pay is not more guarantees but more insurance against downside loss so that we can empower more risk-taking.

This can be achieved not only through government policies but via private market mechanisms. The government, for example, could allow workers to average their income over several years when assessing income-tax liability, which smooths out the amount owed, despite transitory income swings. Contingent work platforms, meanwhile, could offer better benefits, such as sick leave or wage insurance, through an insurance model that is mandatory at the firm level, which prevents adverse selection.



Many Americans feel as though they can’t get ahead: the costs of housing, health care, and education keep rising, while their income has not grown much. Income stagnation has become one of the most pressing policy issues. Regarding the degree to which income has stagnated, we often hear the worst possible estimate. But no matter how you measure it, income is not rising as fast as it once did for many Americans. Not only are incomes stagnating relative to our cost of living; there is also a perception that many households are burdened with rising economic risk [1] and live financially precarious lives.

These problems have captured the attention of policymakers who want to find new ways to make work pay more and be more predictable. The usual policy proposals—a higher minimum wage, more unionization, a jobs guarantee, industrial policy, and universal basic income (UBI)—all aim to increase income and reduce risk. We can expect wage stagnation and risk to take on greater urgency in the next few years. Big economic shocks are normally followed by major expansions of the safety net. The Great Depression was followed by the New Deal, and the Great Recession was followed by the Affordable Care Act. If history is any guide, the U.S. government will be tempted to craft new entitlements as we recover from the recession caused by Covid-19.

We are already seeing the child tax credit, as well as infrastructure plans that promise to offer stable union jobs and industrial policy.

Now is an ideal time for a revamp of the safety net. Negative economic shocks often expose the welfare state’s shortcomings that require fixing. Economies evolve, creating new winners and losers and sources of vulnerabilities. Some of the new risks become apparent only after the economic shock occurs. The Great Depression exposed vulnerabilities of a more industrial, urbanized economy in which market shocks were more severe and systematic. Fewer people managed risk within their small local communities; instead, they lived in the relative anonymity of cities, where people were less able and less willing to help neighbors who had fallen on hard times, often because they themselves were experiencing hardship. Industrialization created a greater need for a government-sponsored safety net.

Ideally, we’d rethink the safety net after each shock, adding new entitlements to manage emerging risks and pulling back others that no longer serve Americans in the new economy. But this happens in only one direction: it is far easier to add a new entitlement than it is to take one away. As we debate the future of the safety net, we should be mindful of how permanent it might become and how it can best serve Americans as the economy continues to evolve. And we must fully understand the challenges facing workers, rather than simply accept conventional wisdom.

We must approach problems with a solid intellectual framework that addresses our safety net’s shortcomings. Many Americans are anxious about their economic situation. Upon closer inspection of the data, the problem is not new and excessive risk—but rather, the opposite. Many households are falling behind because they aren’t able to take risks to get a piece of the upside that an economy in transition offers. Our economic institutions have shut many middle- and lower-income households out of the most dynamic parts of the economy. Attempts to further reduce risk from their economic lives will only be counterproductive.

What the Pandemic Revealed About the Economy

Last year’s pandemic revealed the flaws in the current safety net. Many workers and business owners could not work for more than a year. Expanded unemployment benefits and Paycheck Protection Program grants were effective at replacing lost income but did not always reach people when needed. Policy uncertainty around public health–related restrictions introduced more risk into our lives.

Many of these issues were unique to a pandemic that required restrictions on economic activity, which hopefully will not happen again. But those issues also revealed how antiquated our unemployment system is. The system does not usually cover everyone in a modern workforce, such as self-employed and gig workers. [2] It also does not do much for temporary income drops, which bring risk and uncertainty to families. For example, gig, contract, and self-employed workers don’t always lose their jobs—they just get less work. Gig workers often don’t have access to sick leave, workers’ compensation, and the option to buy subsidized health insurance through work. To the extent that the pandemic is accelerating existing economic trends, we can expect more dislocations and long-term unemployment as more employers mechanize and explore remote work. There will probably be less traditional work, featuring more working from home and less connection to individual employers.

The solution to these challenges has been to push policies that some have long favored, even before the crisis. These policies have aimed to increase earnings and ensure steady, well-paid employment for all, through such programs as guaranteed jobs, UBI, and making it harder for firms to hire contractors. In a changing economy that is leaving some workers behind, it’s easy to understand the motivations behind these policies. But such programs are flawed, and better alternatives exist.

Before we alter the safety net by adding new entitlements, we must take stock of how the economy had changed, leading up to the pandemic, in order to understand which new risks actually need to be addressed. The rest of this paper will explain what is true, what is not, and what is exaggerated about wage stagnation; the effects of choices by today’s employees and government policies that favor job stability and security; and how rethinking the safety net could change risk/reward judgments and lead to a more dynamic economy.

Wage Stagnation: Is It True? For Whom? Why?

The desire to make work pay more stems from the perception that, at least before the pandemic, wages aren’t growing as fast as they used to or may not be growing at all—and it often comes down to how inflation is measured. Claims about wage stagnation rely on the CPI-U (Consumer Price Index of Urban Consumers) measure of inflation. Many scholars argue that CPI-U assumes larger inflation than the average household experiences because it does not fully capture what people actually buy, as well as improvements in new products. This makes stagnation seem worse than it is. When the PCE (Personal Consumption Expenditure) measure is used instead, stagnation is less severe; indeed, there is even some wage growth. [3]

“Wage stagnation” can be a misleading term. Most people enter the labor force and experience wage growth over their working years, with the fastest wage increases occurring before they turn 50. Wage estimates of the whole population over time might indicate stagnation but might not capture the fact that most people still get regular raises. Aggregate measure of income over time might instead reflect an older population that normally experiences smaller raises and a changing education composition of the workforce.

A group of economists recently took Social Security [4] records and tracked individual, career-long earnings of people who entered the labor market each year from 1957 to 1983. Among cohorts that entered the labor market between 1957 and 1967, median lifetime wages [5] of men grew about 12%. But for those who started work between 1967 and 1983, lifetime earnings fell 10%. [6] Women had a better experience: their median lifetime wages increased nearly 20% between 1957 and 1967; and 32% from 1967 to 1983. However, an important proviso is that these estimates do not capture the increase of workplace benefits (such as health insurance and paid leave) that have become a larger share of workers’ compensation. When benefits are included, there is less of a male decline, although it is true that wages are still not growing as much as they used to. The stagnation among men, especially less educated men, is caused by a smaller starting wage and slower growth after they begin their careers.

Social Security records are the gold standard by which to measure income over time. Unlike the panel surveys that economists normally use, Social Security records are based on payroll data and there is no attrition, which makes them more reliable. But there are some weaknesses. Labor economists normally measure lifetime income for different levels of education. It is important to control for education because college-educated workers tend to experience much faster wage growth. Social Security records don’t include education data, which may contribute to some of the stagnation. Earlier cohorts experienced huge gains in educational attainment in the 1950s and 1960s. By the 1970s, this trend had leveled off, and educational attainment remained stagnant for a few decades. This may explain some of the stagnation. But without education data, it is impossible to know how much of the stagnation comes from a leveling-off of educational attainment and how much comes from other structural changes in the economy.

To discern the impact of education, I took data from the Current Population Survey Merged Outgoing Rotation Group and estimated median income by age, gender, and education (Table 1). Importantly, these data do not reflect lifetime income; rather, they merely give a snapshot of income at different ages over time. They tell us, for example, how much that people aged 25–34 were paid in 1989, compared with 2019, depending on level of education and gender.

Among women of most education levels and age groups, earnings are higher in 2019, regardless of whether the PCE or CPI-U measure of inflation is used. There are signs of slowing gains: the difference in earnings between 1989 and 2019 are larger for older women and negative for young, college-educated women. This is consistent with the Social Security study, which estimates that women joining the labor force and staying in even after they had children produced huge wage gains. But as this became the norm for the next generation, those gains slowed. The long-term impact of the pandemic may worsen or reverse these gains because long-term school closures forced some women out of the labor force.

Men have had a harder time. The wages of men without a college degree are lower or barely increased for every age group, using both measures of inflation. Social Security data reveal lower starting wages and smaller increases after they are in the labor force, which is why older workers have comparatively smaller wages, too. We see the generational impact: the oldest groups have smaller wage decreases, consistent with the Social Security study, which found that the earlier cohort has larger gains.

The story for college-educated men depends on how inflation is measured. With CPI-U, all but the oldest group of college-educated workers saw a drop. With PCE, college-educated men have mostly seen wage gains. However, the gains become smaller as the workers age, which suggests that even if starting wages are higher, the wage increases during their career may have become smaller. If these trends continue, even college-educated workers may start to see wage stagnation— or, at least, slower growth than in previous generations. The population of college-educated workers has grown, so one would expect a greater diversity of economic outcomes from higher education. The population of high school graduates also changed over time. It became smaller as more people went to college, so the population from the past may not be directly comparable with the population today.

The reports of wage stagnation are often overstated; [7] real wages are growing for most people over the course of their careers. But they are not growing as fast as they used to—and, for some populations, especially uneducated men, wages have indeed shrunk. This may be due to structural factors in the economy, such as more technology and trade, which reduce the value of less skilled labor—one reason that we see more productivity gains going to capital than labor.

Another popular narrative posits that wages have stagnated because fewer employers are able to use their market power to act as monopsonists and push down wages while capturing profits for owners. [8] This explanation is unsatisfying partly because U.S. workers are more likely to work for one of a few large firms—but at a local level, competition for labor has increased. [9] For example, 30 years ago, the only possible job for some workers might have been at a local hardware store; now, a Home Depot or Lowe’s is more likely to be hiring in the area and hires more people. So while we see more concentrated employment on the national level, we see less on a local level, where wages are set. The monopsony explanation does not take into account differences in productivity and how that affects wages. In many industries, a handful of firms are more productive than the rest and pay their workers (of all skill levels) more than the rest of the industry. These same firms, because of their high productivity, also tend to be larger than less productive firms, which tend to pay workers less. If wages were depressed because of market power, one would expect the larger firms that have become dominant employers to be underpaying workers.

Slowed wage growth has coincided with less wage variability [10] and a decline in the rate at which Americans change jobs, which are traditionally drivers of wage growth. [11] Entrepreneurship, which plays a role in wage growth, is also declining. [12]

Americans make fewer changes and take fewer risks, which, unsurprisingly, is reflected in slower wage growth. Like any asset, income goes up more when there is more risk that produces variability. There are benefits to more certainty but also costs—contributing to slowing wage growth.

Less Dynamism Causes Stagnation

The conventional wisdom is that Americans face unprecedented risk in the labor market. Many point to the increasing unpredictability of income from year to year. [13] Americans also worry about job security, with survey data revealing that most people think that job stability has not changed since the 1980s. In fact, jobs have become more stable, [14] and many workers overestimate how unstable their jobs are. [15]

Perhaps as a result of this anxiety, workers are changing jobs less frequently, [16] which is reflected in measures of medium-term tenure—the share of workers in a job for five or more years— which has stagnated or even increased in the last 40 years. [17] Social Security data reveal that fewer people switch jobs in a three-year period: in 1980, 50% of people had been at the same job for three years or longer; by 2010, this figure had increased to 60%. Short-term employment has become less common. In the 1980s, nearly 25.7% of workers had been at their job less than one year; in 2018, only 20.5% had. [18] Today, when you find a job, you’ll likely have it for several years.

Not only are jobs more stable; wage volatility (the normal variance of wages year to year) has been on the decline since the 1980s, i.e., wages are more predictable and stable than they were in the 1980s—except in recessions. During economic downturns, very high and very low earners tend to experience larger and more prolonged negative wage shocks when they lose their jobs; it also takes longer for them to find another job. [19]

Further research using Social Security records finds that the decline in wage variability is explained by the reduced creation of new businesses and jobs, as well as the reduced frequency of job changes. [20] Notably, positive as well as negative income shocks have been declining.

These trends—less job switching, less variability, and wage stagnation—lead to more stable wages but also less wage growth. The trends also may be contributing to inequality. Wage variability declined between 1985 and 2012 for all income groups, except the top 5%. Their income variability is much higher than in other groups and was unchanged between 1985 and 2012. Higher wage variability in this group could largely explain why they have seen larger wage increases.

Several factors are driving the decline in dynamism and the rise in stability. For example, more people work for larger firms, which tend to be more stable. [21] Women are remaining in the labor force, even after they have families, at a higher rate. But these factors can’t explain all the decline in wage growth. Structural changes to the economy may have reduced the gains to risk. Alternatively, the population may have become more risk-averse.

What About Gig Work?

The lack of dynamism may reflect institutional frictions that are holding back growth in a changing economy. People are often surprised to hear that job switching has decreased because of the widespread perception that workers have become fickle and that gig work is becoming the norm.

Actually, gig and contract work has not taken off as much as we might expect. Economists Larry Katz and Alan Krueger estimated that, after accounting for a booming economy, the share of Americans engaged in gig or contract work as their main job increased only 1% between 2000 and 2017. [22] Census data that included self-employed workers who don’t employ others show an increase from about 11% to 15% between 1996 and 2011, and the share of workers with 1099 income increased from about 12.5% to 15% between 2001 and 2011. [23]

Traditional survey data, of the kind that Katz and Krueger used, may not capture the changing dynamics of the labor market because how we measure and consider contingent work is not so straightforward. For example, people in a stable, predictable contract job may consider themselves employees. But tax data are problematic because it is easy to underreport gig work. Katz and Krueger and other economists [24] have pointed out that many workers are unsure how to classify themselves, and we may be underestimating how much work in the economy is contingent. Katz and Krueger observed that more Americans are doing contract and gig work part-time, in order to supplement more traditional income. Indeed, surveys from the Federal Reserve [25] show that the plurality of gig workers do their work as a complement to another job.

That relatively few people actually participate in the gig economy is surprising, not only because it departs from conventional wisdom but because the economy has changed in a structural way that should make nontraditional employment—as well as job switching—more appealing. Work has become more uniform across companies. Everyone uses the same word-processing software, spreadsheets, and so on. Manufacturing jobs have become more mechanized, which means that the work is more similar, no matter where it is done. [26] Earlier, when each firm used different tools or had specific processes—say, a unique way of writing documents, building cars, or internal communication—there was a benefit in building up firm-specific capital and in mastering skills unique to that job. But now, individual capital—skills unique to each person, which can be taken to different employers—has become more valuable.

Today, contract and contingent work, especially for highly skilled workers, tends to pay more than traditional employment, and workers in nontraditional employment report higher levels of satisfaction and that they prefer the flexibility. [27]

Again, some suggest that despite these structural changes, job switching is now less likely to lead to wage increases because of the increased market power of firms. [28] But this explanation is no more satisfying than it was for slowed wage growth. There is still scope for job switching at the local level, where competition for labor has increased.

Benefits such as health care likely make nontraditional employment and job changing too risky for many people. Health care has become a more valuable part of compensation, which ties people to jobs even when switching might mean higher pay and better skills development. Americans have become more reluctant to move for better opportunities, often for cultural reasons, where moving away from friends and family is less common or even discouraged, or because housing in high-growth cities tends to be very expensive. Excessive licensing requirements can make it hard to move across states or start a business that offers services. Many professions require extensive retraining or fees to work in a different state. Non-compete agreements [29] have become more prevalent, even for low-wage workers, which makes job changing or starting a business more costly because some workers must endure a period of nonemployment or restrictions on starting a business. Americans may be intolerant of wage variability because they don’t have adequate insurance to deal with wage shocks.

Risk, Reward, and a Dynamic Economy

For many Americans, wage stagnation stems from stagnation in their economic lives, which produces less risk but also less upside. To break the cycle of stagnation, we would encourage people to take more risks—to look for a new job or even start their own business—in order to reap the returns. But perhaps decreased risk for increased stability is a trade-off that most people will accept. Therefore, before considering how to address this issue through policy, we should ask three questions.

1. Is decreased risk appetite a problem we need to solve?

Perhaps wage stagnation is an acceptable cost for more stability, which has value for workers who may prefer more predictable income. Some economists speculate [30] that firms are more profitable while wages have stagnated because there is more profit variability and firms are keeping the upside while providing their workers with more predictability, which they may prefer.

Reduced risk is a sign of progress that comes from greater wealth and technology. However, while certainty has value, it also means less growth in the future, since innovation and risk-taking are inextricably linked.

It is tempting to want to remove risk from the economy, and eliminating the costs of catastrophic risks or extreme poverty is a laudable goal. Because more vulnerable workers tend to have less wealth and are more prone to job loss, it’s understandable that we would want to give them more security. But removing risk entirely—especially for low earners and the middle class—should never be our objective. Rather, we should encourage more risk-taking.

In the past, innovation and income growth came from entrepreneurship and experimentation at all income levels. Stagnation from lack of risk-taking may be contributing to discontent and populism because people feel stuck in economic inertia, unable to get ahead, while risk-taking has become exclusively the domain of the elite. There is a dignity in some risk; it is how we realize our full potential.

2. Are we doomed to live in a low stagnation or low wage-growth future?

Wage stagnation is a choice. We have created institutions that reduce risk and that have kept workers from experiencing upside potential. We should restore the dynamism that once was central to the U.S. economy and not exclusive to high earners—which requires introducing more risk into workers’ lives. Our institutions can be reformed to encourage risk, as they did in the past.

The opportunity today is to remove the prevailing institutions that favor a long-term, low-turnover labor force, and to encourage—rather than crowd out—risk-taking. If some Americans prefer the stability of traditional employment relationships, predictable wages, and long-term employment, we should expect stagnation to continue for them. But for others, we can remove the constraints that prevent them from taking risks and empower them to take advantage of a changing economy so that they can experience more wage growth.

We should desire an economy that allows the choice of risk versus stability and that does not tip the scale too far toward stability.

3. Will increased risk make already-struggling households more vulnerable?

More risk might seem like a counterintuitive solution, especially for low-income and middle-class Americans who are struggling and falling behind. But we can encourage more risk-taking in a responsible way that produces more upside and protects people from catastrophic downsides, which would require revisiting what our safety net aims to achieve. We should protect people from risks that have gone wrong, or simply from bad luck, while preserving the incentives to take healthy risks.

The evolution of our safety net reflects a belief that only well-off Americans can afford to take risk. Our policies around work (health care, unions) and wealth accumulation (retirement-account investing defaults in government plans) all steer middle and low earners to less risky plans dominated by bond investments, which is not helping them. People at all income levels deserve the chance to take risk and experience more growth.

As noted above, economic and technological changes have made it easier and cheaper to take risks. Finding contract and gig work is easier than ever because of apps that match buyers and sellers; working for oneself no longer requires high start-up costs or an extensive professional network. Gig and contingent work can be a form of entry-level entrepreneurship that allows people to experiment and learn skills while they work more traditional jobs. Many people picture all entrepreneurs as wealthy and high-impact—for example, the tech founders of Silicon Valley; but traditionally, Americans of all income levels have been entrepreneurs. Gig and contract work—which includes much more than driving for Uber or delivering for DoorDash— might serve as an entry point for many future entrepreneurs.

Rethinking the Safety Net: Insurance vs. Guarantees

Many Americans desire more security but don’t necessarily need more safety. Safety ensures the same outcome, no matter what happens. Security offers protection if things go badly. We should strive to offer more security instead of safety, so that everyone feels comfortable with taking more risk.

Any new entitlement should offer insurance, or state-contingent payouts, instead of a guarantee. Instead, many recent policy proposals—such as UBI, industrial policy to bring back steady manufacturing work, empowering unions, child allowances, and guaranteed jobs—would add more government protection, with the aim of eliminating risk altogether by offering guarantees. These proposals are extremely different in character from previous successful income-boosting entitlements and mark a shift in our thinking.

Our existing entitlements that aim to increase wages—for example, unemployment benefits or the Earned Income Tax Credit (EITC)—pay off only if an event happens: the worker loses his job, or his income falls below a certain amount. A UBI or a job guarantee would pay no matter the state of the economy or personal circumstance. Each of these policies is being framed as a form of entitlement. But they are distinct from traditional entitlements, which aim to provide some form of insurance. These new benefits are guarantees that pay, no matter the state of the world—which poses many problems.

A guarantee is much more expensive than insurance because it must be paid in all states of the economy (recession or boom, job loss or pay raise) and be available to all people, no matter their economic circumstances. Guarantees also create very different incentives. Proponents of UBI argue that our income-based insurance programs reduce the incentives to work. For example, the more income you earn, the lower your health-insurance subsidy. Some programs are poorly structured and introduce high marginal tax rates. But well-designed programs such as EITC subsidize income from work and phase out in a way that does not reduce the incentive to work. [31] In fact, these programs encourage work. [32]

Guarantees influence behavior. A common perception about UBI proposals and other guarantees is that if you give someone a solid foundation of income, he will take more risk. [33] But no evidence supports that theory; [34] in fact, people tend to take more risk when they face fear of loss. [35]

We can see evidence of this dynamic from past policies that offered guaranteed jobs, such as the Works Progress Administration during the Great Depression, which discouraged risk-taking or finding private-sector employment. [36] In contrast, offering more insurance increases people’s propensity to take risk because it ensures a payoff even when a risk does not go well. For example, you might be more inclined to move to a new city and start a new job if you know that you have a backup plan—say, that you could return to your old job and way of life if the new plan does not work out. [37]

Similarly, in financial markets, where risk and reward are the native language, we know that increasing risk-free return discourages risk-taking; i.e., the higher the risk-free return, the less risk investors must take to achieve their goals. We see this with policies on individual behavior. Financial derivatives (insurance contracts on financial assets) were also used to increase risk and produced more upside in financial markets. [38] The result was more growth—not only in rich countries but especially in developing and emerging markets, which, before the proliferation of financial derivatives, lacked access to global capital at low prices. The downside of cheaper capital and more integration was more systemic risk and several financial crises. But crises are not inevitable with better regulation. Undoubtedly, even with increased financial instability, there was tremendous benefit in lifting many people out of poverty and toward higher living standards. Risk is the cost for more growth and rising prosperity.

Insurance can create moral hazard or risk-taking without bearing the full costs of risk gone wrong. But any policy will change incentives. If the goal is more insured risk-taking, some moral hazard, especially for lower- and middle-income households that are currently stagnating, may not be a bad thing. We are subsidizing their risk-taking, which could be cheaper and more effective in increasing living standards than subsidizing a certain level of income.

Insurance is cheaper and creates better incentives for taking risk. Efficiently designed insurance programs reduce risk at a low cost through the power of diversification. Consider unemployment insurance: if all people had to save enough in their bank accounts to survive a long spell of unemployment, that would mean less money to spend or invest in more productive assets that could grow the economy. But if all people put a little money into unemployment insurance, they could be paid when they lose work for less cost because everyone contributes, but not everyone loses a job. The insurance industry’s business model depends on diversification, or many people buying insurance and not everyone needing it at the same time. This is why some forms of insurance are better provided by the government. Many people tend to lose jobs during a recession, so it is not viable for the private sector to offer unemployment insurance broadly. There is also a problem of adverse selection: only people likely to lose their jobs may want insurance. [39] The government can diversify across time by issuing debt [40] and require everyone to participate in the program. In such cases, some government insurance is efficient and growth-enhancing. [41]Guarantees, which pay everyone all the time, have no scope for diversification and are not efficient.

Risk also explains why the push to empower unions is misguided. Unions worked by pooling workers of all skills together. Wages and benefits were fairly uniform, which meant that higher-quality workers subsidized lower-skill workers. When individual skills were not so valuable, the stability that unions offered meant that this could be a valuable and efficient arrangement. But it makes less sense now, in a market where individual skills are more valuable, and will worsen stagnation.

A more flexible labor force, in which people change jobs and move more, as well as pursue more entrepreneurship or gig work, could increase wages by exposing workers to more upside risk. But it also means more downside and years of low earnings. The best way to make work pay is not more guarantees but more insurance against income loss so that we can empower more risk-taking.

Offering More Insurance in Private Markets and Through the Welfare State

Changing labor markets, stagnating wages, mounting deficits, and the permanent nature of entitlements all suggest that the best approach to increase pay is more insurance for the downside and empowering workers to take more risk. As the economy continues to evolve, this approach offers more flexibility for workers who want to pursue various types of traditional and nontraditional employment.

Guaranteed jobs or UBI are poorly targeted and do not match the needs of new workers and may even hold them back by offering the sort of guarantees that perpetuate wage stagnation. Instead, the new safety net should offer various programs to smooth out dips in income and offer benefits that are not tied to a single employer, including:

  • Wage insurance—benefits that account for a drop in income, not just a loss of employment
  • Income averaging—tax rates based on income over three or five years, not just a single year, which will make income more stable for workers in variable work arrangements
  • Providing contingent workers the opportunity to receive benefits, such as health care and sick leave, that are not tied to traditional employment

To protect themselves against income risk, Americans have resorted to stagnation. We can provide downside protection in alternate ways—so that Americans can feel more free to switch jobs, try alternative forms of work, or start new companies. The above-mentioned programs are a more cost-effective and efficient way to address the needs of the new labor force than the guarantee-oriented policies that receive more attention. These programs provide options that would provide more robust insurance that can help spur a more dynamic economy. The options are merely a starting point to think more creatively about how to support a changing economy and break the cycle of stagnation.

Better Benefits for Contingent Workers

The U.S. economy was built for people who are employees of a single company for long periods of time. We have built an economy that makes it very difficult for lower- and middle-income Americans to take risk because the forms of insurance that do protect them from the unexpected are often tied to their jobs. Health insurance is generally provided as a benefit by employers, and unemployment insurance and workers’ compensation are tied to employment. Often, it is possible to get subsidized disability and life insurance and access to retirement accounts through one’s employer. Better insurance rates are achieved through diversification, so pooling employees’ risk together is an attractive proposition for an insurance company.

However, Americans who aren’t employees cannot benefit from pooling and diversification, which increases their insurance rates. But there is a way that our institutions could offer the benefits of diversification to Americans who are self-employed, contract workers, or contingent workers.

In a perfect world, we would separate health benefits from employment. One way to do so would be to eliminate the deductibility of employer-sponsored health care and create conditions for a competitive individual market. However, the deduction and workplace benefits are extremely popular, and a large wholesale change in our health-care system is probably politically untenable. But smaller changes can be made to help workers who prefer nontraditional employment. One change, proposed by Seth Harris and Alan Krueger, [42] would be to offer safe-harbor provisions for platforms such as Uber, Instacart, and temp agencies so that they could pool their contingent workers together and offer them the option to buy various forms of high-quality insurance that is normally available only to employees with pretax money. Currently, offering this sort of benefit would be a risk to the platforms because offering the benefits could be interpreted by the courts to mean that the workers are, in fact, employees, which would require these companies not only to offer these benefits to all employees but to incur many associated costs. A safe-harbor provision would offer these platforms the option of extending benefits without incurring that potential liability.

Another change, proposed by Chris Pope [43] of the Manhattan Institute, would be to organically grow the individual market so that workers could have a viable health-insurance option independent from their employer. As of 2019, the federal government has allowed employers to provide pretax funds to their employees in a Health Reimbursement Account, which they can use to buy their own health insurance. Individuals can also buy insurance with pretax money. Currently, this may not be an attractive option because the individual insurance market tends to be expensive and offer spotty coverage, largely because of the adverse selection—only sicker people tend to buy insurance, and it is possible to buy insurance only when a person anticipates large medical costs.

Pope states that we could get less adverse selection (and make insurance on the individual market more attractive) by offering lower premiums to healthy people and allowing them to renew their policies each year, regardless of how their health status changes. Insurance companies could offer discounts on premiums for people who have had their insurance for many years.

Sick Leave

Many contingent workers don’t have access to sick leave. If they or their children are ill, they must miss work and not get paid. For some of these workers, even if they can afford to miss a few days’ income, there is no assurance that they won’t lose their jobs because of missing a few days.

Sick leave could be financed through more short-term insurance. Workers on platforms could pool a fraction of their wages into a fund that would pay out a preset fraction of the average weekly income on days when they are sick. As with sick leave for employees, the number of sick days could be capped. For example, Uber drivers who typically drive 30 hours or more per week could have some of their wages deducted. If the driver or a family member is ill, the platform would pay the employee 75% of his average weekly income for, say, up to five days each year.

Sick leave benefits could be administered by the platform or temp firm. Adverse selection could be a problem, so participation would have to be mandatory within a given firm. However, platforms could choose whether to offer this benefit, and workers could opt in to platforms that offer the benefits that they desire.

Wage Insurance

Unemployment insurance, the backbone of our current safety net, is designed for workers in traditional employment relationships but is poorly suited for the demands of our changing workforce. We should revise our labor safety net to complement, not discourage, economic dynamism. A good place to start would be policies that protect workers from drops in wages. Enabling lowand middle-income workers to take more risk would require offering them more insurance so that they could better handle shocks to their income—perhaps in the form of wage insurance.

Life-cycle economists view lifetime wages as an asset called “human capital.” Young people are rich in human capital and poorer in financial assets because they have a lifetime of earnings ahead. Older people have little human capital but more financial assets. Lifetime income has the same characteristics as financial assets in that its value is sometimes subject to shocks—and, depending on the industry, the shocks are even correlated to financial markets.

Income normally grows over time, rising steeply in an employee’s twenties and thirties, and leveling out a bit in the fifties. The slope and leveling-off tend to depend on education, industry, and gender. As noted above, starting incomes have become lower and the slope of wage growth has become less steep, especially for uneducated men. But for most people, wages are a rising asset that rises faster if one invests in more education.

Like any asset, income is susceptible to shocks—permanent and transitory. A permanent wage shock puts a worker’s wage growth on a higher or lower trajectory for the rest of his career. An example of a permanent wage shock would be losing a job because of trade or technology. If a job no longer exists, the worker would have to take a pay cut in his next job because he would have to learn new skills. Or it could be a big job change or promotion that would mean higher wages for the rest of the worker’s career.

Transitory shocks do not have much impact on wages the next year, or anytime in the future. The shock could be a yearly bonus or commission, or any one-time windfall. For example, if you drive for a ride-share company, the pandemic would be a transitory negative wage shock. It would be a year of low wages—but the wages would probably recover in the next year or two. People who work on commissions or on a bonus structure tend to have more transitory income shocks, but if more people work on contract, their numbers can grow.

Insurance can decrease the costs of income shocks for workers who would like to partake in a more dynamic economy. They are less likely to face permanent wage shocks, since their drops or increases in wages tend to be more transitory.

Transitory Income Shock Insurance

Income can fall because of something idiosyncratic (e.g., a health or caregiving need) or systemic (e.g., a recession that lowers demand for most services in the economy). As the economy moves to nontraditional working relationships, including more gig and contract work and other forms of small-scale self-employment, unemployment insurance will no longer be adequate in cases of systemic economic shock. In such situations, many firms would sooner lay off workers than cut their pay. But contingent work is more flexible. The worker doesn’t necessarily lose his job—he just gets less work. Because pay does not go to zero, the worker might not qualify for unemployment insurance (though partial unemployment benefits exist in some states, under certain conditions). During the pandemic, accommodations were made for drops in wages, and unemployment included gig workers. Whether the changes will be permanent varies by state.

Instead of traditional unemployment insurance, gig, self-employed, and contingent workers need insurance from temporary drops in wages that could be driven by idiosyncratic shocks (unique to the worker) or economy-wide shocks (such as a recession).

Transitory wage insurance could work in several ways. For example, if income were to drop more than 10%, a worker could apply for a subsidy, equal to 50% of his lost wages, that would last until his income rebounds or for a fixed time period, whichever was shorter. The affected worker would have to document that the income drop was due to a shock—say, a caretaking need, health issue, or economic distress in his community. Participation in the program, as with unemployment insurance, would require paying premiums taken from the worker’s wages.

The insurance could be a private-sector or government program. Because shocks are often idiosyncratic to the worker, it is possible for the private sector to offer the insurance. More systemic shocks, such as a recession that reduces demand or a global pandemic, may be more challenging.

However, a pandemic-level event is extremely rare and likely would come with government intervention. So transitory wage insurance could be offered by a private insurance company or the employer as a workplace benefit. The government could offer the same safe-harbor provisions to employers that offer health insurance. It could encourage participation by allowing workers to pay the premiums with pretax money and by regulating the insurance carriers to ensure that they have enough assets to reduce counterparty risk. Wage insurance would be better than saving accounts because insurance offers scope for diversification, which is more efficient and allows workers to get more coverage for less money.

Moral Hazard and Adverse Selection

Insurance can invite moral hazard. Transitory income risk insurance also faces the possibility of abuse, which can be mitigated by limiting the time the insurance will pay out or by a cap on lifetime payments. Claims for payouts would also have to require evidence for the reasons that income has fallen. As stated earlier, if a small amount of moral hazard incites more risk-taking, that is not necessarily a negative thing.

Adverse selection is a larger concern. If workers are not required to pay premiums, it is likely that only people who frequently need the insurance will be willing to pay for it. Thus, wage insurance may not be economically viable if participation is not mandatory. As with sick leave, however, individual platforms could decide to offer insurance and require participation of all those who work 30 or more hours per week. The platform could withhold premiums paid to the insurance company. Participation in the program would be mandatory within a firm that offers it, but not all firms would have to offer it, and workers could decide which firm to work for, depending on their risk preferences. This may lessen adverse selection because an insurance- requiring firm may offer benefits that attract a broader population of workers.

Income Averaging

As we move away from a traditional employment model, more workers will experience more variable income from year to year. Currently, tax bills are based on income in a single year, which can penalize people whose income is high in one year and low the next. For example, suppose you work as a contractor and you earned $50,000 in 2020; your top marginal federal tax rate (not including self-employment tax) would be 22%. If, in 2021, you earn $90,000, your top marginal rate would increase to 24%. But if you’d earned $70,000 in each of those two years, your top marginal rate would be 22% both years.

The current tax code can penalize variable earnings. To address this issue, a provision in the tax code in 1964–86 enabled Americans to pay a tax rate based on the average of their last five years of income. The provision was abandoned with the 1986 tax reform, [44] largely to simplify the tax code. Eliminating loopholes and broadening the base was a major and noble goal of the reform. The provision was regressive: of the 6.1 million tax filers [45] who took advantage of it in 1981, most were higher earners, [46] likely because of a minimum income requirement to qualify. The rule was fairly complicated and involved calculating a five-year average base income, and the tax liability was based on each year’s extra income. Higher earners had a greater incentive as well as the resources to exploit the rule.

Things could be different today. Tax software means that lower-income Americans could take advantage of the program and it could be reinstated without a minimum income, which could offer more income stability and encourage contingent workers to earn more in higher-earning years by reducing marginal tax rates. If the gig- or contingent-worker economy continues to grow, more low- and middle-income earners might take advantage of the provision.


As the economy recovers from the pandemic, we can expect economic trends to accelerate. The existing model of traditional employment that came about during the postwar period may no longer make sense in our modern tech-driven economy. Some 200 years ago, dependence on a single employer as one’s primary risk reducer felt as precarious and unnatural as gig or contingent work feels to many people today. But the ideal labor relationship changes with the economy. Wage stagnation and a decline in wage variability could be symptoms of a changing economy that no longer suits many modern workers and shuts them out from risk-taking and upside potential.

We can restore dynamism and protect workers from hardship by building on our modern safety net, which requires expanding insurance—an effective and efficient way to encourage risk-taking. Instead, we risk doubling down on what’s not working, by pursuing new policies that offer guarantees or pushing workers toward the old model that no longer serves them. These policies will only reinforce stagnation by further eliminating risk.

A better safety net would encourage more dynamic work relationships that come with more risk but also the potential for more reward. But those relationships require different forms of insurance so that workers who take a riskier path have some protection and feel empowered to pursue their full potential—a more efficient, and ultimately rewarding, way to make work pay.


See Endnotes in PDF

About the Author

Allison Schrager is a senior fellow at the Manhattan Institute and a City Journal contributing editor, where her research focuses on public finance, pensions, tax policy, labor markets, and monetary policy. She is also the author An Economist Walks Into a Brothel and co-founder of LifeCycle Finance Partners, LLC, a risk advisory firm.

Previously, Schrager was a journalist at Quartz, led retirement product innovation at Dimensional Fund Advisors, and consulted to international organizations, including the OECD and IMF. She has been a regular contributor to the EconomistReuters, and Bloomberg Businessweek, and her writing has also appeared in PlayboyWiredNational ReviewForeign Affairs, and City Journal.

She received her undergraduate degree from the University of Edinburgh and a Ph.D. in Economics from Columbia University.


Are you interested in supporting the Manhattan Institute’s public-interest research and journalism? As a 501(c)(3) nonprofit, donations in support of MI and its scholars’ work are fully tax-deductible as provided by law (EIN #13-2912529).