Unemployment insurance is one of the few government programs that hardly anybody opposes, at least in principle. After all, what’s not right about providing short-term help to unlucky workers whose jobs have been destroyed through no fault of their own – and in the process, propping up household spending in the economy when it’s most needed. Yet unemployment insurance in America has been in reform purgatory for five decades, slowly deteriorating in effectiveness and integrity until reaching crisis levels.

What went wrong? It started with a program design error, which was there at the outset in 1935 and became more visible over time, slowly but assuredly compromising unemployment insurance. In public programs, this type of flaw is more commonly found in the benefit: the benefit structure has some incentive built in that designers either didn’t see or discounted as unimportant that undermines the program’s support and efficacy. In UI, while the benefit is far from perfect, it is the program’s revenue source that is deeply problematic.

As we’ll see, fixing UI by clearing away the underbrush of inequity and perverse incentives would not be easy even if there was a political consensus about the goals and methods of change. An even more difficult task would involve starting with first principles about who should be insured against labor market risk and how to manage the transition. But it’s worth remembering that our failure to address evolving risk as capitalism itself evolves will carry a substantial and growing price in terms of economic and social cohesion.

The state-run national unemployment system, which was mandated in 1935 as part of the Social Security Act, is funded with taxes. All firms must pay a tax for each employee – a percentage of a capped wage base – that is levied by state legislatures and paid into individual states’ trust funds. The states’ taxes per worker levied on an individual firm increase if enough workers laid off by the firm claim benefits.

Following the Wisconsin model, relief for the unemployed seemed a secondary consideration in the initial design of the system, with the primary goal to deter layoffs. And this reversal of priorities, which required the “experience rating” of taxes linking tax rates to individual firms’ layoffs, is still driving the program and is at the root of many of UI’s problems today.

The significance of this choice wasn’t obvious at first. Unemployment insurance only paid its first benefits in 1938, a rather fortuitous time to start a program funded by a payroll tax since unemployment fell rapidly to near-zero with the ramping up of war production and expansion of the number of workers siphoned off by the military.

Initially, most of the workers covered were in manufacturing. But after the war, with unemployment remaining low and state trust funds fat with payroll revenues, there was little opposition to broadening enrollment to include almost all workers and thereby increasing the states’ financial liabilities in economic downturns. Indeed, the problematic nature of the funding approach only became front and center during the nasty recession of the mid-1970s. But by then there was no appetite among politicians for reforms that would set UI on a course for sustainability.

Inaction in Action

To receive UI coverage, a worker must have (a) lost their job through no fault of their own, (b) worked sufficiently long to qualify and (c) be willing and able to go back to work. The weekly dollar benefit is calculated separately for each worker based on the worker’s average prior earnings. The replacement rate formula varies by state, but typically works out to around 35-45 percent of pay, up to a state maximum benefit. Benefits are limited to a fixed period also set by the state.

In recessions, the pool of workers applying for unemployment benefits spikes as firms lay off workers en masse; their prospects for reemployment depend highly on the broader economic recovery. The 1973-75 recession, the worst since UI was created, drove the state trust funds toward insolvency. And businesses, struggling to keep their doors open, were slammed by higher experience-rated taxes.

Jolted by these realities, Congress created the National Commission on Unemployment Compensation, a bipartisan, business-worker commission to design comprehensive reform. However, the 1980 final report of the commission, which among other things recommended that experience ratings carry less weight in assessing employer taxes, fell on deaf ears. None of the recommendations for reform were adopted.

Plus ça change … the 1990-91 recession highlighted the same funding weaknesses in UI, and to no one’s surprise, another bipartisan commission was created to give Congress the political cover to pass comprehensive reform. Or, maybe just to kick the can down the road: the Advisory Commission on Unemployment Compensation fared no better than its earlier counterpart. The final report was released in 1996, and none of the recommendations were adopted.

By this point, auditors at the Government Accountability Office had begun to take notice. They released their own analysis in 1993 warning that unemployment insurance was failing to meet its basic objectives. States, by the GAO’s reckoning, had been responding to episodic revenue squeezes by reducing wage-replacement rates and restricting eligibility for benefits. But state governments do not answer to Congress’s watchdog, and their incentives to hold down UI costs apparently outweighed incentives to maintain coverage by raising tax rates or broadening the tax base. 

A Well-Served Constituency

The federal government has little leverage over the way UI works in individual states. States design the benefit formula, set the maximum benefit and weeks of duration along with the tax schedule that finances those benefits, and process the claims. If you take as a reasonable starting point that businesses prefer lower tax bills to higher ones, the gradual deterioration of UI coverage and benefits is understandable. Indeed, the financing for the program is effectively designed to discourage states from maintaining adequate benefits or ensuring equitable access to the program.

Under the current system, employers pay two unemployment taxes. The federal tax is a token $42 per employee (0.6 percent on first $7,000 of earnings) and finances program administration. The state taxes, which fund benefits, run as high as 10 percent on bases of $7,000 to $50,000. As noted, state taxes also vary with firms’ experience ratings and may also change subject to the sufficiency of funding of the state’s trust fund.

A major goal of the system is to be “countercyclical” – that is, to amass revenues when the economy and wages are growing. Then, during recessions, those funds are drawn down to reduce the hardship of claimants, and as a bonus to support spending across the economy and to prevent more layoffs. Once the recession is over, rinse and repeat.

Since the late 1970s, however, states have flirted with trust fund insolvency with increasing frequency because they are loath to raise the employer tax. Both federal reform commissions noted how deeply problematic the trust fund solvency issue was becoming – and equally important, that there was method to the madness: states were choosing to keep trust fund balances at critically low levels to please employers.

The 2007-9 recession laid bare both the extent of the financing problem and how states were inclined to respond. Job loss during this time was deep and prolonged – the economy shed an enormous number of jobs over multiple years. And state unemployment trust funds did not meet the moment. By 2010, 34 states had completely depleted their trust funds and had to take loans from the federal government that, in some cases, took another decade to pay off – all financed through unemployment payroll taxes.

The Government Accountability Office had been both clear and prescient: states need to raise tax rates, modernize tax structures, index their wage bases to keep up with costs and explore robust revenue-raising strategies to get their trust funds to a healthy level before the next economic downturn. Many states have done the opposite, however, following the path of least political resistance by slashing benefits and discouraging applications rather than raising taxes. Two standout examples were North Carolina, which saw massive reductions in benefit outlays, and Florida, whose governor admitted in a televised interview that the UI application had been intentionally made difficult to complete in order to keep people from applying.

The strategy pursued by most states was simple and, by its own terms, successful; the share of unemployed workers actually receiving benefits has been falling for 70 years, driven by the failure of eligible workers to file claims.

The states’ behavior is understandable if not commendable. They worry that raising business taxes just as the economy is coming out of a recession both slows recovery and angers businesses eyeing a return to profitability. Yet that is the only other option under current financing arrangements. So it shouldn’t be surprising that states, aware they are always on the edge of a race to the bottom on taxes with other states seeking to poach businesses and jobs, respond by neglecting the trust funds.

Even if a firm is not inclined to leave a state because of taxes, no employer likes to plan around fluctuating and unpredictable levies. And since firms’ tax rates rise when workers successfully file claims, most employers do all that is legal to curtail applications. Their incentives are reflected in hiring behavior: experience-rating discourages risky hires, such as young workers. But there’s direct action, too.

That’s where Talx, now Equifax Workforce Solutions, fits in. Unemployed workers who claim benefits have to be certified by their former employers that they weren’t fired for cause. Workforce Solutions, which by its own estimate manages a quarter of all unemployment claims on behalf of employers in the U.S., delays that certification process so that some workers fail to press their claims or simply move on to another job. Equifax advertises that it reduces benefits by 2.3 weeks, on average, reducing experience-related charges by 18 percent. This type of savings is only possible by slowing the payment of legitimate claims.

The cumulative effect of low taxes, low benefits, barriers to access and discourage- People waiting to be helped at the Baltimore, Maryland, Welfare Office in January of 1975. First Quarter 2025 75 ment of claims varies by state. A few states have bucked the trend and kept their unemployment insurance systems well-funded, generous and robust. But most are in varying stages of decline. Indeed, they all learned through the pandemic experience that their state’s individual stance on UI didn’t matter when it counted the most because Uncle Sam came to the rescue.

And a Well-Learned Lesson

In every recession since 1957, Congress has intervened to prop up failing state UI trust funds. The scale of the pandemic and the degree of penury to which most state UI funds had sunk made the 2020 CARES Act – with its extra weeks of UI benefits and extra weekly payments along with subsidies for workers otherwise ineligible – a Rubicon event. In effect, the federal government reminded the states that however low a state program sinks, when the national economy needs a boost Congress will step in. Unemployed workers were aided even in states that had callously allowed benefits to wither.

The CARES Act bailout was necessary; after all, the pandemic crisis left one in five workers without jobs. But it further exposed the weakness of having 53 separate state programs (50 states, plus DC, Puerto Rico and the Virgin Islands) with 53 separate computer systems that do not communicate with one another.

Hackers flooded states with fraudulent UI applications. The GAO placed the unemployment insurance system on its high risk list – the collection of government programs most vulnerable to waste, fraud and abuse – a couple years later. Even now, in the wake of years of unemployment in the blissfully low 4 percent range, just 19 states meet the trust fund threshold of “minimally solvent” and four states are still in arrears to the federal government for loans taken out when their trust funds ran dry during the pandemic.

The maximum benefit across states ranges from $275 a week to over $1,000, with the duration of benefits either fixed for all workers (often a uniform 26 weeks) or varying depending on the state unemployment rate (from one to 26 weeks). Thus workers who earn the same amount and who work the same amount could find themselves laid off and eligible for benefits hundreds less per week for only a fraction of weeks total – because of where they live. But when the economy needs relief, Congress doesn’t discriminate between programs – a further incentive to states to let their programs decline.

Conclusions, Not Beliefs

The practical takeaways from this dreary picture:

• States have had nearly 50 years of minimal federal intervention in UI policy, and they have demonstrated that, left on their own, many will not maintain their programs in good health in terms of solvency or benefit adequacy in a race-to-the-bottom environment.
• States have enjoyed nearly 70 years in which federal supplemental funding for UI during recessions was standard operating procedure, teaching them that a bailout for workers is always coming regardless of how badly they have managed their programs.
• Financing UI with an experience-rated tax on employers encourages states to pursue a low-benefit strategy and to block legitimate access to UI.
• Separate, out-of-date state computer systems operating with little coordination render UI extremely vulnerable to fraud.
• The glaring differences in the generosity of individual state UI systems put millions of workers at great financial risk from layoffs.

This grim landscape suggests Washington has two broad alternative courses of action: either dictate reform to the states in a way that sets a national, predictable uniform tax rate, wage base and benefit formula for their programs or absorb the 53 state benefit programs into a federal system with uniform benefits.

Now, most of the weaknesses with the unemployment system are tied to its tax structure. So while reform presents daunting political challenges, it is pretty obvious what needs to be targeted. The ideological critique of the benefits side of the UI equation is that workers on the dole are less motivated to find work. This has some truth: workers receiving UI on average take longer to find new jobs, but they still take them. And it is hardly a rock on which to stand against providing a reasonable safety net for the unemployed.

The impact of raising the weekly benefit or extending the weeks of benefits has a restrained impact on job searches. Even during the pandemic when Uncle Sam was adding a whopping $600 a week to benefits, the estimated impact on labor supply was relatively modest. In part, that’s because most people want to work and expect to work. In part, it’s because the value of a good permanent job is much higher to most people than the value of a time-limited benefit that is typically less than half their prior wage.

Consider, too, that the financial cover provided by generous UI allows workers to be choosier, finding jobs closer in quality and pay to the ones they lost. That serves society as well as the individuals involved because it helps to preserve the economy’s stock of knowledge, skill and experience that inevitably deteriorates when workers en masse take jobs for which they are overqualified.

As to how to proceed with reform, any of the prior blueprints – whether they be bipartisan, left of center, right of center or technocrat in origin – offer numerous suggestions similar to those in the two reform commissions’ reports ignored last century. Most reform proposals maintain the individual state systems but remove major differences, dropping the problematic experience rating and crafting changes in benefit amounts, duration and delivery better suited to the integrated U.S. economy.

Devising satisfactory fixes for UI’s current failures, then, would mostly require political will rather than analytic insight. But building a future-proofed system that grows in effectiveness rather than tying itself in knots to serve an evolving labor market is more problematic, forcing confrontation over issues at the heart of what unemployment insurance is intended to do, and for whom. I raise two of these issues – how to define fault in job loss and how to differentiate independent contractors from traditional employees – that require more than a look back at halfcentury-old reform projects to resolve.

Fault Lines

Not all instances of unemployment are covered by UI – only those in which the worker is determined to be without fault. And not all workers are insured – only those with minimum time worked.

Start with the issue of fault. First, fault is rarely the primary cause of job loss, or predictive of how easy reemployment will be for the jobless. Unemployment varies in a lot of ways that are easy to identify – age, tenure, industry, geography, the regional unemployment rate, the economic dynamic of the job destruction – that have more power to predict how traumatic the job loss will be. Yet today, a 20-year-old who loses a job as a retail clerk when a store closes during a recession has access to the same benefits as a 55-year-old production worker whose factory shuts down permanently.

Even with a zealous dedication to identifying fault as the arbiter to insurance access, the second problem is that it is hard to correctly identify. Fault (or rather lack of fault) has value – it certainly adds the veneer of worthiness to the beneficiary population – but fault is rarely clear-cut as a matter of law. For example, some states insure “good cause” quits, like a worker who must move to avoid domestic violence or because a spouse’s job is relocating. Advocates have argued that good-cause quits should also include abrupt changes to work schedules, hours or conditions. Or, going a logical step further, that an individual whose small business has failed should be able to collect unemployment.

Say reform spelled out all the good-cause quits to accompany no-fault layoffs. How exhaustive of the myriad individual circumstances would it be, and how enforceable? The more ways to be classified “without fault,” the more ways to distort layoff and quitting behavior.

Keep in mind, even a system that ended employer experience-rating but maintained the need to adjudicate fault has this problem because benefits will still be tied to employer certification. True, without a tax penalty linked to experience-rating, employers would no longer be penalized for no-fault layoffs. But this may not give them adequate incentive to be forthcoming about fault. For example, they could say any job separation was a no-fault layoff as a way to avoid the pressure to pay severance.

The worry, of course, is that without a fault requirement, workers could quit and take benefits whenever they’d like, however often they’d like, and would have less incentive to work or maintain employment. If that concern is front and center, though, it would be possible to design the scheme that limited benefits by other criteria, such as frequency of application.

While at it, a redesigned scheme could take into account the actual determinants of labor market risk among the unemployed while addressing the specter of mass job loss through technological advances like artificial intelligence. As currently designed, UI has little to offer the tens of millions of workers whose jobs may eventually be replaced by artificial intelligence.

Specifically, one could structure a worker’s benefit amount and weeks of eligibility so they reflect age, job tenure, current unemployment rates and years of work experience along with the number of claims made across their working lives. One might also allow workers to collect their benefit as a lump sum to pay toward retraining costs or to cover the cost of relocation to markets with lower unemployment. One could also add a triage system so that repeat claimers or long-term claimers are recognized early and shunted expeditiously into job-search counseling.

The bottom line: paying a fixed weekly amount for a fixed set of weeks was far more appropriate when UI was limited primarily to manufacturing workers coping with cyclical layoffs. Our diverse, service-dominant workforce needs a flexible system to match their widely varying circumstances.

Workers Without Employers

Incorporating independent contractors into UI would require collecting taxes from their employers and designing a way to determine what it means for a contractor to “lose” a job. Both are tricky, and even a simple walk through the practical issues raises myriad more.

To start, employers can be located in different states than the contractors they hire, raising the question of which trust fund would collect UI taxes. If states continue the practice of providing very different levels of generosity in benefits, this matters considerably. One way around the problem: substitute a self-employment tax in a manner similar to Social Security’s required contributions.

That, however, is much easier said than done. Nine states don’t collect income taxes and therefore lack a technical infrastructure to collect another direct tax. Plus, this approach would add to the existing problematic incentives to classify employees as independent contractors.

Then there’s the problem of the sheer diversity of workers and economic circumstances in the market for independent contractors – everything from gig work allotted via an app (think Uber) to sole proprietors running small businesses. Just who should get UI benefits and when is far from clear-cut when there’s no precipitating event like a layoff. Indeed, for many independent contractors working multiple contracts, a complete termination of earnings is less likely than a dramatic reduction.

One might add to the mix a “claim window” for independent contractors that could open up, say, when regional unemployment or firm closures exceeded threshold rates. This would raise a whole other issue, though, if sole proprietors would become eligible for cash aid during a recession but a business with even a single employee would not. This would incentivize small businesses to not expand beyond self-employment – and more generally highlight the reality that aid for businesses during most recessions is essentially nonexistent.

Okay, why not take the next step and establish an insurance system for businesses as well as workers during recessions? To put it another way, rather than trying to get contractors to fit into a traditional employment system, treat contractors as small businesses worthy of aid in some circumstances.

Helping independent contractors and businesses during recessions has an unfortunate track record, however, as both Covid- 19’s Pandemic Unemployment Assistance and Paycheck Protection Program were quickly mired in fraud. But cobbling together emergency programs with few strings attached in a crisis is not the same as carefully designing a permanent insurance program. And though problematic, this approach seems a worthy extension of the idea of unemployment insurance in an economy whose vitality depends heavily on rapid turnover in small businesses.

* * *

It always comes back to risk: what is the labor market risk and is it insured? In the U.S., this question has only been addressed twice in broadest terms: first in 1932 in Wisconsin when the first UI system was set up, and then during the pandemic in 2020, when the failings of that system were deep enough to require massive intervention extending to business as well as workers.

Reform could be about fixing the system we have – no easy task, as we’ve seen. Or it could be a chance to address the question of risk afresh, including everyone whose income is subject to influences beyond their control, with all the lessons in hand.