The Leave You Can Have but Still Cannot Use
This series examines paid sick leave systems and a deceptively simple question: why can workers still work while sick after paid leave access expands? The answer is not that leave has no effect. It is that formal coverage is only the first link in a chain that runs through eligibility, income, workload, fear, timing, and the consequences of staying home. Understanding that chain helps readers judge whether a policy reduces presenteeism, rather than assuming that an entitlement automatically removes it.
The contradiction is built into the word “access”
“Access” sounds like a binary condition. A worker either has paid sick leave or does not. That distinction matters, especially for people who would otherwise lose income by staying home. But it is not the same as being able to use leave when illness arrives.
The economic subject here is not a benefit in isolation. It is the design of a paid sick leave system: who is covered, how much income is replaced, when payment begins, what proof is required, how long protection lasts, and who ultimately bears the cost. Those choices influence whether a worker stays home, goes to work sick, returns too quickly, or postpones care. They also shape what happens to coworkers, customers, employers, and the wider public.
The distinction is visible in the coverage numbers. The IZA review reports that the share of U.S. private-sector jobs with paid sick leave rose from 63 percent in 2010 to 77 percent in 2023, as state and local mandates spread and employers increasingly offered leave voluntarily. The Bureau of Labor Statistics’ 2024–25 survey, using a different population and measure, reports that 71 percent of wage and salary workers had access to paid leave at their sole or main job, up from 66 percent in 2017–18. These figures should not be spliced into one trend line: one concerns jobs in a cited private-sector series, while the other concerns workers’ reports in a specific survey module. Still, they point to the same broad fact. Coverage has expanded, but a sizable group remains uncovered, uncertain, or unable to turn coverage into safe time away from work.
That last distinction is the important one. A worker may have leave on paper but face a large practical price for using it. The price may be lost income if the benefit replaces only part of wages. It may be a supervisor’s disapproval, a denied request, or the fear that absence will damage future hours or employment. It may be the knowledge that no one else can cover a shift. In a job that cannot be done remotely, the worker may also have no halfway option between showing up and losing pay.
Presenteeism is the name commonly given to working while ill. It is not simply the opposite of absenteeism. A person who stays home may impose a cost on an employer through an unfilled shift or lost output. A person who works sick may also produce less, face a greater risk of injury, and expose other people to infection. The economically relevant question is therefore not “How do we eliminate absence?” It is “Which choice creates the smaller total cost in this situation, and does the leave system make that choice possible?”
A worker’s decision has consequences beyond the worker
Paid sick leave changes the choice facing a sick worker. Without income protection, staying home can mean an immediate financial loss. With protection, the worker can recover without bearing the entire loss alone. That can be valuable even when no one else is affected: illness is a risk against which people may reasonably want insurance.
Contagious illness adds another layer. A worker may not know exactly when they are infectious, and an employer or coworker cannot perfectly observe that condition from the outside. The parties therefore cannot fully negotiate over the risk. If a sick employee comes to work, the resulting infection can reach people who never agreed to accept it. Paid leave gives the worker an economic reason to self-isolate during at least part of that period. That is why the IZA review describes contagion as one of the strongest rationales for government intervention.
This does not make every absence socially beneficial. A generous system can also make it easier for some people to claim leave when they could have worked, a problem economists describe as moral hazard. The system must then balance two errors. The first is inefficient absenteeism: a worker stays away even though the value of their work exceeds the benefit of resting. The second is inefficient presenteeism: a worker attends even though their productivity is low, their health may worsen, or they may infect others.
The two errors do not respond symmetrically to policy. Making leave harder to claim may reduce some unnecessary absence, but it also raises the cost of staying home with a genuine illness. Making leave more generous may reduce work while sick, but it can increase the number of people who take leave. A system that focuses only on absence rates can therefore mistake more responsible behavior for worse performance. A system that focuses only on generosity can ignore the resources used by unnecessary absence and the incentives that sustain it.
The central idea is a usable option set. Paid sick leave is effective only to the extent that a worker can recognize the benefit, qualify for it, request it, afford any remaining income loss, and take the time without facing prohibitive workplace consequences. The rules may be identical on paper for two employees while their real options differ sharply. One may have a flexible schedule, a substitute colleague, and enough savings to absorb a delay. The other may work in a high-contact service job, have no remote alternative, and depend on every scheduled hour.
The difference matters for public health as well as fairness. Remote work can reduce some opportunities for transmission, and the IZA review notes that full days worked at home rose substantially in the United States after the pandemic. But many jobs remain fully onsite, including occupations with low sick-leave coverage such as food preparation and serving. For those workers, the practical meaning of paid leave is especially important: there is no change in physical proximity unless the worker can actually stay away.
The measures are connected, but they are not interchangeable
A careful discussion must separate at least five observations. Access asks whether a worker reports that leave is available. Use asks whether time away was taken. Unmet need asks whether a worker needed leave but did not take it. Infection activity can show whether illness spread at the population level. Presenteeism asks whether someone worked while sick, often with consequences for productivity or transmission. Each measure illuminates a different point in the chain.
The BLS release illustrates why the distinction matters. In 2024–25, 21 percent of wage and salary workers took paid or unpaid leave during an average week, and four percent took leave because they were ill or needed medical care. During an average month, eight percent needed leave but did not take it. Among those with unmet needs, about 40 percent needed time for their own illness or medical care. The leading reported reasons for not taking leave included having too much work, fearing negative employment consequences or having a request denied, and being unable to afford the loss in income.
Those figures are strong evidence that access and use are not identical. They are not, however, a direct count of people who worked sick. Someone who needed leave but did not take it may have worked, delayed treatment, used another arrangement, or had a non-illness reason for needing time away. The data support a plausible route into presenteeism without proving that every unmet need became a day worked while sick.
This distinction also changes how to read evidence about mandates. If a policy causes sick-leave use to rise, that is not automatically evidence of waste. Newly covered workers may finally be able to stay home when infectious. If infection activity falls, that is evidence consistent with less contagious presenteeism, but it is not a complete measure of every worker’s experience. If access rises while unmet need remains, the policy may have improved the choices of some workers without removing the constraints faced by others.
The review’s reported research follows this logic. Studies of U.S. paid-sick-leave mandates found declines in population influenza-like illness after adoption, while other work found that newly covered employees took more sick days and that unmet sick-leave needs fell. Those findings fit together: a system can move some workers from contagious presence to legitimate absence even as it increases recorded leave use. The effect is a change in the composition of behavior, not the disappearance of illness or pressure.
That is the question this series will keep asking. When a leave policy appears successful, who could use it, what did using it cost, when did the protection begin, and which outcome changed? The answers require looking beyond the entitlement itself—to unequal coverage, practical barriers, benefit design, and the architecture that distributes risk. Paid sick leave can reduce presenteeism. It cannot eliminate the workplace and economic pressures that make working sick the easier choice for some people.