How recession fears, financial uncertainty, and chronic stress reshape the way Americans think, work, and make decisions.
There is a moment many Americans now know well. The number on the gas pump climbs past where it stopped last week, and a small, involuntary calculation begins. It is not only what does this cost but what does this mean. Is this temporary? Is it the start of something? Should I be doing something differently?
That second calculation, the one about meaning, is where the economy stops being a set of figures and starts behaving like a psychological threat.
By the standard measure of how households feel, 2026 has been an unusually heavy year. The University of Michigan’s Index of Consumer Sentiment fell to 44.8 in May, a record low, as supply disruptions tied to the conflict with Iran pushed gasoline prices higher. Nearly 40 percent of respondents mentioned gas prices without being asked. The final October reading was 48.0, which the survey’s director, Joanne Hsu, noted was about 15 percent below February, just before the conflict began.
The American Psychological Association’s latest Stress in America survey, fielded from late July through August among a national sample of 3,132 adults, captures the same weight from another angle. Seventy-seven percent rated the cost of living a very or somewhat significant source of stress, and 68 percent agreed, strongly or somewhat, that they had become accustomed to living with uncertainty. The report’s subtitle is The stress of uncertainty.
Surveys like these record what people report, not what happens in their bodies. But research increasingly suggests that how people expect the economy to behave can affect their bodies and minds well before the feared outcome arrives, and sometimes whether or not it arrives at all.
Uncertainty is a stressor in its own right
In 2016, a University College London team led by Archy de Berker designed an experiment to separate two things that usually travel together: bad outcomes, and not knowing whether they are coming. Forty-five participants guessed whether a snake was hidden under a rock. When a snake appeared, they received a mildly painful electric shock to the hand. A computational model estimated each participant’s uncertainty on every trial.
Stress tracked uncertainty. Self-reported stress, pupil dilation, and skin conductance peaked when the chance of a shock was near 50 percent and were lowest when a shock was either very unlikely or nearly certain. Not knowing was worse than knowing, even when knowing meant pain.
The study used a small sample, brief sessions, and a sting-like shock, not months of rising prices or rumors of layoffs, so applying it to recessions is an analogy rather than a demonstration. Still, the analogy is useful. An economy in which recession is perpetually “possible,” debated nightly but never confirmed or ruled out, resembles the experiment’s most stressful condition: a coin that never stops spinning.
The authors also pointed to an adaptive side. Participants whose stress tracked uncertainty more closely were better at predicting shocks. Stress is part of how the brain learns about danger. The trouble starts when the danger cannot be learned because the signal never resolves.
The worry that outlasts the event
If uncertainty carries its own cost, then the fear of losing a job might have consequences that do not depend on actually losing it. Sociologists Sarah Burgard, Jennie Brand, and James House tested this in a 2009 study in Social Science & Medicine, drawing on two nationally representative U.S. longitudinal datasets covering periods of roughly three to ten years. Persistent perceived job insecurity predicted poorer self-rated health in both samples, and more depressive symptoms in one.
The association held after the researchers accounted for actual job loss and unemployment, and after controls for demographic and job characteristics, earlier health, and a general tendency to report things negatively. Burgard offered several explanations: prolonged ambiguity about the future, the inability to act until the feared event happens, and the lack of the institutional supports that exist for people who have actually lost work.
This is observational research, and self-rated health is an imperfect outcome. But it points to something economic coverage tends to miss. Unemployment figures count the people who have lost jobs. They do not count the much larger group who go to work each day wondering whether they will be next.
The bandwidth tax
Financial pressure may also change how people think, not only how they feel. In a widely cited 2013 paper in Science, Anandi Mani, Sendhil Mullainathan, Eldar Shafir, and Jiaying Zhao proposed that money worries consume mental resources that would otherwise be available for other tasks.
In a New Jersey shopping mall, roughly 400 participants thought through a hypothetical expense before taking cognitive tests. When the expense was small, lower- and higher-income participants performed similarly. When it was large, lower-income participants performed worse and higher-income participants did not. In India, the team tested 464 sugarcane farmers before and after harvest, and the same farmers scored higher after being paid. The authors argued that this reflected attention capture: a pressing concern pulls the mind toward it and leaves less “bandwidth” for everything else.
The finding has been challenged. In 2016, economists Leandro Carvalho, Stephan Meier, and Stephanie Wang randomly surveyed low-income U.S. households just before or just after payday. Despite clear differences in available cash, the two groups performed similarly on cognitive tasks. The difference may reflect short payday cycles versus deeper, longer-lasting scarcity, or an effect smaller and more context-dependent than first reported. For now, the bandwidth hypothesis is best read as a plausible mechanism under active debate, not a settled law.
When your livelihood depends on your attention
For developers and other knowledge workers, these findings raise a pointed question, because attention is not incidental to the job. It is the job. Debugging a race condition, holding a system’s architecture in mind, or reviewing a pull request carefully all depend on sustained, uninterrupted focus.
None of the studies above tested software developers, and nothing in them shows that job insecurity degrades code quality. If financial or job-related worry captures attention, as Mani and colleagues proposed, then work that requires holding many details in mind at once could be among the first to feel it, while urgent, well-defined tasks keep getting done. That is a hypothesis, not a finding.
There are signs the worry is common in the field. In Devographics’ State of Devs 2026 survey, 2,578 of the 5,352 developers who answered a question about workplace issues, about 48 percent, said they had experienced job insecurity at some point in their careers. The sample is self-selected, drawn largely from people who took earlier Devographics surveys and weighted toward web development, and the question covered entire careers rather than the present moment. The figure describes that community, not the profession as a whole.
The UCL experiment invites one further, more speculative extension. In the lab, stress helped people learn when a threat was coming. Whether anything similar happens at work has not been tested, and the leap from laboratory physiology to workplace behavior is a large one. Still, a possible parallel is easy to recognize: refreshing layoff trackers, rereading a manager’s message for hidden meaning, scanning company channels for hints. If habits like these are attempts to resolve uncertainty, they may also draw on the focus the work depends on.
What can be controlled
None of this means Americans are collectively impaired. People vary widely in how they respond to uncertainty, and surveys capture mood more than capacity. The burden is also uneven. APA surveys have long found that lower-income adults, younger adults, and parents report more money-related stress, and in 2026, 59 percent of adults agreed, strongly or somewhat, that they felt they were doing everything “right” and still falling behind.
The research does suggest where relief might come from. In the UCL experiment, and in decades of earlier animal research, predictability and a sense of control reduced stress responses. That points less toward individual willpower and more toward the conditions people live in: employers who say clearly what is and is not changing, policies that soften the cost of a single financial error, and support that reaches people who are anxious before they are unemployed. Burgard’s team noted that programs built for displaced workers miss those who still have jobs but fear losing them. APA’s 2014 data adds one more factor: adults who reported having emotional support said they managed financial stress in healthier ways than those who did not.
Perhaps the larger question is what happens to a society when millions of people spend more mental energy preparing for what might go wrong than building what could go right. The economy may or may not tip into recession in the coming months. For many Americans, the uncertainty has already arrived, and the mind does not wait for the official data to respond.
This is Part 1 of America Under Pressure: The Psychology of Uncertainty. In Part 2, we’ll examine how election coverage, political uncertainty, and digital platforms compete for our attention during an already stressful period.
References
American Psychological Association. (2015, February 4). APA survey shows money stress weighing on Americans’ health nationwide [Press release]. https://www.apa.org/news/press/releases/2015/02/money-stress
American Psychological Association. (2026). Stress in America™ 2026: The stress of uncertainty. https://www.apa.org/pubs/reports/stress-in-america/2026
American Psychological Association. (2026). Stress in America™ 2026: Topline data. https://www.apa.org/pubs/reports/stress-in-america/2026/topline-data.pdf
Burgard, S. A., Brand, J. E., & House, J. S. (2009). Perceived job insecurity and worker health in the United States. Social Science & Medicine, 69(5), 777–785. https://pmc.ncbi.nlm.nih.gov/articles/PMC2757283
Carvalho, L. S., Meier, S., & Wang, S. W. (2016). Poverty and economic decision-making: Evidence from changes in financial resources at payday. American Economic Review, 106(2), 260–284. https://doi.org/10.1257/aer.20140481
de Berker, A. O., Rutledge, R. B., Mathys, C., Marshall, L., Cross, G. F., Dolan, R. J., & Bestmann, S. (2016). Computations of uncertainty mediate acute stress responses in humans. Nature Communications, 7, 10996. https://doi.org/10.1038/ncomms10996
Devographics. (2026). State of Devs 2026: Career. https://2026.stateofdevs.com/en-US/career/
Mani, A., Mullainathan, S., Shafir, E., & Zhao, J. (2013). Poverty impedes cognitive function. Science, 341(6149), 976–980. https://doi.org/10.1126/science.1238041
University of Michigan Surveys of Consumers. (2026, May 22). Consumer sentiment, final results for May 2026. https://data.sca.isr.umich.edu/fetchdoc.php?docid=81238
University of Michigan Institute for Social Research. (2026). As concerns over purchasing power dominate, consumer sentiment moves sideways. https://isr.umich.edu/news-events/news-releases/as-concerns-over-purchasing-power-dominate-consumer-sentiment-moves-sideways/
Top comments (12)
Absolutely outstanding article! One of its most powerful insights is that uncertainty itself can become a psychological burden long before a crisis actually happens. The distinction between losing a job and living every day with the fear of losing one is particularly profound. I also love the concept of the “bandwidth tax” and how financial anxiety can quietly consume the very mental energy people need to perform, create, and move forward. This is an important reminder that economic well-being is about far more than numbers; it is also about mental space, predictability, and a sense of control over one's future. A deeply thought-provoking contribution that deserves a much wider conversation!
This piece is careful about the distance between what each study shows and what it's being asked to support. The "hypothesis, not a finding" framing in the developer section is the right register, and it's rare to see the payday replication failure included alongside the original bandwidth result.
A few thoughts:
The strongest thread is that uncertainty is distinct from bad news. The de Berker result (stress peaking at 50 percent shock probability) suggests why "recession is possible" may feel worse than "recession is here." A confirmed downturn at least lets people replan. An unresolved one keeps the planning loop open. That also explains why the Burgard finding is so important: perceived insecurity predicted worse health even after controlling for actual job loss. The anticipation carries its own cost.
The Mani and Carvalho disagreement may be less of a contradiction than it looks. Your explanation (short payday cycles versus deeper scarcity) is the most plausible reconciliation. If bandwidth capture depends on a concern being pressing and unresolved, then a predictable payday may relieve it only partially, because the underlying shortfall is still there. The hypothesis may hold for acute, salient shocks (the large hypothetical expense, the pre-harvest farmers) and fail for chronic, habituated scarcity, where people have built routines around the constraint. That would also be consistent with the APA finding that 68 percent say they've become accustomed to living with uncertainty. Habituation might blunt the cognitive tax while leaving the health tax intact. Those are separable mechanisms, and the article could say so more directly.
The "refreshing layoff trackers" passage deserves a sharper distinction. Information-seeking under uncertainty can be adaptive or ruminative. The UCL result suggests stress helps calibrate predictions when feedback arrives. Checking a tracker that updates daily with no resolution gives the loop nothing to close. A useful test is whether the checking produces a decision. If it changes what you do (updating a résumé, building a cushion), it's planning. If it only restarts the worry, it's the spinning coin. That's a practical question a reader can ask themselves, and it connects the lab finding to behavior without overclaiming.
The control and predictability section is where the piece's policy conclusion lives, and it could carry more weight. The point about employers communicating what is and isn't changing is concrete and actionable, especially for knowledge workers where attention is the product. Silence from management isn't neutral. By the logic of the UCL study, it's the 50 percent condition. A manager who says "there will be no layoffs this quarter, and I'll tell you by March if that changes" has converted an open-ended threat into a bounded one, even without any good news.
A minor note on the survey figures. The Michigan index and the APA data measure different things (expectations and sentiment versus self-reported stress), and both are subject to partisan and framing effects. You flag that surveys record reports, not physiology, which is good. It might be worth a sentence acknowledging that sentiment readings have moved with political context in recent years, so the record low is informative but not purely an economic signal.
Do you plan to bring the Burgard framework into Part 2? Election-season uncertainty seems like a natural test of whether the "anticipation without resolution" mechanism generalizes beyond jobs.
This captures something critical that standard economic metrics completely miss: the cognitive toll of anticipation.
We talk constantly about GDP, CPI, and unemployment rates, but we rarely measure the daily mental energy drained by the potential of bad news. The point on the "bandwidth tax" hits particularly hard for knowledge workers—when your primary tool is sustained focus, chronic ambient stress isn't just a mental health issue; it directly degrades the exact resource your job requires.
The UCL experiment analogy is spot on. It’s not just the bad outcome that breaks people's focus—it’s the hypervigilance of sitting at 50% probability, refreshing Slack, scanning layoff trackers, and parsing management's tone for hidden signals.
Clear communication from leadership and structural transparency aren't just HR buzzwords—they are essential guardrails against collective burnout. Exceptional write-up.
The takeaway here for leadership is clear: opacity isn't neutral. In the absence of clear data, the brain fills in the gaps with worst-case scenarios, dragging team productivity down with it.
As the study notes, pain that is predictable causes less acute stress than perpetual ambiguity. Leaders can't always guarantee economic stability or ward off market shifts, but they can control transparency. Clearly communicating what is known, acknowledging what isn't, and establishing predictable check-ins can significantly lower the cognitive burden on teams during turbulent times.
This hits on something many of us feel daily but rarely quantify: background economic anxiety acts like a continuous background process consuming system memory.
When applied specifically to engineering and knowledge work, the "bandwidth tax" concept is particularly resonant. Debugging complex systems requires deep contextual loading into working memory; if part of that mental bandwidth is constantly occupied by subtle hyper-vigilance—parsing ambiguous Slack messages, checking market trends, or internalizing job insecurity—our total cognitive capacity takes a silent hit.
The distinction between actual job loss and the anticipation of it is also a crucial point for engineering leaders to grasp. Ambiguity is inherently high-friction. Transparent communication from leadership about runaway roadmaps or shifting financial goals does more than just align team milestones—it literally frees up the cognitive cycles necessary for developers to do high-level problem solving. Looking forward to Part 2 on how election and media cycles compound this baseline noise.
This is a particularly important perspective because it highlights something economic discussions often overlook: the psychological cost of uncertainty can arrive long before the financial consequences themselves. I found the distinction between actual unemployment and the persistent fear of losing a job especially compelling. For knowledge workers, attention is one of our most valuable professional resources, and constantly worrying about what might happen next can quietly consume it. I also appreciate the author's caution in distinguishing established research from hypotheses about developers. Perhaps organizations should recognize that providing clarity and predictability isn't just good management; it's an investment in the mental bandwidth and long-term performance of their people.
Thank you for this. The line that stayed with me is the distinction between the unemployment figures and the much larger group who go to work each day wondering whether they will be next. Burgard’s finding that persistent job insecurity predicted poorer health even after accounting for actual job loss suggests the cost of the worry is real, and rarely counted.
I also appreciated the de Berker study. Stress peaking when the chance of a shock was near 50 percent, and being lower when the outcome was nearly certain, even if painful, explains a lot about why an economy where recession is always “possible” but never confirmed wears people down. Your honesty about it being an analogy rather than a demonstration made it more convincing, not less.
Giving the bandwidth hypothesis fair treatment, including the Carvalho payday study that challenged it, was a good call. It’s rare to see a popular piece present a contested finding as contested.
The point about predictability and control reducing stress responses is where I’d put the emphasis. It shifts the question from individual resilience to the conditions people live in. Employers who say plainly what is and isn’t changing cost very little, and the research suggests they could ease a real burden.
This is a powerful perspective on how economic uncertainty affects more than just our finances. The idea that not knowing what comes next can be more stressful than facing a confirmed outcome really resonates. What stands out most is the hidden cost of financial anxiety: the mental energy consumed by worrying about job security, rising expenses, and an unpredictable future is energy that could otherwise go toward creativity, productivity, and personal growth. I also appreciate the distinction between actual unemployment and the psychological burden of fearing it. Sometimes the greatest challenge isn't the crisis itself, but living in constant anticipation of one. Greater transparency, predictability, and emotional support are not just workplace benefits; they're essential to people's overall well-being.
This is an excellent exploration of how financial uncertainty affects far more than our wallets—it quietly consumes our attention, emotional energy, and sense of control. The distinction between experiencing a setback and constantly anticipating one is especially powerful. People can be employed, meeting deadlines, and appearing productive while privately carrying the stress of not knowing what comes next. I also appreciate that the article acknowledges the limits of the research rather than overstating its conclusions. Its strongest takeaway is that predictability, transparency, and emotional support aren't luxuries; they're essential conditions for people to think clearly, work effectively, and feel secure enough to move forward.
The de Berker experiment reframes something I’d felt but never had language for: the worst state isn’t bad news, it’s a coin that never stops spinning. That maps closely to how a “maybe layoffs, maybe not” period feels at work. Certain bad news at least lets you act, while ambiguity keeps you in standby mode.
I appreciate that you flagged the bandwidth finding as contested. The payday replication is a good reminder that a catchy mechanism can be smaller and more context-dependent than the headline suggests. Still, the Burgard result stands out to me, since perceived insecurity predicted worse health even after controlling for actual job loss. That’s the group no one counts.
For developers specifically, I’d suggest the practical lever is managers and teams, not individual resilience. Since predictability is what reduces the stress response, a few things help a lot:
Say what is and isn’t changing, even when the answer is “we don’t know yet, and here’s when we will.”
Keep planning horizons visible, so people aren’t decoding roadmap changes for hints about their own jobs.
Protect focus time, because attention is the scarce resource when worry is already drawing on it.
On the habit of refreshing layoff trackers: I’d guess a lot of that is an attempt to resolve the uncertainty yourself. A small counter-practice is to set a specific time to check news or channels, rather than checking in the gaps between tasks, so the worry gets a container instead of leaking into deep work.
Thanks for the careful sourcing and for being clear about where the studies stop and the extrapolation begins.