Hybrid Work Is Not the Question: Fairness, Trust and Performance Are

Few workforce debates have absorbed as much senior-leadership attention over the last three years as the question of how many days employees should spend in the office. Return-to-office mandates, anchor days, three-two splits, and full-remote experiments fill headlines and board agendas, and they increasingly surface in engagement-survey results too. The problem lies in the framing. Days-in-office policy, taken on its own, does not explain whether a workforce performs well or badly. The factors that do explain it, fairness, trust, and the quality of performance management, cut across every working pattern, and they are the ones routinely left out of the policy-level conversation. This article sets out why the hybrid argument has been mis-specified, what the evidence actually shows, and where HR teams would do better to look.

The Misframed Debate: Days in the Office as a Proxy for Performance

The current hybrid-work discussion tends to treat location as a causal input to productivity, as if moving employees from two office days to four would, in itself, produce stronger business results. The CIPD’s Flexible and Hybrid Working Practices in 2025 report, drawing on a survey of 2,050 UK employers and 5,017 employees, found that 74% of organisations had hybrid working in place (down from 84% in 2023), that 65% required a minimum number of workplace days, and that 14% of employers allowing hybrid working planned to introduce or increase mandatory days (CIPD, 2025). Taken together, the figures point to a sector that is gradually tightening its arrangements again.

What the report also makes clear, and what the headlines tend to miss, is that flexibility remains a central driver of attraction and retention, and that how hybrid arrangements are designed and managed matters to employees far more than the headline number of office days (CIPD, 2025). Organisations are pulling hard on a fairly weak lever. The stronger ones, perceived fairness, trust in management, and the quality of performance systems, are barely being touched.

What the Evidence Actually Says About Hybrid Work

The empirical evidence on hybrid work is now considerably more settled than the rhetoric suggests. Bloom, Han, and Liang (2024), writing in Nature, reported the results of a pre-registered six-month randomised controlled trial involving 1,612 graduate employees at a large Chinese technology firm. Employees assigned to a hybrid schedule, two days a week working from home, showed no measurable effect on their performance grades across the following two years of reviews, while quit rates fell by a third. The strongest retention effects appeared among non-managers, women, and employees with long commutes, and job satisfaction rose significantly as well.

The finding is consistent with a wider body of research. Choudhury, Foroughi, and Larson (2021), in the Strategic Management Journal, used a natural experiment at the US Patent and Trademark Office to show that moving from work-from-home to work-from-anywhere increased individual output by 4.4% without degrading quality. Barrero, Bloom, and Davis (2023), writing in the Journal of Economic Perspectives, documented the persistence of remote and hybrid working at around a quarter of paid working days across advanced economies, with employees consistently valuing flexibility at the equivalent of an 8% pay rise.

None of this means every organisation should adopt the same hybrid model. It means the broad claim that hybrid working harms performance does not hold up against rigorous evidence. When hybrid models do fail, the cause usually sits in how the surrounding conditions of fairness, trust, and performance management have been designed, rather than in the remote days themselves.

Fairness: Proximity Bias and the Two-Tier Workforce

One of the most consistent findings in the distributed-work literature is that remote and hybrid employees are evaluated less favourably than their in-office colleagues, regardless of actual output. Elsbach, Cable, and Sherman (2010), writing in Human Relations, demonstrated that “passive face time”, meaning simply being seen at work without any actual interaction, produced spontaneous trait inferences of dependability and commitment. Observers drew these inferences without realising they were doing so, and the inferences shaped performance evaluations even when actual work output was held constant.

Golden and Eddleston (2020), in the Journal of Vocational Behavior, extended the picture to objective career outcomes, finding that telecommuting was associated with reduced salary growth and slower promotion, particularly for those who teleworked more intensively and lacked compensating developmental support. The effects were not uniform. Where managers actively sponsored their remote staff, the penalties shrank. Proximity bias is by now well evidenced, and it operates in the ordinary evaluation decisions organisations make rather than in some hypothetical edge case.

There is also a fairness problem in how all of this is distributed across the workforce. Flexible working is used disproportionately by women, carers, and disabled employees, so location-based distortions in evaluation do not land evenly. A policy that leaves proximity bias unaddressed ends up creating a two-tier workforce, in which those able to be in the office gain an advantage and those who cannot lose out, whatever the organisation’s stated commitment to inclusive working. Bakhshalian and Reddington (2017), in PEP® Dr: Unlocking the Puzzles of Peak Engagement and Performance, argue that engagement rests on whether employees see the way work is organised, allocated, and rewarded as fair. Hybrid policy decisions that ignore that perception tend to register in engagement data well before they show up in productivity figures.

Trust: The Psychological Contract Under Pressure

The second overlooked variable is trust, and specifically the psychological contract — the set of unwritten expectations employees and employers hold about their mutual obligations. Rousseau (1995) and Robinson and Rousseau (1994), in the Journal of Organizational Behavior, showed that perceived breaches of the psychological contract are associated with reduced trust, lower organisational commitment, higher turnover intentions, and withdrawal behaviours. Breaches are particularly damaging when employees feel that terms have been altered unilaterally and without adequate explanation.

Return-to-office mandates, introduced without consultation after several years in which flexibility had been encouraged or simply assumed, are close to a textbook case of how a psychological-contract breach happens. The exact number of days required is almost beside the point. What does the damage is the sense that the deal has been rewritten from one side, without warning. Detert and Burris (2007), writing in the Academy of Management Journal, found that whether employees are willing to raise concerns depends less on whether formal feedback channels exist and more on whether they have seen managers respond constructively to input in the past. A mandate handed down without visible consultation therefore carries two costs at once: the loss of flexibility itself, and the message that employee input is neither wanted nor acted on.

The job demands-resources model set out by Schaufeli and Bakker (2004) in the Journal of Organizational Behavior is a useful way to read this. Autonomy over when and where to work is a well-established job resource. Take that resource away without adding anything to compensate, and the balance of demands and resources tips, which the model links to lower engagement and higher burnout. Engagement data gathered after an RTO mandate often tells exactly this story, provided the analysis is sensitive enough to pick it up.

Performance: Measuring Output Instead of Presence

The third neglected variable is performance management itself. A good deal of the resistance to hybrid working rests on performance systems built for an earlier era, when visibility stood in for productivity because output was hard to observe directly. Where those systems have not been redesigned, introducing hybrid work does not get rid of the visibility proxy. It just makes it less reliable than it already was.

Elsbach, Cable, and Sherman’s (2010) work on passive face time describes this problem exactly. If supervisors infer character traits from presence, and those inferences feed into evaluations, the organisation is left with a measurement system that cannot fairly capture what distributed workers actually produce. Forcing the workforce back into the office to suit the limitations of that system is the wrong way round. The system itself is what needs rebuilding, so that it can register output regardless of where the work happens.

A smaller but growing literature sets out how that rebuilding can be done. Neeley’s (2021) research at Harvard Business School on distributed work argues that performance management in hybrid settings should be built around clear outcomes, explicit expectations, and regular structured feedback, rather than around ambient observation of who happens to be at their desk. This is a substantive change in how managers work, well beyond a change of vocabulary. It changes what managers pay attention to, how objectives get set, and how progress is reviewed. Where that change has not been made, hybrid working will look as though it produces weaker performance data. The real problem in those cases is the measuring instrument, which no longer fits the arrangement it is being asked to assess.

What This Means for Employee Engagement Measurement

These same three variables, fairness, trust, and performance, are also where engagement-survey instruments tend to be under-specified. A standard survey will routinely ask whether employees feel able to do their job well and whether they feel valued. It will far less often ask about perceived proximity bias, about whether recent policy changes feel consistent with what employees were previously led to expect, or about the quality of performance conversations in a hybrid context. The result is that the data most likely to explain why a hybrid arrangement is working, or failing to, is rarely gathered in the first place.

A more useful measurement approach starts by treating location policy as context rather than as the outcome being measured. That lets an analyst test whether engagement differences are concentrated among remote, hybrid, or on-site staff, and, more importantly, whether those differences actually track perceptions of fairness and trust rather than location as such. It is worth including validated items on psychological-contract fulfilment and perceived managerial support too, so that the effect of a policy change can be judged against a meaningful baseline rather than guessed at. Closed-ended data then needs pairing with thematic analysis of free-text responses, which is usually where proximity-bias concerns, gaps in performance conversations, and the first signs of trust erosion show up, often well before any of it reaches the turnover figures. As Bakhshalian and Reddington (2017) put it, engagement is a contextual phenomenon. The same hybrid policy can produce very different experiences from one function to the next, and only careful mixed-method measurement can tell a genuine policy problem apart from a local-leadership one.

Frequently asked questions

Conclusion

The question of how many days people spend in the office has been argued over so loudly, and for so long, that it has crowded out the questions that matter more. On the rigorous evidence, well-designed hybrid working sits somewhere between neutral and positive for performance, and it is clearly positive for retention. Where hybrid arrangements disappoint, the policy is seldom the real cause. What is usually missing is some combination of fair evaluation that counters proximity bias, a psychological contract that survives contact with change, and performance management that looks at output rather than at who happens to be visible. Organisations that get those conditions right often find the policy question more or less answers itself. Where they are left unaddressed, the same debate tends to resurface every year or two, while the factors that genuinely drive performance are allowed to drift.

For organisations seeking expert support with engagement measurement, workforce analytics, and the design of policies that take fairness, trust, and performance seriously, Elmira Bakhshalian offers workforce insight consulting grounded in more than a decade of applied research and peer-reviewed publications on engagement and workforce performance.