Job Demands-Resources Model: Designing and Delegating Well

The JD-R model explains why people burn out or thrive.

There is something quietly clarifying about a framework that says the answer to “why are my people struggling?” is almost always some version of “too much of the wrong things, not enough of the right ones.”

That may sound obvious, but obvious is not the same as simple, and the Job Demands-Resources model developed by Arnold Bakker and Evangelia Demerouti is considerably more precise than that summary suggests. It has become one of the most widely used frameworks in occupational health psychology for good reason: it is flexible enough to apply to any job, specific enough to generate actionable insight, and supported by a volume of empirical evidence that most organisational frameworks can only aspire to. Gotta love a good evidence base.

The model is also, when applied honestly, pretty humbling for leaders. Because what it consistently reveals is that the conditions most affecting whether people burn out or thrive are largely within the gift of the person running the team.

Origins and Core Logic

The JD-R model emerged in the early 2000s from a dissatisfaction with the frameworks that had preceded it. Karasek’s Demand-Control model and Siegrist’s Effort-Reward Imbalance model (see some other articles on workplace stress ) had both made important contributions to understanding workplace stress, and both were empirically well-supported. But they shared a limitation that Bakker and Demerouti found constraining: they specified a small, fixed set of working conditions as the key predictors of strain and motivation. Karasek focused on demands and control. Siegrist focused on effort and reward. Both were elegant. Both were, in the view of the model’s architects, too narrow for the complexity of actual working life.

The JD-R model’s central innovation was to propose that every occupation has its own particular constellation of risk and protective factors, and that these can be organised into two overarching categories: job demands and job resources. Rather than specifying which demands and resources matter in advance, the model provides a structure for identifying them in any given context. The framework is universal. The content is local. This flexibility is precisely what makes it so useful.

Job Demands: What Takes It Out of People

Job demands are defined as those physical, psychological, social, or organisational aspects of a job that require sustained effort and therefore carry physiological and psychological costs. The key word is sustained. A single difficult week is uncomfortable. Chronic exposure to conditions that drain energy without adequate recovery is the pathway to burnout.

The specific demands that matter differ by role.

• For a palliative care nurse, emotional demands are primary. The accumulated weight of accompanying people through death and grief, day after day is huge.
• For a software developer working across three time zones, cognitive demands and chronic context-switching take the lead.
• For someone in a matrix organisation who reports to three managers with competing priorities, role conflict and role ambiguity are the defining stressors.

The point is not that demands are inherently bad, but that they cost something. And when the cost is not offset by adequate resources or recovery, people move from tired to depleted to burned out.

It is worth being clear about what exhaustion in this context actually means. This is not ordinary tiredness that a good night’s sleep addresses. The chronic exhaustion that sits at the core of burnout is a state in which the person’s capacity to engage, to care, to respond to what the job asks of them, has been eroded below the level at which it can replenish naturally. The demands have not just been high; they have been high for long enough that recovery itself becomes compromised. The phrase health impairment process is used to describe this, or the health impairment pathway.

Job Resources: What Keeps People Going

Job resources are those aspects of the job that help people achieve their goals, reduce the costs associated with demands, or provide opportunities for growth and development. If demands are what the job takes from people, resources are what the job gives back.

The research has consistently identified several resources that appear across a wide range of occupations: autonomy and decision latitude, social support from colleagues and supervisors, quality feedback on performance, opportunities for skill development and learning, role clarity, and a meaningful connection between the individual’s work and something larger. These resources serve a dual function in the model. They satisfy basic psychological needs for competence, connection, and self-direction (see self-determination theory) and they are also practically instrumental: they help people do their jobs better, which itself generates a sense of efficacy and progress.

The evidence is unambiguous that resources predict engagement, the positive, fulfilling state of work characterised by vigour, dedication, and absorption. Schaufeli and Bakker (2004) demonstrated that resources were the primary predictor of engagement, operating through what the model calls the motivational process: a distinct pathway, running parallel to but separate from the health impairment process. Where demands deplete, resources energise. The two processes operate simultaneously, and the quality of someone’s working experience is determined by the balance between them.

We have seen this play out in practice more times than we can easily count. Teams where the workload is genuinely punishing but people remain highly engaged, because they have real autonomy over how they work, a leader who gives them timely and honest feedback, and genuine development opportunities. And teams where the objective demands are unremarkable but people are disengaged and drifting, because the role offers nothing by way of growth, the feedback is non-existent or worse, and the day-to-day work feels disconnected from anything that matters. The model gives precise language to these observations. Once you look at teams through it, it’s so easy to see what’s going wrong.

The Dual Process: Why Both Pathways Matter

The elegance of the JD-R model lies in holding both processes together. It does not claim that work is simply stressful or simply motivating. It proposes that both processes operate simultaneously, and that a person can, and often does, experience high demands and high resources at the same time.

This has a direct implication for how leaders should think about their teams. Someone who is working hard and feels supported, challenged but not overwhelmed, stretched but not depleted: that is a person operating in a healthy demands-resources balance. The same person, stripped of the resources that have been sustaining them, will not simply become less engaged. They will begin to move along the health impairment pathway even as the demands stay constant, because what has changed is not what the job is asking of them, but what it is giving them in return.

This is why headcount reductions and restructures that eliminate management layers, coaching relationships, or development programmes (e.g. to reduce costs) often produce the burnout spike that nobody in the finance case anticipated. The demands did not increase. The resources disappeared. The model predicts exactly what happens next.

The Buffering Hypothesis: Resources as Protection

One of the most practically important propositions in JD-R theory is the buffering hypothesis: the idea that job resources can directly mitigate the damage of high demands. In plain terms, demanding conditions are less harmful when adequate resources are also present. A teacher facing large classes and curriculum pressure will suffer less if she also has genuine autonomy over how she teaches, a supportive head of department, and specific feedback on what is working. The demands have not changed. The resources change what those demands do to her.

Bakker and Demerouti (2007) provided evidence for this interaction across multiple studies, showing that resources such as autonomy, social support, and performance feedback moderated the relationship between demands and burnout. The buffering effect is not unlimited. There comes a point at which demands overwhelm any conceivable level of resources but, within the normal range of working conditions, it is a robust finding.

For leaders, this may be the single most actionable insight in the entire model: you may not always be able to reduce the demands your team faces, but you can almost always increase the resources available to them. And doing so genuinely matters.

Challenge Demands vs. Hindrance Demands: Not All Pressure Is the Same

One of the most significant refinements to the original model came from Crawford, LePine, and Rich (2010), who argued that not all demands function in the same way. Some demands, while effortful, are experienced as opportunities for mastery and growth. Others are experienced as obstacles. They’re experienced as things that get in the way without offering anything in return. The distinction matters enormously in practice.

Challenge demands: workload, time pressure, job complexity — tend to predict both strain and engagement. They cost energy, certainly, but they carry the potential for satisfaction, competence, and the sense of having done something genuinely difficult. Within reasonable bounds, they can be motivating precisely because they stretch people. This is not a contradiction; it is how growth works.

Hindrance demands: role ambiguity, organisational politics, bureaucratic obstruction, job insecurity — predict strain but not engagement. They drain energy without offering any compensating sense of progress or achievement. They are the demands that make people ask “what is the point?” rather than “this is hard but worthwhile.” Crawford and colleagues’ meta-analysis confirmed this distinction empirically: hindrance demands have no positive relationship with engagement, while challenge demands, within limits, do.

If you have ever wondered why your team seems to come alive under intense project pressure but slowly die under pointless administrative process, this distinction explains it. The goal is not fewer demands per se. It is fewer of the wrong kind of demands (the ones that take without giving) and a clearer understanding of which is which. Many of the hindrance demands that characterise large organisations (duplicated approval chains, unclear decision rights, meetings that exist because they have always existed) can be reduced without affecting output at all. They are simply waste. They’re often really just organisational overhead that costs people something real and produces nothing in return.

What This Means for Delegation

The JD-R model has a specific and under-appreciated relevance to how leaders delegate. Most conversations about delegation focus on efficiency: get the right work to the right person, free up your own time, develop your team’s capability. These are real benefits, but the model points to something more fundamental and, we think, actually far more helpful.

Delegation is one of the most powerful tools a leader has for shifting a team member’s demands-resources balance. Done well, it increases autonomy (a primary resource) by giving people real ownership of meaningful work rather than a small slice of a task the leader has effectively already decided. It creates conditions for skill development and learning. It provides a natural feedback mechanism as people navigate the work themselves. These are not side effects of good delegation; they are the point of it, from a JD-R perspective. The act of genuinely handing over a piece of work, with the authority to make decisions about it, moves resources toward the person doing it.

Done badly, delegation does the opposite. The leader who delegates the task but retains all the decisions has increased the demands on their team member (more to do, more to be accountable for) without increasing the resources (no real autonomy, no new skill, no meaningful ownership). This is a pattern we observe constantly, and the JD-R model names what makes it damaging: it creates demands without the resource that would give those demands value. The person is working harder with no more autonomy, no clearer feedback, and no genuine growth opportunity. They have gained a burden, not a development experience.

Similarly, job design decisions like which work sits in which role, how broadly or narrowly roles are defined, how much latitude people have over how they approach their work, are fundamentally demands-resources decisions. A role that is too narrow offers few demands of the challenge variety and few resources of the growth or autonomy variety. A role that is poorly specified creates hindrance demands (role ambiguity, role conflict) that no amount of individual resilience will fully address. Getting these design choices right is a leadership responsibility, and the JD-R model provides the diagnostic frame for doing so with some rigour.

Personal Resources: What People Bring

The original JD-R model focused exclusively on job characteristics, but subsequent developments incorporated what Bakker and Demerouti (2017) called personal resources. These are the psychological characteristics individuals bring to their work. Self-efficacy, optimism, resilience, and organisation-based self-esteem all function analogously to job resources within the model (see HERO / psychological capital). They predict engagement, help buffer the impact of demands, and are themselves influenced by the availability of job resources in a reciprocal relationship.

This reciprocity is important. People with stronger personal resources tend to perceive fewer demands as threatening, are better at accessing the job resources available to them, and experience greater engagement. This in turn strengthens their personal resources further. The spiral can also operate in reverse: when sustained demands erode personal resources, the person’s ability to draw on available job resources diminishes, demands loom larger, and the downward cycle accelerates.

There is a practical implication worth stating directly, because it is often misapplied. Coaching, stretch assignments, and confidence-building leadership behaviours can meaningfully develop personal resources, and this development genuinely helps. But personal resources should never be treated as a substitute for adequate job resources. This is something we feel very strongly about.

Sending someone to a resilience workshop while doing nothing about the structural conditions that are grinding them down is not a development intervention. It is an abdication of responsibility dressed up as one. We see this conflation all too often in organisations that have discovered wellbeing as a strategic priority without examining what is actually making their people unwell.

What the Evidence Shows

The empirical base for the JD-R model is, by any reasonable standard, substantial. Since the original 2001 paper, the model has been tested in hundreds of studies across dozens of countries and virtually every sector imaginable. Some of our friends and associates do a lot of work with JD-R too and their experience is always very positive.

Furthermore, meta-analyses have consistently confirmed its core propositions: demands predict burnout through a health impairment pathway, resources predict engagement through a motivational pathway, and the interaction between them explains additional variance beyond what either predicts alone.

Cross-cultural studies have been particularly encouraging for the model’s ambitions. The basic dual-process structure appears to hold across Western (it originated in the Netherlands) and non-Western contexts, across public and private sectors, across manual, service, and professional work. The framework appears to be universal. The specific demands and resources that matter most are local, which is precisely what the model predicts. This combination of structural robustness and content flexibility is what has made it useful in so many different settings.

The most methodologically sophisticated strand of JD-R research in recent years has moved beyond cross-sectional surveys to diary studies tracking day-to-day fluctuations in demands and resources. This work has shown that the model operates not just at the level of chronic work conditions but at the level of daily experience: on days when resources are higher than usual, engagement rises and the impact of demands is buffered; on days when resources are depleted, the same demand level produces greater strain. The model is not just a description of stable patterns but a dynamic account of how people experience their working days.

Why This Matters for Leaders

What the JD-R model offers that most frameworks do not is a genuinely diagnostic tool. Rather than prescribing a fixed list of things you should do to improve wellbeing or engagement such as more autonomy, less workload, better feedback, it invites you to ask two deceptively simple questions: what are the demands my people face, and what resources do they have to meet those demands?

The answers are specific to your context, and that specificity is the point. A team of healthcare workers and a team of software engineers face different demands and need different resources. A newly formed team grappling with role ambiguity has different needs from an established team drowning in workload. The model does not tell you what you will find; it tells you how to look.

In practice, this usually means sitting with your team and asking questions that leaders do not ask often enough: what about your work feels genuinely energising, and what feels relentlessly draining? Which pressures feel like they are making you better at what you do, and which just feel like obstacles? What do you have enough of, and what are you consistently short of? These are not complex questions. They require genuine curiosity about other people’s working reality and a willingness to hear answers that may implicate your own decisions. Sometimes, though, these are questions that are best asked by a non-leader in a listening session or through facilitators in anonymous ways.

We have worked with leaders who used a simple demands-resources mapping exercise and were genuinely surprised by what emerged. One team’s primary stressor was not workload, as the leader had assumed, but a chronic absence of communications. Members in this team had no clear sense of whether what they were doing was valued or even noticed. Another team’s most cited resource was not the formal development programme the organisation had invested heavily in, but an informal peer support network in Slack that had emerged organically. The model, applied honestly, surfaces things that assumptions and surveys alone rarely reach.

Applying the Model: What to Look For and What to Do

For leaders inclined to put this into practice, the starting point is always an audit of the actual conditions your team is working under, not the conditions you believe or hope are in place. Demands tend to creep. Workload increases incrementally. Roles expand without formal acknowledgement. Hindrance demands, things like the bureaucratic layers, the unclear decision rights, the meetings that no longer serve a purpose, accumulate over time because nobody actively removes them. The mapping exercise requires deliberate attention, because the steady state of most organisations is a slow drift toward higher demands and eroding resources.

On the demand side, the most useful distinctions are between what is genuinely necessary and what has accumulated through inertia, and between what stretches people in ways that generate growth and what simply obstructs and frustrates. Not every pressure can or should be reduced. Challenge demands that are developmentally appropriate and adequately resourced are part of what makes work meaningful. Hindrance demands, things like role ambiguity, political friction and unnecessary complexity, can usually be reduced without any loss to output or quality. Doing so pays dividends that show up in engagement, retention, and performance.

On the resource side, the first question is sufficiency: are the resources currently available to your team adequate for the demands they face? This is not a question about whether you have provided what the organisation nominally offers, things like training programmes, performance review cycles, formal feedback mechanisms. It is a question about what people actually experience in day-to-day working life.

• Does your team have real autonomy, or nominal autonomy that collapses the moment their judgement differs from yours?
• Are they receiving feedback that is specific and timely enough to be useful, or feedback that arrives in an annual review and serves primarily administrative purposes?
• Do they have genuine development opportunities, or just a training budget that nobody has time to use?

The second resource question is equity: whether the resources available in your team are distributed evenly. Some individuals will have access to a supportive mentor relationship, a flexible schedule, or an unusually clear role that others do not. Resource inequity is a dimension of fairness that leaders frequently overlook, partly because it is less visible than workload inequity. When one person’s role is much better-resourced than another’s, the gap shows up eventually. It will eventually appear in differential engagement and in differential retention, but by the time it is visible it has often been present for a long time.

On delegation specifically: think deliberately about whether the work you are handing over comes with genuine authority or is still tethered to your decisions at every significant point. Delegation that increases accountability without increasing autonomy has added a demand without providing the resource that would give that demand meaning. The question to ask is whether the person you are delegating to will be able to develop genuine competence and ownership through the work, or whether they will simply be executing tasks that happen to sit on their to-do list rather than yours.

Criticisms and Limitations

The model’s strengths are also, in certain respects, its weaknesses,.

The most persistent critique concerns the model’s flexibility. Because it does not specify which demands and resources matter in advance, critics have argued that it risks becoming a classification system rather than a theory. It’s a taxonomy system capable of labelling anything as a demand or a resource after the fact, but not capable of predicting, before the fact, which factors will matter in a given context.

A related issue is the genuinely messy reality that some job characteristics refuse to sit neatly on one side of the demand-resource line. High autonomy is typically a resource, but in a role without adequate support or guidance, autonomy can tip into isolation and overwhelm. Emotional engagement with meaningful work is typically sustaining, but the same engagement can become a source of exhaustion when the emotional demands are unrelenting. The same job characteristic can be experienced as a resource by one person and a demand by another, depending on their skills, preferences, and circumstances. The model assumes cleaner categorisation than lived experience always permits.

Measurement complexity is a further legitimate concern. Because the model is context-specific, there is no single standardised questionnaire that captures all relevant demands and resources across all jobs. Researchers must select or develop measures appropriate to each population, which introduces variability across studies and makes cross-study comparison more difficult than it would be with a fixed-variable framework.

Finally, and most fundamentally, most JD-R research remains cross-sectional, which makes the causal direction of the relationships genuinely ambiguous. Correlation is not causation and all of that.

• Demands may cause burnout > But burnout also causes people to perceive demands as more threatening.
• Resources may generate engagement > But engagement may also motivate people to actively cultivate resources.

The model now acknowledges these reciprocal, dynamic processes, but the empirical work of establishing causal direction is still, in places, catching up with the theoretical ambition.

These are all real limitations, but they do not significantly undermine the model’s practical utility.

Learning More

To go deeper, explore these resources on people-shift.com:

• Wellbeing: What It Is and How to Improve It
• Albrecht’s Four Types of Stress
• Employee Engagement
• Resilience in the World of Work
• The Relationship Between Performance and Pressure
• Herzberg’s Two Factor Theory of Motivation

At PeopleShift, we regard the JD-R model as one of the most practically useful frameworks in occupational psychology, precisely because it refuses to offer a one-size-fits-all answer. It does not tell you that autonomy is the solution, or that workload is the problem, or that resilience training will fix everything. Instead, it tells you to look at the conditions people actually work under, and to take seriously both what is draining them and what is sustaining them.

Like all these models, the diagnostic work itself doesn’t really add value or change things. It’s the conversations and interventions that are built on the back of the diagnostics that lead to positive changes for individuals and organisations.