Leadership capabilities vs competencies comes down to one test, and it is the one I apply to every framework I am handed. A competency is a skill a person can demonstrate when conditions are calm. A capability is the judgement that still works when they are not. The Center for Creative Leadership drew the same line in its 2022 revision of the Leadership Gap Indicator: it replaced the word competence with capability, which it defines as integrating knowledge and skills and adapting to meet future needs.
Most leadership frameworks measure the first thing and then report it as the second. The cost shows up in promotion decisions. A technically excellent person is moved into a role they cannot hold, and the reason is often not missing skill. What is missing is the capacity to carry judgement, ambiguity and other people's trust at the same time.
This piece sets out the definitions, a table you can use to tell the two apart, five tests I apply to any leadership standard, and the evidence rules that stop a framework sliding back into a skills list. The articles on what a leadership capability framework is and on capacity versus capability cover the neighbouring questions.
What the research says about leadership capabilities vs competencies
Start with what the literature actually says, because the terms are used loosely. A 2017 review in Cogent Business & Management traces the idea of competency to McClelland's work in the 1970s and describes it as the ability to apply knowledge, behaviours, personal attributes and skills to the tasks of a specified role. The same review notes that there is ambiguity in how competency is defined. If your organisation cannot say which definition it uses, its competency library will mean something different in every room.
CCL is more precise. In its Leadership Gap Indicator guide, competence is the possession of the skills, knowledge and capacity to fulfil current needs. Capability involves integrating knowledge and skills as well as adapting and flexing to meet future needs. The guide adds that capabilities are not about skills needed for specific jobs, job descriptions or positions. A capability model, it says, shows which characteristics leaders need to demonstrate to execute the business strategy.
A 2026 bibliometric analysis of 1,993 publications on leadership competencies, indexed in Scopus between 1960 and 2025, reaches a similar place from the other direction. Its authors note that many institutions still treat competency models as static models, and they conclude that leadership competencies should be treated as adaptive, context-responsive capabilities. Put those three sources together and the direction is clear. The research is moving from fixed lists of skills towards capability that changes with the organisation.
| Dimension | Competency | Capability |
|---|---|---|
| Unit of description | A skill or behaviour that can be listed, for example giving feedback or managing a budget | A pattern of judgement and action that integrates several skills and adapts as conditions change |
| Evidence | A demonstration, a course completed, a rating from a manager | A decision the person owned, with a consequence attached to them, seen across more than one situation |
| Behaviour under pressure | Often present in a calm room and absent under deadline pressure, conflict or missing data | The point of the definition: it is what remains when the pressure does not let up |
| How you assess it | Scored against a library, tested in an exercise or role-play | Judged from outcomes and patterns over time, using decisions, stakeholder results and repeated 360 evidence |
| Who owns it | Usually learning and development, through a catalogue | The leader's manager and the executive team, through real assignments and the operating rhythm |
The table is my working summary, not an industry standard. Where a row says a competency is usually owned by learning and development, I mean that is how the work is commonly organised, and I am not criticising the people who do it. The point of the last row is that capability cannot sit with one function, because the evidence for it lives inside the work.
Why a competency library stops predicting who you can trust
A competency library tells you whether someone can name the right behaviours in an interview. It tells you very little about whether they will make the unwelcome call when the comfortable one is easier and nobody will notice for six months. I've sat across from leaders who could recite every competency on the wall and still froze the moment a decision actually cost them something: political capital, a friendship, a comfortable narrative about themselves.
Leaders work in two worlds at once. There is the role world, which is what the role requires to produce results through other people: judgement on incomplete information, ownership of decisions, influence without formal authority, and the unglamorous work of shaping a culture nobody assigned you to shape. There is also the skills world, which is what a person can personally do. Competency models live almost entirely in the skills world. Capability lives in the role world. Build only for the skills world and you get leaders who are fluent and still not dependable under load.
That gap shows up in three places. Development turns into a training calendar, so activity is mistaken for progress. Promotions go to whoever is most visibly expert and not to whoever is most reliably accountable for enterprise outcomes. And performance conversations turn political, because nobody defined what good looks like at this level and under these constraints, so the loudest opinion wins. If you recognise the first of these, why leadership development programmes fail goes further into the mechanism.
Is the problem the person or the system around them?
Many gaps that get diagnosed as skill gaps are capability gaps, and many capability gaps turn out to be system gaps. I saw that in one Design step. The leadership team took its problem to be capability. Tracing how work and decisions moved turned up another constraint: several functions had incentives working against one another, so sensible people made choices that collided. I walked the team through the pattern with their own examples, and they altered the incentive.
That is why I do not assess capability as a personal attribute on its own. Meeting cadence, decision rights, the way conflict is handled and whether ownership is clear or only assumed all constrain or amplify what a leader can do. In my view a capable leader can look weak inside a broken operating rhythm, and a moderate one can look strong inside a well-designed one. Assess the person without assessing the leadership operating system around them and you will coach the individual when you should have redesigned the structure.
Five tests that separate a capability from a competency
When someone hands me a leadership standard, I put five questions to each line. They take a few minutes and they expose most of what is wrong with a competency wheel.
The capability lens
- Does it hold under pressure?: A competency is proven in a calm room. A capability is proven when the room is not calm: compressed deadlines, conflicting stakeholders, incomplete data. If the behaviour only appears when conditions are friendly, it is a skill.
- Is the evidence a decision or a demonstration?: Competency evidence is a role-play, a completed course or a manager's tick. Capability evidence is a decision that was owned, with a consequence attached to the person who made it.
- Does it scale with complexity or break?: A negotiation technique works on a small deal and a large one. The capacity to negotiate under board scrutiny with three conflicting stakeholders is a harder thing, and most frameworks pretend it is the same skill applied more.
- Would it survive without being told what to do?: A competency can be prompted in the moment. A capability is what a leader does when nobody is in the room to prompt them. If it collapses without a script, you have trained a skill.
- Is it visible in outcomes a year later?: The honest test of a capability claim is whether the outcome appeared months after the training budget was spent: decision speed, retention in the team, depth of succession. Workshop feedback forms measure none of those.
Run your own standard through the five tests and mark each line as skill, capability or unclear. The unclear lines are usually the adjectives. "Strategic" and "resilient" cannot be tested. "Makes the trade-off explicit and commits the team to it" can.
How I build capability evidence that cannot be gamed
A capability standard earns its place only if it survives a real promotion decision. CCL's guide sets out a sequence for building a capability model that I follow in outline: articulate the organisation's strategy, identify the capabilities that map to it and to existing internal resources such as competency models, review and select from a capabilities library, then administer the survey. My additions concern what counts as evidence and how the standard gets used each week.
- Start from outcomes, not adjectives: Describe what leaders must produce: execution that holds under pressure, a culture that does not fracture when things get hard, decision cycles that are fast and still accountable, successors who are genuinely ready. Adjectives cannot be falsified. Outcomes can.
- Define capability by level: Frontline leaders stabilise execution and coach daily performance. Functional leaders integrate across teams and build capability through the managers beneath them. Enterprise leaders set direction, own risk at system level and develop the successors who will replace them. If one set of words covers all three, the document is a poster.
- Write evidence rules in advance: Decisions made and owned, not plans proposed. Stakeholder outcomes, not meetings attended. Patterns across time, not one strong quarter. This is the difference between a framework that reduces subjectivity in a promotion decision and one that only gives the subjectivity better vocabulary.
- Route it into the leader's actual week: Turn each expectation into a practice the leader is accountable for. Review decisions after the event, honestly, and not on a hypothetical. Assign live work that produces evidence, because workshop attendance only produces a certificate.
Measurement is more common than most executives assume. A 2023 benchmark from Cornell's Center for Advanced Human Resource Studies, run with 14 partner companies, found that every one of them used a leadership measurement tool, and that all of them measured people leaders at every level. The same report says the change curve was steep, that most companies took at least a year to scale the measurement, and that the consensus was to be transparent and patient when introducing it. If your organisation has no capability standard yet, plan for a year and not a quarter.
One test of whether the work has landed is whether people one level down can say what their leader is deliberately developing in them. After 12 months in one engagement I saw leaders running their own development conversations, and 8 managers had moved into larger roles. The CHRO tracked this through internal leadership development plans. That is capability evidence: it sits in the system and does not depend on me being in the room.
Where Leadership Capability Architecture fits
I developed the 5D Leadership Model and Leadership Capability Architecture. The architecture treats capability as a property of the organisation and not only of its individuals, and it has three steps. Design diagnoses constraints and aligns systems to strategy. Embed builds capability into the leadership workflow. Scale means leaders developing leaders. A competency library lives at the first step at best. The glossary definition of Leadership Capability Architecture and the full guide to the architecture set it out in more detail.
The same reasoning explains why I read a competency framework as a hypothesis. It is a list of behaviours someone once thought mattered. The difference between a leadership framework and a leadership architecture is the system that tests the list against decisions. To see where your leaders stand against a capability standard, the leadership assessments are the practical starting point, and the systems fast-scaling companies need show what has to sit underneath.
Coaching matters here, with a condition. Coaching that reinforces a skills checklist is an expensive way to rehearse. Coaching that builds judgement, decision ownership and the stomach for accountability moves a leader from competent to capable. The evidence on what leadership coaching can add to a business is worth reading with that condition in mind. CapabilityAI is my AI leadership coach, built on this capability-first view.
The distinction to keep
A capability belongs to a different category from a competency, and calling it a bigger one hides that. A competency is something a person has. A capability is something a person can be trusted with. That difference is why competency-based systems keep producing leaders who look ready on paper and are not ready in the room.
You cannot train your way to capability. You earn it through decisions that carried real consequences, reviewed honestly and repeated under increasing complexity. Coaching can speed that up. A course alone cannot. A competency library will survive an audit, and a capability architecture will survive a hard year. Only one of those is the test your organisation actually faces.