| Composition model | What it requires to be valid | What it means for virtuous capital |
|---|---|---|
| Additive | Higher-level construct = simple average | Requires justification, not automatic assumption |
| Consensus | Agreement/consistency among individual scores | Same average, different variance = genuinely different organizational condition |
| Emergent | Arises from interaction, not individual scores at all | Systems and incentives can undermine individual virtue entirely |
Chan's (1998) typology identified several distinct composition models describing how a construct measured at the individual level relates to, and potentially aggregates into, a construct at a higher level of analysis such as the team or organization. The simplest, additive composition, treats the higher-level construct as genuinely equivalent to the average or sum of individual scores, appropriate when the underlying theoretical meaning of the construct does not change across levels. More complex models, including what Chan termed consensus composition, require not only a central tendency but also agreement or consistency among individual scores to justify treating the higher-level construct as meaningful, since low agreement indicates the individual-level scores are capturing something that does not actually cohere into a genuine shared organizational property.
This distinction matters directly for how virtuous capital should be understood across the individual and organizational levels. If organizational virtuous capital were governed by simple additive composition, an organization's aggregate virtuous capital would be adequately represented by the average of its individual leaders' virtuous capital scores, and an organization with individually high-scoring leaders would necessarily show correspondingly high organizational virtuous capital. Chan's framework indicates this assumption requires justification, not automatic acceptance, and the evidence on organizational virtuousness reviewed elsewhere suggests the actual relationship is considerably more complex than simple averaging would capture.
Applying Chan's consensus composition model to virtuous capital specifically suggests that organizational-level virtuous capital depends not only on the average level of individual leader virtue but on the consistency of that virtue across the leadership population. An organization with a high average individual leader virtue score but substantial variance, some leaders scoring very high and others scoring considerably lower, presents a genuinely different, and research on organizational climate more broadly suggests less favorable, organizational condition than an organization with the same average but low variance, consistent virtue across the leadership population, even though the simple average would be identical in both cases.
This matters because low consistency in individual leader virtue specifically produces exactly the kind of unpredictable, locally variable organizational experience that Kozlowski and Klein's (2000) broader multilevel theory identifies as undermining genuine organizational-level properties: an employee's actual experience of organizational virtue depends heavily on which specific leader they happen to report to, meaning the organization does not have a coherent, shared organizational-level virtuous capital at all, but rather a collection of highly variable individual-level experiences that an aggregate score, computed as a simple average, would misleadingly present as a single, coherent organizational property.
Beyond the consistency question, multilevel theory identifies a further complication: some organizational-level properties are genuinely emergent, arising from the interaction and coordination among individuals rather than existing as a direct aggregate of any individual-level property at all. Organizational virtuous capital plausibly includes emergent components of this kind, whether the organization's formal systems, incentive structures, and decision-making processes reinforce or undermine the virtue that individual leaders bring to their roles, a systems-level property that cannot be measured or captured by aggregating individual leader assessments alone, however those assessments are aggregated or weighted for consistency.
An organization with individually virtuous leaders operating within formal systems and incentive structures that reward short-term extraction over principled stewardship is likely to show organizational-level outcomes inconsistent with what individual leader virtue alone would predict, because the emergent, systems-level component of organizational virtuous capital is actively working against what the individual-level component would otherwise produce. This is precisely the gap this brief's title identifies: individually virtuous leadership is a necessary but demonstrably insufficient condition for organizational virtuous capital, because the aggregation process connecting the two levels is neither simple nor purely additive.
Consistent with Chan's and Kozlowski and Klein's frameworks, genuine multilevel virtuous capital assessment requires measuring at least three distinct things rather than treating individual leader assessment as sufficient on its own: the average level of individual leader virtue, using the individual-level capital measures already established for this purpose; the consistency of that virtue across the leadership population, since low consistency indicates a fragmented rather than genuinely organizational-level property regardless of average level; and the systems-level, emergent component, whether formal organizational structures and incentives reinforce or undermine individual virtue, a component that individual-level assessment, however aggregated, cannot capture on its own.
The multilevel framework this brief has reviewed has a direct sequencing implication for organizations investing in both individual leader virtue development and organizational virtuous capital: developing individual leaders without addressing consistency across the leadership population, and without examining whether formal systems and incentives reinforce or undermine that individual virtue, risks producing exactly the gap this brief's title identifies, genuinely improved individual scores that do not translate into the organizational-level property the investment was ultimately meant to build. An organization that invests heavily in individual leader development while leaving substantial inconsistency across its leadership population unaddressed, or leaving misaligned incentive structures unexamined, may see individual-level assessment scores improve without corresponding improvement in the organizational-level outcomes virtuous capital theory associates with genuine organizational virtue.
This suggests organizations should sequence virtuous capital investment to address all three components this brief has identified, average individual level, consistency across the population, and systems-level reinforcement, rather than concentrating investment in individual development alone and assuming the organizational-level property will follow automatically. Organizations that measure and address consistency and systems alignment alongside individual leader development are genuinely applying the multilevel theory this brief has reviewed with the completeness the underlying theoretical framework actually requires, rather than treating individual leader development alone as a sufficient proxy for the organizational-level outcome that development investment is, in the end, ultimately intended to produce.
Because the emergent, systems-level component of organizational virtuous capital this brief has described operates independently of individual leader assessment, organizations relying exclusively on individual-level virtue measurement have no direct visibility into this component at all, and typically discover its absence only indirectly, through organizational outcomes, turnover, erosion of stakeholder trust, ethical lapses despite individually principled leadership, that a purely individual-level measurement approach cannot anticipate. Building direct measurement of the systems-level component, examining whether incentive structures, promotion criteria, and resource allocation processes actually reward the behaviors individual virtue assessment identifies as desirable, or instead reward behaviors that undermine it, provides the diagnostic visibility that outcome-based discovery, learning about the gap only after it has already produced consequential organizational harm, does not.
This kind of systems audit is a genuinely different measurement exercise than individual leader assessment, examining organizational artifacts and processes directly rather than individual behavior or self-report, and it requires expertise and methodology distinct from what individual-level virtue assessment already established elsewhere in this framework provides. Organizations building genuinely comprehensive virtuous capital measurement should treat this systems audit as a necessary, distinct measurement investment, not an assumed byproduct of strong individual-level assessment scores.
An organization discovering, through this kind of systems audit, that its performance evaluation criteria implicitly reward short-term results achieved through methods inconsistent with the virtue its individual leaders otherwise demonstrate has identified a specific, addressable misalignment between the individual and systems components of virtuous capital, precisely the kind of gap the composition framework this brief has reviewed predicts will otherwise remain genuinely invisible to individual-level assessment alone, surfacing only once it has already produced a consequential organizational outcome.
The case this brief has developed for consensus and emergent composition over simple additive composition should not be treated as automatically correct for every aspect of virtuous capital; Chan's own framework emphasizes that the appropriate composition model for a given construct is an empirical and theoretical question requiring justification specific to that construct, not a default assumption applicable uniformly across every organizational property under study. Some dimensions of virtuous capital may genuinely aggregate closer to the simple additive model this brief has argued against as a general default, and organizations should test which composition model actually fits their own data rather than assuming consensus or emergent composition applies uniformly across every dimension of virtuous capital without verification.
The practical discipline this suggests is examining the actual variance structure of individual-level virtue scores across an organization's leadership population before assuming a particular composition model, rather than defaulting to either simple averaging or the more complex models this brief has emphasized without empirical grounding for the specific organization and construct dimension being assessed. This is genuinely more methodologically demanding than simply assuming a single composition model applies universally across every construct, but it is the discipline Chan's original framework actually calls for, and applying a composition model without this justification risks the same kind of unexamined assumption this brief has argued against in the case of simple additive aggregation specifically.
The evidence and theory this brief has reviewed identifies a specific, actionable gap in how organizations commonly reason about the relationship between individual leader virtue and organizational virtuous capital: the two are related but not simply additive, and an organization with individually strong leaders can still lack genuine organizational-level virtuous capital if that individual virtue is inconsistent across the leadership population or is undermined by systems-level, emergent factors that individual-level assessment alone cannot detect. Organizations serious about building genuine organizational virtuous capital, not merely a favorable individual-level leader assessment average, need to measure consistency and systems-level reinforcement explicitly, using the multilevel composition framework Chan's and Kozlowski and Klein's research establishes, rather than assuming that developing individually virtuous leaders is sufficient on its own to produce the organizational-level property that development investment is ultimately meant to build.