The Invisible Tax: How Broken Attribution Systems Destroy Organizational Culture

A program designer named Claudia built the patient navigation framework that a health system presented to its board for six months. When her name came up in a leadership meeting, the Chief Medical Officer asked who she was. Claudia had designed, tested, piloted, and revised the framework that this executive had been using to demonstrate organizational progress for half a year. The CMO was not being dismissive. The CMO did not know.

That gap, between the person who builds something and the person who gets associated with it, is not a communication failure. It is an attribution system operating exactly as designed, which is to say: it runs on defaults, and the defaults favor proximity to power over distance from it, presentation over production, and narration over origination. Castilla and Benard (2010) documented something that helps explain why this pattern is so durable. In a set of experiments that should have made every HR department in the country uncomfortable, they found that organizations explicitly committed to meritocratic values were more likely to favor men over equally qualified women in compensation decisions. The mechanism was cognitive: when decision-makers believed the system was fair, they felt less need to examine their own judgments. The word "meritocracy" functioned as a permission structure for unexamined attribution. The label did the work so the people did not have to.

The director who presented Claudia's framework to the board did not experience himself as someone who took credit. He experienced himself as someone who synthesized, who translated technical work into strategic narrative, who served as the connective tissue between the people who built things and the people who funded things. That is what the role asked of him, and he did it well, and the doing looked, from his chair, like leadership. It looked that way because it was leadership, in the specific sense that organizations define leadership: the capacity to frame, contextualize, and represent. The problem is not that the director was dishonest. The problem is that the system counted his framing as the contribution and Claudia's framework as the raw material the contribution was made from, and neither the director nor the system had any mechanism for noticing that the accounting was wrong, because the accounting matched every assumption about what counts as high-level work.

Claudia's health system talked about fairness constantly. It was one of their stated values. The departments that talked about it most were the ones where attribution was most distorted, because confidence in the system substituted for examination of the system, and the distance between what an organization believes about how it assigns credit and how it actually assigns credit is where careers get quietly rearranged.

Heilman and Haynes (2005) clarified the mechanism through a series of experiments that have influenced how every practitioner in this space thinks about ambiguous contribution. When a man and a woman worked together on a task and the outcome was ambiguous, meaning no one had specified who did what, credit was disproportionately given to the man. When the contribution breakdown was made explicit, the disparity disappeared. The variable was documentation. In the absence of specific attribution, evaluators filled the gap with whatever cognitive shortcuts were available, and the shortcuts tracked existing assumptions about who does what kind of work. The experiment was clean. The implication is not: every undocumented contribution in every organization is subject to the same gravitational pull toward whoever the evaluator finds most default-credible, and default credibility is not distributed randomly.

The organizational cost of this pattern is specific and cumulative in ways that are difficult to see precisely because they compound slowly. One misattribution is a frustration. Twelve misattributions over eighteen months is a professional trajectory that no longer matches the contribution record. The person who narrates best gets promoted. The person who built what the narrator described gets a lateral move and a reputation as "solid," which is the adjective organizations use for people whose work they consume without tracking. Over a career, the compounding produces six-figure differences in earnings and five-to-ten-year differences in career progression, driven by narration rather than capability. Correll, Benard, and Paik (2007) demonstrated that perceived contribution directly affected promotion and compensation recommendations, which means the attribution gap is not a feelings problem. It is a compensation problem, a promotion problem, and eventually a retention problem, because the people whose work keeps disappearing into other people's summaries eventually leave for organizations that happened to notice.

The deeper structural issue is that the wrong people end up in the wrong seats. When attribution runs on defaults rather than documentation, the people who get promoted are the people who are most visible, and visibility tracks factors that have nothing to do with the quality of the work: physical proximity to decision-makers, cultural similarity to the evaluating group, comfort with self-promotion in contexts where self-promotion carries uneven social costs. Rivera (2012) found that cultural matching, the degree to which a candidate's presentation resembled the evaluator's own background, predicted hiring outcomes more powerfully than objective qualifications. The same dynamic operates in attribution: the people most likely to receive credit in ambiguous situations are the people who look, sound, and present like the people doing the crediting. The system produces its own evidence for why things are the way they are, which is what makes it so resistant to the kind of awareness-based interventions that most organizations try first. Knowing the pattern exists does not interrupt the pattern if the knowing happens at the conscious level and the pattern runs at the level of reflex.

Bowles, Babcock, and Lai (2007) added a dimension that makes clean advice nearly impossible, and the impossibility matters more than any prescription. They studied advocacy penalties: the social costs incurred by people who advocate for their own credit and compensation. The findings were gendered and consistent. Women who advocated for themselves were perceived as less likable and less hireable than women who did not. Men faced no equivalent penalty. A professional for whom self-advocacy carries a social tax is being told, simultaneously, that their work is invisible and that making it visible will cost them something the system charges unevenly. That bind does not resolve through a workshop on self-advocacy, because the workshop addresses the individual's behavior without touching the structural reality that makes the behavior costly.

There is a practice that begins to change the internal experience of this pattern, even before the structural reality shifts. It is remarkably simple. It requires no confrontation, no organizational buy-in, no conversation with anyone. Ten minutes a week, a private running record: dates, contributions, context, evidence. Not a resume. Not a trophy case. A record of what actually happened, maintained close to the moment it happened, before organizational memory has time to edit it.

Claudia, when asked to list her contributions from the prior year, named four projects. Her actual record, reconstructed from emails and shared drives, showed eleven. She had forgotten seven projects she led. Seven. The organizational silence around her work had edited her own memory. When no one reflects contributions back, the person who made them begins to forget, and the forgetting is the deepest tax of all, because it means the system has succeeded in making the invisibility feel natural. The record interrupts that. Not by making the work visible to the organization, which is a different problem with a different set of tools, but by making the work visible to the person who did it, so that the creeping doubt about whether the contribution was real, whether the pattern is imagined, whether the frustration is proportional, gets answered with dates and evidence instead of eroding further into silence.

The full framework, including the strategic decisions about when to use that record, when to advocate, and when to leave, is at [counterplay-ten.vercel.app](https://counterplay-ten.vercel.app). J. Fraser works with professionals navigating broken attribution systems.