The study that follows comes from the kind of data treasure you rarely see in public with nine years of monthly records from a large civilian workforce, covering more than 70 000 people. Conducted by a team at George Mason University led by Bryan Adams, Valentín Vergara Hidd, and Eduardo López, the work builds a new framework for understanding how people drift through the internal ladders of big organizations. It is a rare blend of granular social life and big system thinking, stitched together by careful data craftsmanship and a willingness to ask the boringly important question: why do people switch jobs inside a company in the first place?
Rather than treating an organization as a simple machine that matches supply and demand for the right skills, the authors propose that teams and social bonds live at the heart of mobility. They show that when people move, they often do so not to chase the next shiny role, but to reconnect with coworkers they have known before. In the Army Acquisition Workforce, a huge slice of moves — about one in three — involve reuniting with past teammates. That is a striking pattern, one that challenges conventional models of internal labor markets and points toward a social microdynamics logic at scale.
A new lens on internal mobility
Traditional approaches to internal mobility tend to treat workers as interchangeable units in a vacancy driven market. You study who needs a job, who can fill it, and how the right skill set migrates from one box to another. This paper pushes back on that assumption by insisting on a richer map of the inside of organizations. Teams are not just labels on an org chart; they are living micro communities where people train together, share tasks, and build reputations. The authors therefore construct a time evolving map of teams, with rules that govern when teams form, when they disperse, and when a group of people moves together between teams. This allows the authors to track the social footprint of a move as a function of who worked with whom and for how long.
One of the clever ideas in the paper is to separate two kinds of moves. Uncoordinated moves are single individuals shifting from one team to another, while coordinated moves involve two or more people moving as a group. This distinction matters because coordinated moves are more likely to be a management decision, while uncoordinated moves reveal more about individual agency and social influence within the team structure. When you watch movements through the lens of teams rather than individuals, a very different story emerges about how work gets done and how careers unfold.
To test their ideas, the researchers built a family of random null models that preserve different sets of facts about the organization. The strength preserving model keeps the flow of moves proportional to how much labor is being moved in and out of each team. The occupation transition model goes a step further by preserving not just where people move, but the kinds of jobs and skills involved in those moves. The third model, called the reuniting preserving model, keeps the eye on a social feature that emerges from the data itself: the tendency for moves to reunite former coworkers. By comparing real transitions to these models, the authors quantify how much social forces are actually shaping internal mobility versus how much is explained by supply demand or occupational niches.
Reuniting moves reveal social capital
The big reveal sits in what they call reuniting moves. In the AAW dataset, reunions are not a rare curiosity; they are a dominant feature of mobility, and they persist under careful statistical tests. The researchers define a reunifying move as one that creates a new connection between two teams in which people who previously worked together end up reuniting in a new setting. When you tally all moves across the nine year window, roughly a third of uncoordinated moves lead to reunions. That is not a marginal effect; it is a robust pattern with clear structural implications for how organizations should think about mobility and career pathways.
Two social factors strongly influence the likelihood of a reunion. First, the smaller the team two coworkers shared, the more likely they are to reunite later. Second, and perhaps more powerfully, the longer two people spent working together, the higher their chances of reuniting in the future. In other words, time together builds social capital that remains legible and influential long after the original collaboration ends. This mirrors the intuitive sense that trusted working relationships compound over months and years, but now we have empirical gravity wells showing how that force pulls people back together inside a large bureaucracy.
To make sense of the scale, the authors quantify reunions in a system wide way. They introduce two magnitudes that measure how well a given mobility model lines up with observed moves: a forward consistency that asks how well the model predicts future transitions given the past, and a reverse consistency that tests the logic in the other temporal direction. Across the trio of models, the reunion aware model consistently outperforms the others, and it outperforms even when the authors strip out the obvious cases where no moves happen at all. In short, reuniting behavior is not a fringe phenomenon; it is central to how a large organization mutates its own internal structure over time.
These results do more than refine a theory. They suggest that an internal labor market is not just about filling vacancies or matching people to roles; it is also a social ecosystem where the history of collaboration matters. Referrals, informal networks, and the reputational capital accrued by being part of a historically cohesive team become a kind of internal currency. The researchers point to referrals as a likely mechanism behind reunions, but the data also open the door to a broader question: how should organizations cultivate and benefit from these social dynamics without letting them distort merit and opportunity?
Why this matters for the future of work
What counts as a breakthrough in this paper is not just the discovery that people reunite with past co workers. It is the demonstration that social microdynamics — the day to day, month to month life of teams — can dominate the broad strokes of internal mobility. If you only modeled mobility by matching skill codes to vacancy postings you would miss a powerful, human layer that can explain a sizable portion of who ends up where inside an organization. The authors show that a model tuned to reunions captures much more of the observed transitions than models that preserve only labor market supply and demand or occupation switching. This is a reminder that people are not moving packages on a conveyor belt; they are moving within a social system that has memory, trust, and a social calculus about where they want to work and with whom.
From a practical standpoint, the work is a blueprint for how large organizations could enhance their internal mobility analytics. If HR analytics can map team structure and social histories with the same care as payroll or job classifications, organizations could forecast not only vacancies but also the social viability of transitions, the likely tax on team cohesion, and the probabilities that certain reunions will occur. That does not mean redesigning who you hire or who you promote into a vacuum; it means acknowledging that social capital is a measurable, consequential resource inside the talent landscape. In an era where talent is one of the scarcest assets, markets inside organizations could become as important as markets between organizations.
Of course the study centers on the Army Acquisition Workforce, a huge civilian workforce that operates with private sector style mobility under a vacancy system. The data come from a civilian subset of the US Army, and the authors are careful to frame their findings as a generalizable methodological approach rather than a one size fits all rule. The key takeaway is not a policy prescription for the Army alone but a new way to think about internal mobility as a social microdynamics problem. The lessons feel deeply relevant to any large organization that relies on teams for execution and learning — tech firms, research labs, government agencies, and multinational corporations alike.
In the end the authors argue for a new generation of vacancy systems that explicitly incorporate teams, time together, and the social fabric of coworker bonds. The social capital in a workplace is not just a soft asset; it is a channel through which information, referrals, and collaboration flow. The RSP model that foregrounds reunions provides a more accurate lens on how moves happen and why they cluster around familiar faces. If organizations begin to treat reunions as a first class citizen in their internal mobility thinking, they may unlock more productive teams, stronger networks, and career paths that feel less like hopping from one box to another and more like a designed journey through a living social ecosystem.
As the authors note, this is a step toward a broader theory that integrates team dynamics with workforce forecasting. The data show a stubborn pattern: people want to work with people they know, and the longer they have worked with them, the more that bond matters. If that remains true, the future of internal mobility may look less like a staircase and more like a web of long standing collaborations, where each move is not just a position change but a choice to re knit a familiar social fabric within a changing organization.
For now, the study stands as a reminder that inside the sprawling walls of large institutions, the human impulse to reconnect with old colleagues can be the quiet engine of career paths and organizational adaptation. The lead researchers at George Mason University and their colleagues have given leadership teams a language to talk about these forces and a toolkit to study them with the same seriousness with which they study budgets, headcounts, and skill inventories. The next decade of work in organizational analytics may well be measured in how adept we become at pairing people with teams and timing their reunions just as carefully as we time promotions and projects.
Ultimately this work invites a modest re framing: internal mobility should be understood as a social process as much as a labor market. The reunions that shape who moves where are not accidents of career luck; they are a structured, measurable feature of how teams live and evolve. If we can map that social terrain with the same clarity we map salaries and skill codes, we may unlock a future in which internal career paths feel more human, more predictable, and more humane at scale.
