Co-invention networks, bridge inventors, and inter-firm mobility
Ch. 21Inventor Mechanics
Innovation is shaped not only by what is patented but by the people who patent it — how inventors collaborate, how they move between firms, and how they migrate across borders. This chapter examines the interpersonal and inter-firm dimensions of inventor activity: the co-invention networks that reveal stable research teams, the bridge inventors who connect otherwise separate organizations, and the mobility patterns through which talent and tacit knowledge circulate across firms, states, and countries.
Inventor Co-Invention
In addition to organizational partnerships, individual inventors form collaborative networks. When the same inventors repeatedly appear together on patents, the pattern indicates stable research teams that persist across projects and years. These co-invention ties constitute some of the most productive relationships in the innovation ecosystem.
Figure 1
632 Prolific Inventors Form 1,236 Co-Invention Ties in Fragmented Team Clusters
Co-invention network among prolific inventors, with edges representing shared patents and node size indicating total patent count.
Loading visualization…
Co-invention network among inventors with significant collaboration ties. Edges represent shared patents; node size indicates total patent count. The network is more fragmented than the organizational co-patenting network, with many small, tightly connected teams.
The increasing connectivity of the co-invention network is consistent with potentially faster knowledge diffusion, though it may simultaneously create path dependencies in innovation direction.
Collaboration Network Structure
The structure of innovation collaboration has evolved considerably over the study period. By analyzing co-inventor relationships as a network, it is possible to measure the connectedness of the innovation ecosystem. The average degree (number of collaborators per inventor) and network density indicate whether innovation is becoming more or less collaborative over time.
Figure 2
Average Inventor Degree Rose 2.5x, From 2.7 in the 1980s to 6.8 in the 2020s
Average co-inventors per inventor and average team size by decade, measuring collaboration network connectivity
Loading visualization…
Summary statistics of the inventor collaboration network by decade. Average degree measures the typical number of co-inventors per active inventor. Both metrics exhibit sustained increases, indicating a progressively more collaborative innovation ecosystem.
Rising average inventor degree reflects both larger team sizes and more extensive cross-organizational collaboration, consistent with a more interconnected innovation network over time.
Bridge Inventors
Some inventors serve as critical bridges connecting otherwise separate organizations and technology communities. These "bridge inventors" have patented at 30 or more distinct organizations, potentially facilitating the transfer of knowledge and practices between firms.
Inter-Firm Mobility
Beyond the structure of co-invention networks, the movement of inventors between organizations, across state lines, and across national borders constitutes a parallel channel for knowledge diffusion. Tracking these mobility patterns reveals how talent and tacit knowledge circulate through the innovation ecosystem.
Talent Flows Between Companies
When inventors file patents at different organizations over their careers, they create The movement of inventors between organizations, tracked by consecutive patent filings with different assignees. Net talent flow reveals which companies are gaining vs. losing inventive talent. that may transfer knowledge between companies. The chart below shows the 20 organizations with the highest total inventor movement (among 50 in the dataset), distinguishing between inventors gained and inventors lost.
Figure 3
143,524 Inventor Movements Flow Among 50 Major Patent-Filing Organizations
Inventor movements into and out of major patent-filing organizations (top 20 by total flow shown), based on consecutive patents with different assignees within 5 years.
Loading visualization…
Movement of inventors into and out of major patent-filing organizations, based on consecutive patents with different assignees (gap of 5 years or fewer). Organizations are ranked by total flow (inflow + outflow). The bidirectional nature of many flows suggests active talent cycling within industry clusters.
Large technology companies tend to be net talent importers, drawing inventors from smaller firms and universities. The bidirectional nature of many flows is consistent with active talent cycling within industry clusters.
International Inventor Mobility
Inventor mobility data reveal how researchers and engineers move across national borders over the course of their careers. International mobility rates have risen steadily and now surpass domestic interstate rates, reflecting the increasingly global nature of innovation talent flows.
Domestic Inventor Mobility
The state and city patent output documented in the Domestic Geography chapter capture where innovation occurs, but not how inventors move across these domestic regions over the course of their careers. Tracking individual inventors across their patent histories reveals patterns of geographic mobility — the manner in which innovators relocate between states, carrying tacit knowledge and professional networks with them.
Global Inventor Migration Flows
Beyond country-level filing volumes, the map below visualizes the dominant cross-border migration corridors that connect national innovation ecosystems. The United States emerges as the central node, linking East Asian, European, and other innovation systems through flows of researchers and engineers.
Inventor Mobility: Event-Study Evidence
The preceding sections document the volume and direction of inventor mobility, but do not address its consequences for individual productivity. An event-study design estimates changes in patenting around the timing of firm moves by tracking each inventor's output from t-5 to t+5 years relative to the move event. By centering outcomes on the move year, this approach controls for secular trends and life-cycle effects, providing estimates of how changing organizational context is associated with citation impact. The panel includes inventors who moved exactly once between top-50 assignees, with at least two patents in both the pre- and post-move windows.
Figure 4
Citation Impact Dips at the Time of Move, Then Recovers Within 2 Years
Mean 5-year forward citations by year relative to firm move (t=0), with 95% confidence interval, for inventors who moved once between top-50 assignees.
Window5y
Loading visualization…
Event-study design tracking inventor citation impact from t-5 to t+5 around a firm move. The shaded band represents the 95% confidence interval. The dip at t=0 reflects the disruption of changing organizational context; the recovery by t+2 suggests that mobile inventors adapt quickly and may benefit from access to new knowledge and resources.
The transient dip in citation impact around the move year, followed by full recovery within two years, suggests that inter-firm mobility is associated with short-term productivity dips followed by recovery. Post-move citation outcomes are on average comparable to or higher than pre-move levels, potentially reflecting exposure to new organizational knowledge and resources.
Figure 5
Moves to Higher-Quality Firms Show Highest Post-Move Citations: 3.4 (Up) versus 3.2 (Down and Lateral)
Mean 5-year forward citations by year relative to move, stratified by move direction (up to higher-quality firm, lateral, or down to lower-quality firm).
Loading visualization…
Event-study results stratified by the direction of the move relative to firm quality (measured by mean citation impact). Moves to higher-quality firms show the largest post-move gains, while moves to lower-quality firms exhibit a more muted recovery, consistent with organizational context being associated with inventor productivity.
The direction of the move matters: inventors who move to higher-quality firms experience the largest post-move citation gains, consistent with organizational resources and peer quality being associated with individual inventor productivity.
Figure 6
Network Position and Citation Impact Among 4,935 Top Inventors
Mean 5-year citations and degree centrality by quintile, 4,935 most prolific inventors
Loading visualization…
Average 5-year forward citations and network degree centrality by quintile among 4,935 top inventors. Higher centrality quintiles correspond to inventors with more co-invention connections.
International Mobility: Rates and Flows
A deeper look at international inventor mobility reveals both the aggregate rate of cross-border movement over time and the specific bilateral corridors through which inventors circulate. The following analyses track the share of inventors who change country across successive 5-year windows and identify the dominant country-pair flows.
The collaboration networks and mobility patterns documented in this chapter reveal the human infrastructure of innovation — the teams, bridges, and talent flows through which knowledge circulates. The next chapter, Geographic Mechanics, examines the spatial dimensions of these dynamics: how cross-border collaboration has evolved, how bilateral innovation corridors like the US-China partnership have grown, and how emerging technologies diffuse geographically from pioneering cities to secondary hubs.
Data coverage: January 1976 through September 2025. All 2025 figures reflect partial-year data.