Professional Esports International Migration Panel (League of Legends), 2009-2026

A complete-observation panel of international labour migration in professional League of Legends esports: 20,648 players, 85,869 team-to-team moves, and 2,550 cross-region international moves, observed daily from 2009-10-01 to 2026-07-22. Research on high-skilled migration has a standing problem: the productivity of the migrant is not observed. Wages stand in as a proxy, and selection bias is fought rather than solved. This market is different. The timing of every move is recorded to the day, nationality is observed separately from residency, productivity is observed per match, and career termination is dated. Location-specific human capital also approaches zero, since the game is identical software everywhere, making esports the limiting case of the claim that occupations with little location-specific human capital show the largest migration responses. Over seventeen years, 13 foreign nationals joined a Korean professional roster against 420 Korean nationals departing for foreign rosters, a ratio of roughly 30 to 1. The median stay of a foreign entrant is 11.2 months; no observed stay exceeded five years. Korean permanent residency requires five years of continuous residence, and the league exempts from its import quota only players holding citizenship or permanent residency. The threshold is longer than the career. The release includes a frozen alias-resolution snapshot. Nationality is missing for 17.9% of move events, and in the inbound cell this is decisive: of 30 unresolved players, 27 turned out to be Korean nationals returning home rather than foreigners arriving. Without that normalization the dataset overstates foreign inflow threefold. Every adjudication is recorded per case, and superseded figures are retained with the reason each was withdrawn. Derived from Leaguepedia (lol.fandom.com) via Special:CargoExport. Reproducible with the Python standard library; no third-party dependencies, no API key, no access restriction. Birth dates are truncated to birth year in the derived panel. See docs/known_issues.md in the archive for limitations, including right-censoring of stay durations and uneven wiki coverage.