sanghabahn/K-ETNIC-Appendix: Online Appendix

Online appendix for the master's thesis "Less Visible Peer Networks and Stock Return Predictability: Text-based Evidence from Korea" (Sangha Bahn, Chung-Ang University, 2026). The dataset (peer_groups_2024_fixedbase2000.csv) lists the ten nearest text-based peers for all 2,233 firms listed on KOSPI/KOSDAQ in 2024. Peers are identified by cosine similarity in the K-ETNIC industry space — Doc2Vec embeddings of Korean annual-report business descriptions (사업보고서), trained on fiscal-year 2000 reports and frozen (the paper's headline model). Each peer is flagged as same- or different-KSIC (two-digit) relative to the focal firm; 56.7% of all peer slots are cross-industry. This is the complete listing summarized in Appendix III of the thesis. See the README for column definitions.