emergence of power-law and other wealth distributions in crowd of heterogeneous agents - jake xia.pd
Academic paper by Jake J. Xia proposing a feedback-loop multi-agent model in which agents’ heterogeneous response functions (Bi) and wealth/power (Ai) co-evolve via a reward function, producing concentrated distributions including power-law, lognormal, and other concentrated forms. The model links synchronization and trend-following to emergence of ‘super-agents,’ compares US vs Japan income/wealth distributions, and argues lower-tail PDF behavior is critical to distinguish distribution types.
Questions this document answers
Jake J. Xia; Harvard Management Company; Research Affiliate at Massachusetts Institute of Technology; email jjxia@mit.edu.
A 2010 working paper (revised edition).
US: 0.45; Japan: 0.33.
U.S. Census Bureau, Table A-4a (2023).
2022 Richmond Fed survey (presented via Federal Reserve Economic Data, St. Louis Fed) for 2024 Q2 update.