Machine Learning Approaches to Detecting Labor Market Monopoly in Platform Economy

As the platform economy expands across various sectors, growing concerns arise about potential monopolistic behaviors and labor market power imbalances. Platform companies use sophisticated algorithms to control key business decisions, including worker assignments and compensation, which can mask unfair practices targeting workers who have few alternative employment options. This paper explores a unified machine learning framework for detecting labor market monopoly in platform-based ecosystems. We employ a combination of econometric and clustering methods, including Difference-in-Differences (DID) and Gaussian Mixture Models (GMM), to analyze wage trends, identify susceptible subgroups, and evaluate whether algorithmic interventions depress worker earnings below competitive levels. While we illustrate our approach through a synthetic case study involving three cities, the methodology is broadly applicable to diverse platform markets. Our results expose significant wage suppression in treatment cities after an algorithmic shift and reveal heterogeneous impacts across worker segments. The findings highlight how machine learning approaches can uncover concealed monopsonistic behaviors, offering actionable insights for policymakers, regulators, and researchers seeking to foster transparency and fairness in the digital labor market.