The Determinants of Self-Employment - Evidence from Self-Employed Persons in Poland

Purpose: The primary purpose of the research is to build a model that makes it possible define categories entities through which self-employment used as form tax optimization. This would facilitate creation more efficient instruments prevent false without discouraging real entrepreneurship. Design/Methodology/Approach: method survey carried out from September 2020 on representative sample 400 Polish self-employed persons representing specific trades. Subsequently, machine learning classification algorithms were implemented find factors characterizing with different tax-optimization attitudes. Findings: results have shown and regression tree (CART) bootstrap aggregation (bagging) models can be helpful determine whether person likely use optimization method. three main determining are person’s attitude towards risk, industry in they work, length their business experience. Practical Implications: performed made identify affect people’s reducing burden, particular, verify scope taxpayers becoming reduce burden (treating optimization), there way shape rules encourage entrepreneurship while minimizing risk abuse such forcing employees become or people registering avoid taxation. Originality/Value: I proposed opposite than pervious approach: this among extend knowledge about toward self-employed. approach should understanding process individual decision-making between an employee