This thesis explores the integration of transaction costs into portfolio optimization models, addressing a significant gap in traditional financial theory where such costs are often overlooked or simplified. By treating transaction costs as dynamic and volatile elements, this research enhances the realism and practical applicability of portfolio construction, leading to more efficient and effective investment strategies. The work is structured across three main chapters.

Chapter 1 introduces the concept of implementation shortfall variance and integrates transaction cost covariance into the mean-variance utility function. Through statistical modeling and simulations using S&P 500 data, it demonstrates how accounting for transaction cost variance can improve portfolio performance.

Chapter 2 extends the analysis to equity factor portfolios, a critical component in asset management. It presents a novel optimization
framework that reduces transaction costs while preserving the core characteristics of these portfolios. By applying this methodology to both single and multi-factor portfolios across global markets, the research shows that transaction-cost-optimized portfolios can achieve superior net returns, particularly in less liquid markets.

Chapter 3 continues this approach to develop factor-enhanced market portfolios optimized for transaction costs. By
integrating these factor strategies within a benchmark-relative context, the research provides actionable insights for asset managers, suggesting that traditional models which overlook or simplify transaction costs may be suboptimal.

Overall, this thesis makes significant contributions to the field of finance by emphasizing the importance of transaction costs in portfolio optimization and offering practical tools and models that can enhance portfolio performance in a realistic and cost-effective manner.