Portfolio Optimization with Transaction Costs

Investors often update their portfolios at regular time intervals by trading stocks, but there are costs associated with these trades. This project seeks to limit these transaction costs by controlling the portfolio turnover (absolute change as a fraction of book size) between time periods. The result is a multiperiod optimization problem with quadratic objective function and non-smooth constraints. The resulting portfolios outperformed benchmark portfolios in both expected utility and actual portfolio value.