Portfolio Theory and Cone Optimization

This paper will discuss portfolio optimization, Quadratic Programming (QP) and Second Order Cone Programming (SOCP). We will use simulated and empirical data to compare the two optimization routines. Daily data for SP500 stocks from 2005 to 2010 was used to show that a 20-days rebalanced portfolio strategy with an expected portfolio return of 60 percent of the maximum expected return for all stocks produced an 8.4 percent return premium on an annual basis if we used QP and 11.2 percent return premium on an annual basis if we used SOCP