Portfolio Selection: An Analytic Approach for Selecting Securities from a Large Universe
探讨在收益率不完全相关时，如何利用现代投资组合理论通过分散化降低风险，并回顾了Markowitz和Sharpe的模型，指出当前实施中仍存在的问题。
Where rates of return are perfectly correlated, risk reduction through diversification cannot be achieved. Where rates of return are less than perfectly correlated, however, then, to the extent that these interrelationships can be known, modern portfolio theory provides a framework in which risk reduction through diversification can be achieved. Markowitz was the first to give rigorous content to the concept of portfolio diversification [14], and to introduce a formulation for treating portfolio selection as a mathematical optimization problem. In order to facilitate application of his own covariance approach, Markowitz first suggested [15, pp. 96–101], and Sharpe later developed a market model formulation according to which it is assumed that the rates of return on various securities “are related only through common relationships with some basic underlying factor” [18, p. 281]. More than 25 years have passed since Markowitz introduced his original formulation, and the literature dealing with the portfolio selection problem that he identified has grown considerably since then. Unfortunately, many problems remain which prevent full and effective implementation of this framework for investment analysis.