specifically in large-cap value securities. Like many asset-management organizations, the manager-selection group seeks to isolate the right style of investment using both measurable metrics and intangible considerations. Measurable metrics include the manager's official database classification and quantitative screening; the main qualitative consideration is a subjective overlay by the investor based on industry knowledge and subjective interpretation of the quantitative data. The manager's self-classification should be used, but not to the exclusion of other metrics. A manager might believe that its investment style is, for example, truly large-cap value, but that manager's definition of value may differ dramatically from other value managers or, in certain cases, from the basic characteristics of traditional value benchmarks. The quantitative screening typically eliminates this type of manager from the analysis. At this stage quantitative screens embrace a number of factors, including style consistency, risk-adjusted performance over rolling periods of time, performance consistency, downside risk for drawdown) analysis, and substyle analysis (to differentiate, for example, a deep value manager from a relative value manager). These screens are analogous to an asset manager's search of databases for companies meeting specific return on equity, net income, and valuation criteria. There are two key factors that the manager-selection team looks for in identifying successful managers from a quantitative perspective: superior risk-adjusted performance in various market environments and over a full market cycle and consistent results relative to an appropriate benchmark. Any quantitative screen, no matter how powerful and robust, should serve only as a starting point; for both asset managers and manager-selection groups, quantitative pitfalls abound. For manager pickers, survivorship bias (the tendency of database providers to expunge the entire track record of products that shut down) represents a significant drawback of manager databases. Furthermore, pure quantitative screens fail to provide fundamental information about the investment manager, the experience of the team, the team's length of time at the firm, the depth of firm resources, the ownership structure of the firm, and so forth. A returns-based screen will also exclude information on how the track record was built-for example, the assets under management. These pitfalls speak to the need for a qualitative overlay to the screening process. Often asset-management teams will qualitatively add and delete companies from their screens based on their history with the management team, prior experience investing in the security, or industry knowledge. Similarly, the manager-selection process should allow for making subjective additions to the list and deletions from the screened list of managers. Such qualitative overrides could derive from any number of sources: There may be a new team from a different organization taking over a bad track record that should be added to the list of managers. There may be a manager with an outstanding track record that screened very well but lost several key investment professionals and should therefore be eliminated from the list. A great team with a solid track record might be housed within a parent company that the manager-selection team knows to be ineffective at retaining talent; such a manager might be eliminated from the process. In summary, a good manager-selection screening process, like a good asset-management investment process, should take into account both observable, objective performance and risk criteria and less formal but equally important qualitative considerations. Managers that