The allure of putting together a portfolio of undervalued gems has long fascinated, and often frustrated, stock-market investors. A common idea is that astute analysis and emotional fortitude can help achieve this goal while the subsequent failure of many of these gems is accepted as simply part of the unpredictable changes in the market and human nature.

Most investors—whose time, interests and resources fall short of the requisite commitment to succeed at stock-picking—sensibly select a professionally managed account to assume that role. But in cases of underperformance, successful manager-picking, like stock-picking, often seems quite random.

One consideration is whether periods of underperformance are inevitable. Many studies have shown that managers whose long-term performance is top-quartile will go through interim bouts of underperformance, and this periodic shortfall can be quite severe. Understanding underperformances is something that should be explored well before the fact.

One reason for manager underperformance is that strong markets are often characterized by rallying large-cap stocks, which in turn feed into the momentum of capitalization-weighted indices. So if a manager makes bets outside the large-cap components of an index, relative performance may lag during a rally.

But, subjecting a manager to tighter guidelines and index-tracking constraints may also hurt performance, especially in the light of fees and transaction costs. Also, passive index-like portfolios assume the risk exposures of the market at a given point in time, which may not be suitable for every investor.

Given all of these factors, deciding how to evaluate manager performance and how tightly to benchmark performance should be driven by several considerations. Here are a few things to keep in mind.

The potential to capture value: Opportunity in a given asset class can be expressed qualitatively (e.g., micro-cap stocks are under-researched) and also measured quantitatively (e.g., diversity of returns within a given asset class). The greater the perceived inefficiency within an asset class, the greater the potential for adding value through active management.

Confidence in a specific manager: Confidence requires understanding the philosophy and methodologies a manager uses to exploit opportunities. A thorough analysis of historical portfolio holdings and returns provides evidence of how well investment implementation matches up with a manager's advertised philosophy and process. Basically, this step comes down to separating skill from luck.

Aggregate portfolio exposure: Just as diversifying across stocks may reduce risk, using a variety of complementary active managers offers additional safety. But grouping active managers in a simplistic way may compound risk through multiple or unintended biases. Managers who rely on similar tools, metrics and investment philosophy may duplicate exposures to broad market drivers such as energy pricing, interest-rate changes, currency and market momentum. Considering individual manager performance within the context of an overall manager mix requires experienced judgment and comprehensive statistical analysis of aggregate portfolio exposures.

The complex exercise of performance evaluation should include a strong focus on understanding the drivers of individual manager performance and the interaction of manager portfolios over time. Only then will the expectation and potential of actively managed portfolios be fully realized.

Tim Knepp, CFA,serves as Chief InvestmentOfficer of Encino, Calif.-based Genworth FinancialAssetManagement (a unit of GenworthFinancial), as well as chairman of the firm's investment management executive committee. He can be reached at tim.knepp@genworth.com.