Designing the Ruler: Why Benchmark Construction Deserves Committee-Level Scrutiny

When an institutional investor hires an active manager, the conversation typically centers on the manager: track record, team turnover, capacity, fees. The benchmark, by contrast, is treated as plumbing—necessary, technical, and largely settled. That assumption deserves retirement. Across US endowments, public pensions, and corporate plans, benchmark customization has become a quiet negotiation in which the measuring stick is reshaped to flatter the runner. The result is a growing class of mandates where underperformance is not merely tolerated but structurally designed out of the evaluation.

The Rise of the Bespoke Index

The institutional toolkit has expanded well beyond the S&P 500 and the Bloomberg Aggregate. Index providers now offer a menu of exclusions, tilts, and overlays that allow sponsors to strip out sectors, factors, or geographies with the ease of a subscription setting. Some of these refinements are genuinely principled—carbon screens, tobacco exclusions, and faith-based mandates reflect legitimate policy choices made at the governance level.

The problem arises when exclusions are made not by the board, but by the manager, and not for policy reasons but for performance reasons. A manager struggling in a lagging sector has a rational incentive to propose a benchmark that omits it. Once the sector is removed, the manager is no longer judged against the very environment that produced the weakness. The alpha appears—because the denominator changed.

How Carving Becomes Strategy

Consider the mechanics. A large-cap equity mandate benchmarked to the full Russell 1000 underperforms for three years, dragged down by a poorly timed energy allocation. In a renegotiation, the manager proposes a custom composite: the Russell 1000 ex-Energy, rebalanced quarterly. The sponsor's staff, eager to retain an otherwise capable team, agrees. On paper, the manager's relative results improve immediately—without a single change to the portfolio.

Multiply this across sectors, factors, and geographies, and the benchmark becomes a moving target calibrated to each period's weaknesses. In the limit, a sufficiently carved benchmark guarantees that any strategy will look competent. The measurement exercise inverts: instead of the benchmark judging the manager, the manager's history shapes the benchmark.

There is a second, subtler cost. Carved benchmarks corrupt the sponsor's total-portfolio arithmetic. Asset allocation assumes each sleeve is measured against a common, investable standard. When sleeves are judged against bespoke composites, the sponsor can no longer tell whether aggregate active risk is being compensated or merely obscured. The board sees a mosaic of managers each beating their reference—while the overall fund underperforms its policy benchmark.

Questions Boards Should Ask Before Signing

Benchmark design warrants the same diligence as manager selection, because in practice the two are inseparable. Before approving a customized reference, committees should insist on clear answers to the following:

• Who proposed the exclusion? If the manager initiated it, treat the request as a performance negotiation, not a technical amendment.
• Is the exclusion investable? A benchmark the sponsor cannot actually own—because the excluded sector remains a legitimate opportunity set—distorts the active risk budget.
• What did the manager's history look like against the uncarved index? Restate the full track record both ways. If the improvement is an artifact of the carve, that fact belongs in the diligence memo.
• Does the exclusion serve policy or performance? Screens adopted by the board for values or risk reasons are legitimate. Screens adopted to improve relative optics are not.
• What is the sunset provision? Custom benchmarks should carry a review date and a pre-agreed reversion to the standard index if the rationale erodes.

Restoring the Ruler's Integrity

None of this argues against customization. Sophisticated sponsors should absolutely tailor benchmarks to reflect genuine policy constraints, liability structures, and values. The discipline lies in who decides and why. Exclusions should originate in the investment committee, be documented with the same rigor as an allocation decision, and apply prospectively—not retroactively to rescue a struggling mandate.

Some institutions have begun requiring that any proposed benchmark change be evaluated by a party independent of the manager relationship, much as audit committees firewall auditor selection. Others mandate dual reporting: every carved mandate is also tracked against its unadjusted parent index, with both lines appearing in board materials. The second line costs nothing and reveals a great deal.

The deeper principle is simple. A benchmark is a promise about what success means. When that promise can be rewritten by the party being measured, it stops being a measurement and becomes a marketing document. Boards that treat benchmark construction as a governance decision—rather than a technical footnote—protect not just the integrity of individual mandates, but the coherence of the entire portfolio's account of itself. The ruler should never be drawn by the hand it measures.