Guide

Active share explained

Harbor Capital's 2024 active equity manager review compared two large-cap growth funds in the same Morningstar category. Both charged 0.85% expense ratios, marketed stock-picking skill, and beat the Russell 1000 Growth index in three of the prior five years. Holdings overlap with the index looked high on a casual glance — but only one fund was a genuine stock picker. Fund A held 85 of the top 100 benchmark names at nearly index weights; its active share was 12%. Fund B ran a concentrated 45-name portfolio with meaningful sector tilts; active share was 78%. Fund A was a closet indexer: investors paid active fees for passive-like exposure. Trustees terminated Fund A, renegotiated the fee schedule on Fund B, and added active-share floors to the investment policy statement. Active share, introduced by Cremers and Petajisto (2009), measures how much a portfolio's weights differ from its benchmark — the static complement to tracking error, which measures how much returns differ over time. Together they answer: “Is this manager actually different, and does that difference earn its keep?” This guide defines the formula, interprets the 0–100% scale, connects active share to the information ratio and performance attribution, works a Harbor Capital manager review example, provides a metric decision table, lists pitfalls, and ends with an allocator checklist alongside our factor investing guide.

What active share measures

Active share answers a simple question: if I line up portfolio weights and benchmark weights name by name, how different are they? It is a holdings-based, point-in-time metric — not a return statistic. A fund can have high active share but low tracking error if its off-benchmark bets cancel out, or low active share but moderate tracking error if it overweights a few names that move sharply.

For a universe of N securities where both portfolio and benchmark hold non-zero weights (after mapping to a common investable set):

Active Share = ½ ∑_i=1^N |w_p,i − w_b,i|

where w_p,i is the portfolio weight and w_b,i is the benchmark weight in security i. The factor of one-half ensures the metric ranges from 0% to 100%: if every weight difference sums to 200 percentage points (full reversal), active share equals 100%.

Reading the scale

0–20%: Near-clone of the benchmark. Often labeled closet indexing when paired with active fees.
20–60%: Moderate differentiation. Many enhanced-index and factor-tilt strategies live here.
60–100%: High conviction. Concentrated stock pickers, sector rotators, and long-short equity often exceed 80%.

Cremers and Petajisto's empirical work found that U.S. mutual funds with active share above 60% and sustained for several years tended to outperform closet indexers before fees — but results vary by market, era, and asset class. The metric is diagnostic, not a guarantee of alpha.

Active share vs tracking error vs holdings overlap

Allocators often confuse three related but distinct measures:

Metric	Input	What it captures	Blind spot
Active share	Weights at a date	Static benchmark overlap	Ignores return correlation; snapshot can change next month
Tracking error	Return time series	Volatility of excess returns	Two portfolios can diverge in returns with similar weights if betas differ
Holdings overlap (%)	Names held in common	Count of shared securities	A fund can hold 90% of index names at tiny weights and still have high active share if weights differ

Active weight (sometimes called relative weight) is another useful companion: for each name, w_p,i / w_b,i shows whether you are 2× overweight or 0.5× underweight. Active share aggregates those differences; active weight highlights the few names driving conviction.

The information ratio divides mean active return by tracking error. High active share with a strong information ratio suggests differentiated bets that paid off. High active share with a negative information ratio suggests expensive differentiation that destroyed value — still better than paying active fees for a closet indexer with the same negative IR.

Closet indexing and the fee puzzle

A closet indexer charges active-management fees while maintaining benchmark-like weights. Investors get beta exposure they could buy for 3–10 bps in a passive index fund, plus unrewarded active risk from tiny overweights that rarely compound. Closet indexing persists because:

Career risk: PMs hug the benchmark to avoid firing in bad years.
Capacity: Large AUM forces managers toward benchmark weights to avoid moving markets.
Marketing: “Beat the index” narratives hide low active share on fact sheets that omit the metric.
Benchmark mismatch: A fund compared to the wrong index can look active when it is not (or vice versa).

Regulatory attention has grown: several European jurisdictions now require active share or similar overlap disclosure. U.S. SEC focus on “name brand” active funds with S&P 500 correlation above 0.95 has pushed allocators to audit overlap internally even when filings are vague.

When low active share is acceptable

Not every low-active-share product is a rip-off. Enhanced index strategies deliberately target 20–40% active share with explicit factor tilts and lower fees than traditional active. Transition portfolios temporarily hug benchmarks during manager changes. Multi-manager sleeves may show low active share at the sleeve level while the total fund is differentiated. The issue is misalignment: if you pay 80 bps for 15% active share, you should know you are buying mild tilts, not deep research alpha.

Harbor Capital manager review worked example

Harbor Capital's investment committee audited three U.S. large-cap equity managers in Q1 2024. Each reported against the Russell 1000. The committee computed quarterly active share from 13F-equivalent holdings (mutual funds from N-PORT; separate accounts from custodian tapes) and paired it with three-year tracking error and information ratio.

Manager	Active share	Tracking error	Information ratio	Fee	Verdict
Alpha Core	14%	1.2%	0.15	0.85%	Closet indexer; terminate
Beta Select	72%	4.8%	0.42	0.85%	Genuine active; retain, fee renegotiation
Gamma Enhanced	38%	2.1%	0.28	0.35%	Enhanced index; fit for core allocation

Alpha Core's top ten holdings matched the Russell top ten within 50 bps each — active share of 14% driven almost entirely by micro overweights in mega-caps. Attribution (see our attribution guide) showed positive selection in only two quarters; outperformance was beta drift when large caps led. Beta Select's 72% active share came from a 45-name portfolio, 18% cash in one quarter (benchmark is fully invested), and a 12% overweight to semiconductors. Tracking error of 4.8% was intentional; IR of 0.42 justified the risk budget. Gamma Enhanced at 38% active share and 35 bps was repriced as a factor tilt sleeve, not a stock-picker.

The committee added policy language: active mandates must maintain active share ≥ 50% on a rolling four-quarter average or move to a lower fee tier. Enhanced sleeves are capped at 45% active share with explicit factor benchmarks.

Computing active share in practice

Data alignment

Garbage in, garbage out. Before applying the formula:

Map ADRs and share classes to a single parent security.
Handle cash and derivatives: Treat cash as zero benchmark weight or use a cash-adjusted benchmark.
Rescale weights so portfolio and benchmark each sum to 100% on the same investable universe.
Include shorts for long-short funds (negative weights increase active share correctly).
Match benchmark version (Russell 1000 vs Russell 1000 Growth) to the fund's stated objective.

Frequency and stability

Active share jumps when managers trade earnings seasons or rebalance. Report quarterly snapshots plus a trailing four-quarter average. Sudden drops may signal style drift toward indexing after bad performance — a red flag for process discipline. Pair with turnover: high active share with 200% annual turnover may be churn, not conviction.

Multi-asset and bond portfolios

The formula extends to bond indexes (sector weights, duration buckets) and multi-asset blends if you define a composite benchmark with consistent weight mapping. Illiquid holdings without benchmark analogs require proxy weights or exclusion with documented rationale.

Metric decision table

Question you are asking	Best metric	Why
Is this fund a closet indexer?	Active share	Direct weight divergence from benchmark
How volatile are excess returns?	Tracking error	Return-based risk budget
Did active bets earn their risk?	Information ratio	Return per unit of tracking error
Where did alpha come from?	Brinson / security attribution	Decomposes allocation vs selection
How many shared names?	Holdings overlap count	Quick screen; insufficient alone
Is the manager trading too much?	Turnover + implementation shortfall	Links activity to costs

Common pitfalls

Using the wrong benchmark — a value fund measured against a growth index shows artificial active share.
Ignoring cash and derivatives — unhedged futures can inflate or deflate active share versus a cash benchmark.
Single-date snapshots — one quarter's 80% active share may follow a year at 25%; use rolling averages.
Confusing overlap count with active share — 95% name overlap can still mean 40% active share if weights differ.
Expecting high active share to guarantee alpha — differentiation must be skillful, not random.
Neglecting capacity — a PM may legitimately compress active share as AUM scales; fees should compress too.
Comparing across asset classes — EM active share distributions differ from U.S. large-cap norms.

Allocator checklist

Compute active share quarterly on the fund's stated benchmark universe.
Require four-quarter rolling average active share in manager reporting.
Pair active share with tracking error and information ratio on the same report.
Flag active share below 30% on mandates charging > 50 bps as closet-index candidates.
Verify benchmark version matches prospectus (not a convenient peer index).
Reconcile cash, ADRs, and derivative exposures before calculating weights.
Review attribution to see whether differentiation is allocation or selection.
Compare fees to active share tier: enhanced index fees for < 45% active share.
Monitor active share drift after poor performance years (career-risk hugging).
Document termination triggers in the investment policy statement.

Key takeaways

Active share = half the sum of absolute weight differences vs the benchmark, ranging 0–100%.
It measures static conviction; tracking error measures return divergence over time.
Closet indexers charge active fees for benchmark-like portfolios — active share exposes them.
High active share alone is not alpha; pair with information ratio and attribution.
Policy floors, rolling averages, and fee tiers align incentives with genuine differentiation.