Guide

Calmar ratio explained

Two funds both return 10% per year with identical Sharpe ratios. Fund A never drops more than 8% from a peak; Fund B suffers a 35% drawdown on the way there. For a pension board that must avoid forced selling during crises, those paths are not equivalent. The Calmar ratio (named after Terry Young's California Managed Accounts Reports newsletter, also called the MAR ratio) compresses return and worst-case pain into one number: annualized return divided by maximum drawdown. Commodity trading advisors (CTAs), hedge fund allocators, and risk officers use it when tail losses matter more than day-to-day volatility. This guide defines maximum drawdown, walks through the Calmar formula and common lookback windows, compares Calmar to Sharpe, Sortino, and information ratio, works a Harbor Capital trend-following sleeve example, lists pitfalls, and provides an allocator checklist alongside position-sizing discipline.

Maximum drawdown: the denominator that matters

Maximum drawdown (MDD) is the largest peak-to-trough percentage decline in a portfolio's equity curve over a measurement window. If a fund rises from $100M to $140M, then falls to $98M before recovering, the drawdown from the $140M peak is (140 − 98) / 140 = 30%. Maximum drawdown records the worst such episode in the period — not average volatility, not a one-day loss, but the deepest hole an investor who bought at the wrong peak would have endured.

Drawdowns matter because they compound asymmetrically: a 50% loss requires a 100% gain to break even. Large drawdowns also trigger real-world constraints — margin calls, redemption gates, career risk for managers, and behavioral panic selling by clients. For a deeper treatment of measurement and recovery math, see maximum drawdown explained.

Peak-to-trough vs rolling drawdown

Calmar traditionally uses the single worst peak-to-trough episode over a fixed window (often 36 months). Some practitioners report rolling 12-month drawdowns or average drawdown instead — those are different metrics and should not be mixed into a Calmar calculation without relabeling. Always confirm whether MDD is gross or net of fees and whether it includes the full track record or only the post-inception window after a strategy change.

The Calmar ratio formula

The standard definition over a lookback period (commonly 36 months):

Calmar ratio = annualized return / |maximum drawdown|

Both numerator and denominator must cover the same window. If you annualize three years of cumulative return as (1 + R_total)^1/3 − 1, the maximum drawdown should be the worst peak-to-trough within those same three years — not the all-time worst drawdown from a longer history unless you extend the return window to match.

Worked arithmetic

Suppose a fund gains +8%, +12%, and −4% over three calendar years (compounded total ≈ +17.7%). Annualized return ≈ +5.6%. Its worst peak-to-trough in that window was −18%. Calmar ≈ 5.6 / 18 = 0.31. A second fund with the same annualized return but only −9% max drawdown scores Calmar ≈ 0.62 — twice the drawdown efficiency on identical headline return.

Lookback windows

Three-year Calmar is the industry default for CTAs and many hedge fund databases. Five- and ten-year windows smooth luck but may bury a recent risk-model blowup. Rolling 36-month Calmar charts show whether crisis performance is improving or deteriorating — a point estimate without context can hide a strategy that was excellent pre-2020 and fragile since.

How to interpret Calmar values

Unlike Sharpe, Calmar has no clean theoretical distribution, so thresholds are empirical heuristics from allocator practice:

Below 0.25 — return is thin relative to worst drawdown; crisis pain likely dominates the investor experience.
0.25 – 0.50 — acceptable for diversified multi-strategy books; common for equity-heavy portfolios after a bad cycle.
0.50 – 1.0 — strong drawdown efficiency; typical of well-run trend followers or balanced funds in favorable regimes.
Above 1.0 — exceptional on paper; verify window length, leverage, and whether MDD reflects a full stress cycle.

A Calmar above 1.0 over only 18 months that missed a known crisis year is marketing, not evidence. Pair any headline Calmar with the date range, net vs gross returns, and whether the maximum drawdown occurred at the start or end of the window (end-of-sample drawdowns can depress Calmar even when recent returns recovered).

Calmar and leverage

Leverage scales returns and drawdowns together in idealized models, leaving Calmar roughly unchanged if volatility targeting is perfect. In practice, leverage introduces gap risk, funding costs, and nonlinear liquidations — so two funds with identical unlevered Calmar can diverge sharply once margin and capacity constraints bite. Always inspect gross exposure and worst single-month losses alongside Calmar.

Calmar vs Sharpe, Sortino, and information ratio

Metric	Numerator	Denominator	Best when
Calmar	Annualized return	Maximum drawdown	Crisis-sensitive allocators, CTAs, drawdown-constrained mandates
Sharpe	Excess return vs risk-free	Total volatility (σ)	Broad fund comparison, mean-variance optimization inputs
Sortino	Return above MAR hurdle	Downside deviation	Asymmetric return profiles, retiree spending floors
Information ratio	Active return vs benchmark	Tracking error	Evaluating active equity managers vs an index

Sharpe penalizes upside volatility equally with downside — a fund that spikes +30% in a good month looks worse on Sharpe than one that grinds +2% monthly, even if both have similar drawdown profiles. Calmar ignores path volatility entirely and focuses on the single worst hole. A strategy can show Sharpe 1.2 and Calmar 0.3 if returns are smooth until a sudden −40% event — exactly the pattern that breaks VaR models tuned on recent calm data.

For benchmark-relative mandates, use information ratio for skill and Calmar for absolute crisis tolerance — institutional portfolios need both lenses.

Worked example: Harbor Capital trend sleeve

Harbor Capital runs a 12% risk-budget allocation to a managed-futures trend sleeve alongside core equities and bonds. Over the 36 months ending Q1 2026 (net of fees), the sleeve delivered:

Compounded return: +22.4% total → annualized ≈ +6.9%
Maximum drawdown: −11.2% (mid-2024 equity-correlation spike)
Calmar ≈ 6.9 / 11.2 = 0.62

The core equity book over the same window: annualized +9.1%, max drawdown −19.4%, Calmar ≈ 0.47. Sharpe favored equities slightly (0.78 vs 0.71) because monthly equity volatility was lower until the drawdown cluster — Calmar flagged the trend sleeve as more crisis-efficient per unit of worst loss.

Allocator decision

Harbor's investment committee set a minimum Calmar floor of 0.40 on any diversifying alternative before increasing weight above 10%. The trend sleeve cleared the hurdle; a long-volatility overlay that showed Calmar 0.18 on the same window was held at a 2% experimental weight pending a longer track through a rates shock. The committee also required rolling 36-month Calmar charts in quarterly reports — not a single point-in-time number from the best three-year slice in fund history.

Decision table: when Calmar is the right metric

Your question	Start here	Also check
Can this CTA justify its slot in a crisis-aware portfolio?	36-month Calmar net of fees vs 0.5 hurdle	Worst monthly loss, correlation to equities in drawdowns
Fund A vs Fund B with similar Sharpe	Side-by-side Calmar on identical window	Recovery time after MDD, tail risk (CVaR)
Levered strategy marketing deck	Calmar on net returns including financing costs	Unlevered Calmar, margin call history
Retirement drawdown sustainability	Calmar of withdrawal-adjusted portfolio	Sequence-of-returns risk, safe withdrawal rate
Crypto or venture allocation	Calmar with full-cycle window including −70% episodes	Sharpe is often misleading on fat-tail assets

Who uses the Calmar ratio

CTA and managed-futures databases

BarclayHedge, NilssonHedge, and similar CTA indices publish Calmar alongside Sharpe because trend programs earn lumpy returns and occasional extended flat periods — maximum drawdown captures the pain of whipsaw regimes better than monthly σ alone.

Fund-of-hedge-funds due diligence

FoHF analysts rank managers by Calmar within style buckets (equity long/short, global macro, event-driven) before diving into factor exposures. A macro fund with Calmar 0.9 and Sharpe 0.6 may be preferable to a market-neutral fund with Sharpe 1.1 and Calmar 0.35 if the allocator's pain threshold is drawdown-driven.

Risk parity and volatility-targeting shops

Portfolios that scale exposure to hit a volatility target still face drawdown clusters when correlations spike to one. Calmar validates whether vol targeting actually improved crisis outcomes or merely smoothed monthly returns before a correlated crash.

Common pitfalls

Mismatched windows — annualized five-year return divided by three-year MDD inflates or deflates Calmar arbitrarily.
Survivorship bias — databases drop blown-up managers; published Calmar medians look better than live allocator experience.
Short track records — three calm years can produce Calmar above 1.0 that collapses in the first real stress event.
Gross vs net — gross Calmar overstates skill; fees widen drawdowns in down years.
Ignoring recovery time — two funds with identical MDD but different recovery speeds impose different liquidity and psychological costs.
Single-episode luck — one avoided crash from cash timing creates heroic Calmar that does not repeat.
Illiquid marks — private credit or PE funds with stale NAV understate drawdowns and overstate Calmar until a repricing event.
Replacing Sharpe entirely — Calmar ignores path risk between peaks; use both metrics plus tail measures.

Allocator checklist

Define Calmar lookback (typically 36 months) in the investment policy statement.
Compute on net returns with numerator and denominator over the identical window.
Report rolling 36-month Calmar quarterly, not only inception-to-date.
Pair Calmar with Sharpe, Sortino, and maximum drawdown duration (months to recover).
Stress-test by shifting the window to include the worst known crisis (2008, 2020, 2022).
Compare Calmar within style peer groups, not across unrelated strategies.
Document whether MDD is peak-to-trough on daily or monthly marks.
For levered funds, report unlevered Calmar alongside gross exposure metrics.
Set explicit Calmar floors for diversifying sleeves before increasing allocation weight.

Key takeaways

Calmar divides annualized return by maximum drawdown — return per unit of worst historical loss.
It answers crisis questions Sharpe cannot: how much upside did you earn relative to the deepest hole?
Three-year net Calmar near 0.5+ is a common allocator hurdle for diversifying strategies; verify the window includes stress.
High Sharpe with low Calmar signals tail-risk strategies that look smooth until a sudden drawdown cluster.
Always match return and MDD windows, use net returns, and report rolling Calmar alongside recovery time.