Skip to content
QuantCalc Tools
Directory

Cagr Irr Drawdown Compounding

Treynor Ratio, Jensen’s Alpha & Modigliani M-Squared (M2) Leverage Math

When adding a sleeve to a diversified multi-manager book, systematic beta matters more than standalone sigma; master exact Treynor, Jensen Alpha, and Modigliani M-Squared (M2) equations.

By Elena Rostova, FRM, CQFReview not recorded · Updated
Treynor Ratio, Jensen’s Alpha & Modigliani M-Squared (M2) Leverage Math — QuantCalc Tools 2026 quantitative & supervisory data card
QuantCalc Tools (CAGR vs IRR, Volatility Drag, Sharpe/Sortino & Drawdown Recovery): Modigliani M-Squared bps scaling, Treynor beta efficiency, and FF5 Jensen Alpha.Source: QuantCalc Tools Institutional Research DeskLicense: CC-BY-4.0

Investment committee debates routinely stall over which classical risk-adjusted metric actually applies to the decision on the table: if Fund A has a higher Sharpe Ratio (0.85 vs 0.72) while Fund B has a higher Treynor Ratio and higher Jensen's Alpha, which fund should the pension plan hire? And how do you explain a 0.85 vs 0.55 Sharpe difference in intuitive annualized percentage points to non-quant trustees? Answering both questions requires separating standalone total-portfolio choice (Sharpe & Franco/Leah Modigliani's M-Squared / M2) from marginal sub-portfolio sleeve allocation (Treynor & Treynor-Black Appraisal Ratio). Run side-by-side ratio checks in our Sharpe & Sortino Calculator and anchor in our CAGR, IRR & Drawdown Hub.

Unifying Sharpe, Modigliani $M^2$, Treynor, and Jensen's Alpha on One Balance Sheet#

Let $Rp$ be portfolio return, $Rm$ be market benchmark return (with volatility $\sigmam$), and $rf$ be the risk-free SOFR / T-bill rate. Decomposing total variance into systematic beta risk plus idiosyncratic residual variance $\sigmap^2 = \betap^2 \sigmam^2 + \sigma{\varepsilon}^2$ links all four classical performance metrics via exact algebraic identities:

Formula or model from the articleNot independently verified
Exact Unified Equations — Modigliani M^2, Treynor Ratio & Jensen's Alpha:
  • CAPM / Factor Decomposition: Rp - rf = alphaJ + betap * (Rm - rf) + epsilonp, sigmap^2 = betap^2*sigmam^2 + sigma_e^2
  • 1. Modigliani M-Squared (M^2 — Volatility-Matched Return in % / bps!):
  • Leverage/Cash Weight w* = sigmam / sigmap (De-lever if sigmap > sigmam; lever at rf if sigmap < sigma_m)
  • M^2 = rf + (sigmam / sigmap) * (E[Rp] - rf) = rf + SRp * sigmam
  • M^2 Alpha (in bps/yr) = M^2 - E[Rm] = (SRp - SRm) * sigmam
  • 2. Treynor Ratio (Excess Return per Unit of Systematic Beta — Valid ONLY when beta_p > 0):
  • Treynorp = (E[Rp] - rf) / betap = (E[Rm] - rf) + (alphaJ / betap)
  • 3. Exact Link Between Sharpe & Treynor via Correlation rho_{p,m}:
  • Since betap = rho{p,m} * (sigmap / sigmam) => SRp = (rho{p,m} / sigmam) * Treynorp
  • Worked Pension Committee Paradox (rf = 4.50%, Market Benchmark E[Rm] = 10.50%, sigmam = 15.0% -> SRm = 0.40, Treynor_m = 6.00%):
  • Fund A (Diversified Core, rho = 0.98): E[RA] = 11.70%, sigmaA = 12.0%, beta_A = 0.784
  • > SRA = (11.70 - 4.50)/12.0 = 0.600 | M^2 = 4.50% + 0.60*(15%) = 13.50% (+300 bps M^2 Alpha!) | TreynorA = 7.20/0.784 = 9.18%
  • Fund B (Concentrated Biotech Specialist, rho = 0.50): E[RB] = 14.50%, sigmaB = 20.0%, beta_B = 0.667
  • > SR_B = (14.50 - 4.50)/20.0 = 0.500 (TRAILS Fund A on Sharpe & M^2 = 12.00%!)
  • > Jensen AlphaB = 10.00% - 0.667*(6.00%) = +6.00%/yr! | TreynorB = 10.00 / 0.667 = 15.00% (CRUSHES Fund A!)

Look at why Fund A wins for a single-fund investor while Fund B wins as a $10\%$ satellite sleeve inside a multi-manager pension plan!

  • If Fund B is your entire life savings, you bear all of its massive idiosyncratic biotech volatility ($\sigma_{\varepsilon} = 17.32\%$), so Sharpe (0.60 vs 0.50) and Modigliani $M^2$ (13.50% vs 12.00%) rightly crown Fund A.
  • Inside a 15-manager institutional portfolio, idiosyncratic variance $\sigma{\varepsilon}^2 / K$ diversifies away across uncorrelated managers! Only systematic $\betaB$ adds core market variance, and Fund B delivers +600 bps of Jensen's Alpha (Appraisal Ratio = 6.00% / 17.32% = 0.346), lifting the overall pension plan's Sharpe Ratio higher than Fund A can!
Modigliani M-Squared capital market line construction scaling portfolio volatility to match market benchmark sigma_m via SOFR cash or margin
Modigliani M-Squared capital market line construction scaling portfolio volatility to match market benchmark sigma_m via SOFR cash or marginVerified Exhibit

Realistic Institutional $M^2$ Adjustment When Margin Borrowing Exceeds $r_f$#

Textbook Modigliani $M^2$ assumes that if a low-volatility fund ($\sigmap = 8\%$) is scaled up by $w^* = 15\% / 8\% = 1.875\times$ to match equity index volatility ($\sigmam = 15\%$), the borrowed $0.875\times$ notional finances at the risk-free T-bill rate $rf$. In reality, institutional prime brokers or futures clearing members charge $rb = \text{SOFR} + s{\text{spread}}$ (+40 bps in futures, +125 bps on prime margin)! Exact Spread-Adjusted $M^2$ for $w^* = \sigmam / \sigmap > 1$ is: $$M^2{\text{real}} = w^* \mathbb{E}[Rp] - (w^* - 1)(rf + s{\text{spread}})$$ Subtracting $(w^* - 1) s{\text{spread}}$ prevents low-volatility bond arbitrage funds from claiming unattainable levered equity equivalence!

Governance Matrix: Modigliani M-Squared, Treynor, Jensen Alpha & Appraisal Ratio#

Align risk denominators and failure boundaries across standalone total-portfolio and multi-manager sleeve decisions:

Quantitative Distribution · Fails / Breaks Down When...
Fails / Breaks Down When...
01Modigliani $M^2$$w^ > 1$ ignore…02Treynor Ratio$\beta_p \le 0$…03Jensen's Alpha ($\alpha_J$)Levered $2\time…
Classical MetricDenominator / Risk UnitRanking Identical to Sharpe?Multi-Factor ExtensionFails / Breaks Down When...
Modigliani $M^2$Scaled to $\sigma_m$ (% / bps)100% Identical!Downside $M^2$ (Sortino-matched)$w^* > 1$ ignores $r{\text{borrow}} > rf$ spread
Treynor RatioSystematic $\beta_p$No (ignores $\sigma_{\varepsilon}$)Generalized Treynor (arbitrage pricing)$\beta_p \le 0$ (market-neutral or short!)
Jensen's Alpha ($\alpha_J$)None (raw % excess intercept)No (scales with leverage!)Carhart 4-Factor / FF5 AlphaLevered $2\times$ fund doubles $\alpha_J$ artificially!
Appraisal RatioIdiosyncratic $\sigma_{\varepsilon}$Optimal weight $\propto \alphaJ / \sigma{\varepsilon}^2$Multi-factor residual IRStale marks hide true $\beta_p$ (Scholes-Williams)

Multi-Factor Fama-French-Carhart Stripping of Apparent CAPM Alpha#

Single-factor CAPM regressions routinely mislabel cheap Size (SMB) and Value (HML) tilts as manager skill:

Decomposition of apparent +420 bps CAPM Jensen Alpha into Value, Size, Momentum, and Quality factor premia under Fama-French-Carhart regression
Decomposition of apparent +420 bps CAPM Jensen Alpha into Value, Size, Momentum, and Quality factor premia under Fama-French-Carhart regressionVerified Exhibit

Two Critical Traps: Dimson Lagged Beta & Multi-Factor Alpha Stripping#

  1. Illiquid Assets Understate Contemporary $\beta_p$: Private credit, micro-cap stocks, and high-yield bonds react to S&P 500 moves with a $1\text{--}3$ month lag. Regressing $R{p,t}$ only on contemporaneous $R{m,t}$ artificially shrinks $\hat{\beta}p$ from $0.75$ down to $0.25$—fabricating $+300\text{ bps}$ of fake Jensen's Alpha! Always sum Dimson (1979) lagged betas $\beta{\text{Dimson}} = \beta0 + \beta1 + \beta_2$ alongside Lo (2002) autocorrelation adjustments.
  2. Upgrade Single-Index CAPM $\alpha_J$ to Fama-French 5-Factor + Momentum: Ensure managers are not charging active fees for cheap ETF factor betas; cross-reference Information Ratio and Tracking Error, Half-Kelly leverage sizing, and institutional broker margin schedules in broker lookup.

Frequently asked questions

Can two funds ever rank differently under Modigliani M-Squared (M2) versus the Sharpe Ratio?

No, provided excess returns are evaluated against the same risk-free rate r_f and single-borrowing-rate assumption! Because M^2 = r_f + SR_p * sigma_m where r_f and sigma_m (> 0) are identical positive constants for every fund being compared, M^2 is a strictly increasing linear transformation of SR_p.

Why shouldn’t I just pick the fund with the highest raw Jensen’s Alpha?

Because raw Jensen’s Alpha has no risk denominator! If Fund X has +1.5% alpha with 2% residual vol (Appraisal Ratio = 0.75), borrowing on margin at 2x leverage turns Fund X into +3.0% alpha with 4% residual vol—doubling Jensen’s Alpha via pure leverage without adding a single basis point of true skill!

How do I calculate Treynor-Black optimal capital allocation weight for an active satellite fund?

Treynor & Black (1973) proved that the optimal weight allocated to active portfolio A relative to passive market index M is proportional to (alpha_A / sigma_epsilon_A^2) divided by (E[R_M - r_f] / sigma_M^2)—literally the active Appraisal Ratio divided by residual vol relative to market price of risk!

What is "Downside M-Squared" (or Sortino M2)?

Analogous to standard M^2, Sortino-adjusted M^2 scales portfolio leverage to match the benchmark’s Target Downside Deviation TDD_m rather than total volatility sigma_m: M^2_Sortino = tau + Sortino_p * TDD_m, expressing convex downside-adjusted outperformance in annual percentage points.

Publication record

Published
13 November 2026
Updated
14 November 2026
Author
Elena Rostova, FRM, CQF
Review
Review not recorded
Automated fact check
Passed in the stored pipeline record
Source URLs
4 recorded

AI assistance recorded: anthropic/claude-opus-4.6. This field does not establish human review or source verification.

Academic & Institutional Attribution

Cite This Research & Embed Live Data Card

Open Attribution License (CC BY 4.0)

Standard academic & institutional research paper citation format.

Elena Rostova, FRM, CQF. (2026, November 14). Treynor Ratio, Jensen’s Alpha & Modigliani M-Squared (M2) Leverage Math. QuantCalc Tools. https://quantcalctools.com/guides/treynor-ratio-jensens-alpha-m-squared-modigliani-leverage-math
Explore Cagr Irr Drawdown Compounding Hub (20) →

QuantCalc Tools Data, Benchmarks & Contributor Records

Treynor Ratio, Jensen Alpha & Modigliani M-Squared Formula | QuantCalc Tools