Negative Private Equity Alpha

This paper draws exclusively on data and analysis from four SMP Financial research documents: the European Family Offices Compendium v17, the Annex B Quantitative Portfolio Analysis v7.0 (21-year backtest of the consensus European family office portfolio), the Private Equity Markets Compendium v1.1, and the Public Markets Compendium v1.4. No new primary data is collected. The paper’s purpose is to present the evidence on private equity alpha that emerges from cross-reading these documents, to examine its causes, assess the critical caveats, and set out the implications for portfolio construction. This document contains no investment advice and is intended for professional and institutional readers.

A GUIDE FOR NON-PROFESSIONAL READERS

What is alpha?

Alpha is the investment industry’s term for return that cannot be explained by the general movement of markets. If you invest in a fund that simply rises and falls with the stock market, that is “beta” — pure market exposure. If the fund consistently earns more than the market over time, after adjusting for how much market risk it took, that extra return is alpha. It is what you pay active fund managers to produce.

Alpha is measured against a benchmark — a reference portfolio that represents the “free” alternative. For this analysis the benchmark is a Global 60/40 portfolio: 60% invested in a global equity index fund and 40% in a bond index fund. This is the simplest, cheapest, most liquid portfolio a sophisticated investor could hold. Any strategy that claims to be better than this must earn positive alpha to justify its added complexity, illiquidity, and cost. If it earns negative alpha, it is, by definition, worse than the free alternative on a risk-adjusted basis — regardless of what its absolute return looks like.

What is private equity?

Private equity (PE) is the ownership of companies that are not listed on a stock exchange. Instead of buying shares that trade daily on the London Stock Exchange or New York Stock Exchange, a private equity investor buys a stake in a company directly — typically as part of a fund managed by a specialist firm. The fund holds those stakes for five to ten years, trying to improve the companies and then sell them at a profit.

The standard promise of private equity is that it earns more than public stock markets over the long run, in exchange for two things you give up: liquidity (you cannot sell your stake whenever you want) and transparency (the value of your investment is not updated daily — it is estimated periodically by the fund manager). Historically, this promise has been partly fulfilled. Top-performing PE managers, over long periods and measured carefully, have outperformed public equity markets by perhaps two to four percentage points per year. But “top-performing” and “measured carefully” are doing a great deal of work in that sentence.

Why does observing negative alpha here matter?

This analysis does not look at one PE fund in isolation. It examines the actual portfolio that European family offices — some of the wealthiest and most sophisticated private investors in the world, with average assets of €1.70 billion — collectively hold. Their portfolio contains approximately 19% private equity and 36% private markets in total. The question being asked is simple: has this allocation, as a whole, earned a return that justifies its complexity, cost, and illiquidity relative to what a basic index approach would have delivered?

The answer, based on a 21-year backtest, is no — but with important qualifications. The portfolio delivered a gross return slightly below a simple 60/40 index portfolio (7.1% versus 7.2% per year) while taking 31% more risk. After costs, the shortfall widens. In the worst years — the 2008 financial crisis, the 2022 rate shock — it fell harder than the benchmark rather than providing the protection that diversification is supposed to deliver. The added complexity produced negative alpha of 1.16% per year.

This finding matters beyond the family office world. Pension funds, endowments, sovereign wealth funds, and high-net-worth individuals across Europe are being actively encouraged — through new regulations such as ELTIF 2.0 and through fund marketing — to increase their allocations to private markets. The promise is always the same: better returns, genuine diversification, lower volatility. This analysis examines whether that promise is supported by the data for a real, large-scale institutional portfolio over two decades. It is not a verdict on private equity as a concept; it is evidence that the average execution of the PE allocation thesis, in the way European family offices currently implement it, has not earned its keep.

How to read this document

The document is in six parts. The first two sections present the data — what the numbers show and why. Section 3 is the most important caveat: the analysis uses market prices that update daily, whereas family offices actually use estimated valuations that update quarterly. This makes the backtest harsher than the experience family offices themselves report, and the gap is quantified carefully. Section 4 tests whether external data from McKinsey, Cambridge Associates, and Dimensional Fund Advisors tells the same story. Section 5 works through the practical implications for portfolio construction. Section 6 states the conclusions plainly.

Technical terms are used where necessary but explained when first introduced. A glossary is not included in this document; the companion SMP Financial Public Markets Compendium and Private Equity Markets Compendium each contain full glossaries.

THE CENTRAL FINDING

−1.16% Annualised alpha vs Global 60/40, 2005–2025

The consensus European family office portfolio, holding 24% in private equity and private credit (PE 21%, private credit 3%), generated negative alpha of 1.16% per annum against a simple two-asset benchmark over a 21-year backtest. Source: Annex B Quantitative Portfolio Analysis.

133% Down-capture ratio

In years when the Global 60/40 fell, the family office portfolio fell 35% harder. Private markets allocations amplified rather than dampened drawdowns. Source: Annex B.

0.44 Sharpe ratio (gross) vs 0.54 for 60/40

The portfolio earned less return per unit of risk than a plain stock-bond portfolio, despite carrying materially more complexity, illiquidity, and cost. Source: Annex B.

0.95 PE-to-MSCI World correlation

The listed PE proxy (LPX Major Market Index) was 95% correlated with global equities — near-perfect co-movement. Private equity diversification, on this evidence, was illusory. Source: Annex B.

“The extra complexity of private markets and alternatives did not, on this sample and these proxies, earn its keep relative to a plain two-asset portfolio.” — Annex B Executive Summary

1. What the Data Actually Says

1.1 The Backtest Portfolio

The analysis originates in Annex B of the SMP Financial European Family Office suite — a 21-year quantitative backtest (2005–2025) of the consensus median portfolio derived in Section 11 of the main Compendium from its two principal allocation surveys (UBS; Goldman Sachs/Campden), within a ten-source survey base. That portfolio — designated the Analysis Base Portfolio (ABP) — holds the allocations that European family offices actually report holding: 30% developed market equities, 19% private equity, 12% real estate, 14% DM fixed income, 3% private credit, 5% hedge funds, 4% EM equities, 3% EM fixed income, 2% infrastructure, 2% gold, and 5% cash plus a 1% Art sleeve carried at cash-equivalent characteristics (an effective 6% cash weight).

Private markets therefore comprise 36% of the ABP by capital weight (PE 21%, real estate 10%, private credit 3%, infrastructure 2%). This is not a theoretical allocation: it is what European family offices with average AUM of €1.70 billion actually hold, synthesised across UBS, Goldman Sachs, J.P. Morgan, and Campden Wealth survey data.

All asset classes are proxied using listed, mark-to-market indices for the backtest: LPX Major Market Index for PE, EPRA Nareit Developed for real estate, Credit Suisse (now Morningstar LSTA) Leveraged Loan Index for private credit, MSCI World for DM equities, and so forth. The portfolio is rebalanced annually to target weights. This is an important constraint: it means the backtest is harsher than family offices’ own reported figures, for reasons examined in Section 3.

1.2 The Full Performance Picture

The ABP’s 21-year performance record is summarised below alongside its two benchmarks.

Table 1. Full metric suite: Analysis Base Portfolio vs benchmarks, 2005–2025 (21 annual observations). Net A = gross −0.44% p.a. (UCITS/ELTIF route); Net B = gross −0.72% p.a. (institutional). Source: Annex B Quantitative Portfolio Analysis, Table B2.1.

Several features of this table are individually significant.

• The ABP’s gross CAGR of 7.1% falls short of the 60/40’s 7.2% even before costs — and after realistic costs (Net A: 6.6%; Net B: 6.4%) the shortfall widens on an absolute return basis.

• The ABP runs 14.5% volatility versus the 60/40’s 11.1% — 3.4 percentage points of additional annualised risk — while delivering a gross return 0.1 percentage points below the benchmark.

• The Sharpe ratio deterioration is substantial: 0.44 for the ABP versus 0.54 for the 60/40 — the family office portfolio earns approximately 18% less return per unit of risk despite 31% more volatility.

• The worst-year drawdown is -33.5% for the ABP versus -22.1% for the 60/40 — a difference of 11.4 percentage points in the single worst calendar year.

• The percentage of positive calendar years is 76% for the ABP but 86% for the 60/40 — the family office portfolio has negative returns in one year in four, against approximately one year in seven for the simpler benchmark.

• The CVaR at 95% — the expected loss in the worst 5% of outcomes — is -25.5% for the ABP versus -19.0% for the 60/40, a 6.5-point tail risk differential.

1.3 Alpha, Beta, and the Benchmark-Relative Picture

The headline alpha figure — −1.16% per annum against the Global 60/40 — is derived from OLS regression of ABP excess returns on 60/40 excess returns over the 21-year sample.

Table 2. Benchmark-relative statistics, 2005–2025. Source: Annex B, Table B3.1.

Reading the information ratio

An information ratio of 0.11 vs the 60/40 means the ABP generates 0.11% of active return for every 1% of active risk taken. This is very low — professional active managers typically target IRs of 0.3 to 0.5 or better — but it is at least positive. Against the MSCI World, the IR is -0.54: the ABP actively destroys risk-adjusted return relative to a pure equity benchmark. This asymmetry is explained by the ABP’s construction: it takes equity-like risk but holds enough fixed income and cash to dilute pure equity upside, without the diversification benefits that a genuine alternatives allocation would provide.

2. Why: The Mechanics of Negative Alpha

2.1 Risk Concentration, Not Diversification

The primary cause of the ABP’s negative alpha is not that private markets are bad assets in absolute terms. It is that the private market assets held in the ABP — as proxied by listed instruments — behave like leveraged equity rather than genuine diversifiers. The risk decomposition from Annex B makes this concrete.

Table 3. Capital weight vs risk contribution (% of total portfolio volatility). Total portfolio volatility: 14.5%. Source: Annex B, Table B4.1.

Three numbers from this table define the problem. Private equity holds 21% of capital but contributes 37.3% of portfolio risk — a risk-to-capital ratio of 1.78x, the highest of any asset class. Real estate holds 10% of capital but contributes 12.8% of risk. Combined with DM and EM equities, the growth complex (DM equity + EM equity + PE + RE) holds 63% of capital but contributes approximately 89% of total portfolio volatility.

The reason for PE’s outsized risk contribution is its correlation to the equity market. The LPX listed PE proxy shows a correlation of 0.95 to the MSCI World — near-perfect co-movement. Real estate (EPRA Nareit) correlates at 0.80. Private credit (leveraged loan proxy) correlates at 0.71. These are not alternative assets in any meaningful statistical sense: they are leveraged or credit-sensitive equity exposures. A portfolio that is 36% ‘private markets’ but where those private markets track the public equity market at 0.71 to 0.95 correlation is not more diversified than a pure equity portfolio — it is concentrated in the same risk factor at a higher overall leverage level.

The diversification ratio

Annex B calculates the ABP’s diversification ratio at 1.12 — a modest improvement over a concentrated portfolio, but far below what the asset class breadth would suggest. A genuinely diversified portfolio of 11 asset classes with meaningful low-correlation assets would typically show a diversification ratio of 1.3 to 1.6 or higher. The 1.12 figure reveals that the portfolio’s breadth is taxonomic (eleven named classes) but its risk structure is monolithic (equity beta).

2.2 The Return Decomposition Evidence

The PE Markets Compendium draws on StepStone Group analysis cited by McKinsey to decompose where private equity returns have historically come from. For deals completed between 2010 and 2022, leverage and multiple expansion together accounted for 59% of total PE returns. Operational value creation (revenue growth and margin improvement) accounted for 41%.

This decomposition is central to understanding the alpha question. Alpha in private markets was never primarily generated by the superior operational skill of PE managers. It was generated by two macro tailwinds: (i) falling interest rates that reduced the cost of leverage and mechanically increased the value of leveraged assets; and (ii) expanding valuation multiples as risk appetite grew in the decade of quantitative easing. Both tailwinds have reversed or materially weakened.

Table 4. PE return decomposition: historical sources and current status. Source: StepStone Group analysis cited in McKinsey Global Private Markets Report 2026 (reference [A] in PE Compendium); author-constructed status column.

The implication is stark. The historical PE premium was significantly a leveraged beta trade dressed as alpha. When you hold listed PE proxies in a portfolio and compare to a 60/40 benchmark, you are stripping away the smoothing that conceals the true beta exposure, and the result is negative alpha. The beta is present — PE loaded 1.78x more risk per unit of capital than the portfolio average — but the alpha that was supposed to compensate for taking that risk is absent.

2.3 Regime-by-Regime Evidence

Breaking the 21-year backtest into five market regimes reveals where the ABP’s performance deficit is most acute.

Table 5. Annualised regime returns and ABP vs 60/40 differential. Source: Annex B, Table B6.1.

The pattern is unambiguous. The ABP outperforms the 60/40 only during the QE expansion (2010–2019) — precisely the regime characterised by falling rates, expanding multiples, and abundant cheap leverage: the three macro tailwinds that StepStone’s analysis identifies as accounting for 59% of historical PE returns. Remove those tailwinds, and the ABP underperforms in every other market regime. It underperforms badly in both crisis episodes (GFC, inflation shock), underperforms modestly when public equity surges (AI cycle), and captures upside but less efficiently than pure equity in the pandemic cycle.

The QE decade as statistical outlier

The family office PE allocation appeared to work during 2010–2019 because the conditions of that decade — near-zero rates, quantitative easing, expanding multiples — specifically favoured leveraged buyout structures. That regime is over. Both the PE Compendium and the Public Markets Compendium document that the macro tailwinds driving the PE premium in that decade have materially changed: the ECB raised rates to 2.25% in June 2026; entry multiples are at record highs; and leverage is more expensive. The one regime that generated positive alpha for the ABP was the exception, not the rule.

3. The Critical Caveat: Appraisal Smoothing

3.1 The Smoothing Problem

The most important qualification to the negative alpha finding is the smoothing distortion inherent in how private market assets are valued. Annex B explicitly addresses this and it must be given full weight.

Private equity, private credit, and real estate allocations in family office portfolios are not valued daily or even monthly. They are valued using lagged appraisal-based methods: periodic assessments of the underlying company or asset value by fund managers, typically quarterly, reflecting recent transaction comparables and earnings multiples. These valuations lag market reality — they understate losses in downturns and understate gains in sharp recoveries. The result is systematically smooth return series that understate true volatility.

The ABP backtest uses listed-market proxies (LPX PE, EPRA Nareit, LSTA Leveraged Loans) which are mark-to-market and carry no smoothing. To quantify the gap, Annex B applies an AR(1) smoothing filter (parameter φ=0.4) to the PE, private credit, and real estate series, simulating appraisal-based valuation.

Table 6. Smoothing distortion: mark-to-market vs appraisal-simulated returns. AR(1) filter φ=0.4 applied to PE, private credit, and real estate. Source: Annex B, Table B8.1; methodology following Geltner (1991).

The implication is direct. If a European family office were to report its portfolio’s Sharpe ratio using its own internal valuations (appraisal-based), it would likely report 0.59 rather than the true mark-to-market 0.44 — a 34% overstatement of risk-adjusted performance. It would report volatility of approximately 11% rather than 14.5% — understating true risk by approximately 3.5 percentage points. It would observe shallower drawdowns in its own reporting than the underlying economic reality of its assets.

This is not fraud — it is methodology

Annex B is explicit that appraisal smoothing ‘is not fraud — it reflects genuine valuation methodology.’ GAAP and IFRS accounting for illiquid assets requires managers to use their best estimate of fair value, typically informed by comparable transactions and earnings multiples. This produces smoother series than daily market prices. The problem arises when these smoothed series are used to compare performance against mark-to-market benchmarks, or when smoothed Sharpe ratios are presented as evidence of risk-adjusted performance superiority over liquid alternatives. The honest accounting requires acknowledgement of both the smoothing and what it conceals.

3.2 The Honest PE Performance: Mark-to-Market vs Appraisal

The smoothing analysis means there are two versions of the PE alpha story:

The mark-to-market truth (ABP backtest)

Negative alpha of -1.16% p.a. vs 60/40. Lower Sharpe (0.44 vs 0.54). Worse drawdowns (-33.5% vs -22.1%). This is the honest underlying economic picture of what the assets are actually worth in real time.

The appraisal-reported experience

Sharpe appears to be 0.59. Volatility appears to be 11.0%. Drawdowns appear shallower. This is what family offices see in their own portfolio reports and what LP due diligence exercises typically analyse.

The gap between these two pictures is the information problem at the heart of the PE alpha debate. LPs who evaluate private market performance using appraisal-based NAV data are systematically comparing smooth private returns against volatile public benchmarks — and concluding that private markets are superior diversifiers. The mark-to-market analysis reveals this conclusion is significantly an artefact of the measurement method.

3.3 What Cambridge Associates Says About Horizon-Adjusted Performance

The PE Markets Compendium draws on Cambridge Associates’ US PE/VC Benchmark Commentary (H1 2025) — the most rigorous publicly available fund-level net IRR benchmark using actual fund cash flows. The CA picture offers a more nuanced reading than the ABP backtest alone.

Over long horizons — ten years and beyond — the CA US PE Index shows a Public Market Equivalent (mPME) premium of approximately 200 basis points over the Russell 3000 and 480 basis points over the MSCI World on a cash-flow-adjusted basis. Over shorter horizons (five years), performance has been mixed: the CA PE Index has generally outperformed small-cap benchmarks but equalled or underperformed large-cap indices. In H1 2025 specifically, the US PE Index earned 3.9% on a six-month basis, with buyouts at 3.6% and growth equity at 4.9%.

The critical distinction between the ABP backtest and the CA PME analysis is methodology. The CA mPME compares actual fund cash flows (capital calls and distributions) to what those same cash flows would have earned if deployed in public markets on the same schedule — a like-for-like comparison that corrects for the timing distortion inherent in IRR. The ABP backtest, by contrast, uses listed proxies with daily mark-to-market pricing and annual rebalancing — which captures the beta accurately but cannot replicate the specific cash flow and holding-period dynamics of actual PE fund investing.

Reconciling the two findings

The CA PME data showing a 200–480bp long-run premium and the ABP backtest showing -1.16% alpha are not straightforwardly contradictory. They measure different things. The CA PME captures actual fund-level returns net of fees, using cash flow timing — and shows genuine outperformance at the asset class level over 10+ year horizons, heavily concentrated in top-quartile managers. The ABP backtest uses listed PE proxies, which are mark-to-market but capture the asset class beta more cleanly. The two methods bracket the truth: actual PE investing by skilled LPs in top-quartile funds has historically outperformed public markets on a long-horizon, cash-flow-adjusted basis. But for the average LP holding a portfolio that looks like the European family office consensus, the mark-to-market evidence suggests the extra risk is not being adequately compensated.

4. External Corroboration

4.1 McKinsey: Three Consecutive Years of Buyout Underperformance

The PE Markets Compendium’s Section 2 draws on McKinsey’s Global Private Markets Report 2026 for the most current industry-level performance data. The McKinsey finding is unambiguous and extends the ABP backtest’s conclusion to real, realised fund-level IRR data.

In 2025, buyout funds underperformed US and global public equities for the third consecutive year, generating pooled IRRs of approximately 7–8% (top-quartile basis) against the S&P 500’s 18% and the MSCI World’s 22%. Over the full 2015–2025 period, average buyout pooled IRR was approximately 6% — compared with the S&P 500’s 15% total shareholder return and the MSCI World’s 13% over the same period. Buyout fund IRRs reached a post-2002 trough for the 2022–2025 period at 5.7% pooled.

The single most damaging data point is the 2015–2017 vintage: funds from this cohort are generating approximately 2% IRRs. These are funds that paid peak-cycle multiples, held assets through the rate shock of 2022–2023, and are now struggling to exit. A European family office that committed capital to PE in 2015–2017 — the most natural vintage years for a programme building during the QE era — is holding assets generating approximately 2% IRR, against a risk-free rate that has been above 4% in the US for the past two years.

Table 7. Current PE performance data against public market benchmarks. Sources: PE Markets Compendium Section 2; McKinsey Global Private Markets Report 2026 [A]; Cambridge Associates [D]; Campden Wealth as cited.

4.2 Dimensional: 42 Years of Private Credit vs High Yield

The Public Markets Compendium’s Section 5 reports a finding by Dimensional Fund Advisors that independently corroborates the ABP’s negative alpha on the private credit component of the portfolio.

Using 42 years of MSCI data (1980–2022), Dimensional’s Co-CIO Savina Rizova found that public high-yield bond benchmarks outperformed private credit on a RECTA framework (risk-adjusted, excess, correlation, timing, and alpha basis). Private credit’s IRR of approximately 10% over that period appeared attractive in absolute terms, but when the correlation structure, timing of cash flows, and risk adjustment are properly applied, public high yield performed comparably or better — with superior liquidity.

This finding directly interrogates the private credit allocation within the ABP (3% weight, 2.2% risk contribution). The ABP’s private credit proxy — the Credit Suisse Leveraged Loan Index — shows a 0.71 correlation to the MSCI World and a standalone Sharpe of 0.31. It contributed modestly to the portfolio’s risk but generated insufficient return to justify its illiquidity. The mean-variance optimiser in Annex B’s Refined ABP reduces the private credit allocation to zero (from 3%) — an artefact of in-sample optimisation, but consistent with the Dimensional finding.

4.3 Goldman Sachs Family Office Report: LP Behaviour Responding

The most direct evidence that sophisticated LPs are responding to this data is the shift in PE allocation reported by Goldman Sachs. The 2025 Goldman Sachs Family Office Investment Insights report records that PE allocations among surveyed family offices fell from 26% in 2023 to 21% in 2025 — a five-percentage-point reduction in two years, the largest reallocation away from any single asset class in the dataset.

The Goldman report explicitly attributes this to ‘market dislocations, dispersion, and outperformance in public equities creating an attractive opportunity set relative to private equity.’ Family offices are shifting capital back toward public equities — the same assets that the ABP backtest shows have a superior Sharpe ratio (0.52 for MSCI World gross) and, at the current juncture, competitive forward return expectations.

The survey data assembled in the main Compendium corroborates this cross-sectionally: on the consistent basis adopted in v17.0 of the suite — UBS’s European allocation set against Goldman Sachs’ primary global release — both principal sources report private equity at 21% (Goldman’s figure down from 26% in 2023), and the consensus median is 21%. UBS’s own global series shows PE essentially flat (16% in 2019; 17% in 2025) even as developed-market equity allocations rose from 23% to 27%: the marginal allocation has gone to public, not private, equity. The consensus median of 21% therefore sits at the centre of the survey readings — and represents a meaningful overweight relative to what the mark-to-market performance data would suggest is warranted for the risk consumed.

5. Implications for Portfolio Construction

5.1 The True Cost of the PE Allocation

The ABP’s 21% PE allocation consumes 37.3% of total portfolio risk while generating negative alpha against a simpler benchmark. Translating this into portfolio construction language: for every unit of risk budget available, PE is 1.78x more expensive than the portfolio average, but generates returns that do not compensate for that cost.

The opportunity cost is concrete. Annex B’s mean-variance optimisation shows that a disciplined reweighting within ±5 percentage points per asset — adding 5pp to DM equities, adding 5pp to gold, trimming 5pp from PE, trimming 5pp from real estate — lifts the annualised CAGR from 7.1% to 7.9% at the same volatility, improving the Sharpe ratio from 0.44 to 0.50. These are in-sample estimates subject to estimation error, but the direction is clear and consistent with the structural arguments.

5.2 What the DMS 125-Year Data Implies

The Public Markets Compendium’s Section 1 grounds the discussion in 125 years of the DMS database. The long-run equity risk premium of 4.3% over bills and 3.1% over bonds represents the baseline compensation available from public equity markets without any illiquidity premium, manager selection risk, or fee drag. Private equity’s historical case rests on earning this public market premium plus an additional illiquidity premium of approximately 200–400 basis points — the PME outperformance documented by Cambridge Associates over 10+ year horizons in skilled hands.

The current evidence suggests that illiquidity premium is substantially not being captured by the consensus family office PE allocation for three reasons: first, the premium is concentrated in top-quartile managers that most LPs cannot consistently access; second, the macro conditions (leverage, multiple expansion) that drove 59% of historical PE returns have structurally changed; and third, the smoothing of reported PE performance obscures the true risk being taken, making the net premium appear larger than it is on a mark-to-market basis.

5.3 Three Scenarios for How the Alpha Gap Closes

The negative alpha finding is a statement about the past 21 years as proxied by listed instruments. It is not necessarily a permanent condition. There are three scenarios in which PE alpha could reassert itself.

Table 8. Scenarios for PE alpha reassertion. Author-constructed; probability assessments are directional, not quantitative.

5.4 What the Optimised ABP Implies

Annex B’s Refined ABP — the maximum-return portfolio at ABP volatility within ±5pp per-asset constraints — reduces PE by 5 percentage points (21% to 16%), real estate by 5 (10% to 5%), DM fixed income by 4 and cash by 5, and eliminates the 3% private credit sleeve; the proceeds go to DM equities (+5pp), gold (+5pp), infrastructure (+5pp), EM fixed income (+5pp), and small additions to EM equities and hedge funds. The result is an 80 basis point CAGR improvement (7.1% to 7.9%) at the same risk level — a figure independently reproduced from the Table B10.1 return matrix.

The optimiser’s logic is transparent: DM equities at 0.95 correlation to PE offer essentially the same beta as PE, but more cheaply (no illiquidity premium demanded, lower fees, daily liquidity, better drawdown profile). Gold at 0.26 correlation to world equities is the only asset in the portfolio that genuinely diversifies equity risk — and is underweighted at 2%. The survey data independently confirms that European family offices are actively increasing gold allocations, consistent with this finding.

The two moves the optimiser makes that require caution are eliminating private credit entirely (in-sample overfitting to one bad year, 2008) and doubling EM fixed income (flattered by one good year, 2009). These are estimation-error artefacts, not structural signals. But the reduction in PE and RE and the increase in DM equity and gold are structurally motivated and robust to methodology.

5.5 The European Family Office Context

For a European family office with an average AUM of €1.70 billion, the practical stakes of this analysis are significant. A 21% PE allocation represents approximately €357 million per office. Because the −1.16% alpha is measured at the whole-portfolio level, its cost is properly stated on total assets: approximately €20 million per year in foregone risk-adjusted return for the average office, relative to what a simpler benchmark would have generated — before even accounting for the higher fee drag of the private market vehicles.

Figure: Exhibit 3. The practical stakes for the average European family office. Author-constructed from Annex B and Section 5.5. — see the designed PDF for the exhibit.

The compendium data also establishes that the DPI drought — distributions as a share of PE AUM at a historic low of 6% for the twelve months to June 2025 — is creating a liquidity constraint that is invisible in the appraisal-based reported returns but very real in the cashflow experience. A family office that committed €357 million to PE at the 21% consensus weighting is receiving approximately €21 million in annual distributions from that sleeve (6% DPI rate), against a risk-free rate of 2.25% (ECB deposit rate, June 2026) on the same capital of approximately €8.0 million. The income from the PE allocation is barely covering the risk-free opportunity cost of the illiquidity, let alone generating a premium for it.

5.6 The Consensus+ Construction

The findings above motivate a named construction: Consensus+, the consensus portfolio with the negative-alpha private-markets sleeves replaced by public-market allocations. Recomputed directly from the Annex B annual return matrix (Table B10.1) on the same annual-rebalancing basis, three variants illustrate the effect. Substituting the PE and private credit sleeves entirely into DM equities produces a gross CAGR of 7.4% at 12.9% volatility (Sharpe 0.50; worst year −29.3%; alpha −0.39%; down-capture 116%). A balanced substitution into DM equities, fixed income, gold and infrastructure produces 7.1% at 11.6% volatility (Sharpe 0.52; worst year −25.1%; alpha −0.11%; down-capture 101%). Annex B’s Refined ABP, which retains a 16% PE sleeve within ±5pp constraints, produces 7.9% at 14.5% (Sharpe 0.50).

Figure: Exhibit 1. The consensus portfolio, the Refined ABP, and two Consensus+ substitution variants, gross, 2005–2025. Author-constructed; recomputed from Annex B Table B10.1 with annual rebalancing. — see the designed PDF for the exhibit.

The honest in-sample reading: Consensus+ closes most, not all, of the negative-alpha gap (−1.16% falls to between −0.1% and −0.4%), cuts the worst calendar year by four to eight percentage points, and reduces down-capture from 133% toward 101–116% — with the equity variant modestly above, and the balanced variant broadly level with, the plain 60/40’s 7.2%. It does not, on gross in-sample figures, exceed the 60/40’s Sharpe ratio of 0.54. The structural case therefore carries the remainder, and it is substantial: the cost drag modelled in Annex B (−0.44% to −0.72% per annum) applies almost entirely to the private sleeves being replaced, so the net comparison favours Consensus+ by more than the gross; implementation requires no top-quartile manager access; liquidity is daily; and the in-sample caveats of Section 5.4 apply with less force because the substitution is justified structurally — by the 0.95 correlation of the listed PE proxy to world equities — rather than by return-chasing.

Figure: Exhibit 2. Annualised alpha versus the Global 60/40 before and after substitution of the private-market sleeves. Author-constructed from Annex B Table B10.1. — see the designed PDF for the exhibit.

For Irish-resident implementation there is a further leg, developed in Annex C: directly held public equities fall within the CGT regime (33%, annual exemption, full loss relief, and the step-up in basis at death), whereas fund-wrapped private-market exposure typically falls within the 38% exit tax regime, with eight-year deemed disposals, no loss offset, and no step-up. The substitution therefore compounds three ways — gross risk-adjusted return, fees, and tax. A full specification of the Consensus+ portfolio and its implementation is the subject of a companion note.

6. Conclusion

The evidence assembled across the four SMP Financial research documents is consistent and mutually reinforcing. The consensus European family office portfolio, as backtested over 2005–2025 using mark-to-market proxies, generated negative alpha of 1.16% per annum against a simple Global 60/40 benchmark. It did so at materially higher volatility, a lower Sharpe ratio, worse drawdowns in every major down episode, and a down-capture ratio of 133%. The headline CAGR of 7.1% gross fell short of the 60/40’s 7.2% even before costs — a shortfall obtained while taking 31% more volatility.

The primary cause is structural: private equity and real estate proxies correlate at 0.95 and 0.80 to world equities respectively, making them leveraged equity rather than genuine diversifiers. The portfolio concentrates risk rather than spreads it. The one regime in which the ABP outperformed — the QE expansion of 2010–2019 — was precisely the regime in which leverage and multiple expansion (59% of historical PE returns) were most powerful. That regime has ended.

The critical caveat is the smoothing distortion. Real family office private market holdings, valued on appraisal rather than mark-to-market, will show Sharpe ratios approximately 0.15–0.20 higher, volatility approximately 3.5pp lower, and shallower drawdowns than the honest underlying economic picture. This is not fraud; it is the consequence of genuine valuation methodology. But it means the negative alpha finding is harsher than family offices’ own experience — and that the experience itself is optimistic relative to the true risk being taken.

The CA PME data over 10+ year horizons shows that private equity, assessed on actual cash-flow timing, has historically delivered a genuine premium of 200–480 basis points over public market equivalents at the pooled index level — a premium heavily concentrated in top-quartile funds. This is real and should not be dismissed. But it is concentrated in top-quartile manager selection, requires patient long-horizon capital, and is increasingly dependent on operational value creation rather than financial engineering — a harder and less scalable capability.

The practical conclusion for a European family office reviewing its allocation is not ‘exit private equity.’ It is: be honest about what private markets are actually delivering and why; assess whether your manager relationships are top-quartile or median; stress-test the portfolio using mark-to-market rather than appraisal-based valuations; and consider whether the current allocation to PE (21%) is appropriate given that listed DM equities offer the same beta at lower cost, better liquidity, and without the structural headwinds of record entry multiples, elevated rates, and a historic DPI drought.

The public market data assembled in the Public Markets Compendium provides the comparison point. Global equities (MSCI World) have delivered 9.1% gross CAGR over the same 2005–2025 period at 17.3% volatility and a Sharpe of 0.52. International small value returned 38.6% in 2025 alone. Bonds have re-emerged as competitive on a risk-adjusted basis with forward yields around 4%. The opportunity cost of the PE allocation — defined as what the same capital could have earned in liquid public markets — is not theoretical. It is quantified.

“The case for PE rests on the genuine illiquidity premium and lower realised volatility of true (unlisted) private markets, which a listed-proxy backtest cannot capture. The refinements identified here would improve risk-adjusted efficiency without departing from the consensus that defines [the portfolio].” — Annex B Executive Summary

The honest summary in three lines

On a mark-to-market basis, the consensus PE allocation generates negative alpha. On an appraisal basis, it looks better than it is. The long-run case for PE is real but is concentrated in skilled manager selection and requires honest accounting of the risk being taken.

Appendix. Sources

This paper draws exclusively on four SMP Financial documents and the primary sources cited within them. No new primary data is collected or analysed.

Table A.1. Source reference summary. All data author-paraphrased from cited sources.

SMP Financial Ltd is regulated by the Central Bank of Ireland. Registration number C48338. This document does not constitute investment advice or a financial promotion. It is a research paper intended for professional and institutional readers. Past performance is not a reliable indicator of future results.