Free-Rider Aggregate Ownership
Derived from the Free-Rider Portfolio Framework — collective intelligence, aggregation ownership, relentless friction minimisation.
Own the weighted-average outcome of everyone trying to pick the winner — and let the market’s price discovery work for you at near-zero cost.
Headline finding
The core claim holds: across five historical crises the conservative aggregate drew down 13.6–33.5 points less than the concentrated winner-pick (GFC: −18.1% vs −51.6%). But the protection is not what a casual reading suggests — the bond aggregate damps through low volatility, not crisis-negative correlation: in the 2020 fast crash BND ended −1% with a −9% path low, and in 2022 it fell −13% WITH equities — a regime the framework never names.
The framework
The economic thesis the portfolio is built on — stated as falsifiable claims with evidence grades.
Markets require a minority of participants to spend money on discovering information; their trading moves prices toward fair value, and an index holder inherits the result without paying the discovery cost. The framework therefore replaces "which asset will win?" with "own the weighted average of all attempts to answer that question": global cap-weighted equity as the aggregation of discovery, a broad bond index as the admission of uncertainty, and a hard rule that every activity must justify its cost — because friction, not insight, is what most reliably compounds.
Pillars
Falsifiers
- F1 — If the aggregate of active funds persistently (≥ 10 years, net of costs) beats the cap-weighted market they trade in, the free-rider premise fails: the arithmetic of active management would have to be wrong.
- F2 — If a mechanical non-cap-weighted rule (e.g. equal weight) dominates cap-weighting on risk-adjusted return persistently across markets and decades, "selection is outsourced" loses its force as a reason to hold the cap-weighted aggregate.
- F3 — If aggregate productive capacity stagnates for decades (global real growth ≈ 0), the system-growth premise fails and the aggregate compounds nothing — diversification then only redistributes a flat pie.
- F4 — If stock–bond correlation stays persistently positive (an inflation-dominated regime), the "own many futures" damping channel is structurally impaired — 2022 was a one-year preview of exactly this failure mode.
Read the full framework summary8 sections ▾
The free-rider structure
Many complex systems need only a minority of participants to discover information; the majority can benefit from the resulting structure. In markets the active minority pays for research, forecasting and monitoring, and their trading impounds what they learn into prices. A passive holder of the aggregate inherits those conclusions continuously, without paying the discovery cost. The objective is therefore not to outperform the discovery process — it is to own its output.
Own the aggregation, not the prediction
Most participants ask "which asset will win?". The framework asks what happens if you own the weighted-average outcome of all participants attempting to answer that question. The supporting arithmetic is Sharpe’s: before costs, the average active dollar must earn the market return (active holdings sum to the market); after costs c it earns strictly less. The aggregate is thus not a bet against active skill — it is a claim on its collective output, minus nothing.
Diversification as a knowledge strategy
The framework reads diversification not primarily as risk management but as epistemics: an explicit admission that the future is unknown. Instead of predicting, own many futures simultaneously. What the paper does not say — and the stress evidence adds — is that this damping is regime-dependent: a broad bond aggregate cushions a slow deflation crisis, barely moves in a fast crash, and falls with equities in a rate shock. "Many futures" is not "all futures".
Natural selection inside the index
The market continuously performs selection: poor performers shrink, lose influence and eventually drop out of a capitalization-weighted index; successful entities grow and gain weight. The index holder receives this adaptation automatically — selection is outsourced. This is also why the framework bets on the system rather than any entity: companies fail, while knowledge accumulates and productive capacity expands. The bet is "human productive capacity grows over time", not "this company succeeds".
Friction, optionality, time
Every friction — fees, taxes, trading costs, behavioural mistakes, complexity overhead — subtracts from the compounding base every single year, which is why the framework treats cost removal as the highest-certainty improvement available. Structures should preserve liquidity and adaptability (optionality compounds, rigidity decays), and the holding period should be extended before any complexity is added: long horizons magnify small, reliable advantages.
Meta-exposure and the generalized form
The framework’s hierarchy of abstraction runs: pick a stock → own many stocks → own the market → own the mechanisms that discover and reallocate capital. Each step requires less prediction. The same template generalizes beyond investing: identify who performs discovery, who bears its cost, and how discoveries are aggregated — then position to benefit from the aggregate (markets, science, open-source ecosystems, venture portfolios).
What the stress evidence adds
The core claim survives testing: across five historical crises the bond-anchored aggregate drew down 13.6–33.5 points less than a concentrated winner-pick of the ex-post winning market, and the synthetic worst-profile gap is 16.6 points. But the protection mechanism is not what a casual reading suggests. The real crisis correlation between world equity and the broad bond aggregate is slightly positive; the cushion comes from the bond sleeve’s low volatility, not from it rising when equities fall. In the 2020 fast crash the aggregate bond index ended at −1% with a −9% path low; in 2022 it fell −13% alongside equities — a regime the paper never names.
Limits
The framework is silent on the joint-fall regime (stocks and bonds down together, 2022-style), treats the broad bond index as an unconditional diversifier, and its single most load-bearing lever — cost minimisation — is the one thing a market-data stress test cannot verify (no fees, taxes or behavioural gaps are modelled here). The cost claim rests on external evidence, not on this analysis. And the honest counter-finding: the concentrated winner-pick won every recovery and the full 18-year window on return; this framework claims the drawdown side only.
From framework to portfolio
If the return source is the system’s collective intelligence rather than anyone’s prediction, the portfolio follows almost mechanically: world cap-weighted equity is the aggregation of discovery (P1/P6), a broad bond index plus a cash sleeve own the futures equity does not cover (P4/P7), and everything else — tilts, hard assets, themes — is a prediction plus friction, which P3 says to remove. The only decision the framework leaves open is the equity/bond split, so the tiers express exactly that: how much uncertainty you admit, not which future you predict.
The portfolio
Three funds, no tilts, no leverage. VT (total world, cap-weighted) carries the aggregation claim; BND (broad US bond aggregate) is the admission of uncertainty; BIL (T-bills) preserves optionality. Deliberately NO gold, crypto or factor sleeves: each would be a prediction plus friction, which the framework forbids — and the cost of that purism is visible in the 2022 row of the evidence, where there is no third diversifier to carry.
| Building block | Role / regime defended | Conservative | Balanced | Offensive |
|---|---|---|---|---|
| World equities, cap-weighted (VT) | The aggregation of all discovery — the system’s collective intelligence as return source | 40% | 60% | 90% |
| Broad bond aggregate (BND) | The "many futures" sleeve — damps through low volatility, not crisis-negative correlation | 50% | 35% | 10% |
| T-bills / cash (BIL) | Optionality reservoir — the only sleeve that carried in 2022 | 10% | 5% | 0% |
How the critical sizes were chosen
- The equity/bond split is the single decision the framework leaves open — the tiers are degrees of admitted uncertainty (P4), not market views. 60/40 (balanced) is the canonical expression.
- Every omitted sleeve is deliberate: gold, crypto or factor tilts would each be a prediction plus friction (P3). The price of that purism is concentrated in the rate-shock regime, where stocks and bonds fall together and nothing else carries.
- BND, not long Treasuries: the framework wants the broad aggregate, not a duration bet. The evidence shows what that choice costs — a much weaker crisis hedge than long bonds (GFC +8% vs TLT’s +25%; COVID −1% vs TLT’s +14%).
The stress evidence
Regime-switching SV factor model — 50 synthetic paths × 6 profiles, 252-day horizon, calibrated on 2008–2026 (4,510 trading days), plus 5 real historical episodes in both rebalancing conventions. The tested claims are the paper’s pre-registered testable implications (it contains no numbers of its own). Descriptive: assumptions stated, not a forecast.
Tiers — real backtest
| Tier | Return p.a. | Vol p.a. | Sharpe | Max DD | Worst regime |
|---|---|---|---|---|---|
| Conservative Bond-anchored aggregate ownership (40/50/10) | +5.6% | 8.7% | 0.67 | −20.7% | Rate shock −22% · p95 −31% (moderate) |
| Balanced The classic 60/40, world-aggregate flavour | +6.9% | 12.5% | 0.60 | −31.3% | Crisis shock −23% · p95 −34% (moderate) |
| Offensive Near-full equity aggregate (90/10) | +8.5% | 18.5% | 0.53 | −45.5% | Crisis shock −32% · p95 −47% (fragile) |
Real backtest over the common window (2014–2026). Worst regime = synthetic profile with the deepest median drawdown; p95 = 95th-percentile drawdown across the n=50 paths (the tail, not the typical case). The grade reflects the median — read it together with the p95.
Synthetic regimes — drawdown depth
Cells: median max-drawdown across n=50 synthetic paths. p95 = 95th-percentile drawdown (the tail — roughly the worst 1-in-20 path). p5 = 5th-percentile 1-year return. Colour ∝ median drawdown depth.
The hedge that breaks
The broad aggregate (duration ~6, including credit) is a much weaker crisis insurer than long Treasuries: it carried in the slow GFC deflation (+8% while equities halved) and was flat in Q4 2018, but in the 2020 fast crash it ended at −1% with a −9% intra-episode low (credit spreads + dash-for-cash), and in 2022 it fell −13% alongside equities. Its damping works through low volatility, not counter-movement — the real crisis correlation with world equity is slightly positive.
Real historical episodes
| Episode | Window | Conservative | Balanced | Offensive | Single-thesis |
|---|---|---|---|---|---|
| COVID crash Fast crash: BND did NOT cushion (−1%, path low −9%). | Feb–Mar 2020 | −14.3% | −20.9% | −30.8% | −27.9% |
| 2022 rate shock Stocks and bonds fell together — BND −13%. The framework’s blind regime. | full-year 2022 | −13.5% | −15.4% | −17.9% | −33.2% |
| 2018 Q4 selloff | Sep–Dec 2018 | −6.4% | −10.2% | −15.9% | −21.9% |
| Current regime Calm regime — every tier positive; the concentrated pick leads. | Oct 2025–Jun 2026 | +5.2% | +7.2% | +10.2% | +16.3% |
| GFC (proxy) Proxy: VT→SPY. The slow deflation crisis where BND did insure (+8%). | Oct 2007–Mar 2009 | −18.1% | −30.4% | −48.9% | −51.6% |
Buy-and-hold end return per episode. Single-thesis = Winner-pick benchmark (100% QQQ — the ex-post winning market).
Document-claim verdicts
| # | Document claim | Verdict | Conf. | Note |
|---|---|---|---|---|
| B1 | Owning the aggregate avoids winner-pick concentration drawdowns | Supported | High | Synthetic worst-profile gap 16.6pp; real crisis gaps 13.6–33.5pp — both pre-registered thresholds met in every episode. Post-hoc paired hardening on seed-matched paths: the winner-pick drew down deeper in 100% of replicas in 5 of 6 profiles, 88% in the rate shock. |
| B2 | "Own many futures" — the bond sleeve carries in equity crises | Partially supported↳ Overstated | High | GFC yes (BND +8%, sleeve +4.0pp); COVID fast crash no (BND −1%, path low −9%). Real VT–BND crisis correlation is slightly POSITIVE (+0.10) — damping comes from low vol, not counter-movement. |
| B3 | The framework is silent on the joint-fall regime (stocks + bonds down together) | Partially supported↳ Understated | High | 2022: BND −13%, the bond-heaviest tier lost −13.5%, and more bonds barely helped (conservative vs offensive only 4.4pp apart). Synthetically, rate shock is the WORST profile of the most conservative tier. |
| B4 | "Time is the primary edge" — long horizons absorb what one-year windows cannot | Descriptively plausible | Medium | All tiers positive over ~18 years (+5.6% to +8.5%/yr) through three crises, while a single 252-day stress window can cost −21%. Direction consistent; the decades claim itself is one sample path, not provable here. |
| B5 | Friction minimisation is the highest-certainty lever | Not testable | Low | No fees, taxes or behavioural gaps are modelled in this engine — the framework’s most load-bearing lever rests on external evidence, not this analysis. |
| B6 | Free-riding on price discovery works as a mechanism | Not testable | Low | A theory claim (Grossman–Stiglitz logic), not a backtest subject — no active-vs-passive cost data in scope. |
Caveats
- Single-factor model with mean R² 0.94 — the four-name universe is so equity-dominated that the factor nearly equals the market; idiosyncratic structure is thin.
- Tail bands indicative: ~17 independent 252-day blocks (effective sample size ≈ 17).
- GFC is a proxy (VT→SPY: US-only stands in for world equity; world fell similarly or deeper in 2008).
- Rebalancing convention matters in the GFC row: daily constant-mix ended up to 4.2pp worse than buy-and-hold; real threshold rules lie between.
- Monte-Carlo spread: rate-shock medians of the balanced/offensive tiers vary ±4pp across seeds; the high-vol benchmark varies up to 5.8pp (reported separately, not mingled with the tiers).
- No fees, taxes or tracking difference modelled — the framework’s central cost lever (B5) is untested here by design.
- All figures in USD.
- Calibration window 2008–2026 — longer than thesis #1 (2014–2026, BTC-bounded), so regime tables are not directly comparable across theses.
Found this useful?
I run bespoke stress analyses and build portfolio frameworks like this one on request.
Commission your own →Descriptive stress-test case study — not investment advice, and not a suitability or recommendation statement. The framework is a falsifiable hypothesis; the portfolio is an illustrative, rule-based expression of its principles for independent analysis. Stress results are model-dependent and not a forecast of future market behaviour.