Investment question and why it matters

A single signal horizon imposes the same clock on securities that may discover prices at different speeds. The decision-relevant question is therefore not simply whether a volatility-curve dislocation predicts subsequent cross-sectional returns. It is whether the dislocation's useful life varies with liquidity—and whether a representation spanning responsive and persistent information can operate across both a broad US equity universe and its more liquid nested subset.

That distinction matters for portfolio construction. A view calibrated only to rapid repricing may discard slower-moving structure in the broad universe; a persistent view may remain stale among securities where information is incorporated quickly. The supplied evidence motivates this tension, but does not resolve it causally.

Mechanism and testable hypothesis

The proposed mechanism begins with heterogeneous price discovery. Liquid securities tend to incorporate new information quickly, so a volatility-curve dislocation there may have a shorter useful life. Less liquid securities can adjust more gradually, allowing a related dislocation to persist.

A dual-horizon representation joins a responsive view of recent repricing with a persistent view of structural effects. The blend does not create information. Its purpose is to reduce the timing mismatch that can arise when a single cross-section contains materially different adjustment speeds.

This mechanism implies three tests. The nested liquid subset should favor more responsive information; the broad universe should retain value from persistence; and the blended representation should transfer more consistently across the two settings than either component alone. Failure on any of those comparisons would weaken the liquidity-conditioned decay explanation.

Testing methodology and historical evidence

The appropriate design is point-in-time and cross-sectional. Form each universe with information available at the decision date, lag every input, and assign liquidity groups without future information. Compare responsive, persistent, and blended representations in the broad universe and the nested liquid subset. Reserve holdout periods and alternative universe definitions in advance, then examine calendar stability, volatility regimes, liquidity buckets, sector and style neutralization, portfolio breadth, trading activity, concentration, and realistic implementation costs.

Across the supplied 2019–2023 simulation, every reported calendar year produced a positive historical return. Separately supplied headline statistics report an 11.65% historical return, a 2.35 risk-adjusted return, and 11.48% trading activity. These observations support continued investigation; they do not establish that liquidity caused the results or that the results will persist.

The reporting layers require reconciliation. The headline and reconstructed historical-return measures appear to use different conventions. Annual peak-to-trough decline fields do not reconcile with the aggregate maximum peak-to-trough decline, and March 2020 is absent from the monthly table. The daily coverage reports no omitted observations, which does not resolve the missing monthly entry. These differences should be settled before the evidence is treated as validated research.

Portfolio role and diversification logic

The most defensible use is as a candidate cross-sectional ranking input inside a broader equity model. Liquidity can inform how much weight to place on responsive versus persistent information, while the two components can also serve as diagnostics: if their relative value does not change across liquidity tiers, the proposed mechanism loses credibility.

A market-neutral implementation is a possible testing setting, not an established outcome. No supplied evidence demonstrates diversification against an existing portfolio, independence from other signals, or resilience after exposure neutralization. Any allocator considering the signal would need incremental tests against the actual portfolio, including correlations, marginal risk contribution, overlap, concentration, and performance after market, sector, size, volatility, and liquidity controls.

Implementation constraints and frictions

The public evidence does not provide a complete treatment of transaction costs, financing, borrow availability, market impact, or capacity. Those omissions are especially important because the hypothesis deliberately spans liquidity tiers. The segment expected to preserve a dislocation for longer may also be the segment in which trading is more expensive and capacity is lower.

Implementation review should translate trading activity into explicit cost assumptions, vary market-impact and financing inputs, test borrow constraints and short-sale availability, and measure concentration and portfolio breadth by liquidity bucket. Timing rules must reflect decision-date data and executable delays. Capacity should be evaluated under stressed participation assumptions rather than inferred from historical returns or aggregate breadth. A simulated benefit that disappears under plausible costs or cannot be implemented at the intended capital scale is not decision-useful.

Failure conditions and limitations

The thesis would be weakened if the liquid subset does not favor responsive information, the broad universe does not retain value from persistence, or the blend does not improve transfer across the two universes. It would also fail if alternative liquidity definitions reverse the ordering, pre-reserved holdout results do not persist, or performance vanishes after sector and style neutralization.

Other rejection conditions are economic rather than statistical: plausible trading costs, market impact, financing, borrow constraints, or short-sale frictions could consume the simulated benefit; combining horizons could dilute meaningful extremes; or changes in market structure and volatility regimes could break the relationship. Universe selection, survivorship effects, or data revisions may explain part of the association.

This is a historical simulation confined to a broad US equity universe. It cannot establish causality and may not transfer to other markets or security types. The supplied evidence is pending independent review, contains unresolved reporting differences, and lacks the out-of-sample, regime, exposure-neutralization, and cost-stress evidence needed before deployment. Public reporting intentionally omits proprietary formulas, identifiers, and implementation details, so the article alone is not sufficient to reproduce the underlying study.

Research attribution: Pro Quant Trade Research Team.