Benchmark-Neutral Pricing Measure

• Benchmark-Neutral Pricing Measure is a probability measure that defines asset prices as martingales when expressed in units of the growth-optimal portfolio (GOP), ensuring minimal arbitrage-free valuation.
• It extends classical risk-neutral valuation by using the GOP as the numéraire, yielding lower prices—often 25–50% less—and effective hedging strategies in long-dated and incomplete market contexts.
• The approach provides a unified pricing and risk management framework, supported by explicit models like the Minimal Market Model, which facilitates accurate derivative pricing and hedging.

The benchmark-neutral pricing measure is a probability measure in mathematical finance under which all asset prices, when expressed in units of the growth-optimal portfolio (GOP), become martingales. This approach, sometimes called the benchmark approach or GOP-adjoint measure, generalizes and extends classical risk-neutral valuation, especially in settings where an equivalent martingale measure may fail to exist or risk-neutral pricing overstates long-dated contract values. Benchmark-neutral pricing is minimal in an economically meaningful sense: it yields the lowest model-consistent, arbitrage-free price for replicable claims and provides a robust pricing and hedging framework for incomplete markets, long-term insurance products, and asset classes where classical risk-neutral methods are inadequate (Platen, 19 Jun 2025, Platen, 2024, Schmutz et al., 24 Jun 2025, Chen et al., 26 May 2026).

1. Theoretical Foundation and Definition

Let $(\Omega,\mathcal{F},\{\mathcal{F}_t\},\mathbb{P})$ be a filtered probability space. The growth-optimal portfolio $S^*_t$ is the strictly positive, self-financing portfolio that maximizes the expected logarithmic growth rate:

$$\mathbb{E}^{\mathbb{P}}\big[\log(X_T/X_t)\mid\mathcal{F}_t\big]\le\mathbb{E}^{\mathbb{P}}\big[\log(S^*_T/S^*_t)\mid\mathcal{F}_t\big]$$

for any admissible wealth process $X$. The benchmark-neutral measure $\mathbb{Q}^{\mathrm{BN}}$ is defined by using the GOP as numéraire. If $S^{**}_t$ denotes the extended GOP (e.g., including the riskless asset), and $S^*_t$ the stock-only GOP, the density process is

$\Lambda_t = \frac{S^*_t}{S^{**}_t}$

$$\frac{d\mathbb{Q}^{\mathrm{BN}}}{d\mathbb{P}}\Big|_{\mathcal{F}_t} = \Lambda_t$$

This density is a true $\mathbb{P}$-martingale under mild integrability conditions in typical models such as the minimal market model (MMM), ensuring that $S^*_t$0 is equivalent to $S^*_t$1 (Platen, 19 Jun 2025, Platen, 2024).

2. Benchmark-Neutral Pricing Formula

The primary pricing result is that for any contingent claim $S^*_t$2 with finite expected payoff in GOP units, the time-$S^*_t$3 price is

$S^*_t$4

This yields a unique $S^*_t$5-martingale for the benchmarked price process and is minimal among all self-financing supermartingales replicating $S^*_t$6 (Platen, 19 Jun 2025, Platen, 2024, Sun et al., 2019, Chen et al., 26 May 2026). When the claim is replicable, this provides the minimal arbitrage-free price:

$S^*_t$7

This formula is robust to market incompleteness and does not require the existence of an equivalent risk-neutral measure.

3. Comparative Analysis with Classical Risk-Neutral Valuation

Risk-neutral pricing requires the existence of an equivalent martingale measure $S^*_t$8 under which discounted asset prices (denominated in the riskless asset $S^*_t$9) are martingales:

$$\mathbb{E}^{\mathbb{P}}\big[\log(X_T/X_t)\mid\mathcal{F}_t\big]\le\mathbb{E}^{\mathbb{P}}\big[\log(S^*_T/S^*_t)\mid\mathcal{F}_t\big]$$0

However, in models where $$\mathbb{E}^{\mathbb{P}}\big[\log(X_T/X_t)\mid\mathcal{F}_t\big]\le\mathbb{E}^{\mathbb{P}}\big[\log(S^*_T/S^*_t)\mid\mathcal{F}_t\big]$$1 is only a strict local $$\mathbb{E}^{\mathbb{P}}\big[\log(X_T/X_t)\mid\mathcal{F}_t\big]\le\mathbb{E}^{\mathbb{P}}\big[\log(S^*_T/S^*_t)\mid\mathcal{F}_t\big]$$2-martingale (as in the MMM), no such $$\mathbb{E}^{\mathbb{P}}\big[\log(X_T/X_t)\mid\mathcal{F}_t\big]\le\mathbb{E}^{\mathbb{P}}\big[\log(S^*_T/S^*_t)\mid\mathcal{F}_t\big]$$3 exists. By contrast, the benchmark-neutral measure always exists (provided the GOP exists) and often yields significantly lower prices—typically 25–50% below classical risk-neutral prices for long-maturity options and insurance contracts (Platen, 19 Jun 2025, Platen, 2024, Rudd et al., 2018, Chan et al., 2010). Risk-neutral delta hedging in these settings often leads to systematic over-hedging and pronounced profit-and-loss drift over long horizons.

4. Minimal Market Models and Explicit Formulas

The benchmark approach has been extensively developed using the Minimal Market Model (MMM), in which the GOP is modeled as a time-transformed squared Bessel process of dimension four:

$$\mathbb{E}^{\mathbb{P}}\big[\log(X_T/X_t)\mid\mathcal{F}_t\big]\le\mathbb{E}^{\mathbb{P}}\big[\log(S^*_T/S^*_t)\mid\mathcal{F}_t\big]$$4

where $$\mathbb{E}^{\mathbb{P}}\big[\log(X_T/X_t)\mid\mathcal{F}_t\big]\le\mathbb{E}^{\mathbb{P}}\big[\log(S^*_T/S^*_t)\mid\mathcal{F}_t\big]$$5.

This admits explicit transition densities and closed-form expressions for derivative prices. For example, the price at time $$\mathbb{E}^{\mathbb{P}}\big[\log(X_T/X_t)\mid\mathcal{F}_t\big]\le\mathbb{E}^{\mathbb{P}}\big[\log(S^*_T/S^*_t)\mid\mathcal{F}_t\big]$$6 of a European put with strike $$\mathbb{E}^{\mathbb{P}}\big[\log(X_T/X_t)\mid\mathcal{F}_t\big]\le\mathbb{E}^{\mathbb{P}}\big[\log(S^*_T/S^*_t)\mid\mathcal{F}_t\big]$$7 and maturity $$\mathbb{E}^{\mathbb{P}}\big[\log(X_T/X_t)\mid\mathcal{F}_t\big]\le\mathbb{E}^{\mathbb{P}}\big[\log(S^*_T/S^*_t)\mid\mathcal{F}_t\big]$$8 is

$$\mathbb{E}^{\mathbb{P}}\big[\log(X_T/X_t)\mid\mathcal{F}_t\big]\le\mathbb{E}^{\mathbb{P}}\big[\log(S^*_T/S^*_t)\mid\mathcal{F}_t\big]$$9

with closed-form expressibility using the noncentral chi-square law, as well as for zero-coupon bonds and various long-dated products (Platen, 19 Jun 2025, Platen, 2024). Benchmark-neutral formulas are available for complex insurance derivatives, including variable annuities (Sun et al., 2019) and variance swaps under the $X$0 volatility model (Chan et al., 2010).

5. Hedging, Risk-Minimization, and Working Capital

Under the benchmark-neutral measure, any self-financing admissible portfolio in GOP units evolves as a local (or true) martingale. For nonreplicable or complex claims, benchmark-neutral risk-minimizing hedging strategies arise via the Galtchouk–Kunita–Watanabe decomposition, producing a dynamic hedge in tradable assets and a minimal unhedgeable residual. Working capital for a diversified portfolio of liabilities can be monitored and algorithmically refinanced, with asymptotically vanishing per-contract risk under diversification (Schmutz et al., 24 Jun 2025).

In insurance contexts, benchmark-neutral pricing avoids "insurance–finance arbitrages of the first kind": as long as premiums do not exceed benchmark-neutral prices, no unbounded profit with bounded risk is possible via joint insurance–hedging strategies (Schmutz et al., 24 Jun 2025).

6. Practical Applications, Extensions, and Model Estimation

Benchmark-neutral pricing is particularly advantageous in the following settings:

• Long-term guarantee and insurance contracts (variable annuities, GMWBs, long-dated zero-coupon bonds), where it results in lower and more robust capital requirements (Platen, 2024, Sun et al., 2019).
• Regulatory valuation frameworks, where BN pricing provides the model-consistent lower bound; prices above this admit arbitrage if the GOP is tradable (Platen, 19 Jun 2025).
• Markets lacking NFLVR or admitting strict local martingale phenomena.
• General incomplete-market situations, where entropy-minimization principles select the measure "closest" to the real-world law, interpolating between expectation- and replication-based methodologies (McCloud, 2020, Chen et al., 26 May 2026).

Empirically, estimation of the GOP can be approached by maximizing log-returns or by data-driven SDF learning schemes (Chen et al., 26 May 2026). Recursive marginal quantization and related algorithms enable fast, accurate valuation of Bermudan and path-dependent contracts in the benchmark-neutral framework (Rudd et al., 2018).

7. Connections and Distinctions in the Literature

The benchmark-neutral approach encompasses and generalizes classical risk-neutral valuation, Arrow-Debreu state pricing, and forward measures, with the GOP emerging as the universal numéraire underpinning real-world pricing. In incomplete or bubble-prone environments, it provides a canonical pricing law when risk-neutral measures fail or are non-unique (Chen et al., 26 May 2026). Recent literature relates entropy-penalized, model-calibrated, or robust-pricing methods to tilting or centering around the benchmark-neutral measure, instead of the physical or risk-neutral law (McCloud, 2020, Chen et al., 26 May 2026).

A summary table of key benchmark-neutral measure properties:

Property	Benchmark-Neutral Measure	Classical Risk-Neutral Measure
Numéraire	Growth-optimal portfolio (GOP)	Riskless savings account
Existence in incomplete or bubble markets	Yes	Not always
Minimality of price	Yes (for replicable claims)	Not minimal (often larger)
Arbitrage-free	Yes	Yes (if measure exists)
Pricing formula	$X$1	$X$2

The benchmark-neutral measure thus acts as a unifying foundation for modern arbitrage theory, risk management, and financial engineering in general market settings (Chen et al., 26 May 2026, Platen, 19 Jun 2025, Platen, 2024).