Economic Modelling · Valuing benefit guarantees using simulation

Risk-Neutral vs Real-World Valuation: When to Use Each

Updated 11 October 2026 · Fact-checked

Risk-neutral scenarios grow every asset at the risk-free rate and discount at that rate. They give market-consistent prices for guarantees. Real-world scenarios use realistic expected returns, including risk premiums. They are used for capital, risk and profit projections. Same payoff, different purpose: price with risk-neutral, assess risk with real-world.

Understand Risk-Neutral vs Real-World Valuation

A benefit guarantee is an option written by the insurer. To value it by simulation, you first need scenarios of future asset returns and interest rates. There are two ways to build them, and they answer different questions.

Risk-neutral (market-consistent) scenarios are built so that the price of any traded asset equals the expected discounted payoff. Every asset has expected return equal to the risk-free rate. You discount at the risk-free rate. The result matches what the market would charge to hedge the guarantee. You are not saying markets will earn the risk-free rate. It is a pricing device that works because of no-arbitrage.

Real-world scenarios try to reflect what you actually expect to happen. Equities have a higher mean return than bonds, volatility reflects your view of the future, and you may include mean reversion or fat tails. You use them to answer questions such as: how much capital do I need to hold to be 99.5% sure of meeting claims? What is the probability of ruin? What profit will I earn on average?

The key point: the risk premium is the difference. Moving from real-world to risk-neutral changes the drift, not the volatility. Volatility is the same under both measures for the same asset in a standard model.

Calibration differs too. Risk-neutral models are calibrated to current market prices, for example the yield curve and implied volatilities of options. Real-world models are calibrated to historical data and expert judgement about long-term averages. You check a risk-neutral generator with a martingale test: discounted asset values should have an average equal to their starting value.

Key rules to remember

Risk-neutral price: V₀ = E_Q[ e^(−rT) × Payoff ]; Expectation under the risk-neutral measure Q, with constant risk-free rate r. With a stochastic rate, discount along each path.
Risk-neutral drift for a share: dS = r S dt + σ S dW_Q; Drift is r (less dividend yield q if there is one). Volatility σ is unchanged.
Real-world drift for a share: dS = μ S dt + σ S dW_P; μ = r + risk premium. Same σ as above.
Market price of risk: λ = (μ − r) ÷ σ; Links the two measures: dW_Q = dW_P + λ dt.
Martingale test: E[ e^(−rt) S_t ] = S₀; Check by simulation: average the discounted value across paths and compare with S₀ using the standard error.
Standard error of simulated mean: SE = s ÷ √N; s = sample standard deviation of discounted values, N = number of scenarios. A result within about 2 SE of S₀ passes.

How to solve Risk-Neutral vs Real-World Valuation questions

Use this method for any question asking you to choose, build or check a scenario set.

1Identify the purpose. Pricing, hedging or market-consistent valuation means risk-neutral. Capital, solvency, ruin probability or expected profit means real-world.
2State the drift. Risk-neutral: risk-free rate (minus dividend yield). Real-world: risk-free rate plus a risk premium.
3State what stays the same: volatility and correlations are normally unchanged by the measure change.
4Choose the discount rate. Risk-neutral: risk-free rate along each path. Real-world: a rate suited to the task, often the risk-free rate for liabilities, or a risk-adjusted rate for profit.
5Say how the model is calibrated: market prices (yield curve, implied volatility) for risk-neutral; historical data and judgement for real-world.
6Validate the output. Run a martingale test for risk-neutral scenarios and compare with S₀ using the standard error. Check real-world scenarios for realistic means, volatilities and tails.
7Calculate the result as the average discounted payoff, and comment on sampling error and limitations.

Quickest way: Purpose-first decision and martingale check

When to use it: Use in multiple-choice questions and short written parts where you must pick the scenario type or test a generator quickly.

Ask: am I pricing or assessing risk? Pricing means risk-neutral, risk assessment means real-world.
For a martingale test, compute the average of discounted values, then compare with the starting value.
Work out SE = s ÷ √N. If the gap is more than about 2 SE, the generator fails.
If asked about drift, remember: only the drift changes between measures, not the volatility.

Common mistakes in Risk-Neutral vs Real-World Valuation

Saying risk-neutral scenarios predict that assets earn the risk-free rate.
The name suggests a belief about actual returns.
Fix: Say it is a pricing tool. Expected discounted payoff under Q equals the market price. It does not forecast returns.
Using real-world scenarios to price a guarantee that is hedged in the market.
Real-world returns feel more realistic.
Fix: Price for hedging with market-consistent scenarios. Use real-world only for capital and risk questions, or for unhedged risk.
Changing volatility when moving between measures.
Students assume everything shifts with the measure.
Fix: In the standard model, only the drift changes. Volatility is the same under P and Q.
Passing a martingale test by eye without the standard error.
A small difference looks acceptable.
Fix: Compare the gap with SE = s ÷ √N. Use a stated tolerance such as 2 SE.
Discounting risk-neutral payoffs at the expected real-world return.
Mixing up the discount rate with the asset's return.
Fix: Under Q, discount at the risk-free rate. The risk premium is already removed from the drift.
Calibrating a risk-neutral model to historical data only.
Historical data is easy to obtain.
Fix: Calibrate to current market prices, such as the yield curve and option-implied volatilities, so the model reproduces them.

Worked examples

Example 1

A share has S₀ = ₹200. A risk-neutral generator with r = 6% p.a. continuously compounded produces 10,000 one-year scenarios. The mean of the share prices at t = 1 is ₹212.60 and the standard deviation of the discounted prices is ₹50. Does the generator pass a martingale test at about 2 standard errors? (No dividends.)

Show the solution

Discount the mean price: 212.60 × e^(−0.06).
e^(−0.06) ≈ 0.94176, so the discounted mean ≈ 212.60 × 0.94176 ≈ ₹200.22.
Standard error = 50 ÷ √10,000 = 50 ÷ 100 = ₹0.50.
Difference from S₀ = 200.22 − 200 = ₹0.22.
0.22 ÷ 0.50 = 0.44 standard errors, which is below 2.

Answer: The generator passes the martingale test. The discounted mean is about ₹200.22, within 0.44 standard errors of ₹200.

Example 2

A life insurer sells a maturity guarantee. Explain which scenario type it should use (a) to set the price charged for the guarantee which it will hedge, and (b) to decide how much capital to hold against the guarantee. Also state how the share drift differs, if μ = 9% and r = 6%.

Show the solution

(a) The price must match the cost of hedging in the market. Use risk-neutral scenarios, calibrated to market yields and implied volatilities, and discount at the risk-free rate.
In the risk-neutral scenarios the share drift is r = 6%.
(b) Capital depends on how bad outcomes could actually be. Use real-world scenarios with realistic returns, volatility and tails, and measure a percentile or tail measure of the cost.
In the real-world scenarios the share drift is μ = 9%.
The risk premium is 9% − 6% = 3%. With σ, say 20%, the market price of risk is 3% ÷ 20% = 0.15. Volatility stays at 20% in both sets.

Answer: (a) Risk-neutral scenarios with drift 6% for pricing and hedging. (b) Real-world scenarios with drift 9% for capital. Only the drift differs, by the 3% risk premium. Volatility is unchanged.

Exam tips

Start every answer by naming the purpose. Examiners award marks for linking the scenario type to the use.
Write the drift under each measure explicitly, and state that volatility is unchanged.
For a martingale test, always give the standard error and a clear pass or fail conclusion.
Mention calibration sources: market prices for risk-neutral, history and judgement for real-world.
In computer-based questions, show the formula for the discounted average, the code or Excel steps and the result.

Practice questions from Valuing benefit guarantees using simulation

Risk-Neutral vs Real-World Valuation in other exams

The same ground in other exams, if you are preparing for more than one or want another angle on it.

FRM Part IIReal-World vs Risk-Neutral Default Probabilities

Risk-Neutral vs Real-World Valuation: frequently asked questions

What is the main difference between risk-neutral and real-world scenarios?

Risk-neutral scenarios use the risk-free rate as the drift for all assets and are calibrated to market prices. Real-world scenarios use realistic expected returns with risk premiums and are calibrated to history and judgement. The first is for pricing, the second for risk and capital.

When should I use real-world rather than market-consistent valuation?

Use real-world scenarios for capital requirements, ruin probabilities, percentile measures and expected profit projections. Use market-consistent valuation when the aim is a price consistent with the market cost of hedging.

How does a martingale test work?

Discount each simulated asset value back to time 0 and average over all scenarios. The average should equal the starting price within sampling error. Use the standard error, s ÷ √N, to judge the gap.

Does volatility change between the two measures?

In the standard models, no. The change of measure alters the drift by the market price of risk times the volatility. The volatility itself is the same.