Skip to content

Level III Core · Asset Allocation to Alternative Investments

Approaches to Modeling Alternatives in Asset Allocation

Updated 8 October 2026 · Fact-checked

Modeling alternatives means choosing a method to set their weights in a portfolio. Mean-variance optimization is simple but fails with smoothed, illiquid returns. Monte Carlo handles liquidity and path effects, factor-based approaches look at underlying risks, and liquidity-aware methods treat illiquidity as a constraint or cost.

Understand Approaches to Modeling Alternatives in Asset Allocation

Alternatives such as private equity, real estate and hedge funds are hard to model because their returns are not like listed asset returns. Many are valued by appraisal or manager estimates, not by daily trades. They also cannot be sold quickly, and their payoffs may be skewed.

The first problem is smoothed returns. Appraisal values move slowly, so reported returns show positive serial correlation. This lowers measured standard deviation and understates correlation with equities. The asset then looks like it has high return, low risk and good diversification. Mean-variance optimization (MVO) will then overweight it. The result is an allocation that rests on a data artifact.

The second problem is illiquidity. MVO assumes you can rebalance freely at no cost. With private assets you commit capital, wait for drawdowns, receive distributions, and cannot rebalance when you want. A single-period model ignores this.

So candidates should know the main approaches. Mean-variance is the base case and is usable only after fixing the inputs, for example by desmoothing returns and adding constraints on weights. Monte Carlo simulation generates many paths for returns, cash flows, contributions and spending. It can model capital calls, distributions, liquidity needs and non-normal outcomes, and it reports the probability of meeting goals. It depends on its assumptions, and it is harder to explain to a client.

Factor-based approaches treat assets as bundles of risk factors such as equity beta, credit, interest rate, inflation and illiquidity. You allocate to factors, then map factors to assets. This shows hidden overlap: a private equity fund is largely equity beta plus an illiquidity premium. It supports risk budgeting across the whole portfolio. A traditional asset class approach instead groups assets by similar characteristics and legal form and sets weights for each class. The risk is that classes with different labels share the same risk drivers. Liquidity-aware approaches add a liquidity budget, a cap on illiquid assets, and stress tests of cash needs. Some also add an explicit liquidity cost or premium.

Key rules to remember

Desmoothing (Geltner) of appraisal returns
R*(t) = [R(t) − φ × R(t−1)] ÷ (1 − φ)
R(t) is the observed smoothed return, φ is the smoothing parameter (the weight on the prior value), and R*(t) is the estimated true return. φ is often estimated from the first-order autocorrelation of the observed returns. Valid for 0 ≤ φ < 1.
Effect of desmoothing on volatility
Compute σ(true) as the standard deviation of the desmoothed series R*(t)
Do not scale observed volatility with a shortcut factor. Desmoothing raises measured volatility when φ > 0, so check that the standard deviation of R* is higher than that of the observed returns.
Portfolio variance (two assets)
σp² = w1²σ1² + w2²σ2² + 2 w1 w2 ρ σ1 σ2
Understated σ and ρ from smoothing reduce σp² and bias MVO toward alternatives.

How to solve Approaches to Modeling Alternatives in Asset Allocation questions

For any question on modeling alternatives, identify the problem in the data or the client, then match the method to it.

  1. 1Read the client's objectives and constraints first: liquidity needs, horizon, spending, and regulatory limits.
  2. 2Identify the asset's return features: appraisal-based or smoothed, illiquid, skewed, or fat-tailed.
  3. 3State the effect on inputs: lower measured volatility, lower correlation, so the Sharpe ratio and diversification are overstated.
  4. 4Choose the approach that fixes the problem: desmooth and constrain for MVO; Monte Carlo for cash flows and goal probability; factors for hidden overlap; liquidity budgets for funding needs.
  5. 5If a calculation is requested, show the formula, the inputs and the result, with the correct sign and units.
  6. 6Give the recommendation and one reason tied to the client, plus one limitation of the method, using the command word asked (identify, explain, justify).

Quickest way: Problem-to-method matching

When to use it: Use this when a vignette lists a weakness and asks which approach or adjustment fits best.

  1. Smoothed or appraisal returns: desmooth; expect higher volatility and correlation.
  2. Capital calls, distributions or liquidity needs: Monte Carlo or liquidity-aware modeling.
  3. Different labels but same risk drivers: factor-based view.
  4. Weights look too high in MVO: add constraints or use resampling and adjusted inputs.
  5. Always link the answer to the client's liquidity and horizon.

Common mistakes in Approaches to Modeling Alternatives in Asset Allocation

  • Saying smoothing raises measured volatility.

    Students mix up appraisal smoothing with noisy market prices.

    Fix: Remember that smoothing hides moves, so measured volatility and correlation are too low and the Sharpe ratio is too high.

  • Applying the desmoothing formula with the wrong sign or forgetting to divide by (1 − φ).

    The formula is memorized loosely under time pressure.

    Fix: Write R* = (R − φ × prior R) ÷ (1 − φ) and check that the result is more volatile than the input.

  • Treating φ as exactly equal to the autocorrelation of observed returns.

    Students read φ as a statistic rather than a model parameter.

    Fix: Say φ is the smoothing parameter, often estimated from the first-order autocorrelation of observed returns. Use the value the question gives you.

  • Claiming Monte Carlo gives the optimal portfolio.

    Simulation sounds more advanced than MVO.

    Fix: Say it evaluates outcomes, cash flows and goal probability for given assumptions, and it is only as good as those assumptions.

  • Treating the factor-based approach as ignoring asset classes.

    Students think factors replace the whole process.

    Fix: Say factors describe risk drivers, then are mapped back to investable assets and used to check diversification and risk budgets.

  • Recommending MVO results without client constraints.

    The model output looks precise.

    Fix: Cap illiquid weights and test the plan against the client's liquidity needs and horizon before recommending.

Worked examples

Example 1

A private real estate index reports returns of 4.0% in the prior period and 6.0% in the current period. The estimated smoothing parameter φ is 0.50. Using the Geltner formula, estimate the unsmoothed current-period return.

Show the solution
  1. Formula: R* = (R − φ × prior R) ÷ (1 − φ).
  2. Numerator: 6.0% − 0.50 × 4.0% = 6.0% − 2.0% = 4.0%.
  3. Denominator: 1 − 0.50 = 0.50.
  4. R* = 4.0% ÷ 0.50 = 8.0%.

Answer: The unsmoothed return is 8.0%.

Example 2

A pension fund's optimizer assigns 30% to private equity because its appraisal-based volatility is low and correlation with equities appears modest. The fund must pay benefits from the portfolio and has uncertain capital calls. Explain two problems with the result and recommend how to model the allocation.

Show the solution
  1. Problem 1: appraisal returns are smoothed, so volatility and correlation are understated and the return per unit of risk is overstated, which inflates the weight.
  2. Problem 2: MVO is a single-period model that assumes costless rebalancing, so it ignores illiquidity, capital calls and the fund's benefit payments.
  3. Fix inputs: desmooth the returns, which raises volatility and correlation, then re-run the optimization with a cap on illiquid assets.
  4. Add Monte Carlo simulation with capital calls, distributions and benefit payments to test the probability of meeting liabilities and the liquidity shortfall risk.
  5. Use a factor view to check that private equity is not duplicating the fund's existing equity risk.

Answer: The 30% weight is inflated by smoothed data and by ignoring illiquidity. Desmooth the returns, cap illiquid weights, and use Monte Carlo with cash flows to confirm the fund can meet benefits and capital calls.

Exam tips

  • Link every weakness to its effect on inputs: lower volatility, lower correlation, higher Sharpe ratio, overweighting.
  • Show the desmoothing formula and the numbers; a correct number alone earns credit, but working protects you from slips.
  • When asked to justify a method, name the client's constraint it addresses, such as liquidity needs or horizon.
  • For compare questions, state one strength and one limitation of each approach in a short line.
  • Answer only the number of points asked, in the order asked.

Approaches to Modeling Alternatives in Asset Allocation in other exams

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

Approaches to Modeling Alternatives in Asset Allocation: frequently asked questions

Why does MVO overallocate to alternatives?

Smoothed returns understate volatility and correlation, so alternatives appear to offer high return with low risk. The optimizer then gives them large weights. It also ignores illiquidity and rebalancing limits.

What is the difference between an asset class and a factor-based approach?

An asset class approach groups investments by similar characteristics and sets weights per class. A factor-based approach allocates by underlying risk drivers such as equity, credit or inflation. It can reveal overlap that labels hide.

How do you treat illiquid assets in asset allocation?

Limit their weight, desmooth their returns, and test the plan against cash needs. Monte Carlo can model capital calls and distributions. A liquidity budget keeps the investor able to meet obligations.

What does desmoothing do to the numbers?

It reverses the averaging effect of appraisals. Estimated volatility and correlation with other assets rise, and the Sharpe ratio falls. The allocation to the asset usually drops.