How should illiquid assets like private equity and real estate be incorporated into asset allocation?
For CFA Level III, I know that many institutional portfolios hold significant allocations to illiquid assets. But the standard mean-variance framework assumes you can trade continuously. How do practitioners handle the challenges of incorporating PE, real estate, and infrastructure into their allocation models?
Incorporating illiquid assets into asset allocation is one of the most practically important — and theoretically challenging — topics in CFA Level III. The standard MVO framework breaks down when assets can't be traded freely.
Key Challenges:
- Stale pricing / smoothed returns — Private equity and real estate valuations are updated quarterly (at best), creating artificially low measured volatility and low correlations with public markets. This makes illiquid assets look like a 'free lunch' in an optimizer.
- Illiquidity premium estimation — Investors demand compensation for locking up capital. The question is: how large is this premium, and is it enough to justify the constraints?
- Commitment pacing — You can't simply buy $500M of PE on day one. Capital is committed and drawn over 3-5 years. Actual exposure lags target allocation significantly.
- Rebalancing impossibility — You can't sell PE to rebalance. The allocation will drift based on capital calls, distributions, and NAV changes.
Adjustments for MVO:
| Issue | Adjustment Method |
|---|---|
| Smoothed returns | Unsmooth using Geltner (1993) method: back out true volatility from autocorrelation in reported returns |
| Understated correlations | Use unsmoothed returns to recalculate; typically doubles the correlation with public equity |
| Illiquidity premium | Add 1-3% to expected return (varies by asset type and market conditions) |
| Rebalancing constraints | Model as 'locked' positions that can't be traded; use simulation rather than single-period MVO |
Example — Edgemont University Endowment:
Target: 25% private equity, 15% real estate, 60% public markets.
Smoothed PE statistics: 12% return, 8% volatility, 0.3 correlation with equities.
Unsmoothed PE statistics: 12% return, 22% volatility, 0.7 correlation with equities.
Using smoothed inputs, the optimizer would suggest 40%+ in PE — clearly unrealistic. Unsmoothed inputs produce a more reasonable 20-25% allocation that reflects the true risk.
Practical Approaches:
- Commitment pacing models — Plan vintage year commitments to maintain target allocation through the J-curve. If the target is 25% PE, you might need to commit 8-10% of NAV annually because not all commitments deploy at once.
- Liquidity bucketing — Divide the portfolio into a liquid bucket (public markets) and an illiquid bucket (PE, RE, infra). Only the liquid bucket is subject to rebalancing.
- Secondary market sales — A growing secondary market allows some (discounted) liquidity for PE fund interests.
Exam focus: CFA Level III tests whether you can identify the smoothing bias, explain how to correct for it, and discuss why naive MVO over-allocates to illiquid assets.
For allocation modeling practice, check our CFA Level III materials on AcadiFi.
Master Level III with our CFA Course
107 lessons · 200+ hours· Expert instruction
Related Questions
What exactly is the Capital Market Expectations (CME) framework and why does it matter for asset allocation?
How do business cycle phases affect asset class return expectations?
Can someone explain the Grinold–Kroner model step by step with numbers?
How do you forecast fixed-income returns using the building-blocks approach?
PPP vs Interest Rate Parity for forecasting exchange rates — when do I use which?
Join the Discussion
Ask questions and get expert answers.