Risk

Volatility, Drawdown, VaR and Sharpe in Illiquid Assets

Volatility measures the dispersion of returns, drawdown measures peak to trough loss, value at risk estimates a loss threshold at a given confidence level, and the Sharpe ratio measures excess return per unit of volatility. All four rely on observable, frequently priced returns, so they tend to understate risk in illiquid assets whose reported values are smoothed appraisals.

What does volatility measure?

Volatility is the standard deviation of periodic returns, usually annualized. It measures how widely returns vary around their average and is the most common single number used to describe investment risk. Monthly volatility is annualized by multiplying by the square root of twelve, which assumes returns are independent from month to month. Volatility treats upside and downside surprises equally and says nothing about the shape of the distribution, so two assets with the same volatility can have very different exposure to large losses.

Volatility is useful for comparing liquid assets and for building portfolios, because it combines with correlations to estimate portfolio risk. Its limits appear when returns are skewed, when large losses cluster in crises, or when prices are not observed often. A strategy that sells insurance like protection may show low volatility for years and then suffer a sudden large loss, which a volatility estimate based on calm periods will not anticipate. For that reason, volatility should be supplemented with measures that focus on the tail of the distribution.

What is maximum drawdown?

Drawdown is the percentage decline from a previous peak in value to a subsequent trough. Maximum drawdown is the largest such decline over a measurement period. It captures something volatility misses: the path of returns and the experience of an investor who bought at the top. Drawdown also connects directly to practical constraints such as margin calls, redemption pressure, covenant tests and the investor's willingness to stay invested. A related measure, time to recovery, records how long the portfolio takes to regain its prior peak.

Maximum drawdown is backward looking and depends on the specific history observed. A fund with a short track record that has not experienced a crisis may show a small drawdown simply because it has not yet been tested. Comparing drawdowns across strategies is only meaningful over the same periods and with similar pricing frequency. Investors should therefore read drawdown statistics alongside the length and composition of the track record. A manager that has operated only in rising markets has not yet demonstrated how it behaves when liquidity disappears and asset prices fall together.

How is value at risk calculated and what are its limits?

Value at risk estimates the loss that a portfolio is not expected to exceed over a given horizon with a given confidence level. A one month 95% VaR of $3 million means that, under the model's assumptions, losses should exceed $3 million in only about one month out of twenty. VaR can be calculated parametrically from volatility and an assumed distribution, historically from actual past returns, or by simulation. It is widely used for risk limits and regulatory capital because it summarizes risk in a single dollar figure.

VaR's main weakness is that it says nothing about how bad losses are beyond the threshold. Two portfolios with the same VaR can have very different losses in the worst 5% of outcomes. Expected shortfall, also called conditional VaR, addresses this by averaging losses beyond the VaR level. Both measures depend on the assumed distribution and historical data, and they tend to understate risk when the data omit severe stress periods or when correlations rise in a crisis. Backtesting, which counts how often actual losses exceed the VaR estimate, provides a practical check on model calibration.

What does the Sharpe ratio tell you?

The Sharpe ratio divides the return above a risk free rate by the volatility of returns. It measures how much excess return an investment provides for each unit of total risk taken, and it is a standard way to compare managers and strategies. Higher is better, all else equal. Because it uses volatility as the risk measure, it inherits every weakness of volatility: it rewards strategies with smooth but negatively skewed returns, it can be manipulated through return smoothing or infrequent pricing, and it does not distinguish between upside and downside variation.

Variants address some limitations. The Sortino ratio uses downside deviation instead of total volatility, and the Calmar ratio divides return by maximum drawdown. None of them fix the core problem that inputs must be reliable, frequently observed returns. A Sharpe ratio computed on appraisal based quarterly returns is not comparable to one computed on daily market prices. When comparing a liquid and an illiquid strategy, practitioners should calculate both ratios on the same pricing frequency and adjust the illiquid series for smoothing first. Otherwise the comparison systematically favors whichever asset is priced less often.

Why do risk metrics understate risk in illiquid assets?

Private equity, private credit, real estate and infrastructure are valued periodically using appraisals or models rather than daily market prices. Appraisers anchor to prior values and recent transactions, so reported returns are smoothed: they respond slowly and partially to market changes. Smoothing reduces measured volatility, reduces measured correlation with public markets and reduces measured drawdown, which makes illiquid assets look safer and more diversifying than they are. The economic risk of owning a leveraged building or company does not disappear because nobody prices it daily.

Analysts address smoothing by unsmoothing returns, often by assuming that reported returns follow a first order autoregressive process and backing out the underlying return series. Under that assumption, true volatility can be estimated by scaling observed volatility by a factor that depends on the autocorrelation of reported returns. Another approach uses public market proxies, such as listed real estate or leveraged equity indices, to estimate the risk of comparable private exposures. Each method relies on assumptions, so results should be presented as a range rather than a single number.

How should practitioners use risk metrics responsibly?

Risk metrics are best treated as a set of complementary lenses rather than a single answer. Volatility and correlation support portfolio construction, drawdown describes path dependent pain, VaR and expected shortfall set limits, and Sharpe style ratios compare efficiency. For illiquid assets, practitioners should adjust for smoothing, examine leverage and refinancing exposure directly, and run scenario analysis based on economic stresses rather than relying on historical volatility. Liquidity risk itself deserves separate measurement: unfunded commitments, redemption terms and the time required to sell assets can matter more in a crisis than any statistical estimate. Clear disclosure of how each metric was calculated, over what period and with what pricing frequency, prevents misleading comparisons.

Computing risk metrics and adjusting for smoothing

1. A liquid portfolio has a monthly standard deviation of 4.0%. Annualized volatility = 4.0% x square root of 12 = 4.0% x 3.464 = 13.86%.
2. Annual return is 9.0% and the risk free rate is 4.0%. Sharpe ratio = (9.0% - 4.0%) / 13.86% = 0.36.
3. One month parametric 95% VaR on a $50,000,000 portfolio = 1.645 x 4.0% x $50,000,000 = 6.58% x $50,000,000 = $3,290,000, ignoring the expected return.
4. The portfolio rose to $120 per unit and later fell to $90. Drawdown = ($90 - $120) / $120 = -25%.
5. A private fund reports 6.0% volatility with first order autocorrelation of 0.5. Unsmoothed volatility = 6.0% x square root of (1.5 / 0.5) = 6.0% x 1.732 = 10.39%. With 9.0% return and a 4.0% risk free rate, the reported Sharpe is 5.0% / 6.0% = 0.83, but the adjusted Sharpe is 5.0% / 10.39% = 0.48.

How underwriters apply this

• Use volatility, drawdown, VaR or expected shortfall, and Sharpe together, since each misses something.
• Adjust appraisal based returns for smoothing before comparing them with liquid assets.
• Check that the measurement period includes at least one significant stress episode.
• Measure leverage, refinancing exposure and liquidity terms directly for illiquid holdings.
• Prefer expected shortfall over VaR when the loss distribution has fat tails.
• Disclose calculation method, period and pricing frequency with every reported metric.

Frequently asked questions

Keep reading

Educational content written by practitioners. Not investment, accounting, legal or tax advice.