Maximum Drawdown (MDD)
Maximum Drawdown measures the largest peak-to-trough decline in portfolio value over a specified time period. It captures the worst-case cumulative loss an investor would have experienced and is one of the most intuitive and widely used measures of downside risk.
Overview
Maximum Drawdown is a path-dependent risk measure that reflects the actual investment experience of an investor. Unlike VaR or standard deviation, which are based on return distributions, MDD is computed from the cumulative wealth path and directly measures the largest loss from a historical peak.
MDD is particularly important for hedge funds, commodity trading advisors (CTAs), and any strategy where investor psychology and capital preservation are paramount. A large drawdown not only erodes capital but can trigger margin calls, redemptions, and behavioral mistakes. For example, a 50% drawdown requires a subsequent 100% gain just to break even.
The measure is also used in performance evaluation through the Calmar Ratio, which divides annualized return by maximum drawdown, providing a risk-adjusted performance metric that penalizes strategies with deep drawdowns.
Mathematical Formulation
Drawdown at Time t
Let denote the portfolio value at time . The drawdown at time is the percentage decline from the running maximum:
Note that always (it is zero when the portfolio is at an all-time high and negative otherwise). The running maximum is sometimes called the "high-water mark."
Maximum Drawdown
The Maximum Drawdown over the full period is the most negative drawdown observed:
Equivalently, MDD can be expressed as the maximum peak-to-trough loss:
where is the peak date and is the subsequent trough date. MDD is typically reported as a positive percentage (e.g., "the maximum drawdown was 35%").
Calmar Ratio
The Calmar Ratio relates annualized return to maximum drawdown, providing a risk-adjusted performance measure:
A Calmar Ratio above 1.0 indicates that the annualized return exceeds the maximum drawdown, while values above 3.0 are generally considered excellent. The ratio is commonly calculated over a trailing 36-month window. Unlike the Sharpe Ratio, the Calmar Ratio directly penalizes large peak-to-trough losses rather than overall volatility.
Recovery Time
The recovery time (or drawdown duration) measures how long it takes for the portfolio to recover from the trough back to the previous peak:
where is the peak date and is the trough date. If the portfolio never recovers within the observation period, the drawdown is said to be "unrecovered." Recovery time provides important context beyond the magnitude of the drawdown.
Worked Example
Consider a portfolio with the following monthly values (in thousands):
| Month | Value | Running Max | Drawdown |
|---|---|---|---|
| Jan | $100 | $100 | 0.0% |
| Feb | $110 | $110 | 0.0% |
| Mar | $105 | $110 | -4.5% |
| Apr | $88 | $110 | -20.0% |
| May | $95 | $110 | -13.6% |
| Jun | $112 | $112 | 0.0% |
| Jul | $108 | $112 | -3.6% |
The Maximum Drawdown is 20.0%, occurring from February ($110) to April ($88). The peak was in February and the trough was in April. The portfolio recovered to a new high in June ($112), so the recovery time was 4 months (February to June).
If the annualized return over this period were 15%, the Calmar Ratio would be: .
Advantages & Limitations
Advantages
- Highly intuitive:Directly answers "What was the worst loss from peak to trough?" -- easily understood by all investors.
- Path-dependent: Captures the actual investment experience, including the sequence and duration of losses.
- No distributional assumptions: Computed directly from the price path without assuming any return distribution.
- Behavioral relevance: Reflects the psychological impact of sustained losses, which often triggers investor withdrawals.
Limitations
- Sample-dependent: MDD is a single extreme observation and increases monotonically with the length of the observation period.
- Not forward-looking: Past maximum drawdown may not predict future worst-case losses, especially after regime changes.
- No probability attached: Unlike VaR, MDD does not indicate the likelihood of such a drawdown recurring.
- Single event: Focuses on one worst episode, potentially ignoring multiple significant but smaller drawdowns.
- Frequency sensitivity: MDD computed from daily data will generally be larger than MDD from monthly data for the same period.
References
- Magdon-Ismail, M., & Atiya, A. F. (2004). "Maximum Drawdown." Risk Magazine, 17(10), 99-102.
- Chekhlov, A., Uryasev, S., & Zabarankin, M. (2005). "Drawdown Measure in Portfolio Optimization." International Journal of Theoretical and Applied Finance, 8(1), 13-58.
- Grossman, S. J., & Zhou, Z. (1993). "Optimal Investment Strategies for Controlling Drawdowns." Mathematical Finance, 3(3), 241-276.
- Young, T. W. (1991). "Calmar Ratio: A Smoother Tool." Futures Magazine, 20(1), 40.
- Magdon-Ismail, M., Atiya, A. F., Pratap, A., & Abu-Mostafa, Y. S. (2004). "On the Maximum Drawdown of a Brownian Motion." Journal of Applied Probability, 41(1), 147-161.