Maximum Drawdown (MDD) Explained

Investors often focus on average returns and standard deviation (volatility), but these metrics miss something crucial: the worst loss you might endure. Maximum Drawdown (MDD) answers the question: "What's the worst-case scenario I should prepare for?"

Quick definition: Maximum Drawdown (MDD) is the largest peak-to-trough percentage decline a portfolio has experienced during a specific period. It measures the worst outcome in a historical record.

Key Takeaways

• MDD is the single worst drawdown from peak to trough in historical data
• MDD provides a realistic worst-case benchmark for position sizing and asset allocation
• A portfolio's MDD is more informative than average returns alone
• MDD varies dramatically across asset classes, time periods, and strategies
• Understanding MDD helps you size positions for drawdowns you can actually tolerate

The Definition and Calculation

Maximum Drawdown is calculated as:

MDD = (Trough Value − Peak Value) / Peak Value × 100

Example: A portfolio reaches $250,000 (the peak). Later, it falls to $100,000 (the trough) before recovering.

MDD = (100,000 − 250,000) / 250,000 = −60%

The MDD is −60%, or "a 60% drawdown."

Key point: MDD measures the single worst incident in the historical record. It's not an average or a probability—it's the worst thing that actually happened.

How MDD Differs from Volatility (Standard Deviation)

Many investors confuse volatility with drawdown risk, but they're distinct metrics.

Standard Deviation (Volatility): Measures the typical dispersion of returns around the average. A portfolio with 15% volatility means returns typically vary by 15% from the average in any given year.

Maximum Drawdown: Measures the worst sustained loss from peak to trough, not typical variation.

Example comparison:

Portfolio A:

• Annual returns: +10%, −8%, +12%, −5%, +11%
• Average: +6%
• Standard Deviation: 8.5%
• Maximum Drawdown: −8%

Portfolio B:

• Annual returns: +0%, +0%, +35%, −50%, +50%
• Average: +7%
• Standard Deviation: 28%
• Maximum Drawdown: −50%

Both have positive average returns, but Portfolio B has much higher volatility and a much larger drawdown. Volatility and MDD are related but not identical. You need both metrics to understand risk.

MDD Across Asset Classes

Historical Maximum Drawdowns tell us the worst case each asset class has experienced:

Asset Class	Period	MDD
S&P 500 (stocks)	1926–present	−89% (1929–1932)
NASDAQ	1970–present	−83% (2000–2002)
US Bonds (10Y Treasury)	1926–present	−22% (1980–1981)
Emerging Markets	1988–present	−61% (2008–2009)
Real Estate (REITs)	1972–present	−68% (2008–2009)
Gold	1968–present	−65% (1980–2001)
Bitcoin	2011–present	−89% (2021–2022)

Notice:

• Stocks have the highest historical MDD (−89%), but only once every century or so
• Bonds are much lower MDD (−22%), making them portfolio stabilizers
• Emerging markets are volatile, higher MDD than developed markets
• Bitcoin matches stock-market worst-case behavior, yet is much newer

A diversified 60/40 (stock/bond) portfolio's MDD is typically 30–35%, significantly lower than all-stocks but higher than bonds alone.

Rolling Drawdowns vs. Historical MDD

This distinction is important:

Historical MDD: The worst drawdown that actually occurred in recorded data. For the S&P 500 since 1926, that's −89%.

Forward-Looking Worst Case: "What's the worst drawdown that could occur?" This is harder to estimate but is often 10–20% worse than historical MDD, because markets can surprise to the downside.

Financial models sometimes use "rolling MDD analysis." This calculates the MDD across all overlapping periods in the data. For example, if you calculate MDD for each 10-year period over 50 years, you see how worst-case varies across different eras.

The 1926–1950 rolling MDD is much worse (−89% from the 1929 crash) than the 1980–2020 rolling MDD (−57% from 2008). This matters for forward projections: relying on 1980–2020 data might underestimate future tail risk.

Using MDD to Size Positions

MDD is most useful for position sizing. Here's the logic:

1. Determine your actual maximum loss tolerance in dollars: If you have $500,000 and can't afford to lose more than $150,000, that's a 30% loss tolerance.

2. Calculate the maximum position size from MDD: If a stock has a historical MDD of −60%, you can't put more than 50% of your portfolio in it (because a −60% position in a 50% position = −30% portfolio loss).

The formula: Max Position Size = (Your Tolerance MDD) / (Asset MDD)

Example:

• Your tolerance: 30% portfolio loss
• Stock MDD: 70%
• Max position size: 30% / 70% = 43% of portfolio

This means you can position up to 43% in that stock before hitting your personal 30% loss tolerance if the worst case occurs.

However, there's a catch: position MDD can be worse than the underlying asset's historical MDD, because specific events not yet seen could occur. Always use a margin of safety—maybe sizing at 60% of the calculated maximum.

MDD and Regime Shifts

Historical MDD assumes the future will be like the past. Regime shifts can break this assumption.

Example: REITs before 2008:

• Historical MDD (1972–2007): −35%
• 2008–2009 MDD: −68%

The 2008 financial crisis was regime shift—a period when REITs behaved worse than historical data suggested they could. The same applies to stocks, emerging markets, and bonds at various points.

This is why stress testing and scenario analysis matter more than pure historical MDD reliance.

MDD and Psychology

MDD is critically important for psychological reasons. A portfolio that experiences a 20% decline is easier to hold than a 50% decline, even if both fully recover. The deeper the MDD, the more likely investors are to panic-sell.

Consider two portfolios with identical 10% annual returns:

Portfolio A: Steady 10% annually, zero drawdowns (hypothetically)

• Investors hold easily

Portfolio B: Grows 20% in some years, −30% in others, but averages 10%

• Investors panic during the −30% years and often sell

The average is identical, but the psychological experience is different. This means portfolio design should consider not just expected returns but the path to those returns.

A portfolio with lower MDD (say, a 60/40 split with −30% MDD) versus a higher MDD portfolio (say, all stocks with −50% MDD) might underperform by 0.5% annually, but the psychological benefit of the lower drawdown often makes the lower-MDD portfolio the better choice for individuals.

Historical MDD Recovery Analysis

Knowing the worst drawdown helps estimate recovery time.

From earlier articles, we know:

• 30% MDD requires 43% gain = 3.5 years recovery at 10% growth
• 50% MDD requires 100% gain = 7 years recovery at 10% growth
• 70% MDD requires 233% gain = 12 years recovery at 10% growth

If you have a 40-year investment horizon and experience a 50% drawdown early, recovery time of 7 years is manageable. But if you're retiring in 10 years, a 50% drawdown is catastrophic. This is sequence-of-returns risk.

Limitations of MDD

MDD is backward-looking: It tells you the worst that happened historically, not the worst that could happen. The next crisis might be worse than anything in the data.

MDD is path-dependent: A portfolio that declines 30%, recovers, then declines 30% again shows an MDD of 30% even though you experienced −30% twice. The total pain might be greater.

MDD ignores recovery time: An MDD of −50% that recovered in 2 years feels different from an MDD of −50% that took 10 years. Traditional MDD doesn't distinguish.

MDD is sensitive to the data period: Using S&P 500 data from 1926 gives an −89% MDD. Using data from 1980 onward gives a −57% MDD. The chosen period dramatically changes the metric.

Real-World Applications

Hedge Funds: Often marketed on low MDD. A hedge fund with 5% annual returns but −15% MDD might be marketed as "lower risk" than stocks at 10% returns with −50% MDD. Some investors buy this marketing; others correctly note that 5% annual returns are worse than 10% even if drawdowns are lower.

Risk Parity Portfolios: Aim for low MDD by equalizing risk across asset classes. Typical target: −20% to −25% MDD. This comes at the cost of lower long-term returns than stock-heavy portfolios.

Tail Risk Hedging: Some investors explicitly buy insurance (options strategies) to reduce MDD at the cost of 0.5–2% annual returns. Whether this trade-off is worthwhile depends on psychological tolerance.

Common Mistakes

Assuming MDD = future risk: Historical MDD of −30% doesn't mean the future can't produce −60%. It just means −60% hasn't happened yet in that dataset.

Using MDD to justify overconcentration: "This stock only had an −40% MDD, so I'll put 50% of my portfolio in it." But if company-specific risk causes a −70% loss, your portfolio is devastated.

Ignoring recovery time in MDD analysis: A −20% MDD recovered in 6 months is much less painful than an −20% MDD that takes 5 years. Pure MDD ignores this.

Confusing MDD with "expected" loss: MDD is the worst historical loss, not the typical loss. Most drawdowns are 10–25%, not the MDD extreme.

FAQ

Q: How does MDD compare to Value at Risk (VaR)? A: VaR estimates the loss that could occur in a specific percentage of scenarios (e.g., 95% confidence that loss won't exceed 15%). MDD is the actual worst loss that occurred. VaR is forward-looking probability; MDD is backward-looking fact.

Q: What's a "good" MDD? A: Depends on your tolerance. For most investors, a portfolio with MDD of −25% to −35% is acceptable. MDD below −20% requires lower stock allocation. MDD above −50% requires strong conviction and long time horizon.

Q: Should I optimize my portfolio for lowest possible MDD? A: Not exclusively. A portfolio designed solely to minimize MDD might have such low expected returns that you fail to reach your financial goals. Balance MDD tolerance with return needs.

Q: How often does MDD repeat? A: Sometimes years pass between extreme MDD events. The −89% 1929 MDD didn't occur again in the 80+ years of data since. The −57% 2008 MDD was the largest since 1929. One severe MDD per 20–30 years is typical.

Q: Can I use MDD to predict future returns? A: No. Low historical MDD doesn't predict future returns, and high MDD doesn't predict future losses. They measure past volatility, not future performance.

Related Concepts

Calmar Ratio: A metric that divides annual return by MDD. Useful for comparing strategies: higher returns with same MDD wins; same returns with lower MDD wins.

Ulcer Index: A volatility metric that penalizes time underwater (days below previous peak) more than pure drawdown. Captures the psychological pain of prolonged underperformance.

Drawdown Duration: How long the portfolio stays in negative territory. Some MDD events recover in weeks; others take years. This asymmetry is important.

Summary

Maximum Drawdown is the reality check for portfolio design. It asks: "What's the worst-case scenario I've experienced or could experience?" For most investors, an honest assessment of drawdown tolerance—and designing a portfolio around that tolerance—is more important than chasing an extra 0.5% of annual returns through excessive risk.

In the next article, we examine what happens psychologically during a crash—the emotional journey that makes surviving drawdowns so difficult.

Next

The Emotional Phases of a Crash →