Skip to main content

Why Average Returns Lie

If an investment reports a "10% average annual return" over five years, your intuition says: multiply the $100,000 you invested by 1.10 five times, and you'd have about $161,051. But walk into a bank and ask for your balance after that five-year period, and you might find only $130,000 or $90,000, depending on how the returns were actually distributed. The average return of 10% was truthful; your actual outcome was not.

This is not deception. It's mathematics. Specifically, it's the gap between arithmetic mean (simple average) and geometric mean (actual growth rate). That gap grows wider with volatility, and it silently eats wealth every single year.

Quick definition: Average returns refer to the arithmetic mean—the sum of all returns divided by the number of periods. Actual returns follow a geometric progression where each year compounds on the previous year's ending value, which is systematically lower when volatility is present.

Key Takeaways​

  • The arithmetic mean (simple average) is always equal to or higher than the geometric mean (actual return)
  • This gap widens dramatically with volatility—a portfolio with the same average return but higher volatility will compound to much less
  • Two portfolios can have identical arithmetic mean returns but vastly different actual wealth outcomes
  • Volatility drag is not a minor adjustment; it scales with the square of volatility, compounding over years
  • Most reported returns use arithmetic means, which systematically overstate what you actually made
  • Understanding this gap is essential for evaluating fund performance and comparing investment strategies

The Simple Example That Changes Everything​

Let's walk through an example that feels intuitive at first, then shocks you once you see the actual numbers.

You have $100,000. A mutual fund reports it had these annual returns:

Flowchart​

  • Year 1: +60%
  • Year 2: -40%
  • Year 3: +60%
  • Year 4: -40%
  • Year 5: +60%

The arithmetic mean of these returns is:

(60% + (-40%) + 60% + (-40%) + 60%) / 5 = 60% / 5 = 12%

The fund could legitimately advertise "12% average annual return." And you might think: After five years, my $100,000 should become $100,000 × (1.12)^5 = $176,234.

Let's calculate what actually happened:

  • End Year 1: $100,000 × 1.60 = $160,000
  • End Year 2: $160,000 × 0.60 = $96,000
  • End Year 3: $96,000 × 1.60 = $153,600
  • End Year 4: $153,600 × 0.60 = $92,160
  • End Year 5: $92,160 × 1.60 = $147,456

Your actual balance is $147,456, not $176,234. The fund's "12% average" compressed you into a $47,456 gain—not a $76,234 gain. The difference is $28,778—nearly 31% less than the simple average promised.

The actual (geometric) mean return is:

(1.60 × 0.60 × 1.60 × 0.60 × 1.60)^(1/5) - 1
(0.9216)^(1/5) - 1 = 0.9840 - 1 = -1.6%

Your actual return was negative 1.6% annually, despite a positive 12% arithmetic mean. The swings ate the gains. This is why averages lie.

The Mathematics of Volatility Drag​

The relationship between arithmetic mean (AM), geometric mean (GM), and volatility (σ) is expressed by this approximation:

GM ≈ AM - (σ² / 2)

This formula is elegant and terrifying. It says: for every unit of variance (squared volatility), you lose half that much in actual returns.

Let's apply this to a realistic portfolio scenario. Suppose you're comparing two portfolios:

Portfolio A (Low Volatility):

  • Average annual return: 8%
  • Standard deviation: 6%
  • Variance: 0.0036

Portfolio B (High Volatility):

  • Average annual return: 8%
  • Standard deviation: 12%
  • Variance: 0.0144

Both have the same 8% arithmetic mean return. But the volatility drag differs:

  • Portfolio A: Geometric mean ≈ 8% - (0.0036/2) = 8% - 0.18% = 7.82%
  • Portfolio B: Geometric mean ≈ 8% - (0.0144/2) = 8% - 0.72% = 7.28%

Over a 30-year period, this small difference explodes:

  • Portfolio A: $100,000 × (1.0782)^30 = $771,462
  • Portfolio B: $100,000 × (1.0728)^30 = $677,801

Portfolio B's higher volatility, despite identical average returns, costs you $93,661 in final wealth. That's a 12% difference in outcome driven purely by volatility drag.

And notice: both are materially below what the naive arithmetic mean would predict ($100,000 × (1.08)^30 = $1,006,266). The arithmetic mean overstates actual returns for any portfolio with any volatility at all.

Why This Happens: The Compounding Asymmetry​

Remember the asymmetry from the previous article: a 50% loss requires a 100% gain to recover. This is the same principle at work.

When you flip a coin for portfolio returns:

  • Heads: +30% (multiply by 1.30)
  • Tails: -30% (multiply by 0.70)

The arithmetic mean is: (30% + -30%) / 2 = 0%

But the geometric mean is: $(1.30 \times 0.70)^0.5 - 1 = (0.91)^0.5 - 1 = 0.954 - 1 = -4.6%$

A portfolio with "zero average return" (heads half the time, tails half the time) actually declines 4.6% per period. The asymmetry of multiplication ensures that a -30% loss does more damage than a +30% gain undoes.

This is not a small effect. Over 10 coin flips:

  • Arithmetic expectation: 0% × 10 = 0% (you'd expect no change)
  • Actual median outcome: $(0.91)^10 - 1 = -37.6%$ (you've lost nearly 40%)

Real-World Evidence: Fund Performance Reports​

Most mutual funds and ETFs report returns using arithmetic means, which is technically honest but systematically misleading. Let's examine real behavior.

Consider a volatile technology fund that experienced these returns during a 5-year period:

  • Year 1: +35%
  • Year 2: +15%
  • Year 3: -20%
  • Year 4: +40%
  • Year 5: +10%

Arithmetic mean: (35 + 15 - 20 + 40 + 10) / 5 = 16%

The fund can advertise "16% average annual return." Impressive, right?

Let's calculate the actual outcome:

  • End Year 1: $100,000 × 1.35 = $135,000
  • End Year 2: $135,000 × 1.15 = $155,250
  • End Year 3: $155,250 × 0.80 = $124,200
  • End Year 4: $124,200 × 1.40 = $173,880
  • End Year 5: $173,880 × 1.10 = $191,268

Your actual balance is $191,268. The geometric mean return is:

(1.35 × 1.15 × 0.80 × 1.40 × 1.10)^(1/5) - 1 = 2.2956^(1/5) - 1 = 1.1894 - 1 = 18.94%

Wait—in this example, the geometric mean (18.94%) actually exceeds the arithmetic mean (16%). How?

This happens when the volatility is "lucky"—losses occur when the portfolio is smaller, and gains occur when it's larger. This is the best case for volatility. Most of the time, volatility does the opposite: losses occur on a large base (hurting more), and gains occur on a smaller base (helping less). When that's the case, geometric mean trails arithmetic mean by the full variance amount.

The point: the arithmetic mean alone tells you almost nothing about your actual return.

The Time-Weighted vs. Money-Weighted Problem​

There's a second layer of deception in return reporting. Funds report "time-weighted returns," which ignore the timing of cash flows. Your actual return—the "money-weighted return"—depends on when you added or removed money.

For example, suppose a fund has these annual returns:

  • Year 1: +100%
  • Year 2: -50%

Time-weighted return: (100% + (-50%)) / 2 = 25% average

But what if you invested $100,000 at the start of Year 1, and added $100,000 at the start of Year 2?

  • End Year 1: $100,000 × 2.0 = $200,000
  • Start Year 2: $200,000 + $100,000 = $300,000
  • End Year 2: $300,000 × 0.50 = $150,000

Your money-weighted return is: IRR of cash flows = -10% per year.

The time-weighted return was +25% (average). Your actual return was -10%. You lost money overall, despite positive average returns, because you added capital right before the loss year.

This is why "fund performance" and "your actual performance" are often very different numbers.

The Distribution That Matters: When Losses Come​

All volatility is not equal. The timing and sequence of returns matter enormously.

Consider two sequences for a $100,000 investment:

Sequence A (Loss First):

  • Year 1: -30% → $70,000
  • Year 2: +50% → $105,000
  • Year 3: -10% → $94,500
  • Average return: 3.3%
  • Actual return: -1.8% annualized

Sequence B (Gain First):

  • Year 1: +50% → $150,000
  • Year 2: -30% → $105,000
  • Year 3: -10% → $94,500
  • Average return: 3.3%
  • Actual return: -1.8% annualized

Both have identical arithmetic means (3.3%) and identical ending values ($94,500). But the emotional experience and the intermediate values are completely different. Sequence A feels like recovery. Sequence B feels like deterioration.

For an investor who needs to withdraw money mid-period, the sequencing is catastrophic. If you need $30,000 in Year 2 from Sequence A, you withdraw from $105,000 (taking less than 30% of the portfolio). If you withdraw from Sequence B in Year 2, you withdraw from $105,000 (identical). But if you need to withdraw in Year 1, Sequence A forces you to sell from a $70,000 portfolio (taking 43% of it), while Sequence B lets you withdraw from $150,000 (taking only 20%).

Sequence of returns risk is why retirement timing matters so much, and why "average returns" are useless for planning.

How to Spot When Averages Are Lying in Reports​

When you read performance claims, look for these red flags:

Red Flag 1: Only arithmetic mean, no volatility data

If a fund reports "8% average annual return" but doesn't report standard deviation or variance, you're missing critical information. A fund with 8% arithmetic mean and 5% volatility is dramatically different from a fund with 8% arithmetic mean and 20% volatility.

Red Flag 2: Inconsistent reporting timeframes

A fund that reports 1-year, 3-year, 5-year, and 10-year returns might look great on some timeframes and mediocre on others. The cherry-picked timeframe is often the one that looks best.

Red Flag 3: No drawdown data

Knowing maximum drawdown (the worst peak-to-trough decline) tells you how much pain you might endure. A fund with 8% average return and 5% maximum drawdown is vastly superior to a fund with 8% average return and 40% maximum drawdown.

Red Flag 4: Time-weighted returns without cash flow adjustments

If you're evaluating your own portfolio, ignore the reported time-weighted returns and calculate your money-weighted return (internal rate of return). That's your actual return.

A Visual Framework: The Volatility Penalty​

Here's a practical table showing the drag at different volatility levels, assuming identical 10% arithmetic mean returns:

VolatilityArithmetic MeanEstimated Geometric Mean20-Year Final Value
0%10.0%10.0%$673,200
5%10.0%9.875%$653,900
10%10.0%9.5%$611,700
15%10.0%8.875%$557,200
20%10.0%8.0%$485,000
30%10.0%5.5%$302,000

Notice: at 30% volatility (which is normal for growth-oriented portfolios), your actual return is 4.5 percentage points below the reported arithmetic mean. That compounds to a 55% smaller final portfolio over 20 years.

Common Mistakes​

Mistake 1: Comparing arithmetic means without volatility

If Fund A reports 9% returns and Fund B reports 8.5% returns, Fund A looks better. But if Fund A has 25% volatility and Fund B has 8% volatility, Fund B will likely deliver more actual wealth.

Mistake 2: Using simple averages to project future portfolio values

The most dangerous mistake is taking a fund's "average return" and compounding it linearly. If a fund reports 10% average return, don't assume $100,000 becomes $110,000 per year. Each year's growth starts on a different base depending on the previous year's volatility.

Mistake 3: Ignoring sequence risk in retirement planning

If you're retiring and living off portfolio withdrawals, the order of returns matters more than the average return. A sequence of poor returns early in retirement is catastrophic, even if average returns over 30 years are solid.

Mistake 4: Not separating luck from skill

A fund might have higher actual returns than its peers. But if that fund achieved higher returns through lucky sequencing (gains happened to come in "good" years), the advantage might disappear if sequencing reverts.

FAQ​

Q: Is there a way to avoid volatility drag?

A: Not entirely, but you can minimize it. Lower-volatility portfolios automatically experience smaller drag. Additionally, rebalancing (selling winners, buying losers) can reduce the volatility penalty by buying low and selling high.

Q: Does compound interest fix this problem?

A: Compound interest is what creates the problem. If returns were linear (additive), averages would work perfectly. But compounding is multiplicative, which is why volatility's asymmetry matters.

Q: Should I always choose the lowest-volatility investment?

A: Not necessarily. A higher-volatility investment might have higher arithmetic returns that more than offset the volatility drag. But you must compare geometric means, not arithmetic means, to decide.

Q: How much volatility is "too much"?

A: That depends on your goals and time horizon. For a 30-year investor, moderate volatility (10-15%) is usually acceptable if it provides higher expected returns. For a 5-year investor with a specific goal, even 10% volatility might be too much.

Q: Can averages ever be useful for investing decisions?

A: Yes, but only in combination with volatility and risk metrics. A complete performance picture includes arithmetic mean, standard deviation, Sharpe ratio (return per unit of risk), maximum drawdown, and period-by-period returns.

Q: Why do funds report arithmetic means instead of geometric means?

A: Partially because arithmetic means are simpler to calculate and more intuitive to investors. Partially because they're higher, making the fund look better. The SEC doesn't mandate geometric means, so most funds stick with arithmetic.

  • 50% Loss Needs 100% Gain to Recover: The same mathematical asymmetry that makes averages lie also makes recovery asymmetric. See 50% Loss Needs 100% Gain to Recover.
  • Arithmetic Mean vs. Actual Investor Return: A deeper dive into the specific relationship between these two metrics. See Arithmetic Mean vs. Actual Investor Return.
  • How Volatility Silently Eats Returns: Volatility drag is the compounding engine of portfolio underperformance. See How Volatility Silently Eats Returns.
  • Sequence of Returns Risk: For retirees, the order of returns matters more than the average.
  • Standard Deviation and Risk: Understanding volatility metrics is essential to evaluating performance claims.

Summary​

Average returns lie because the arithmetic mean (simple average) is not the same as the geometric mean (actual compounded return). The gap between them grows with volatility—a systematic drag on wealth that compounds every single year. Two portfolios can have identical arithmetic mean returns but vastly different actual outcomes, depending on how volatile each is.

This is why volatility management is not a luxury; it's a necessity. A fund with 8% arithmetic mean return and 5% volatility will compound to vastly more wealth than a fund with 8% arithmetic mean return and 20% volatility, all else equal. The second fund's "average returns lie" in your favor, but your actual dollars received tell a different story.

When evaluating investment performance, always ask: what is the geometric mean return? How much volatility is present? What is the maximum historical drawdown? And crucially: given my time horizon and cash flow needs, how much sequence risk am I exposed to? The arithmetic mean alone answers none of these questions.

Next​

Arithmetic Mean vs. Actual Investor Return →

Authority sources: