CAGR Explained: Why "Average Returns" Lie to You

Start with the puzzle from the intro, with real numbers. You invest 10,000:

The arithmetic average return is (+100 − 50) / 2 = +25% per year. Your actual compound growth is 0% — you ended where you began. Both numbers are "true"; only one describes your money.

The gap exists because percentage gains and losses are asymmetric: a 50% loss requires a 100% gain just to break even. A portfolio that falls 30% needs +43% to recover; one that falls 90% needs +900%. Averaging the percentages hides this asymmetry. Compounding — multiplying the growth factors — exposes it: 2.0 × 0.5 = 1.0, i.e. no growth.

This is why every serious performance figure in finance — fund factsheets, index returns, GDP growth — is quoted as a compound annual rate, and why you should treat any "average annual return" claim with suspicion until you know which average it is.

CAGR answers one precise question: at what constant yearly rate would my money have had to grow to get from the starting value to the ending value? It is the geometric mean of the yearly growth factors, expressed as a single smooth rate:

CAGR = (Ending Value / Beginning Value)^(1 / years) − 1

Example: 10,000 grows to 25,000 over 7 years:

CAGR = (25000 / 10000)^(1/7) − 1
     = 2.5^0.1429 − 1
     = 0.1399  →  ≈ 14.0% per year

Key properties worth internalizing:

• CAGR only needs three numbers — start value, end value, and time. The path in between does not change it.
• CAGR ≤ arithmetic average, always. They are equal only when every year's return is identical. The gap between them grows with volatility.
• CAGR assumes a single lump sum at the start and no cash flows in between. If you invested monthly (a SIP) or added and withdrew money, CAGR is the wrong tool — you need XIRR (covered below).
• Fractional years are fine: use years = days / 365.25. A stock that went from 100 to 130 in 18 months has a CAGR of 1.3^(1/1.5) − 1 ≈ 19.1%, not 30%.

The gap between the arithmetic average and CAGR has a name — volatility drag — and a remarkably useful approximation:

CAGR ≈ Arithmetic Average − (Volatility² / 2)

Example: average return 12%, volatility (std. dev.) 20%
CAGR ≈ 0.12 − (0.20² / 2) = 0.12 − 0.02 = 10%

Two funds with the same 12% average return but different volatility end up in very different places over 20 years on 10,000:

Same advertised "average", half the final wealth. This is why volatility is not just an emotional problem — it is a direct mathematical cost to compound growth, and why diversification (which lowers volatility more than it lowers average return) genuinely is the closest thing to a free lunch in investing.

CAGR silently assumes all your money went in on day one. Most real investors add money over time — monthly plans, bonuses, occasional withdrawals. For any situation with multiple cash flows, the correct measure is XIRR (extended internal rate of return): the single discount rate that makes all your dated cash flows consistent with the final value.

• Use CAGR when: one deposit, no flows — comparing two funds' published performance, measuring a stock you bought once, evaluating "this index went from X to Y in n years".
• Use XIRR when: monthly SIP, irregular top-ups, partial redemptions, dividends reinvested manually — i.e. almost any real portfolio. Every spreadsheet has it: =XIRR(values, dates).

The difference is not academic. Consider a 5-year monthly SIP where the market did all its rising in the final year: the fund's 5-year CAGR might be 10%, but your XIRR could be 18% (most of your units were bought cheap) — or the reverse. Judging a SIP by the fund's CAGR, or a lump sum by your friend's XIRR, produces nonsense comparisons.

One more trap: annualizing short periods. A fund up 6% in three months is sometimes advertised as "26% annualized" (1.06⁴ − 1). Extrapolating one good quarter to a year is marketing, not measurement — regulators in several countries prohibit annualizing sub-year returns for exactly this reason.

A checklist to run every time someone shows you a return number:

• 1. Which average? "Average annual return of 15%" — arithmetic or compound? If the document does not say CAGR (or "annualized"), assume the flattering one was chosen.
• 2. Which window? Performance measured from a market bottom looks heroic; the same fund measured from the prior peak may be mediocre. Check whether the start date is doing the heavy lifting — ask for since-inception and multiple windows (3y, 5y, 10y).
• 3. Nominal or real? A 12% CAGR during 8% inflation is a 3.7% real return (1.12/1.08 − 1). Long-horizon goals — retirement, education — should always be planned in real terms.
• 4. Before or after costs and taxes? Index returns include no fees. A fund citing the index's CAGR while charging 1.8% annually is quoting you a number you cannot receive.
• 5. Survivorship? "Our funds averaged 14%" often excludes the funds that were closed or merged after performing badly. Fund families rarely advertise their discontinued products.
• 6. Point-to-point or investor-weighted? The fund's CAGR is not what its investors earned. Money tends to pour in after good years; investor-weighted returns (XIRR of aggregate flows) routinely lag fund CAGR by 1–2 percentage points.

None of this requires distrusting everyone — it requires knowing that a single return number always compresses away information, and the choice of which information to compress away is rarely random.