CAGR Explained: How to Measure Growth Over Multiple Years

Define the beginning, ending and elapsed years

The beginning value is the amount at the start of the measurement interval. The ending value is the amount at its end. Use comparable values: revenue should be compared with revenue measured on the same basis, and an investment balance should be compared with a balance that represents the same holdings and cash-flow assumptions.

The number of years is elapsed time, not the number of year labels in a list. From the start of 2021 to the start of 2024 is three years, even though four year labels can appear on a chart. When the dates are not whole years apart, state how you converted the interval into years. Small differences in the period can change the annualized rate.

A beginning value must be greater than zero for this ordinary ratio-based calculation. A positive ending value produces the usual growth or decline interpretation. Negative starting or ending values, such as a business metric that crosses from loss to profit, do not fit the ordinary positive-value CAGR comparison well. Describe those changes directly instead of forcing a percentage to look meaningful.

The formula finds an equivalent annual multiplier

CAGR = [(ending value / beginning value)^(1 / years) − 1] × 100

First divide ending value by beginning value to get the total growth factor. Then take the root corresponding to the number of years. Subtract one to turn the yearly multiplier into a yearly rate, and multiply by 100 for percentage display. This is a geometric calculation; dividing the total percentage gain by the number of years is a different operation.

You can check the result by rebuilding the ending value: beginning value × (1 + CAGR as a decimal)^years. If that reconstruction does not match the ending value within rounding precision, review the inputs or the rate units. A rate of 10% is 0.10 in this expression, not 10.

Worked example: ₹20,000 becomes ₹26,620

Suppose a value starts at ₹20,000 and ends at ₹26,620 after three years, with no additions or withdrawals for this comparison. The ending-to-beginning ratio is 26,620 ÷ 20,000 = 1.331. Its cube root is 1.1. Subtracting one gives 0.1, or a CAGR of 10%.

The reverse check is ₹20,000 × 1.1 × 1.1 × 1.1 = ₹26,620. This gives a useful interpretation of the answer: three years of smooth 10% growth would reach the same endpoint. It says nothing yet about the actual balance observed between those dates.

CAGR is not absolute return or an arithmetic average

Absolute return describes the total percentage change over the whole interval. Here it is (26,620 − 20,000) ÷ 20,000 × 100 = 33.1%. Both “33.1% over three years” and “10% annualized” describe the same endpoints, but they answer different questions. Dropping the time period makes the first number ambiguous.

Dividing 33.1% by three gives about 11.03%, not the 10% compound rate. That division does not account for each year’s change building on the previous year’s balance. Similarly, an arithmetic average of observed annual returns need not reconstruct the ending value. Always label which measure you are using.

The annualized number hides volatility

Both paths in the table have a 10% CAGR, yet one falls from ₹30,000 to ₹18,000 during the second year. Someone who needed to withdraw money at that point would experience the paths differently. A single endpoint-based rate cannot show drawdowns, the timing of losses or how long recovery took.

Before comparing two histories, look at the interval and the values within it. Choosing a particularly weak starting point or a particularly strong ending point can make a headline rate look attractive. A short interval can also make an annualized number seem more dramatic than the small actual change warrants. Include the period and context alongside the CAGR.

Deposits and withdrawals change the question

If a balance rises from ₹20,000 to ₹40,000 because you added ₹20,000, the doubled ending balance is not proof of a 100% investment gain. Plain CAGR does not separate contributions from returns. With multiple dated cash flows, a method designed to account for those cash flows is needed; simply using the first and last account balances can be misleading.

For regular monthly contributions, the SIP guide explains a contribution-and-compounding model. It is a projection under stated assumptions, rather than a way to measure every possible real investment history.