What Return Should You Assume?

Real and nominal returns

A nominal return is the raw percentage an investment gained. A real return is what remains after inflation — the change in what the money can actually buy. Over one year the distinction is minor. Over thirty it dominates everything.

Real = (1 + nominal) ÷ (1 + inflation) − 1

  7% nominal, 3% inflation
    exact:         (1.07 ÷ 1.03) − 1  =  3.883%
    subtraction:    7% − 3%           =  4.000%

The consequence of getting this wrong compounds. €500 a month for 30 years at 7% nominal reaches €609,985. The same contributions expressed in today’s purchasing power — a 3.883% real return — reach €339,878. Both are correct answers to different questions, and they differ by €270,107, or 44%.

Where the familiar numbers come from

The figures that circulate — 7%, 8%, 10% — are not properties of investing. Each is a specific measurement of a specific market over a specific window, and knowing which one matters.

• Around 10% nominal is the frequently quoted long-run US large-cap average. It includes inflation and reinvested dividends, and it describes the single best-documented and arguably most fortunate equity market of the twentieth century.
• Around 6–7% real is roughly that same figure with inflation removed. This is the number most retirement research is built on.
• Around 5% real is a common planning figure that applies a haircut for a globally diversified portfolio rather than a US-only one, and for the possibility that the past century was unusually favourable.
• Lower still is what appears when fees, taxes, and a non-US market history are all applied. Studies of other developed markets have generally produced materially lower long-run figures than the US.

None of these are wrong. They measure different things, and the number that belongs in a projection depends on which market you are actually invested in, whether the projection is real or nominal, and what has already been subtracted.

What has to come off

Index returns are gross. An investor receives less, and the gap is not small over decades.

• Fund fees. The ongoing charge comes off the return every year, whatever the market did.
• Platform or account fees. Charged separately by many brokers, often as a percentage of assets.
• Tax. On dividends, on gains, or both, depending on the jurisdiction and the account.
• Transaction costs. Spreads and commissions on every purchase, which matter more for frequent contributions.
• Cash drag. Money sitting uninvested between contribution and purchase earns nothing.

A 7% gross return with a 1.5% total cost is a 5.5% return to the investor, and over thirty years that difference is far larger than it sounds — the fees guide works through exactly how much.

Average returns and actual returns

There are two ways to average a series of returns, and projections need the second one.

Arithmetic mean = the simple average of each year's return
Geometric mean  = the constant rate that produces the same
                  final result as the actual sequence

+50% then −50%:
  arithmetic mean = 0%
  actual outcome  = €100 → €150 → €75, a loss of 25%
  geometric mean  = −13.4%

The geometric mean is always the lower of the two, and the gap widens with volatility. A projection built on an arithmetic average will systematically overstate the result, which is one reason quoted long-run averages should be treated as an upper bound rather than a central estimate.

How the choice is usually approached

There is no correct number, because the return is not knowable in advance. What can be described is how people generally handle that:

• Working in real terms, so every figure in the projection is in today’s money and no inflation adjustment is needed at the end.
• Using a figure below the headline historical average, on the reasoning that fees, taxes, and global diversification all reduce it.
• Running the projection more than once at different rates, and treating the spread as the answer rather than any single run.
• Checking how sensitive the conclusion is — if a plan only works at 8% and fails at 6%, that is information about the plan rather than about the return.