What Real Return Should You Assume for Stocks? About 5%, Not 7%

A globally diversified equity portfolio has historically earned about 5 percent real per year over the long run. Not the 6 to 7 percent number popularized by U.S.-only history. Two pieces of the gap are simple to explain. The U.S. was an outlier. And terminal-wealth math compounds at the geometric mean, not the arithmetic mean. A planning number that uses 6 to 7 percent real for a globally diversified portfolio is borrowing optimistic numbers from the most successful market of the most successful century, then compounding them with the wrong statistic.

This is the short version of a longer story covered in Do 200 Years of Stock Returns Still Matter? The argument here is narrower: where each number actually comes from, why the gap between 5 and 7 is real and persistent, and what to do about it in a financial plan.

Where the “6 to 7 Percent Real” Number Comes From

The popular “stocks return 7 percent real” rule of thumb has a clear lineage. Jeremy Siegel’s Stocks for the Long Run popularized roughly 6.5 to 7 percent real for U.S. equities over very long horizons. CRSP-based Ibbotson series anchor U.S. modern returns at a similar level. UBS Global Investment Returns Yearbook 2026 reports U.S. equities compounded at 6.6 percent real per year from 1900 through 2025 (UBS Yearbook 2026 public summary).

That number is real and verified. It also describes one country in one century. Reusing it as the assumption for a globally diversified portfolio is a categorical mismatch.

What the Global Record Actually Shows

Outside the United States, the picture changes materially. Edward McQuarrie, citing the Dimson-Marsh-Staunton Credit Suisse Yearbook in his 2024 SSRN paper (McQuarrie SSRN, p. 13): $1 invested in U.S. equities in 1900 grew to $1,937 by the end of 2020. The same dollar invested in equities outside the United States grew to just $179, less than a tenth of the U.S. value. The annualized real return on world ex-U.S. equities over those 120 years was 4.4 percent.

A global market-cap-weighted blend of the two depends on the weighting. With U.S. equities at roughly 60 percent of world market cap and ex-U.S. at 40 percent, the blend lands at 0.60 × 6.6 + 0.40 × 4.4 ≈ 5.7 percent real. The historical DMS World index has run closer to 5 percent because the U.S. weight in the index was lower across most of the sample. Either way, the global figure sits well below the U.S. alone.

Recent decades reinforce the point. UBS Yearbook 2026 reports that global equity investors earned 3.5 percent real annualized over the first quarter of the 21st century (2000 through 2024), materially below the 20th-century U.S. norm.

Geometric vs. Arithmetic: The Methodology That Explains Half the Gap

The other source of confusion is the difference between the arithmetic mean of annual returns and the geometric mean (the compound annual growth rate). For long-run investor experience, the geometric mean is the rate that drives terminal wealth.

Jordà, Knoll, Kuvshinov, Schularick, and Taylor (2017) produced one of the most thorough long-run global return databases: The Rate of Return on Everything, covering 16 advanced economies from 1870 through 2015. Their Table 3 reports two equity numbers for the full sample:

Arithmetic real: 6.89 percent per year
Geometric real: 4.64 percent per year

The 225-basis-point gap is the volatility drag from compounding. With annual equity volatility around 17 percent, the arithmetic-to-geometric drag is approximately σ²/2 ≈ 0.17²/2 ≈ 1.45 percentage points per year. The math is not a curiosity. A reader who plugs an arithmetic average into a compounding model gets a materially wrong terminal wealth answer because the arithmetic mean ignores the path effect of negative returns.

Half of the gap between “7 percent real” and “5 percent real” comes from this single methodological choice. Anyone quoting a long-run equity return should specify which statistic they are using.

Realized Returns May Overstate Forward Expectations

A subtler caveat applies even to the verified geometric numbers. Twentieth-century realized stock returns were boosted by what finance researchers call rerating: a roughly one-time set of improvements in corporate governance, accounting disclosure, investability of foreign markets, broader diversification, and falling required returns as risk premia compressed. Some of the past return reflected expanding valuation multiples and shrinking discount rates, not repeatable cash-flow growth.

The cleanest way to see this is through the Gordon-growth decomposition: long-run nominal return ≈ shareholder yield + nominal earnings growth + the annualized impact of any change in the price/earnings multiple. When multiples expand over a sample window, the realized return exceeds what shareholder yield plus growth alone would have produced. That excess does not repeat in a new investor’s forward window unless multiples expand again from the new starting point.

Aswath Damodaran’s annual implied equity-risk-premium work captures this directly. His January 2026 data update puts the S&P 500 implied ERP at 4.23 percent over the Treasury bond rate. Combined with current real bond yields, that implies a forward U.S. equity real return materially below the backward-looking 6.6 percent. Vanguard’s 2026 capital markets outlook lands in the same neighborhood: 4 to 5 percent nominal annualized for U.S. equities over the next 5 to 10 years, which corresponds to roughly 2 to 3 percent real at a 2 percent inflation assumption.

The implication for planning is direct. Even if you correctly use the global geometric historical mean, forward expectations may be lower because the conditions that produced part of the 20th-century equity premium were one-time improvements rather than repeatable mechanisms. Do Stock Valuations Still Matter? unpacks the forward-return decomposition with an interactive calculator.

Stocks Are Not a Reliable Short-Run Inflation Hedge

A common companion claim to “stocks return 7 percent real” is that equities reliably hedge inflation. The long-run record is more nuanced.

UBS Yearbook 2026 finds that equities and bonds have both delivered higher real returns when inflation was low and growth was high. High inflation impaired real equity performance. The Jordà et al. paper is more specific: outside the interwar deflation period, equity returns have co-moved negatively with inflation in almost all eras (FRBSF WP 2017-25, p. 4).

Stocks have historically beaten inflation over long horizons. They have not been a dependable short-run hedge against an inflation shock. The 1970s are the most-cited counterexample: U.S. equities had negative real returns through that decade even though long-run real returns recovered after the early 1980s. For investors with specific multi-year liabilities tied to inflation, direct CPI-linked instruments (TIPS and I Bonds) do the hedging job that equities historically have not. See The Best Inflation Hedges Are Boring.

What This Means for Your Planning Number

A globally diversified equity portfolio anchored on 4 to 5 percent real in retirement projections, rather than 6 to 7 percent, produces three concrete differences:

More conservative withdrawal rates. Sustainable withdrawal rates depend directly on assumed forward returns. Lowering the equity assumption from 7 to 5 percent real shifts the SWR analysis meaningfully toward more cautious numbers. See Safe Withdrawal Rate.
Larger required savings to hit a given FI number. Lower assumed returns translate directly into higher required contribution rates and longer accumulation horizons. See Your FI Number for the math that ties spending and SWR to the portfolio size you actually need.
Plan robustness across regimes. A plan built on U.S. 20th-century returns assumes that the next 30 years will look like the most favorable century of the most favorable market. Building the plan on 4 to 5 percent real means it survives a wider range of historical regimes, including the ones that produced ex-U.S. and post-1970s outcomes.

A defensible sensitivity test runs the plan at three return levels: 6 percent real (optimistic, U.S. 20th-century historical), 4.5 percent real (central, globally diversified long-run mean), and 2.5 percent real (conservative, post-2000 global realized). If the plan only works at 6 percent, it is a bet on one country’s century rather than a plan.

Frequently Asked Questions

Why use 5 percent real instead of the U.S. historical 7 percent?

Two reasons. First, the 7 percent figure is U.S.-only and century-specific; the global record is closer to 5 percent because non-U.S. equities returned roughly 4.4 percent real over the same span. Second, the 7 percent figure often comes from an arithmetic mean rather than a geometric mean; geometric (CAGR) is the rate that drives terminal-wealth compounding, and historical geometric means cluster closer to 5 percent.

What’s the difference between geometric and arithmetic returns?

The arithmetic mean is the simple average of annual returns. The geometric mean is the compound annual growth rate. With positive volatility, the geometric mean is always lower than the arithmetic mean by approximately σ²/2 (variance over 2). For equity returns with annual volatility around 17 percent, the gap is roughly 1.4 percentage points per year. Use the geometric mean for terminal-wealth calculations.

Should I use the same return assumption across the whole horizon?

Long-run averages are the right anchor for long-run plans, but forward returns over the next 5 to 15 years can deviate materially based on starting valuations. With CAPE near 40 in early 2026, Vanguard projects U.S. equities at 4 to 5 percent nominal (about 2 to 3 percent real) over the next decade. A defensible approach is to use a lower number for the first 10-year window and the long-run 4 to 5 percent geometric average thereafter. See Do Stock Valuations Still Matter? for the CAPE-based forward-return logic.

Are stocks really not an inflation hedge?

Over multi-decade horizons, equities have beaten inflation comfortably. Over shorter windows during inflation shocks, they have often performed poorly in real terms. The 1970s in the United States are the canonical example. For specific inflation-linked liabilities, direct hedges (TIPS, I Bonds) are more reliable. Equities are growth assets that have outpaced inflation on average; they are not a contractual inflation hedge.

Where do calculators that say “use 7 percent” get that number?

Most popular retirement calculators anchor on U.S. 20th-century history. Some use arithmetic means rather than geometric means. Some quote nominal rather than real returns and do not adjust for inflation. Each of those choices pushes the headline number upward by 1 to 2 percentage points relative to the global geometric real number that actually drives long-run terminal wealth.

Related Guides

Do 200 Years of Stock Returns Still Matter? covers the regime-dependence story behind the lower global number, McQuarrie’s data-archaeology correction to Siegel, and a calculator that compares terminal wealth across five historical regimes.
Do Stock Valuations Still Matter? covers the CAPE-based forward-return decomposition that captures the rerating caveat directly, with an interactive calculator.
The Case for Global Equity Diversification covers why U.S.-only history is upward-biased and how to think about owning the world.
Safe Withdrawal Rate covers how forward-return assumptions feed directly into sustainable retirement spending.
Your FI Number covers how the assumed return determines the portfolio size you need to hit financial independence.
How to Start Investing uses 5 percent real as the default assumption in its compound-interest calculator and ties it back to the same DMS World data.
The Stock Market Rewards Optimists, But Not the Naïve Ones distills Dimson’s Rational Reminder interview and includes a stress tester that shows how much your plan depends on the equity risk premium assumption.

Key Takeaways

A globally diversified equity portfolio has historically earned about 5 percent real per year over the long run, not the 6 to 7 percent often quoted from U.S.-only history.
The U.S. compounded at 6.6 percent real from 1900 to 2025 (UBS Yearbook 2026). Equities outside the United States compounded at 4.4 percent real over a similar span (DMS via McQuarrie). The market-cap-weighted global blend lands closer to 5 percent.
Geometric and arithmetic returns are not interchangeable. Jordà et al. report 16-economy equity arithmetic real at 6.89 percent and geometric real at 4.64 percent for 1870-2015. Terminal-wealth math uses geometric returns.
Realized 20th-century returns were boosted by one-time rerating effects (governance, disclosure, falling required returns). Forward expectations may be lower than backward realizations even at the global geometric mean.
A planning baseline of 4 to 5 percent real, with sensitivity testing at 2.5 and 6 percent, is more defensible than anchoring on the U.S. 20th-century 7 percent figure. A plan that only works at 7 percent real is a bet on one country’s century.