J.P. Morgan vs Research Affiliates: Which Capital Market Assumptions Should You Use?

The Two Sources at a Glance

Both are genuinely free to access. J.P. Morgan's full 100-page PDF report circulates widely and is accessible with registration on their asset management website. Research Affiliates' interactive tool is completely open — no registration required.

Where They Agree

Before diving into disagreements, it's worth noting the areas of consensus. When two independent teams with different methodologies reach similar conclusions, that convergence carries weight.

International equities outperform US

Both sources expect developed ex-US and emerging market equities to deliver higher returns than US large-cap stocks over the next decade. J.P. Morgan forecasts 7.4% for EAFE and 7.7% for EM. Research Affiliates sees similar figures — approximately 7.7% nominal for developed ex-US and 7.5% for emerging markets.

Fixed income is back

After a decade of near-zero yields, both see bonds offering meaningful returns again. US aggregate bonds sit around 4.6\u20134.7% expected return from both providers — a genuine allocation-worthy return for the first time since before the GFC.

The 60/40 portfolio works again

Both sources, by their own methodologies, suggest that balanced portfolios will deliver reasonable returns going forward — though they disagree on exactly how much (more on this below).

The Big Disagreement: US Large-Cap Equities

This is where it gets consequential. The single largest allocation for most investors — US large-cap equities — is where these two sources diverge most dramatically.

Why They Disagree: A Philosophical Divide

The 3.6-percentage-point gap on US equities isn't a data error. It reflects a genuine intellectual divide about what drives long-term returns.

J.P. Morgan: fundamentals first

J.P. Morgan decomposes expected returns into building blocks: revenue growth, buyback yield, dividend yield, valuation drag, margin pressure, and dilution. For US large-cap, they see 6.0% revenue growth, a healthy buyback yield, and only a modest −2.0% valuation headwind — resulting in a 6.7% total return. Their view is that US corporate profitability, buyback culture, and AI-driven productivity gains justify structurally higher valuations than historical averages.

Research Affiliates: valuations first

Research Affiliates centres their model on the Shiller CAPE ratio. With US CAPE at roughly 34 — nearly double the long-term average — their model assumes it reverts halfway toward the historical mean over 10 years. This valuation compression alone subtracts approximately 2.5% annually from returns, pulling the total down to around 3.1%.

Rob Arnott's team argues that starting valuations are the single strongest predictor of subsequent 10-year returns, and that narratives about “this time is different” — whether the internet in 2000 or AI in 2025 — have historically been poor justifications for ignoring valuation.

Who Has the Better Track Record?

This is harder to answer definitively than you might think.

J.P. Morgan has 30 years of published assumptions, giving them the longest track record in the industry. Their estimates have been broadly reasonable over full cycles, though like everyone, they didn't predict the magnitude of US outperformance driven by mega-cap tech concentration.

Research Affiliates claims their valuation-dependent model improves out-of-sample forecasts by 30% versus naive historical averages. Their AAI tool has a 10-year track record — still relatively short for evaluating 10-year forecasts, but their underlying methodology (CAPE-based reversion) has decades of academic support dating back to Robert Shiller's original research.

The honest answer: neither has been consistently right, and both have added value versus assuming the future will look exactly like the past.

What Should You Actually Do?

1. Don't pick one — use both

The disagreement itself is informative. If you build a portfolio that works under both sets of assumptions, you're building something more robust than if you optimise for either one alone. Run your portfolio through both and see how sensitive your outcomes are.

2. Use the disagreement as a stress test

If your retirement plan only works under J.P. Morgan's 6.7% US equity assumption and breaks under Research Affiliates' 3.1%, you have a concentration risk problem regardless of who's right.

3. Consider the consensus

Both sources agree that international equities and fixed income offer competitive returns right now. The strongest conviction should come from where they converge, not where they diverge.

4. Watch the update cadence

J.P. Morgan publishes annually in October. Research Affiliates updates monthly. If markets move significantly mid-year, RA's estimates will shift with them while J.P. Morgan's remain fixed. Neither approach is strictly better — annual stability is useful for institutional planning, while monthly updates are more responsive to reality.

The Bigger Picture

The J.P. Morgan vs Research Affiliates debate is really a proxy for a deeper question: do valuations matter more than fundamentals for predicting long-term returns?

History suggests both matter, but over different horizons. Fundamentals (earnings growth, productivity, demographics) dominate over 20\u201330 years. Valuations (CAPE, price-to-book, yield) dominate over 7\u201312 years. Since both providers forecast roughly 10-year returns, valuations should carry significant weight — which arguably gives Research Affiliates the more theoretically grounded approach for this specific horizon.

But theory and practice diverge. US markets have sustained elevated CAPEs for over a decade now, and investors who bet on mean reversion in 2015 have been wrong for 10 years running. J.P. Morgan's willingness to incorporate structural change (AI, buyback culture, intangible-heavy business models) may better capture a genuine regime shift.

The intellectually honest position is uncertainty. And the practical response to uncertainty is diversification — across assets, across geographies, and across assumption sets.

Summary: At a Glance

Both are valuable. Neither is complete. Use them together, stress-test your assumptions, and build portfolios that survive both worldviews.