Your 30-Year Plan Is Built on a 7-Year Forecast: What the Horizon Mismatch Costs (2026)

Does it matter that a 30-year plan uses a 7-year forecast?

The plan

One plan, run 11 ways. Only the return path differs between arms; volatilities and the correlation matrix are held constant throughout, so the comparison isolates the horizon treatment rather than confounding it with a change of risk model.

What the mismatch costs

“Stretched” applies the publisher's expected returns to all 30 years. “Own horizon” applies them for the number of years that publisher forecasts, then the 30-year J.P. Morgan set for the remainder.

30-year success rate — the share of 10,000 paths that reached age 95 with money left. The J.P. Morgan row is identical under both treatments because it is already a 30-year forecast; that identity is the study's own control.

How to read it

Three things follow, and the third is the one that matters in practice.

At twenty years the mismatch barely registers

Charles Schwab's 20-year set covers two thirds of the plan, and correcting the last third moves success by 0.4 points. Stretching a twenty-year forecast over thirty is close enough to harmless that it is not worth the argument. Source: www.schwab.com, checked 2026-09-16.

At ten years it is material but not decisive

Between 5.1 and 7.2 points across Vanguard, BlackRock, Invesco. It moves the answer without changing the conclusion: on these forecasts a 4.0% withdrawal still lands well below the success rate the same plan gets from the 30-year set, correction or no correction. Sources: corporate.vanguard.com, checked 2026-09-16; www.morningstar.com, checked 2026-09-16; www.invesco.com, checked 2026-09-16.

At seven years, applied to a mean-reverting forecast, it dominates

GMO goes from 5.0% to 45.7% — the difference between a plan that almost always fails and one that is merely uncomfortable. GMO publishes a 7-year mean-reversion view (www.gmo.com, checked 2026-09-16); stretching it over 30 years assumes the reversion never finishes, which is a far stronger claim than its publisher makes.

Where the money ends up

Median balance at age 95, nominal dollars. A median of $0 means more than half of paths were exhausted before the end — which is why the 7-year row shows $0 under both treatments even though its success rate moves by 40.7 points: the correction rescues paths without lifting the middle of the distribution above zero.

What this implies for a plan

The fix is a term structure, not a different forecast: apply the near-term view for the years it covers, then a long-run set for the remainder. That is what the right-hand column above does, and what the planner's short-term and long-term assumption tabs do.

It matters most where it is least visible. A near-term forecast has its largest effect on the first decade of retirement, which is exactly where sequence-of-returns risk concentrates — so a horizon treatment that is wrong at the start is wrong in the years that count most. And correcting it does not make the near-term view irrelevant: even applied over its own 7 years and no further, GMO's forecast still costs 41.4 points against the J.P. Morgan set alone. The near-term view genuinely matters. Stretching it merely overstates it.

Where this study came from

From a mistake in our own work. The 4% rule stress test ran six published forecasts across 30 years when only one of them is a 30-year forecast, and its most-quoted row was the 7-year one. That page now carries the limitation, the corrected figures and a link to each publisher's own horizon; this study is the general version of the same correction, so the size of the effect can be read off directly rather than inferred.

Download the data

Every figure on this page, in both formats. Published under CC-BY-4.0 — use it with attribution.

Methodology

Assumptions and limitations

• This does not say which forecast is right. It compares horizon treatments, not forecasters. Every arm is run on its publisher's own numbers as published.
• No other income. No Social Security, pension or part-time earnings — a bare portfolio, which makes the arms comparable and the plan harsher than most real ones.
• No taxes or fees, so the return path is the only moving part.
• The anchor is not an endorsement. J.P. Morgan's set is used after the near-term window because it is the only 30-year forecast in the roster, not because it is the best one.
• One plan. $1,000,000, 60/40, 4.0% withdrawal, 30 years. The direction of the effect is structural, but its size depends on the plan — a shorter horizon or a lower withdrawal rate leaves the correction less room to matter.
• Publishers revise. These are the sets the engine carried in September 2026; each row links to the publisher's own page, which is the version to check against.

Frequently asked questions

Does it matter that a 30-year retirement plan uses a 7-year capital market forecast?

It depends entirely on how short the forecast is. On one $1,000,000 plan spending $40,000 a year from 65 to 95, 10,000 paths per arm: stretching Charles Schwab's 20-year set across all 30 years cost 0.4 points of success (82.3% against 82.6%), which is not worth arguing about. Stretching a ten-year set cost 5.1 to 7.2 points. Stretching GMO's 7-year set cost 40.7 points — 5.0% against 45.7% — which is the difference between a plan that almost always fails and one that is merely uncomfortable. Sources: www.schwab.com, checked 2026-09-16; www.gmo.com, checked 2026-09-16.

Why is the mismatch so much worse for a 7-year forecast than a 20-year one?

Two reasons compound. The first is arithmetic: a 20-year set covers two thirds of a 30-year plan, so only a third of the horizon is being invented, while a 7-year set leaves 23 years to invent. The second is what the short forecast says. GMO publishes an explicitly mean-reverting 7-year view, so stretching it assumes the reversion never finishes — a far stronger claim than its publisher makes. Source: www.gmo.com, checked 2026-09-16.

What is the fix?

A term structure, not a different forecast. Apply the near-term view for the number of years it was published for, then a long-run set for the remainder. That is what the corrected column in this study does, and it is what the planner's short-term/long-term assumption tabs do.

Does correcting the horizon make the near-term forecast irrelevant?

No, and that is the part most likely to be misread. Applied correctly — over its own 7 years and no further — GMO's view still costs 41.4 points against running the 30-year set alone (45.7% against 87.1%). A weak first decade matters precisely because sequence-of-returns risk concentrates there. Stretching it merely overstates it. Source: www.gmo.com, checked 2026-09-16.

Which 30-year forecast should the remaining years use?

This study uses J.P. Morgan because it is the only 30-year set in the roster, not because it is the best one. Any long-run set would do; the point of the comparison is the horizon treatment, and the same anchor is used in every arm so the arms differ only in their near-term years. Source: am.jpmorgan.com, checked 2026-09-16.

Does this change the conclusion of the 4% rule stress test?

It changes one of its numbers materially and leaves the rest standing. That study ran every forecast across all 30 years, so its GMO row read 5.0% where the horizon-correct figure is 45.7%. The ten-year rows move by 5.1 to 7.2 points and keep their ordering, and the overall finding — that today's published forecasts put 4% well below its historical success rate — survives the correction. The stress test now carries the limitation and these figures. Source: www.gmo.com, checked 2026-09-16.

Related research

Changelog

• v2026.1 (2026-09-18) — initial release. 5 forecasts at 7, 10, 10, 10, 20 years plus the 30-year anchor, two horizon treatments each, 10,000 paths per arm.

Last updated 2026-09-18. Dataset license: CC-BY-4.0. QuantCalc is an independent retirement-planning research project. The publisher names above identify the capital-market assumptions the engine uses; QuantCalc is not affiliated with, endorsed by, or sponsored by any of those firms, and all trademarks belong to their respective owners. Educational research, not financial, tax, or legal advice.

@misc{quantcalc2026your30yearplanisbuiltona7yearforecastwha,
  title  = {Your 30-Year Plan Is Built on a 7-Year Forecast: What the Horizon Mismatch Costs (2026)},
  author = {{QuantCalc Research}},
  year   = {2026},
  url    = {https://quantcalc.app/research/cma-horizon-mismatch-2026/},
  note   = {Accessed <date>}
}