The Fiduciary Retirement Plan File: What to Document and Why

A retirement plan is a recommendation, and a recommendation a fiduciary makes should be reconstructable later. If a client questions the advice in three years — or a regulator asks how a number was produced — the planning file has to answer: what did you assume, where did the assumptions come from, and how did you compute the result. A probability-of-success figure with no documented basis is hard to defend, no matter how sound the underlying analysis was.

This guide is a practical checklist of what belongs in the file behind a Monte Carlo retirement plan, and why each item is there.

What the file is for

Run your own numbers — FREE

10,000 Monte Carlo simulations. Forward-looking forecasts from BlackRock, JPMorgan, Vanguard, GMO, Schwab, Invesco. No account needed.

Try QuantCalc Free →

The planning file is not the client report. The client report communicates the recommendation; the file substantiates it. The two have different audiences. The client wants to know what to do; a reviewer — a compliance officer, a successor advisor, the client's attorney — wants to know whether the process was reasonable and the inputs were sourced. A good file lets someone who was not in the room reconstruct the analysis without calling you.

The checklist

A defensible retirement plan file should capture each of the following:

1. The client inputs. Spending path (including any phased changes — go-go, slow-go, no-go years), time horizon and the longevity assumption behind it, current portfolio value and allocation, income sources and their start dates (Social Security, pensions, annuities), and any one-time inflows or outflows.
2. The return and inflation assumptions, with sources. Not just the numbers — where they came from. If you used published capital market assumptions, name the source and the horizon. If you compared across several, record the range. (See the companion post on forward-looking vs historical capital market assumptions.)
3. The sampling regime. How many scenarios were run, the path count, whether returns were drawn correlated or independent, and how inflation was modeled. A 10,000-path correlated run and a 100-path independent run can produce different headline numbers from the same inputs; the file should say which one you used.
4. The result and its precision. The probability of success and the confidence interval around it, so the reader knows whether the number is stable or noisy.
5. The levers and sensitivities. What happens to the result if the client spends more, retires earlier, or markets disappoint early. Documenting sensitivity shows the recommendation was stress-tested, not point-estimated.
6. The date and the version of the engine or tool. Assumptions and tools change; the file should pin the analysis to the moment it was produced.

Why the methodology matters as much as the number

Two advisors can report the same 87% and have done very different work. One sampled 100 independent paths off a single historical average; the other ran 10,000 correlated paths across several published forecast sources and reported a confidence interval. The headline is identical; the defensibility is not. The methodology is what distinguishes a number you can stand behind from one you merely produced.

This is why the sampling regime line on the checklist is not a technicality. Independent return draws understate tail risk because they let a bad equity year and a bad bond year avoid coinciding — yet 2022 showed they can occur together. A file that documents correlated sampling is recording a more conservative, more realistic process.

Making documentation a byproduct, not a chore

The friction with documentation is that it is usually manual — copying assumptions into a memo after the fact, which is exactly when it gets skipped. The more sustainable approach is to use a tool that emits the documentation as a byproduct of running the plan.

QuantCalc's Advisor PRO tier ($249/year) produces two separate artifacts from a single client run: a white-label PDF report for the client, and a four-page methodology supplement for the file. The supplement lists the forecast sources, the sampling regime, the path count, the confidence interval, and the standard disclosures — the items on the checklist above — without re-entering anything. The report goes to the client; the supplement goes in the file. You can see both formats in the live sample at quantcalc.app/advisors.

Frequently asked questions

Isn't the client report enough for the file?

Usually not. The client report states the recommendation; the file has to substantiate it — the sources, sampling regime, and sensitivities a reviewer needs to judge whether the process was reasonable. Those details are typically omitted from a client-facing document by design.

How detailed does the assumption sourcing need to be?

Detailed enough that someone else could reproduce the analysis: the source of the return assumptions, the horizon, and — if you compared several — the range across them. "We used a reasonable long-term return" is not reconstructable; "we ran the plan against six published forecast sources and reported the range" is.

Does QuantCalc generate the file documentation automatically?

Advisor PRO emits a four-page methodology supplement alongside the client report from the same run, covering the sources, sampling regime, path count, confidence interval, and disclosures. It is built to drop straight into the planning file.

QuantCalc is an independent educational tool, not affiliated with, endorsed by, or sponsored by any referenced firm. Return assumptions are derived from publicly available research. Not financial advice.