The Assumptions page holds every knob that drives a projection, in one place: the presets, a grid of returns, volatility and allocation, then inflation and simulation. Start with a preset - it sets most of the grid in one move. Adjust the individual rates when you have a specific view, not just to make the plan look better.

The grid gives each asset a row and each attribute a column, so what stocks return, how much they move and how much of them you hold read across a single line. Cash shows a dash under volatility because the engine treats it as a fixed rate. Volatility and correlation dim in Historical Simulation mode, which resamples real market years and ignores them - the values are still there and apply again the moment you switch back.

The defaults sit deliberately between two reference points, and it is worth knowing where. They are below what markets have actually delivered: an 8% stock return against 2.54% inflation is about 5.3% a year after inflation, where every thirty-year stretch since 1872 returned between 6.1% and 9.7%. They are also well above what the major forecasters currently expect over the next ten years, which is nearer 3.5-5.5% before inflation. Neither of those is the wrong number - they answer different questions, and a retirement plan runs long enough that valuations have time to revert.

Assumptions belong to one plan. Every plan you hold carries its own set, which is what lets two plans be genuinely different scenarios - and also means changing your mind about a rate does not reach your other plans by itself. Two Apply to all plans buttons at the foot of the page exist for exactly that: one for the rates and allocation, one for the simulation parameters. They are separate because the two are stored separately, and pushing one is not pushing the other.

Changing them from a results page

Three of these numbers - stocks, bonds and inflation - also sit in a small strip at the top of the pages that show what they produce: the Dashboard, Cash Flow, Time Machine, the What-If Explorer and the optimizers. Nudge one there and the page in front of you moves, instead of coming here, changing it, and going back to see what happened.

It is the same setting, not a preview and not a scenario. A change made there is a change to the plan, and it appears on this page and everywhere else the moment you make it. While you are experimenting, the strip offers Revert, which puts all three back to what they were when you started.

If it is in the way, hide it with the X. The toolbar button beside the plan name brings it back, and remembers which you chose.

Market assumptions

Three named starting points - Cautious, Baseline and Favorable - each setting returns, inflation and volatility together. Unlike the individual rates below, this one is available on every account, free included.

Each shows what your portfolio earns after inflation, blended through your after-retirement allocation - so it moves when you change that mix, and all three move together. The selected one also spells out the two halves, because the surprising half is invisible otherwise: Cautious assumes higher inflation as well as lower returns, so a 6% return can read as 2.9% after inflation and look wrong until you see the 3.0% doing the other half of the work.

That is deliberate, and it is the whole idea. The stretch that ruined retirements beginning in 1966 ran 5.4% inflation against perfectly respectable nominal returns, so a cautious view that lowered inflation alongside returns would barely be cautious at all.

The cautious figures are the ten-year outlook currently published by Vanguard and the other major forecasters. The favorable ones stay inside what markets have really delivered since 1872 - an assumption that beat all of recorded history would not be an assumption. Volatility and the stock-bond correlation come from the same 1872-2025 dataset the historical simulations use, which is why Cautious also turns the correlation positive: in the 1966-1995 window bonds fell alongside stocks rather than cushioning them.

This is not the same as looking at your bad case. Whichever you pick, the simulation still runs two thousand different market sequences around it, and the low percentiles of that run are already your bad case. Choosing Cautious moves what counts as an ordinary market, which then has its own good and bad outcomes around it. The two stack, so reading Cautious as the pessimistic scenario double-counts the pessimism.

You can edit them. Change a rate while Cautious is selected and it becomes your Cautious - switch to Baseline to compare, switch back, and your version is still there. Nothing is overwritten, because your edits belong to the preset you were on. An edited preset says so on the card and offers a reset back to the published figures.

Edits are per plan, so one plan's Cautious is not another's. Apply to all plans under the grid pushes your definitions to every other plan you own, the same way the rates themselves do.

Only the edits are Plus. Choosing between the three is available on every plan.

Custom shows when a plan's values match none of the three - an imported plan, or one built before the presets existed. Ordinary editing no longer produces it, since the preset absorbs the change.

Presets never touch your asset allocation: how much of your portfolio sits in stocks is a decision about your own risk tolerance, not a view about markets.

Average returns

Three rates, all arithmetic annual averages:

• Stocks - default 8%. The historical long-run total return on US large-cap stocks (S&P 500 with dividends reinvested) is roughly 9-11% before adjusting for survivorship and selection bias. Forward-looking forecasts are substantially lower - the major firms currently publish 3.5-5.5% for the next decade, mostly because US valuations are high today - but those are ten-year views, and a retirement plan is longer than that. 8% sits between the two, which is what the Cautious and Favorable presets bracket.
• Bonds - default 4.5%. Close to long-run intermediate-term Treasury total returns. Higher than recent decades but lower than the 1980s yield environment.
• Cash - default 2.5%. Roughly aligned with long-run cash / money-market yields. If you have a specific high-yield savings rate, set it on each Cash account individually - that per-account APY overrides this default.

These are average assumptions. Each Monte Carlo iteration draws random returns around them; the median path will compound slightly below the arithmetic mean because of volatility drag (an inherent property of any distribution with non-zero variance). See Monte Carlo simulations.

Common pitfall: setting stocks to 10-12% to make the plan look better. That inflates your success rate on screen without changing what's actually likely to happen - it just hides risk. If you want to see how the plan holds up under a gloomier or sunnier outlook, use the Market assumptions card above rather than typing a number: it moves inflation and volatility to match, which a lone change to the stock rate does not.

Asset allocation

You set two mixes, each across stocks, bonds and cash, and each row must total 100%. Bonds is the remainder - type a stocks and a cash figure and bonds fills the gap - and the Set cash to the remainder links under the grid, one for each period, set cash to whatever is left over.

• Pre-retirement. The blend the engine uses while you're still earning. Most users keep this stocks-heavy (e.g. 60/40 default) because the long horizon absorbs volatility.
• Post-retirement. The blend that takes over at your retirement age. Conventional advice is to step down the stock share at retirement (e.g. 50/50 default) to reduce sequence-of-returns risk in the early drawdown years.

The shift happens at your retirement age - the engine doesn't currently model a continuous glide path, just the two phases.

Cash appears in two places and they do different jobs. The cash share here is how much of the portfolio sits in cash; the cash return on the Average Returns card is what that share earns. A Cash account carrying its own APY overrides that rate for the account's own balance.

Inflation rates

Two rates, applied differently across the plan:

• General inflation - default 2.54%. Roughly the long-run US CPI average. Used as the default growth rate for recurring expenses and as the basis for today's-dollars to future-dollars conversion.
• Medical inflation - default 3.36%. Healthcare costs have historically risen faster than the CPI; the engine offers a separate rate so healthcare expenses, Medicare premiums, IRMAA tiers, and LTC costs can grow at their own pace.

Each recurring expense line picks one of these or a custom rate - see How inflation works in the plan.

Taxable accounts: turnover

Funds trade inside your account even when you do not, and the gains they realize are taxed to you each year. Turnover is the share of each year's growth in your taxable investment accounts treated as realized and taxed that way, while you still hold the account. The rest of the growth is taxed only when you withdraw, against your cost basis. The default is 20%. Broad index funds and ETFs realize far less, often close to 0%; actively managed funds realize more. It applies to every taxable investment account in the plan, and setting it is part of Plus.

Simulation parameters

These tune how the simulation generates uncertainty around your average return assumptions:

• Simulation mode. The first control on the card, because it decides which of the others apply. Monte Carlo draws each year's returns from a lognormal distribution, and uses the standard deviations and correlation below. Historical Simulation resamples contiguous blocks of real 1872-onward market data, and uses the block length instead. The same toggle sits on the Monte Carlo chart; it is one setting shown in two places. See Historical market simulations for the trade-offs.
• Stock volatility (sigma) - default 17%. The standard deviation of annual stock returns. Higher sigma = wider outcome bands and more frequent extreme years.
• Bond volatility (sigma) - default 8%. Same idea for bonds. Lower because bond returns have historically been less variable.
• Stock-bond correlation - default -0.20. The correlation coefficient between annual stock and bond returns. Negative values mean stocks and bonds tend to move in opposite directions, which is the long-run historical pattern. Plays a meaningful role in how the 60/40 or 50/50 portfolio smooths over time.
• Use fixed seed. When on, every run uses the same pseudo-random sequence, so results are perfectly reproducible across reloads and plan tweaks - changes you see are because of your inputs, not random variation. Turn it off if you want to see how much the result depends on luck of the draw across runs.

Restoring

There is no longer a restore button for the rates themselves, and that is because there is nothing left for it to do: your edits belong to the preset you are on, so the plan already holds what its preset says. The reset that does mean something - back to the published figures for that preset - sits on the preset itself, and appears once you have edited it.

One detail worth knowing on Asset Allocation: bonds has no default of its own because it is the remainder, so restoring clears any bonds figure you typed and lets it go back to filling the gap. That is why the row lands back at 60/40/0 rather than keeping your bond number.

Apply to all plans

Each card can push its values into every other plan you hold. Use it when you have changed your mind about something that is not really a scenario - a preferred inflation rate, a lower long-run return - and want the change to reach the plans you are not currently looking at.

What it touches:

• Only that card's fields. Applying inflation rates does not carry your allocation or your simulation parameters with it. Everything else in each target plan is left exactly as it was.
• Only plans you own and have not archived, and not the plan you have open, which is the source. Archived plans are left alone deliberately - you set them aside - and plans shared with you are never written to, because they belong to somebody else.

The confirmation lists the plans by name, so you can see what you are about to change rather than approving a count. There is no undo, and each plan it touches will recalculate its results the next time you open it. If a plan cannot be written - an encrypted plan while your session is locked, say - it is skipped and named, and the rest still go through.

Iteration count

Fixed at 2,000 iterations. More iterations smooth the percentile bands but add no new information past a few thousand draws. The figure is large enough that the 20th-80th percentile band stabilizes well across reruns.

When to adjust the defaults

The defaults work for most users. Consider changing them when:

• You hold a clearly different long-run market view (e.g. you assume lower returns going forward because of valuation levels).
• You're modeling a non-standard portfolio - higher cash allocation, all-bond mix, alternative investments.
• You want to stress-test against an aggressive crash by raising the sigma values temporarily.

Otherwise, the most useful experiments tend to be on the input side - retirement age, Social Security claiming age, conversion strategy - not the assumption knobs.