Which assumption matters most
Every plan rests on assumptions that are guesses: returns, inflation, house prices, mortgage rates, spending and pay. The sensitivity chart shows which of them the result depends on most, so attention goes where it changes the answer.
The rules
Each assumption is moved down by a fixed step and then up by the same step, with everything else unchanged:
| Assumption | Step each way |
|---|---|
| Investment returns | 1 percentage point |
| Inflation | 1 percentage point |
| Home prices | 1 percentage point |
| Mortgage rates | 1 percentage point |
| Living costs | 10% |
| Pay | 10% |
Rows appear only when they apply: no home, no home price or mortgage row; no pay, no pay row. The rows are ranked by the swing, which is net worth at the end of the plan with the assumption moved up less net worth with it moved down, in today's dollars. A run that stops before the end because money runs out is ranked first.
What the chart measures
The steps are not equal in size or in kind, so the chart compares these particular changes, not the importance of each assumption in general.
Inflation is different from the others. A change to inflation moves the rate for tax brackets, living costs, pay, rent, carrying costs and recurring events together. It leaves investment returns, home price growth and mortgage rates at their nominal rates. Higher inflation therefore means costs rise faster while every asset earns less after inflation, and the end amount is also divided by a larger factor to restate it in today's dollars. That combination is why it is often the largest bar.
What the engine does
- Reruns the whole plan twice for each assumption, through the same calculation as the main projection, so tax, withdrawals and borrowing respond to each change.
- Divides each run's final net worth by (1 + that run's inflation) for each year of the plan.
- Lists first any assumption whose change stops the plan.
- Starts from the plan's own values, so the chart changes with them. The default return and inflation are described in expected returns.
What it deliberately does not
- It does not move two assumptions together, apart from inflation's built-in effects.
- It does not say how likely each change is. For that, see Monte Carlo, which varies returns and home prices by their historical spread.
- It does not include assumptions outside the six, such as tax rates or retirement age.