Spending Guardrails and the 4% Rule

Almost every conversation about retirement spending starts from one number, and almost nobody who repeats it can say where it came from or what it was actually claiming.

This page is about that, and about the alternative that has largely replaced it in planning software — spending guardrails. The short version is that the two are answering different questions, and the profession is openly split on whether the guardrail version cuts spending sooner than it needs to.

Where does the 4% rule come from?

From a single study, and it was narrower than its reputation.

William Bengen published “Determining Withdrawal Rates Using Historical Data” in the Journal of Financial Planning in 1994. He took historical US market returns, applied them to a portfolio, and asked what starting withdrawal rate — raised each year for inflation and never otherwise adjusted — would have survived every rolling thirty-year window in the record.

Note what that question contains. A starting rate. A fixed dollar withdrawal thereafter. A thirty-year horizon. US historical returns. A specific stock-and-bond mix. Change any of those and the answer changes.

Note also what it is not. It is not a prediction, it is not a guarantee, and it was never a recommendation to withdraw a fixed sum for thirty years without looking up. It is a statement about what the historical record would have tolerated, and its own author has revised the figure more than once as the method was extended.

The reason it took hold anyway is that it is a single number you can hold in your head, and there is real value in that. It gives a household a rough sense of scale — of whether it is in the neighborhood or nowhere near it — in one division. That is a genuine use, and it is about the only one it is fit for.

What is wrong with a fixed withdrawal?

Nothing, until the portfolio falls. Then the rule keeps doing exactly what it was told.

A fixed real withdrawal ignores what the portfolio is worth. After a deep decline the same dollar amount is a much larger share of what remains, so the withdrawal rate rises precisely when the balance is least able to support it. The worked example above is nothing more than that arithmetic.

This is the same mechanism as sequence-of-returns risk, seen from the spending side rather than the returns side. A poor stretch early in retirement is dangerous mostly because withdrawals continue through it, converting a temporary decline into a permanent reduction in the capital that has to produce the rest of the income.

Which points at the obvious response: react. A household that reduces its withdrawal after a decline is in a materially different position from one that does not. That is the whole idea behind guardrails.

What are spending guardrails?

A rule agreed in advance about when spending changes and by how much.

The best-known formal version is the set of decision rules Jonathan Guyton and William Klinger published in the Journal of Financial Planning in 2006, under the title “Decision Rules and Maximum Initial Withdrawal Rates.” They set an upper and a lower boundary around the current withdrawal rate. If the rate drifts above the upper boundary — which happens when the portfolio has fallen — the withdrawal is cut by a set amount. If it drifts below the lower boundary, because the portfolio has done well, the withdrawal is raised.

The point of writing it down beforehand is behavioral as much as mathematical. It converts “we should probably spend less” — a conversation nobody wants to have in the middle of a bad market — into a rule that was agreed when nobody was frightened.

The practical consequence people find most surprising is that an adjusting rule generally supports a higher starting withdrawal than a fixed one. That is not a free lunch and it is not a trick. A fixed rule has to be conservative enough to survive the worst historical sequence without ever adapting, because it has promised never to adapt. A rule that can cut has bought itself that safety margin by accepting real reductions later.

Which means the reductions are the mechanism, not a malfunction. A household that takes the higher starting figure and then declines to make the cut when the rule calls for it has kept the risk and given away the protection.

If it helps to have the numbers in one place, I keep them all in the Ultimate Retirement Guide.

Do guardrails actually work better?

This is the honest answer, and it is that qualified people disagree in public.

Michael Kitces has argued directly against the Guyton-Klinger formulation — his 2024 piece is titled “Why Guyton-Klinger Guardrails Are Too Risky For Retirees” — on the grounds that a rule triggered by the withdrawal rate fires when the rate crosses a line, whether or not the underlying plan is actually in trouble. His alternative is to trigger on the plan’s own risk measure rather than on the rate, so the cut arrives when the household’s probability of success has genuinely deteriorated.

That is a substantive disagreement about the trigger, not about the idea of adjusting. Nobody serious is arguing for never adapting. The argument is about what should set the boundary, and reasonable practitioners land in different places.

So treat any confident claim here — including the ones repeated most often — as a position rather than a settled finding. The research is a couple of decades old, it is built on one country’s market history, and the two most-cited frameworks disagree with each other. That is a normal state for a young field and it is not a reason to ignore it. It is a reason to be suspicious of precision.

What does this mean for a specific household?

Three things that survive the disagreement, because all sides accept them.

A withdrawal plan that cannot adapt is making a promise the market did not agree to. Whatever the trigger, the capacity to adjust is what separates a plan from a hope.

Decide the rule before you need it. The value of a written boundary is almost entirely that it exists in advance. A rule invented during a decline is a decision made under exactly the conditions that produce bad decisions.

Ask what a cut would actually mean. A household whose spending is mostly discretionary can absorb a reduction that would be impossible for one whose spending is mostly fixed. The same rule is a mild inconvenience for one and unworkable for the other, and no published study knows which you are.

None of this is a recommendation, and no withdrawal rule protects against every outcome. Which rate, which boundary, and which trigger belong to a plan built on your own figures — including the tax cost of where the money comes from, which is a separate question and often a larger one than the rate itself.