Guardrails (Guyton-Klinger): flexible spending rules, with a worked example
The 4% rule asks you to spend a fixed, inflation-adjusted amount no matter what markets do. That's safe, but it means choosing a low starting number to survive the worst case. Guardrails flip the deal: start higher, and agree in advance to trim spending when your portfolio gets stressed. Runway's withdrawal guardrails toggle implements a version of this, and this post shows exactly what it buys and what it costs.
What are the Guyton-Klinger guardrails?
Jonathan Guyton and William Klinger introduced a set of decision rules in "Decision Rules and Maximum Initial Withdrawal Rates" (Journal of Financial Planning, March 2006). The two best known act as "financial guardrails" when market conditions cause the initial withdrawal rate to rise or fall significantly:
- Capital preservation rule: if the withdrawal rate has risen 20% above the starting rate, "the current year's withdrawal is reduced by 10 percent."
- Prosperity rule: if it has fallen 20% below the starting rate, spending is raised by 10%.
The authors report that initial withdrawal rates of 5.2–5.6% are sustainable at their 99% confidence standard for portfolios with at least 65% in stocks, and 4.6% with 50% in stocks. They are candid about the cost: the capital preservation rule's drawback is "chiefly the lost purchasing power each time the rule is triggered."
What does Runway implement?
Runway implements those two rules, with the published 20% bands and 10% adjustments, on top of its inflation-adjusted baseline, plus a limit of its own (next section). It does not model the paper's other rules. The first is a freeze on inflation increases after a down year; in a plan measured in today's dollars that would be a real cut about the size of inflation, so leaving it out means Runway's cuts come only from the capital preservation rule. The second is a sell-the-winners rule, which Runway replaces with its own account-ordering rules. The third is a late-life exception under which the capital preservation (cut) rule stops 15 years before the planning age; Runway keeps that rule on to the end of the plan. Turn it on with the guardrails toggle in Explore Your Plan or What-If Scenarios; when a rule fires, the year shows up in Today's Actions.
One detail to know: Runway's withdrawal rate is your planned yearly spending divided by the accounts that fund your plan, before any Social Security or pension income. A new income source therefore does not lower the rate, and the rules respond to your portfolio, not to the amount you net out of it each year.
Does Runway limit how far guardrails can cut or raise spending?
Yes: guardrails never take spending more than 20% below, or 10% above, your spending target, however many times a rule fires. Each trigger still moves spending 10%, but in the published rules every adjustment becomes the base for the next, so five cuts in a row would land 41% below where you started. Runway stops at the limit; if a rule would go past it, spending stays at the limit and that year's note says so. The limit is Runway's own design choice, not part of the Guyton-Klinger paper. It is what keeps the worst case a bounded cut instead of an open-ended one, and the odds table below shows what it costs.
How do advisors apply guardrails?
Mostly as the published rules, with judgment on top, and we found no standard cap on the total adjustment. What the sources say:
- The original paper. Guyton and Klinger tested the cut size and found that "withdrawal reductions above 10 percent (say, 15 percent and 20 percent) did not noticeably improve the probability of success"; the adjusted amount becomes the base for the next year, and the capital preservation rule expires 15 years before the end of the plan. They set no limit on the total.
- Practitioners. In Kitces's description of using the rules with clients, Guyton and Klinger expected 2 to 5 cuts in a 40-year retirement, the advisor reviews guardrails quarterly, and for market-driven breaches waits two calendar quarters before insisting on a change. That article sets no floor or cap on cumulative cuts.
- Critics. Kitces's analysis of the rules finds that cuts compound: in historical tests spending ended 28% below the original plan for a retiree just before the 2008 crisis, 36% below after the dot-com peak, 45% below in the Great Depression and 54% below in the 1970s. Its alternative, risk-based guardrails, sizes each adjustment by the plan's probability of success and cut only about 3% in the 2008 case.
- A different kind of limit. Vanguard's dynamic spending research limits each year's change instead of the total: in its example, spending can rise at most 5% or fall at most 2.5% from the prior year, as summarized by My Money Blog (we haven't checked the figures against Vanguard's paper). Narrow limits behave like a fixed-dollar rule and wide ones like a fixed percentage of the portfolio.
Runway's 20% and 10% are a middle path: far less than the 28% to 54% total cuts the critics found, while keeping most of the protection in the table below.
What do guardrails do to the odds?
Same household every row: the sample couple (Pat 65 and Alex 64, $1,460,000 saved), Social Security at 69 and 68, 2,000 simulated market paths, changing only the yearly spending and whether guardrails are on, with and without Runway's limit on how far they can move spending:
| Yearly spending | Chance money lasts, off | Chance money lasts, on (Runway's limits) | Chance money lasts, on (no limit) | Median ending balance, off | Median ending balance, on (Runway's limits) |
|---|---|---|---|---|---|
| $95,000 | 97.1% | 100.0% | 100.0% | $2,194,630 | $2,294,890 |
| $110,000 | 77.8% | 99.1% | 100.0% | $929,223 | $1,570,727 |
| $125,000 | 41.6% | 88.4% | 100.0% | $0 | $955,330 |
The pattern is clear, and easy to misread. A high success rate with guardrails doesn't mean spending is guaranteed; it means the money rarely ran out because spending was cut when needed. The limit is the visible cost: at $125,000 a year, 100.0% with no limit becomes 88.4%, because spending can't be cut as far. The table doesn't show how much was cut. That's the next section.
What do the cuts look like?
Take the $110,000 household and add a bad start: markets fall 20% a year for two years beginning at Pat's age 66. Without guardrails the plan keeps spending $110,000 a year and the money is gone before the end (ending balance $0). With them, with and without Runway's 20% limit on cuts:
| Pat's age | Yearly spending target, Runway's limits | Yearly spending target, no limit |
|---|---|---|
| 66 | $110,000 (crash begins) | $110,000 (crash begins) |
| 67 | $99,000 (cut 10%) | $99,000 (cut 10%) |
| 68 | $89,100 (cut 10%) | $89,100 (cut 10%) |
| 69 | $88,000 (stopped at the 20% limit) | $80,190 (cut 10%) |
| 70 | $88,000 | $72,171 (cut 10%) |
| 71 onward | $88,000 | $64,954 (cut 10%, then stable) |
| What happens to the money | Runs short at age 94 (ending balance $0) | Ending balance $645,793 |
With the limit, the rule fires three years in a row and spending settles at $88,000, 20% below where it started. With no limit it keeps firing five years in a row, each cut compounding on the last, and settles at $64,954, 41% below where it started, in exchange for a plan that finishes with $645,793. Runway's sample includes long-term-care costs in each person's last years, which is why the limited plan runs short at 94 even though it avoided the early crash. If a 20% cut is the most you would accept in a crisis, guardrails make a $110,000 start safer but not safe in every market; if you could accept deeper cuts, a higher limit would protect more. That is the question to ask before using them.
How do guardrails fit with the 4% rule and other approaches?
The 4% rule (see what the research really says) is the fixed-spending baseline; guardrails let you start above it by accepting variability. They pair naturally with a spending floor you've chosen in advance, and with the idea that early retirement years matter most. See why the first ten years matter. For where your own number lands, see how much you can safely spend.
Test guardrails on your own plan. Flip the toggle, see the spending path and the odds. Free to start, with the full What-If tools in Pro.
Try the plannerHow were these numbers computed?
The simulation rows are Runway's Monte Carlo (2,000 paths, fixed seed 2026) on the sample household with default market assumptions, each person planned through age 95, with the long-term-care cost left out, changing only spending and the guardrails flag. The "no limit" column runs the published rules with Runway's limit on total cuts and raises switched off; the other guardrail columns include it (cuts stop at 20% below the target, raises at 10% above). The spending-path table is a single deterministic projection with a −20%-a-year crash for two years beginning at Pat's age 66 and $110,000 of yearly spending, which does include Runway's long-term-care episodes (one for each person, $129,575 a year for three years in each person's last three years, so they affect the ending balance); the dollar amounts are the engine's alerts, in today's dollars. The Runway app now shows these same plans in future dollars (each year raised by the inflation assumption), so its figures run larger than the ones here. The 20% bands and 10% steps are the published figures; the 20% and 10% limits are Runway's own. All are fixed in the engine and not adjustable. The research summary is from the sources linked in it; the Vanguard figures are from a secondary summary.