Skip to content
Menu
Finabase Learn

Turn savings into retirement income Lesson 4 of 9

Retirement · Turning savings into income

Guardrails withdrawal strategy: Guyton-Klinger rules tested from 2007

The Guyton-Klinger decision rules as published in 2006, simpler percent-of-portfolio and floor-and-ceiling rules, and their spending paths from January 2007.

On this page
  1. Why vary spending at all
  2. The Guyton-Klinger rules as published
  3. Two simpler variable rules
  4. The test: January 2007 on SPY and AGG
  5. Spending volatility against depletion risk
  6. Pitfalls and limits
  7. The rules compared
  8. How Finabase does it
  9. Frequently asked questions
  10. Summary
  11. Check your understanding
  12. Exercise

A guardrails withdrawal strategy starts retirement spending at a fixed amount, raises it with inflation like the The 4% rule: what Bengen found, what it assumes and where it breaks, and then changes it in steps when the portfolio drifts too far from plan. If markets fall and this year's withdrawal becomes too large a share of what is left, spending is cut. If markets rise and it becomes too small a share, spending is raised. The guardrails are the two thresholds that trigger those changes.

The idea is a trade. A retiree who accepts that spending may be cut in bad years can start higher than one who insists on never cutting, because the cuts protect the portfolio when it is most vulnerable. The best-known version is a set of decision rules published by Jonathan Guyton and William Klinger in 2006. This lesson states those rules as the authors wrote them, sets two simpler variable rules beside them, and runs all of them, with fixed 4% and 5% withdrawals, on the same $1,000,000 portfolio of SPY and AGG from January 2007, a retirement that met the 2008 crash in its second year. The result is less flattering to guardrails than their reputation, and the reason is instructive.

Why vary spending at all

A fixed real withdrawal is simple to live with, because the income never changes, and dangerous for the same reason. When a portfolio falls early in retirement the withdrawal keeps coming out of a smaller base, which is how Sequence of returns risk: why the order of returns matters does its damage. The only way a fixed rule can protect against that is to start low enough to survive the worst sequence on record, which is why 4% is built for the worst case and leaves a surplus in most others.

Variable rules move the adjustment to the moment it is needed. They ask the retiree to take a pay cut after bad markets, sometimes a large one, in exchange for a higher starting income and a lower chance of running out. Whether that is a good trade depends on how much of the budget can actually be cut. Spending on travel and gifts can be trimmed; rent, insurance and food cannot.

The Guyton-Klinger rules as published

Guyton published a first set of rules in the Journal of Financial Planning in October 2004. In March 2006 he and Klinger tested a revised set with Monte Carlo simulations over 40-year retirements, using return data from 1928 to 2004 and from 1973 to 2004, and portfolios of 50%, 65% and 80% stocks. The paper keeps four rules.

  1. Portfolio management rule. After a year in which an asset class has risen above its target weight, the excess is sold and held in cash for future withdrawals. Withdrawals come first from those overweight holdings, then from cash, then from bonds, then from stocks.
  2. Withdrawal rule. Each year's withdrawal is the previous year's raised by inflation (CPI), except that there is no raise after a year in which the portfolio's total return was negative and the withdrawal rate would be above the initial rate. A missed raise is never made up. (The 2004 rules also capped any yearly raise at 6%; the 2006 paper drops that cap.)
  3. Capital preservation rule. If the current withdrawal rate, this year's withdrawal divided by the current balance, has risen more than 20% above the initial rate, this year's withdrawal is cut by 10%. With a 5% start the trigger is a current rate above 6%. The cut becomes the base for later years. The rule stops applying 15 years before the end of the planning horizon, for example after 85 for someone planning to 100.
  4. Prosperity rule. If the current withdrawal rate has fallen more than 20% below the initial rate, 4% for a 5% start, this year's withdrawal is raised by 10%, and the raise also becomes the base for later years.

The paper's conclusion was that, with these rules and at least 65% in stocks, starting withdrawal rates of 5.2% to 5.6% were sustainable at a 99% confidence standard over 40 years, and that with 50% in stocks the maximum fell to as low as 4.6%. The authors report that applying the rules consistently effectively eliminated the risk of running out of money, at the cost of the lost purchasing power each time spending was cut.

Two simpler variable rules

Percent of portfolio. Each January, spend a fixed share of the balance, here 5%. The rule can never run the portfolio to zero, because it always takes a fraction of what remains, but income moves one for one with the market: a 25% fall in the portfolio is a 25% pay cut the next year.

Floor and ceiling. Spend the same 5% of the balance, but never less than a floor or more than a ceiling, both set in real terms against the first year's amount. In the test below the floor is 90% and the ceiling 120% of the first year's $50,000, raised with inflation. The floor limits the pay cut; the ceiling stops spending racing ahead after a boom. These parameters are this lesson's illustration, not a published rule.

The test: January 2007 on SPY and AGG

The portfolio is $1,000,000 in 60% SPY and 40% AGG with dividends reinvested, rebalanced monthly, from January 2007 to September 2026. Spending is set each January and taken in twelve monthly instalments; the rest earns the month's return. Inflation raises use CPI-U from December to December. The Guyton-Klinger plan starts at 5% and applies the withdrawal, capital preservation and prosperity rules each January with a 40-year horizon, so the capital preservation rule is live for the whole test. Its portfolio management rule is replaced by monthly rebalancing, which keeps the comparison on spending rules alone. There are no fees or taxes. Because inflation raises here are yearly, the fixed 4% plan's numbers differ slightly from the monthly-indexed test in the The 4% rule: what Bengen found, what it assumes and where it breaks lesson.

Yearly spending under four rules, $1,000,000 in 60% SPY / 40% AGG from 2007 $30k $40k $50k $60k $70k 2007 2010 2013 2016 2019 2022 2025 2027 Fixed 4%, raised with CPI Fixed 5%, raised with CPI Guyton-Klinger, starting at 5% 5% of each January balance Spending set each January, in December 2006 dollars (CPI-U); 2026 runs to September Real yearly spending
Real spending per year under four rules, $1,000,000 in 60% SPY and 40% AGG from January 2007 to September 2026, in December 2006 dollars (CPI-U). Spending set each January; no fees or taxes. Computed by Finabase.
Rule (start January 2007) Real spending, lowest year Highest year Largest one-year real cut Total real spending to Sept 2026 Real balance, Sept 2026
Fixed 4%, raised with CPI $40,000 $40,000 none $790,000 $1,195,093
Fixed 5%, raised with CPI $50,000 $50,000 none $987,500 $817,488
Guyton-Klinger from 5% $41,470 $50,000 10.0% (2009) $882,120 $998,656
5% of each January balance $36,094 $55,938 25.4% (2009) $907,249 $1,012,569
5%, floor 90%, ceiling 120% $45,000 $53,206 15.4% (2023) $919,957 $954,546

All figures are in December 2006 dollars. Total spending counts 2026 as nine months.

What happened year by year

In January 2009 the portfolio stood at $749,936 after the 2008 crash, and the Guyton-Klinger plan's inflation-raised withdrawal would have been 6.94% of it. Two rules fired at once. Because 2008's return was negative and the rate was above 5%, there was no inflation raise; and because 6.94% was above the 6% guardrail, the withdrawal was cut 10%, from $52,054 to $46,849. In real terms spending fell from $50,000 to $45,010.

What followed is the point of the test. The portfolio recovered, but the cut was permanent: the rules never restore a cut, they only allow a separate 10% raise if the rate falls below 4%. It never did. In January 2022, after a strong decade, the current rate was 4.10%, just above the trigger, and the year that followed brought losses in both stocks and bonds. Two more skipped raises, in 2019 and 2023 after the losing years 2018 and 2022, left the plan spending $41,470 a year in real terms from 2023, barely above the fixed 4% plan's $40,000.

The fixed 5% plan, meanwhile, never cut anything and still had $817,488 in real terms in September 2026. In hindsight the 2008 crash was not deep or long enough to break a 5% fixed withdrawal from a 60/40 portfolio within 20 years, so the guardrails' protection was bought and not needed. The guardrails plan spent $105,380 less in real terms over the period and kept $181,168 more.

The percent-of-portfolio rule shows the opposite failure mode: it never threatens the capital but passes every market move straight to income, from a 25.4% real cut in 2009 to $55,938 in 2022 and back down 25% in 2023. The floor-and-ceiling version spent more than the guardrails plan in total and kept less, sat on its floor of $45,000 for most of the years to 2017, and still had to absorb a 15.4% cut in 2023 when it fell from $53,206 back to the floor.

Spending volatility against depletion risk

Every rule here sits on one line. At one end, fixed real spending gives a perfectly smooth income and the full risk of running out. At the other, a fixed percentage of the balance cannot run out but makes income as volatile as the portfolio. Guardrails sit near the smooth end with a safety valve; floor-and-ceiling rules sit nearer the variable end with limits on how far income can move.

No point on the line is right in general. One historical path, especially a favourable one like 2007 to 2026, cannot rank them: it shows how each rule behaves, not which one works. Monte Carlo studies such as Guyton and Klinger's, and the Monte Carlo simulation: testing a plan against thousands of futures lesson's approach, judge rules across thousands of paths, including ones far worse than 2008, and that is where guardrails earn their higher starting rate.

Pitfalls and limits

  • Cuts compound. A 10% cut that is never restored lowers every later year. Three skipped raises and one cut left the 2007 plan 17% below its starting real income.
  • Rules interact. In January 2009 the withdrawal rule and the capital preservation rule fired together. The order in which rules are applied changes the result slightly; the paper applies the other rules to the adjusted amount.
  • The trigger is a cliff. A current rate of 4.10% gets nothing; 3.99% gets a 10% raise. Small differences in timing or fees can flip a decision.
  • Missing the portfolio rule. Guyton and Klinger's results include their portfolio management rule. A test without it, like this one, measures the spending rules alone.
  • Real budgets have floors. A 10% cut in total withdrawals can be a much larger cut in discretionary spending once fixed costs are paid.
  • Taxes and RMDs. Withdrawals from tax-deferred accounts are taxable, and required minimum distributions may force withdrawals larger than a rule calls for; the money can be reinvested in a taxable account but not left in the IRA.

The rules compared

Rule Starting rate in this test Income changes when Can run out? Main cost
Fixed real (4% rule) 4% Never, apart from inflation Yes, in bad sequences Low starting income, large surplus in good periods
Fixed real at 5% 5% Never More likely than at 4% Higher depletion risk
Guyton-Klinger guardrails 5% Current rate leaves the ±20% band; inflation raise skipped after losing years Rarely, in their simulations Permanent cuts; raises need a large rally
Percent of portfolio 5% Every year, with the balance No Income as volatile as the portfolio
Floor and ceiling 5% Every year, within limits Yes, if the floor is spent long enough Some volatility and some depletion risk

How Finabase does it

Finabase is a research app: it does not run a withdrawal plan, manage an account or know a reader's taxes. The guardrail test needs two numbers that Finabase does show. The Portfolio page gives a tracked portfolio's total market value, and dividing the year's planned withdrawal by that value is the current withdrawal rate that every rule above watches. The Economy page shows the consumer price index used for the inflation raise. The test on this page was run on the same monthly price and dividend history of SPY and AGG that their symbol pages show.

Frequently asked questions

What are the Guyton-Klinger guardrails?

They are two rules from Guyton and Klinger's 2006 paper: cut this year's withdrawal by 10% if the current withdrawal rate is more than 20% above the starting rate, and raise it by 10% if the rate is more than 20% below. With a 5% start the band runs from 4% to 6%.

What starting withdrawal rate do guardrails allow?

In the authors' simulations, 5.2% to 5.6% for portfolios with at least 65% in stocks over 40 years at a 99% confidence standard, and as low as 4.6% with 50% in stocks. These are model results, not guarantees, and depend on following the rules every year, including the cuts.

Is the capital preservation rule used for the whole retirement?

No. It stops 15 years before the end of the planning horizon, for example after 85 for someone planning to 100. The authors found that keeping it in the final years cost purchasing power without improving the chance of success.

Is a guardrails strategy better than the 4% rule?

It depends on what the retiree values. Guardrails allow a higher starting income at the cost of possible cuts; the 4% rule gives a stable income at the cost of a lower start. In the 2007 test, guardrails spent about 12% more in total than fixed 4% and ended with about 16% less.

Summary

Guardrails let a retiree start spending above the 4% rule by agreeing in advance to cut when the portfolio struggles. Guyton and Klinger's 2006 rules skip the inflation raise after losing years, cut spending 10% when the current withdrawal rate rises more than 20% above its start, and raise it 10% when the rate falls more than 20% below; in their simulations that supported 5.2% to 5.6% starts over 40 years with at least 65% in stocks. On SPY and AGG from January 2007 the rules did their job in 2009, cutting real spending from $50,000 to $45,010, but the cut was never restored, and by 2023 the plan was spending $41,470 a year while a fixed 5% plan that never cut was still solvent. Percent-of-portfolio rules swing the other way, with a 25.4% real pay cut in 2009. Every variable rule exchanges some steadiness of income for some safety of capital.

Check your understanding

4 questions on this lesson. Each answer comes with the reason, and nothing is scored.

  1. A retiree starts at a 5% withdrawal rate under the Guyton-Klinger rules. In January the planned withdrawal is 6.5% of the balance. What does the capital preservation rule do?

    1. Nothing, because the rate is still below 7%
    2. Cuts this year’s withdrawal by 10%, because the rate is more than 20% above 5%
    3. Stops withdrawals for a year
    Show the answer

    B. Cuts this year’s withdrawal by 10%, because the rate is more than 20% above 5% The trigger is a current rate more than 20% above the initial rate, 6% for a 5% start. The cut is 10% and becomes the base for later years.

  2. Under the 2006 withdrawal rule, when is the yearly inflation raise skipped?

    1. Whenever inflation is above 6%
    2. After a year with a negative portfolio return, when the withdrawal rate would be above the initial rate
    3. Every other year
    Show the answer

    B. After a year with a negative portfolio return, when the withdrawal rate would be above the initial rate The 2006 paper dropped the 2004 cap on raises and kept a freeze after losing years, but only when the withdrawal rate is above its starting level. A skipped raise is never made up.

  3. In the 2007 test, why did the guardrails plan end up spending less than a fixed 5% plan that never ran out?

    1. Its 2009 cut and three skipped raises were permanent, and the current rate never fell far enough to trigger the prosperity rule
    2. It started at 4%
    3. It paid higher fees
    Show the answer

    A. Its 2009 cut and three skipped raises were permanent, and the current rate never fell far enough to trigger the prosperity rule The rules restore income only through a separate 10% raise when the current rate falls more than 20% below the start. In January 2022 the rate was 4.10%, just above the 4% trigger.

  4. What is the main drawback of spending a fixed 5% of each January balance?

    1. It can run the portfolio to zero
    2. Income moves with the market: in the 2007 test real spending fell 25.4% in 2009
    3. It ignores inflation in good years
    Show the answer

    B. Income moves with the market: in the 2007 test real spending fell 25.4% in 2009 Taking a share of what remains can never exhaust the portfolio, but every market fall becomes a pay cut of the same size.

Exercise

One task to do in Finabase. Nothing is graded or saved.

Open Portfolios. Signed out, the page shows an example portfolio with a Market value. Set a first-year withdrawal of 5% of it. Under Guyton and Klinger’s rule, spending is cut 10% when the current rate rises more than 20% above the start, above 6%. Suppose the portfolio falls 10%, then 20%: work out the withdrawal rate each time (same dollars ÷ new value). Which fall trips the guardrail?

Open Portfolios
What you should find

A 10% fall lifts 5% to about 5.6%, inside the guardrail; a 20% fall lifts it to 6.25% and triggers the 10% cut. Ignoring the withdrawal itself, any fall of more than about 16.7% does it. In the lesson the cut came in 2009, taking real spending from $50,000 to $45,010, and it was never restored.

Sources and further reading

  1. Jonathan T. Guyton and William J. Klinger, “Decision Rules and Maximum Initial Withdrawal Rates,” Journal of Financial Planning (March 2006)
  2. Jonathan T. Guyton, “Decision Rules and Portfolio Management for Retirees: Is the ‘Safe’ Initial Withdrawal Rate Too Safe?”, Journal of Financial Planning (October 2004)
  3. William P. Bengen, “Determining Withdrawal Rates Using Historical Data,” Journal of Financial Planning, vol. 7, no. 4 (October 1994)
  4. U.S. Bureau of Labor Statistics, Consumer Price Index for All Urban Consumers (CPIAUCSL), via FRED; no index was published for October 2025
  5. Finabase chart data — monthly prices and dividends of SPY and AGG, January 2007 to September 2026; simulations computed by Finabase