Which withdrawal strategy held up in the worst markets?

The short answer

We ran four withdrawal strategies through every 30-year U.S. retirement since 1928, with the same savings, the same starting spending and the same history. Starting at 4.5% of $1 million ($45,000 a year, raised with inflation):

• Fixed spending lasted in 58 of 69 start years, and in 0 of the 10 hardest.
• Guardrails lasted in 68 of 69 start years, and in 9 of the 10 hardest.
• Two buckets lasted in 62 of 69 start years, and in 3 of the 10 hardest.
• Guardrails + cash bucket lasted in 69 of 69 start years, and in 10 of the 10 hardest.

The plans that adjusted spending lasted far more often in the worst markets, and paid for it with leaner years. Cash buckets on their own did less for whether the money lasted, but they largely stopped forced selling after a fall. No strategy won on every measure.

The four strategies, with the same rules for each

Fixed spending

The first year’s withdrawal, raised with inflation every year and never otherwise changed: the 4% rule approach. 60% stocks, 40% 10-year Treasuries.

Guardrails

The same 60/40 mix. If the withdrawal rate rises 20% above where it started, spending is cut 10%; if it falls 20% below, spending rises 10%. Spending never goes below 80% of the starting amount.

Two buckets

Four years of withdrawals held in Treasury bills and 10-year Treasuries, the rest invested 80% in stocks. Spending comes from the cash bucket, which is refilled only when investments are near their previous high. Spending never changes.

Guardrails + cash bucket

Guardrails decide how much to spend; the bucket decides where it comes from. In a down year with too little cash, spending dips one guardrail step instead of selling.

These are the same rules the methodology describes for every Century Tested calculator.

How often each strategy lasted

$1 million, 30 years, U.S. data 1928–2025, rebalanced yearly, no taxes or fees. The 10 hardest start years are those with the lowest 30-year safe withdrawal rate at 60% stocks: 1937, 1962, 1964, 1965, 1966, 1967, 1968, 1969, 1972, 1973. See the worst years to retire.

The 10 hardest start years, one by one

Starting at 4.5% ($45,000 a year), this is what each strategy had left after 30 years in today’s dollars, or the year it ran out. Hardest first.

1966 could sustain only 3.73% a year with fixed spending, so every plan starting at 4.5% was under strain. Replay these years with your own numbers in the 1966 and 1973 walkthroughs.

What the flexibility cost

Fixed plans never cut, until some of them ran out. Guardrail plans cut to as little as $36,000 a year, but raises after good markets meant the typical guardrail retirement spent more in total than fixed spending. In the hardest start years the cuts came first, so the flexible plans spent less in total there, and kept going.

Selling after a fall: what buckets changed

Selling investments right after a down year locks in losses. Over 30 years, the median retirement did that this many times at 4.5%:

• Fixed spending: 8 years
• Guardrails: 8 years
• Two buckets: 2 years
• Guardrails + cash bucket: 2 years

Buckets paid down-year spending from cash instead, which is their main job. See how bucket size changes this in the bucket calculator.

Ending balances, and a caveat

Median savings left after 30 years, in today’s dollars, at 4.5%: fixed spending $1,182,393; guardrails $1,084,311; two buckets $1,891,386; guardrails + cash bucket $1,621,761.

Bucket plans invest 80% in stocks outside the cash bucket, about 66% of savings at the start at 4.5%, against 60% for the other two. Part of their larger ending balances comes from that higher stock share, not only from the bucket.

What this test can’t tell you

• It uses U.S. history only, and there are fewer than four independent 30-year periods since 1928. It’s a stress test, not a forecast.
• Taxes, fees, Social Security and other income are left out so the strategies compare on equal terms. Your own plan includes them.
• Whether leaner years are acceptable depends on how much of your spending is flexible. That’s a personal choice the data can’t make for you.

Compare strategies on your own numbers: fixed vs dynamic withdrawals, guardrails vs buckets and the guardrails calculator.

Download the data

Every start year from 1928 to 1996, each strategy, at 4%, 4.5% and 5%: whether it lasted, years lasted, ending and lowest balance, lowest yearly spending, total spending, cuts and years selling after a drop.

Download as CSV