What 1966 Retirees Can Teach Anyone Retiring in 2026

If you wanted to design the worst possible decade to start a retirement, you would have a hard time topping the seventeen years from January 1966 through August 1982. The Dow Jones started 1966 at around 985. By August 1982, after sixteen and a half years of stagflation, oil shocks, two recessions, and the worst inflation in modern American history, the Dow was sitting at — get ready — about 777. Lower than where it started. For sixteen and a half years. Not flat. Lower. Adjusted for the staggering inflation of those years, the real value of a dollar invested in stocks in 1966 had been cut by something like two-thirds by 1982. Brutal, right?

That seventeen-year window is the canonical worst case in retirement income research. The 1966 retiree is the cohort whose story economists, planners, and authors keep coming back to, because it teaches a lesson that shows up nowhere else in the historical data with quite the same force. The lesson is simple. The lesson is uncomfortable. And the lesson is not theoretical — it has direct, immediate, practical implications for anyone retiring in 2026. Let me walk you through it.

The 1966 retiree's situation, on paper

Imagine a man who retired on January 1, 1966. Call him Walter. Walter had saved diligently for forty years. He had a balanced 60/40 portfolio — sixty percent in U.S. stocks, forty percent in intermediate-term Treasuries — which was the conventional advice of the day. He decided to follow the planning guideline of his era: withdraw four-and-a-half percent of his starting portfolio in year one, then increase that dollar amount by inflation each year. If his portfolio was $200,000 (a substantial nest egg in 1966 dollars, equivalent to maybe $1.8 million today), his year-one withdrawal was $9,000.

For decades of historical data before 1966 — testing this withdrawal pattern against every rolling thirty-year window from 1926 to the early 1960s — that 4.5% rate had survived comfortably. Most cohorts ended thirty years with money to spare. So Walter had reason to feel safe.

Now here's the thing about the average return Walter actually got. Run the numbers across all thirty years from 1966 to 1995. The annualized return on a 60/40 portfolio over that whole window was roughly seven percent. Seven percent. Not a disaster. Not a bull market, but not a disaster. Average. Within the range of historical norms.

And yet — Walter's portfolio ran out before year thirty. Bengen's research, which we'll come back to, identified the 1968 retiree as the worst case in the data; the 1966 retiree was nearly as bad. The math could not catch up. Same average return that worked for cohorts before and after. Different sequence. Different outcome.

What killed Walter's plan: the 1973–1974 bear market

The single biggest blow came in 1973–1974. The S&P 500 lost roughly 48% in real terms over those two years — one of the worst back-to-back declines in U.S. market history, made worse by the fact that inflation was running at 8% to 12% during the same period. So Walter's stock portfolio, in real purchasing power, dropped by nearly half. His bond portfolio also lost real value because inflation was eroding the principal faster than the bonds were paying.

By 1975, only nine years into retirement, Walter's portfolio in real terms was a fraction of what he started with. And here is the lethal part — he had been withdrawing the inflation-adjusted equivalent of his year-one $9,000 every single year. Those withdrawals didn't pause for the bear market. They didn't shrink. They scaled up with the brutal inflation of the era. By 1975, his nominal annual withdrawal was probably around $14,000 a year, against a portfolio that had been gutted.

So Walter was selling shares to pay grocery bills, in a market that had cratered, in a currency that was losing purchasing power monthly. The forced selling locked in losses. Each share sold at the bottom was a share that didn't get to ride back up when the market eventually recovered. By the time the great bull run of 1982-2000 finally arrived, Walter's portfolio was too small to participate meaningfully.

The lesson: average return is the wrong number to plan with

This is the foundational insight of retirement income research. Two retirees can have the exact same thirty-year average annual return, withdraw the same dollar amounts in the same pattern, and end up with completely different outcomes. The variable that makes the difference isn't the average. It's the order.

Bill Bengen, a financial planner who later became a household name in retirement research, formalized this finding in his 1994 paper "Determining Withdrawal Rates Using Historical Data." He back-tested every rolling thirty-year window from 1926 to 1992 to find the worst-case starting withdrawal rate that survived all thirty years. He called it SAFEMAX. The number he landed on was 4.15% — set by the 1968 retiree, with the 1966 cohort close behind. That 4.15% rounded down to "the 4% rule," which has been the conventional planning starting point ever since.

It is worth understanding that the 4% rule isn't a safety guarantee. It is a backtest of one country's market history, anchored on the worst case in that data. Subsequent research from Wade Pfau, Morningstar, and others has run the math against current valuations and bond yields and produced safe-withdrawal estimates ranging from 3.0% to 4.0% depending on the year and assumptions. The safe rate is not a constant. It moves.

What 1966 means for someone retiring in 2026

The fair question: is 2026 a "1966-style" environment? The honest answer is — we don't know yet, and that's the point. Nobody knew in 1966 either. The 1966 retiree had no special warning that the next sixteen years would be the worst stretch for a balanced portfolio in the modern era. They retired into an economy that looked normal and ran straight into a buzzsaw.

What we can say about 2026: equity valuations are elevated by historical standards. Bond yields are higher than they were in 2021 but still modest by long-run norms. Inflation has been declining from the 2022-2023 peak but isn't yet back to the comfortable 2% the Fed targets. The risk-of-bad-sequence environment isn't identical to 1966, but it is also not obviously safer. Anyone retiring into 2026 should plan as if a 1966-like decade is possible, because the cost of being wrong on the optimistic side is very high and the cost of being wrong on the cautious side is just a smaller portfolio at age 95.

The four lessons Walter would offer

If Walter could come back and talk to a 2026 retiree, here is what I think he'd say. He's not real, but the lessons from the historical data he stood in for absolutely are:

• Don't sell shares to pay bills during a down market if you can possibly avoid it. This is the most important lesson. The 1973-1974 forced selling is what made his sequence shock irreversible.
• Build an income floor before you need it. Cash, bonds, MYGAs, sometimes a SPIA — enough to cover essential expenses for at least five years without touching the equity portfolio. The bucket strategy. Walter didn't have one because the concept didn't exist in widely-published form yet. You do.
• Pay attention to the sequence as it unfolds, especially in the first ten years. A retiree who sees a bad sequence developing in years one through five has options — reduce withdrawals temporarily, delay non-essential spending, find part-time work, defer Social Security claiming if they haven't yet. The 1966 retiree didn't have a framework for these adjustments. You do.
• Don't trust "average returns will save me." They won't. Average returns are a math abstraction. You and I live through specific years, in specific orders. The order is what determines whether the math saves you.

What this looks like in practice

The ironic, useful fact about the 1966 cohort is that their pain became the next generation's planning curriculum. The 4% rule, sequence-of-returns risk, the bucket strategy, the income floor framework, the Retirement Red Zone concept — all of it is descended from researchers staring at the 1966-1982 data and asking, "How do we keep the next Walter from running out?"

The answer, it turns out, is structural. You can't predict whether 2026 is the next 1966. You can build a plan that holds up if it is. A plan with an income floor that covers essentials. A withdrawal strategy that doesn't force selling at the bottom. A reasonable starting withdrawal rate calibrated to current conditions. And a willingness to adjust if the first five years of retirement turn out to be uglier than expected. That's not a guarantee. It's a structure that converts a worst-case decade from a fatal blow into a survivable shock. That's the difference between Walter's outcome and what we know how to build now.

Get that structure right and you can typically sleep at night, even in years that look like 1966. That's the whole point of all of this.

The bottom line

The 1966 retiree had a normal thirty-year average return and still ran out of money before year thirty. The reason wasn't bad luck. The reason was the order in which the returns showed up — bad early, brutal mid-decade, recovery too late. Forced selling during the 1973-74 bear market locked in losses that the eventual bull run couldn't undo. The lesson translates exactly to 2026: build an income floor, don't be forced to sell at the bottom, calibrate your withdrawal rate to current conditions, and stay flexible in the first ten years. Walter would have wanted those tools. We have them. Don't waste the lesson he paid for.