The man on the dock is the whole video. Weekday morning, thermos, a rod he is not really watching, sixty-something years old and finished with work for good. Everybody my age can picture him, because we all knew one.
Here is the figure that did not make the script. In 1976, when the Social Security Administration went and counted what older Americans were actually living on, 26% of units aged 65 and over had any private pension income at all. 94% had a Social Security cheque. An aged unit is a married couple or a single person, and the arithmetic does not soften: three in four of the people already retired in the year I was describing were living without the pension the video treats as the era’s standard equipment. (Income of the Population Aged 55 and Older, 1976.)
The man on the dock existed. He was a minority, and I told you he was the norm.
This page is everything I found while building that video, including the places where what I found argued with me. The sources are grouped under the claim each one belongs to, and the section at the bottom is the one to read if you read only one.
The video: Retirement at 62 Didn’t Die. It Was Killed. · the channel
The pension most people did not have
Coverage and receipt are two different things, and the gap between them is where the nostalgia lives. At the high-water mark, somewhere between 1975 and 1980, roughly four in ten private-sector workers were participating in a defined benefit plan. Participating is not the same as qualifying. In 1972, only 32% of private plan participants had a vested benefit, meaning a legal right to anything at all if they left. ERISA moved that hard: by 1979 it was 48%. (Vesting of Private Pension Benefits in 1979, Social Security Bulletin.)
Sit with the 1972 number for a second. Two-thirds of the people inside the golden-age pension had not yet earned a penny of it, and the ten-year cliff the video mentions is exactly why. The law that fixed this arrived in 1974, seven years before the 401(k) went live, and it fixed it for the generation that came after the man on the dock.
The plan count is the wrong number
I used the plan count in the video because it reads like a crime scene, and it does: 103,346 plans in 1975, a peak of 175,143 in 1983, down to 56,405 by 1998, and 46,233 in 2023. Run the participant numbers beside it and the shape changes. Active participants in those plans went 27.2 million in 1975, 29.9 million in 1983, 22.9 million in 1998, 11.1 million in 2023. (DOL Private Pension Plan Bulletin, historical tables.)
Between 1983 and 1998 the number of plans fell by 68% and the number of people in them fell by 24%. Most of the plans that vanished in that stretch were small ones, a doctor’s practice or a family firm with a dozen employees. The PBGC insured more than 90,000 small plans in 1985 and fewer than 19,000 by 2002, and small plans have disappeared faster than any other group every year since. (PBGC, Trends in Defined Benefit Pension Plans.) The usual explanation is the cost of complying with ERISA, which falls hardest on an employer with nobody to administer it. The industrial pensions, the ones at the plant and the phone company, were still standing.
Which means the collapse I was describing is not a thing that happened to your father. It is happening now, to you. Half the fall in active participants comes after 1998. The first cohorts to live the whole story are retiring this decade, and defined contribution plans went from 11.2 million active participants in 1975 to 96.4 million in 2023, so nobody has to guess what replaced it.
The paragraph, and how much it can carry
The history is solid. Section 401(k) arrives in the Revenue Act of 1978, written for executive deferrals, debated by nobody. Ted Benna reads it in 1980, sees that it can be turned on ordinary payroll, and launches the first one on 1 January 1981 at his own firm. In 2011 he told CBS News that his creation had become a monster. (CBS News, 29 November 2011.) He has repeated it for fifteen years and he means it.
The causation is where I would push back on my own script. Economists who have tried to measure why firms abandoned defined benefit plans do not find one lever. Aaronson and Coronado, writing for the Federal Reserve in 2005, attribute a large share of the shift to the composition of the workforce and to technical change, with the pattern varying by industry in a way a single tax paragraph cannot explain. (FEDS 2005-17.) Manufacturing shrank. Union density fell. Workers changed employers more often and a portable account suited them better. All of that was happening with or without Benna.
The paragraph is a cause. I called it the cause, and it is not.
What 1983 actually did to the number 62
This is the correction I care most about, because it is in the title. The 1983 amendments did not abolish retirement at 62. The earliest eligibility age is still 62 today, unchanged since 1961, and anybody can walk in and claim tomorrow.
What changed is the price. When the full retirement age was 65, claiming at 62 paid 80% of your primary insurance amount. With the full retirement age at 67, which applies to everyone born in 1960 or later, claiming at 62 pays 70%. (SSA, benefit reduction for early retirement.) Same door, same age, 10 percentage points less on the other side of it, for life.
That is a sharper sentence than the one I wrote, and it is the version I would use again. Retirement at 62 was not killed. It was repriced, quietly, by a Congress that knew a benefit cut is easier to pass when it arrives as an actuarial table.
The retirement age itself
The Center for Retirement Research tracks the age at which more than half of a birth cohort has left the labour force. For men it was 66 in 1962, 64 in 1975, and it bottomed at 63 in 1983 before climbing back to 65 in 2024. For women it went from 53 in 1962 to 63 in 2024. (Average Retirement Age for Men and Women, 1962-2024.)
So the male retirement age fell by about three years across the pension era and has since given back two. The cliff I implied is a slope, and for women it runs the other way entirely, which is a different video and probably a better one.
The savings figure I used
The 185,000 dollars is real and it is narrower than it sounds. It is a conditional median, which means it describes households aged 55 to 64 that have a retirement account, and says nothing about the ones that do not. The Federal Reserve found 54.3% of all American families held a retirement account of any kind in 2022. (SCF 2022 bulletin.) The Center for Retirement Research puts it plainly for older households: only about half of them have a 401(k) at all, and among those that do, the combined 401(k) and IRA median was 204,000 dollars in 2022. (CRR, 401(k)/IRA holdings in 2022.)
So when I said half of households have less than 185,000 dollars, that was wrong in the direction that flatters my argument. Half of the households that have an account have less. Count everybody and the true midpoint is far below it. I should have said so.
The Social Security figure I used
I said 73% of seniors lean on Social Security for most of their income, and that number comes from household surveys. Bee and Mitchell at the Census Bureau matched the survey against tax records and found the survey captures 48% of retirement income for people 65 and over. 46% of the time, somebody drawing retirement income reported none. Median household income for that age group came out 30% higher on the administrative data, 44,400 dollars against 33,800. The share of beneficiaries said to depend on Social Security for 90% or more of their income fell from 36% to 18%. (Do Older Americans Have More Income Than We Think?, Census working paper 2017-39.)
Retirees are poorer than they should be. They are not as poor as the survey says, and the gap is widening, because the money that surveys miss is exactly the money the 401(k) era produces: irregular withdrawals from an account rather than a cheque that arrives on the same day every month.
The five moves, checked
The earnings record on ssa.gov is free and benefits are computed from your highest 35 years, so a missing year is a permanent reduction and correcting it costs nothing.
The claiming figure holds. At a full retirement age of 67, age 62 pays 70% and age 70 pays 124%, and 124 divided by 70 is 1.77, hence 77% bigger. It is guaranteed and inflation-adjusted, and it is still wrong for plenty of people, which is why the video says decide it on purpose rather than telling you to wait.
The unclaimed employer match is the weakest of the five. The number everybody quotes traces to a Financial Engines study from 2015, which is old enough that I would not put a figure on screen, so the script says billions and leaves it there.
The government lookup at lostandfound.dol.gov is real and it is not finished. It was built from a voluntary request to plan administrators plus IRS Form 8955-SSA data, and that form is not updated when somebody closes an account, so it produces false positives as well as gaps. (NAPA, December 2024.) Search it, and if nothing comes up, that is not proof you have nothing.
The fee number comes from the Department of Labor’s own booklet: 0.5% against 1.5% over a 35-year career, and the ending balance differs by about 28%.
What did not survive
The share of retirees collecting a private pension has not fallen. It was 26% of aged units in 1976 and 28% in 2014. (SSA Income of the Aged Chartbook, 2014.) This is the most awkward fact on the page and I have thought about it for a week. The likeliest reading is that pension receipt lags pension coverage by about thirty years, so the flat line describes people who earned their benefits before the freeze, and the fall is in front of us rather than behind. That is still not what the video implies, which is that the loss has already landed.
Old age got dramatically less poor, not more. The poverty rate for Americans 65 and over was 28.5% in 1966 and 9.2% in 2017, a fall of nearly 70%. The supplemental measure, which counts medical costs, puts it at 14.1%, which is the number to argue with. (CRS, Poverty Among the Population Aged 65 and Older.) Whatever was taken away, this was not taken away, and it happened because the 1972 amendments tied Social Security to inflation, rather than because of anything a company did.
Retirement risk is not a straight line up. The National Retirement Risk Index, which estimates the share of working-age households that will not maintain their standard of living, read 31% in 1983, rose to 53% in 2010, and came back down to 39% in 2022, the lowest since 2004. (CRR.) The long trend is bad. The recent trend is not, and anyone drawing a single descending arrow from 1975 to today is drawing from memory.
The 2 trillion dollars in lost accounts is a model, not a count. It comes from Capitalize, a company whose business is moving old 401(k) accounts, and it is built from Form 5500 filings and estimated abandonment rates. (Capitalize, The True Cost of Forgotten 401(k)s.) The order of magnitude is probably right. The precision is a press release.
And the 1970s pension broke its promises too. Bethlehem Steel and United Airlines handed theirs to the federal insurer, which pays a capped benefit, and thousands of people found out in their sixties what capped meant. A 401(k) has never done that to anyone, because there is nothing in it to default on. That is the one argument for the new deal I cannot get around, and I left it in the video for that reason.
The video this comes from: Retirement at 62 Didn’t Die. It Was Killed.