The Truth About Social Security and Dead PeopleThe Truth About Social Security and Dead People
There’s a reason millions of people are not listed as deceased in the main database when they should be, but it's not the fraud that Elon Musk and Trump say it is.
February 24, 2025
(Bloomberg Opinion) -- For more than a decade, Social Security’s inspector general has been on a quest to get dead people off the benefit rolls. The effort started before former corporate lawyer Gail Ennis took the job in 2019, appointed by President Donald Trump, but picked up speed afterward. This work has consisted mostly of audits comparing Social Security records with death records kept by the states, other federal agencies and even different Social Security databases.
In a related but somewhat different audit that has been in the news lately, Ennis in 2023 also looked into the 18.9 million entries in the Social Security Administration’s “Numident” master file of all assigned Social Security numbers that had (1) a birth year of 1920 or earlier, and (2) no record of death. As of 2020, only 86,000 Americans were that old, leaving about 18.8 million dead people not listed as dead by Social Security. As you may have heard, these undead hordes were discovered recently by Elon Musk’s Department of Government Efficiency, with Musk implying and President Trump claiming outright that tens of millions of dead people are collecting Social Security benefits.
This is false. My Social Security number is in the Numident file, and if you have one, yours is, too. That doesn’t mean we’re receiving benefits. Ennis — who was, remember, a Trump appointee — found that a perfectly reasonable-sounding 44,000 of those with birth dates in 1920 or earlier were actually getting payments from Social Security. In 2015, her predecessor, Patrick P. O’Carroll Jr. (a George W. Bush appointee), had looked into those with Numident birth dates before June 16, 1901, making them older than any known living person at the time, and found 6.5 million with no death recorded but only 266 receiving benefits, all but 13 of whom turned out to have incorrect birth dates on file.
This is a modal window.
The media could not be loaded, either because the server or network failed or because the format is not supported.
Almost all the others had died and had their benefits cut off many decades ago, before Social Security developed its current electronic death-reporting system. Both Ennis and O’Carroll urged the Social Security Administration to do something about these millions of inactive but officially alive centenarian account holders, but agency officials objected that updating all the records would cost millions of dollars and result in little or no reduction in benefit payouts.
The death-record comparisons conducted by the inspector general’s office, on the other hand, have led the SSA to update records and cut off benefits for thousands of (dead) people. They offer a fascinating window into how big the problem of dead Social Security recipients actually is, and how much work it takes to fix it. At least 38 of these audits have been conducted since 2014, with the results usually going something like the matching exercise with 16.1 million California death records from 1905 to 2017 that was completed in 2021. It found:
245 dead people whose children or spouses or others with access to their bank accounts were still receiving their benefits, which the Social Security Administration canceled.
52 beneficiaries “who appear to be deceased” that the SSA was still looking into.
438,460 people who were listed as dead by California and not Social Security but weren’t receiving Social Security benefits.
89 people who were listed as dead in California records but were still alive, although 12 of them died while the SSA was checking into their cases.
The inspector general estimated that $21.3 million had been paid out in error to the 245 dead people, and an additional $8.3 million to the 52 possibly dead ones. This particular audit report didn’t go into any detail about 89 people incorrectly listed as dead, but others offered fun commentary such as:
“An OI review determined the beneficiary was not deceased based on interviews with family members and the beneficiary.”
“SSA determined it had incorrectly issued the same SSN to two people.”
“SSA stated that one beneficiary is likely alive and the victim of an identity thief.”
The totals I added up from all the matching exercises were:
44,152 dead or possibly dead people who were receiving benefits or had received them after death.
$875 million in improper or possibly improper Social Security payments.
834,612 dead or possibly dead people who weren’t listed as such in the Numident file, but also weren’t receiving benefits.
726 people who were listed somewhere as dead but were in fact alive.
This does not represent an exhaustive accounting, of course. There are still 23 states that haven’t had their deaths data compared with Social Security’s, there are surely undiscovered data-entry errors and a few deaths that aren’t recorded anywhere, plus people continue to die. But the totals above do give a sense of the scale of the problem. In a country where every year more than 3 million people die and Social Security pays out more than $1.3 trillion in benefits, it doesn’t seem very big.
It’s also mostly in the past tense: More than half the dead people and almost a quarter of the estimated overpayments were from a single audit of recipients whose benefits had already been suspended because the SSA suspected they had died but hadn’t marked them as dead in the Numident file or tried to collect overpayments. On the audits that identified continuing improper payments, the SSA acted quickly to stop them, saving an estimated $150 million in the 12 months after the audits.
There are, of course, other sources of potential Social Security fraud. The most persistent appear to involve the Social Security Disability Insurance and Supplemental Security Income programs, where verifying that recipients meet eligibility requirements is a more expensive and time-consuming process than checking whether people are dead, with the financial or political costs or both sometimes outweighing the benefits. An inspector general’s report from 2015, for example, found that from 2008 through 2013, the SSA spent $323 million collecting $109.4 million in low-dollar overpayments (defined as overpayments smaller than the average cost of recovering an overpayment), most of them to Supplemental Security Income recipients. And Inspector General Ennis quit last year amid serious political blowback from an anti-fraud program run by her office that piled large fines on unsuspecting recipients of benefit overpayments. There’s surely more that can be done to combat fraud at the agency. There just don’t seem to be a lot of easy pickings.
But what about all those undead centenarians — of whom there are now 20.8 million, according to numbers released by Musk — in Social Security’s Numident file? Why are they still not marked as dead after multiple attempts by inspectors general to make that happen?
The reasons given by SSA officials have been that adding death information to all the records would cost too much ($5.5 million to $9.7 million, it estimated in 2015), require regulatory changes to carry out, result in the possible release of personal information about still-living people to the agency’s publicly available Death Master File and “be of little benefit to the agency” because few if any payments are going to dead people born before 1920.
In their reports, Ennis and O’Carroll both objected that the incorrect records did impose costs on government agencies and financial institutions that rely on Social Security’s deaths data, as these inactive accounts are easy targets for identity theft. Between 2006 and 2011, 66,920 of the Social Security numbers registered to people born in 1901 or earlier had wages, tips, and self-employment income associated with them — meaning that people born a lot more recently than that, and probably lacking in authorization to work in the US, had used them to get jobs. Between 2016 and 2020, 139,211 of the numbers registered to people born in 1920 or earlier did.
These 206,131 workers reported $11.6 billion in income, which hints at another possible reason the Social Security Administration hasn’t been super-aggressive about clearing dead people and other dodgy entries out of the Numident file — the up to $1.4 billion in payroll taxes on that $11.6 billion in earnings. These people won’t be receiving benefits, as the SSA has moved all the earnings into its Earnings Suspense File, so their contributions basically represent free money for the Social Security Trust Fund.
That money may not have been worth all the havoc that Musk & Co. are now wreaking on Social Security’s reputation, though. As with dead people on voter rolls, who are hardly ever found to have voted but are regularly cited by those who wish to sow doubt about election integrity, millions of dead people with no death record in the Numident file are a liability that the Social Security Administration can probably no longer afford.
More From Bloomberg Opinion:
Democrats Need to Channel Their Inner DOGE: Matthew Yglesias
Americans Want Higher Taxes for Social Security: Kathryn Edwards
A Dose of DOGE (Sans Elon) Is What Europe Needs: Lionel Laurent
Want more Bloomberg Opinion? OPIN <GO>. Or subscribe to our daily newsletter.
To contact the author of this story:
Justin Fox at [email protected]
About the Author
You May Also Like
