Reusing numbers would be a bad idea. It would create subtle and difficult programming bugs.
After you die, SSN’s are public. Having multiple people with the same number would mess up genealogy searches.
A bigger problem would occur with databases where there is a relationship between an insured person or similar, and beneficiaries, and contingent beneficiaries, with a mixture of current and deceased/inactive customers.
As for the Y2K type of problem, yes, it would exist, But databases are mostly more flexible than they used to be. Adding another digit, or going alphanumeric, is not as hard as it was.
Since they ought to be treated as opaque strings, the only coding issue is validation. Which ought to live solely in the UI layer and maybe some inter-system boundaries like ETLs.
TIL the real reason for the difference between the first 3 digits in my and my wife’s SSN. She’s 5 years older, but her first 3 are 5xx, mine are 4xx. I always thought it might have been because of reused numbers, nope, just the difference between me getting mine in Texas, she getting hers in California. Ignance fought!
When I was going to the University of Wisconsin, in the 1980s, it was when people were a lot more lax about public usage of SSNs; our student ID numbers were our SSNs, with a check digit appended to the end.
When professors would post test results, typically as computer printouts on their office doors, the results would be listed by our student ID numbers. The vast majority of the ID numbers (and, thus, SSNs) began with “3,” as that was what SSNs issued in Wisconsin started with. However, we had a lot of Minnesota residents at the school (as the UW and the University of Minnesota had a reciprocity agreement on in-state tuitions), and all of the Minnesota students’ ID numbers started with “4.”
The SSA instituted the “Enumeration at Birth” program in 1989, and it’s largely been the norm since – particularly since you now have to place your kids’ SSNs on your tax forms to claim them as dependents.
Our test results were posted the same way, usually with the last 4 or 5 digits. There were a few people whose numbers we figured out, because they were known curve wreckers, but there was one class where the curve wrecker was ME, and of course my friends knew about it but I otherwise kept my mouth shut about it.
I remember when that came in. IIRC I was young 20s / college-ish. And had had my SSN since I was 15-ish & getting ready for my first W2 job.
About 15 years of births got SSNs that year.
The public explanation for the change was to stop the very widespread practice of separated or divorced parents each claiming all their kids as deductible dependents.
Once you had to put down the SSN of each kid you claim, IRS could trivially detect two people on two returns claiming the same kid.
Or claiming completely fictitious kids, which was also said to be a popular tax cheat.
IRS dryly reported that next year the number of dependent kids in the USA had dropped by millions. Hmmm.
The IRS also requires detailed explanations if you count more than (approximately) 22 dependents. I’m not sure what the rationale is behind that number, but this too is another reason we need SSNs to declare someone as a dependent.
When I was a prof, one place I worked at had a carbon-form thing. Several sheets, one for the prof, a couple for the registrar and a special extra one. That sheet had a tear off strip on the side. You were supposed to tear it off and post it. It was just full SSNs and the grade.
As a CS prof I was extremely uncomfortable posting it. But one had to otherwise there would be a stream of nervous students banging on my door to get their grade.
Since it was alphabetically ordered, it didn’t take much of a genius to figure out who was who in a mid-sized class.
Another one here whose college ID number was his SSN. That’s how I memorized my SSN, since I had to use my student ID number much more frequently than my SSN.
Younger people are horrified by this story, just as they are amused by stories of programming in Fortran on punched cards (or for you EEs out there, running SPICE simulations on punched cards - why do you think they call the input a SPICE deck?).
The Canadian SIN would be analogous. We had hourly (union) and staff (non-union) people where I worked, and people could transition from one to the other, so the only record tying people’s two payroll records together for historical purposes was a database of SIN, dates, and employee numbers. Worse yet, they at one point re-used employee numbers until they realized the compuer problems, so SIN was the only thing tying records to a person.
Another thing, probably less of a problem in the USA, was that Revenue Canada started tidying up their databases, and we started getting mail “this employee from now on will use only this SIN, not that one.” A lot of people, particularly from the Maritime provinces, had been registering 2 or more SIN’s in the early days, so they would earn money using one while collecting unemployment insurance on the other. Issues like that had to hand-corrected in the company database.
I assume not one of the problems the SSN had.
There were also complaints that Hockey Canada used SIN’s to identify individual youth hockey players, to avoid people switching teams and possibly using disallowed incentives to change teams, etc.
Unique and ubiquitous identifiers and accompanying “Big Brother” implications was a long discussion in one of my computer science classes.
Our VAX IDs were our student ID minus the first digit IIRC. Since I worked for computer services I also had UCSBPM.
I just missed using punch cards, but I do remember running SPICE and later IGSPICE. A VAX computer was named SPICE (another was named ANSYS)
(Programs used by Electrical and Mechanical Engineers respectively)
OP seems pretty clear-headed about what they are asking. Phone numbers can be recycled because you only need it for as long as you are using it. Even though it’s assigned to a person, it’s not a personal identifier, it’s a device identifier. An SSN is a unique personal identifier that stays with you your whole life, and beyond.
It would, and it’s inevitable. Maybe not today, maybe not tomorrow, but soon and for the rest of your life.
And not just the input format itself will have to be changed, but also the validation rules. And it will still have to support the original format. Even if SSA re-issues new-format SSNs to everybody with an old one, that isn’t going to work for dead people. SSA keeps a “death list” but I don’t think they keep a list of whom to contact about that person. So old numbers will probably stay as-is.
About about half of the 1 billion possible numbers are still available. They are currently issuing new ones at a rate of about 5 million a year. So at current rates, we’ll run out in about 100 years. But the issue rate is likely to grow, so let’s say [pulls number out of ass] 75 years. This is probably one of those boiling frog problems that the government won’t do anything about until it’s an emergency.