When will we run out of Social Insurance Numbers?
Recently I had a question: why does nobody talk about running out of Social Insurance Numbers? Every Canadian, permanent resident in Canada, or someone with a work visa receives a SIN. These numbers are actually a finite resource and it has been bugging me about why this never comes up.
This is kind of a shitpost so don’t take it as completely factual, but some effort was put into guessing things.
What is a SIN?
Beginning in 1967, all Canadians, permanent residents living in Canada, and those with a work visa have been issued a Social Insurance Number. It is used as an identifying number permitting you to pay taxes and access government services. It also has horrible feature creep such as being used for your credit history, but that is for another time.
It is a nine-digit number with some special properties.
046 454 286
While the above is fictitious, it is a valid SIN. However, it does not mean that SINs range from 000 000 000 to 999 999 999 (or a whole
billion of numbers) are available as there is some very basic validation using the very simple Luhn algorithm.
Because of this validation, it actually means there are only 100 million valid SINs, but there are some rules with how it is handled.
The first digit of a SIN indicates the type of number it is. A SIN starting from 1 through to 7 means it is issued based on the region
you were registered in–not born, but often they align. For example, someone with a SIN starting with a 3 for sure means the person
got their number while living in Québec, whereas 7 could either mean British Columbia or Yukon, but also something else of which I
will get to in a moment.
For the rest of the numbers, 0 is reserved for the Canada Revenue Agency, 8 is for business numbers, and 9 is given to people
who are here on a work visa of some sort.
With these rules in mind plus the algorithm that validates the numbers, there’s only a pool of 70 million total allocated. However, while that might seem like a lot, especially considering projected populations by the mid-century, there quite a few factors to consider with how the whole system works.
Population broken down by regions
The 1966 census put Canada at just over 20 million people. When the 2021 census was conducted, we reached over 37 million, but it is likely in the 2026 census, we’ll surpass 42 million.
Using Statistics Canada’s projections, Canada’s population by 2050 could be with a maximum growth scenario, just shy of 55 million people. There’s a low-end of the scale, but even then population projections show that we will be flirting with the 50 million mark regardless.
Here’s how it all breaks down by province or rather, Social Insurance Number areas.
| SIN Region | Area Covered | 1966 | 2021 | 2050 Proj. |
|---|---|---|---|---|
1 |
Nova Scotia New Brunswick Prince Edward Island Newfoundland and Labrador Ontario* |
1,974,758 | 2,409,874 | 3,453,600 |
2, 3 |
Québec | 5,780,845 | 8,502,000 | 9,954,700 |
4, 5 |
Ontario* | 6,735,481 | 13,991,643 | 21,710,495 |
6 |
Northwestern Ontario* Manitoba Saskatchewan Alberta Northwest Territories Nunavut |
3,630,000 | 7,047,520 | 12,192,605 |
7 |
British Columbia Yukon |
1,888,056 | 5,266,932 | 7,602,700 |
If you wish to view the 1966 census, the Government of Canada has a document you can review.
If you wish to view the 2021 census, a non-PDF page can be browsed as well.
Ontario and Northwestern Ontario
Ontario is a special province with special privileges because it is the most populous, so it is given 4 and 5. However, not all of Ontario is
included in the scheme.
Northwestern Ontario is lumped in with the 6 region, which shoves out about 230,000 people from the pool. This number has remained
stable since the earlier census so it doesn’t really have as much of an impact as one might think. I’ve pooled its population with those
regions to reflect this scenario.
However, the province is now eating into what is available for the Maritime provinces and Newfoundland and Labrador, which is region 1. I
did not reflect the population of those provinces but will talk about how I address this later.
British Columbia and Yukon’s problem
So there are 0, 8, and 9 numbers that are available, but they’re not intended to be used geographically. However, 8 has been
exhausted and now is dipping into 7. This means that the available pool for the province and its northern territorial neighbour is ever so
slowly shrinking in availability.
Death
The thing not mentioned thus far is that SINs outside of 9 are unique to each person, meaning that when a person dies, their number is
supposed to go with them. This implies that numbers are exhausting themselves as population grows. Let’s figure out how many people have
died in each region since 1967.
From 1991 up to 2021, 7,415,242 deaths have been recorded, but if you go up to 2024, you can add an additional million to that count. As the baby boomers continue to age and die off, the deaths will continue to increase and that has to be taken into consideration for the future.
Statistics Canada only keeps track of deaths by place of residence and not from where they were born. It has been difficult to track down official numbers on deaths pre-1991, but Macrotrends has it per 1,000 people. Based on that data, we’re looking at about 10 million people having passed away since the SIN system has been implemented.
That is one whole block of numbers that we cannot use any longer.
When will we hit exhaustion?
Here’s how I think the current pool looks like health-wise with the death figure added in. We’ll assume one SIN per counted person because the number of non-citizens would not affect the outcome that much regardless of what someone might tell you.
| SIN Region | Available SINs | In Use | Retired | % Remaining |
|---|---|---|---|---|
1 |
10,000,000 | 2,409,874 | 647,503 | 69.4% |
2, 3 |
20,000,000 | 8,502,000 | 2,284,380 | 46% |
4, 5 |
20,000,000 | 13,991,643 | 3,759,378 | 11.2% |
6 |
10,000,000 | 7,047,520 | 1,893,580 | 10.5% |
7 |
10,000,000 | 5,266,932 | 1,415,158 | 33.1% |
It can be safe to assume that 18.38 million SINs are still available in the system. However, I cannot ascertain how much of region 7
has been affected by the exhaustion in 8, so it is probably safe to assume that it is lower than that.
There are about 459,000 registered corporations federally, with an additional 36,200 more not for profits. The same report suggests that each year, anywhere between 12,000 and 24,000 are dissolved. That’s just for the federal side of things and I am certain that the number of much higher when taking into consideration how many are registered just to individual provinces.
This is a bit beyond my knowledge area, but considering that the 8 block has been exhausted, I am certain the 33% figure for region
7 is significantly lower in availability.
Ontario began to dip into region 1 at some point in the past decade so it’s difficult to say how much has already been consumed
without guesswork. With a projection of up to 7.7 million more residents by 2050, there’s certainty for the province to exhaust that
region’s availability even if a slow growth scenario comes into play.
So I have no real way to address the actual use but I imagine that considering what is available in its original regions, it’s probably under 10% right now, but not for much longer.
Québec though with its regions 2 and 3 has plenty of room to grow as it is only expected to grow a meagre 1.4 million. Even with
expected number retirements, there should remain some room to take on what Ontario is expected to undoubtedly need. Is this a political
possibility? I imagine not, but who knows there.
Here’s the big problem: Western Canada.
British Columbia is expected to grow by 2.3 million people in the coming two decades, which might seem like the biggest problem, but it’s actually everybody else that is going to have a larger impact.
The Prairie provinces plus Northwestern Ontario and the two eastern territories are potentially going to grow by 12.1 million. Alberta alone
is suggested to hit 8.1 million people, putting it 600,000 ahead of its western neighbour. Region 6 includes these provinces and that sort of
growth will for certain exhaust their allocation of SINs.
Based on all of this, it is extremely unlikely to occur in the next decade. I would not be surprised to hear talk about in the 2030s and for sure expect by the 2040s that the way we handle the Social Insurance Number system will require a change.
How to fix this
Like all problems including climate change and resource exhaustion, this is a problem for the future and for sure the Gen Z types who are entering the workforce will be managing this with the generation behind them rolling their eyes at their solutions.
Let’s let this dumb millennial suggest a few and what each has as for a problem.
Just add a digit
This is probably the most sensible idea and can still be verified the same way. It would add about a billion more SINs to the available pool by making each region allocated with 100,000,000 numbers. This should be fine until the United Nations has to take over every country because climate change causes the oceans to rise many metres, causing ruin to national governments and to also ensure political unity in the face of the Martian Congressional Republic.
But that is not going to happen for a few more centuries, so no need to worry.
From a technical standpoint, it would require most to permit a tenth digit. You can grandfather in existing number by appending or prefixing a zero at the end depending on implementation.
This is how I would do it and is therefore unlikely to happen.
Reuse numbers
Don’t.
Because of feature creep such as credit ratings being dependent on these numbers, I would not recommend this. Ten million SINs have likely been retired merely to due to death alone since its implementation and that number is going to be significant considering how many there are left to go.
This would for certain lead to fraud of some sort.
I don’t expect this to occur.
Come up with a new system
From a technical standpoint, this is probably the second worst idea after reusing. Imagine being told that the nine digit number you know off the top of your head from having filled out job applications a dozen times as a teenager is now going to be replaced with something else.
I am willing to bet that some consultants will make retirement by 35 if they were to get the government on board with this one, so expect this one to be floated towards the top unfortunately.
This is the most expensive and will make private corporations a lot of money, so it is probably the most likely.
Ditch the algorithmic check
It’s not hard to generate a SIN number. A simple Python script is often enough.
def check(sin):
def dof(n):
return [int(d) for d in str(n)]
digits = dof(sin)
odd = digits[-1::-2]
even = digits[-2::-2]
csum = 0
csum += sum(odd)
for d in even:
csum += sum(dof(d*2))
return csum % 10
def valid(sin):
return check(sin) == 0
for x in range(100000000,800000000):
if valid(sin=str(x).zfill(9)):
print(x)
The above just generates numbers, confirms it against the Luhn algorithm, and prints it out if valid. This code has been written countless times and I am sure sits in various repositories. With a few changes, you can use the above to generate credit card numbers too since they use the same check–I recommend against openly doing this for legal reasons.
If you have a better way to write the above, have at it. I am a terrible software developer.
The actual problem is attaching to a name. However, offline checking of a SIN is useful in some scenarios. The government does not want everybody to all the time check against some central computer.
However, as I write this, I don’t think that I have ever had my SIN confirmed as my own. You cannot even get a plastic card with the numbers on it so does it matter? It does make it difficult to spit out a number on the spot, but like the code example above, you could just randomly pick from one anyway and then make it a problem when it comes to tax time.
Because of a potential fraud angle I am sure that the government could come up with, I don’t think that they will choose this option.
Closing
I am so glad that this is not my issue to tackle ultimately. Perhaps the 2038 computer bug issue will solve this problem for us, which is about the time the federal government being in the midst of taking this issue seriously.