What is the population of New York?

All about the state claiming the vaccine is STILL 93% effective.

“It is ridiculous to set a detective story in New York”, Agatha Christie is credited with saying, “New York is itself a detective story”.

That may be true — yet, I will risk being a little ridiculous as there is a story to tell here. New York, after all, has the most effective vaccines in the world. How curious that this state alone is still claiming that the vaccines are up to 92.9% effective in those older than 65. That seems, at best, counterintuitive given even the most self-absorbed vaccinephiles regularly claim that immunity wanes the quickest in older populations.

We are forced, then, to either suspect our disbelief or assume malfeasance, somewhere along the line. To add to the confusion, New York claims that nearly the entire unvaccinated population, at this point, has had the virus, and a heightened level of natural immunity in the unvaccinated population is nothing in the face of regular old alpha-calibrated vaccine antibodies, which, in the rest of the known world, do absolutely nothing in the face of omicron.

Why, one must wonder, would Pfizer’s CEO disregard such strong data when discussing transmission? He should have just pointed to the Empire State to discredit the naysayers. See, the Pfizer vaccine is the superior brand, you just need more boosters, am I right? Or maybe he, too, does not trust what New York is selling.

In any case, there seems to be a mystery afoot.

I will start off by saying I am hesitant to take the case numbers in New York at face value in the first place. I do not trust them. I do not trust the people working for the state. I do not trust their senior level officials. I do not trust their politicians.

The state has heavily politicized the virus more than anywhere else in the world. I am skeptical about who is collecting their data and how it is collected. But let’s pretend, for a moment, that data collection is not the problem, fundamentally, and instead, the problem lies in the representation of fairly collected data. Yes, that’s like playing poker with a known cheat and hoping he has no aces up his sleeve, but data collection can not be the only problem here.

It is interesting that the state does not present a breakdown of raw cases by vaccination status. Instead, they provide case rates per 100k for all populations and merely vaccine effectiveness when broken down by age. A vaccinephile may claim that this is a “fair” way to represent cases. It evens the playing field, they might say, or we would expect more heavily vaccinated populations to have more cases even if the vaccine does work!

That, indeed, is a standard defense when the majority of cases, deaths, or what-have-you are in vaccinated populations. But presenting the data in that way is about obfuscation, plain and simple.

What is the population of New York?

One of the biggest red flags I have been noticing a lot lately is there are an alarming number of jurisdictions that claim vaccinated populations larger than the total populations. As more and more places successfully push vaccines, this is becoming more and more frequent. It should be a hint that the unvaccinated denominator (and probably the vaccinated) is, simply put, incorrect.

New York is no exception to this oddity. In the 65-74 age group, they claim over 99% of the cohort is vaccinated, but hovering over the data we find that their claim is actually over 104% of the cohort. Yikes.

Why is this happening?

Ah.

Their base population is using 2019 American Census Survey population estimates. Now, we are not just going back one or two years, as we have seen public health officials do with the Canadian data, but a full three years worth of population growth. Or is it just three years of population growth that we are missing?

For years, people have been claiming the population of New York is in decline. With crime, corruption, and left-wing politicians, that seems reasonable. Except, it turns out the population decline in New York is a myth. As Noah Smith notes,

“People move to New York because that’s where the jobs are, and jobs move to New York because that’s where the skilled workers are. For New York to go into a sustained decline would take a massive catastrophe — something big enough to remove the nucleus of companies and skilled workers completely. Covid simply wasn’t that”.

And that seems to be a fair assessment. Indeed, the ACS estimates have for years been predicting a decline in the population, yet come census time, the population in New York did not decline. In fact, it’s population boomed. The ACS estimates projected a population of 19,336,776, but the census population ended up being 20,201,249 — a difference of nearly 900,000. An absolute miss. The only state where the ACS estimates similarly missed the mark was New Jersey (both states population growth was under-counted by about 4.5%).

There appears to be something horribly wrong with the assumptions that the ACS population estimates in the state. The vintage 2021 estimates of the population also projected another decline of 360,000 people, so the story goes. I do not believe either estimates. They are missing something. Is the scarcely regulated movement of illegal immigration the difference? It remains unclear to me.

Such a population miss, along with an aging population, was bound to create a scenario where the unvaccinated are hugely under-counted in the numbers. I would even note that the census likely misses the 80-100,000 homeless individuals on the streets of New York, who will be included in the vaccination numbers (incentives, heh) and maybe more than once (who hasn’t thought about it, really?), yet count against the unvaccinated denominator. These individuals likely fall most densely in the 18-64 population bracket.

Furthermore, as the denominators of unvaccinated populations approach 0, the effect of a single case is magnified non-linearly.

Here’s an example. Say a single case occurs in two populations. Population A has 100 people and population B has 1000 people. The case rate per 100k in A is 1000, and just 100 in B. But if we remove 50 people from each population, then the case rate in A becomes 2000, but the case rate in B only changes to 105.26.

When we do not count the unvaccinated correctly, we are essentially just removing population. The less unvaccinated there are to begin with, the more profound the effect.

Over 65 years old

While the data New York is using claims that 3,213,534 individuals are older than 65, and only 13,175 of those are unvaccinated, we know it is incorrect. They also claim that there are only 15,474,107 individuals over the age of 18 as opposed to the census number of 16,088,135 (and remember, up to 100,000 homeless individuals are not counted here).

Unfortunately, the census does not delve into the actual ages of the population, at least not to the granularity we require. ACS estimates from 2020 (prior to the census) are that 16.5% of the population is over 65, and the vintage 2021 estimates are that 16.9% of the population is over 65. Interestingly, the Cornell population estimates, which have been more accurate in terms of predicting the overall population growth (they predicted 20,146,131 in 2020) claims 17.3%.

In any case, at the low end, there are probably 3,333,206 individuals over 65 as of 2020 in New York or 3,475,835 at the high end. This translates to an increase in the unvaccinated population base of between 10.08 to 20.91 TIMES MORE than is currently estimated. And that’s in 2020. The over 65 population has likely grown since then. With Cornell’s estimate for 2022, this number jumps to an unvaccinated population over 65 of 35.99 times greater.

The Cornell numbers seem quite high to me especially since New York is often a place people go to work when they are young, but we actually have no good way of verifying the true population, which is part of the problem in the health authorities trying to do the same with little extra information. The myth that governments are bastions of information sharing is just that, a myth, so they may have used the ACS estimates out of convenience more than anything… though, they should have been aware the numbers were way off and, at the very least, put a big, bold disclaimer.

Over 18 but less than 65

For younger populations, the census does tell us the population of 16,088,135, and this population would depend on the lower and upper bounds of the over 65 populations, so the population between 18 and 65 is likely between 12,612,300 and 12,754,928. The unvaccinated population would then be 1,509,918 to 1,652,546… or 1.30 to 1.43 times more than reported. Plus maybe 100k homeless people and any adjustments for immigration that occurred to the state between 2020 and 2022.

How does this affect reported Vaccine Effectiveness in those 18-64?

This is an interesting one as we do not directly observe case numbers, but, by understanding ratio dynamics, we can indirectly observe how vaccine effectiveness would change even without case numbers.

Recall, vaccine effectiveness is essentially just a measurement of relative risk. We expand the ratio of risk in vaccinated arm / risk in unvaccinated arm even further. Risk in each arm = cases / population in the respective arm.

Let’s me try to explain this intuitively with an example using the 18-64 age group. Unfortunately, New York data breaks effectiveness down into 2 age groups, which we do not have information on. They are relatively similar though, so let’s assume the effectiveness is right in the middle at 62%. That would mean that VE = (1 - 0.38) * 100.

Expanding: 0.38 = risk vaccinated / risk unvaccinated. So in order to get, let’s say, 0% VE as a baseline, either the risk in the vaccinated needs to increase or the risk in the unvaccinated needs to decrease. Or a combination of both. Each risk is also a ratio in and of itself that represents cases over population. Since the top ratio represents the vaccinated population, we can assume it is constant. What would cause the bottom ratio to fall? Either less cases in the unvaccinated population or a larger unvaccinated population.

Just for illustrative purposes, let’s look at the New York official data and find some case numbers that would give us a ratio of 0.38. Any combination will do. For example, 8242 cases in the fully vaccinated population, 10046672, and 2500 cases in the unvaccinated population, 1,158,191.

(8242/10,046,672) / (2500/ 1,158,191) = ~0.38.

So what happens when the unvaccinated population is larger? Recall that we estimated the unvaccinated population for 2020 is 1.3 to 1.43 times larger than the New York official data suggests.

(8242/10,046,672) / (2500/ (1.3 * 1,158,191)) = ~0.494

(8242/10,046,672) / (2500/ (1.43 * 1,158,191)) = ~0.543

Plugging that ratio into our VE equations of 1-relative risk ratio and VE with the larger unvaccinated population is then between 45.7% and 50.6%.

Unanswered questions

Now, that’s likely still an overestimate. Remember, we did not include the homeless population or behavioral differences. There are a number of other questions that need to be answered before we can say there is some vaccine effectiveness in the younger populations of New York, such as:

1. Whether this may also have to do with mandates. Are unvaccinated still forced to test to work there?

2. Testing locations. Is the New York data reporting from, say, Walgreens, who younger age groups may be more likely to test at?

3. Boosters — in other data sets, those with boosters do by far worse than any other group. Could this be an under-boosted population?

4. Reinfections — New York does not add these to the data, IIRC, and there have been hundreds of thousands of reinfections in the past few months.

5. Population movement and growth. This seems like the most obvious one that could result in hundreds of thousands of unvaccinated people not being counted in the denominator. Remember, this is data from 2020. All those aged 16 or 17 have now joined the 18+ age group and many others in this age group may have moved while seeking work.

6. Differences in natural immunity levels.

7. Do people have other suggestions?

How about vaccine effectiveness in the 65 and older age group?

Where the difference is astounding is in the oldest age group. Recall the earlier chart I posted that showed the vaccine being 92.9% effective in this age group. We can do the same calculations¹ for this group using the multipliers we calculated earlier that show the unvaccinated population for 2020 is somewhere between 10.08 to 20.91 greater than the official data lets on. It turns out by running the calculations the real VE would be between -48.4% and 28.5%.

Yikes. That is not anywhere close to the reported VE and that is before even adjusting for population growth. That is an enormous difference that we only need to adjust for the official census data to arrive at. No thinking person who is in charge of this data and has likely stared at it far longer than me should have missed this, yet here we are.

1. For those who would like to try, the “official” numbers for those over 65 are 2,915,789 vaccinated and 13,175 in the unvaccinated. I used 1272 cases in the vaccinated and 81 cases in the unvaccinated as a hypothetical reference to get to 92.9% VE.