This might sound ridiculous, but sometimes I suspect it might be a waste of time for market participants to obsess over every incremental data point — every new study — documenting the evolution of vaccine efficacy.
While you could argue that the durability of the protection provided by the shots is all that matters, there’s a sense in which that’s only accurate in a scenario where they stop reducing the odds of hospitalization and death.
To be sure, nobody wants to hear that efficacy against infection (and especially symptomatic infection) decreases materially over a relatively short timeframe. But as long as vaccinated people aren’t being hospitalized regularly with severe disease, nobody is going to blame the trader in you for ignoring the daily virus headlines, even if some folks might blame the human in you for exhibiting something like callous disregard for an ongoing global tragedy.
The Delta variant introduced considerable uncertainty into an already ambiguous macro environment, but generally speaking (and give me some rope here), there’s not much in the way of compelling evidence to suggest that the strain is capable of completely evading the vaccines on the way to creating an acute crisis among the fully vaccinated.
In a new note, Goldman looked at the evidence. And before you roll your eyes, be apprised that the bank didn’t offer any medical opinion — they merely documented the results of recent studies.
“Some recent estimates of vaccine protection from infections have been much lower than the headline results last fall,” the bank wrote. The simple figure (below, from Goldman) is very useful for those who find it difficult to keep apprised of the dizzying array of incoming data from disparate studies conducted across various locales.
The obvious takeaway is that efficacy against infection diminishes quite significantly, while protection against symptomatic infection and, crucially, hospitalizations, wanes much less.
The actual numbers vary depending on the study and the vaccine, which is what makes this so difficult to track in real time. It feels like every other morning there’s a new set of competing headlines.
Goldman averaged across what they called the “most robust studies” and determined that efficacy against infections falls from ~80% one month after full vaccination to just 55% five months later, while protection against hospitalizations stays generally steady at ~90% after six months. Between Moderna, Pfizer and Astra, Moderna offers the highest rate of protection — or at least based on Goldman’s analysis of study results.
While “no study found a statistically significant increase in hospitalization risk,” the bank wrote that “given the high starting level of hospitalization efficacy even a small decline could have a large impact on the number of breakthrough hospitalizations.” So, that’s something to keep in mind.
As far as booster shots go, the bank noted that “early Israeli evidence suggests that booster shots return vaccine protection against infections to high levels and should raise protection against hospitalizations even higher.”
Meanwhile, in a separate note, the bank’s Zach Pandl said that in Goldman’s view, “the market is worrying too much about global cyclical risks from Delta outbreaks and China’s slowdown, and our Fed forecast is still more dovish than the market’s.”
2022 might be “less friendly,” Goldman reckoned. But for now, carry on, I suppose. And remember: You may be a bull or a bear, but “you are not a horse.”
Here is something that is often overlooked when reviewing vaccine efficacy data. Because outcomes differ substantially by age, as you dig further into the detals, it is important to compare by age cohort.
For example, I wanted to see if vaccinated persons have better outcomes should they get Covid (i.e. are “breakthrough” cases). So I pulled the data for my state.
– For “all” confirmed Covid cases, the hospitalization rate is 5.4% and the case fatality rate (CFR) is 1.15% (14,871 hospitalized and 3,155 deaths, from 273,896 cases).
– For “breakthrough” Covid cases, the hospitalization rate is 5.48% and the CFR is 0.86% (523 hospitalized and 82 deaths, from 9,539 cases).
5.43% vs 5.48% and 1.15% vs 0.86%, hmm. Does that mean than vaccination does not improve outcomes for breakthrough cases? Wrong – that is a totally mistaken inference.
In each age cohort, the breakthrough cases have substantially lower hospitalization rate and CFR than the “all” cases. This is even true for the older cohorts who have the highest hospitalization rate and CFR whether vaccinated or not:
– 50-59 y/o breakthrough cases have 50% lower hospitalization rate and 80% lower CFR, compared to “all” 50-59 y/o cases.
– 60-69 y/o breakthrough cases have 43% lower hospitalization rate and 68% lower CFR, compared to “all” 60-69 y/o cases.
– 70-79 y/o breakthrough cases have 33% lower hospitalization rate and 59% lower CFR, compared to “all” 70-79 y/o cases.
– 80+ y/o breakthrough cases have 16% lower hospitalization rate and 64% lower CFR, compared to “all” 80+ y/o cases.
What’s going on? If each cohort individually benefits, why do the overall numbers seem to show no benefit?
Explanation: the vaccines are Covid-protective for everyone, but the oldest cohorts get relatively less protection than the youngest cohorts. Makes sense, older persons have weaker immune systems generally and more of them are immunocompromised, so there is less for the vaccines to train and boost. So if you are going to be a breakthrough case, you’re more likely to be older and older Covid cases have worse outcomes whether vaxed or not. In essence, the breakthrough case population is biased to the most vulnerable cohorts.
A lot of people don’t understand the cohort bias effect. Most media and anti-vaxxers certainly don’t.
Unfortunately, data collection in the US is poor and we tend to see all cohorts lumped together. Israel and the UK have better data, being more centralized. But you can usually dig out the age cohort data on at least a state basis, if your state isn’t suppressing it.
Also, it is important to remember that overall outcomes are weighted averages. Each cohort not only displays different results but the older cohorts are smaller and contribute less weight to overall averages. When I was in grad school I took Intro to insurance. The prof came in the first day and began to explain the basic principle of all insurance underwriting, that victim identities don’t matter, only the numbers do. He produced a toy gun and turned his back to the class, which had 50 students. He pointed the gun over his shoulder an started waving it about. He said, “Imagine this gun is real and I have one shot. What are the odds I might kill someone?” The odds, of course, are less than or equal to 2%, 1/50. What will happen is only important to the one it happens to. To the insurance company they only have to know the odds and collect enough in premiums from the 50 people to pay the claim of the loser. The odds of a breakthrough are what they are, not what they are reported to be. The only thing that really matters is whether or not any one of us, fully vaccinated, gets the plague anyway, and how badly. The virus doesn’t care who gets it and neither does the medical community. It either spares any one of us or it doesn’t. Statistics my friends, it’s the law.
It seems like the odds of killing a specific person would be 1in 50 but the odds of killing “someone” would be a lot higher. If I roll a ball at 12 skittles, the odds of hitting one specific skittle might be 1 in 12 but the odds of hitting any skittle is much greater.
No matter what the odds, the insurer doesn’t care and it won’t matter to you what the odds are unless you die.
Thank you Heisenberg for this article.
Interestingly, there is recent data to suggest that as the Covid virus mutates, human memory B-cells (maybe T-cells also) in vaccinated people are also mutating in a good way — to support longer-term immunity. The concept that humans have evolutionary adaptive immune systems is an important concept to ponder. Nonetheless, the people that make a choice to ignore the reality of vaccinations supporting immune function are basically just a group of lemmings that are herding themselves into extinction — which in a very large perspective, is a good thing for all the survivors.