I started wearing disposable gloves and a mask in early March. Almost everyone was looking at me as if I was some kind of a freak. Today, I don’t wear a mask when I go out and people are looking at me as if I am an idiot.
Most of the time it makes perfect sense to do what the crowd is doing. In investing this is called trend following. It is one of the most profitable investment strategies. If you had bought the S&P 500 Index in 2009 AFTER the index increased 20% from its recent bottom and didn’t sell until the S&P 500 Index declined 20%, you would have returned around 230% in 11 years.
You didn’t have to be an investment genius to achieve this performance. All you have to do is to buy after the market is up 20% and sell after the market is down 20%. If you have some sort of talent in identifying the turning points in the S&P 500 Index slightly in advance, you can achieve even better returns. The S&P 500 Index gained more than 27% since its intra-day bottom on March 23rd. Hedge funds’ top 5 stock picks usually perform much better than the S&P 500 Index. You aren’t going to believe what I am going to tell you now, but it is 100% true: The top 5 stocks among hedge funds lost only 3.2% year to date, whereas the S&P 500 Index lost more than 13%. So, our free recommendation to you is buy the following 5 stocks instead of index funds: Amazon.com Inc. (NASDAQ:AMZN), Facebook Inc. (NASDAQ:FB), Microsoft Corp (NASDAQ:MSFT), Alibaba (NYSE:BABA), and Alphabet Inc. (NASDAQ:GOOGL).
We launched Insider Monkey’s investment newsletters more than 7 years ago and this was the first time in 7 years we intervened. That was because we saw the recession coming at the end of February (read our article here). Here is how I did it.
I read a lot of academic articles and used incomplete data reported by other countries to estimate the parameters of this new coronavirus. We encountered deadlier viruses like SARS, MERS, and Ebola over the last 20 years, but they fizzled out. So, I especially tried to figure how COVID-19 managed to elude our virus fighters. The answer was asymptomatic transmission.
This meant the only way to stop this virus at our borders was to test everybody. Unfortunately, tests aren’t 100% accurate (some tests fail 20-30% of the time to detect someone who is actually infected), so we also have to quarantine them for a period of up to 2 weeks and retest them. That’s what China is doing today to international travelers.
Obviously we weren’t doing any of this and we didn’t close our borders to all inbound traffic, so this meant only one thing: the new coronavirus was freely spreading.
The first parameter I estimated was the life cycle of this virus. I estimated that it took an average of 5-6 days from infection to the onset of symptoms, another 5-6 days from the onset of symptoms to hospitalization, and an average of 14 days from hospitalization to death. Check out the image below from an academic article that was published a week or so ago for the actual distributions:
The second parameter I estimated was the new coronavirus’ infection fatality rate. Initially I used a 0.5% estimate, but later on I switched to 1%.
The third parameter I estimated was the new coronavirus’ doubling rate. Data from other countries indicated that the new virus doubles every 3 days when it is freely spreading. Obviously, once we start implementing social distancing, lockdowns, and other policies, this parameter would change.
Then, I proceeded to setup a very simple model. When someone dies from COVID-19 in a given country, it means this person was infected approximately 24 days. However, since only 1 out of 100 people who were infected 24 days ago dies, we can estimate that there were 99 other infected people 24 days ago as well.
It takes one infected person about 18 days to spread the virus to 100 people. This implies that when a country reports its first COVID-19 death, the virus has been spreading within that country freely for about 6 weeks.
If a country had 100 infected people 24 days ago, this number would double to 200 on day -21, 400 on day -18, 800 on day -15, 1600 on day -12, 3200 on day -9, 6400 on day -6, 12800 on day -3, and 25600 on day 0. Basically, when a country reports its first COVID-19 death, we know that there must be around 25,000 other infected people within the country spreading this virus freely.
This is only if we successfully identified the first COVID-19 death. If there were other unnoticed COVID-19 deaths earlier, the number of infected people would have been even higher.
On March 10th the COVID-19 death toll in the United States stood at 28. Most people weren’t alarmed by this figure. I was terrified. It meant there are potentially 700,000 people already infected with the virus and spreading it.
On March 20th I published an article with the title “Hell Is Coming: Here is the Mathematical Proof”. The U.S. death toll stood at 205 on the morning of March 20th, so the article made the following prediction:
“Except a few educated people, no one has any idea that there are already around 2 million infected people in America today and the American death toll will exceed 15,000 in just 24 days. If we don’t take strict measures, we will be reporting 1000 deaths per day in just 3 weeks. The attacks on 9/11 killed around 3000 people. We will be reporting a 9/11 every three days. That’s why we say “hell is coming”. This is a mathematical certainty. It is inevitable.”
It was indeed inevitable. Luckily, one governor after another took action and imposed “shelter in place” orders. The first state was California (March 19), followed by New Jersey (March 21) and New York (March 22). On March 23rd there were 9 states with statewide “shelter in place” orders. This number jumped to 21 states on March 26th and 30 states on March 30th. Finally, Florida joined all other most populous states and ordered its residents to stay at home effective April 3rd.
Social distancing and lockdowns work.
Let’s say you live in New York City and got infected right at the moment before the state went into lockdown. On average it will take you around 5-6 days to develop symptoms. If you check out the actual distribution curve, you will see that it may take as long as 2 weeks. The probability density function in the above image also tells us that it may take you up to an additional 15 days from the onset of symptoms to get hospitalized. During this 25 day period you may potentially be infectious and can spread the virus to other people (this is the extreme case; most patients aren’t infectious after 10 days or so).
This also tells us that almost all of the people who were infected at the beginning of a statewide lockdown won’t be infectious by day 25. Though it is still possible that these people infected one of their family members around day 5 to day 10, and their family members will pose a threat to the public for the next 25 days. It is also possible that some of the infected people were among the “essential workers” and continued to spread the virus.
However, most people are extremely cautious right now. There are definitely a small number of infected left by day 25 of a statewide lockdown, but these people won’t be able to spread this virus as easily because there aren’t a lot of opportunities. There will still be a few hundreds of new infections in a state with 20 million residents though. That’s why we need to continue the current lockdowns for another few weeks to make sure that the new infections fizzle out.
Today is day 21 of the lockdown in New York State, the epicenter of the U.S. pandemic. I estimate that there are only a few hundreds of new infections. I am not talking about the numbers announced by governor Cuomo. The new cases announced by Cuomo are the infections that they were able to identify and that actually initiated 1-3 weeks ago. I am talking about the new infections that happened today. We have very few of them.
That’s why I am not afraid anymore. The probability of catching COVID-19 in New York City today is probably 100 times smaller than the probability of catching it a month ago. Yet, most people are 100 times more concerned about getting infected now than they were a month ago.
That’s because yesterday more than 2000 Americans lost their lives to COVID-19. The good news is that our death toll isn’t doubling every three days anymore. The death toll on April 10th was only 45% higher than the death toll on April 7th. One day ago that figure was 52%. This means we are only 3-5 days away from a peak in daily death toll.
If we get there sooner, the eventual death toll would be around 50,000. Remember, it takes an average of 24 days to die from a coronavirus infection. That roughly means that half of the deaths will take fewer than 24 days and the other half will take longer than 24 days (I know the distribution isn’t symmetric and some states will peak later than others).
Overall, though, once the daily number of deaths peak and the daily death toll starts to go down, we can estimate the total number of COVID-19 deaths at the end of May by doubling the total death count on the day of the peak. For example, if we peak on April 13th and the death count is 25,000, then we can roughly estimate that the death toll would be 50,000 by the end of May.
If we peak on April 15th at a death toll of 30,000, this will indicate that the death toll would be 60,000 by the end of May. This is what Dr. Fauci was talking about a few days ago.
Obviously all of these estimates assume that we don’t end the current “shelter in place” orders too early, but even if we do, the new wave of deaths will show up in statistics in June and July.
The bottomline is this. We are 100 times safer now than we were a month ago. President Trump and his pandemic response team did a fantastic job over the past few weeks limiting the number of COVID-19 deaths to around 55,000. We won the first battle. I feel sorry for the English, Dutch, and Swedish people though. Their responses weren’t as timely and appropriate as President Trump’s and it is now a mathematical certainty that a larger percentage of their population will die from COVID-19 by the time things settle in a couple of months.
Now we have a different battle to fight. We can’t lower our guard.
The total number of confirmed cases around the world is approaching 2 million. There are millions more that we weren’t able to identify. Anywhere from 20% to 50% of these unidentified COVID-19 infections are asymptomatic or mildly symptomatic. We shouldn’t let anyone into our country (including U.S. citizens) without a 14-day quarantine and a negative test result at the end of the quarantine period.
This restriction should also apply to the residents of the states that didn’t have “shelter in place” orders. People who show any symptoms should be promptly tested, isolated, and all of their contacts should also go through the same process.
Since asymptomatic patients spread this virus very effectively, everyone in public places should be required to wear masks.
There are more than 100 different companies working on a COVID-19 vaccine right now. It is very likely that at least one of those companies will succeed within a year.
These are probably the darkest days of this pandemic and I am certain you will be able to see what I am seeing right now in a couple of weeks.
We did well and saved millions of American lives. This is the most resilient virus we’ve come across over the last 100 years, but the American people proved to be more resilient than the coronavirus.