13 Specific COVID-19 Predictions To Guide Fundamental Analysis

Summary

• Human investors are prone to mistakes of cognitive bias, especially in volatile markets driven by an unprecedented risk factor - the COVID-19 virus.
• I explain how to use Bayesian inference to avoid cognitive biases. This process allows specific, verifiable, and update-able predictions to guide the investment narrative.
• I outline 13 specific predictions about the virus outbreak, the economy, and the market that guide my own fundamental analysis.
• Over the next month or so, I expect additional market volatility, with the S&P 500 ultimately declining below 2,400 (17x 2020 EPS of $140) as GDP growth and earnings estimates are revised lower.
• Toward the end of 2020, I expect economic conditions and the market to improve as we all learn to live with the virus and treatments are on the horizon.

Why I Need to Make Predictions

I am a fundamental business analyst, not an economist nor an epidemiologist. But right now, every company's fundamentals have been thrown into question by COVID-19 - an exogenous shock to both supply and demand unprecedented in modern markets. So even as a fundamental analyst, it is impossible to invest responsibly without forming an opinion about how this risk will unfold. This article shares the forecasts I am using to inform my own fundamental analysis.

My goal is to avoid the twin dangers of greed and fear that create cognitive biases and cause bad decisions. To do this, I use a technique called Bayesian inference. The general procedure is to make specific, verifiable predictions to which I assign confidence levels. These confidence levels can then be updated in a calculated way as new evidence becomes available. I then use these predictions to support or change my investment narrative.

The framework I am using for my COVID-19 analysis is divided into three parts:

1. How the virus itself progresses. This is largely a matter of epidemiology but can cross into how sociological and political decisions impact spreading.
2. How the economy responds. This is impacted by #1 but also by how producers and consumers react to the risk, regardless of the actual severity of the virus.
3. How the markets respond. This is impacted by #1 and #2 but also by how investors react to uncertainty.

I will lay out the general narrative I am using to invest followed by the specific predictions I am using to support my narrative.

The Investment Narrative

Regarding the virus itself. In the northern hemisphere, the virus will worsen significantly through March before moderating during the spring and summer. However, the virus will then worsen in the southern hemisphere and re-emerge in the northern hemisphere in the fall. At that point, there will be hope for a vaccine or effective antiviral treatment, and the situation will begin to come under control globally.

Regarding the economy. The virus will cause economic disruption globally. Governments will attempt to curtail activity to contain the virus, creating additional disruption and volatility due to simultaneous shocks to both supply and demand. Q1 real GDP growth will decelerate relative to Q4 for many regions, with Q2 real GDP growth then turning negative. Eventually, the response will shift from containing the virus to protecting the most vulnerable. This will allow economic activity to improve sequentially in Q3 and Q4. By Q1 2021, most regions will return to positive GDP growth.

Regarding the market. This is the most difficult to forecast because it involves getting the previous two factors correct but also predicting how sentiment among market participants will evolve in the face of significant uncertainty. In general, I predict equity markets will experience further weakness from their closing prices on March 6, 2020.

Specific Predictions

These are the specific predictions I'm using to test my investment narrative via Bayesian inference. I will update these periodically to hold myself accountable.

Number of cases. I predict the number of cases as reported by the World Health Organization (WHO) will be at or above the following levels on the following dates (if WHO uses a range, I predict my estimate will be at or above the midpoint of the range).

• The number of total confirmed cases outside China will exceed 500,000 on 3/31/2020. Confidence level: 80%
• The number of total confirmed cases outside China will exceed 2 million on 4/31/2020. Confidence level: 75%
• The number of total confirmed cases outside China will exceed 10 million on 12/31/2020. Confidence level: 70%
• The number of total confirmed cases in the U.S. will exceed 10,000 on 3/31/2020. Confidence level: 80%
• On or before 5/31/2020, the number of new cases reported outside of China will decline for three consecutive days at least once. Confidence level: 65%

Treatments. I predict treatments will be in sight by the end of 2020.

• A Phase II clinical trial for a COVID-19 vaccine will be completed by 9/30/2020, as recorded by clinicaltrials.gov. Confidence level: 75%
• Either the U.S. Food and Drug Administration (FDA) or the European Medicines Agency (EMA) will approve a COVID-19 vaccine by 3/31/2021. Confidence level: 85%
• Either the U.S. Food and Drug Administration (FDA) or the European Medicines Agency (EMA) will approve a COVID-19 antiviral treatment by 12/31/2020. Confidence level: 80%

Response. I predict a rather dramatic social and political response in the U.S. and Europe, which could unsettle markets given the unusual nature of these types of events. Evidence so far includes actions on the part of the Italian government to quarantine much of Northern Italy as well as comments by the director of the U.S. National Institute of Allergy and Infectious Diseases on 3/8/2020 that suggest the U.S. would be willing to consider more aggressive mitigation strategies.

• Within the U.S., there will be at least one attempt to enforce mandatory quarantines on at least a city-wide basis by 4/15/2020 (verifiable by a news report from a major national news outlet). Confidence level: 70%
• NYC will close its public schools for at least one day on or before 4/15/2020. Confidence level: 70%

Economy. I predict a slowing through at least Q2 2020, with a recovery by at least Q1 2021. This is driven by simultaneous shocks to supply (evidenced by the sharp drop to sub-50 in China's PMI) as well as shocks to demand as customers and employees quarantine themselves. These predictions will be confirmed by bea.gov.

• U.S. real GDP growth will be less than 1.5% for Q1 2020. Confidence level: 75%.
• U.S. real GDP growth will be negative for Q2 2020. Confidence level: 75%
• U.S. real GDP growth will be positive for Q1 2021. Confidence level: 70%

Market. The S&P 500 will close below 2,400 on or before 4/15/2020. Confidence level: 65%. This will be driven by gradually deteriorating GDP estimates and downward revisions to S&P 500 earnings estimates as economists and analysts lower their forecasts, particularly as the Q1 earnings season approaches in April. My S&P 500 forecast assumes a 17x multiple on S&P 500 earnings that decline by about 15% in 2020 to $140.

How I Am Investing

My predictions become less confident the further we get from the science of the virus itself. While I believe the market has further room to decline, my confidence in that prediction is lowest of all.

For this reason, my investment objective is to reduce my exposure to broad indices and heavily exposed sectors (travel, tourism, energy, materials, most financials), but to avoid shorting and also to avoid COVID-19 "story stocks" given the risk of over-hype.

Meanwhile, I am making a list of high-quality companies with strong fundamentals that have been too expensive for the last several years and might now become attractive.

Analyst’s Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Comments

[B Brooks]

In a lot of ways these predictions are looking really conservative now. I think they were good predictions to begin with, but now we are going to be flying past most of these numbers. Which I understand was to be expected since he was assigning probabilities well above 50%.

US crossed 10,000 cases in less than half the given time and are currently sitting over 5x the predicted amount with 80% confidence interval.

Cases outside of China will cross over 500,000 on the 26th or 27th. This leaves several days to spare and increases the likelihood of hitting the end of April target.

There are plenty of city and at least a few I believe statewide MANDATORY quarantines.

NYC schools were cancelled around 5 days after the prediction and around a month before the cutoff date. The cancelled school for over a month, when the prediction was for at least a day, and I would be shocked if it doesn't get extended through the rest of the year.

GDP numbers are too early to tell but are looking accurate.

Market dropped below the cutoff as predicted with 65% confidence.

The only other thing I will add is that the case numbers are way off, but not for the reason people think. Yes their are a lot of people with mild symptoms not getting tested, but there is also a bottleneck. 2 weeks ago tests were taking 1-3 days to get confirmed. 1 week ago it was 4-5 days. Now we are above a week. My wife got tested a week ago, and they told her 4-5 days. From another source I have heard that Colorado has been taking 7-10 days. These numbers are on such a huge time delay that it will take 4 weeks from change in actions to trickle down to number of cases. Then an extra week to trickle down to the number of deaths. We are about 1 week into substantial changes and that means these numbers are going to keep running exponentially for about 3 more weeks before it starts to flatten.

[grok42]

Hi @Viceni Investing,

Really like your approach. Reminds me a bit of the book "Superforecasting". The book describes a similar approach, but less formal on the math.

In regards to the virus, it may be important to note that China was able to contain the COVID-19 under worst case conditions. Their new infection numbers were only 26 yesterday, which I think is quite remarkable. South Korea also appears to be achieving containment. Those two examples tell us developed nations can indeed contain the outbreak.

The risk to me seems to be in developing nations with large populations and weak health care systems such as India, Pakistan, Bangladesh, SubSaharan Africa, etc. African countries were able to contain Ebola, so maybe containment is possible in developing countries. If out of control epidemics do occur in those countries, then the question becomes can they be mostly isolated until a vaccine is ready?

Currently Iran is the top concern, imho. From the social media discussions one can get the impression the epidemic is out of control.

Once a vaccine shows up, it would seem to me to be game over for COVID-19 in countries that can afford to vaccinate everyone. There is one firm in Israel that had been working on similar vaccines for chickens that claims they will have an approved vaccine for COVID-19 within 3.5 months. Moderna has already provided a vaccine to NIH for testing. My sense is the window for a vaccine is 4 to 10 months.

So much depends as you note on the spread of the virus until a vaccine shows up. Your model might benefit from having a best case, base case, and worst case scenarios with probabilities for each.

The virus is not all that deadly, imho, compared to SARS, Ebola, etc. There is a fair amount of evidence that the death rate is more like .9% or less, maybe only .5%. The Diamond Princess case is perhaps a useful real world test where everyone was tested and outcomes known. 82% of folks that catch the virus have no or mild symptoms. Most of the mortality is in folks above 65 with co-morbidities. Bottom line is the world may be overreacting a bit on the seriousness of the virus. Of course if you are older with compromised health, it is certainly more worrisome.

Would encourage you to post updates periodically. Would be interesting to see how the model evolves and how close it comes to the reality. Sure sounds like a useful approach.

Best, grok42

[Managing]

Here is my model:

This has the potential to be the first and only pandemic in the history of mankind to be arrested.

Or not.

* Likelihood that no one has any clue about the timing and pace of this disease: Confidence level: 100%.

Best guess if you are trying to time this, is to look at South Korea, they have transparent data and are testing everyone they can. Mind you, they are the "best case" scenario.

[xpan]

I love specific predictions, everybody does. But, when you say that the confirmed cases will be more than half million by the end of March, it doesn't match the trend. A lot of people are saying they don't trust China data, but there's no evidence the data are not true.
When you say the SPX EPS 140 and PE 17, you totally lost my trust. 140 is a wild guess, you are simply baseless. 17 multiple means you don't know how the market works.

[cross]

interesting thoughts. l look forward to tracking the realizations of these predictions in the coming months. Thanks

[Illius]

I like your method. You create a "model" of the world that is based on reason and historical precedent. This is a very good way to sort though the mayhem of so many variables. The model may be wrong, somewhat wrong, just a little wrong or possibly right. It can be changed with new information, tweaked or discarded altogether.

This is actually exactly how science works. People who do not understand science or don't do it professionally do not understand this, and often criticize scientists for being "wrong." . Sometimes a model is deep profound and correct (General Relativity), or it evolves sequentially into a complete and correct theory (as was the history of quantum mechanics). In any case,
sequential modeling of reality is in itself a scientific method (usually stated as the logic of proving or falsifying a hypothesis). I prefer "model" because it may involve many hypotheses
simultaneously. In any case, it's good science, and I think good investment strategy.

[ColdwaterRod]

Re: "Regarding the market. This is the most difficult to forecast because it involves getting the previous two factors correct but also predicting how sentiment among market participants will evolve in the face of significant uncertainty. "

What effect on investor sentiment do you see monetary or fiscal stimulus as having - if any? How about bond yields going 'lower for longer' in relation to SP 500 dividend yields?
How about eventual CVD-19 'burnout' similar to trade war rhetoric burnout?

[ColdwaterRod]

I guess it depends on your retirement horizon and cash level. Buffet likes banks and airlines at this level. (And more at a lower level, I'm sure!). So, this is a lifetime opportunity which value investors salivate over. (I do like your strategy of missing both extremes--at least until the dust settles, and depending on your time horizon).