Equity Valuation – Part 1 Without COVID-19

To understand the current equity market valuation, a good place to start is to do a quick analysis of what market value looked like at the start of the year. We can then look to separate the impact of COVID on the economy and markets. At the time of writing, the S&P is heading towards 3,000.

PE ratio

A simple and decent method of equity valuation is to look at the PE ratio overtime – price of equities divided by their annual Earnings. For the S&P, it is currently just over 20, which is roughly the middle of the range of the past 5 years. Furthermore, it looks even more reasonable when compared with fixed income where bond yields are close to nothing.

Over the same period the stock market has risen strongly. In the chart below, I show that the value of the S&P has doubled in the last decade, as have quarterly earnings (S&P EPS)

Where does earnings growth come from?

The answer is of course that companies are part of the economy and as the economy grows then corporate earnings also grow. In the chart below, we see uninterrupted US GDP growth for the past 10 years and also a rising S&P based upon growing earnings.

This must all be consistent then?

This simple look at equities implies that the rapid rise in equities over the past decade has been fully justified by the fundamentals in the economy.

But I played a little trick in the charts above.
While the stock market, GDP and corporate earnings have all gone up over the past 10 years they have not gone up by the same amount.

  • Nominal GDP has risen by 50%
  • Corporate earnings have doubled
  • The US stock market has tripled

The part I want to look at more closely is the difference between GDP growth and the far more rapid growth in corporate earnings.

Any Reasons to be cautious?

It is remarkable to have such a large rise in earnings when compared to the overall economy and one would expect to see profits as a share of GDP to have risen significantly over the same period. With decent overall growth, then you might expect growth in profits to be better, and to offset this the share of the economy going to workers would be reduced. This would fit the idea that while growth has been good, more of the benefits go to the capital and capital owners and less to workers; in this recovery a lack of real wage growth is often cited as a concern in the face of rising equity prices.

When I look at the data however, it does not fit this intuition. This is a chart of corporate profits as a share of GDP over the past 60 years.

We can see it strongly mean-reverting and so the recent pattern is exactly what we would expect to see. In a recession wages are sticky and do not readily fall; it is companies that have profitability issues. We all know that companies go bust in recessions and in 2001 profits fell to 7% of GDP. In the early stages of a recovery, it is profits which rebound the fastest, wages remain subdued as there is still high unemployment so the gains from GDP growth go to businesses, and by 2014 profits as a share of GDP had nearly doubled to 12%. But later in the cycle this reverses, and profit margins are squeezed, which they have been for the past 5 years.

This does not make sense!

I have just told you two contradictory stories. One is that corporate earnings have been rising rapidly over the past few years, far faster than GDP growth. The other is that profits as a share of GDP have been falling as we would expect late in the economic cycle.

The reason I can tell you two completely different stories is that I have two different data sources. The first is the earnings (S&P EPS) as they are reported by companies. The second is profits as they are recorded in the income method GDP data – known as the NIPA data (National Income and Product Accounts).

These two ways of measuring earnings are not exactly the same. For example, the S+P is only 500 companies whereas NIPA represents the entire US corporate sector. There are also differences in accounting and tax. But logically they are highly similar, and it is no surprise that historically they track very well. Here is a chart from the late 80s until 2014 showing the rise in earnings (EPS) as measured by companies and as measured by the National Accounts. We can see that in the long run, they track pretty well but there are some periods of divergence for instance around the time of the dot com boom in the late 90s.

If I draw the more recent history, we can again see this divergence. The reported earnings have been rising rapidly (white line) while the NIPA measure of profits has been stagnating (orange line).

The relationship is easier to see if we take them as a ratio. In the chart below we see a large spike around the Dot Com bubble (EPS growing more than NIPA), a large spike down during the Great Recession (the opposite) and another large spike higher recently.

These differences are so large that they require an explanation.

  1. Dot Com spike

A bubble emerged in the late 90s with very high PE ratios i.e. companies were expensive compared to earnings. In addition, these reported earnings were inflated; a famous example you may recall was Enron who were highly “creative” with how they recorded and reported their earnings. When the bubble bursts and we move into recession then these accounting methods are not sustainable, and we see the rapid fall in reported earnings and the ratio of reported earnings to NIPA data renormalizes.

There is a danger in that earnings are presented to us by corporates in the most flattering version they can create. In a bull market, there are opportunities to keep presenting this managed version, perhaps similar to how people curate their Instagram feed. Through heavy use of filters and selective framing, someone might look as though they are very attractive with an opulent lifestyle. The recession is the equivalent of when you meet them in person and realise that the reality is not exactly what was promised.

  1. Great Recession

This rapid fall in reported earnings is easily explained as a result of huge write-downs taken by financial firms. This is a good example of an item that is recorded by companies as a change in earning, but not included in GDP data. Once the write-downs have finished, the ratio between reported earnings and NIPA profits renormalizes.

  1. Now

I have been searching for a good explanation of this divergence and am yet to find one. One plausible idea is stock buybacks, but this is not true as they are adjusted for in the earnings data. Other sources of divergence such as tax are real but do not come close to explaining the large difference.

Could the NIPA data be wrong?
This is data that will get revised, but it would take something extraordinary for GDP revisions to change corporate profits by the 40% divergence we have seen to EPS

Could it be financial accounting manipulation?
Some argue the rules are so much stricter now, so it is not possible. Surely what we learned from the last crisis is that Rating Agencies having strict rules on how to make a security AAA that enabled smart bankers to arbitrage those rules. I do not believe we could ever have rules so strict that smart bankers cannot find ways to optimise them. This does not mean that people must be lying or breaking the law. Bear in mind virtually everything that Enron did to inflate their earnings was legal.

Could it be offshoring profits?
There could be something here and it is a very murky and complex area. We know that the large tech companies have found ways to limit their onshore earnings, keeping profits in countries where they have to pay no tax but this is hardly a new phenomenon. It is something I will be looking into as an explanation.

Could it just not be a problem?
Maybe this is the first time we have ever seen a large, rapid, permanent shift in the relationship between reported earnings and NIPA data. Maybe. Most people in effect seem to be assuming this and financial markets are not at all concerned.

What if the NIPA data is right?

If the NIPA data is correct, then the PE ratio is currently far higher than 20. In the previous two cycles it was reported earnings that correct, not the NIPA data, and the timing of the correction is during a recession. As Warren Buffet said “You only find out who is swimming naked when the tide goes out”.


My view of value before COVID was that it was reasonable to belong to equity markets early in 2020. I was aware of the NIPA data divergence as an issue, but it has been an issue for a long time and it has not been a market driver. The trend towards higher earnings and higher equity prices had been very strong and I believed that it would take an event or catalyst to reverse it. If we are heading into a significant recession, then this may be the time we understand if the relationship between reported earnings and GDP profits will reconnect.

3 paths forward for the economy

The pandemic is affecting the entire world, but we are seeing very different approaches to dealing with it in different countries. I have identified 3 broad strategies each of which will have very different outcomes in terms of the economic outlook. In this post I am not looking to evaluate the approaches from an ethical or public health perspective, but purely on what the likely economic results will be.

Path 1 – Control the Virus e.g. New Zealand

This is the approach I have been advocating and which I spoke about on my recent IAN talk. Please click below if you are interested in watching.


This approach has been very successful at controlling the virus in some countries such as New Zealand, Norway, Singapore, China and Australia. All of these countries now have low levels of infection and can look at a material reopening of their economies. Other countries have been trying to play catch up, are past the peak in their infections and are maintaining lockdowns until the level of infections comes to a low level. This group of countries includes Canada and most of the EU.

This has not been an easy path to follow and for it to be successful it does not only take the good public policy but also it critically requires huge public support for measures which involve massive changes in behaviour. The countries that have been successful have given clear and detailed public health messages which have been adhered to due to high levels of public support and trust in the country’s government and institutions.

The economic impact of COVID has been very severe but these countries are the most likely to hope to experience COVID as being economically similar to a natural disaster and see a sharp rebound in economic activity. Bernanke, ex-Chairman of the Federal Reserve in the US, likened COVID to a snowstorm. In a snowstorm, we halt much of the economic activity but once the snow melts, we can go back to how we were before.

Waiting until the level of infection is very low has two big economic benefits

  1. The number of people infected is low, which means that anyone you interact with is highly unlikely to have the virus and so all the allowed activities are in effect very safe. This means people are confident to go back to work and do the allowed activities as you reopen.
  2. There is room to experiment with the reopening and how it impacts R0. If the impact is larger than you expected for an activity, then you have time to reverse the policy as the absolute number of infections remains very low. There is no danger of the virus getting out of control quickly.

With COVID the world will not be the same as before, but it may be possible to reopen large parts of the economy while keeping the rate of transmission and level of infections low.

Path 2 – Herd Immunity e.g. Sweden

I have used Sweden as my example but at the moment, to be honest, it is my only one and is a global outlier. It is the only country I know which has chosen to follow this path with high public approval and consent which makes all the difference to the potential economic outcome.

The choice to manage the spread of the virus with the longer-term goal of herd immunity is clearly possible. The economy is not shocked by the strict lockdowns and only a milder form of restriction is made to slow rather than to reduce the virus. Sweden has high levels of public support with 70% of the population supporting its government’s approach. With public trust and support the economy continues to function with a far milder slowdown in economic activity.

The data from the antibodies tests in Stockholm last week were broadly in line with my expectations but completely shocking to the Swedish authorities. With only 7.3% of Stockholm residents testing positive for the antibodies we can imply a mortality rate of 1.5-2.0%. This is slightly higher than the results from New York, France, Spain and the UK which all suggested 1.0-1.5% but I would suggest is within a margin of error given the relatively small sample size.

From outside Sweden we might think that this result would lead to a change in policy and a dramatic fall in public support but so far this does not seem to be the case at all. This implies that Sweden is uniquely happy to follow this route and I cannot say that the economic outcome should be poor. Whether it is better or worse economically than the path chosen by other countries such as Norway is not clear yet, but I see no reason to think it materially better or worse.

While I said that Sweden is my only example of herd immunity with broad public support it is not my only example of a country pursuing this strategy. A good example is Brazil where the government under Bolsonaro has deemed COVID to be “a little flu” and will not pursue any public health measures. But the population is not supportive of this plan which will lead to severe economic disruption. Without consent the population is afraid and so its behaviours change, both on the supply side as people do not want to go to dangerous workplaces, and on the demand side as a lack of confidence leads to a decline in spending and investment.

It is important to note that from a public health perspective Path 2 is the default strategy. Anything less than a strongly concerted effort to control the virus will make it continue to spread and grow. But going down this path with consent like Sweden has a very different economic outcome to doing it without consent.

Path 3 – Neither of the above e.g. UK, US

This occurs when there is a major divide between those who want to follow the two paths above. A majority of the population would prefer to take Path 1 and control the virus but a significant minority, including the government, prefer Path 2 to reopen the economy even if the virus is not controlled. This is the situation in the US and increasingly clearly also in the UK.

For both the US and the UK we went into lockdown late, made the lockdown not very strict and are moving to ease restrictions before the level of infection is low and declining. There has been some attempt to control the virus, but it has been half-hearted and thus ineffective.

Trump gets pretty uniformly panned in the UK for his approach but in a way, it has a certain brutal honesty to it. He does not really pretend that he cares about public health, he is very clear that he cares about the economy. He does not argue that it is safe to reopen, he argues that the economy comes first. The US public has a majority in favour of controlling the virus but has a significant minority that strongly prefers Path 2. Those people are also Trump’s support base and so politically it should not be surprising that he is pursuing a policy path which they support even if a majority of the broader country do not. Trump has never really been a President who wanted to govern for everyone, he governs for his base and his interests. It remains to be seen if he can generate widespread support for his approach.

The UK political landscape is somewhat different. We have a clear majority in favour of following Path 1 but again a significant minority who prefer Path 2. Looking at the opinion polls I do not see the strong divide in the country we had with Brexit. 73% of the UK population wants clear advice with detailed guidance, not the muddling appeal to “common sense” that we are getting. It is true that Conservative voters, Leave voters and older people are more likely to want to reopen earlier but there is a majority in all categories for a focus on public health and caution on reopening. But importantly the relatively small minority who are very keen on a rapid reopening are the small subset who are the type of people who are members of the Conservative Party and also the type of people who are Conservative MPs and in government.


I am sure there are many in the UK government who passionately want to follow Path 1, perhaps after his ICU experience that even includes Boris Johnson. But we also have Dominic Cummings who showed what he thought of the “Stay at Home” slogan by travelling 250 miles to Durham, and Rishi Sunak who is arguing for a swift reopening.

The effect of a split in government and a lack of clear direction is that the default strategy of letting the virus run away again becomes the effective policy. This is in effect the same policy that Trump has adopted but confusingly the rhetoric is entirely the opposite. When Trump argues for reopening, he talks about individual liberty and the economy. When the UK government argues for school reopening it does not say what everyone knows i.e. it is an economic priority. Instead, Gove tells us it is “safe”, and Williamson uses emotional blackmail on teachers to tell them their pupils need them. This contradiction between rhetoric and reality leads to further erosion of public trust.

To follow Path 1 the government needed to build a consensus in early March for measures which the public initially did not think we needed. It could still attempt to use the majority support for prioritising public health now to try again to make more effective policy. Instead, it is using the rhetoric of public health to sell the policies of reopening. Even now the slogan is

“Stay Alert – Control the Virus – Save Lives”

This is a slogan for Path 1. But all the policies are moving towards Path 2. I do not think this has a good outcome. The result will not be the economic recovery they hope for as without trust and security our confidence is so impaired that no recovery is possible.

If we combine the rhetoric of a government focused on controlling the virus, with the actions of a government that will not do so, I expect to see both a rising infection rate and also a misfiring economy. In this case, perhaps the government will admit its mistakes and refocus on being effective in controlling COVID. Or perhaps they will claim they have done everything they can, that controlling the virus was not possible as people did not “Stay Alert”, and we should move down Path 2.

Which of those sounds more in character for this government?

Does policy matter?

It is all too easy to become fatalistic when faced with a massive problem like COVID-19. One could just assume that there is nothing we can do; all the deaths and economic destruction are inevitable and we just have to get on with things. I do not agree. Good policy is possible and the contrast between the approach of the UK and NZ is a useful demonstration of this.

The chart below compares actual daily deaths in the UK (orange) with a projection for the next few months (blue). The timing and size of the initial peak were dictated by the decision to go into lockdown on March 23rd. As inputs to the projection, I allow R0 to stay around 1 for a few weeks and then creep up to 1.3. From early June daily deaths begin to rise rapidly again; if we want the next peak to be at a similar level to the last, then a further lockdown will be required soon.

As an alternative scenario, let’s look at the model if we had gone into lockdown on March 9th. This was the date that Italy went into lockdown whereas here in the UK, Boris Johnson spent the Daily Briefing explaining that he was still shaking hands with everyone, including COVID patients in hospital.

Blue again is the model, this time based upon the earlier lockdown assumption; orange actual deaths in the UK. The difference may appear shocking but comes directly from the exponential growth of virus cases. As opposed to peaking around 1,000 deaths, we would have peaked at below 200. Rather than the current many hundreds of deaths per day, we would have perhaps 10.

This exponential growth has some simple intuition around it. Early on during the 2-week period when we were told to wash our hands, but not to socially distance or stay home, I estimate R0 to be 2.0. R0 of 2 means the number of infections doubles every 5 days – therefore in 2 weeks the number of infections would double 3 times i.e. 2, 4, 8 –2^3. If instead the R0 were 0.8, then the number of people infected would change by a factor of (0.8)^3 – which is to roughly halve. Comparing these, entering lockdown earlier would have halved cases rather increasing by a factor of 8 – a total of a 16-fold impact on the number of cases. This is why the blue line above is so much lower than the orange line.

The impact of this decision has not just been seen in the extra tens of thousands of deaths, it also makes a huge difference in how we can loosen the lockdown. In the projection above, I still assume a post lockdown R0 of 1.3. If the overall level of infections is much lower, then in absolute terms the growth in the virus is hugely lower. This would mean we would have far more room to experiment with lockdown easing measures such as school reopening without risking an imminent large outbreak.

This scenario shows to me that policy choices really matter. The example of NZ where they did exactly this shows that it is possible in the real world. They went into lockdown BEFORE they had their first death from COVID. The infection levels were low and now they have effectively eradicated the virus with a total of 21 deaths. This is why the UK government does not want to talk about international comparisons or mistakes they have made. We had plenty of warnings and examples from other countries on what we should do, but the UK has made catastrophic policy decisions that have cost tens of thousands of lives and made it hard to contemplate the removal of the lockdown. However, we do not have to continue making terrible decisions and better policy is possible.

Model update

We have had some higher quality data recently which we can use to compare to my model assumptions and projections

Update on Mortality Rate

I used 1% in my models from early April. This still looks a good estimate but now I think rather than the range being 0.5-1.0% it is perhaps 1.0-1.5%

The reason is that we now have had 3 high quality pieces of data from large samples of people from areas of high infection to see how many people have been infected. I have already written about the New York results which suggested something a little over 1% mortality rate. Spain found that 5% of its population have had it and France 4.5% In Spain this result means that if we just take the raw COVID death data we have a mortality rate of 1.1%. If we take a more reasonable assumption that this data is underreported and use the excess death data instead then the true Infection Mortality Rate would be perhaps 1.4%.

I used the updated mortality rate of 1.1% to relook at the path of deaths in Spain and from that estimate what R0 must have been.

To get this fit we need an R0 of close to 3 early in the infection, dropping to 2 as people became aware of the pandemic and then to 0.8 after the lockdown on March 13th. This is the same as my models from last month.

Update on R0

Public Health England has recently suggested that the overall R0 in the UK is currently just below 1 which is a combination of being far below this in London (perhaps 0.4) and likely above 1 in other parts of the country such as the North East.

Source: FT

Since we are relaxing lockdown it seems reasonable to project that the R0 will rise from here. In the projection below I have put the future R0 at just 1.2 and if we want the next peak in deaths to be the same as the last one then we would need to go into lockdown in early July i.e. about 6 weeks’ time. Then we would come out of lockdown again in early September i.e. we spend more time in lockdown than we do in periods of minor easing and we never reach Boris’ Phase 2 of further easing of lockdown restrictions.

Is there a more optimistic projection?

To be much more optimistic we need to find a way to limit transmission without going back into lockdown. We must find more targeted ways to change behaviour rather than the very blunt tool of making us all stay inside our homes. This could be Test, Trace and Isolate or better public information on how to limit transmission.

How else could the projection be changed?

The simplest way is that policy is different. If lockdown becomes more severe then the R0 never rises and deaths keep falling. If instead we choose to carry on along Boris’ path of further lockdown relaxation then the R0 would be higher again and the number of deaths would rise far more rapidly. If we choose not to go back into lockdown then the number of deaths just keeps rising exponentially.

What is safe?

As expected, many countries are now relaxing their lockdowns as case numbers and deaths move past the peak. In the UK on Sunday, the PM took 15 minutes to explain what was going on although spectacularly failed to explain anything. The debacle, watched by 30 million people, has left us with vague and often contradictory public health messaging about how we should all behave, which is the worst possible approach.

This article provides a summary of the advice and attempted “clarifications” from the past 3 days: https://www.huffpost.com/entry/how-boris-johnsons-stay-alert-message-unravelled-in-24-calamitous-hours_n_5ebb4d71c5b6b58e4cc98f7c

With the government failing to provide credible expert guidance, we have been left largely on our own. Tempted to rely on our “common sense”, humans are terrible at intuitively understanding risk, especially where statistics and brand-new problems are involved. We love to draw analogies e.g. “it is just a flu” even if these analogies are not very good. My approach is to research the virus as much as I can and share my current thinking on what I’m doing to stay safe.

I have been disappointed by the output from governments, health bodies and academia, but have found a few bright people able to communicate the science. Prof Erin Bromage of Dartmouth University is excellent, and I have incorporated much of his recent post in my thinking. https://www.erinbromage.com/post/about-the-author-professor-erin-bromage

Please be careful and do not imagine I am giving you all health advice or that there is certainty on any of these items. This is simply my current understanding of the science and how it applies to behaviour.

How the virus is transmitted

My understanding of how we become infected through surfaces has not changed very much and I have nothing to add to the official guidance. The advice on handwashing and not touching your face, combined with regular cleaning of surfaces seems a highly sensible way to reduce the risk of transmission.

I have changed my thinking around infection through the air. It seems that most cases of transmission come via the air rather than surfaces and given this is the main risk it should be the focus of our public health advice on return to work as it is much harder to manage as an individual. In particular the risk comes from aerosol i.e. the very small particles which last far longer in the air and move further than the 2m social distance guidelines.

Initially I made the mistake of thinking that viral transmission is a bit like a game of tig – if you got too close to an infected person you were “it”. It is often portrayed that you are safe from being “it” or safe from infection if you stay more than 2 metres away from other people, “social distancing”. The more I have read, the more this seems a poor metaphor for the transmission of a virus.

To become infected with a virus, there is a threshold number of virus particles required to enter your system. If there are just a few particles, then your immune system recognises them, bats them away and you are not infected. If there are many, the system is overwhelmed, and the virus freely replicates.

This number must vary by person but let’s assume a threshold of 1000 particles as a reasonable benchmark. Your risk of infection is determined by the number of particles in the air and how long you are exposed for:

Infection = Exposure x Time

If your exposure were 50 particles per minute and you spent 20 minutes in that environment, then you would reach the threshold of 1000 particles and be at serious risk of infection. Higher numbers of particles in the air or longer exposure both of course lead to higher risk.

How many virus particles does an infected person put into the air?

Initially let’s assume we are outside in good airflow and let’s ignore how long the particles stay in the air. The number of particles released by an infected person will vary depending on whether the person is:

  1. Breathing
  2. Speaking
  3. Heavy breathing/singing
  4. Coughing/sneezing

Considering each in turn:

  1. Breathing– not very risky

There is a low amount of viral material, perhaps only 3-20 virus RNA copies per minute. In addition, the water particles you breathe out upon which the virus sits are low velocity and so drop to the ground quickly. This means that even without good social distancing it would take a long time to become infected perhaps many hours, so in this situation social distancing should be very effective.

  1. Speaking – more risky

Speaking generates as many as 10 times as many droplets as breathing i.e. up to 200 per minute which means we could reach our 1000 particle threshold in as little as 5 minutes. Of course, in most conversations we each speak about half the time with pauses, so it seems more reasonable to think of 10-15 mins as a reasonable threshold for dangerous exposure levels. Here the good news is that the velocity of the particles is again low and so social distancing of at least 2m should be effective. Adding more distance of course, would make this even safer and so extended conversation outdoors should be safe.

  1. Heavy breathing/singing – be far more careful
    Playing sport or singing in a choir is significantly more dangerous than simply breathing or speaking. Here the number of particles and their radius of spread is far greater. If someone is simply walking past, then the time of exposure is so short so as to not be too risky, but any activity where you are within a few metres for an extended period will be dangerous.
  2. Cough and sneeze – very risky
    The number of virus particles expelled in coughs and sneezes is enormous and they travel at high velocity so social distancing will not protect you as the particles will go a lot further than 2m. If someone is coughing or sneezing, then you simply have to try to be nowhere near them and hope they sneeze in the other direction.

What does this mean for what we can safely do?
For both breathing and speaking, it takes time to get a dangerous level of exposure and social distancing outside will work. Passing people in the street or someone running past you is not dangerous and I feel less paranoid about passing contact with strangers. I can also be quite confident about talking to people more than 2m away outside, especially in the sunshine, where UV does a fantastic job of killing the virus in the air and on any surface.

But this is notably different from the current government advice which has been very restrictive on people having socially distanced interactions outside in the sunshine. It seems now you are allowed to sit 2m away from people outside but only if you don’t know them. I have found there to be a large difference between what I think is safe and the current government advice. They are still very restrictive on outdoor activities I think are safe and promoting indoor activities I think are still very dangerous.

What about Indoors?

The situation indoors is very different from outdoors and inherently riskier. Let’s consider 5 risk factors:

  1. Type breathing (normal breath, speaking, heavy breathing, cough, sneeze)
  2. Number of people
  3. Size of space
  4. Amount of ventilation
  5. Length of time

As described above socially distanced breathing and speaking are lower risk, but in an enclosed space, particles will continue to circulate in the room so you will not be protected from prolonged exposure. The guidance today from the government on workplace safety, suggests that simply having workers and commuters at least 2m apart on trains and offices will provide a significant level of protection. I do not think this is true.

Here is a sample office floorplan with social distancing, courtesy of WeWork

Here we have an example of an office with people all 2m apart and no open windows. The air in the room will slowly circulate and everyone in the room will receive a high level of exposure.

There are well documented examples of where this has happened. The restaurant below had airflow from right to left and over a 90-minute dinner, Diner A1 infects many people even at other tables, well beyond the 2m threshold.

This example from a call centre is even more compelling.

From only one infected employee over a single week, we see very high level of infection (note only 2 out of 94 infected were asymptomatic). The air circulates around the room over a long period and they are all exposed.

The issue of airflow distributing infection is certainly well known in the medical profession. The following diagram is from the Department of Health HTM 01-05 and shows how instruments and people should move in a sterile environment. But government are not recommending this concept being applied to other workplaces.

Simple advice on a 2m separation indoors is far from sufficient, but this has been overwhelmingly the focus of the government’s return to work advice. If a simple 2m rule indoors is not enough, then we need to better understand what matters and how to make indoor spaces safer. Simply opening public transport, offices, restaurants, and cinemas with social distancing is not going to work.

I would like to find research which gives parameters to turn the risk factors above into an operational risk model.

For example

Normal office

20 people, 140 sqm, speaking, low ventilation, 8 hours = what risk level?

Socially distanced office:

10 people, 140 sqm, speaking, low ventilation, 8 hours = what risk level?

How about a squash court?

2 people, 60 sqm, heavy breathing, low ventilation, I hour = what risk level?

I imagine that the relationships are non-linear and non-intuitive which means I cannot say how much safer a socially distanced office is from a normal one. Nor can I be confident if they are safer or worse than an hour of squash – one could easily be an order of magnitude more dangerous than the other. I need a lot more information before I can make good risk decisions.

Next Questions

Even though our government has given up we cannot.

  • How far do particles travel and how long do they stay in the air? https://www.youtube.com/watch?time_continue=39&v=WZSKoNGTR6Q&feature=emb_logo
  • Size of particles: amount of virus, travel, persistence, types of breathing
  • If someone else has been in a room for an hour how long should I wait before I enter it?
  • Does aircon and airflow make it worse by distributing the particles?
  • How much ventilation is required to make indoor risk more similar to outdoor risk?
  • Do plexiglass or masks actually help much?
  • What level of risk are we aiming for in workspaces and why?
    Medical practice uses the concept of Universal Precaution, but I cannot see that we can all wear PPE

What is the Mortality Rate? Update 2

Given the recently announced antibody tests from NY, I wanted to take a closer look at my model and compare the results.  It is important to be cautious about overinterpreting preliminary results like this; it is using a new test, for a new disease, on a sample for which we do not know how well it represents the population, but it is also the best test we have so far.

In my simple model for daily death, there are 2 inputs: speed of transmission, R0, and mortality rate.  Given that we only know the output, we cannot fully determine the value of both inputs, however to fit the empirical evidence it must be that if the mortality rate is lower then the speed of transmission must be higher and vice versa.  In other words, each mortality rate has an associated path of transmission and thus a path which determines the total number of infections/deaths.

The model can tell us which combinations of transmission and mortality are possible but not what either of them is.  To do that we need to look at other data.  In my first post on mortality, I focused on studies that estimated mortality rate directly, from which we then estimate the transmission rate.  The other approach is to focus on the number of infections to give a rate of transmission and then see what that implies for mortality rate.

When looking directly at studies on mortality, I concluded that 1% was my best estimate.  The New York test is the other approach, looking at the number of infections.  We can now look at what a completely different data set tells us.  I have drawn 3 paths for NY infections below, all of which can fit the basic data we have seen on death rates.

Picture 1

There is clearly a non-linear relationship between % mortality and % infections in the population.  If I take 1% as the base case, then a rate of 0.5% would require a far higher infection rate while a rate of 1.5% instead would not change the number of infected people nearly as much.

My model predicts a mortality rate of 1% and current deaths data, then the population infection rate would be 12% of the NY State population.  This is remarkably close to the estimate from the NY test result of just under 14%.

The NY test could imply mortality rate a little below 1%, but on the other hand we know there is material underreporting of COVID deaths and so, one could equally argue that the mortality rate is a little over 1%, consistent with the same data.  Some media outlets are using the results to suggest the mortality rate is 0.5%, a poor estimate and a bad way to use statistics.  This ignores the important lag between infections and death as well as the underreporting of deaths.

My initial estimate of the mortality rate was around 1% which is confirmed by this test.  This makes all the projections for future waves of infection more likely to be ones we should be confident about.

Is this a surprise?

To me, no.  It is very much in the range of where medical expert consensus had it beforehand.  I do not think I have been proposing a model any different from mainstream medical thought, but I have found it useful to be able to make my own projections and explain the implications.

However, looking at media reports and my Bloomberg inbox, I have only read people saying that this data shows that the mortality rate is lower than expected. Some are very excitedly saying is a “game changer” and “most of our clients think this is a potentially important development”.  If this reaction is common, then it shows that while the news is virtually only about COVID, very little of it is genuinely informative and people’s understanding of the data is extremely poor.

One of the persistent misrepresentations of the data is that the reported case fatality rate is an estimate.  It is not.  It is a statistic.  It is simply confirmed deaths divided by confirmed cases.  Given the low levels of testing, we know that reported cases are a small fraction of actual infections.  Hence it may be used as an estimate, but it is a truly terrible estimate if you simply take the headline number.  Did anyone really think that mortality was “estimated” to be 13% in the UK for people who got infected?   Only if you thought this was an estimate would this “new” estimate of 1% seem lower to you.

Policy implications

The debate on deaths vs economic costs of reopening will continue, but I hope that so many pseudo-experts will now refrain from saying that COVID has a mortality rate close to flu.

Using the estimate of mortality rate from the NY data to imply infection rates in other countries, suggests that herd immunity is a very long way off.  The UK would have infection levels of just 4%.

PS Treatment news

Yesterday the S&P dropped 1.5% on news that Gilead’s drug which they touted as a treatment for COVID-19 had failed its first clinical trial.  It was a major positive news story a week ago, which tells us how people struggle to interpret news and statistics and the media struggles to provide sensible coverage.  As I mentioned on April 20th It is overwhelmingly likely that it is not trueUntil we have some proper testing, I will assume this is a nonsense story”

Similarly, please do not inject or drink disinfectant.  Killing a virus outside your body is not the same as killing it inside your body.  On the bright side, this was a Trump theory that was so bad even Fox News has not been pushing it.

What is the mortality rate? Quick data update

This data point seemed important to me and worth a short immediate note.   We have just had released the reasonably sized antibody test from an area of high infection.

The rate of infection in New York is estimated to be 14%.

This implies around 2.7m people in NY State have had the virus.

A little over 20,000 are reported to have died from COVID.

This gives a mortality rate of 0.8%

We know that the number of deaths is likely to be underestimated due to people dying at home or never being tested, so the actual mortality rate would be higher than this.

Perhaps 1% is still the best estimate.

This data confirms the studies from around the world and means that the models we have been working with look to have the correct parameters.  It makes the idea that the mortality rate is far lower even more implausible.