Enclosed are the latest forecast and infection/fatality data from our real-time COVID-19 tracking systems. We've also included some new analytics, such as benchmarking the change in restaurant bookings around the world to proxy for the intensity of the containment strategy. If you have not read our methodology paper on this forecasting model, you can find it here.
Key results as at 29 March 2020:
- The peak in America's new infection numbers will be realised between April 9 and April 18 assuming it is 50% as effective at containment as South Korea and 75% as effective as Italy. In a more gloomy scenario where the US is only half as effective at containment as Italy, the peak new infection numbers materialise around April 29 (see first screenshot). For the time being, America's fatality rate is growing at a slower pace than the UK, the major European nations, and China's original path (see third screenshot);
Australia's peak new infection numbers are realised between April 4 and April 11 assuming it is 50% as effective at containment as South Korea and 75% as effective as Italy (see second screenshot). If Australia is only half as good as Italy at containment, the peak in new infections is observed around April 20. Interestingly, Australia's infection rate (fourth screenshot) and, in particular, its fatality rate (third screenshot) are tracking better than North Atlantic nations with Australia's fatality rate sitting somewhere between South Korea and Japan.
- In key countries around the world, the change in restaurant bookings is broadly similar, although Australia has noticeably lagged, presumably because of slower convergence to hard containment. This does not, however, appear to have impacted Australia's relative infection/fatality performance as yet, which is slower than most global peers. (The bookings data is sourced from Open Table.)
Of course we see declines as populations are controlled. Is that really worth focusing on given its so obvious? The issue is that you can't take your foot off the shutdown given the virus is so transmissible, it will just come back in successive waves. The only thing that saves this is a cure. A vaccine. Which makes the only metrics worth tracking are the human trials in progress or soon to be in progress.
It is worth focusing, given that (rightly or wrongly) markets will declare victory and buy up. If there is a second wave, then the market will take another dip. The question is whether the government will be strong enough to quarantine overseas arrivals until a vaccine is developed and the population is injected (1.5 years away).
Thanks once again for sharing your tracking analysis; irrespective of the investing insights, I find comfort in seeing detailed data and projections.
I agree Marios. That is what the market is focused on! Possibly I am thinking two-three steps ahead here.
I am personally running a simple global model which has an Easter peak for daily deaths - this relies on new daily (actual, not reported) cases having just peaked, and global carriers peaking later this week (currently could be at around 11m and >25K in Australia). Assumption that lockdowns are successful is critical to all these models. The reported cases data is not great, cross-border, as testing differs, so death data is the most reliable (though also has classification and co-morbidity issues) . What happens on re-opening is a different question entirely, and China will be the first case study. Progress on medicine can also change the game entirely.
I think the figures re april peak are optimistic. Do you mean peaks in %'s or peaks in deaths. Even peaks in %'s, is not a true peak as the death count or the case count will continue to rise, far beyond April...Also not to mention the extreme lack of testing in many developing nations like India and so on. Also not to mention, equities people will wait to see evidence that the virus doesnt reappear when they reopen, this virus already has caused so much angst, people wont be silly enough to buy until it is proven. Look at China - simply not able to reopen anything despite having peak of cases for nearly 2 months.
A good 9 minute explanation of the maths behind the Corona virus pandemic, which would be useful for anyone trying to gauge where we are in the cycle, can be found on Youtube. Here's a link: https://www.youtube.com/watch?v=Kas0tIxDvrg
Regarding Australian cases, have you considered the slowdown in infection from closing the borders, vs the potential up-tick in infection from community transmission? I’ve not seen community transmission only statistics. However if this number begins to grow at an exponential growth rate of, say 20%, then this will be A devastating blow for Australia. The slowing curve due to the strict quarantine could be a false dawn.
Everyone 's looking at case numbers, but seems to me we should be looking at the "active numbers". ie; case numbers less death and recovered. This gives a very different picture, and is more pertinent, in my view. Also the growth trajectory deaths graph is misleading as it doesn't take account of the population numbers. Should be looking at active cases per 1M population. Lies, dam lies and statistics.
The market might declare a temporary victory should there be a peak in the rate of new infections but it seems to be overlooking the fact that now the virus is at large in the community, only one of two things can happen for each of us - we are eventually infected (and recover or die) or we are vaccinated. One of those two things must now happen to each of us for this to be over. As long as we have finite capacity in our health system, containment policy must remain in effect until the community has either been vaccinated or has caught the virus and developed immunity or died. How long will it take to develop a vaccine and inoculate the entire community so they can return to work? And how long will it take until the market reaches the conclusion that the damage to GDP through people being isolated will continue long after the curve has been flattened? That gives me all the perspective I need on this recent bounce.
maybe too late to mention these things but I have an app that is publicly available which grabs data from NSW Health. My last view of it showed that there has been a bell curve drop in the internationally imported cases and its community transmission we have to worry about here - free app and analysis available here (check out the By Source menu option on the left): https://app.powerbi.com/view?r=eyJrIjoiMjlmN2Y4MmYtZTU3Mi...
for those interested in active case types you can see that in this global focused (and again, free) app I built to track this stuff (along with confirmed cases, recoveries and deaths): https://app.powerbi.com/view?r=eyJrIjoiNzhjMzdkMTktYjRjYi... When in the app, click on Growth - Multi on the left hand navigation column, then select the Active option in the chart page that comes up. The bar chart on the left allows you to select multiple countries (use CTRL + click on each bar), or you could filter from the top. You can then see the various country levels of active cases over time. The last menu option in the app "Days since X" currently allows you to baseline all countries since their 100th confirmed case. This is a type of chart that can help you see how the spread has compared across different regions. I will add a way to compare deaths since 10th case, recoveries since 100th and active cases since 100th. None of this is investment advice (obviously) but hopefully making the data more easily accessible like this can help anyone interested.
Who's the epidemiologist here?