Covid-19 Analysis – updated

  • 0
  • March 27, 2020

FOR INVESTMENT PROFESSIONALS ONLY – NOT TO BE RELIED UPON BY RETAIL INVESTORs

This analysis is the intellectual property of P1 Investment Management Limited and is not to be copied or reproduced without written permission. The information contained herein is intended to be an indicator of the likely UK infection rates. It does not draw any conclusions as to how the markets will behave, nor does it look at wider, global incidents of the pandemic and risks of secondary infections, the risks of which are, at present, unclear. P1 Investment Management will update this research periodically in line with emerging data.


5th May 2020

Dr Quintin Rayer, Head of Research, P1 Investment Management

P1 Investment Management has applied our modelling and quantitative experience to the UK’s response to covid-19 pandemic as a means to aid our investment planning and processes. We are not epidemiologists nor medical experts, but by analysing the publicly available data, we aim to provide insight into the possible outcomes of the UK’s response. We are sharing our research to aid professionals and our industry in making decisions in these uncertain times.

To be clear:

  • We are not medical nor epidemiological experts; this is not medical research.
  • This research is an analytical assessment of publicly available data to improve our understanding of the pandemic and aid us in our investment process.
  • This research is not a tool for timing markets, but research to understand the epidemic and possible evolution in the UK and plan accordingly.

Summary Assessment – Covid-19

  • Based on different measures (dates when number of daily cases exceeded 250 and 500, or when total cases exceeded 1,000), the UK would appear to be approximately: 49-53 days behind China, 13-17 days behind Italy, 5-9 days behind Spain and 3-5 days behind the USA.
  • Based on the best fit data and the progression of events in China, an estimated date for the partial lifting of restrictions in the UK would be 13th June.
  • Less robust estimates based on outbreak dates in China suggest earlier dates of 27-31 May. The next three-week review data for the UK government’s restrictions would be expected to be 7th May.
  • The peak in UK daily new cases might have been expected to be around 15-19 April based on Chinese data, 3-13 April based on Italy; or tentatively 7-26 April based on Spain. It is unclear whether daily cases have peaked in the USA.
  • The peak in UK daily new cases may be passed, covering the period around 5-19 April, which appears to be supported by the above estimates from other countries. If this is the case, the UK peak has been longer than that of China and Italy, but shorter than that of Spain and the USA.
  • The UK experience appears to be more closely following that of Italy or one-fifth of the USA new daily cases (“USA5”, see below), and less like Spain or the USA.  Recent UK experience appears to be deviating from that of China as the UK experiences a more extended peak.

Best Fit Daily Incidence Charts

When comparing the UK with the USA, an obvious point of difference is the relative population of the USA (327.2 million) compared with the UK (66.4 million). The population of the USA is approximately five times that of the UK. A simple adjustment to allow for this might be to divide the number of daily new cases in the USA by five, for the comparison. Interestingly, the same approach does not appear relevant to the UK-China comparison (the population of China is 1,409 million), where the COVID-19 outbreak was largely contained within Hubei province (capital Wuhan) which has a population of 58.5 million, and is thus more directly comparable with the UK.

Presenting the data on a log scale, helps comparison between the USA and other countries.  

Method

Defining the epidemic start dates is not straightforward. For example, the first data we have for China comprises 259 cases (23/01/2020). In comparison, the first data for Italy comprises 1 case (20/02/2020), and the first data for the UK is 4 cases (23/02/2020). It is thus clear that these are not equivalent comparisons. 

We took the daily data and determined the best fit by shifting the dates relative to the UK. We were only able to do this for the period where UK data overlaps that of other counties. The resulting best fit chart is shown above. 

The best-fit approach, using the available data, helps allow for lack of initial unreported data in China (or other nations), as it seeks to match the significant trends over the full range of UK data available. Although in this case, only the current growth in UK daily cases can be used, which may limit the accuracy of the results.  

This approach should be a more reliable method, rather than focusing on when the first cases were reported, particularly considering the differences between the number of cases first reported in China (259) compared to the numbers of initial cases in Italy and the UK (1 and 4 respectively). Even between the Italy and UK early cases, direct comparisons for “start dates” are challenged by the rapid growth in Italian cases (on consecutive days 1, 17, 58, 78, 72…) compared with slower initial growth in the first UK cases (on consecutive days 4, 0, 0, 0, 3…).

Country Days before daily cases exceeded 250 (date) Days before daily cases exceeded 500 (date) Days before accumulated cases exceeded 1000 (date)
UK 0 (14/03/2020) 0 (18/03/2020) 0 (14/03/2020)
China 51 (23/01/2020) 53 (25/01/2020) 49 (25/01/2020)
Italy 13 (01/03/2020) 17 (01/03/2020) 14 (29/02/2020)
Spain 5 (09/03/2020) 9 (09/03/2020) 5 (09/03/2020)
USA 4 (10/03/2020) 5 (13/03/2020) 3 (11/03/2020)

Based on the above, from the start of the outbreak, it would seem reasonable to estimate that the UK is approximately  

  • 49-53 days behind China 
  • 13-17 days behind Italy 
  • 5-9 days behind Spain 
  • 3-5 days behind the USA 

When will covid-19 restrictions be likely to end?

The Chinese have partially lifted the lockdown on Wuhan on 8th April. Based on the best fit of the UK data to China this gives an estimated UK date based on similar response of 13th June. Italy has announced it will further relax restrictions on 4th May. Based on the best fit of the UK data to Italy this gives an estimated UK date based on similar response of 19th May.

However, following its five tests, the UK government may exercise greater caution than Italy. A reason for doing so is not hard to glean from perusal of the daily new case data between the two countries.

 

Country Lockdown partial relaxation date Number of daily new cases at lockdown partial relaxation date Average number of daily new cases in the 7 days prior to partial relaxation date (excluding partial relaxation date)
China 08/04/2020 63 35.4
Italy 04/05/2020 (Current figures to 04/05/2020 used right) 1389 1863

Italy is still experiencing an average 53-fold higher number of daily new cases over the preceding week than China. Motivations for partial relaxations of restrictions may be based on economic and governmental financial resilience, as well as epidemic control.

If the other epidemic start date estimates are used, as China is estimated to be 49-53 days ahead of the UK, this would suggest estimated UK dates for the partial lifting of restrictions of 27th May to 31st May. The reason that these apparently larger number of days give earlier dates is that the Chinese dataset effectively starts part way into the epidemic. Similar estimates based on Italy suggest estimated UK dates for the partial lifting of restrictions of 17-21 May. These results are summarised below.

Country Estimated partial relaxation of UK restrictions based on…
  Best fit data Alternative measure earliest estimate Alternative measure latest estimate Date range
China 13th June 27th May 31st May 27th May to 13th June
Italy 19th May 17th May 21st May 17th May to 21st May

Clearly the earlier dates for partial relaxation of restrictions would be expected to be associated with greater risk of a second outbreak, or lesser degrees of relaxation of controls. On 16th April the UK government indicated that restrictions would continue for at least three weeks (i.e. until 7th May).

When is the UK peak of the epidemic likely to be?

To estimate the peak in the epidemic, other countries’ periods of peak daily cases were considered, to get a sense of how far into each national epidemic the peak occurred. This required data from countries where the peak has already occurred. The country data considered is summarised below.

Country Has peak been reached? Dates of peak Period of peak (days) Equivalent UK dates
UK Possibly at, or just past, peak 05/04/2020 – 19/04/2020 [?] 15 [?] 05/04/2020 – 19/04/2020 [?]
China Yes 09/02/2020 – 13/02/2020 5 15/04/2020 – 19/04/2020
Italy Peak appears to have been passed 19/03/2020 – 29/03/2020 11 03/04/2020 – 13/04/2020
Spain Possibly peak appears to have been passed 23/03/2020 – 11/04/2020 20 07/04/2020 – 26/04/2020
USA Probably not 01/04/2020 and ongoing [?] 34+ 13/05/2020 to [?]

 

While the Chinese epidemic has clearly peaked, the large spike in cases on 12/02/2020 adds to difficulties in identifying the true peak period. The Chinese analysis is now based on an adjusted dataset. While this has addressed the large spike in Chinese daily new cases on 12/02/2020, it is likely to have introduced additional uncertainty to the Chinese figures in the run up to that period. On this basis, the peak in the Chinese epidemic is estimated to have occurred between 09/02/2020 – 13/02/2020. For the UK, these dates correspond to the period 10/04/2020 – 14/04/2020. The peak in the Chinese epidemic lasted 5 days.

The peak of the Italian epidemic appears to have been reached between dates 19/03/2020 – 29/03/2020. For the UK these dates correspond to the period 03/04/2020 – 13/04/2020.  The similar data for Spain and the USA is presented in the table above.

In principle, apart from challenges dealing with the spike in cases on 12/02/2020 for China, the Chinese and Italian data should be more reliable for estimating the peak in new cases than the Spanish data, since it is less clear that daily cases in those nations have really peaked.  

It would suggest that the peak numbers of daily cases in the UK would be likely to be between

  • 15th – 19th April based on China.
  • 3rd – 13th April based on Italy.
  • Tentatively, 7st – 26th April based on Spain.

At the time of writing, UK data is available up to 26th April. It is possible that the UK peak was around the period 5-19 April.

It is also worth noting the range of periods for which the epidemic peaks have lasted in different nations. Five days in China, 11 days in Italy and 20 days for Spain. It seems likely that the strength of governmental policy response (social distancing, restrictions on interactions and so-called ‘lockdowns’) and the degree to which they are followed by citizens, may determine the length of the period during which peak daily cases occur. It also suggestions that caution will be required in interpreting when then peak period for new cases has ended.

Which Countries’ Outbreaks are likely to be most like the UK?  

The UK’s outbreak appears to be lagging between 3 and 53 days behind other nations and reports of alarming statistics and tragic cases are appearing. It is therefore natural that we would wish to know how closely the UK’s experience is following other nations.

The chart below shows the best-fit data expanded to focus on the current UK experience.

From the chart, the UK lies within a group of nations comprising Italy, Spain, and USA/5. The UK experience appears to differ from that of the unadjusted USA, which initially has fewer reported daily cases, but where numbers have recently significantly overtaken daily cases from China, Italy, and Spain at the same stage.  

Daily Deaths Data

Based on the best-fit lags from the daily new cases data, daily deaths are presented in the chart below.

Daily deaths in the UK would appear to be broadly following those of Italy and Spain.

Uncertainties

  • The data estimates will become better as more information rolls in daily.   
  • More refined data, particularly in the case of earlier Chinese cases would require a reworking of the analysis.  
  • Policy responses have not been the same in all countries, which will mean differences, and the UK future policy response may be different.  
  • Research may yield additional methods to manage covid-19 which will impact the data. 

Dr Quintin Rayer

Posted By Dr Quintin Rayer

Quintin is a Chartered Fellow of the Chartered Institute for Securities and Investments, a Chartered Wealth Manager and holds a Physics degree from Imperial College London and a Physics doctorate in atmospheric physics from Oxford University and is a Fellow of the Institute of Physics.