Godlak stata monkey

Share this post

Covid-19 vaccinations outcomes

godlak.substack.com

Covid-19 vaccinations outcomes

Focus on countries with the highest % of vaccinations (data source: covid.ourworldindata.org)

Godlak 🥨
Jul 11, 2021
Share
Share this post

Covid-19 vaccinations outcomes

godlak.substack.com

Currently, here are the countries with the highest figures for % of administered doses of covid-19 vaccine(s):

Image

Let do 3 things:

  1. represent graphically the curves of vaccinations, PCR, PCR+, hospitalizations and deaths for these countries where data is available for all variables

  2. represent graphically scatter and linear fit plots for %PCR+ and deaths with/from covid-19 against % of fully vaccinated people

  3. represent graphically the coefficients from regressions of %PCR and deaths with/from covid-19 against % of fully vaccinated people


Curves of vaccinations, PCR, PCR+, hospitalizations and deaths

Due to data availability, UAE & NLD are not displayed; due to data availability, some curves cannot be displayed

PCR up, PCR+ up, deaths down

PCR+ chaotic but relatively stable, deaths up !

PCR+ flat, PCR data stops in June (?), hospitalizations and deaths flat

PCR up, PCR+ not yet (?), deaths up !

PCR & PCR+ down, deaths down

PCR & PCR+ in tango, currently down; deaths up !

PCR & PCR+ in tango up, hospitalizations and deaths flat

PCR up, PCR+ down (interesting…), deaths down but picking up (?)

PCR & PCR+ tango down, hospitalizations & deaths down

PCR+ down, no PCR data, deaths data difficult to interpret…

PCR & PCR+ in tango up, hospitalizations & deaths down

Overall, it’s all over the place… Very different countries, high vaccinations %, very different outcomes…


Scatter and linear fit plots for %PCR+ and deaths with/from covid-19 against % of fully vaccinated people

Due to data availability, 2 countries are dropped from this analysis

Timspan from 1/1/2021 until latest available data is used

As above, all variables are 7 days moving averages; fully vaccinated people % is lagged by 14 days

%PCR+ is used to take into account the PCR intensity and the number of PCR+ in one metric

Weekly dummies are included in the regressions

Negative relationship between vaccinations and %PCR+ and deaths in Malta

Flat link between vaccinations and %PCR+ and a positive link between vaccinations and deaths in Chile

Negative link between vaccinations, %PCR+ and deaths in Uruguay

Positive link between vaccinations, %PCR+ and deaths in Mongolia

Flat link between vaccinations & %PCR+ and fitted negative one (but one can observe a flat one as well from dots) between vaccinations & deaths in UK

(timid) Negative link between vaccinations, %PCR+ and deaths in Qatar

Linear fits tend to show a negative link between vaccinations, %PCR+ & deaths in Canada, but looking at the dots one can see a rather flat link

Flat link between vaccinations, %PCR+ and deaths in Singapore

Flat link between vaccinations & %PCR+ in Denmark, also between vaccinations & deaths although linear fit show a negative line

Overall, things are again all over the place, confirming the initial graphical representations with various curves


Coefficients from regressions of %PCR and deaths with/from covid-19 against % of fully vaccinated people

Coefficients from robust regressions are show graphically below

All variables are 7 days moving averages; fully vaccinated people / 100 is lagged by 14 days

How to read this ? For instance, in the case of Malta (MLT iso code): the coefficient is close to 0 in both cases (w/r to %PCR+ or deaths) - coefficients values are on the Y axe - while, according to p-values, the coefficient for %PCR+ is significant while for deaths it is not (the closer the p-value is to 0, the more significant is the coefficient).

Overall, most of the coefficients are very small when they are significant.

Over the period under investigation, for instance in Qatar, an average increase in the 7 days moving average of fully vaccinated people of 1 point is associated with a reduction in the 7 days moving average of %PCR+ of less than 0.004, 14 days later.

Similarly, for instance in Urugay, an average increase in the 7 days moving average of fully vaccinated people of 1 point is assocaited with a reduction in the 7 days moving average of deaths (per million) of slightly more than 0.2, 14 days later.

Overall, the links between vaccinations and %PCR+ or deaths are statistically rather weak in those countries and over this time span.

Share
Share this post

Covid-19 vaccinations outcomes

godlak.substack.com
Previous
Next
Comments
Top
New
Community

No posts

Ready for more?

© 2023 Godlak 🥨
Privacy ∙ Terms ∙ Collection notice
Start WritingGet the app
Substack is the home for great writing