7

In a related question the OP linked to this Knoema site to show that Spain has had no excess deaths in 2020 from Covid-19:

enter image description here

I'm skeptical because the site doesn't cite any source for their info.

Furthermore, one can easily find (alas also unsourced) "alternative facts" for mortality in Spain, which aren't monotonic in the previous years, e.g. countryeconomy.com:

enter image description here

Additionally EuroMOMO has this graph for excess deaths for Spain (alas in a big and ever-updated page)

enter image description here

There's no 2020 total I can easily get from EuroMOMO, but eyeballing it, it seems to me that integrating the large spike should be noticeable in contrast to previous years, i.e. the increase in 2020 vs 2019 should be more than 2019 vs 2018 etc.

Someone in comments pointed out to Spain's own MOMO, which no doubt is where from EuroMOMO gets their data, but I don't see an obvious total there for the year 2020 either. Spain's MOMO however does offer the raw data in CSV. Adding the defunciones_observadas field (which seems to be the best as it has a date reporting correction) for todos (both sexes) for 2020, I get 465,535, which is a higher number than Knoema reports (by approximately 32,400 deaths). Furthermore, if do the same exercise for 2019 (data is in the same CSV), I get 394,339 deaths for 2019, which is lower than either of the other sources (Knoema ~428K or countrydata ~417K) for that pre-pandemic year. I'm not 100% sure I'm adding the right field though.

So, is that 433.13K (from Knoema) the true number of deaths in Spain in 2020? The same question goes for 2019: is 428.44K the true number of deaths in Spain in 2019?

The Knoema difference between these years is 4.69K, whereas Spain's MOMO (if I'm using the right field from their dataset and I did not mess up the addition(s)) results in a difference between 2020 and 2019 of some 71,196 deaths, an order of magnitude more than Knoema reports.

Improve this question
15
  • 1
    the big spike equals to about 40 excess deaths? Doesn't seem like a lot and would suggest that the 433k number could be right.
    – user1721135
    Jan 18, 2021 at 0:26
  • 1
    @user1721135: that's a z-score. They don't give the raw numbers, as far as I can tell (at least not per country).
    – Fizz
    Jan 18, 2021 at 0:43
  • 1
    Because I'm not even convinced that the website (and hence, the claim) is notable. You, however, are skeptical about their claim and want to learn whether their number is valid… Just saying. A bit tongue in cheek: Would you accept an edit in which I change the line referring the lacking source to "I'm skeptical because they make their source only available after signing up, which I want others to do in order to answer my question"?
    – Schmuddi
    Jan 18, 2021 at 12:18
  • 1
    Today, January 18th, I would not be surprised if a reasonably legitimate site published data for 2020 based on estimates, not taking Covid into account at all.
    – gnasher729
    Jan 18, 2021 at 14:44
  • 1
    At this point it is impossible to know how many deaths Spain had in 2020. It takes more than 18 days (today in January 18) to find and count all the bodies. There may be estimates out there, but it will take at least two more months until we have something resembling a final count. Counting deaths is much more complicated than what one might imagine.
    – Stand with Gaza
    Jan 18, 2021 at 18:17

2 Answers 2

Reset to default
12

Thanks to Taladris' comment, the data source for Knoema's number is found on another page:

The 2019 Revision of World Population Prospects is the twenty-sixth round of official United Nations population estimates and projections that have been prepared by the Population Division of the Department of Economic and Social Affairs of the United Nations Secretariat.

LATEST UPDATE (3/26/2020)

So it seems these (433.13K deaths in 2020 and 428.44K in 2019) numbers are UN projections made in 2019.

I've tried to find the exact number in the UN's files and they turned out to be (both) estimated and interpolated. They have some kind of model that predicts/estimates them for groups of five years, e.g. for Spain I can find enter image description here 1,985K for the 2010-2015 and 2,106K for 2015-2020 respectively. And then that data is interpolated to annual data. In the screenshot below you can see Spain's interpolated numbers in the bottom right corner coinciding with the Knoema data

enter image description here

The (Excel) fields are formatted not to show decimals, but these are present in the UN data as 428.436 and 433.131 respectively for 2019 and 2020.

I'm not 100% sure of this since they marked everything as estimates in that file, but UN's own data sources page for Spain says

Spain

Population: Total population and distribution by age and sex estimated to be consistent with the population by age and sex of the (a) 1950, 1960, 1970, 1981, 1991, 2001, 2011 censuses; (b) official estimates through 2017; and with estimates of the subsequent trends in fertility, mortality and international migration.

[...]

Overall Mortality: Life expectancy at birth and age pattern of mortality based on: (a) official estimates through 2014; (b) registered deaths by age and sex available through 2016; (c) International estimates used up to 2017.

It's a bit unclear if the total mortality is inferred from censuses or by adding the age group data. In any case, none of these UN figures is anything but an estimate for Spain starting with 2017.

Improve this answer
1
  • 12
    Wow... so "Coronavirus didn't have an effect on mortality in Spain! For proof, check out these projected mortality numbers from before they knew about coronavirus!" Nice.
    – Ben Barden
    Jan 20, 2021 at 17:08
7

The raw data are available from the Spain MOMO site at the "Datos" tab. The link labeled "aquí" here will give you a CSV file of deaths per date, either National or by region, starting 2018-12-31.

If I have done the calculations correctly, the current total for all-cause deaths in 2020 is 465,535. I suggest the original question is based on incomplete data. (The CSV data will probably be revised, too, but not enough to change the big spike in deaths.)

Improve this answer
8
  • 11
    There is a disturbingly common habit among covid skeptics to quote incomplete data as evidence covid mortality is low which suggests both motivated reasoning and severe data illiteracy at the same time.
    – matt_black
    Jan 19, 2021 at 11:09
  • 1
    Careful analysts understand the data and compile consistent time series for excess mortality. Like the Financial Times (who have excess mortality in spain as >70k so far).
    – matt_black
    Jan 19, 2021 at 11:11
  • 1
    This doesn't add any substance that isn't already part of the question. Its more of a comment than an answer.
    – Caleb
    Jan 20, 2021 at 12:04
  • @Caleb That's because this is a case of a question with its own answer in it.
    – Rekesoft
    Jan 20, 2021 at 13:04
  • 1
    @Caleb The question was edited to incorporate this answer (without credit).
    – Andrew Lazarus
    Jan 20, 2021 at 16:37

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .