As 2021 comes to a close, I’m having a distinct sense of déjà vu. A year ago, I wrote a post celebrating COVID vaccines as a triumph of science. And I noted that vaccine discovery is a collective endeavour. The cumulative number of major vaccines tracks closely with the cumulative number of scientific papers, a bellwether for humanity’s collective knowledge. Here’s the trend:
This trend is a beautiful reminder that science is a social endeavour. No matter how novel, every discovery builds on the work of others. Science, in other words, is the quintessential public good. And so are vaccines.
Looking at the roll out of previous vaccines, however, I noted a perverse (but unsurprising) trend. Instead of chasing people, vaccines chase dollars. Pick any vaccine you want, and you’ll find that the corresponding vaccination rate tracks with GDP per capita. Sadly, the roll out of COVID vaccines is no exception. Here’s the latest data:
What I did not anticipate, back in December 2020, was that a significant portion of rich-country citizens would choose to remain unvaccinated. Although anti-vaxers are a source of frustration for those of us who want to get on with a post-pandemic life, there is an upshot. It seems plausible that by avoiding vaccines, rich anti-vaxers are freeing up some supply for poor countries. In a year like 2021, you have to take whatever solace you can get.
Support this blog
Economics from the Top Down is where I share my ideas for how to create a better economics. If you liked this post, consider becoming a patron. You’ll help me continue my research, and continue to share it with readers like you.
Stay updated
Sign up to get email updates from this blog.
This work is licensed under a Creative Commons Attribution 4.0 License. You can use/share it anyway you want, provided you attribute it to me (Blair Fix) and link to Economics from the Top Down.
Sources and methods
Data for the number of scientific papers comes from Sci-Hub, via Library Genesis. The raw data comes as an SQL database dump. If you’re interested, I built an R function that can parse this data. Check it out at Github.
Data for new vaccine dates is from Wikipedia.
Data for COVID vaccination rates is from Our World in Data. GDP per capita data is from the World Bank. For both sources, I’ve used the most recent data (in each country) that is available.
Blair, why do you use GDP in the second graph?
《The time has come to discard ‘real’ GDP and to elevate new measures that actually address important social problems.》
https://economicsfromthetopdown.com/2019/12/15/why-we-should-abandon-real-gdp-as-a-measure-of-economic-activity/
Are you saying GDP measures wealth accurately?
I used nominal GDP here, which is a decent measure of relative income. What I refrain from using is ‘real’ GDP.
If you used real GDP, would the result be very similar?
From your earlier blog, linked in my previous comment:
《For all that it purports to say, Gross Domestic Product (GDP) fails to explain anything relevant about the world.》
《When there is uncertainty in a measurement, the appropriate response is to report it openly. 》
Why are standard errors for each of the surveys that go into the GDP (nominal or real) calculation simply thrown away?
《Prices are a tool for distributing resources. The proper place for prices, then, is for understanding economic distribution.》
Why am I struggling to reconcile this with nominal prices used in nominal GDP calculation? Are the nominal prices closer to true value, as you see it?
《Likewise, we can compare the income of individuals or groups of individuals. The meaning is in the comparison, not the aggregate value itself.》
Is nominal GDP a measure of world power? But what does that have to do with me as an individual American who feels powerless to change anything? What is the uncertainty associated with the single dot labeled “USA” in the GDP graph?
Put another way, do you assume ergodicity in these graphs?
Hi RSM,
Thanks for the questions. Hopefuly my responses will clarify things.
Yes, they would be similar. When comparing GDP across countries in the same year, converting to ‘real’ GDP amounts to accounting for ‘purchasing power parity’. But since nominal GDP varies so much, variation in purchasing power (however you calculate it) doesn’t affect results all that much.
Sampling error for nominal GDP is often reported (at least in the US) and is typically trivially small. The systemic (or epistemic) error in ‘real’ GDP, however, is rarely if ever reported. Doing so would undermine confidence in the measure, which is a key part of our current ideology.
There is no such thing as ‘true value’. There are market prices … full stop. When I compare nominal GDP across countries, I am comparing the international distribution of income. This income says nothing about the scale of ‘production’ (whatever that means).
Nominal GDP (per capita) is a measure of relative income … assuming you compare it to GDP per capita in other countries. If you think that income is caused by power (which I do), then yes, GDP is a metric of power. But (crucially) it is an average measure. Since individual income varies so much, the average doesn’t say much about any one person.
I don’t really understand the question.
RSM has a point – looking forward to your answer
Blair you say in your admirably thorough reply below that nominal GDP reports trivially small standard errors, but how much do you know about their sampling methodology, imputations, data augmentation, adjustments, etc.? Reading Jacob Assa’s publications on GDP, am I wrong in ascribing to nominal GDP the same fundamental flaws to which the inflation adjustments of ‘real’ GDP add but icing to the cake, as it were? (Doesn’t Assa’s description of the subjectivity and neoclassical model assumptions that go into imputing the income-equivalent of owner-occupied housing, for one example, apply equally to nominal and ‘real’ GDP?) Does the latter simply add yet another layer of subjectivity to an already wildly subjective figure?
Can you see how I believe nominal GDP, too, should have uncertainty around it that should undermine your confidence in using it for policy purposes? For example, aren’t financial flows and income mostly ignored by even nominal GDP? How much financial income is left unreported and unrepresented in your graph? (May I repeat another commenter’s recommendation of Perry Mehrling’s work on money flows?)
When you say you are comparing the international distribution of income, how much faith do you have in the measurements? Why do gross financial flows dwarf economic transactions measured by GDP? For example, are trillions of dollars exchanged each day in repo markets producing income for traders that mostly won’t show up in GDP, nominal or real? If they hold their income in financial assets that they can convert to dollars only if and when they want, don’t they hold a lot of power that GDP does not even measure?
《Since individual income varies so much, the average doesn’t say much about any one person.》
So what would you like me to take away from the graph? What are you really saying? What if you relabeled GDP per capita “Power per capita”? (But aren’t you still leaving financial markets out of your measure of power?)
Should the ~$60000 per year for USA GDP per capita have a lower error margin of at least $50,000?
As for ergodicity, may I recommend Lars Syll’s blog?
《if one takes the number of crimes committed by black people in a certain day divided by the total number of black people, and then follows one random-picked black individual over his life, one would not find that, e.g. each month, this individual commits crimes at the same rate as the crime rate determined over the entire ensemble. Thus, one cannot use ensemble statistics to properly infer what is and what is not probable that a certain individual will do.》
In other words, what have your graphs got to do with my individual life, and my choices?
Thanks for the detailed response. About the ‘uncertainty’ in nominal GDP, we have to be careful about the terminology that we use. When I said that ‘uncertainty’ was trivially small, I meant the ‘sampling’ uncertainty that arises from inferring something about the population from a smaller simple. This uncertainty is usually trivially small.
Then there is ‘systemic’ uncertainty … or perhaps more appropriately called ‘epistemic’ uncertainty. This epistemic uncertainty stems from the various ways we define nominal GDP in the first place. As Jacob Assa correctly argues, what does and does not get included in nominal GDP is largely arbitrary, and inherently political.
The same goes with any definition of ‘income’. Are capital gains ‘income’? In my view, yes. In terms of the national accounts, no. Steve Roth is doing interesting work on this front, trying to untangle the various forms of income.
When we move to international data, there are loads of problems, as different countries define nominal GDP slightly differently. As to you question about how much faith I have in these measures, I think that it depends on the question you’re asking. Is the World Bank data the correct measure of income? No. Is it a measure. Yes, and it is useful for comparing income in an order of magnitude sense.
For instance, I suspect that regardless of how you defined and measured income, you’d find a strong correlation between income per capita and the vaccination rate.
As far as what you should take away from my graph, I think I made that clear in the post. Vaccines chase dollars (roughly construed).
About ergodicity (as described by Syll), I think this is better described as ‘central tendency’. Does income per capita describe a tight central tendency for citizens of a country? No … income inequality is usually so large that the average income doesn’t tell us much about any single individual.
Aggregate analysis is rarely meant to say much about individuals. But to answer your question, the data says that if you live in a rich country, you have a far higher probability of being vaccinated than if you lived in a poor country.
Blair, thanks for your answers, but may I still ask: what relevance does “vaccines chase dollars” have to an unvaccinated individual in the US making a personal, not an economic, decision?
Perhaps I can spin a tale in which rich countries have more tax money to give private vaccine manufacturers, but whose taxes have been raised to pay for vaccines?
Does US power really lie in the ability to create dollars? Can Treasury sell new bonds, often at higher prices, faster than it pays off maturing bonds, with plenty left over to pay for vaccines?
May I advance that US power lies in its reserve currency status, and the Fed put (an implicit promise to markets that it will draw on its unlimited store of liquidity to act as a value buyer)? Can I spin a tale where GDP whether ‘real’ or nominal is too arbitrary, so shouldn’t we look at money markets to find the real power relationships?
I’m not sure how to respond to this question, as statistics can’t tell us much about individual decisions.
Yes, capitalization is probably a better indicator of power than GDP. But that sort of data is far more difficult to get (especially for poor countries) … hence the quick and dirty analysis here using World Bank data for GDP (which is really easy to get).