Assumptions:
A. Many EA positions have a large surplus of people applying to them (eg. research)
B. Everyone applying for EA jobs aims to do good.
C. Most people in high earning professions will not be earning to give.
Here I take EA jobs to mean any job directly contributing to a certain cause.
Take person 1 and 2, where person 1 is marginally more skilled than person 2 at everything.
So person 1 is the job applicant who gets hired, and person 2 is the next best option.
When working directly for EA organizations, the added value of person 1 working there compared to person 2 would be fractional.
When working in a high paying profession and aiming to earn to give, person 1 will donate much more money to good causes than person 2, who according to assumption C would be a non EA applicant.
Following this, for areas with many applications, would earning to give not be the more advantageous option?
For contributing directly:
For earning to give:
So a direct contribution is worth it if:
When taking the median value from the 2018 80k hours survey on how much a new hire is worth in donations it comes out to $1,000,000 per year.
Assuming a 5% skill gap that would make the Added direct value:
When looking at rough guesstimate salaries for earning to give, along with a significant dose of eyeballing, it seems that this value can quite easily be surpassed.
Discussion:
It seems to me then, that the place where direct contribution can be beneficial would be those functions where assumption A does not hold. And at that point, it would follow the normal advice for finding a career as can be found on 80k.
From the 2019 survey this would be mostly operational/ management positions. Oddly enough from the same survey it also shows that the EA community would need more researchers, which doesn't jive with the perception of a PhD excess.
Assumption B seems either guaranteed trough the structure of the job, or trough the application process.
The counter argument here is that it frees that person up to do some other important job, but while assumption A holds, this wouldn't be significant.
Assumption C is mostly based on anecdotal observation but seems to hold true.
A lot of the numbers used in my findings are anecdotal rather than extensively researched.
My personal shortcomings:
I admit that I’m quite ignorant to the inner workings of the research world, and to how to write philosophical argumentation.
Along with that there is quite a spread amongst the values from 80k's research that due to my lack of skill in working with uncertainties I did not account for.
Please argue with my points regardless of the sloppy style that they are delivered in.
Advice on writing is also appreciated but possibly better delivered in a method other than a reply.
I aim to improve my overall skills and abilities, and if there is something here that seems particularly faulty and could use more work, I would be happy to hear it!
Sources:
80k's research into donation value compared to direct work (2018 is the last version I found the relevant table):
2019 EA leaders survey
It looks like you're comparing a situation where an EA applies to an EA organization (competing against other EAs) to a situation where the EA applies for earning to give, competing against non-EAs. You argue that the counterfactual difference is larger if the EA gets the high-earning job instead of a non-EA because for the EA role, the next-best candidate would also do something impactful if they get the role.
This is true when you look at it very narrowly (only look at the impact difference for that one specific job that people applied to, their first job application). However, consider what happens in each case after the other person gets rejected. The non-EA who gets rejected for the high-earning job will do something else where they presumably won't have an outsized impact, either. By contrast, the other EA person who also applied to the direct work role will likely continue to apply to impactful roles. (They might even consider earning to give as a fallback option.)
So, once you consider further effects (second job applications, etc.), it becomes clear that the consideration you highlight loses most of its relevance. (It only applies to the degree that you getting the EA job slows down other EAs' career trajectories or adds some chance that they give up on impactful roles altogether, being discouraged.)
0.05⋅1000000=50000See also this article.
I could imagine that the organizations here were asked to compare the person they actually hired to the next-best candidate. So, there's probably no discounting – the impact is estimated to be 1 million in donation/grantmaking equivalents.
The reason the values can be so high is because earning to give is only impactful if there are shovel-ready interventions. To get shovel-ready interventions, you need people doing direct work. To convert money into direct work, you need more direct work (e.g., grantmakers or headhunters or senior staff running hiring rounds and doing onboarding). In a funding landscape where organizations never have to neglect core priorities in order to fundraise, it isn't easy to replace direct work with money. Eventually, there have to be enough people to do all that direct work.