I’ve written a series of posts where I discuss remuneration/compensation and demandingness in effective altruism. Here I briefly summarise that series and explain how the different posts fit together.
My key claim is that remuneration in effective altruism should, on average, be substantial. That’s mainly because of incentive effects, which I expect to outweigh the financial costs (partly thanks to the improved funding situation in effective altruism). In part, my series can be seen as a response to the recent articles on the EA Forum that express worries about greater effective altruist spending and increased remuneration.
I think effective altruism has already been moving in the direction of higher average remuneration for the last couple of years, and I suggest that development should continue. To an extent, I think it represents a convergence with standards and norms on the regular labour market. Effective altruists are sensitive to monetary incentives just like other people—maybe more so than is sometimes acknowledged.
I discuss several counter-arguments; i.e. arguments for lower remuneration. I am relatively critical of the argument that effective altruists should use willingness to work for low remuneration as a costly signal of value-alignment. By contrast, I’m a bit more ambivalent about reputational arguments.
Besides substantial average remuneration I also argue for remuneration variance. Specifically, I argue that effective altruist funders should use monetary incentives to encourage people to take particularly impactful jobs. Given the likely large differences in impact between jobs, I expect that to be worth it.
I also discuss more general and conceptual issues in several posts. I argue that just as effective altruists should be neutral between different causes, so they should be neutral between the use of different resources (e.g. time vs money), as well as between different mindsets (e.g. a frugality mindset vs other mindsets). It seems to me that effective altruists aren’t always neutral in these senses, but that we sometimes cling on a bit nostalgically to the mindset and the approaches that the movement had at the start. Instead, I think we should be as open to changing our minds on these issues as we are regarding cause selection. I also show that to achieve resource neutrality, we can conceptualise use of our time in terms of its potential monetary value.
These are just some broad qualitative sketches, similar in style to the posts I respond to. I try to define the conceptual landscape and share some intuitions. Obviously this is not nearly as strong evidence as hard data would be. I think it could be valuable if some effective altruist researchers—e.g. researchers with an economics training—studied these issues in more detail. They could collect data on effective altruist remuneration and try to estimate effects of different remuneration levels on impact. Since effective altruism is growing, it seems increasingly plausible to have researchers dedicated to studying such issues. It also seems to me that effective altruists have devoted far less time to these issues than to cause prioritisation. Just as we do data-driven research on the cost-effectiveness of different causes, so we should do data-driven research on the cost-effectiveness of different remuneration levels. Such research could potentially also help make discussions about remuneration levels less emotional and more detachedly focused on impact.
The posts in chronological order:
Thanks to Pablo Stafforini, George Rosenfeld, and especially Ryan Carey and Daniel Eth for very helpful comments on this series of posts.