An argument against costly signalling in effective altruism
This is an expansion of a comment that I wrote on James Ozden’s EA Forum post The dangers of high salaries within EA organisations.
Ozden effectively argues as follows:
Staff at effective altruist organisations need to be value-aligned in order to make the right kinds of decisions; e.g. regarding what research to prioritise and whom to hire.
People who are willing to accept a lower salary are much more likely to be value-aligned.*
This is a pro tanto reason for lower salaries.
Thus, Ozden effectively argues that willingness to accept a lower salary is a costly signal of general value-alignment. Recently, there have been multiple similar discussions about costly signalling in effective altruism.
I think that a problem for this argument is that willingness to accept a lower salary isn’t as strong evidence of value-alignment as it may seem. I think there are non-trivial numbers of highly committed effective altruists - who would make very careful decisions regarding what research questions to prioritise and tackle, and who would be very careful about hiring decisions - who would not be willing to work for a lower salary. Conversely, I think there are many people - e.g. people from the larger non-profit or do-gooding world - who would be willing to work for a lower salary, but who wouldn't be very committed to effective altruist principles. Relatedly, in research on the general population led by Lucius Caviola, we found that inclinations towards different aspects of effective altruism (what we called “expansive altruism” and “effectiveness-focus”) came apart (though admittedly it’s not straightforward how this generalises to applicants to effective altruist jobs). In my view, these different observations suggest that willingness to work for a lower salary need not be as useful a costly signal as it may seem.
Costly signalling can also be used as an argument for a vegan diet. For instance, in a recent discussion Onni Aarne argues that consuming factory farmed animal products indicates “moral unseriousness”. That may suggest that effective altruist applicants who are not vegans are less likely to make good prioritisation or hiring decisions - and that veganism is therefore a useful costly signal of value-alignment. But I would question that - my guess is that it’s no better a signal than willingness to accept a lower salary.
In short, I don’t think there is a unidimensional “value-alignment” trait that you can observe from willingness to accept a lower salary or a vegan diet. Instead, correlations between different traits that may be useful for effective altruists are likely much weaker. That means that it’s probably better to try to gauge, e.g. potential applicants’ tendencies to make impartial prioritisation or hiring decisions (and other important traits) directly. This is hard, granted, but it’s more informative than trying to infer these tendencies indirectly, via looking at very different traits.
However, this argument doesn’t conclusively show that effective altruists shouldn’t engage in costly signalling. It is just a pro tanto argument, which must be weighed against other potential arguments. In particular, I haven’t discussed how salary and diet may affect effective altruists’ reputation outside of the community. I may return to that topic at a later point.
* Note that Ozden isn’t arguing that effective altruist salaries should be very low in absolute terms. He uses the term “moderate”.