The latter seems substantially better than the former by my lights (well, substituting âacross the boardâ for âlet the market set pricesâ.)
The standard econ-101 story for this is (in caricature) that markets tend to efficiently allocate scarce resources, and you generally make things worse overall if you try to meddle in them (although you can anoint particular beneficiaries).
The mix of strategies to soft-suppress (i.e. short of frank collusion/âoligospony) salaries below market rate will probably be worse than not doing soâthe usual predictions are a labour supply shortfall, with the most able applicants preferentially selecting themselves out (if I know Iâll only realistically get $X at an EA org but I can get $1.5X in the private sector, thatâs a disincentive, and one that costs the EA org if they value my marginal labour more than X), and probably misallocation issues (bidding up wages gives a mechanism for the highest performers to match into the highest performing orgs).
Itâs also worth stressing the âHave a maximum, but ask the applicant to make a first suggestion; donât disclose wages; discourage employees sharing their salary with other employeesâ isnât an EA innovationâthey are pretty standard practice in salary negotiation on the employer side, and they conspire to undermine employee bargaining position. Canny employees being confronted with âhow much do you need?â may play along with the charade (âI need $10k more for leisure and holidays which areâpromise! - strictly necessary to ensure my peak productivity!â) or roll their eyes at the conceit (âSo I need $X in the sense you need to offer $X or I wonât take the jobâ).
Moreover:
âPaying by necessityâ probably gets into legal trouble in various jurisdictions. Paying Alice more than (similarly situated) Bob because (e.g.) she has kids is unlikely to fly. (More generally, the perverse incentives on taking âpay by necessityâ at face value are left as an exercise to the reader).
Heavily obfuscating compensation disadvantages poorer/âless experienced/âless willing negotiators. I think I remember some data suggesting there are demographic trends in these factorsâinsofar as it does, it seems likely to lead to unjust bias in compensation.
Typical sentiment is that employees rather than employers are the weaker party, at greater risk of exploitative or coercive practice. I donât understand why in EA contexts we are eager to endorse approaches that systematically benefit the latter at the expense of the former.
Not trying to push down the ceiling doesnât mean you have to elevate the floor. People can still offer their services âat a discountâ if they want to. Although this still a bad idea, one could always pay at market and hope employees give their âunnecessaryâ money back to you.
Iâm a big fan of having some separation between personal and professional life (and I think a lot of EA errs in eliding the two too much). Insofar as these arenât identicalâinsofar as âGreg the humanâ isnât âGreg the agent of his EA employers willâ, interests of (EA) employer and (EA) employee wonât perfectly converge: my holiday to Rome or whatever isnât a âbusiness expenseâ; the most marginal activity of my org isnât likely to be the thing that I consider the best use of my funds. Better to accept this (and strike mutually beneficial deals) rather than pretending otherwise.
Iâd add that I think thereâs something to be said in favor of a needs-based model in the early stages of a startup. For as long as youâre heavily funding-constrained, it allows you to hire a greater number of people at a given cost. (This is essentially first-degree price discrimination; maximizing producerâs surplus (â altruistic utility) can IMO be a good idea under some circumstances.) One could argue that even then, promising EA startups should (and will) be paid better, but Iâm not sure this always works out in practice.
The latter seems substantially better than the former by my lights (well, substituting âacross the boardâ for âlet the market set pricesâ.)
The standard econ-101 story for this is (in caricature) that markets tend to efficiently allocate scarce resources, and you generally make things worse overall if you try to meddle in them (although you can anoint particular beneficiaries).
The mix of strategies to soft-suppress (i.e. short of frank collusion/âoligospony) salaries below market rate will probably be worse than not doing soâthe usual predictions are a labour supply shortfall, with the most able applicants preferentially selecting themselves out (if I know Iâll only realistically get $X at an EA org but I can get $1.5X in the private sector, thatâs a disincentive, and one that costs the EA org if they value my marginal labour more than X), and probably misallocation issues (bidding up wages gives a mechanism for the highest performers to match into the highest performing orgs).
Itâs also worth stressing the âHave a maximum, but ask the applicant to make a first suggestion; donât disclose wages; discourage employees sharing their salary with other employeesâ isnât an EA innovationâthey are pretty standard practice in salary negotiation on the employer side, and they conspire to undermine employee bargaining position. Canny employees being confronted with âhow much do you need?â may play along with the charade (âI need $10k more for leisure and holidays which areâpromise! - strictly necessary to ensure my peak productivity!â) or roll their eyes at the conceit (âSo I need $X in the sense you need to offer $X or I wonât take the jobâ).
Moreover:
âPaying by necessityâ probably gets into legal trouble in various jurisdictions. Paying Alice more than (similarly situated) Bob because (e.g.) she has kids is unlikely to fly. (More generally, the perverse incentives on taking âpay by necessityâ at face value are left as an exercise to the reader).
Heavily obfuscating compensation disadvantages poorer/âless experienced/âless willing negotiators. I think I remember some data suggesting there are demographic trends in these factorsâinsofar as it does, it seems likely to lead to unjust bias in compensation.
Typical sentiment is that employees rather than employers are the weaker party, at greater risk of exploitative or coercive practice. I donât understand why in EA contexts we are eager to endorse approaches that systematically benefit the latter at the expense of the former.
Not trying to push down the ceiling doesnât mean you have to elevate the floor. People can still offer their services âat a discountâ if they want to. Although this still a bad idea, one could always pay at market and hope employees give their âunnecessaryâ money back to you.
Iâm a big fan of having some separation between personal and professional life (and I think a lot of EA errs in eliding the two too much). Insofar as these arenât identicalâinsofar as âGreg the humanâ isnât âGreg the agent of his EA employers willâ, interests of (EA) employer and (EA) employee wonât perfectly converge: my holiday to Rome or whatever isnât a âbusiness expenseâ; the most marginal activity of my org isnât likely to be the thing that I consider the best use of my funds. Better to accept this (and strike mutually beneficial deals) rather than pretending otherwise.
Iâd add that I think thereâs something to be said in favor of a needs-based model in the early stages of a startup. For as long as youâre heavily funding-constrained, it allows you to hire a greater number of people at a given cost. (This is essentially first-degree price discrimination; maximizing producerâs surplus (â altruistic utility) can IMO be a good idea under some circumstances.) One could argue that even then, promising EA startups should (and will) be paid better, but Iâm not sure this always works out in practice.
Other than that, I agree.
Well put Gregory, you nicely captured a lot of concerns I have about the âpay by (reported) necessityâ model.