Monetising time-donations
Effective altruism is often framed in terms of monetary donations. In particular, discussions of effective altruist sacrifice are often framed in terms of monetary donations: of whether to donate 1% of one’s income, or 10%, or more.
In recent years, there’s been a shift in emphasis in effective altruism towards direct work: towards using one’s time, rather than one’s money, for altruistic purposes. One might think that that shouldn’t affect the way we think about altruistic sacrifice too much; that we could just replace “income” with “working hours” or “waking hours” and ask whether we should spend 1%, or 10%, or more, of them for altruistic purposes (e.g. by working for a charity).
However, in my view that analogy doesn’t work too well. It seems to assume that we should have two jobs - one which we pursue for self-interested reasons, and one which we pursue for altruistic reasons (this could be volunteer work). But most effective altruists think that that’s not an effective strategy, and instead only have one job, which they typically have both altruistic and self-interested reasons for pursuing (to varying degrees). Therefore, I think that thinking in terms of “donating 10% of one’s time” is usually not a helpful way of thinking.
Is there a better way of thinking of these issues in a movement largely focused on direct work? A standard approach in economics is to monetise all sorts of values or benefits: to turn them into money. You could, for instance, monetise some effective altruist’s work through asking what the maximum amount a rational and well-informed donor would in principle be willing to pay for it.* If the effective altruist has a lower salary, they are effectively thereby making a donation:**
Where:
I: Impact of direct work in dollars.
S: Salary in dollars.
D: Monetised donations in dollars
Thus, if someone’s annual impact is valued at $1m, and their salary is $100,000, they are effectively making a $900,000 donation.
We may also want to know how much someone who works directly donates in terms of fractions of their total monetised resources (including their human capital) - similarly to how someone may donate 10% of their income. One might think that one should just divide their monetised donations by their monetised impact:
In the example above, that would mean that they are donating 90% of their total monetised resources ($900,000/$1m).
But I think that’s not quite the relevant metric. We are not ultimately interested in what fraction of their actual monetised impact that they donate - which is what Formula 2 expresses - but rather in what fraction of their potential monetised impact that they donate. Our human capital is determined by the impact of the highest-impact job that we could have had, not by the impact of our actual job. And these often come apart, as people take a job that’s lower impact, e.g. because they find the highest-impact job lower status or less fun. Suppose, for instance, that our effective altruist could have taken a job with a $10m impact, but that they chose not to take it for self-interested reasons. In that case, they are reducing the impact they could have had with 90% for self-interested reasons. And our formula needs to reflect that. We thus shouldn’t have the monetised impact of their actual work in the denominator, but replace it by monetised impact of the most effective work they could have had.***
Where:
P: Impact of the most effective potential direct work in dollars.
In the example the effective altruist is thus donating 9% ($900,000/$10m) of their total monetised resources, not 90%.
We can define two different factors which together decide how much of their total monetised resources they donate:
The first factor is Formula 2: the fraction of their actual monetised impact that they donate. Call that the salary factor. The second factor is their actual job’s impact as a fraction of the impact that they would have had at the most impactful job that they could have taken. Call that the job selection factor. In this example, the salary factor is 90%, whereas the job selection factor is just 10%. Thus the salary factor barely reduces their monetised donations at all, whereas the job selection factor reduces them by an order of magnitude.
Of course, these numbers are just hypothetical, but I think that they are actually not unrealistic.**** I think that some effective altruist jobs have a very high impact, and that they plausibly are paid nowhere close to their monetised impact. Moreover, I think that people often choose lower-impact jobs for self-interested reasons. Thus, I think that very often, the job selection factor reduces effective altruists’ monetised donations more - and often substantially more - than the salary factor. To the extent that that’s true, we should focus relatively more on people choosing the highest-impact job they could have chosen, and relatively less on them being frugal and earning a low salary. My sense is that if we can make people take higher-impact jobs by paying higher salaries for such jobs, we should often do so.
Lastly, let me make a clarification regarding the concepts of sacrifice and demandingness. Here, I’ve focused on what fractions of their monetised resources that people who work directly donate. But that has at best a very weak connection to how demanding effective altruism is in psychological terms. For instance, suppose that someone actually works at their potential highest-impact job, and that its monetised impact is $10m. Moreover, suppose that they earn a salary of $100,000 and very much enjoy their job. They are then giving away 99% of their monetised resources, but at the same time, it’s intuitively not a very demanding job - as their personal level of well-being may even be higher than it would have been, had they taken a job on the regular labour market for non-altruistic purposes. Such examples have arguably become pretty common as the total amount of funding in effective altruism, and the total monetised effective altruist impact, have grown. Thus the issue of what fraction of their monetised resources that people who work directly donate and the issue of the psychological demandingness of their involvement in effective altruism should be kept distinct.
* How to determine this willingness to pay is quite tricky, but I won’t go into the details of that here.
**Any monetary donations that this effective altruist worker makes should of course also be taken into account. The below formulas effectively use post-donation salaries: the amount that they keep for themselves, after donations have been deducted.
*** It’s true that when effective altruists discuss how much money to donate (e.g. in the context of the Giving What We Can pledge), they usually talk about fractions of their actual income, and not fractions of their potential income. One might think that one, analogously, should talk about monetised time-donations in terms of the fraction of our actual monetised impact that we donate. But as discussed in this paragraph, I would reject that: fractions of potential monetised impact seems like a more relevant measure.
**** More empirical studies of these issues would be great.
A 2019 EAGx talk of mine discusses similar themes, but with a different framing.
Thanks to Ryan Carey and Daniel Eth for their insightful comments.