ONLY FOR PROF DAN

The Lessons of Effective Altruism Jennifer C. Rubenstein*

The Most Good You Can Do, Peter Singer (New Haven, Conn.: Yale University

Press, ),  pp., $ paper, $ cloth.

Doing Good Better, William MacAskill (New York: Avery, ),  pp., $

paper, $. cloth.

C arol Sue Snowden worked for thirty years as a librarian at the Columbus

Metropolitan Library in Columbus, Ohio. She led a quiet, frugal life,

spending money mostly on books, which were her passion. When she

died, she donated the money she had saved—over $ million—to the Columbus

library and seven local schools. Most of us would look upon this generosity

with admiration, but according to a new movement called Effective Altruism

(EA), Snowden got it wrong. While she was right to donate her money, she should

have instead directed it to an organization that does the most good overall.

Roughly three quarters of U.S. households donate to charity, representing a

total of almost $ billion in . Some of these donors are what we might

call “passionate philanthropists,” like Snowden: they make a deliberate choice to

donate a large amount of money to an issue or organization about which they

care deeply. Others are what social psychologists call “warm glow” givers. They

contribute smaller amounts of money to several different organizations in re-

sponse to dramatic images of suffering, requests from friends or relatives, and/or

to support causes that are close to their hearts. According to EA, both groups are

misguided. Individuals should donate to organizations that do the most good—

that is, organizations that maximally improve the welfare of sentient beings.

While most organizations that Effective Altruists think will do the most good ad-

dress social problems, such as severe poverty, some Effective Altruists support

*I thank Jeff Flynn, Iason Gabriel, Andrew Gates, Ted Lechterman, Jennifer Petersen, Allison Pugh, and Denise Walsh for very helpful comments on a previous version of this essay, and Brittany Leach for excellent research assistance.

Ethics & International Affairs, , no.  (), pp. –. ©  Carnegie Council for Ethics in International Affairs doi:./S

511

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0892679416000484 Downloaded from https://www.cambridge.org/core. University of Waikato Library, on 29 Nov 2017 at 03:36:30, subject to the Cambridge Core terms of use, available

more exotic causes, such as slowing human aging, figuring out how to blow up

asteroids headed toward Earth, and preventing the “robot apocalypse.” Other ac-

tivities endorsed by Effective Altruists include choosing a career path that will do

the most good, not eating meat (especially chicken), and donating a kidney.

Two recent books by philosophers—The Most Good You Can Do, by

Peter Singer, and Doing Good Better, by William MacAskill—aim to introduce

EA to an educated lay audience. Despite the more incremental title of

MacAskill’s book, both authors enthusiastically endorse EA’s commitment to

doing the most good. As MacAskill writes, the question that Effective Altruists

ask themselves is “how can we . . . succeed in having the greatest positive impact

we can?” (MacAskill, p. ; see also MacAskill, p.  and Singer, p. vii).

Both books show that EA is a better approach to donating than passionate philan-

thropy and warm-glow giving. However, the distortions that arise from EA’s com-

mitment to maximization (doing the most good), its narrow focus on improving

individual welfare (doing the most good), and the disturbing “hidden curriculum”

that it teaches (discussed below) suggest the need to move beyond it.

Founding Moments

Effective Altruism had two founding moments. In the mid-s, MacAskill and

another philosophy graduate student at Oxford, Toby Ord, began researching the

cost-effectiveness of charities fighting global poverty. In  they started the EA

“meta-charity” Giving What We Can, which encourages people to donate at least

 percent of their income to the charities with the greatest positive impact. At

around the same time, two hedge-fund analysts, Holden Karnofsky and

Elie Hassenfeld, independently founded another organization, GiveWell, which

focuses on in-depth research to identify the charities that do the most good—a

determination that incorporates, but is not limited to, being highly cost-effective

(MacAskill, p. ). Thus, EA came of age simultaneously in Oxford, where it has

been especially popular among young analytic philosophers, and in Silicon Valley,

where it has been especially popular among young tech entrepreneurs (GiveWell is

located in San Francisco). Not surprisingly given this history, the membership of

EA organizations has been, and remains, predominantly young, white, well-

educated, able-bodied, and (to a lesser degree) male. While Singer is at pains to

offer portraits of Effective Altruists who do not fit this description (ch. ), EA

is clearly a product of, and continues to inhabit, these milieus.

512 Jennifer C. Rubenstein

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0892679416000484 Downloaded from https://www.cambridge.org/core. University of Waikato Library, on 29 Nov 2017 at 03:36:30, subject to the Cambridge Core terms of use, available

EA has grown rapidly. At the time of this writing, Giving What We Can has

, members who have collectively donated more than $ million to “highly

effective” charities and have pledged to donate more than $ million over the

course of their careers. GiveWell estimates that it has directed more than $

million to what it sees as highly effective charities since its founding in .

While these numbers are small compared to overall levels of charitable giving,

they are significant in absolute terms. They are also only one manifestation of

EA’s influence, which also includes shifting scholarly and public debates about

giving, changing how charities conceptualize and measure their effectiveness,

and influencing how some young people choose a career path.

Singer rightly describes himself as a “parent” of the EA movement. In his semi-

nal  essay, “Famine, Affluence, and Morality,” he argued that one should

donate money to address serious suffering “up to [and perhaps even beyond]

the point at which by giving more one would begin to cause serious suffering

for oneself and one’s dependents.” EA retains Singer’s emphasis on the

demanding moral duty to donate, but adds something new: using reason and

empirical evidence to ensure that one’s donations do the most good possible.

So while in his  essay Singer argued that individuals in wealthy countries

should direct their donations to help victims affected by the “emergency” of “cons-

tant poverty, a cyclone, and a civil war” in East Bengal, in The Most Good You

Can Do he clearly supports an argument that MacAskill makes explicit—that

individuals should generally not provide emergency relief in response to high-

profile disasters because it tends to do less good than other interventions

(MacAskill, ch. ).

Singer and MacAskill are both utilitarians. While they seek to build a big tent

and persuade people who do not share their utilitarian commitments, the shadow

of utilitarianism looms large over both books. Both authors view all human indi-

viduals as having equal and great moral worth, a commitment shared by many

moral theories. However, their focus on maximally increasing individual welfare

as the most appropriate way to enact this commitment is most closely associated

with utilitarianism. EA can therefore trace its ancestry not only to Singer but also

to earlier utilitarian philosophers, such as John Stuart Mill and Jeremy Bentham,

who argued for bringing about “the greatest amount of good for the greatest

number.”

the lessons of effective altruism 513

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0892679416000484 Downloaded from https://www.cambridge.org/core. University of Waikato Library, on 29 Nov 2017 at 03:36:30, subject to the Cambridge Core terms of use, available

A Philosophy and a Movement

While it has a philosophical lineage, EA is not only a set of philosophical ideas. It

is, as Singer emphasizes, “a philosophy and a movement” (p. , emphasis added).

It consists of both abstract principles and efforts to enact those principles, includ-

ing EA meta-charities, EA-influenced charities, EA blogs, discussion groups, chap-

ters, TED Talks, conferences, scholarly papers, and websites offering career

guidance. This proliferation raises the sticky question of how books that seek to

popularize EA, such as those under consideration here, should be evaluated.

Should we focus only on the core ideas of EA, for example that individuals should

do as much good as possible? Or should we also include what happens when these

ideas are put into practice?

I propose that, as a philosophy and movement that focuses on consequences, EA

should be evaluated in light of all of its overall effects, even those that run counter to

its stated intentions. For example, EA meta-charities are in principle open to en-

dorsing any organization that can be shown to do the most good. In practice, how-

ever, they systematically prioritize organizations that provide public health

interventions and cash transfers over those that engage in political advocacy,

because the effects of the former are easier to measure and compare than those

of the latter (MacAskill, pp. –). While many within EA are attempting to ad-

dress this issue, it cannot be dismissed on the grounds that it is unintended. The

regular but unintended effects of EA are features—not minor bugs—of the system.

The Argument for Effective Altruism

Singer and MacAskill make at least three important arguments in favor of

EA. These arguments are primarily empirical, but they have powerful normative

implications. First, there are huge differences in the good that can be done by

different interventions (MacAskill, ch. ). Second, these differences are often mea-

surable. While these measurements are typically quite rough, the differences

among various kinds of interventions are sometimes so great that they can still

be captured using highly inexact methods. For example, a donation of $,

can be used to train and provide one guide dog for one blind person in the

United States, or it can be used to fund surgery to prevent blindness (from tracho-

ma) for five hundred people in a poor country (MacAskill, p. ). Singer and

MacAskill’s third important argument is that most people in wealthy countries

514 Jennifer C. Rubenstein

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0892679416000484 Downloaded from https://www.cambridge.org/core. University of Waikato Library, on 29 Nov 2017 at 03:36:30, subject to the Cambridge Core terms of use, available

are very well-off in absolute terms, and vastly better off than many people in poor

countries (MacAskill, ch. ). As MacAskill points out, a worker in the United

States who earns $, a year is still in the top  percent of earners globally

(p. ).

Together, these three arguments suggest that most adults in wealthy countries

can, without immiserating themselves, donate enough to make a meaningful dif-

ference in the lives of many poor people. Even if MacAskill’s estimation that it

costs $, to save a life in the developing world (p. ) is optimistic, it casts pas-

sionate philanthropy and warm-glow giving (as well as the refusal to donate at all)

in a new light. If a young adult with trachoma, facing a lifetime of blindness, asks

you why you are funding one five-hundredth of a guide dog rather than a surgery

to enable her to see, what would you say? Singer and MacAskill do not show that

there are no good answers to this question. But they do show, powerfully, that it is

a question we should ask.

More prosaically, the idea that donors should do the most good generates in-

sights about everyday donating practices that are relevant for anyone who cares

about the effects of their donation, not only Effective Altruists. Donors are fre-

quently told that, when deciding where to donate, they should consider how

much of an organization’s budget is devoted to overhead versus program costs.

EA tells us that this figure is virtually meaningless; what matters is the amount

of good that an organization accomplishes (and the harm it inflicts) per dollar

spent, regardless of whether that dollar is categorized as overhead or program

cost (MacAskill, ch. ). Likewise, while donors frequently prioritize causes in

which their donation will address all or almost all of a problem, EA suggests that

we should focus instead on the absolute size of our contribution. As MacAskill

writes, “It’s not the size of the bucket that matters, but the size of the drop” (p. ).

Criticisms of Effective Altruism

Despite the power and usefulness of its arguments, EA has no shortage of detrac-

tors. I turn now to discussing three lines of criticism. The first is a common ac-

cusation that, I suggest, EA has the resources to parry; the latter two pose a

more serious challenge.

Intimate Donating

One common criticism of EA involves what we might call “intimate donating,”

that is, donating that is intended not only to support the aims of the recipient

the lessons of effective altruism 515

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0892679416000484 Downloaded from https://www.cambridge.org/core. University of Waikato Library, on 29 Nov 2017 at 03:36:30, subject to the Cambridge Core terms of use, available

organization but also, and sometimes primarily, to express or do something within

an intimate relationship. For example, when I donate to my niece’s school

dance-a-thon for the American Cancer Society or participate in the Ice Bucket

Challenge at the behest of an old friend, a big part of what I am doing is express-

ing my love for my niece and connecting with my friend. What can EA say to do-

nors who understand and accept that they are not doing as much good as possible,

but take themselves to be doing something else that is also morally valuable?

EA’s main response to this question has been developmentalist. At the height of

the Ice Bucket Challenge craze, MacAskill urged people to conceive of their par-

ticipation in it not as “‘doing [your] bit’ but instead as taking one small step to-

wards making altruism a part of your identity.” That is, while participating in

the challenge does not do the most good (because the recipient organization,

the ALS Association, is not one of the most effective charities), it could be part

of an individual’s process of moving toward doing the most good. Throughout

their books, both Singer and MacAskill describe how aspiring, or even tentatively

aspiring, Effective Altruists can move gradually toward a more EA-consistent life-

style, and they encourage their readers to follow suit.

This developmentalist response seems to imply that full compliance with EA

principles is incompatible with intimate donating. This is not the case, however.

As Singer notes, “Typical effective altruists leave themselves time and resources

to relax and do what they want” (p. ). So as long as donations made via the

Ice Bucket Challenge are classified as “luxury spending,” and therefore do not dis-

place donations (or other activities) that do do the most good, there is no reason to

object to them (p. ). The same point holds for other donations that do not do

the most good.

A critic might rejoin that there are important differences in kind between par-

ticipating in the Ice Bucket Challenge and, say, buying oneself an expensive stereo.

Effective Altruists do not need to deny this, however. Their argument suggests that

the bolded boundary on our moral maps ought to be drawn in a different place

than we usually draw it, not that other kinds of boundaries do not exist. They

aim to shift attention away from the distinction between all kinds of do-gooding

(including doing the most good) and self-interest, and toward the distinction be-

tween doing the most good and everything else (including both self-interest and

other kinds of do-gooding). This does not commit them to denying the existence

of salient distinctions within the latter category.

516 Jennifer C. Rubenstein

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0892679416000484 Downloaded from https://www.cambridge.org/core. University of Waikato Library, on 29 Nov 2017 at 03:36:30, subject to the Cambridge Core terms of use, available

If all donations did about the same amount of good (and no harm), there would

be no reason not to deploy one’s charity dollars to, in effect, do double duty by

both expressing love or enacting friendship while also contributing to the broader

social good. However, EA tells us that because there are vast differences in the

good that different kinds of donations do, this kill-two-birds-with-one-stone strat-

egy involves a steep moral cost. The more general point that I take from this is not

that leading a life fully devoted to nurturing deep friendships—or to playing

music, training guide dogs, or undertaking basic scientific research—is wrong-

headed, as some Effective Altruists imply, but rather that such a life sacrifices

what Effective Altruists argue persuasively is a distinctive and important value:

doing as much good as you can.

Effective Altruism Is Anti-Political

A second and more serious criticism of EA is that it offers technical and economic

solutions to what is, in fact, the political problem of severe poverty. Among other

things, critics allege that EA is biased against advocacy for political change. As

evidence, critics typically point out that, as noted above, most organizations

identified as “most effective” by EA meta-charities focus on public health and

cash transfers, not political advocacy. One explanation for this (which Effective

Altruists readily acknowledge) involves measurement bias: because EA aims to

do the most good, it needs to measure and compare the good that different inter-

ventions accomplish (MacAskill, pp. –). This objective helps to drive EA’s

singular focus on doing good and its reticence to consider other values, which

are harder to measure and commensurate. It also gives EA a reason to conceptu-

alize “doing good” as increasing individual welfare: because individual welfare can

be measured (sort of) in “quality-adjusted life years” or similar units, the effects of

initiatives that aim to increase individual welfare can be compared—for example

by using randomized controlled trials (MacAskill, pp. –). Because public

health interventions and cash transfers are more amenable to these approaches

than political advocacy, EA meta-charities have tended to focus on them.

Why is EA’s relative inattention to political advocacy objectionable? Critics

offer two distinct reasons. One is that political advocacy addresses “root causes,”

and so is more cost-effective and sustainable over the long term than the

“Band-Aid” solutions that EA supports, and thus is better than the approaches

that EA endorses by its own criteria. However, I think EA can respond to this ver-

sion of the objection fairly easily. Given choices between sure bets that do less

the lessons of effective altruism 517

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0892679416000484 Downloaded from https://www.cambridge.org/core. University of Waikato Library, on 29 Nov 2017 at 03:36:30, subject to the Cambridge Core terms of use, available

good and long shots that do more good, and between projects that do less good

but have effects that are easier to measure and projects that probably do more

good but have effects that are harder to measure, EA can choose the latter options.

That is, it can put relatively more weight on the former of two values that it al-

ready endorses: doing the most good and certainty about doing good.

Another objection to EA’s inattention to political advocacy is much more chal-

lenging. According to this objection, the problem is not simply measurement bias;

it is a more thoroughgoing anti-political sensibility. Indeed, while both Singer and

MacAskill explicitly discuss political advocacy campaigns in their books (Singer,

pp. –; MacAskill, pp. , –), both seem to see the world fundamentally

through the lens of individual needs and welfare, rather than power, inequality,

injustice, exploitation, and oppression. This sensibility seems to contribute to,

and not (only) result from, a judgment that interventions aimed at addressing

the latter sorts of issues are risky and unlikely to succeed. For example, Singer

writes that “it isn’t clear that making the rich richer without making the poor

poorer has bad consequences, overall. It increases the ability of the rich to help

the poor, and some of the world’s richest people, including Bill Gates and

Warren Buffett, have done precisely that” (p. ; see also p. ). But the salient

question about making the rich richer is not, as Singer implies, whether things

would be better under global communism (p. ); it is, rather, about whether

and how extreme economic inequality contributes to political inequality and

domination.

Altering this anti-political sensibility and its downstream effects would, I think,

require a transformational shift in EA. One way to enact this transformation,

which I discuss in more detail below, would be for EA to explicitly incorporate

other values in addition to increasing individual welfare, such as justice, fairness,

and inclusion. (Singer notes that some Effective Altruists already care about these

other values as ends in themselves, but he himself seems not to [p. ]). This more

pluralistic approach is philosophically coherent and provides a powerful basis for

critiquing passionate philanthropy and warm-glow giving. The bigger challenge

involves its implications for EA as a movement. Among other things, EA organi-

zations such as GiveWell and Giving What We Can currently offer (a) concrete,

easy-to-follow guidance to donors about where to donate and (b) the opportunity

to read and/or engage in discussions that draw heavily on the tools of economics

and analytic philosophy. Shifting the central objective of EA to “doing a lot of

518 Jennifer C. Rubenstein

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0892679416000484 Downloaded from https://www.cambridge.org/core. University of Waikato Library, on 29 Nov 2017 at 03:36:30, subject to the Cambridge Core terms of use, available

good and attending to other values” would require a more extensive rethinking of

both offerings.

To be clear, adopting a more political sensibility is not the same as designating

more organizations that engage in political advocacy as “most effective.” Even

when it involves empowering marginalized groups, political advocacy by outsiders

can amount to democratically illegitimate meddling. It can also distort the agen-

das of domestic organizations or individuals that have a better understanding of

local conditions than the external organizations that fund them. Even advocacy

aimed at reforming the harmful policies and practices of a donor’s own govern-

ment, which some critics of EA have endorsed, is not without moral danger.

For example, Section  of the Dodd-Frank Act requires U.S. companies to

show that minerals they import from conflict-prone areas in the Democratic

Republic of the Congo are not “conflict minerals.” Yet despite good intentions, ac-

tivists from both the DRC and the United States have argued that not only does

this do little to stem the conflict, it also makes life much worse for the miners and

their families. All of which is to say that the question should not be “How can

EA meta-charities tweak their criteria so they end up endorsing at least some or-

ganizations that do political advocacy?” but rather “How can EA as a whole ensure

that its analysis is attentive to political dynamics—even if it turns out that the best

role for outsiders is addressing symptoms rather than underlying causes?”

Effective Altruism’s “Hidden Curriculum”

My own deepest worries about EA pertain not to the interventions that it sup-

ports, but rather to the self-understandings and sensibilities that it cultivates in

its members via its “hidden curriculum.” A hidden curriculum is the set of “un-

written, unofficial, and often unintended lessons, values, and perspectives” that are

taught alongside an explicit official curriculum. EA teaches a hidden curriculum

to its members through its arguments and images, the emotions it evokes, and the

relationships that it fosters and discourages.

As I discuss below, EA’s hidden curriculum includes at least four lessons: ()

Effective Altruists are heroic rescuers; () doing good is largely an individualistic

project; () doing the most good does not require listening to those affected by the

issues one is trying to address; and () anger is not an appropriate response to

severe poverty. I am not suggesting that Effective Altruists consciously believe

these lessons or teach them explicitly; nor can I offer here evidence about the ex-

tent to which they have been taken up. But I do think that EA conveys them.

the lessons of effective altruism 519

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0892679416000484 Downloaded from https://www.cambridge.org/core. University of Waikato Library, on 29 Nov 2017 at 03:36:30, subject to the Cambridge Core terms of use, available

Moreover, these lessons are not mere window-dressing; they serve important

functions in helping EA attract and retain members. Let us look at them more

closely.

Lesson : Effective Altruists are heroic rescuers. EA’s hidden curriculum teaches

people in wealthy countries to think of themselves as heroes and rescuers. In the

opening paragraph of his book, Singer recounts a realization that one of his

students had that led him to adopt an EA lifestyle: “Suppose you see a burning

building and you run through the flames and kick a door open, and let one

hundred people out. That would be the greatest moment in your life. And I could

do as much good as that!” (p. ). Singer tries to motivate his readers to become

Effective Altruists by encouraging them to adopt this same outlook—to see that

they, too, can easily rescue many people. MacAskill does the same thing: “You

pass a burning building, kick the door down, rush through the smoke and flames,

and drag a young child to safety. . . . If you saved several people’s lives [in this way]

. . . you’d be a hero.” And then he adds, “But we can do far more than that”

(pp. –).

This kind of high-drama emergency rescue scenario is powerfully motivating

(which might be why Singer consistently invokes such scenarios in his work).

However, it might also encourage aspiring Effective Altruists to think of them-

selves as rescuers, and the people they wish to assist as helpless victims more gen-

erally. This conception of “self and other” can have several negative and

distorting effects. It can make it harder for the self-described rescuer to notice

the ways in which she has contributed to and/or benefited from the problems

she seeks to address, and it can lead her to discount the insights of the “victims.”

The desire to have personally made a quantifiable difference can contribute to a

bias in favor of interventions, such as public health interventions, that can be plau-

sibly parceled out in ways that give credit to individual donors (for example, “Your

donation will save ten lives!”).

An obvious response to this criticism is that EA should simply use different tac-

tics to motivate its members. While I agree that this is a good idea, a central

strength of EA is that it eschews several other motivational strategies that are

equally or more objectionable—for example, showing images of pitiful-looking

“identifiable victims.” While some form of political solidarity is likely preferable,

the fact that EA focuses on doing the most good means that Effective Altruists

have to be ready to cut bait if a particular organization or cause ceases to do

520 Jennifer C. Rubenstein

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0892679416000484 Downloaded from https://www.cambridge.org/core. University of Waikato Library, on 29 Nov 2017 at 03:36:30, subject to the Cambridge Core terms of use, available

the most good, which significantly complicates political solidarity as a motivational

strategy.

Lesson : Doing good is a largely individualistic project. A second lesson in EA’s

hidden curriculum is that doing good is an individualistic project. Of course,

heroes are lone figures, so this lesson overlaps with the previous one. But here I

emphasize that EA assimilates the role of donor to two other individualistic roles.

One is the savvy customer: MacAskill notes that donors in rich countries can

provide “x” more benefit by donating to the most cost-effective charities than

to other ones. This, MacAskill exclaims, “might be the most amazing deal you’ll

see in your life” (p. ). The other role is the virtuous self-improver: This idea—

that Effective Altruists ought to be engaged in a continual process of

self-conscious self-improvement—emerges out of the developmental story that

both MacAskill and Singer tell, discussed above. As Singer puts it, Effective

Altruists are people who “like to challenge themselves, to do a little better this year

than last year” (p. viii).

All three of these aspirational figures—the hero, the savvy consumer, and the

virtuous self-improver—direct the donor’s focus back toward herself, and away

from collective practices of joint action with others. I worry that if doing good

is conceived of as not only the most important project but also as akin to heroism,

bargain-hunting, and self-improvement, then more solidaristic modes of action—

which have historically undergirded many projects that have significantly im-

proved the human condition, such as struggles for civil rights—will cease to be

easily imaginable.

Lesson : Efforts to do the most good do not require listening to or working

alongside those affected by the issues one is addressing. While EA is individualistic

in some respects, it is also a “community” (MacAskill, p. ). The EA meta-charity

Giving What We Can intentionally fosters community among its members both

to motivate them and to reduce the social burdens of an EA lifestyle (such as not

going out to expensive restaurants). Yet unlike many other identity-based

communities, the EA community systematically excludes the very people who are

its central objects of concern. While alleviating severe poverty is a primary aim of

EA, poor people generally do not have enough money or time to become Effective

Altruists. Likewise, while there are exceptions, poor people and activists in poor

countries often find themselves already enmeshed in the particular issues that

the lessons of effective altruism 521

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0892679416000484 Downloaded from https://www.cambridge.org/core. University of Waikato Library, on 29 Nov 2017 at 03:36:30, subject to the Cambridge Core terms of use, available

affect their communities, such that one of EA’s central questions—Which of all of

the issues and causes in the world should I support in order to do the most good?

—is not salient for them.

As a result, when members of the EA community discuss issues among them-

selves, they are not talking to poor people or to activists who work closely with

them (as far as I have been able to discern). Nor does there seem to be any effort

to put the EA community in contact with activists, civil society groups, or NGOs

based in poor countries. I tend to think that this is objectionable in itself. But in

addition, an important limitation of the methods that EA uses to evaluate different

interventions, such as randomized control trials, is that they do a poor job of cap-

turing long-term, diffuse, and subtle political and social effects—for example, on

intergroup social relationships, state capacity, and civil society mobilization.

Engagement with those affected and others with more direct experience of the is-

sues that EA seeks to address would likely be extremely valuable in helping them

account for these effects. While Effective Altruists have increasingly noticed the

limitations of their formal models and the need to rely more on “intuition” and

“reality checks,” they do not seem to have structures in place to take up expe-

riential, anthropological, or humanistic modes of knowledge (a notable exception

is the Open Philanthropy Project’s work on the history of philanthropy).

In addition to potentially undermining EA’s efforts to achieve its own goals, this

sort of insularity can contribute to a more generalized myopia. For example, at a

recent global conference on Effective Altruism an event organizer declared in an

introductory session that he “really [does] believe that Effective Altruism could be

the last social movement we ever need.” Given that EA has virtually no interest

in, or capacity for, addressing issues of injustice, exploitation, and oppression, one

wonders whether more time spent listening to people facing those issues might

have led the speaker to a different conclusion.

Lesson : Anger is not an appropriate response to the issues that EA addresses.

While Effective Altruists are often accused of being cold rationalists, and while

they do prioritize reason over emotion, they do in fact recognize a role for

emotions (Singer, p. ; see also pp. , , –, , and MacAskill, pp. , ).

What is the emotional content of EA? Recall Singer’s student, noted above, who

declared that saving lives “would be the greatest moment” in his life (see also

p. ). Singer also describes another Effective Altruist who gives “cheerfully”

(p. ). Singer adds that “effective altruists don’t see a lot of point in feeling guilty.

522 Jennifer C. Rubenstein

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0892679416000484 Downloaded from https://www.cambridge.org/core. University of Waikato Library, on 29 Nov 2017 at 03:36:30, subject to the Cambridge Core terms of use, available

They prefer to focus on the good they are doing. Some of them are content to

know they are doing something significant to make the world a better place” (p.

viii). Finally, Singer emphasizes that one can be an Effective Altruist without

feeling grim and depressed about living on less money than one would otherwise

(pp. –, ch. ). (Anecdotally, when I submitted a blog post to the Giving What

We Can Blog, I received instructions for authors explaining that “because we are

welcoming, we use photos of our members: smiley, happy people.”)

There thus seems to be little room for emotions such as anger, guilt, and self-

doubt in EA. While Singer and MacAskill do not say that these emotions are bad

or illegitimate, they strongly emphasize the value and appropriateness of their op-

posites (MacAskill, p. ). Of course, it is true that, as Martha Nussbaum argues,

anger can direct energy to retribution rather than constructive efforts for change.

However, as Nussbaum also writes, “Anger is sometimes justified and right. It is

an appropriate response to injustice and serious wrongdoing.” On a cognitive

view of emotions, being angry about an injustice can be partly constitutive of

recognizing it as such. Likewise, guilt can be a part of acknowledging that one

has wrongly contributed to suffering, and self-doubt can reflect humility and an

understanding that one does not have all of the answers.

Perhaps, then, EA could create a bit more space for these emotions? Among

other things, recognizing the appropriateness of some kinds of anger would be

consistent with a more politicized analysis of the issues that EA addresses: what

looks like a regrettable misfortune when we focus only on individual welfare (a

child gets malaria) becomes recognizable as an injustice (to which anger is an ap-

propriate response) when we look at the underlying structures that contribute to it

(an externally enabled corrupt ruling elite that drains resources from government

public health budgets). Openness to negative emotions may be challenging, how-

ever, insofar as members of EA are attracted to its cheery emotional register, and

anger at injustice can be difficult to square with willingness to cut bait if an inter-

vention does not seem to be doing the most good.

EA’s hidden curriculum is deeply troubling, in part because it is not window-

dressing: it does important work, especially motivational work. But this should not

cause us to lose sight of EA’s central insight, described above: we should be think-

ing rigorously, comparatively, and collectively about how individuals can use the

money in their possession to improve the world, including (but not only) by in-

creasing individual welfare.

the lessons of effective altruism 523

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0892679416000484 Downloaded from https://www.cambridge.org/core. University of Waikato Library, on 29 Nov 2017 at 03:36:30, subject to the Cambridge Core terms of use, available

Another Way Forward?

What might it look like for a philosophy and social movement to grow up around

this insight, without EA’s narrow focus on either maximization (doing the most

good) or welfare (doing the most good)? Can we imagine an organization, very

roughly analogous to GiveWell, that would offer guidance to individual donors

who reject warm-glow giving and passionate philanthropy, and are looking in-

stead for guidance on donating that is rigorous, empirically grounded, compara-

tive, and attentive to consequences but that also acknowledges and incorporates a

range of hard-to-measure substantive and procedural values, including justice,

equality, fairness, and empowerment of those directly affected?

Critics of charity will object that in a just world private individuals would not be

able to accumulate enough money to make voluntary donating a significant driver

of social change, and in the current (unjust) world the wealthy are unlikely to sup-

port reforms that would limit their ability to accumulate. I think this objection is

overstated, both because an effective social practice of donating does not require

the excessive accumulation that we associate with such megadonors as Bill Gates

and Warren Buffett, and because individuals often donate to support political re-

forms that will reduce their own ability to accumulate.

Any effort to create a more pluralistic, less maximizing GiveWell would of

course learn from and build on the work of many existing organizations, such

as Solidaire, which funds social change organizations in the United States;

Thousand Currents (formerly IDEX), which gives grants to grassroots organiza-

tions in poor countries and accepts donations from the general public; and

Shack/Slum Dwellers International and WIEGO, both of which empower and

give voice to very poor people. Such an effort would also build on the insights,

experiences, and arguments of EA, which bring into view the following crucial

questions that would have to be answered in order to ascertain whether the sort

of GiveWell alternative that I have been describing would be possible or beneficial:

• How can we create criteria for evaluating organizations, interventions, or

causes that () make comparison possible, but () are responsive to the

plural goals of the organizations being evaluated, the people they aim to

assist, and the donors who will use these criteria to inform their donating

practices?

524 Jennifer C. Rubenstein

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0892679416000484 Downloaded from https://www.cambridge.org/core. University of Waikato Library, on 29 Nov 2017 at 03:36:30, subject to the Cambridge Core terms of use, available

• What are the precise mechanisms whereby donations from individuals in

wealthy countries can promote not only individual welfare but also justice,

equality, and/or other values? What are the best sources for learning about

these mechanisms?

• How can organizations based in the Global South accept donations from

individuals in the Global North without compromising their actual and

perceived democratic legitimacy?

• Given that several of the negative lessons in EA’s hidden curriculum

emerge from its efforts to motivate its members, would this alternative

organization simply serve as a resource for donors who are already moti-

vated? How would it balance solidarity and partnership with the willing-

ness to shift gears?

Singer and MacAskill have done an outstanding job of presenting EA to a wide

audience. For those of us who see the promise in EA, but also recognize its lim-

itations, their books are best seen as contributions to a much larger conversation.

NOTES

 “Librarian Quietly Saved $ Million For Gift Back To Library,” Huffington Post, September , , www.huffingtonpost.com////library-donation_n_.html.

 Amount donated by individuals according to Charity Navigator, www.charitynavigator.org/index.cfm? bay=content.view&cpid= (accessed September , ). Estimations of the percentage of the U.S. population that gives range from  percent of households (Philanthropy Roundtable, www.philanthro pyroundtable.org/almanac/statistics/) to  percent of adults (Gallup, “Most Americans Practice Charitable Giving, Volunteerism,” December , , www.gallup.com/poll//americans- practice-charitable-giving-volunteerism.aspx). Both sites accessed September , .

 The term “robot apocalypse” is from Amia Srinivasan, “Stop the Robot Apocalypse,” London Review of Books , no.  () pp. –. See also the Open Philanthropy Project, “Global Catastrophic Risks,” February , www.openphilanthropy.org/research/cause-reports/global-catastrophic-risks/global- catastrophic-risks (on asteroids); and Zander Redwood, “Living to ,: An Interview with Aubrey de Grey,” , Hours, April , , hours.org///living-to--an-interview-with-au- brey-de-grey/ (on aging).

 See the Giving What We Can website, “Our History,” www.givingwhatwecan.org/about-us/history/.  See the GiveWell website, “GiveWell’s Impact,” www.GiveWell.org/about/impact.  Peter Singer, “Famine, Affluence, and Morality,” Philosophy and Public Affairs , no.  (), pp. –.

 Julia Driver, “The History of Utilitarianism,” in Edward N. Zalta, ed., Stanford Encyclopedia of Philosophy, Winter  Edition, plato.stanford.edu/archives/win/entries/utilitarianism-history/.

 It also means that the positions of EA organizations, and discussions of EA on blogs and websites, have already moved beyond some of the ideas in these books. For example, while MacAskill describes “earn- ing to give” (choosing a high-earning job so that you can give away more money) as a “solid bet” (pp. –), the EA organization that proposed it (, Hours) has pedaled back its support for this approach. See William MacAskill, “, Hours Thinks that Only a Small Proportion of People Should Earn to Give Long Term,” , Hours, July , , hours.org///- hours-thinks-that-only-a-small-proportion-of-people-should-earn-to-give-long-term/. While this re- view responds primarily to the version of EA presented by MacAskill and Singer, readers who want an up-to-date understanding of the positions and debates circulating within EA should consult blogs and websites of EA and EA-affiliated organizations, such as GivingWhatWeCan.org and GiveWell.org.

the lessons of effective altruism 525

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0892679416000484 Downloaded from https://www.cambridge.org/core. University of Waikato Library, on 29 Nov 2017 at 03:36:30, subject to the Cambridge Core terms of use, available

 Some empirical scholars doubt the veracity of these two arguments. See Angus Deaton, Responding to Peter Singer in the Boston Review online, July , , bostonreview.net/forum/logic-effective- altruism/angus-deaton-response-effective-altruism.

 The best critical discussions of EA (some more critical than others) include those on Amia Srinivasan, in the “Thread for Discussing Critical Review of Doing Good Better in the London Review of Books,” September , , effective-altruism.com/ea/nw/thread_for_discussion_of_critical_review_of_doing); Iason Gabriel, “Effective Altruism and Its Critics,” Journal of Applied Philosophy , no.  (); Dylan Matthews, “I Spent a Weekend at Google Talking with Nerds About Charity. I Came Away . . . Worried,” Vox, August , ; Derek Thompson, “The Greatest Good,” Atlantic, June , ; and Robert Simpson, “Moral Renegades,” New Rambler Review of Books, July , , newramblerreview. com/book-reviews/philosophy/moral-renegades.

 William MacAskill, “The Cold, Hard Truth About the Ice Bucket Challenge,” Quartz, August , , qz.com//the-cold-hard-truth-about-the-ice-bucket-challenge/.

 The term “luxury spending” is not Singer’s, but he cites it approvingly.  Deaton, Responding to Peter Singer, July , .  Ted Lechterman, “The (Effective) Altruist’s Dilemma,” unpublished manuscript (August ).  For example, Nandini Deo and Duncan McDuie-Ra describe how, in order to please external funders,

NGOs in India prioritized interventions aimed directly at ending violence against women, rather than addressing the underlying cause of such violence, which was women’s lack of access to employment. Deo and McDuie-Ra, The Politics of Collective Advocacy in India: Tool and Traps (Sterling, Va.: Kumarian Press, ), ch. .

 Deaton, Responding to Peter Singer, July , .  Laura Seay, “What’s Wrong with Dodd-Frank ? Conflict Minerals, Civilian Livelihoods, and the

Unintended Consequences of Western Advocacy,” Center for Global Development Working Paper , January , www.cgdev.org/publication/what%E%%s-wrong-dodd-frank--conflict- minerals-civilian-livelihoods-and-unintended.

 See the website of the Glossary of Education Reform, “Hidden Curriculum,” edglossary.org/hidden- curriculum/.

 Peter Singer, The Life You Can Save (New York: Random House, ).  Gabriel, “Effective Altruism and Its Critics.”  The GiveWell Blog, “Why We Can’t Take Expected Value Estimates Literally (Even When They’re

Unbiased),” blog.givewell.org////why-we-cant-take-expected-value-estimates-literally-even- when-theyre-unbiased/, and “Some Considerations against More Investment in Cost-Effectiveness Estimates,” blog.givewell.org////some-considerations-against-more-investment-in-cost-effec- tiveness-estimates/. Both posts updated July , .

 Open Philanthropy Project, “History of Philanthropy,” www.openphilanthropy.org/research/history- of-philanthropy?expand_menu=.

 Matthews, “I Spent a Weekend at Google Talking with Nerds About Charity.”  Martha Nussbaum, Anger and Forgiveness: Resentment, Generosity, Justice (New York: Oxford

University Press, ).  Martha Nussbaum, Political Emotions: Why Love Matters for Justice (Cambridge, Mass.: Belknap Press,

), p. .  Mathew Snow, “Against Charity,” Jacobin, August , , www.jacobinmag.com///peter-

singer-charity-effective-altruism/.  Solidaire, solidairenetwork.org/.  Ibid.  “Thousand Currents: Exchanging Grassroots Brilliance,” as described on the homepage of the

International Development Exchange (IDEX), www.idex.org/.  Shack/Slum Dwellers International (SDI), “About SDI,” sdinet.org/about-us/what-we-do/; and

Women in Informal Employment: Globalizing and Organizing (WIEGO), wiego.org/.  Deo and McDuie-Ra, The Politics of Collective Advocacy in India.

526 Jennifer C. Rubenstein

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0892679416000484 Downloaded from https://www.cambridge.org/core. University of Waikato Library, on 29 Nov 2017 at 03:36:30, subject to the Cambridge Core terms of use, available