Effective altruism

William Macaskill, Doing Good Better

Peter Singer, The Most Good You Can Do

These two books provide a good introduction to the emerging movement of ‘effective altruism’ which is making waves among philanthropists, students and some development wonks too. Effective altruists seek to use reason, data and evidence to maximise the amount of ‘good’ they can do in their lives. They donate a large share of their income and engage in detailed technical discussions about which charities are most effective. There are some interesting lessons for aid donors here, but also some serious questions about effective altruism, including the way it tends to portray development as a financial and technical problem, rather than a political or social one.

In 2004 the New Yorker magazine published a profile of an extraordinary man. Zell Kravinsky, a real estate magnate, had not only given away almost all of his $45 million fortune to charity, but decided to donate one of his kidneys to a stranger. Kravinsky, according to Peter Singer in The Most Good You Can Do, “puts his altruism in mathematical terms”. The risk of dying as a result of the donation was estimated at 1 in 4,000. This meant, to Kravinsky, that if he withheld his kidney from someone who would otherwise die, he would be valuing his own life at 4,000 times that of the stranger – a valuation he described as “obscene.” Kravinsky says that the reason people question his kidney donation is that “they don’t understand math.”

Kravinsky is one personal example – admittedly a rather extreme one – of a growing social movement which calls itself effective altruism. Effective altruists are committed to do as much good as they can with their lives. They tend to live modestly and donate as much as they can to charity – at least the traditional 10% of their income, often much more. They debate which charities are the most effective ones to support, based on quantitative measures such as ‘cost per life saved’. They try to choose the most effective career to do good, often concluding that they can have more of an impact by taking a highly paid job in a bank or a hedge fund and giving away most of their salary than by working directly for a charity or other ‘good cause’. And a small but growing number are also choosing to donate part of themselves– blood, bone marrow or even a kidney- to someone who needs it more than they do.

The two books under review set out the ‘creed’ of effective altruism (there is something almost like a community of believers about the movement) but they do so in quite different ways. Doing Good Better, by William Macaskill, a young associate professor of philosophy and the coiner of the ‘effective altruism’ term in 2011, comes across as a practical guide for the young, committed graduate who wants to know how to do the most good with their life. The focus is on how to distinguish between ways of helping which are highly effective from those which are ineffective – both in terms of careers and charities. There are checklists of questions you should ask before supporting a particular cause or charity, and the book concludes with some tips about how to become an effective altruist. Peter Singer’s The Most Good You Can Do is more reflective and philosophical, although still very readable. The author has been writing on the ethics of donations to charity since 1972 so is something of an intellectual father for the movement. The Most Good You Can Do reviews the emergence of the effective altruism movement, and looks into what motivates effective altruists and how they justify what they do.

Global poverty reduction is not the only ‘cause’ considered by Macaskill and Singer, but it features high on the list – especially as saving lives in poor countries is much cheaper than saving them in rich ones. In deciding which interventions are most cost-effective in fighting poverty, Macaskill puts a lot of emphasis on randomised controlled trials (RCTs) – a methodology imported from medicine where there is a group of people randomly selected to receive the ‘treatment’ (i.e. development project) and another group who do not, so the effects of the project can be more clearly identified. In fact, GiveWell – an organisation set up by effective altruists to rate charities according to their impact and find those which deliver “real change for your dollar” – recommends only four ‘top charities’ all of which are backed by RCT studies.

RCTs have been subject to a lot of debate in development circles in recent years, much of it critical. I have some sympathy for those critiques, but before I get on to them I’d like to consider the positive contribution that effective altruism can bring to development agencies. Firstly there is a welcome emphasis on really understanding in detail what an NGO is achieving. Sometimes in development agencies we struggle to articulate what a particular project or programme is actually doing in practice – for example, it sounds much better to say your project is “building capacity” than to say you are buying computers and running training courses. Also, whilst staff in development agencies go through a lot of hoops before approving new projects, including cost benefit analysis and estimates of value for money, too often these are rhetorical exercises based on weak underlying numbers, so the emphasis on ‘hard evidence’ vetted by a completely external organisation is attractive (at least the basis of GiveWell’s judgements is clear, even if you don’t agree with it). There’s also a welcome emphasis on following up to ensure effective delivery. The Against Malaria Foundation, for example, gets points on GiveWell because it has “strong processes for ensuring that nets reach their intended recipients and monitoring whether they are used over the long-term” – to avoid classic problems such as nets being sold on or used for fishing. Finally it’s notable that GiveWell’s top rated NGOs are all focused on one particular kind of intervention. There may be good reasons why development agencies and large NGOs try to do a little bit of everything – but there are surely some benefits to focusing on one thing and doing it extremely well.

This said, there are aspects of the effective altruists’ worldview that to me are somewhat troubling. As others have pointed out, effective altruists have to date – and especially on GiveWell – tended to focus on charities which provide cost-effective ‘fixes’ such as distributing insecticide treated bed nets or deworming pills. They do not seem to have engaged much with what to me are the bigger factors which influence global poverty – for example the broad-based economic growth which is lifting millions of people out of poverty in countries like China and Vietnam; the corrupt political systems in many developing countries which ensure the benefits of development go mainly to the elite; and the ways rich countries’ actions and policies can act to keep poor people poor –through, for example, support for undemocratic leaders, the arms trade or tax havens. Effective altruists will point out that their movement is not only about GiveWell, and that there are other initiatives which are considering how effective altruists can support what they call ‘systemic change’ – for example the Open Philanthropy Project. But this has not to date looked at any of the structural factors which underpin world poverty.

The other troubling aspect of the movement is the way it explicitly discourages people from basing their charity on an emotional connection. For example Macaskill says despite having visited a Fistula hospital in Ethiopia he decided it would not be appropriate for him to support the Fistula Foundation as he believed there were other charities who would make more effective use of his money. I don’t disagree with that decision in itself, but in a world where people are becoming more individualised, and genuine connections and empathy seem to be rarer and rarer, there is something concerning about subjecting every emotional impulse to a cost-benefit calculation. I’d rather urge people to hold on to their impulses to be kind to their neighbours and help people they know, rather than exclusively basing their desire to do good on sending money abroad.

Effective altruism is an interesting movement which has the potential to shake up the development business in some ways. The presence of an independent organisation like GiveWell may increase the pressure in development agencies for better measurement of their effectiveness. But in responding to this pressure, it will be important to remember that the numbers do not tell the whole story. For me, evaluation of development programmes’ effectiveness has to involve not only quantitative calculations of ‘cost per life saved’ but also a thorough, qualitative assessment of the ways in which people’s lives are affected by programmes – including especially their unintended consequences.

Advertisements

2 thoughts on “Effective altruism

  1. Hey, Justin! Thanks for a great article – all around a pretty fair review. I appreciate that you gave credit to EA for a few positive developments, particularly in how we handle and use evidence.

    Responding to this criticism that EA does not work on political or systemic change, I’ll say a couple of things. Links all at the end.

    First, Effective Altruists absolutely are engaging with political/systematic change within global poverty, in many different ways. Here are a few examples that come to mind: Open Philanthropy has worked significantly on land use reform (US poverty) and international barriers to migration (global poverty). Open Borders is a group that is loosely associated with EA that tries to promote the removal of barriers to migration. Wave is an app co-founded by an EA that allows migrants to send remittances back home for far less than through traditional money sending services. GiveDirectly has greatly strengthened and popularized the empirical case for a “basic income,” an idea that EAs generally support. Thomas Pogge, a philosopher, has argued the moral case against tax havens, arms trade, and many other causes that EAs are generally sympathetic to. I realize that some of these cases won’t strike you as systemic/political change, but I’d argue that 1) they are popularizing and building the coalition for longer-term political change and that 2) political action doesn’t usually resemble traditional political actions like protesting/organizing/etc. at first.

    Second, the EA movement believes its comparative advantage is in supporting causes that are highly evidence-based and scaling them up – this generally biases against more speculative causes like political/systematic change. However, organizations like the Open Philanthropy Movement are trying to figure out what political/systemic actions they can take that actually are effective. Doing this is tremendously difficult, especially in the face of weak evidence. I can point you to other individual effective altruists that I know working on this type of thing, particularly in the animal rights movement, a cause that effective altruists generally feel is extremely important and more amenable to traditional political actions. EAs understand that the rewards for political and systemic change are high, but it’s very difficult to do effectively. That’s why we’re putting so much energy into getting it right!

    Third, even if EA doesn’t figure out how to do systemic/political change well, I think the contributions we’ve made to development and global health are already pretty huge. EAs use evidence in ways that I feel are much more responsible / effective than most traditional development agencies. If a larger portion of international aid were allocated in ways that EAs would like, we would have many, many more times impact than we currently do.

    I hope this is helpful!

    https://80000hours.org/2016/02/doing-good-through-for-profits-lincoln-quirk-and-wave/
    http://www.openphilanthropy.org/focus/us-policy/land-use-reform
    http://openborders.info/
    http://www.vox.com/2016/1/28/10860830/y-combinator-basic-income

    Liked by 1 person

    1. Thanks Scott – glad you liked the article and some good thought provoking responses.

      First I’m glad to hear that EAs are keen on basic income. This is something I’ve considered in a previous review post (https://developmentbookreview.com/2016/03/11/review-james-ferguson-give-a-man-a-fish/). If Ferguson is right, the idea of a basic income could have far-reaching political implications, and the case for it is pretty strong. I do think there is a difference though between the political effects of a cash transfer programmes supported by aid agencies (manna from heaven, if you like) and a basic income programme supported by the state (demanded by the people as their rightful share of national wealth). In the long run I think the political effects Ferguson envisages are more likely to come into effect if the transfers are provided as a right rather than a charity handout.

      Secondly, in terms of systemic change more broadly, it’s interesting to hear about the initiatives that EAs are taking. I guess the difference between us comes down to the need to have evidence of effectiveness for what are inevitably going to be speculative actions. If the civil rights activists in the 60s or demonstrators in Eastern Europe in 1989 had waited for solid evidence before they did anything, arguably a lot more harm would have been done while they were waiting. In development, taking on that lesson may mean fewer experiments (i.e. gold standard RCTs) and more experimentation (i.e. doing something, monitoring closely what effects it seems to have, and trying something else if it doesn’t work). The distinction between experiments and experimentation is something Lant Pritchett has emphasised – see this blog post from Duncan Green: https://oxfamblogs.org/fp2p/lant-pritchett-v-the-randomistas-on-the-nature-of-evidence-is-a-wonkwar-brewing/

      Having said this, I think development agencies, and individual staff within them, often hide behind a statement like “it’s just not possible to get evidence on whether x is effective – but we should still do it because blah blah blah”. This is not what I am saying. The experimentation approach is hard to get right and it’s really difficult to know what effects your project is having in the short term. In both approaches there is a need for better data and more effective M&E. But I would just argue that sometimes we need to take a punt without knowing what effects we are going to have and monitor very well as we go along. Also as I hint at the end, it’s really important to understand ‘side effects’ of development programmes – which are often not factored in to RCTs and could be a way of making them even more rigorous.

      Thanks again for your thoughtful response!

      Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s