How much do people pursuing earning to give actually give?

Introduction

The purpose of this document is to estimate:

1. How much do people pursuing earning to give donate, and how much can we expect them to in the immediate future?
2. How much more giving has 80,000 Hours caused?

The second part fits into our upcoming impact evaluation.

Summary

How much do people pursuing earning to give donate?

• We estimate there’s at least 100 people pursuing earning to give in the effective altruism community, based on survey data from the effective altruist organisations.
• From our own surveys, we’ve found 39 people pursuing earning to give whose career plans have been changed by 80,000 Hours.
• A random sample of ten of these 39 reported total donations to high impact and meta-charities over the last three years of £210k.
• The top three donors we know among the group of 39 gave £230k over the last three years.
• The members of the random sample of ten estimate they will donate £1.6m over the next three years to high impact charities and meta-charities. The top five donors we know expect to donate £2.1m over the next three years. If a significant number continue earning to give, donations will be substantially higher after 2016 due to rapid expected earnings growth.
• The majority of the donations are expected to be to effective altruist organisations, followed by GiveWell recommended charities.
• The estimates are complicated by: (i) the difficulty forecasting salary (ii) the chance of mass drop outs from earning to give (iii) biases in reporting (iv) dependence upon a couple of individuals, who account for a large fraction of the donations.
• Overall, our best guess estimate is that the group of 39 has donated £230k – £400k over the past three years, and will donate about £2m over the next three years (with an 80% confidence interval of £500k – £4m).
• The entire earning to give community of around 100 to 150 is likely to be donating about two to three times as much.

How much more giving has 80,000 Hours caused?

• We asked the random sample of ten from among the group of 39 to estimate how much of their expected giving to effective charities is due to 80,000 Hours. On average, they estimated 30%.
• We searched the group of 39 for the donors who attributed the most donations over the next three years to 80,000 Hours’ influence. Collectively, the top five attribute £565k. More detail on each individual is in the Appendix.
• These estimates are complicated by all the issues that complicate the overall estimates of donations, plus additional biases and the difficulty of separating our influence from the rest of the effective altruism community.
• Based on the survey data and these considerations, our overall estimate is that 80,000 Hours has caused £50k of donations over the past three years. We expect to cause £500k of donations over the next three years, and substantially more after that, although our estimates are highly uncertain. The increase over the next three years is because more of the community will soon start work, and others are expected to see rapid salary growth during their first years of employment.

Section 1: How Much Is the Earning to Give Community Giving, and Where?

How many people are pursuing earning to give?

For the purpose of this evaluation, we take ‘pursuing earning to give’ to mean aimng to take a high earning career, aiming to give over 10% of your income to charity, focusing on the effectiveness of your giving, and being involved with the effective altruism community.

Counting the people identified in our survey

Our survey had 205 respondents. By reading their career plans, we judged who was pursuing earning to give as their best guess career strategy. We estimated there were about 50.

This agrees with the number of people in the survey planning to enter high earning professions (about 70, of which ~80% identify as effective altruists).

We’ve identified around another 20 from prior knowledge, making a total of 70.

This is highly likely to be an underestimate, because there are probably people who have heard the arguments (which have reached millions of people) but are not in touch with us.

What about the rest of the effective altruism community?

How many active members of the effective altruism community are there? 470 people in the 2013 Less Wrong survey identified as effective altruist. In 2012, GiveWell tracked 1270 donors who gave over $1000 to one of their recommended charities. As of March 14th, there are 457 members of Giving What We Can. 205 people filled out our impact survey, and which 80% said they were active supporters of effective altruism. In total, this adds up to 2,360 people, but because there’s significant overlap and not all of these people are effective altruists, we estimate there’s about 1,000 – 2,000 active effective altruists. There’s far more people interested in the ideas and partially engaged. For instance, there’s about 4,000 people on The Life You Can Save mailing list.

What proportion might be pursuing earning to give? GiveWell tracked 129 donors giving more than $10,000 last year. The average Less Wrong effective altruist only donates 3.5% of their income, suggesting a relatively small proportion pursue earning to give. About 35 Giving What We Can members have pledged over 20% of their income or taken the Further Pledge. Collectively, this gives another 100-200 big givers. Accounting for overlap and some not pursuing high earning careers, we’d roughly estimate there’s another 50-100 people pursuing earning to give in the effective altruism community.

Overall, we’re confident there’s at least 100 people pursuing earning to give. This is about 5-10% of the active effective altruism community.

How much do the people engaged with 80,000 Hours who are pursuing earning to give donate?

A random sample of ten

We’ve identified 107 people who have plausibly significantly changed their career plans due to 80,000 Hours (see our upcoming impact evaluation for more detail). Of these, we counted 39 who seem to be pursuing earning to give as their best guess strategy.

We randomly sampled 10 of 39 people pursuing earning to give. We asked them:

Roughly, how much have you donated over the last three years?
Where did you donate to?
Roughly, how much do you expect to donate over the next three years?
Where do you plan to donate?
How much of the above do you attribute to 80,000 Hours?

The questions were often followed by an exchange to clarify the questions. We ran our final estimates by each person, encouraged them to point out inaccuracies, and made updates based on their feedback.

Here are the results for their historical and projected giving:

*Key: 80k = 80,000 Hours, GW = GiveWell, GW reccs = GiveWell recommended charities, CFAR = Centre for Applied Rationality, MIRI = Machine Intelligence Research Institute, ACE = Animal Charity Evaluators, CSER = Centre for the Study of Existential Risk, THL = The Humane League, EA orgs = effective altruist organisations, X-risk = existential risk.

*Dollars were converted to pounds at a rate of $1.6 per pound.

We agreed to keep the identities confidential. However, we can say (7) is Patrick Brinich-Langlois and (9) is Peter Hurford.

Our main conclusion is that the distribution is very skewed. The historical mean is about 20 times larger than the median, and around 80% of the total comes from the top 10%. We think this reflects the population in general. First, because the distribution of income is highly skewed. Second, we believe there are about four people with long-term earnings prospects similar to number (10) in the total population of 39.

Tracking individual donations

Since the distribution is so skewed, we can make a reasonable estimate of the total by just tracking the top couple of donors.

Based on our current knowledge, we contacted about ten of the people we think are most likely to be among the top historical and future donors. We also asked them to list their past and expected future donations. Note that money saved to donate later was included in the totals, but this constituted under 10% of the total.

Of the donors we know, who changed plans due to 80,000 Hours, the top three donors over the last three years are:

1. Number (10), £200k
2. Number (16), £30k
3. Number (7), £40k

Total: £270k

Over the next three years, the top five donors we currently know expect to give the following to high impact charities:

1. Number (10), £940k
2. Number (17), £420k
3. Number (16), £360k
4. Number (13), £300k
5. Number (12), £100k

Total: £2,120k

Note that these estimates aimed to include the probability of changing jobs or quitting earning to give.

Beyond 2016:
Most of the people we spoke to expect to be donating significantly more after 2016, if they keep earning to give, because their earnings will grow rapidly. Several people are entering professions which offer a real chance of earning £1-20m per year within a decade. Of course, making any forecast is difficult, because there might be large drop-outs from earning to give.

The vast majority of these donations are currently expected to go to GiveWell recommended charities or effective altruist organisations.

Overall estimates

Based on the above data,what can we conclude about the donations of the group of 39 people pursuing earning to give who changed changed their career plans due to 80,000 Hours?

Key considerations:

• Expected future earnings of the group. In particular, what are the expected earnings of the most high earning people? What are the chances of someone in the group becoming very wealthy?
• Systematic risks to pursuing earning to give – might everyone quit?
• Sources of bias:
  • Are people overconfident about their expected earnings and level of donations?
  • Are they biased upwards because they want to support 80,000 Hours?

Overall, we think the estimates for the next three years are more likely to be optimistic than pessimistic, so are inclined to reduce them. On the other hand, we think there’s a reasonable chance someone is very successful and makes a very large donation. We also think there’s some chance lots of people stop earning to give. So, our overall confidence interval is fairly wide.

Beyond the next three years, uncertainty increases dramatically. On the one hand, expected earnings are much higher. On the other hand, it becomes harder to estimate the risk of drop out. In the future, it may be helpful to make some estimates of expected lifetime earnings for the group, and combine these with drop out rates measured over the next few years, to make an estimate of expected lifetime donations.

Our overall subjective estimates of total donations from the group of 39:

• Over the past three years:
  • Mean estimate: £300k
  • 80% confidence interval: £230k – 400k, or £6k – £10k per person.
• Over the next three years:
  • Mean estimate: £2m
  • 80% confidence interval: £500k – £4m, or £13k – £100k per person
• Annually beyond 2016: £200k – £10m per year

What can we infer about the whole earning to give community?

The data above is based on the 39 people pursuing earning to give who report changing their career plans due to 80,000 Hours. There seems to be about 100 people pursuing earning to give in total. How much is the entire group donating? One way to estimate is to extrapolate. However, there are likely to be some differences between the two groups. These are some speculations about how the wider community differs from our sample:

• They donated more over the last three years, because they include more established people pursuing earning to give, like Jeff Kaufman and Julia Wise and Brian Tomasik). In contrast, those influenced by 80,000 Hours are younger and new to earning to give.
• They may have higher ‘tail’ donations, since it seems to include more people entering finance and tech entrepreneurship.

Section 2 – How much more giving has 80,000 Hours caused?

This is very difficult to work out, since it requires estimating what the relevant people would have done if 80,000 Hours had never existed. In this section, we present some survey data, outline the key considerations, and give our overall estimate.

Survey data

Data from the random sample of ten

The main method we pursued was asking the random sample of ten to estimate how much of their donations to effective charities they attribute to 80,000 Hours, either through encouraging them to earn more or to donate a larger fraction of their income. We defined effective charities as GiveWell, Giving What We Can or ACE recommended charities, as well as effective altruist organisations.

This gave the following results:

On average, the respondents attributed about 30% of their donations to the existence of 80,000 Hours.

Data from the largest donors

Of the donors we know well, who changed plans due to 80,000 Hours, the only donor who attributed more than £10k of donations to 80,000 Hours over the past three years is Number 10.

Over the next three years, the five donors who attribute to the most to 80,000 Hours are:

1. Number (10), £280k
2. Number (17), £105k
3. Number (14), £100k
4. Number (12), £45k
5. Number (13), £35k

Total: £565k

Note that you can see descriptions of how 80,000 Hours changed their career plans in the Appendix. also note that these estimates aimed to include the probability of changing jobs or quitting earning to give.

Key considerations

Estimating how much extra giving 80,000 Hours has caused requires taking account of the estimates above, and then estimating a difficult counterfactual: what would have happened if 80,000 Hours didn’t exist? This involves disentangling the influence of 80,000 Hours from others in the effective altruism community.

We think the self-reported estimates, however, are likely to be a good starting point. The relevant individuals aimed to isolate our influence from others in the effective altruism community. The main danger with these is that they’re biased upwards, because the people in the sample want to exaggerate the importance of 80,000 Hours, telling us what we want to hear. We aimed to combat this exaggeration by encouraging people to point out inaccuracies, but it’s likely that some overstatement remains.

The overall totals for the past and next three years are highly sensitive to Number 10’s donations and degree of influence, since they determine about 80% of the totals. Also note that the influence mainly took place before 80,000 Hours was officially founded, though involved two key members of the pre-founding team during our volunteer phase in 2011.

Considering the group more broadly, however, we think the average of 30% of donations due to 80,000 Hours is fairly reasonable in this time frame. There must be some group responsible for the influence, and since 80,000 Hours has been the largest promoter of earning to give among the effective altruism community, and one of the main groups growing the community, it seems reasonable that some proportion of the impact is due to us. Moreover, as you can see from the studies in the Appendix, many of these people have been highly engaged with 80,000 hours over the last couple of years.

Beyond 2016, it’s harder to say, since it’s difficult to know how the growth of effective altruism in general will unfold relative to 80,000 Hours. If effective altruism grows rapidly independently of 80,000 Hours, then in the long-term these people would have probably have earned to give otherwise. (Similar to how if GiveWell hadn’t been so good at charity evaluation, Giving What We Can or another group would have invested more in research, and might have ended up recommending the same charities several years later).

Overall estimate of additional donations due to 80,000 Hours

Based on all these considerations, and the considerations in the first section, these are our overall subjective estimates:

• Over the past three years:
  • Mean estimate: £50k, £1.3k per person
  • 80% confidence interval: £10k – £100k, or £0.25k – £3k per person
• Over the next three years:
  • Mean estimate: £500k, or £13k per person
  • 80% confidence interval: £100k – £1mn, or £5k – £26k per person
• Annually beyond 2016:
  • 80% confidence interval: £20k – £1m, or £0.5k – £25k per person

What about donations to 80,000 Hours itself?

The group of 39 has donated about £100k to 80,000 Hours in the last three years, and can be expected to donate a similar amount annually going forward.

Conclusion

On average, people pursuing earning to give report they have already donated thousands of pounds to charity. Overall, it seems likely that the existing community of about 100 people pursuing earning to give will soon be donating over a million pounds per year to high impact charities. There’s evidence 80,000 Hours has caused a significant fraction of these donations.

Questions for further work

• We’ll continue to track people overtime, so we can measure how the donations compare to predictions, and improve our estimates of drop out rates.
• It could be helpful to do further analysis of the base rates for lifetime earnings to better understand the long-term trajectory.

Appendix-individual studies of career change

10. Anonymous

Situation before: Student. Considering a wide range of careers, with the top contenders being doctor and maths professor. Member of Giving What We Can, pledging 10% of his income to cost-effective charities.

How 80,000 Hours intervened: Through involvement with us and other groups, he started to consider earning to give or working directly in ‘effective altruist’ nonprofits. We gave him extensive feedback, advice and discussion of his plans, in person and in writing, weighing up the relevant arguments and providing relevant information, over the course of months. We encouraged him to pursue earning to give, due to his unusually large earning potential. Note that the decision was made before the official founding of 80,000 Hours, but involved our pre-founding team.

How his plans changed: He decided to pursue earning to give, in part due to 80,000 Hours and in part due to advice from others in CEA. He plans to lead his career on the basis of where he can make the most difference.

What he’s doing now: He is working for a proprietary trading firm. He was able to donate $40,000 in his first three months, and plans to donate upwards of $100,000 after his first full calendar year. He agrees that this was the right decision. He continues to receive research from one of our trustees on where to donate, and we’ve introduced him to several other members of 80,000 Hours. He donates to 80,000 Hours among other organisations. He has also helped to convince a recently retired Partner of his company to support 80,000 Hours financially.

He attributes 30% of his donations to 80,000 Hours’s pre-founding team.

11. Sam Bankman-Fried

Background MIT – Physics, graduates in 2014

What were they planning before? Student. Considering politics, journalism and academia. Highly concerned with making a difference.

How did their plans change? Became more in favor of earning to give, at least in the short-run, and more concerned about causes that benefit the long-run future.

What are they doing now? Completed an internship at a proprietary trading firm and intends to take up a job offer in the summer. Intends to donate all earnings he doesn’t need to live on and do well in his job. Final year at MIT. He plans to lead his career on the basis of where he can make the most difference.

How and why did the change come about? Sam was convinced through discussion with 80,000 Hours staff; introduction to the 80,000 Hours community; and lectures on the core concepts. Participating in the 80,000 Hours community contributed to finding his finance internship.

How much have they donated over the last three years? None

How much do they plan to donate over the next 3 years? ~100k, though ~$1m per year from 3 years time if he doesn’t burn out. He attributes 20% of these donations to 80,000 Hours.

12. Andrew Farmer

Background Major in Math at MIT.

What were they planning before? Identified as a utilitarian. Choosing between a career in software engineering or quantitative trading. Had job offers within both.

How did their plans change? Became more convinced of the importance of donations compared to the direct impact of working in software. Within finance, became more inclined to take the option with the highest discounted earnings, rather than the best lifestyle. Believes they’ve become more likely to stick to their altruistic aims.

What are they doing now? Decided to take a job in quantitative trading and donate over half of his income to GiveWell recommended charities and meta-charities. Highly concerned with choosing the career in which they can have the largest impact, including long-run effects.

How and why did the change come about? Was persuaded by the importance of donations compared to direct impact through discussion with a member of 80,000 Hours. Received career coaching and was given estimates of direct impact, which confirmed this impression. Was also encouraged to consider the discovery value, career capital and degree of fit with both options, and was told about crucial considerations around existential risk.

How much has he donated over the last 3 years? Zero

How much does he intend to donate over the next 3 years? £300k He attributes ~15% of his total impact to 80,000 Hours.

13. Adam Gleave

Background Studying computer science at the University of Cambridge, graduates in 2015.

What were they planning before? A career in finance or software.

How did their plans change, and what are they doing now? Became more inclined to donate a substantial fraction of their income to charity. Became more inclined towards finance in order to earn more. Plans to donate to GiveWell recommended charities, organisations aiming to mitigate existential risk, or effective altruist organisations. Has accepted an internship in quantitative trading.

How and why did the change come about? Found out about the arguments in favour of donating through 80,000 Hours and Giving What We Can. Now an active participant in the 80,000 Hours community in Cambridge.

How much do they plan to donate over the next 3 years? £50k, and substantially more in later years that if he continues in finance. He attributes 70% of future donations to 80,000 Hours.

14. Matt Gibb

Background: Oxford University – PhD in Computational Biology 2012

What were they planning before? Student. Planning to work in start-ups.

How did their plans change, and what are they doing now? Decided to donate 33% of his income, and the value of the equity of any start-ups he creates, to effective charities. Plans to lead his career on the basis of where he can make the most difference.

Now he’s the co-founder of a start-up, Dropkic.kr, which matches investors with crowdfunding projects. All of the founders have made a legally binding agreement to donate 33% of the proceeds of a future sale of the company.

How and why did the change come about? Matt came to an 80,000 Hours talk, was exposed to the idea of earning to give, and was immediately convinced. He started volunteering for 80,000 Hours as one of the founding 6 members.

How much have they donated over the last three years? £2-3k to AMF and SCI.

How much do they expect to donate over the next three years? £100k in expectation, though likely to be near zero.

For more, see this interview

15. William Denver

Background
William is working at an education nonprofit he co-founded, and recently completed a PhD in Maths a top-5-ranking US graduate school.

What were they planning before?
William was planning on doing a Maths postdoc, but was considering other options.

How did their plans change, and what are they doing now?
William ended up increasing his confidence that working in finance was a good option for him, he applied to a financial trading firm, did an internship with them, and now has two offers from top finance firms that he is choosing between.

How and why did the change come about?
Through one-on-one coaching with 80,000 Hours and a conversation with someone in the 80,000 Hours network who works at a financial trading firm. 80,000 Hours gave William more confidence that pursuing a finance path was right for him.

How much does he intend to donate over the next three years? About £94. He attributes £7k (7%) of this to 80,000 Hours.

In their own words:
“The reassurance of talking to people who were really serious about career advising meant I could focus on the most important aspects of my job search, and this in turn meant I had less doubt and more intent on what I was doing. I imagine this made it much more enjoyable than it would have been otherwise.

I actually wasn’t planning to change careers when I started talking to 80k; I was just applying to things for exploratory reasons, but I found a job that was just so awesome that I had to take it. I was so busy at the time, having 80k to narrow my search made it meant I actually had time to do my interviews. It’s possible I would have missed this opportunity if I’d been less focussed on my best options.”

16. Ramit Nehru

Background University of Chicago, Economics, worked as a quantitative financial analyst for several years.

What were they planning before? Choosing between doing a medical degree to go into medical research and maximising earnings in finance to do earning to give. Most likely to do medical research. Highly concerned with making a difference.

How did their plans change? Now most likely to take the earning to give route in finance. More likely to donate to meta-charities.

What are they doing now? Applying for roles in finance and medical research internships, while continuing as a quantitative financial analyst and studying medicine part time.

How and why did the change come about? During a case study with 80,000 Hours, he learned that: the earnings in finance are higher than he thought, in medical research it’s not uncommon to become stranded mid-career, there are reasons against thinking the medical research path is better than earning to give and that there are good donation opportunities within meta-charities.

Full write up of the case study is here.

How much has he donated over the past three years? £19,000 to his own nonprofit. He doesn’t attribute any of this to 80,000 Hours.

How much does he intend to donate over the next three years? £280,000 if he stays in his current job. He intends to give to GiveWell recommended charities and effective altruist organisations. If he switches into trading, he would donate roughly an extra £150,000. There’s roughly a 40% chance of this, of which 20% is due to 80,000 Hours. So, expected future donations are £340,000, of which £30,000 are due to 80,000 Hours.

17. Anonymous

We can’t share the details of this case, but they expect to donate £420k to GiveWell recommended charities in the next three years (though it’s likely to be much less). They attribute 25% of this to 80,000 Hours, or £105k.