A Framework for Thinking about the EA Labor Market
Note: Views mine, not my employer’s. Thanks to David Reinstein, Peter Hurford, and others who provided helpful feedback; any errors are mine alone.
Since 2015, the EA community has increasingly discussed talent constraints. That discussion generally hasn’t taken place in the language often used to discuss labor markets, that of labor economics. I argue that standard frameworks like labor supply and demand models can provide considerable insight into why EA organizations and the broader EA community experience shortages (and surpluses) of various skills and how to resolve these imbalances.
Through an economics lens, organizations that offer non-competitive compensation should generally expect to find it hard to hire the best candidates. It’s well known that the nonprofit sector as a whole pays significantly less than the for-profit sector, but there isn’t much hard data about how compensation at EA organizations compares to what different types of candidates could alternatively earn. I propose surveying EA organizations to get actionable information about their competitiveness, and using that data to guide future decisions.
Defining Talent Constraints
When we talk about “talent constraints” in EA, what do we mean by the term?
80K has offered a working definition of “talent constrained”:
“An organization is talent constrained when, for someone who could take (a reasonably important) job at that organization, they would typically contribute more to that organization by taking the job than earning to give.”
They go on to note that this definition is imprecise and somewhat problematic:
“While we think this framing can sometimes be useful, it also has some problems. For example, this definition seems less useful when an organization’s best potential hires don’t have very high earning potential and wouldn’t be very good funders.”
This imprecision has the potential to create ambiguity and confusion. I propose an alternative definition:
An organization is talent constrained when it doesn’t have (and/or can’t hire reasonably easily) the people it needs despite offering competitive compensation.
This definition (which I’ll use for the remainder of this post) may not be perfect, but it has the advantages of being simple and applicable across organizations, causes, roles, and geographies. Perhaps most importantly, it’s an intuitive definition: if someone hears the term “talent constrained” it will likely conjure up images of an organization that doesn’t have the talent it needs, rather than an organization where (some) potential supporters would be more valuable as employees than donors.
The definition also works in describing more specific skill gaps if one simply states explicitly which skills one is referring to. In general, I suggest people follow 80K’s advice regarding specificity: “It’s nearly always clearer to talk about the specific needs… ideally down to the level of specific profiles of people, rather than talent… in general.”
The importance of competitive compensation
Through the lens of labor economics, which “looks at the suppliers of labor services (workers) and the demanders of labor services (employers), and attempts to understand the resulting pattern of wages, employment, and income”, a talent constrained organization should consider raising salaries. When wages are structurally suppressed, through mechanisms like wage ceilings that limit how much people are paid, labor shortages generally result.
It’s widely understood that the nonprofit sector pays significantly less overall than the for-profit sector, and “much of the literature on pay in nonprofits” addresses this discrepancy. This means that nonprofits routinely lose out on the talent they need because their salaries can’t compete with for-profit firms, as Dan Pallotta argued in his 2013 TED Talk:
“The median compensation for a Stanford MBA, with bonus, at the age of 38, was 400,000 dollars. Meanwhile, for the same year, the average salary for the CEO of a $5 million-plus medical charity in the U.S. was 232,000 dollars, and for a hunger charity, 84,000 dollars. Now, there’s no way you’re going to get a lot of people with $400,000 talent to make a $316,000 sacrifice every year to become the CEO of a hunger charity.”
Exacerbating the problem, while for-profit companies routinely reward the employees who have the largest impact on the bottom line, nonprofit compensation rarely incentivizes employees to achieve more social impact. An analysis of the financial records of over 27,000 US nonprofits found “the pay of nonprofit chief executive officers (CEOs) is strongly predicated on what managers in similar-sized organizations receive… [and] that nonprofit executive compensation is only modestly affected by CEO performance, as measured either by improved fund-raising results or better administrative efficiency.” These dynamics suggests there could be a large pool of talented workers who would be willing to work in the nonprofit sector if the compensation weren’t so low and so misaligned with performance.
Turning to an example from the EA community (where I suspect organizations generally pay less than what comparable nonprofits do), The Centre for Effective Altruism is hiring a new CEO. Should it restrict its search to candidates willing to show their commitment by pledging everything they earn above a modest amount to effective charities? (Purely hypothetical question, I have no reason to think CEA is doing this).
That approach would definitely produce mission aligned candidates. But I think it would be foolish given the extraordinary importance of getting the best possible person in that high-leverage position. What if that person won’t (or can’t) make the kind of sacrifice Pallotta describes? Is it worth losing out on the enormous incremental impact the person could produce?
How competitive are EA organizations?
To gauge whether an EA organization is competitive or not, we can benchmark its salaries against what candidates could earn on the open market, at other nonprofits, or at other EA organizations. These comparisons help define a candidate’s opportunity cost (i.e. what they would forgo by working in the EA ecosystem instead of pursuing other opportunities). And, as the following simplified flowchart shows, they can also help define why talent constraints exist and what can be done about them. (If you’re not sure how to follow the flow chart, there is a guided example in this footnote).
EA labor market supply
Increasing compensation can often help attract talent. But I believe many EAs are resistant to raising salaries as a way to close talent gaps because they conflate willingness to work for low pay with fit for a job. One respondent to the 2017 EA Leaders Survey was explicit about this: “Raising salaries… is unlikely to be very helpful for attracting top talent since the most suitable candidates are also the most altruistic ones.”
While the “most altruistic” candidates might be the “most suitable” all else being equal, for practical purposes all else is never equal. As Holden Karnofsky observed about GiveWell’s experience, “it’s certainly the case that some people require substantially more pay than others (based on different career stages, etc.) even when they buy heavily into the mission.” In other words, willingness to work for below market rate is a poor proxy for both a candidate’s altruism and their job fit.
Talent constraints get exacerbated when we (in Pallotta’s words) “confuse morality with frugality”. Labor supply curves shift inward, reducing the quantity of labor supplied, when candidates have relatively higher earning power in other industries and when barriers to entry are in place (like cultural expectations of deeply discounted wage scales).
Lower salaries don’t just reduce the overall supply of labor, they also skew which types of candidates end up working, or not working, in EA roles. EA leaders who were surveyed about talent in 2018 think the community’s biggest skill gaps are in areas that require significant experience like expertise in government, policy, or AI, or skills in operations or management. These labor shortages are in sharp contrast to the job market for EA roles that require less experience, where even highly credentialed and mission aligned young candidates are finding that it is really really hard to get hired by an EA organization.
The EA community’s youthful demographics no doubt explain much of this discrepancy, and related factors like network and founder effects likely play significant roles as well. These issues make it all the more important that EA organizations reconsider other practices like low salaries that exacerbate these imbalances.
Low salaries make it relatively harder to find experienced candidates than inexperienced ones because of several factors that shift the labor supply curves on a relative basis:
● Earning power in alternative jobs. Experienced candidates generally have better paying alternatives than their inexperienced counterparts (i.e. they have higher opportunity costs). Experienced candidates will often be sacrificing hundreds of thousands of dollars or more to work for an EA organization; that would be very rare for junior candidates.
● Barriers to entry: Experienced candidates are more likely to have dependents and mortgages, and for them working at an EA organization is more likely to involve a psychologically difficult large pay cut. Inexperienced candidates, on the other hand, may see working in EA as the path of least resistance.
● Non-monetary compensation. As Milan Griffes has argued, EA jobs provide scarce non-monetary goods like “social status, life-orientation, a sense of having near-maximal impact, and being part of a value-aligned, elite tribe.” Younger candidates likely perceive more value from these particular factors. (More experienced candidates, by contrast, would likely perceive relatively more value on non-monetary compensation in the form of flexible working hours or paid parental leave).
Learning more about EA’s talent gaps and how to solve them
As 80K puts it, “Skill bottlenecks are a matter of degree” and these degrees can vary significantly depending on the specific skills in question. But we have little hard data to quantify skill gaps across various areas.
I suggest adding new questions to the next talent survey of EA organizations to help capture some of these nuances. These questions will hopefully make it easier to answer questions like: Are EA organizations talent constrained? If so, which sorts of organizations and which sorts of talent? How large are these constraints? What can be done about them?
New questions (which should continue past surveys’ practice of asking the same questions about both a junior and senior hire and anonymizing organizational responses due to the sensitivity of the information involved):
● Generally speaking, how easy/difficult do you currently find it to fill roles at your organization? (Scale of 1 = Very difficult to 5 = Very easy)
● What do you pay current employees relative to what they could earn on open market, including jobs in the for-profit sector? (Multiple choice: More; about the same; 0-10% less; 11-20% less, etc)
● What do you pay current employees relative to what they could otherwise earn in the nonprofit sector (including all nonprofits not just EA organizations)? (Multiple choice: More; about the same; 0-10% less; 11-20% less, etc)
● What do you pay current employees relative to what they could otherwise earn at another EA organization? (Multiple choice: More; about the same; 0-10% less; 11-20% less, etc)
● How much do you plan to offer future hires relative to current employees in similar roles? (Multiple choice: More; Less; About the same. If someone answers “more”: Do you plan to increase salaries for existing employees? What factors into this decision?
● Would any of the following steps be helpful in closing your organization’s talent gaps? (Rate the following options on a Scale of 1 =Not at all helpful to 5 = Extremely helpful):
○ Increasing salaries
○ Investing in recruiting
○ Increased publicity of position
○ Better access to interested candidates
○ Other (please describe)
This information is relatively easy to collect and interpret, corresponds to a real-world decision organizations make (how much to pay), is grounded in observable data (market wages), and is comparable across roles, organizations, geographies, and causes. To mitigate the cost of data collection, I suggest abandoning survey questions about non-traditional labor metrics that seem to be producing noisy data and to generally be causing confusion. If we collect compensation data and respondents indicate which types of roles they’re thinking about as junior and senior (which will vary across organizations), we’ll have a rich and actionable picture how EA compensation stacks up to the competition for various types of skills.
My hunch is that for the most part talent constraints are, and will continue to be, common among EA organizations. The world’s biggest problems require significant and diverse talent, especially as existing organizations continue to grow and new ones emerge. And if EA organizations offer compensation that is as non-competitive as most nonprofits, they should expect chronic struggles attracting the candidates they need (most notably in roles requiring specialized expertise).
To solve this problem, I think EA organizations need to approach compensation in a way that’s oriented around impact rather than convention, even if it sometimes means paying more than most nonprofits would for a role. I suspect it would be very helpful to strategically and opportunistically increase compensation to try to tap a vein of skilled candidates that are willing to work in the nonprofit sector for a discount, but not as big a discount as most nonprofits might expect.  This might require a shift in mindset from viewing talent gaps and funding gaps as oppositional to seeing the latter as a frequent cause of the former. Increasing compensation won’t be a cure all (for an excellent overview of why there can sometimes be practical difficulties in trading money for capacity, see GiveWell’s 2013 blog post We can’t (simply) buy capacity), but my hunch is it would be a helpful first step.
However, there’s no need for the EA community to rely on my intuition or anyone else’s when there’s hard data we can use to guide our decisions. I propose surveying EA employers to see how they pay relative to the nonprofits and for-profits they compete with for talent. This would radically improve our understanding of how much talent EA organizations want at different price levels; in economics terms, we’d be learning what the labor demand curve looks like. I also enthusiastically support proposals to use the EA survey to learn more about EA job seekers, which would improve our understanding of the corresponding labor supply curve.
If we think about the EA labor market in terms of labor supply and demand, we’ll be able to leverage the efforts of many very smart people who have already thought about a lot of the issues we care about. They can help us make informed decisions about crucial questions that will ultimately shape much of the EA labor market: How much should EA organizations compete with the for-profit sector for top talent? How can they fund the talent they need (and other operational expenses)? And if they can’t or won’t become more competitive with for-profit firms, how do EA organizations expect to find the talent they need to achieve their ambitious goals?
I welcome thoughts on how to better calibrate and operationalize “can’t hire reasonably easily” in the most meaningful way and/or ideas for alternative definitions. ↩︎
While I disagree with 80K’s definition of talent constraints and some of their other thinking on the issue, I want to emphasize that this constitutes a very minor critique. 80K has done an excellent job drawing attention to the importance of the EA labor market, and has helped a lot of people achieve much more impact with their human capital than they otherwise would. 80K also deserves applause for recognizing that some of its content was causing confusion and publishing a clarification on their thinking (which includes the definition I’ve quoted). ↩︎
Disclaimer: I’m not an economist. I majored in economics in college and was a research assistant for a few professors, though that was nearly two decades ago. I was admittedly fuzzy on some details so I watched a few refresher videos. I highly recommend EconplusDal’s videos (which have millions of views) and have tried to lean on his expertise where possible. For instance, when I discuss factors that drive shifts in the labor supply curve, the factors I discuss are ones he emphasizes in his video on the topic. I also got very helpful feedback from an economist on an early draft. ↩︎
This was almost certainly true in 2015 when EA organizations paid less and talent gaps first started being talked about. ↩︎
Starting in the upper left hand corner, the first question to ask is whether pay is competitive vs. the for-profit sector. If so, follow the “Yes” arrow to arrive at a diagnosis of a recruiting problem that suggests a need to invest more in recruiting. If that suggestion worked (“it worked”), the talent constraint is resolved. If “it didn’t work”, then the suggestion is to try investing even more in recruiting. If there’s no money available, we follow the “need $” arrow to arrive at a new diagnosis of a funding problem. ↩︎
80K’s most recent advice on this issue is: “In general, if an organization thinks that it’s constrained by a particular role, we think they should seriously consider raising salaries for that role. However, we don’t think this will fully solve the problem, and it isn’t always the right thing to do.” ↩︎
To be clear, the literature on nonprofit compensation includes significant discussion of how the “morality vs. frugality tradeoff” impacts labor supply under different conditions. For example, this article synthesizes empirical and theoretical findings about the nonprofit wage gap: Hallock, K. (2000). Compensation in nonprofit organizations [Electronic version]. Research in Personnel and Human Resources Management, 19, 243-294. http://digitalcommons.ilr.cornell.edu/hrpubs/18/ ↩︎
A shortage of experienced talent seems clearly problematic. One could argue a hyper-competitive job market is a positive thing, but anecdotes like this are obviously worrisome: “I’ve recently graduated from one of the top ~10 universities worldwide, after investing heavily in EA throughout my studies. While a student, EA was the biggest thing in my life… Over the last seven months, I’ve made over 20 unsuccessful job applications (I keep a spreadsheet). This has increased the severity of my depression and anxiety. Over time, I began to shed my identity as an EA, no doubt as a self-defence mechanism.” ↩︎
I’d also suggest surveying a broader set of organizations. The latest survey included some changes that partially addressed concerns that have been raised about the potential for serious selection bias, but much more could be done in this direction. That said, I recognize the EA community has some experts on survey design (Rethink Charity has done a terrific job with the EA Survey for instance), and would welcome (and generally defer to) their suggestions. ↩︎
The 2017 EA Leaders Survey asked respondents “If you tried to spend more money on attracting talent what might you do?” While most respondents provided ideas (typically raising salaries and/or investing in recruiting), about ⅓ noted concerns that a higher salary for a new hire would require raising salaries for existing staff to avoid resentment. My suggested new question intends to get more explicit and quantified information about the extent to which this dynamic is shaping EA hiring decisions and organizational growth trajectories. ↩︎
To make these comparisons, it’s generally helpful to discuss salaries via a certain sized discount (premium) to market, in percentage terms. For example one might say: “Organization X pays senior developers 20% more than what other nonprofits pay, but that’s a 30% discount to the open market so they’re having trouble hiring.” If that message were expressed in dollar terms (rather than percentages), it’d be less comparable to analogous messages about different roles, geographies, etc. ↩︎
Asking organizations how much financial value they place on their last hire appears particularly confusing and problematic. I’d also suggest discontinuing the “discount rate” approach to measuring hiring urgency since it is yielding results that are too noisy to act on. Instead, I’d ask organizations to directly rate how urgent they find hiring on a Likert scale. ↩︎
Here’s a simplified example of my intuition on the right strategy for the EA ecosystem to adopt. Imagine an EA organization is trying to make an important hire with skills that are highly desired by the for-profit sector. If good candidates could earn $75,000/year elsewhere in the nonprofit sector, and $150,000 in the for-profit sector, initially offering in the neighborhood of $100,000 could be a good strategy. At that salary, the EA charity would be extremely attractive to candidates strongly leaning toward nonprofit work. And while candidates would still need to take a substantial $50,000/year discount relative to what they could earn in the for-profit sector, that’s a much smaller discount than other nonprofits would require. So offering $100,000 could attract a lot of people who are on the fence about nonprofit and for-profit work (perhaps for reasons that have nothing to do with their mission alignment or job fit). If we change the example to someone who could earn $1 million (or more) in the for-profit sector, my intuition is much weaker. In principle, I don’t have any objection to nonprofits paying key staff hundreds of thousands of dollars a year if that makes them more impactful overall. ↩︎
Imagine a parallel universe where 80K’s original insights about talent gaps were released with a title and framing of “Why talent gaps may be more important than you think” instead of the “Why you should focus more on talent gaps, not funding gaps”. I think the EA community would be more impactful and on a better trajectory in that universe. For what it’s worth, I agree with 80K’s clarification that “it’s possible for an organization to be mainly constrained by a certain skill, or funding, or both, or neither”. But I don’t think their headline summary of the issue (“skill and funding constraints are not opposites”) captures the essence of the relationship which I see as “funding gaps are a frequent cause of talent gaps, particularly in sectors with structurally non-competitive salaries”. ↩︎
I’m wary of theories that assume “EA exceptionalism.” EA organizations aren’t the only ones staffed by people who believe they are working on the most important cause, and are sacrificing earning power to do so. ↩︎