Summary / Key Takeaways

Over the summer, about a dozen very cool policy professionals shared their takes on impactful US policy careers. The sessions were not recorded, so—in hopes that others find the content as useful as I did—here’s Part 1 of what I learned (companion post here). My main personal takeaways:

• Being absurdly impactful through US policy work is very feasible, and not even very rare in this community. Anecdotes suggest that, if more people with certain strengths (discussed below) go into US policy work, it’s fairly likely that—over the course of their whole careers—they’ll have multiple huge impact wins (See more)
  • Supporting anecdotes (I lack exact numbers): A fairly small number of impact-driven people have gone into US policy to work on some issues of concern to this community. After just a few years of US policy work, a large portion of these people can make compelling cases that they’ve achieved huge wins (e.g. equivalent to getting billions of dollars to causes that many in this community consider high priorities). Also, a large portion have gained fantastic career capital.
  • In other words, good fits have a decent-to-high chance of achieving multiple big wins over their US policy careers. Rather than just hoping to eventually be at the right time and place, they repeatedly put themselves into the right time and place.
• Big impact wins typically require identifying/creating, and taking, non-obvious opportunities. (See more)
  • This means there’s lots of room for counterfactual wins.
  • To have these wins, it’s critical to often ask yourself, “What would be good to see in the world that probably won’t happen by default / if I don’t make it happen? How can I make it happen?” (And then do it!)
  • In fast-paced jobs, you won’t have the time or attention for that unless you very intentionally make sure you do, e.g. by blocking off some time each day or week for focused work on big picture strategy, actively looking for and brainstorming opportunities for impact, and more generally doing what’s most impactful.
• Big impact wins are hard to predict and relationship-driven, so building a broad network is very useful. (See more)
• Powerful people in government are typically far too busy to turn vague proposals into concrete policy, and they only have the authority to pull certain levers. So policy researchers and advocates will be much more impactful if they do the following (See more):
  • Create very specific, shovel-ready recommendations that some individual(s) in government have the authority to implement. (See here for an example.)
  • And/or present such recommendations to people who have authority to implement them.
• Things that make someone a good fit for high-impact policy work in general include the following (See more):
  • Strong communications / people skills
    • (Extraversion isn’t a general requirement, although it does help)
  • The ability to work well with people who have very different views than you
  • Analytical skills
  • Being fine with (and good at!) having most of your impact come from occasionally finding/creating non-obvious opportunities to make good things happen
• (Part 2 here.)

Additional Takeaways

• Potentially promising policy jobs (and some skills that seem especially important for being highly positively impactful in these jobs) include:
  • Congressional staffing (See more: [1], [2], [3])
  • Think tank work (See more: [1], [2], [3])
  • Executive branch work (See more: [1], [2])
  • Policy journalism (Writing skills and prudence seem especially important.) (See more: [1])
  • Political data analytics (Statistics skills and entrepreneurship seem especially important.) (See more: [1], [2])
  • Lobbying / advocacy (Experience and cautiousness seem especially important.) (See more: [1], [2])
  • Running ballot initiative campaigns (Entrepreneurship and cautiousness seem especially important.) (See more: [1], [2])
  • Running for office (Social skills, personal background, and cautiousness seem especially important.) (See more: [1], [2], [3])
  • Working as a lawyer (Entrepreneurship seems especially important.) (See more: Part 2)
• If you want Congress to do something, you should get it into a big spending bill or an obscure bipartisan bill, since these are usually the only substantive bills that Congress passes. (See more)
• Broadly, think tanks have two main paths to influence (See more):
  • Taking broader ideas and connecting them to very specific things that specific authorities can actually implement
  • Shifting broader thinking/conversations/narratives
• Much, perhaps most, of the impact of policy journalism comes—not from having general audiences read an article—but from having a few relevant policy makers read it. (See more)
• Day-to-day policy work often consists of lots of meetings and emails. (See more)

Context

This is Part 1 of a 2-part series; the companion post is here.

• Part 1 (this part) mostly gives descriptive context about US policy careers (as well as some advice that is difficult to separate from these descriptions).
• Part 2 is fully focused on sharing advice.

Most of the information here is my personal takeaways from what was shared in this summer’s US policy careers speaker series. To respect speakers’ privacy and fill in a little context, I’ve removed their personally identifying details and mixed in some information from other sources.

Discussion focuses on US policy careers, but I’d guess there are significant similarities to policy careers in other countries.

Caveats:

• Panelists in the speaker series are not necessarily affiliated with this forum or related organizations, communities, and ideas.
• This is extremely far from being comprehensive.
• It’s very possible that I misheard/misunderstood some things.
  • If something here is critical for a career or advocacy decision, please get it at least double-checked before you use the information.

I’m thankful to the speakers and my co-organizer for sharing their time and experience, as well as to Fiona Pollack, Felipe Calero, and especially for help with note-taking and comments.

Insights into how the policy ecosystem works, how policy professionals can have impact, and personal fit

The policy ecosystem (government + adjacent organizations)

General overview:

We’ll start off with a few points of broad context, and what they mean for doing impactful policy work.

The US federal government is made up of many different individuals and agencies; policy design and advocacy will tend to be much more impactful if it’s informed by relevant actors’ unique interests and authorities.

The US federal government is big and complicated. The executive branch of the government is especially big and complicated. Here’s a relatively simple chart of how the government as a whole is structured (source):

Don’t worry about memorizing all of this now, but for a taste of the complexity, here’s a more detailed chart (source):

And as if that’s not enough complexity, the 15 executive departments on here all have their own agencies.

Why so many agencies? Largely because doing all the things the US government likes to do is a ton of work; someone has to deliver mail, negotiate treaties, build spacecraft, train soldiers, collect taxes, persecute interstate criminals, manage federal land, help fund schools, and so on.

Different agencies tend to have different interests, as well as different authorities (things they can do, levers they can pull). It’s not just that the different branches work very differently from each other; so do different committees and executive departments. For example, the House Science committee and an intelligence committee are interested in very different aspects of AI, and what they’re able to do about it is very different. Even different agencies within the same executive department can be very different from each other.

Understanding these differences within government is very helpful for doing impactful advocacy. If you want someone to fund pandemic preparedness, you probably won’t get much done by calling the Office of Science and Technology Policy; they aren’t in charge of much funding. And if you want someone to fund AI interpretability research, how much you emphasize its relevance to algorithmic bias vs robustness should probably depend on whether you’re talking to the National Science Foundation or the Defense Advanced Research Projects Agency.

The government interacts with many non-governmental agencies, so there are many ways to influence government policy without being a government employee.

The US government also regularly interacts with a bunch of organizations outside of government. These organizations help fill gaps that government employees aren’t able to fill, due to e.g. government employees’ limited time, expertise, and authority. For example, think tanks are places for slower, more deliberate thinking than policymakers have time for, and they’re ways people can get jobs when the politicians they support are out of power. Other government-adjacent organizations include advisory bodies, philanthropic foundations, more technical research institutes, and contractors.

People with power tend to be super busy; this explains a lot.

People with power tend to be super busy. As one speaker put it, they have much bigger portfolios and more crises than they have time to think about. This seems to be a central fact of much US policy making, in the sense that it seems to contribute heavily to many other trends:

• Relationships heavily influence policy professionals’ (policy + hiring) decisions.
  • If you need to know many things, but you don’t have time to figure them out yourself, then it becomes necessary to rely on others for doing much of that information collection and analysis—about whom to trust, whom to hire, what to pay attention to, what decisions to make, and so on.
  • The more you’re tight on time, the more you need to rely on relationships for informing decisions.
• Policymakers make myopic decisions.
  • By default, long-term planning tends to be a relatively low priority for many reasons. So if agencies have to frugally budget time, long-term planning is easy to cut first.
• Lobbyists are highly influential.
  • Policymakers largely have to rely heavily on others for policy information and ideas, and lobbyists deliver on these.
• Simple ideas and narratives are often highly influential.
  • If policymakers don’t have much time to grasp complex ideas, nuanced narratives, and what they mean for policy decisions, they’re left making decisions based on simpler ideas.
• Only clear, short summaries at the beginnings of reports have a decent chance of being read by policymakers.
  • Anything else would take too long.

(Should policymakers expand their staff capacity, to spread out their work? Probably, but they haven’t done that recently. Maybe they’re too busy.)

Now that we have more context on the policy ecosystem, let’s take a closer look at some specific parts of it, as well as some activities that influence the government.

Congress:

• People usually think of working in Congress as working for a Congress member or committee. But it’s also possible to work for some other government agencies that are heavily involved in the legislative process: Congressional Research Service (CRS), Government Accountability Office (GAO), Congressional Budget Office (CBO), legislative affairs offices throughout the executive branch.

  • For details, see footnote.^[1]

• On political polarization:

  • A staffer explains that partisanship is what usually kills people—people get burned out on wanting to accomplish something and having that fail (sometimes at the last moment) for frustrating reasons of partisan politics.^[2]

    • A related upside is that sometimes things go surprisingly well, e.g. a bill that seemed like it had died gets picked up later.

  • At the staff level, it’s not as partisan as it appears from the outside.

    • There’s lots of showmanship and aggressive partisanships for the cameras
    • There’s plenty of bipartisan friendships, even some bipartisan marriages

  • See Matt Yglesias’ article on Secret Congress.

  • Media coverage makes people feel locked in and unable to change positions.

    • But less so with more obscure, impactful legislation.

  • Ideas being labeled as Dem or Rep ideas often has little to do with ideology and much to do with which members introduced/advocated for them.

  • Congressional support agencies are strictly non-partisan.

• Many staffers leave / burn out after a few years, but a significant minority sticks around for a long time.

• What are the different roles in a Congressional office? See footnote.^[3]

• The big ideas tend to start from lobbyists or the White House, not Congress.

• According to a staffer, the show Veep is pretty spot-on in terms of “how absurdist the process can be," "West Wing is not... it's quite idealistic," and Miss Sloane is “realistic on lobbying.”

Executive branch:

The executive branch is where most of the money and people in government are. The Department of Defense is a very big part of it.^[4]

It’s a very diverse place, with many “little kingdoms.” So getting things done often requires much interagency cooperation, through e.g. budgetary influence, charm, authority, political appointees in charge and alliances between them, political capital, other connections, etc. As one government official put it, big interagency coordinations have “complicated success, if they have success at all.”^[5]

Here are just a few of the many spheres in the executive branch. They’re relevant to issues of concern to this community, and they illustrate the differences between different parts of government.

• The national security (NatSec) space:

  • This is the part of the government that focuses on things like war and counterterrorism. So this community is relevant to issues in areas including great power conflict and preventing bad actors from causing catastrophes.
  • Agencies in this space include:
    • The Department of Defense (i.e. the Pentagon), which is huge and coordinates national security activities
    • The National Security Council, which advises the president on national security issues
    • The Intelligence Community (discussed below), much of which is in the Department of Defense
  • Working in this space typically requires you to have a security clearance (see this section in Part 2 on tips for that).
  • Some parts of this space seem to be known for being especially fast-paced and myopic, because they’re so busy dealing with immediate crises and supporting day-to-day decisions (especially those of the president).^[6] Some parts seem to have reputations for hawkishness and zero-sum thinking. But there’s much variety; other parts are different.

• The Intelligence Community (IC):

  • This is the part of the government that collects and analyzes information of potential relevance to national security. They also advise military planning. So this community is relevant to issues in areas including international cooperation and preventing bad actors from causing catastrophes.
  • It’s a community of ~17 agencies, including the CIA, the FBI, and the intelligence agencies of most branches of the US military.

• The R&D ecosystem of government:

  • This is the part of the government that funds, and in some cases carries out, R&D (research and development). It’s especially relevant for issues such as funding research that advances alternative proteins or helps make emerging technologies safe and beneficial.
  • For a list of major agencies in this space, see footnote.^[7]
    • Note that there’s much overlap between the R&D ecosystem and the national security space—about half of the R&D budget is in the Department of Defense, which funds a mix of applied and basic research.
  • Culturally, some parts of this space are relatively slow-paced, academic, and less focused on national security concerns.

Agencies outside of the White House matter; a government official recounts that many decisions in the White House are deliberated on in (and had most important aspects of decided in) sub-agencies.

Think tanks:

One speaker mentioned that the think tank where they work uses reports mostly as excuses to get meetings with policymakers. Policymakers typically don’t have time to read long reports, so they get briefed in these meetings. The think tank and its employees use their relationships with people who work with the policymaker to get these meetings.

Ballot initiative campaigns:

Some US states and cities let people run ballot initiatives: people can propose legislative changes, and if these proposals get enough signatures from people in that jurisdiction (and are within the scope of what ballot initiatives are allowed to do in that jurisdiction), they go on the ballot for voters to vote on in the next election. Then, if over 50% of voters approve of a measure, it passes.

Getting a ballot initiative on voter’s ballots is mostly a matter of money, according to people with lots of relevant experience. If you pay enough people to collect signatures, you’ll get enough signatures.

Currently, long-term-oriented work in running ballot initiatives and related activities is very young, so it involves lots of trying new things and trying to think of new (adaptations of) ideas.

Running for federal office:

Notes from a professional with relevant experience:

• Politics is an environment in which trust is low and public visibility is high.
  • (This makes high-trust communities well-positioned to work together well.)
• Politicians are mainly filtered to have passion for and conformity toward their party’s orthodoxy.
  • Election processes also filter for politicians who are power-hungry and good at fundraising.

Lobbying:

Roughly, lobbying means asking government officials to make or change government policies.^[8]

Paths to impact & related advice

Broadly, US policy work can be very impactful because the US federal government moves lots of money (several trillion dollars each year), and because it’s the only organization with the authority to do certain things.

Before diving into the specifics of certain policy jobs, I’ll start with two more generally applicable takeaways.

Big impact wins are not that rare, for people who are great fits.

_I was very pleasantly surprised by how often speakers have achieved big impact wins, and by how big some of these wins have been. _

It seems to be a common idea in the community that policy work, especially if it’s focused on emerging technologies, is high-risk. For example, the 80,000 Hours article on US AI policy work claims:

Your chances of having a large social impact by working in AI policy is highly uncertain. You would be intentionally taking a big, low probability bet, not just with regard to your own contribution (which is generally uncertain in any cause), but also with regard to the overall case that this area is promising to work in.

Hearing from the speakers makes me more skeptical of that first claim—that even if we assume AI policy or biosecurity policy are promising to work in, an individuals’ chances of making a contribution are slim. Speakers’ experiences seem to suggest that long-term-oriented US policy work is often high-risk in the sense that one may not have impact for a few months or years. But over a whole career, people who are (very) strong fits seem likely to have numerous huge wins.

(A speaker notes: Impact is highly dependent on fit. If someone is an OK fit—if they’re the kind of person who would survive in government—the expected value of their work in government is some value. If someone is an excellent fit—if they’re the kind of person who would thrive and be a policy entrepreneur, figuring out how to create high counterfactual marginal expected value and executing on it—then the expected value of their work is much higher.)

Why do I think this?

• In just a few years, at least a handful of people who are very concerned with improving the impacts of emerging technologies have reached positions in the US government from which they are well-placed to contribute.

• We heard from three speakers in the executive branch; of the two who discussed their impact so far:

  • One who’s been working on this for less than a year said they don’t think they’ve had much direct impact yet.
  • One who’s been in the executive branch for several years says there've been 6 or so moments that would have easily justified their career in government, with more big wins likely to come soon.^[9]

• Another speaker, over the course of just a few months, very plausibly counterfactually secured billions of dollars in government spending for a cause most of us would agree is very important. (And this speaker wasn’t even selected to give a talk based on this—I didn’t know it when I reached out to them.)

• Another speaker made a case that they have a nontrivial chance of having swung a very influential election. (And this speaker wasn’t even selected to give a talk based on this—I didn’t know it when I reached out to them.)

• Even if you have a career in policy that doesn’t have much direct impact, there’ll likely be lots of room for you to use policy experience and connections to advance the efforts of other very thoughtful and altruistic people.

  • The low number of people who are doing that seems to be a big bottleneck on good things happening in this space.
    • The community’s funds have grown a lot; there’s much need for expertise in many domains (including policy and politics) for turning funds into action that helps address pressing problems.

Caveat: there were significant selection and attention effects going on—I was disproportionately likely to hear from, hear about, and remember people with unusually impactful US policy careers (that is, unusually impactful even for people in this community, which is already trying to be unusually impactful). But that caveat is mainly applicable to the first two bullet points above; the others still make a strong case.

How do people have these big wins? That brings us to our next point:

Big impact wins require taking the time to look for non-obvious opportunities:

Big impact wins typically require identifying/creating, and taking, non-obvious opportunities. To have these wins, it’s critical to often ask yourself, “What would be good to see in the world that probably wouldn’t happen by default / if I weren’t there? How can I make it happen?” In fast-paced jobs, you won’t have the time or attention for that unless you very intentionally make sure you do, e.g. by blocking off some sacred time in a day or week for it.

Speakers from a wide range of policy backgrounds—the executive branch, Congress, think tanks, journalism, advocacy, running for office—seemed to generally agree on how counterfactual impact in policy works: most of the time, your work has little to no value relative to what would have happened without you, and occasionally you find/create opportunities to do extremely valuable things.

Those occasional high-value moves aren’t obvious; they won’t smack you in the face, as an impact-driven government official put it. So they require a sense of what would be good to see in the world that probably wouldn’t happen by default / if you weren’t there.

There may be high-value moves both within your job and outside of it.

In very fast-paced jobs, finding ways to pay time and attention to that is crucial for having impact wins—it’s very easy for all your attention to instead go to putting out fires.

(Note that the importance of this for impact doesn’t seem to be at all unique to US policy work; see, for example, a highly successful community builder’s discussion of a similar mindset.)

Big impact wins are hard to predict and relationship-driven, so building a broad network is useful:

It’s typically very difficult to predict in advance which relationships will open high-impact opportunities,^[10] so building a broad professional network—making many bets, in a sense—is often useful.

Having an impact in the executive branch:

Ways of having positive impact in the executive branch include:

• Improving the allocation of funding^[11]

• Improving the allocation of attention (e.g. by raising important issues in meetings)

• Improving regulations

• Improving US interactions with other countries

• Improving who works at government (e.g. through hiring and supporting others’ careers)

  • Lots of the outcome of something like the Cuban missile crisis comes down to having people in the room who can discuss merits and risks well.
    • Getting lots of thoughtful/prudent in government makes it more likely that the person “drawn from the urn” for a crisis will make good decisions.

• Building career capital (connections, experience, reputation, etc)

For a few additional details, see footnote.^[12]

Having an impact in Congress:

• There’s plenty of evidence suggesting that—since elected officials are super busy—Congressional staffers end up having much influence on legislation. This is a big deal; Congress (among other things) allocates trillions of dollars in funding, so some staffers can move billions of dollars (within constraints).
• Working in Congress brings great career capital for policy-related work.

Adding to that:

Broadly, working in Congress can be a promising path to positive impact because Congressional action is necessary for some important policy changes. After all, Congress has a monopoly on some influential powers (e.g. the authority and resources to shift much funding, the authority to create new agencies). On top of that, Congressional action is harder to reverse than presidential action—the next administration can’t easily undo legislation. As a rule of thumb, big, bold, long-term change tends to require Congressional action.

There’s usually just two kinds of substantive bills that Congress passes. If you want to get something done, the best way is usually to get it into one of these kinds of bills:

• Big spending bills^[13]

• Obscure bipartisan bills

Many other bills get dropped “for stupid reasons” or are never pursued with dedication in the first place (they’re mostly for showmanship).

See footnote for additional details.^[14]

Having an impact in think tanks:

People in power have no time to think; the think tanks do it for them—they make the ideas and hand them to them. Broadly, think tanks have two paths to influence:

• Taking broader ideas and connecting them to very specific things that specific authorities can actually implement
• Shifting broader thinking/conversations/narratives

As an example of the latter, one think tank corrected the misconception that China was far outspending the US on AI spending. The think tank was pretty sure this wasn’t true, did research and produced a better estimate, and then this stopped being a talking point.

Because people in power are super busy, they won’t read your 50 or even 15-page report. They might read a 1-page executive summary in your report, or tell one of their staff to read your report. So it’s really valuable to them if you can take ideas and make them very concrete/shovel-ready/easy to execute on. (E.g. “spend this much money on these specific things with these criteria,” not “fund this broad thing”). To be able to do this well, it’s helpful to have a good understanding of which levers different pieces of government can pull.

Another point: as discussed above, impact is often hard to predict, so it’s often more impactful to brief lots of people about research findings than to try to identify and target the few most impactful briefings.

Having an impact in policy journalism:

Policy journalism has multiple paths to impact. Here are some that seem especially promising:

• Getting your articles read by a small number of relevant policy makers
  • Policy makers (and/or their staff) often read and are influenced by news media. This creates opportunities for both positively shaping their opinions (i.e. persuading them) as well as their allocation of attention (i.e. keeping certain issues on their radar).
  • There are well-documented cases of this being very impactful (especially at journalism outlets that are more well-established)—see this Vox article by Dylan Matthews for an overview.
  • Anecdotes suggest policymakers do hear and consider the case you write, even when that is hard to see (e.g. because hidden constraints keep policymakers from implementing your suggestions).
  • Like with some journalists’ more general target audiences, it’s useful to keep in mind that policy makers are often smart and sometimes ignorant about issue specifics.
• Having your articles nudge general public opinion (especially by shaping narratives)
• Building career capital:

  • Making connections, by getting to know people as sources^[15]

  • Learning a lot about how policy processes happen

Additional advice on doing impactful journalism:

• How do some impact-driven journalists choose what to write about? See footnote.^[16]
• Advice for writing about technical topics - see footnote.^[17]
• How to write about unintuitive ideas? See footnote.^[18]

Having an impact in political data analytics:

A political data scientist sees three main ways in which their work has been positively impactful:

• Helping politicians of one political party get elected:^[19]

  • This person helped advise many campaign ads used by one party.
  • They were able to identify a data-backed opportunity for improving ads’ persuasiveness that others hadn’t accounted for, and they helped persuade people to implement this improvement.
  • They think this has a significant chance of having swung that influential election.

• Thinking of new ideas: In collaboration with a few others, this data scientist drew on their expertise to think of new things they could help make happen.

  • In particular, they helped think of running certain ballot initiatives.
  • Since then, their experience in political analytics has helped them contribute good intuitions and models about what ballot initiatives would pass.

• Connections: This person’s job connected them to people who are very competent in policy work, which has been helpful for:

  • Drawing on these people for later campaigns.
  • Sometimes being able to influence, through discussion, what influential people do or talk about.

This data scientist notes that lots of value in their own career required certain choices that could easily have gone other ways, so they wouldn’t recommend political data analytics for most people for impact. Political data analytics work can involve less counterfactual impact, e.g. if it consists of making state-of-race predictions and sorting voter lists. For more details, see the section on personal fit for political data analysts.

Having an impact as an elected official:

How hard is it to win an election for federal office? I don’t know. Some relevant information:

• Drawing on some relevant statistics, a political data scientist guesses that, for certain people (including a nontrivial number of people reading this), chances of getting elected to Congress if they try are about 25%.

• An old 80,000 Hours article estimates, with significant guesswork, that Congressional staffers who are interested in becoming a Congress member have about a 1 in 40 chance of doing so,^[20] while interested top law school graduates have about a 1 in 10 chance.

• (These apparently different estimates are compatible; the political data scientist was talking about a reference class that the article wasn’t considering.)

• One policy professional shared an anecdote of meeting a Congress member who had “no charisma,” and they suggested that elected officials’ talents aren’t as insurmountable as one might guess.

Another point: if you win election to Congress the first time, this doesn’t just get you into office for that term; it also gives a massive boost to your chances of winning future elections. For the past several decades, reelection rates in the US Senate and House of Representatives have hovered around 90%. So the above odds aren’t just estimated odds of someone winning one term in office—they’re close to being estimated odds of someone winning a career as a Congress member.

If you manage to win federal elected office (e.g. become a Congress member), you’ll have plenty of room for counterfactual impact. This is because many people reading this would have both the personal assets and the political wiggle room to help solve pressing problems.

Personal assets:

A professional with relevant experience explains that, while there are plenty of good-hearted people in office, Congress is in short supply of people with certain qualities: people who are evidence minded, careful thinkers, who care about the consequences of their actions, and who have strong truth-seeking drives, as well as people who can prioritize time and effort based on considerations of scale, neglectedness, and tractability.

It also helps (for one’s ability to do better than the person who would otherwise hold the seat) that elected office filters heavily for qualities other than policy making competence; this leaves open plenty of room for improvement along competence dimensions.

Political wiggle room:

As discussed above, elected Congress members—especially those in safe states—are very safe in primaries and general elections. And it’s not like they’re barely scraping by in these elections; of the 435 general elections for US House of Representative seats in 2020, for example, only 77 were decided by margins of less than 10%.

Comfortable prospects of reelection give elected officials some wiggle room in what they do. For example, politicians can often choose whether to be allies or champions of causes. This wiggle room is especially available when it comes to obscure issues; as a political data scientist puts it, you won’t get primaried (defeated in a primary election) for spending political capital on some technocratic thing.

Plenty of politicians in fact do this (spend some political capital on obscure bills). For example, lots of Senators put political capital into the recent tech and manufacturing bill, which isn’t making many headlines. See Matt Yglesias’ discussion of “Secret Congress” for more discussion of these dynamics.

Having an impact through running ballot initiative campaigns:

As noted earlier, getting ballot initiatives on voter’s ballots mostly just requires paying enough signature collectors. That doesn’t mean impact-driven communities should put every ballot initiative they can afford onto ballots; there’s risks to watch out for, including risks of reputational damage, risks that a good-seeming proposal is actually harmful, risks that a good proposal will crowd out even better alternatives, and risks of motivating counter-initiatives (initiatives that try to reverse earlier initiatives).

Still, ballot initiative campaigns can be powerful levers for city-level and state-level change (there are no nationwide ballot initiatives). They’ve been very successful in advancing animal welfare, and they may soon help secure about $50 million in funding for pandemic preparedness research. Ballot initiatives may be promising tools (or inconvenient thorns?) for emerging technology safety research and governance more broadly, given that California and Massachusetts—two states that are home to many leading research universities and tech companies—allow ballot initiatives. It would be easy for such ballot initiatives to accidentally cause harm (e.g. by incentivizing companies to move to different states), so anyone interested in this would be wise to collaborate closely with experts who’ve thought carefully about relevant risks.

What does one need to run a ballot initiative campaign that has lots of positive impact? According to someone who has done this, a campaign director should coordinate several kinds of domain experts, to draw on legal, political, issue-specific, and campaign expertise—see footnote for more details.^[21]

In addition to the above coordination, you need funding, and you need to actually run the initiative (e.g. get signatures).

Having an impact through lobbying:

Some insights from a professional with relevant experience:

• Detailed knowledge of what’s going on inside relevant decision-making bodies is very useful for advocacy/lobbying.^[22]

• Some issues—especially those for which potential benefits would be highly diffused—have many people willing to be allies, but few or no champions, i.e. people dedicated to taking initiative and pushing on the issue. So such advocates can make a huge difference. As one experienced advocate put it, for ensuring good things happen, sometimes “you need someone in the room yelling about how important this is.”^[23]

• Promising tactics for advocacy can include the following, if they are done well—see footnote.^[24]

Personal fit

As discussed in the above section, personal fit is very important for how much impact one has in policy careers.

Before we get into specific policy jobs, these things make someone a good fit for policy work generally:

• Being willing and able to work well with—and ideally get along well with—people who have very different views than you (very important)

  • The US government has all the US represented—many different perspectives.
    • If you can’t stand hanging out with people who have very different perspectives than you, you’re going to have a bad time.
    • As one government official put it, “trying to be in an ideological bubble” is “a great way to not get anything done,” in Congress as well as in the executive branch.
    • So intercultural communication skills and emotional intelligence are very helpful. These skills are also practicable.
  • This (presumably) means being able to work well with people who have different views on hot political issues, as well as with people who don’t share values that are common in the EA community.
  • Multiple speakers over different talks emphasized this point.^[25]

• Strong communication skills

  • People skills

  • Self-censorship skills^[26]

  • Writing well

  • Communicating verbally well (both in informal and formal settings)

• Being able to keep an eye on—and carve out time for—what matters most (see the above discussion of this).^[27]

  • (This is especially challenging in e.g. some jobs at the Department of State, which can be very fast-paced and require juggling tons of information.)

• Being fine with clustered impact, i.e. with occasionally doing very valuable things and usually working on things you may not see as especially high-priority

• Having a high hedonic setpoint^[28]

• Analytical skill (ability to analyze an issue, use evidence to reach conclusions)

• Willingness to operate with mainstream conventions (e.g. wear nice clothes, network)

I was surprised that speakers didn’t much more heavily emphasize social skills/charisma. From their advice, it seems that people can be strong fits for policy work—even if they're not exceptionally charismatic—if they can get by with the social stuff (are pleasant to work with, not shy about reaching out to and talking with people), are OK with the quirks and warts of policy world culture, and are good at job-specific skills (e.g. writing for some jobs).^[29]

A relatively low-cost test way to test your personal fit for policy work is spending time in government-adjacent DC think tanks. You’ll spend time with folks with government experience.^[30] In addition, you’ll get to see how the policy process works; you get a better sense of how the process works end-to-end than you would even in some agencies.

Personal fit for Congress:

• You have to be willing to ask for things and be told no (one staffer made the analogy that it;s like asking people out on dates).
• You have to be willing to make compromises.
• It’s a fast-paced environment, which is great for some people—you’re not slogging away at the same thing for a long time.
• Support agencies are more traditional, can have more academic personalities; you do research and then explain it to people.
• Every Congress member’s office has a unique work culture.
• You have to be willing to work long hours (at least if you’re a regular staffer)—see footnote for more related details.^[31]

Personal fit for think tank work:

What makes think tank work a good fit for someone?

• One person at a think tank explains that they really like their work because they really enjoy interacting with people and learning about different issues.^[32]

• It can require a tolerance for dealing with issues around funders (depending on the think tank).

  • Agendas are often set with what funders are interested in at the time, and funders will often be less knowledgeable than you or have different priorities.

• Traditionally, think tank work (at least in one professional’s experience) required tolerance for being frequently interrupted in your work, as well as being present for others’ work.

  • But this may be less relevant in a post-COVID-19 world.

• The cliche image of a DC think tank: vague talking around issues and never really saying anything of substance, while saying nice things and wearing nice clothes

  • It’s very easy to get lost in that world if you want to.
  • Still, there are lots of work where really good substantive work happens.
    • Even there, you’ll be coming in contact with those people—watch out if that drives you nuts.

• Your experience will be determined by your manager and your team—it’ll be hard to get info from their website/what others have said online.

  • Invest in trying to get connected to people who work there, to learn more.

Personal fit for political data analytics:

What makes political data analytics a comparative advantage for someone? These things help:

• Having a strong statistical background/intuitions
• Programing ability
• Ability to iterate quickly
• Interest in politics
• Not being especially charismatic (that would make more social roles like Congressional staffing more likely to be your comparative advantage)

Personal fit for running for office:

One professional with relevant experience suggested that, while not everyone in this community is a good fit for running for office in the US, some people definitely are.

What makes someone a good fit for being a US political candidate? These things are important, they explain:

• US citizenship (required for US federal office)
• Openness to public visibility
• People skills (enjoying interacting with people; being a warm, likeable person; being a good listener)
• Self-censorship skills
• Ability to context-switch and meet people where they are
• Having good, healthy boundaries

These things are also good signs about your chances of being elected (they help, but they are not necessary):

• If you’ve at some point thought you should run for office
• If someone has at some point told you that you should run for office
• If you’re from (spent a substantial portion of pre-college years in) a small state, especially if it fits your party affiliation
• If you have a law degree
• If you have military experience
• If you’re a person of color (regardless of whether you’d run as a Democrat or Republican)
• If you’re a woman (especially if you’re willing to run as a Democrat)
  • Women are often under-confident in their abilities to run for and win elected office—watch out for that.

Personal fit is very individual-dependent, so talking with someone who has relevant experience is very helpful for assessing fit and getting started—a relevant professional’s #1 recommendation for people who might be good fits (e.g. some of the above things apply to them) is to talk to them. Message me on this forum, and I might be able to put you in touch. Also message me / fill that out if reading this section makes you think of someone who might be a good fit.

If you’re on the fence about getting in touch, please go for it!

What is day-to-day work like?

Generally, typical day-to-day work for policy professionals seems to consist largely of meetings and emails. Some jobs also involve different things; as you might expect, research jobs tend to involve more writing and research, while more senior roles can involve hiring and project/people management. Some policy jobs (e.g. some Congressional staffer / national security jobs) are very fast-paced and involve long hours; others policy jobs (e.g. many research roles in some think tanks and Congressional support agencies) are more moderate-paced and offer better work-life balance. Within a job, busyness varies somewhat depending on current events.

Here’s how a handful of people described their own experiences in various policy jobs:

• Executive branch:
  • Lots of meetings and emails
  • Some jobs: lots of drafting and giving feedback on documents
  • A high-up role in the executive branch:
    • Most important thing in his work is hiring people, help set priorities/budgets, make recommendations for / bring up topics in meetings with influential decision makers
    • Most of the day is not spent on super engaging stuff. Trying to move stuff down the pipeline to his two bosses, helping get information from other agencies
    • Does try to spend time on the most important things
• Congressional staffer:
  • Lots of meetings (because you’re trying to create/advance bills, and so are others)
  • Lots of communication, give and take; you often won’t get exactly what you hoped for
• Congressional Research Service:
  • Most of 450 analysts get an expertise-based slice of the federal governments’ issue interests—”you are the expert” on that topic
  • What you’ll be working on in a day is kind of unpredictable—new tasks come up
  • Academically/intellectually inclined folks work there (among others?)
  • There’s lots of collegial/informative interaction with staffers (and Members?)
  • Days vary by the day and how “hot” your issue is at the moment
    • Sometimes talking with people, sometimes (on slower days) doing reports
  • Tons of writing
  • "Work on legislation" can be a lot of things—identifying portions of existing law that would match the needs of the staff for revision, helping advise on the drafting process to ensure it matches the intent properly, explaining effects of possible amendments, etc. The member/committee always serves as the "hub" of the process.
• Think tanks:
  • One junior research role:
    • Quickly got handed a wide portfolio—a lot but learned a lot
    • In entry-level roles, you’ll likely be doing RA things that support report authors
    • Traveled to hear about national security landscape outside of DC
    • Majority of time was spent “getting a feel for the policy landscape” that could have then used in more reports/leading policy pushes
    • Very predictable/regular interaction with policymakers through events
    • Long timelines - “we’re trying to get this report out in a few months”; was able to plan out research
    • Felt freedom to ask question they felt were most useful and do what they thought they needed to do to figure it out
    • Had little direction/feedback from supervisor or way to track impact wins
    • If the think tank is strapped for cash, it’s very concerned with doing things that look like impact to funders
  • One program manager role:
    • Much less predictable work than a junior research role
    • Managing budgets, meeting agendas, a distributed team
    • Offered cool “new product” of creative events connecting people in a new way
      • That has to do with that particular think tank being less traditional
    • No work-life balance
  • One role as one of the directors of a think tank:
    • Lots of meetings with colleagues, with external people (often just get-to-know)
    • Setting aside 5-10 hrs week on actual research
    • Talks/Q&As are relatively common
    • Lots of navigating bureaucracies and different people’s interests (not just partisan interests)
• Steps early in the process of running for elected office (after talking with the speaker / assessing personal fit)
  • Talk about districts, which party to run for, how you feel comfortable positioning yourself relative to electorate, voters in state, figuring out narrative, making valuable connections with interest groups in area, decision makers in area, areas you need to brush up on (designing plan of action for addressing weak spots).
  • Pay attention to deadlines, paperwork, going around to voters, shaking their hands, etc

Addendum

Edit: When writing Part 2, I realized a little of that content would fit better in this post. To make it easier for people who already read the above to find this extra content, here it is.

More on paths to impact:

• People often overestimate how difficult it is to get certain government jobs; the competition for very consequential government jobs relative to leverage is often really low, especially compared to corporate jobs.
  • Funding you can control in a mid-level federal job (relative to the counterfactual employee) can often be in the billions.
    • E.g. at budget offices
    • Large foundations/philanthropies are similarly high-leveraged.

More details on personal fit:

• Examples of DC-specific phrasings/surprising cultural things that sometimes bother people from this community:
  • People are often pretty wrong in kind of obvious ways that might bother you.
  • People tend to care less about things like people in other countries, animals, future generations.
  • Open-mindedness is lower; caricaturization happens.
  • Standards for evidence and argument are often much weaker than “in more academic and more truth-focused” circles
    • Some stereotypes
    • Silly mistakes about economics, math

More details on what the policy world is like:

• One policy professional mentioned that most policy folks they interact with “are there because they’re trying to make a difference”—there aren’t many sleazy types (there’s more of those in local politics because of weaker selection effects).
  • People in DC are relatively idealistic in that many of them could be making more money elsewhere and are working in government because they think it is the right thing to do, but tend to be less open-minded/cosmopolitan than many people in this community.
• In the White House much decision-making is made on sleep deficits (which is risky)
  • High-consequence decisions are sometimes made by people who haven’t eaten/slept in the last 24 hours.

What’s next?

A companion post with more concrete advice on how to take steps toward these careers is available here.

If you have an additional/different perspective on policy careers that you’d like to share in the comments, please do so!