Daniel sent us this one — he's been thinking about something we touched on in earlier conversations about governance failure in Israel. The term "policy wonk" gets thrown around a lot, often as a kind of put-down. But when you strip that away, what does the skill set actually translate to? What job descriptions do these people hold, what backgrounds do they come from, and how do they build a knowledge base that's genuinely useful to the policymakers they work with? It's a good question, because the term obscures more than it reveals.
It really does. And I think the first thing to say is that a policy wonk, stripped of the pejorative baggage, is someone with deep granular knowledge of how a specific policy domain actually works. Not the theory, not the textbook version — the implementation details. They know which levers actually move when you pull them, which ones are rusted shut, and which ones will trigger three unintended consequences if you're not careful. Tax codes, healthcare financing formulas, education funding mechanisms, regulatory frameworks — these are the landscapes they navigate.
It's not just knowing that something exists. It's knowing which three percent of a thousand-page bill actually matters.
And it's worth distinguishing the wonk from adjacent roles that often get conflated. A lobbyist advocates for a specific interest — they have a client. An academic publishes for peer review — their audience is other scholars. A pure data scientist models without institutional context — they can tell you what the regression says but not whether the agency in question has the staff to implement it. A wonk sits at the intersection of data, institutional memory, and legislative mechanics. They understand the policy substance, the political viability, and the administrative feasibility all at once.
The triple threat. Or the triple headache, depending on your perspective.
It's worth noting — the best wonks are deeply political. Not in the partisan sense necessarily, but in the sense that they understand a policy proposal is worthless if it can't command enough votes. They're not dreaming up optimal solutions in a vacuum. They're designing things that can actually pass and actually work.
Let's get concrete. If someone is described as a policy wonk, what are the actual job titles they might hold? Where do they sit in the org chart?
There are a handful of archetypal roles. On Capitol Hill, you've got the legislative director for a member of Congress — that person manages the entire policy portfolio, juggling everything from defense appropriations to agricultural subsidies, and they're the ones who decide which issues the member prioritizes. Then there's the senior policy advisor, who's specialized in one domain — healthcare, or tax, or foreign affairs. These people typically have ten to fifteen years in their field. Below them, you've got legislative assistants and legislative correspondents — the junior wonks who handle constituent mail on specific issues and draft the first versions of talking points and bill language.
There's a pipeline. You start in the mailroom of policy and work your way up.
And then there's the committee staff — these are the people who actually run the machinery. A committee staff director for, say, the Senate Finance Committee manages a team of maybe thirty to forty policy specialists, coordinates hearings, negotiates bill text with the minority staff, and interfaces with the Congressional Budget Office on scoring. These roles are less visible than the members themselves but often more influential on the actual content of legislation.
Because the member is doing press conferences and fundraising calls while the staff director is on the phone with the CBO at eleven at night trying to figure out if a particular provision will add twelve billion or fourteen billion to the deficit.
That twelve versus fourteen billion distinction can be the difference between a bill living or dying. The Byrd Rule in the Senate reconciliation process has very specific deficit thresholds. If you blow past them, your entire bill can be challenged on a point of order. Knowing exactly where those lines are — that's wonkery in its purest form.
That's the Hill. What about outside of Congress?
Think tanks are the other major ecosystem. You've got policy analysts at places like the Urban Institute, the Brookings Institution, RAND, the Heritage Foundation, the Center for American Progress. These are the people producing the white papers, the policy briefs, the cost estimates that eventually get picked up by legislative staff and turned into actual bills. At the more senior level, you've got fellows and senior fellows — people with fifteen to twenty years of experience who are the go-to experts in their domain. And then there are the campaign-adjacent policy shops — the policy director for a presidential campaign, for instance, who has to translate a candidate's broad vision into specific, costed proposals that can survive a general election debate.
At the state and local level?
A director of policy for a governor or a mayor. These people are translating campaign promises into executive orders and budget proposals. They're working with agency heads to figure out what's actually implementable. In a large city like New York or Los Angeles, the mayor's policy director might oversee a team of ten to fifteen people covering everything from housing to transportation to public health. And at the county level, you've got policy analysts embedded in departments like health and human services who are the ones who actually understand how federal block grants flow down to local programs.
There's a whole ecosystem here that most people never see. These aren't the names in the newspaper, but they're the ones who wrote the thing the names in the newspaper are arguing about.
And let me give you a concrete example of how this works in practice. The Affordable Care Act — love it or hate it — was shaped in large part by a small group of staffers on the Senate Finance Committee who had spent years studying what happened in Massachusetts. The 2006 healthcare reform there, Romneycare, was their blueprint. They knew which design elements had worked — the individual mandate, the subsidies structured as advanceable tax credits, the essential health benefits package. And crucially, they knew which elements would collapse under adverse selection if you got the design wrong. The individual mandate wasn't an ideological choice. It was a technical necessity to prevent the insurance pool from becoming a death spiral where only sick people signed up.
Because without the mandate, healthy people stay out, premiums spike, more people drop out, premiums spike further, and the whole thing implodes.
This wasn't theoretical. They had the Massachusetts data. They could see exactly what the enrollment patterns were, what the risk corridors looked like, what happened when you adjusted the subsidy thresholds. That's wonkery in action. It's not glamorous. It's sitting in a windowless room reading actuarial reports. But it's the difference between a bill that functions and a bill that collapses on contact with reality.
The Clinton healthcare effort in 1993 — that was the cautionary tale they were all trying to avoid, right?
The Clinton plan was drafted in secret by a task force of about five hundred people led by Ira Magaziner, with very little input from the congressional committees that would actually have to pass it. It was over a thousand pages long and incredibly prescriptive about things like regional health alliances. The Senate Finance staffers who worked on the ACA had all studied that failure. They knew they needed to build something that the committees could own, that had buy-in from key stakeholders like the insurance industry and the pharmaceutical companies, and that could survive the legislative gauntlet.
Institutional memory isn't just a nice-to-have. It's the thing that prevents you from repeating a catastrophic failure.
That's one of the key ways wonks build useful knowledge. It's not just reading white papers. It's a systematic process. First, comparative case analysis — studying how ten different states implemented Medicaid expansion and which design choices produced better outcomes on cost, access, and health metrics. Second, institutional memory — knowing that the 1996 welfare reform had a hidden provision that still affects SNAP eligibility today, or that a particular regulatory waiver process was created in 1987 for a completely different purpose but now governs how states can experiment with Medicaid.
The ghosts of legislation past.
And third, relationship networks. The best wonks have the phone numbers of the career staff at the Department of Health and Human Services who know where the bodies are buried in the regulations. The political appointees come and go every four to eight years. The career civil servants have been there for thirty years. They know that a particular regulation was written in 1992 to solve a problem that no longer exists, but it's still on the books and it still constrains what you can do. If you don't know that person, you don't know that constraint.
The knowledge base is part data, part history, and part Rolodex.
The educational path matters too, though it's not the whole story. Most wonks start with a master's in public policy, an MPP, or a related field like economics or a JD with a policy focus. Programs like the Harvard Kennedy School, Princeton's Woodrow Wilson School, Georgetown's McCourt School — these are the traditional feeders. But the critical credential isn't the degree. It's Hill experience. Spending two to four years as a staffer, ideally five to ten. The real education happens in the trenches — drafting bill language, writing committee markups, negotiating with the CBO for score estimates.
The CBO deserves its own moment here, because you mentioned them several times. What actually is the CBO and why does it have so much power?
The Congressional Budget Office employs roughly two hundred and fifty analysts — economists, health policy specialists, tax experts — whose entire job is to produce cost estimates for every major piece of legislation. When a bill is introduced, the CBO "scores" it — meaning they estimate how much it will cost or save over a ten-year window. That score is make or break. If the CBO says your healthcare bill will add three hundred billion to the deficit, you've got a massive problem. If they say it'll save two hundred billion, you've got a massive selling point.
These are career people, not political appointees.
The director is appointed by the House and Senate budget committees, but the analysts are career civil servants. They're not supposed to be partisan. Their methodology is public, their assumptions are documented, and both parties have learned — sometimes painfully — that you can't just dismiss a CBO score you don't like. The 2017 effort to repeal the ACA died in part because the CBO scored various repeal bills as causing twenty-plus million people to lose coverage. Those numbers were politically devastating.
Let's talk about the 2017 Tax Cuts and Jobs Act, because that's another case study in wonkery.
That bill was drafted almost entirely by a handful of staffers on the Senate Finance Committee and the House Ways and Means Committee who had been working on tax reform since the 1990s. These people had been waiting for their moment for two decades. They knew the tax code inside out. They knew exactly which provisions would trigger the Byrd Rule in reconciliation — the rule that says a provision can't increase the deficit beyond the ten-year budget window. That's why several of the individual tax cuts were designed to expire after 2025. It wasn't an oversight. It was a deliberate design choice to fit within the reconciliation constraints.
The fact that we're now having a debate about extending those cuts isn't a glitch. It was the plan all along.
It was the plan. And the people who drafted that bill knew that the corporate tax cut — reducing the rate from thirty-five percent to twenty-one percent — would be permanent, because they could offset it with other revenue raisers over the long term. But the individual cuts had to be temporary to make the math work. That's the kind of detail that only a handful of people in the country truly understood at the level required to draft the legislative text.
We've established what a wonk is and where they come from. But here's where it gets tricky. When you're building an opposition movement from scratch, you don't have a twenty-year institutional memory to draw on. You don't have committee staff. You don't have people who've been marinating in the tax code since the Clinton administration. So how do you build one?
This is the central challenge, and it's where a lot of movements fail. The Tea Party wave in 2010 is the classic cautionary tale. They won elections — sixty-three House seats, a historic landslide — but they had almost no policy infrastructure. No detailed alternatives. No legislative language ready to go. They had a lot of energy and a clear message about what they were against, but when it came time to govern, they were reactive. They could block things, but they couldn't build things.
The "repeal and replace" problem in miniature. They had the repeal part figured out. The replace part was a blank sheet of paper.
Contrast that with the 1994 Republican Contract with America. That was built by a network of wonks at the Heritage Foundation and the American Enterprise Institute who had spent eighteen months drafting ten specific bills with detailed legislative language. They had cost estimates. They had talking points. They had answers to every obvious objection. When Newt Gingrich and the Republican caucus unveiled it six weeks before the election, they weren't just saying "we're against the Democrats." They were saying "here are ten specific things we will pass in the first hundred days." And they did pass most of them.
The Heritage Foundation team — how many people actually wrote that?
About twelve policy analysts working over eighteen months. That's it. A dozen people in a conference room produced the intellectual architecture for a political revolution that reshaped American politics for a generation. The Contract itself was drafted by a core team led by policy analysts who had been working on these issues for years. They knew the legislative mechanics. They knew what could pass. They knew what would score well with the CBO. It wasn't magic. It was methodical, grinding work.
The lesson for an opposition movement is: you need your twelve people in a conference room.
You need them early. You can't scramble to put this together six months before an election. The Contract with America team started working in early 1993, nearly two years before the 1994 midterms. They had time to draft, redraft, vet with subject matter experts, run the numbers, and anticipate every attack line the Democrats would use. By the time the campaign was in full swing, the policy work was done.
Let's bring this international, because the prompt is really about the question of building viable opposition, and the Israeli context is the starting point. What are the examples from other countries of movements that did this well?
The Chilean student movement of 2011 is a fascinating case. It started as street protests — massive ones, hundreds of thousands of students demanding education reform. But unlike a lot of protest movements, they had a dedicated policy team. These weren't just kids with signs. They had education economists, legal scholars, and policy analysts who produced detailed proposals for how to reform Chile's voucher-based education system. They could tell you exactly how the funding formulas worked, what the constitutional constraints were, and what specific legislative changes would be required.
That translated into actual policy outcomes?
When Michelle Bachelet returned to the presidency in 2014, her major tax reform — which raised corporate taxes to fund education — was shaped directly by the policy work the student movement had done. The students didn't just win the argument in the streets. They won it in the ministries, because they had people who could sit down with the technocrats and speak their language.
That's the bridge. You need people who can translate between the energy of a movement and the machinery of government.
The UK provides both a positive and a cautionary example. The 1997 Blair campaign had a dedicated policy unit that spent years developing specific, costed proposals. The New Deal for Young People — a welfare-to-work program for eighteen to twenty-four year olds — wasn't a slogan. It was a detailed program with funding mechanisms, eligibility criteria, and performance metrics. They knew exactly how much it would cost and how they would pay for it.
Versus the Corbyn era.
The Corbyn leadership from 2015 to 2019 had a much weaker policy infrastructure. Manifesto commitments were often drafted in isolation by the leader's office without proper costing or feasibility analysis. The 2019 manifesto in particular had a reputation for promising things that sounded good but hadn't been stress-tested. And when you're in opposition, you're going to be attacked on credibility. If you can't defend the numbers, you lose the argument before it even starts.
The number of special advisers in the UK government grew from thirty-eight in 1979 to over a hundred by 2019. That's a proxy for the professionalization of policy advice, right?
And it reflects a broader trend across democracies. Policy-making has become more complex, more technical, and more data-driven. You can't just wing it with a few clever speechwriters and a good gut instinct. You need people who understand the models, the datasets, the regulatory frameworks. The days of the gifted amateur running a government department are over — or they should be.
Let's get into the data dimension specifically. You mentioned that modern wonks need to be data-savvy. What does that actually mean in practice?
It's not just reading regression tables. It's understanding administrative data quality — knowing which government datasets are reliable and which are garbage. The Census Bureau's American Community Survey, for instance, is the gold standard for demographic data in the United States. It's methodologically rigorous, it has large sample sizes, and it's updated annually. On the other hand, some self-reported crime data — like certain FBI Uniform Crime Reporting categories — have known issues with underreporting and inconsistent definitions across jurisdictions. A good wonk knows the difference.
It's not just "what does the data say." It's "should I trust this data at all.
That's a skill that takes years to develop. You also need to be able to run or at least interpret cost-benefit analysis using tools like the Urban Institute's Tax Policy Center microsimulation model or the CBO's dynamic scoring framework. Dynamic scoring, by the way, is controversial and worth understanding. Traditional scoring — static scoring — assumes that tax changes don't affect overall economic growth. Dynamic scoring tries to model how tax changes affect growth, which then feeds back into revenue estimates. It's more realistic in theory, but it's also more sensitive to assumptions. Small changes in your growth assumptions can produce wildly different revenue estimates.
Which means the wonk who controls the assumptions controls the score.
That's why both parties fight so hard over who runs the CBO and what methodology they use. The wonk fights are often the most consequential fights in Washington, and nobody pays attention to them.
Let's talk about the opposition challenge more directly. If you're building a movement — whether in Israel or anywhere else — and you want to transition from protest to governance, what's the concrete playbook?
Step one: identify three to five priority policy domains. You can't be expert in everything. Pick the areas that matter most to your movement's vision — housing, healthcare, education, corruption reform, whatever it is — and go deep rather than wide. A movement that has detailed, costed proposals in five areas and nothing else is far more credible than a movement that has vague position papers on fifty topics.
Depth over breadth.
Step two: recruit domain experts who are willing to do the grunt work. And I want to emphasize the grunt work part. Comparative analysis — studying how different jurisdictions have tackled the same problem — is tedious. Legislative drafting is tedious. Running cost estimates is tedious. You need people who find that work satisfying, not people who just want to be on television.
The glory-to-grind ratio in policy work is heavily skewed toward grind.
Step three: create a formal policy shop. This is a non-profit or campaign-adjacent entity — a 501c4 in the US context, or the equivalent in other countries — that produces white papers, model legislation, and cost estimates. It needs a budget, staff, and a clear mandate. The Heritage Foundation's Contract with America team had all three. They weren't volunteers working on weekends. They were full-time professionals with salaries and deadlines.
A movement that spends all its money on rallies and advertising and nothing on policy infrastructure is building a house without a foundation.
Step four: build relationships with career civil servants. This is often overlooked because movements tend to be anti-establishment by nature, and career bureaucrats are about as establishment as it gets. But these are the people who know what's actually implementable. They've been in the agencies for twenty or thirty years. They've seen administrations come and go. They know which reforms worked and which ones died on the vine because the agency lacked the capacity to implement them.
It's like adopting a feral cat.
The relationship between a movement and the civil service. You need them, they're wary of you, and if you approach them wrong they'll disappear under the porch and you'll never see them again.
actually not a bad analogy. And the key is respect. Career civil servants aren't the enemy. They're the institutional memory of the state. If you treat them as obstacles to be routed around, you'll lose access to the very knowledge you need to govern effectively.
Let me ask you something about the incentive structure here. Why would someone choose to be a wonk rather than, say, a lobbyist who makes three times the salary?
That's a real tension. The financial sacrifice is significant. A mid-career policy analyst at a think tank might make eighty to a hundred and twenty thousand dollars. A lobbyist with the same level of expertise can make three to five hundred thousand or more. The wonks who stay in the policy world tend to be motivated by something other than money — intellectual satisfaction, public service, the desire to shape outcomes rather than just advocate for an interest.
Or they're just too stubborn to leave.
There's some of that too. But the brain drain to the private sector is a real problem. When the best tax policy minds leave the Hill for K Street, the legislative process loses institutional knowledge. And replacing that knowledge takes years.
Which creates a vicious cycle. The less expertise there is inside government, the more government has to rely on outside lobbyists for information, which makes the lobbyists even more powerful, which makes it even harder to retain talent in government.
This is one of the arguments for strengthening the career civil service and making policy staff positions more attractive. Better pay, clearer career paths, protection from political purges. The wonks who stay in government for thirty years are worth their weight in gold, but the system doesn't always treat them that way.
Let's address some misconceptions head-on. The biggest one, I think, is that policy wonks are just boring academics who don't understand politics.
That's completely wrong, and the best wonks disprove it every day. They understand that policy design is fundamentally about what can pass, not just what's optimal. They know which compromises are fatal and which are cosmetic. They know which stakeholders have veto power and which are just noisy. That's political judgment, not academic abstraction.
Another one: that data analysis alone produces good policy.
Data is only as good as the institutional knowledge that interprets it. A regression model can tell you that a particular intervention reduces hospital readmissions by eight percent. It can't tell you that the regulation required to implement that intervention will be impossible to enforce because the relevant agency has lost thirty percent of its inspectors over the last decade. That's the kind of thing you only know if you've been in the trenches.
The third one: that think tanks are all partisan hack factories.
Some are, and they're explicit about it — Heritage is conservative, CAP is progressive, and they don't pretend otherwise. But many think tanks produce rigorous, non-partisan analysis that's used by both parties. The Urban Institute, RAND, the Brookings Institution — their work is cited in both Democratic and Republican legislation. The partisan ones serve a real function too, by the way. They provide the intellectual ammunition for their side. But not everything is a partisan hit job.
What are the actionable takeaways here? If someone listening to this is involved in building an opposition movement — in Israel or anywhere else — what should they actually do?
The single highest-leverage investment is funding a policy shop. A small team of wonks — start with five to ten people — dedicated to producing detailed, costed alternatives in three to five priority areas. This is more important than advertising, more important than rallies, more important than social media presence. Those things build visibility. The policy shop builds the capacity to govern.
Visibility without capacity is how you get the Tea Party problem — you win and then you don't know what to do with the win.
Second actionable insight: for individuals who want to become useful wonks, the path is fairly clear. Get a policy-relevant graduate degree — an MPP, a JD with a policy focus, an economics degree. Spend two to four years as a legislative staffer or the equivalent in your country's parliamentary system. Then specialize in one domain for five-plus years. The key is building a reputation as the person who knows the obscure detail that will break a deal. You want to be the person that the member of Congress calls at ten PM and says "what happens if we change the phase-in period from three years to five?" and you can answer without looking anything up.
The human search engine.
Third: movements should actively recruit from the career civil service. These are people who have spent years implementing policy. They know what works and what's a fantasy. They're often overlooked because they lack political credentials or campaign experience, but their institutional knowledge is irreplaceable. If you can bring a few senior civil servants into your policy shop — people who've been inside the agencies and know where the skeletons are — you've got an enormous advantage.
They know which doors are unlocked and which ones have alarms.
Which ones are just painted on the wall and don't open at all.
We've made the case for the wonk. But there's a tension here that I think is worth sitting with. How do you rehabilitate the term "policy wonk" in the public imagination? And how do you ensure that the wonk class doesn't become a self-perpetuating elite, disconnected from the people they're supposed to serve?
That's the democratic accountability question, and it's not trivial. The wonk class, like any expert class, can become insular. They go to the same graduate schools, work at the same think tanks, speak the same jargon, and develop the same blind spots. There's a real risk of groupthink — the ACA architects, for all their sophistication, didn't fully anticipate the political backlash the individual mandate would generate, because in their circles the mandate was just obviously necessary.
Sometimes the thing that's technically correct is politically combustible, and the wonk bubble doesn't always see it coming.
That's why you need diversity in your policy shop — not just demographic diversity, but experiential diversity. People who've worked in state government, not just federal. People who've been teachers or nurses or small business owners, not just career policy professionals. People who've lived in the communities affected by the policies they're designing. The wonk class needs to be porous, not a closed guild.
The other piece of this is communication. The best policy in the world is useless if you can't explain it to people in a way that makes sense.
That's a skill set that doesn't always overlap with wonkery. The people who can draft perfect legislative language aren't necessarily the people who can go on television and explain it in plain language. Movements need both. The communicators and the wonks have to work together, and they often don't speak the same language — literally.
The wonk speaks CBO score. The communicator speaks kitchen table.
Someone has to translate between them. That's actually a role in itself — the policy translator. Some of the most effective people in politics are the ones who can take a complex regulatory proposal and turn it into a thirty-second ad that doesn't distort it beyond recognition.
Looking forward, what does AI do to all of this? If policy modeling becomes increasingly automated, does the wonk's role change?
It shifts from manual analysis to interpreting model outputs and understanding their limitations. AI can run a thousand simulations of a tax proposal in seconds. What it can't do is tell you that one of your assumptions is wrong because of a regulatory quirk that isn't in any dataset. The core skill — knowing how policy actually works in practice, knowing the institutional landscape, knowing which questions to ask — becomes more valuable, not less, as the computation gets commoditized.
The wonk becomes the person who knows when the model is lying to you.
Or when the model is telling you something true but irrelevant. An AI can optimize a policy for cost-effectiveness and produce something that's politically impossible. The wonk's judgment — what can pass, what can be implemented, what won't blow up in your face three years later — that's not something you can train a model to do. Not yet, anyway.
We've covered a lot of ground. I think the core argument here is that policy wonks aren't the boring footnote to politics — they're the infrastructure. Without them, you get vibes and slogans. With them, you get a government.
The term itself — I think we should reclaim it. "Wonk" shouldn't be an insult. It should be a job description for the people who do the unglamorous, essential work of figuring out how to actually make things better. Every successful reform, every functional program, every piece of legislation that actually improved people's lives — somewhere behind it was a wonk who did the reading.
The reading nobody else wanted to do.
The thousand pages nobody else wanted to read. The CBO score nobody else wanted to understand. The regulatory history nobody else knew existed. That's the craft.
And now: Hilbert's daily fun fact.
Hilbert: The mantis shrimp has the most complex eyes in the animal kingdom — each eye contains twelve types of photoreceptor cells, compared to the human eye's three. But here's the twist: despite this hardware advantage, experiments in the 1950s showed that mantis shrimp are actually worse than humans at distinguishing between similar colors. They seem to process color information in a completely different way — faster but less nuanced — essentially trading resolution for speed.
They've got the world's fanciest camera and they're using it as a motion detector.
That's unsettling.
This has been My Weird Prompts. Thanks to our producer Hilbert Flumingtop for keeping the show running. If you found this useful, share it with someone who's building something — a movement, an organization, a policy team. And if you want to go deeper on how to measure whether these systems are actually working, check out the Democracy Dashboard episode. We're at myweirdprompts.
See you next time.
