Onpode
Cover art for Why some states extract revenue and build infrastructure while others extract revenue and collapse

Why some states extract revenue and build infrastructure while others extract revenue and collapse

November 14, 2026 · 13 min

Maya Chen & Dr. Nathan Hayes

Nathan, quick thing before we start — did you have a weird week too, or is it just me carrying something heavy into this room? Now, I'll say the material didn't help. Though I suppose that's appropriate — the subject earns the weight. It really does. We're talking about state capacity today — which sounds…

State capacity refers to the institutional ability of a government to raise revenue, enforce laws and contracts, administer policy, and deliver public goods such as education and healthcare.

0:0012:33
Get the next episode on World →

Follow it free — new episodes land in your feed.

Or make your own — any topic, in minutes

More Onpode episodes on World →

About this episode

State capacity refers to the institutional ability of a government to raise revenue, enforce laws and contracts, administer policy, and deliver public goods such as education and healthcare.

Grounded in 12 sources
Algorithmic Governance And The Transformation Of Administrative Law In The Digital State ↗ · doi.org
State Capacity and Gender Inequality: Experimental evidence from Papua New Guinea∗ ↗ · semanticscholar.org
The dual role of state capacity in opening socio-political orders ↗ · tandfonline.com
Raising Revenue Right - World Bank Documents and Reports ↗ · documents1.worldbank.org
Exploring the Dynamic Interplay between Institutions and State ... ↗ · elibrary.imf.org
[PDF] The Worldwide Governance Indicators: 2025 Methodology ... - World Bank ↗ · worldbank.org
Berggruen Institute ↗ · berggruen.org
Global Peace Index 2026 ↗ · economicsandpeace.org
What is state capacity and why does it matter? - Economics Observatory ↗ · economicsobservatory.com
State capacity - Wikipedia ↗ · en.wikipedia.org
Bold investment in the basics of tax administration is key to sustained revenue gains | ICTD ↗ · ictd.ac
Digital public infrastructure and tax: strengthening state capacity with shared digital rails | ICTD ↗ · ictd.ac
Read transcript

Maya Chen: Nathan, quick thing before we start — did you have a weird week too, or is it just me carrying something heavy into this room?

Dr. Nathan Hayes: Now, I'll say the material didn't help. Though I suppose that's appropriate — the subject earns the weight.

Maya Chen: It really does. We're talking about state capacity today — which sounds almost technical until you sit with what it actually means for the people inside these systems. And the place I want to start is the DRC. The Democratic Republic of Congo extracts roughly 55% of citizens' income through state agents — and returns almost nothing. No services. Nothing.

Dr. Nathan Hayes: 55% — that's higher than the top marginal income tax rate in most Nordic countries, which at least come with functioning healthcare and infrastructure.

Maya Chen: Right — and yet we categorize the DRC as a weak state. The Fragile States Index 2024 puts it at 106.7, in the same tier as Somalia at 111.3, South Sudan at 109.0. The framing is failure. But I keep thinking — failure at what? Because something is clearly working.

Dr. Nathan Hayes: And this is where I want to be precise about the definition. State capacity — the institutional ability to raise revenue, enforce laws, administer policy, deliver public goods — is the structural foundation. The DRC demonstrably has one component: extraction. What it lacks is the outward-facing functions. Legal capacity, administrative capacity. Those are absent.

Maya Chen: So the central thesis here is almost uncomfortable — a capable state and a kleptocratic one aren't distinguished by bureaucratic skill. They're distinguished by whom the power actually serves.

Dr. Nathan Hayes: That's the thesis, yes. And it has mechanistic implications I want to walk through, because the development literature often treats weak fiscal capacity and extractive institutions as the same problem. They're not. A state can be highly organized for extraction — that is a form of institutional competence.

Maya Chen: Which is — sort of a dark reframe of the whole 'capacity building' project, isn't it. If the capacity is already there, just oriented differently, then aid that strengthens institutions might be strengthening the wrong ones.

Dr. Nathan Hayes: It raises the causal question I think matters most: does capacity enable legitimacy, or does pre-existing social trust enable the tax compliance that then funds capacity? That arrow isn't settled.

Maya Chen: Mm, and that — yeah, that's the thing we're really trying to work out. Because if the trust has to come first, then no amount of institutional engineering gets you there from outside.

Dr. Nathan Hayes: Right, but — the trust question actually requires we decompose what capacity even is first, because I think we've been using it as one thing when it's three. Fiscal capacity: collect revenue. Legal capacity: enforce contracts, protect property. Administrative capacity: convert money into services. And the mechanism matters — fiscal resources fund administration, administration enables tax collection, legal infrastructure secures conditions for both. Break one pillar and it cascades.

Maya Chen: Three distinct pillars. Okay — so which one does the DRC actually have?

Dr. Nathan Hayes: Parts of fiscal. The extraction part. Not the revenue-building part — there's a distinction. And near-zero legal or administrative capacity. So think of it like a restaurant that's very good at charging customers but has no kitchen. The billing system works. The product doesn't exist.

Maya Chen: Oh — that's it, isn't it. That's the whole thing.

Dr. Nathan Hayes: And here's what that means mechanically: extractive institutions — kleptocracy, asset seizure, monopolization — they're not the symptom of incapacity. They're a — actually, let me reframe — they emerge as the default when a state can't fund legitimate governance. But in the DRC, extraction isn't filling a vacuum. It's the organized product.

Maya Chen: So say you run a small textile workshop in Kinshasa. You pay a local official 20,000 francs monthly — not labeled a tax, just understood. Then a second official seizes your shipment mid-transit, claims you haven't paid him. You go to court. And the court — won't help.

Dr. Nathan Hayes: That's not one failure. That's two separate extraction nodes operating in parallel. Which means — importantly — the state isn't absent. It's present twice, taking twice, delivering nothing.

Maya Chen: And the court being useless — that's the legal capacity pillar completely gone.

Dr. Nathan Hayes: Exactly that. No legal capacity means contracts aren't enforceable, property rights aren't protected — so investment becomes irrational. Why invest long-term when seizure is a credible outcome? That feeds back into fiscal capacity, because now the tax base is suppressed. Which is the low-capacity equilibrium trap — the pillars reinforce each other on the way down just as powerfully as on the way up.

Maya Chen: Compounding returns, but — going the wrong direction. Faster down than up.

Dr. Nathan Hayes: The Berggruen Index, 145 countries, 2000 to 2023 — Helmut Anheier called it the most striking finding of the dataset — most countries don't move between governance tiers over that entire quarter century. Which suggests asymmetry is real. The trap holds.

Maya Chen: So calling the DRC weak is — mm — it's actually the wrong word. It's misdirected. Selectively strong where it most harms people, absent everywhere they need it.

Dr. Nathan Hayes: And 'selectively strong' is actually the key phrase, because it tells you why the trap is self-sealing. Extraction erodes the formal economy — businesses go informal, capital hides, investment dries up. Which shrinks the tax base. Which means extraction becomes the only viable revenue mechanism again. You can't escape the loop because the loop destroys the resources you'd need to escape.

Maya Chen: It forecloses itself.

Dr. Nathan Hayes: Structurally, yes. And then — now, this is where the resource curse enters — natural resources make it worse not better, because oil rents or mineral rents let you skip the legitimacy bargain entirely.

Maya Chen: Wait — skip it how?

Dr. Nathan Hayes: Broad taxation requires — at some point — enough compliance that you can't just take it all by force. And that compliance only comes if people believe the state delivers something. So historically, the revenue-legitimacy feedback loop is what pressures governments toward at least minimal public goods. But resource rents bypass that. You fund the state from the ground, not from your citizens. No compliance needed. No legitimacy bargain.

Maya Chen: So the extraction doesn't even need to function well — you just need the oil to flow.

Dr. Nathan Hayes: Right — and that's why GDP per capita correlates more reliably with institutional quality than with natural resource endowments. Resource wealth without the institutional feedback loop just substitutes rents for the pressure that would have built capacity. You get the money without building anything that could survive the money running out.

Maya Chen: That is — mm — that's almost cruel as a design. The thing that looks like an advantage is actually hollowing out the mechanism that would have created real stability.

Dr. Nathan Hayes: Compounding institutional returns work in both directions — that's the point. High-capacity states: fiscal resources fund better administration, broader tax collection, further institutional investment. Each cycle widens the gap. Low-capacity states with resource rents: the rent substitutes for the cycle, so you never enter it. You're not building a ladder. You're standing on a column that has no rungs.

Maya Chen: And the Berggruen data — four stable worlds of governance, 25 years, rare upward mobility — that's the evidence that the trap holds even at the aggregate level. It's not individual countries being unlucky. It's a pattern.

Dr. Nathan Hayes: Anheier called it the most striking finding of the dataset. 145 countries. A quarter century. And the configurations held. Which means — actually, this is the question I think you've been circling — is 'development' as a ladder just a story we construct backwards from the cases that climbed?

Maya Chen: From the winners. Yeah. We're — sort of narrating the survivors and calling it a roadmap. And the part that makes this messier is coming — because the indices we're using to measure who's climbing and who's stuck? They don't even agree with each other. That's a different kind of trap.

Dr. Nathan Hayes: And it matters mechanically — if you can't locate where a country sits, you can't prescribe what it needs. The measurement problem isn't academic. It has consequences.

Maya Chen: But the measurement problem isn't just — mm — a methodological footnote. Because the International Research Institute looked at composite indices from 2015 to 2024 and found that two countries can trade places by ten or more ranks depending solely on which index you cite. Same underlying data. No change. Just — which tool you picked up.

Dr. Nathan Hayes: Ten or more ranks. That's not noise. That's a different country.

Maya Chen: And the World Bank Worldwide Governance Indicators and the Berggruen Governance Index — they both claim to be measuring the same underlying construct. Institutional quality. But they diverge most sharply in the middle-income band. Which is exactly where the policy questions are most live.

Dr. Nathan Hayes: Now, the mechanism matters here — because what I think is actually happening is that composite indices conflate two structurally different problems. A state weak across all three capacity pillars looks identical, on the index, to a state strong at extraction but weak at service delivery. Same score. Completely different diagnosis.

Maya Chen: Wait — so the DRC could score similarly to a country that's just administratively slow?

Dr. Nathan Hayes: Potentially, yes. And the policy consequences of that conflation are significant. If you can't distinguish selective capacity from uniform weakness, you prescribe the same intervention for both. Which is — actually, this is where the digital tools argument breaks down in a very specific way.

Maya Chen: The digitalization fix.

Dr. Nathan Hayes: The ICTD research is pretty direct on this — sustained revenue gains require skilled personnel, institutional coordination, investment in tax administration basics. Not just digitalization. And what happens when you drop digital tools into a low-capacity state without those preconditions — it doesn't dissolve the constraints. It automates them.

Maya Chen: It automates — wait, so a faster extraction system is still an extraction system.

Dr. Nathan Hayes: For the lowest-capacity states, digital tools are more likely to be captured for predation than to generate any legitimacy feedback. You haven't built trust. You've just made the taking more efficient.

Maya Chen: And that's — mm — that's the same trap as the index problem, sort of. The tool presupposes the thing it was meant to build. The index presupposes we know what we're measuring. The app presupposes there's institutional trust to amplify.

Dr. Nathan Hayes: Which loops back to the IMF causal arrow question — does capacity enable legitimate institutions, or does pre-existing social trust enable the revenue base that builds capacity? Because if trust has to come first, then no index score and no digital rollout gets you there from the outside.

Maya Chen: And we're not resolving that arrow today, are we.

Dr. Nathan Hayes: No. The honest answer is — the correlation between institutional quality and prosperity is real, the direction is real, but the precision of that country-level claim? Softer than the development literature typically presents. We're building policy on indices that disagree at the resolution where policy actually has to work.

Maya Chen: And that's — yeah, that's where I keep getting stuck. Because the Berggruen four worlds of governance, 25 years of data, and they're just... stable. They're not a ladder. They're four separate equilibria. So the question we've been asking — how does a fragile state climb — might be the wrong question entirely. Maybe the real question is what actually disrupts an equilibrium. And I don't know that aid does it. I don't know that institutional transplants do it.

Dr. Nathan Hayes: And the measurement problem makes it genuinely hard to know when a country has even moved. If the World Bank WGI and the Berggruen Index diverge by ten or more ranks in the middle-income band — which is precisely where any country climbing would appear — then what looks like upward mobility might just be an artifact of which index you're watching.

Maya Chen: So we might not be able to see the escape even if it's happening.

Dr. Nathan Hayes: Not reliably. No.

Maya Chen: The textile workshop is what stands out. We started this whole conversation with the DRC — 55 percent extracted, nothing returned — and I framed it as a policy problem. A structural problem. But she's not waiting for Helmut Anheier's dataset to resolve or for the ICTD to publish the right intervention framework. She's deciding today whether to buy a second loom. And that decision — invest or stay liquid, trust the system or protect against it — that's simultaneously the symptom of the trap and the mechanism that keeps it closed. Her rational response to extraction is exactly the thing that starves the tax base that might have funded something different.

Dr. Nathan Hayes: And she's making the economically correct decision, given the information she has. That's what makes the low-capacity equilibrium trap structurally coherent. It's not irrationality at any level. It's rationality all the way down, producing collective ruin.

Maya Chen: Which is — mm — a genuinely uncomfortable place to end up. We didn't find the lever. I'm not sure the lever exists in the form we went looking for it.

Dr. Nathan Hayes: Honest, though. The correlation is real. The direction is real. The precision of what to do about it — not there yet. I'd rather leave it at that than close the loop we didn't actually close.