Why Nations Fail: What Venice’s Rise and Fall Teaches Us About AI Governance

Acemoglu and Robinson’s Why Nations Fail reads like a thousand-year dataset for anyone building institutions around transformative technology. The argument is monocausal by design, and that’s both its power and its limit: the prosperity of a nation is a product of its political and economic institutions. But the real punch of the book is elsewhere: the binary isn’t stable. Institutions flip. And the mechanism of that flip is what matters most.

TL;DR

Why Nations Fail (2012) argues that the wealth gap between nations has a single root cause: political and economic institutions. Countries with inclusive institutions (broad property rights, rule of law, pluralistic politics) generate sustained growth. Countries with extractive institutions (elite capture, restricted access, concentrated power) stagnate or collapse after brief spurts. The engine connecting the two is creative destruction: inclusive institutions tolerate it, extractive ones suppress it. Institutions flip. Venice went from the most inclusive economy in the medieval world to a museum in three generations. The mechanism of that flip, the Serrata, is the single most important case study for anyone thinking about AI governance today.

Acemoglu received the Nobel Prize in Economics in 2024, alongside Robinson and Simon Johnson, largely for this body of work on how institutions shape prosperity.

The Core Framework: Two Kinds of Institutions, One Mechanism

The architecture rests on a simple taxonomy. Economic and political institutions are either inclusive or extractive.

Inclusive economic institutions secure property rights for a broad cross-section of society, allow reasonably free entry into markets, and enforce contracts. Inclusive political institutions distribute power pluralistically and impose constraints on executives. The two reinforce each other: broad political participation generates demand for economic openness, and widely distributed economic power makes political capture harder.

Extractive institutions are their mirror. Power concentrates in a narrow elite. Property rights exist only for insiders. The state apparatus channels resources upward. And the two forms of extraction reinforce each other just as effectively: economic rents fund political entrenchment, and political monopoly protects economic privilege.

The mechanism connecting institutions to growth is Schumpeter’s creative destruction. New technologies, new business models, new entrants displacing incumbents. Inclusive institutions tolerate creative destruction because no single group can block it. Extractive institutions suppress it, because every wave of innovation threatens the elite’s position. This is why extractive regimes can sometimes generate impressive short-term growth (the Soviet Union averaged 6% annually from 1928 to 1960) but cannot sustain it: they systematically prevent the very disruptions that produce long-run prosperity.

Venice: The Parable That Carries the Book

Forget the Industrial Revolution, the North/South Korea contrast, or the Glorious Revolution. The case that best illustrates the full mechanism is medieval Venice.

By the early eleventh century, Venice had become possibly the wealthiest city in the world. Its prosperity rested on a specific institutional innovation: the commenda. This was a contract for single trading voyages. A sedentary investor (the stans) provided capital. A travelling merchant (the tractator) carried the goods, managed the trade, bore the physical risk. Profits were split: 75/25 in a unilateral commenda (where the investor put up all the capital), 50/50 in a bilateral one (where the tractator contributed a third).

What made the commenda matter was not its commercial design. It was what it did to the class structure. A young Venetian with no family wealth could enter the most profitable business in the Mediterranean. Government records from 960, 971 and 982 show that between 65% and 81% of names appearing in elite commercial documents were new in each generation. The economy’s upper reaches were being constantly refreshed.

That commercial openness drove political openness. The Doge’s power was progressively constrained. The Great Council was established. Independent magistrates, courts, a court of appeals, private contract law, bankruptcy law, and proto-banking institutions followed. Venice was building, layer by layer, what looked like the first inclusive society in European history.

But each new generation of commenda-enriched merchants diluted the economic and political power of established Venetian families. The elite who had themselves risen through open institutions now faced displacement by the next wave. So they acted to protect their position, triggering a series of institutional closures that would become known as the Serrata.

The Serrata (1286-1324) was not a single event but a sequence. First, political closure: the Great Council restricted membership to those already serving (1297), then formalised a hereditary aristocracy via the Libro d’Oro (1315). Second, economic closure: the commenda was banned (1314), trade was nationalised via state galleys, and heavy taxes were imposed on individual traders (1324). A police force was created for the first time in 1310.

Venice’s population, which had reached 110,000 by 1330, began to contract. Long-distance trade became the exclusive preserve of a narrow nobility. The institutional machinery that had generated the city’s wealth was dismantled by the very people who had benefited from it. Venice went from economic powerhouse to museum.

Critical Junctures and Institutional Drift

Venice introduces the book’s second major concept: the critical juncture. A moment of systemic disruption that destabilises existing arrangements and opens a window for change in either direction.

The Black Death is the canonical example. It killed roughly half of Europe’s population and created a massive labour shortage. In Western Europe (England especially), this tipped the balance of power toward peasants: serfdom weakened, wages rose, mobility increased. In Eastern Europe, the same shock produced the opposite: landlords consolidated power and imposed the Second Serfdom, the reimposition and intensification of feudal labour obligations that bound peasants more tightly to the land than before the plague.

Why the divergence? Acemoglu and Robinson argue it came down to small pre-existing institutional differences. Western European peasants already had slightly more bargaining power before the plague: English manorial courts gave them some legal standing, and the feudal system was already loosening at the margins. Eastern European landlords had slightly more consolidated control over their peasants and over local political structures. The plague did not create the divergence. It amplified differences that were already there but too small to matter under normal conditions.

This is what Acemoglu and Robinson call institutional drift. Like genetic drift in isolated populations, small differences in institutional arrangements accumulate over time. When a critical juncture hits, those small differences get amplified. England’s slightly weaker monarchy in the 1600s meant that private merchants controlled overseas trade rather than the Crown. Those merchants accumulated economic power, then political power, then forced the Glorious Revolution of 1688, which created the institutional foundation for the Industrial Revolution a century later.

The sequence matters: political institutions shape economic ones, which reshape political ones, in either a virtuous circle or a vicious circle. Virtuous circles are self-reinforcing but not self-perpetuating. Vicious circles are sticky but not permanent. Both can be disrupted at critical junctures, and the outcome depends on who holds what cards when the disruption hits.

The Iron Law of Oligarchy (and Its Exceptions)

The book borrows from Robert Michels, a German sociologist whose Political Parties (1911) argued that all organisations, even democratic ones, inevitably concentrate power in a leadership class. Acemoglu and Robinson apply this as the Iron Law of Oligarchy: the tendency for revolutions against extractive institutions to simply replace one extractive elite with another. Sierra Leone’s independence produced domestic elites who governed with the same extractive logic as the colonisers. Ethiopia’s Derg overthrew the Emperor only to impose equally brutal extraction. If the architecture remains extractive, whoever sits at the top will be shaped by its incentives.

The exceptions prove instructive. Botswana, at independence in 1966, had 22 university graduates and seven miles of paved roads. But it had pre-colonial institutions (the Tswana kgotla system) that included collective decision-making and constraints on chiefs. Its first president, Seretse Khama, chose to build on those inclusive traditions. When diamonds were discovered, revenues were shared nationally. Botswana is now the highest per-capita income country in sub-Saharan Africa.

The lesson: institutional starting conditions at critical junctures determine the range of what leadership can achieve. Khama had institutional raw material. Mobutu did not.

Growth Under Extraction: The China Question

The book’s treatment of China is its most speculative and, a decade later, its most testable. Acemoglu and Robinson argue that Chinese growth since Deng Xiaoping’s reforms is real but inherently limited, because it relies on technology adoption and capital investment rather than creative destruction. Political institutions remain extractive: no independent judiciary, no free press, no pluralistic constraints on executive power.

Their prediction: Chinese growth will slow unless political institutions become more inclusive.

The picture since 2012 complicates this. China has built genuine, world-leading technological strength: dominant positions in battery technology and EVs (BYD, CATL), solar manufacturing (over 80% of global capacity), autonomous vehicles, manufacturing robotics, and increasingly competitive frontier AI research. This is not pure catch-up; much of it is original work at the technological edge. The Acemoglu framework has to reckon with the fact that extractive political institutions have, so far, coexisted with serious innovation in specific sectors where the state sees strategic value.

At the same time, the post-2020 regulatory interventions (Ant Group’s IPO pulled, Didi delisted, the private tutoring sector shut down) suggest real tension. These were not conventional regulatory actions but abrupt assertions of political authority over sectors that had grown powerful enough to operate with significant autonomy. Whether this represents the kind of institutional closure the book predicts, or a different model of state-directed innovation that the framework doesn’t fully account for, remains the open question.

Where the Framework Strains

Two main lines of serious criticism.

First, the monocausal frame. Jared Diamond argued that geography and biogeography create the initial conditions that make certain institutional paths more or less likely. Acemoglu and Robinson acknowledge institutional drift but never fully grapple with why certain regions drifted in certain directions.

Second, the elasticity of “inclusive.” The post-Glorious Revolution Parliament was inclusive by 1688 standards but extractive by 1928 standards. The United States had formally inclusive institutions while operating chattel slavery. Critics have noted this can shade into circularity: successful countries had inclusive institutions, and we know their institutions were inclusive because they succeeded.

A third critique, less commonly raised: the book has little to say about epistemic power concentration. Property rights, contract law, market access: these are the institutional tools it analyses. But AI concentrates a different kind of power, cognitive and epistemic, that maps imperfectly onto existing categories. The commenda worked because capital and labour were separable. When the “labour” is performed by a model and the “capital” is compute, what institutional form distributes the surplus?

The Conceptual Toolbox

Prosperity, Acemoglu and Robinson insist, is always a choice encoded in institutions, never a gift bestowed by geography or culture. Venice did not decline because the Mediterranean shifted or because Venetians lost their commercial instincts. It declined because a narrow elite chose to close the system that had made them rich. The question the book leaves us with is not whether this will happen with AI. It is whether we are building the institutional architecture to prevent it.