Why Autocracies Won
We tried to change them, they are changing us.
In September 2019, Boris Johnson prorogued the British Parliament. Five weeks of parliamentary session, suspended by executive decree, justified by the constitutional convention that governments may seek a prorogation at any point. The Supreme Court ruled the action unlawful in a unanimous judgment — eleven judges, zero dissents. The Prime Minister had attempted to bypass the legislature because the legislature was in his way. The courts stopped him. The system, on that occasion, worked.
What made the episode significant was not the prorogation itself but the reasoning behind it — the reasoning that a significant portion of the public found, if not exactly acceptable, then at least comprehensible. The democratic process had produced a deadlock. The democratic process was generating friction. The democratic process was, in the most literal sense, getting in the way of getting things done. And the question that the prorogation raised — not loudly, not in the terms in which it would need to be raised honestly, but unmistakably — was whether the friction was worth it. Whether the checks, the balances, the procedural protections, the institutional vetoes were producing outcomes proportionate to the delay and the cost they imposed.
The question was not new. It has been building for thirty years. What is new is that we now have a rigorous answer, and the answer is uncomfortable.
The argument that capital requires democracy to flourish was one of the organizing assumptions of the postwar liberal order. The Bretton Woods framework, the Marshall Plan, the construction of the European project, the entire institutional architecture of Western-led globalisation rested on the premise that economic openness and political freedom were mutually reinforcing — that the prosperity produced by liberal capitalism would sustain the political will for liberal democracy, and that liberal democracy would in turn produce the rule of law, the contract enforcement, the stable property rights framework that productive capitalism required. The two were inseparable. You could not have one without the other.
What the past thirty years have demonstrated, with a clarity that the architects of that order never anticipated and their successors have been slow to acknowledge, is that this premise was historically contingent rather than structurally necessary. Capital does not require democracy. It requires stability, rule of contract, and the absence of the kind of friction that slows decision-making, constrains accumulation, or redistributes returns. When democracies provide these conditions efficiently, capital routes through them. When they do not — when democratic accountability produces regulatory complexity, when electoral cycles produce policy uncertainty, when organized labour produces wage floors, when environmental regulation produces compliance costs — capital routes elsewhere.
The elsewhere, for the past thirty years, has increasingly been China.
China's accession to the World Trade Organisation in December 2001 was the moment the assumption broke openly. The architects of the accession — the Clinton and Bush administrations, the WTO membership, the multinational corporations that lobbied for it — operated on the explicit premise that economic integration would produce political liberalisation. History would resolve the tension between capitalism and authoritarianism in capitalism's favour. The Chinese Communist Party would, under the pressure of market forces and the aspirations of an expanding middle class, gradually converge toward the political forms that Western liberal democracies had developed.
The opposite happened. China grew its economy at rates no democratic system has sustained. It built infrastructure at speeds no democratic planning process has matched. It coordinated industrial policy at a scale that no coalition of private actors operating within a market system has achieved. It absorbed the manufacturing base that democratic economies had developed across a century, employed the labour that democratic economies had displaced, and accumulated the capital that democratic economies had released — all while maintaining a political system that made none of the concessions the accession's architects had predicted. The market did not liberalize China. China disciplined the market.
The market did not liberalize China. China disciplined the market. And the global capital system, indifferent to the political form of its host, noticed and routed accordingly.
The global capital system, indifferent to the political form of its host, noticed and routed accordingly. The river does not have preferences. It has physics. Capital flows toward efficiency, toward lower costs, toward faster decisions, toward the absence of the procedural protections that democratic systems exist to maintain. The architects of liberal democracy built the banks — the institutional framework that gave global capital its productive form. They then removed those banks in the name of the river's freedom. The river found the most efficient channel available. That channel ran through Beijing.
The standard defence of democracy against this argument has always been the long game. Yes, autocracies coordinate faster. Yes, they execute more decisively. Yes, they eliminate the friction that slows democratic governance. But — and this was the argument that Western liberals produced with some confidence for most of the postwar period — they also eliminate the conditions for genuine innovation. The dissent. The failure. The lateral thinking. The distributed experimentation that only a relatively open society permits. The innovation premium was democracy's structural compensation for its slowness. Silicon Valley, not a state planning bureau, built the internet. The West, not the Soviet system, won the technological race that determined the Cold War's outcome. Human cognitive freedom, it turned out, was not just a political value but an economic and strategic asset. It was worth the friction.
This argument is now being dismantled. Not refuted — dismantled. From two directions simultaneously, and by the same technology.
The foundation models, the large language models, the infrastructure of artificial intelligence were developed inside liberal democracies — Stanford, MIT, Google DeepMind, Anthropic, OpenAI. The innovation premium produced the technology. The question is what happens next. Authoritarian states are now deploying and scaling AI with a speed and coordination that democratic systems — with their regulatory debates, their distributed decision-making, their institutional caution, their electoral cycles — cannot match. The innovation advantage produced the tool. The efficiency advantage now determines who deploys it fastest and at what scale. China's state capacity for coordinated AI deployment is, on current evidence, substantially greater than the capacity of any democratic government to coordinate a comparable effort. The West invented the engine. It is not obvious that the West wins the race.
The second movement is more fundamental. If the machine can generate novel combinations, test hypotheses, and produce scientific and creative outputs at scale — if it can do, in other words, a substantial portion of the cognitive work that the innovation premium depended on — then the distributed human creativity that justified democracy's slowness becomes structurally less necessary. The strategic asset that compensated for democratic friction is being automated away at precisely the moment when the efficiency argument for authoritarian coordination is at its strongest. The long game argument assumed that human innovation required human freedom. That assumption is no longer secure.
There is a further dimension that the standard political debate has not yet reached, and that is worth naming clearly because it will become unavoidable within a decade. If artificial intelligence concentrates productive capacity in a sufficiently small number of hands while simultaneously displacing human labour across entire categories of work — not the cyclical displacement of previous technological transitions, which created new categories of employment to absorb the displaced, but a structural displacement from which the jobs do not return because the machines do the work permanently, at zero marginal cost — then the political question of the coming decades will not be growth. It will be redistribution. At a scale and speed that liberal democratic institutions were not designed to manage.
The property rights architecture, the constitutional protections for accumulation, the non-majoritarian constraints that the postwar settlement embedded in its institutional framework — all of these were designed to prevent the kind of forced redistribution that the architects of that order had watched destroy European societies in the first half of the twentieth century. They are now, potentially, the obstacle to the only correction adequate to the problem. The institutional forms that could address structural mass unemployment combined with extreme concentration of AI-generated wealth — mandatory redistribution of AI returns, the socialisation of AI infrastructure, state capacity for forced economic reorganisation — require precisely the centralised power and the subordination of property rights to collective need that the postwar constitutional framework was designed to prevent.
The structural irony is not subtle. The system the architects built, having concentrated productive capacity beyond any previous historical precedent while rendering human labour structurally surplus, may require for its own correction precisely the institutional forms their ideology spent a century ruling out. The democratic architects built the banks. The river found its level. And the flood that followed may only be containable by the methods of those who never believed in banks at all.
None of this is an argument for autocracy. The historical record of centralised state models as instruments of human wellbeing is catastrophic and documented. The point is not that authoritarian systems are better. The point is that liberal democracy is losing the arguments it used to win — the efficiency argument, now settled in autocracy's favour by three decades of comparative evidence; and the innovation argument, now being dismantled by the technology that liberal democracies themselves produced. What is left, as a justification for accepting the friction, the slowness, the procedural complexity that democratic governance requires, is the argument from values: that human dignity, political freedom, and the rule of law are worth the economic cost, regardless of the competitive consequences.
That argument is true. It is also, on current evidence, losing.
The communities the river left behind — the deindustrialised regions, the displaced workers, the populations whose economic and cultural identity were simultaneously dissolved by the globalisation project — have drawn their own conclusions about the value of the institutional framework that produced their situation. They have produced the unmediated figures, the executive overreachers, the leaders who treat the democratic process as an obstacle to getting things done. They have done so not because they are irrational but because they have correctly identified that the system as it currently operates is not working in their interests, and that the people who designed and maintain it have consistently prioritised its theoretical virtues over their concrete conditions.
The architects built something remarkable. They then dismantled the conditions that made it work, in the name of the freedom of the thing they had built. And the question that the communities they left behind are now asking — the question that the next decade of AI-driven concentration will make impossible to avoid — is whether the remarkable thing was ever really for them at all.
This essay is drawn from Chapter Four of The Dream of Never: Culture, Capital, and the Refusal of Limits, which traces the hollowing of postwar democratic institutions from the architects' paradox through the discovery of autocratic efficiency to the AI redistribution problem. Available on Amazon Kindle.
michelerovatti.com · Writing since 2012
Related Essays & Book
The following essays explore related themes surrounding capital, culture, democracy, ideology, and contemporary political consciousness developed throughout The Dream of Never.
The Dream of Never
Culture, Capital, and the Refusal of Limits
by Michele Rovatti
A sharp diagnosis of the growth imperative and the cultural logic that made it inevitable — from theology to AI.
📘 Get your copy on Amazon