Mancur Olson’s The Rise and Decline of Nations doesn’t provide a particularly optimistic picture: once your nation has been stable for a while, and may even have risen to wealth, it becomes more and more vulnerable to “institutional sclerosis.” This happens because small groups are better able to overcome free-riding, resulting in their ability to effectively skew the system towards their own interests. As more and more of these groups emerge, survive and are able to reap their rents—protected from that competition which makes for general progress and growth—the overall system deteriorates.

If you take Olson’s work to its logical conclusion, a very effective cure for economic stagnation is a catastrophic war. That is obviously not a desirable solution. But Olson was pointing to a real issue: the longer a society remains stable, the more it gets choked by special interest groups. These “distributional coalitions” aren’t interested in growing the economic pie; they just want to use the government to protect and grow their piece. Over time, their relentless rent seeking makes the entire system sclerotic, and Olson noted that it historically took massive shocks—like the total devastation of Germany and Japan in WWII—to wipe the “institutional slate” clean. Stripped of their entrenched lobbyists, those nations were in formidable positions to unleash economic growth.

However, relying on systemic collapse or war to clear out rent seekers is obviously not a viable policy prescription. We need a peaceful mechanism to achieve this Olsonian clean “institutional slate,” and this is where artificial intelligence enters the picture as a potential systemic shock.

To understand how this mechanism could work, let us apply some systems thinking to a concrete example: the notoriously complex German tax system. Currently, the sheer density of German tax law acts as an artificial barrier to entry, generating massive rents for a specific distributional coalition—the tax consultants, bureaucrats responsible for overseeing taxes, and politicians able to hand out rents to favored groups. Because navigating the bureaucratic maze requires highly specialized human capital, these groups hold a lucrative position. Consequently, they possess a strong incentive to lobby against any meaningful tax simplification, as doing so would destroy their business model.

Artificial intelligence presents an exogenous technological shock that can shatter this stagnant situation. If AI can parse and execute complex tax codes at a fraction of the cost, the economic foundation of the tax consultant industry is effectively erased. As the sector’s revenues dry up, its financial capacity to fund lobbying efforts shrinks simultaneously. Without a powerful, well-funded rent-seeking group actively demanding the preservation of tax complexity, the political friction preventing reform dissipates. In this scenario, technology clears the slate, drastically reducing the coalition’s lobbying power and finally making meaningful legislative reform possible.

Yet, this not only sounds too easy to be true, it also overlooks one thing: the resilience of entrenched coalitions. Olson explicitly noted that distributional coalitions inherently try to slow down a society’s capacity to adopt new technologies to protect their status quo. Before AI can fully erode their lobbying power, incumbent industries should be expected to engage in creative rent seeking. The tax coalition, for instance, is highly likely to lobby the government to mandate that AI-generated tax submissions remain legally invalid unless reviewed and stamped by a certified human professional, citing familiar justifications like “data privacy” or “liability.” That is, those who benefit from rent seeking will cloak their self-interest in noble defences of, say, “tax justice” or invoke the dangers of “algorithmic bias.” Rather than fading quietly, the incumbent coalition will likely pour all its resources into a final, fierce lobbying blitz to regulate AI out of existence before it can scale. This inevitable backlash will make for a difficult political battle in the coming years.

Winning this battle requires us to recognize the broader political economy at play. If we want to escape the trap of institutional sclerosis, we must understand that AI is not a magic bullet. Rather, it essentially is a window of opportunity to weaken and overcome entrenched distributional coalitions. But to use this window and unleash Schumpeterian creative destruction, we must fiercely push back against these coalitions who will try to regulate these new technologies out of existence. This could, for example, involve raising the awareness of the notoriously difficult to organize big groups, i.e., of the citizenry at large. This would certainly be a task for economists and Hayekian second-hand dealers of ideas consciously defending innovation and the freedom to innovate against the rent seekers who wish to protect their rents. What is certain in all of this is that the battle for the clean slate will not be won automatically; it requires action on our part.