For centuries, Europe has been politically fragmented. Counterintuitively, this fragmentation was a blessing, not a curse. The reason was that, unlike other absolute monarchs, such as Chinese emperors, European kings and queens had far less capacity to ban or monopolise new technologies. They were limited by power competition: European countries lived in constant war, and rulers were often quick to embrace what their neighbours sought to suppress.
As Matt Ridley aptly put it in his book, the printing press offers a revealing example. Although it was first invented in China during the Ming Dynasty, where its use was subject to strict state control and censorship, it was Johannes Gutenberg who revolutionised printing. Around 1440, he created the first movable metal type printing press in Mainz – now part of Germany. However, and crucial for our case, as a result of political unrest and guild conflicts in Mainz, Gutenberg moved to Strasbourg – now part of France – where he continued improving his prototype.
The same holds true for intellectual ideas. Thinkers often moved from one country to another when they clashed with national rulers or when their ideas threatened the established order. Hobbes, Locke, Voltaire, and Descartes are among the most well-known examples. Their ideas transformed Europe and proved critical to Europe’s political and economic development. Had these thinkers not been able to move abroad and continue their work, many of the ideas they are now celebrated for might not have been developed in Europe at that time.
Europe’s political fragmentation enabled, rather than hindered, its economic development, by preventing the suppression of inventions and good ideas: from the printing press to property rights, freedom of speech, and the scientific method.
The argument that market fragmentation might be a good thing is, for economists, a difficult pill to swallow. We’ve been trained on the notion that lowering market barriers is fundamental to economic development: it encourages competition and specialisation, increases variety, and reduces prices. In essence, market integration is synonymous with economic prosperity and a better quality of life.
Tackling Europe’s market fragmentation has long been the leitmotif of the EU project. One of the main jobs of the EU is to approve regulations that ensure goods, services, capital, and people can flow freely from one EU country to another. A common set of rules across the 27 member states means less market fragmentation. The EU Single Market is, without doubt, one of the most impressive economic and political achievements that the EU has accomplished.
However, what if an EU regulation is flawed, or simply unnecessary? What if, dare one say it, some EU countries might have preferred not to regulate AI or take a different approach to industrial policy? We may never know the answer. But the lessons from history are hard to miss: the uniform application of a bad regulation – such as banning the printing press – can be far worse than a fragmented market.
So, how do we fix this? We don’t want fragmented markets in the EU, or beyond. But nor do we want damaging regulations. So how do we weed out bad ideas while keeping the good ones flowing?
We Need to Start Thinking Like IT Engineers
Software isn’t developed like traditional products. Tech companies often start with a prototype that is, by any standard, bad. Deliberately so. The goal isn’t to sell the prototype, but to gather early feedback from users. This approach saves money by avoiding the trap of investing heavily in features that customers neither want nor value. This iterative process continues until the software is ready, and often well beyond, with updates issued to fix mistakes or add new features even after release. At the heart of this approach is the belief that not everything can be anticipated, and that partial failure is tolerated, as a feature, not a bug.
We propose a new policy approach that brings the iterative mindset of IT engineers into the framework of Better Regulation. The EU already follows a methodology to assess the costs and benefits of regulation and to model their likely impact on the economy. This approach is undeniably useful, helping policymakers identify and reflect on the potential economic effects of regulation. However, these effects are still modelled estimates, and both known unknowns and unknown unknowns inevitably remain.
The policy will proceed if its estimated benefits outweigh the costs. Yet the EU should aim higher. New EU regulation should not just be better, but the best. To that end, the EU ought to revive the spirit of natural selection in policymaking inherent in the competitive dynamic between European nations. Achieving this requires that new regulations should be tested, and this testing should happen earlier in the process, much like the iterative approach common in IT development. We therefore propose the creation of a regulatory sandbox for regulators.
A regulatory sandbox is a controlled testing environment where firms can trial innovative products, services, or business models under relaxed regulatory conditions and with regulatory oversight. The EU, EU member states, and non-EU countries have already implemented some of these sandboxes. However, in this case, it would be the EU and national civil servants, not firms, who would be testing their new policy ideas.
EU policymakers could select a small (but representative) group of companies and pay them for testing an early-stage version of the regulation. In the sandbox, EU civil servants will actively test new ideas and regulations, much like tech developers test early-stage software. By doing so, they will gather feedback, iterate, and refine regulations before they are fully developed.
Of course, this approach is not without risks. Companies testing the regulation may overstate the costs and minimise the benefits. Moreover, the selection of the companies testing the regulation should be as random as possible to avoid the problem of self-selection. To mitigate some of these issues – which are also inherent in any call for evidence prior to regulation – the testing process should be open and transparent, with the results included in the stakeholder consultation that is already part of the regulatory process. For EU countries, regions, and local authorities, the iterative approach to public policy can also be applied to testing changes in the delivery of public services.
This approach aims to replicate what a fragmented Europe achieved in the late Middle Ages: letting the best ideas rise to the top, while quietly retiring those that fall short. Just as political fragmentation during that time acted as a pressure valve for bad ideas and a springboard for good ones, a regulatory sandbox for regulators can help the EU combine its Single Market ambitions with a culture of regulatory trial and error. In doing so, we avoid the mistakes of the Chinese Ming Dynasty and build a regulatory model fit for the age of innovation.
I fully acknowledge that testing regulation is neither simple nor risk-free. However, for those who view this iterative approach as impractical, the takeaway should be a broader lesson about humility in regulation, especially when it comes to fast-evolving technologies. Even the most rigorous ex-ante cost-benefit analysis can obscure variables beyond the control of policymakers, or miss unknown unknowns that later prove critical. The lesson from history is clear: when in doubt, policymakers should hold back from regulating and limit new technologies until their effects, and needs, are better understood. Regulation should not be Europe’s default response to every new challenge, but a last resort.