The Chicago School believed that markets operate best when they are left to organize themselves, and that state regulation only disrupts this process to the detriment of social welfare. Brazil’s Competition Act for Digital Markets, now under legislative debate, reveals the myopia of the Chicago School’s organizing principle and instead supports the heterodox institutionalist position that state activity, whether through activity or abstention, is always at the heart of market formation, writes Diogo Rosenthal Coutinho.
In the Chicago School account that still organizes competition policy worldwide, markets are default states: self-organizing and efficient unless disturbed. And law should step in only to correct discrete failures, such as monopolization, anticompetitive contracts, or negative externalities. Friedrich Hayek gave this view its philosophical anchor: a market order is cognitively superior to any order a planner could design. Those closest to commercial activity know best how to organize it. On this account, regulation should be limited to market fixing, and the law should show up only after the market has organized itself through the activity of individuals and individuals collected into corporations. Anything more would be politicized and an act from a position of inferior information. Therefore, proactive regulation would hamper the spontaneous order it is meant to protect.
The New Institutional Economics offered a more sophisticated version of the same view. Douglass North and Oliver Williamson brought back institutions into the neoclassical economics of the Chicago School. In this view, institutions help reduce transactions costs or imperfect access to information. So, disclosure and licensing requirements from the Securities and Exchange Commission or Federal Deposit Insurance Corporation could make information more transparent and help investors and customers make more informed decisions. However, state institutions primarily exist to guarantee contracts and define property rights. Empowering state institutions is risky as they are often driven by self-interested actors and special interests that redistribute or rent-seek to the loss of general welfare. Thus, according to New Institutional Economics, institutions best serve society as limited constraints that lower transaction costs and secure property rights. Their followers in development policy turned that insight into a portable checklist of good institutions to be installed anywhere. Both the Chicago School and New Institutional Economics (there was significant overlap in membership) measure law against a benchmark of efficiency defined before any particular market exists.
Markets, however, do not exist in a state of nature, and law does much more than patch them. It helps create and shape them. Digital markets make this point plain.
Digital platforms became infrastructures for commerce, cultural production, and public debate under legal rules that decided who bears liability for content, who controls data, and on what terms rivals can reach users. For example, rules about free speech in the United States allow platforms like Facebook and X to host misinformation, whereas rules in the European Union require these platforms to moderate or even remove some speech, including bullying and large disinformation campaigns, or risk fines. These laws have tailored how digital platforms developed and the impact they have on society. The law preceded and shaped the market activity.
As countries around the world evaluate and respond to the impact that digital platforms have had on competition, consumer protection and welfare, and democracy, getting digital market regulation right therefore requires treating law as deliberate institutional architecture, built in anticipation of the capacities, dependencies, and distributive effects a given market structure will produce. Such a disposition treats regulation and antitrust as public policies that build markets, rather than simply responding to them in patchwork efforts to bandage market harms, in other words.
Brazil’s Congress is currently debating a bill that works this way. Bill 4675/2025, also known as the Competition Act for Digital Markets, would give the country’s competition authority, the Administrative Council for Economic Defense (CADE), a new mandate to designate and regulate systemically relevant digital platforms. The Act would treat market structure itself as the object of regulatory policy. It is an unusually clear specimen of what the political scientist Steven Vogel calls marketcraft: the idea that governments do not just referee markets, they build the rules, infrastructures, and capacities that make markets exist as markets at all.
Markets as institutional architecture
The idea is older than the law and economics movement associated with the Chicago School, which gained prominence during the 1970s. At the turn of the twentieth century, the early institutionalists Thorstein Veblen, John Commons, and Wesley Mitchell, along with the legal realist Robert Hale, relocated economic analysis away from the abstract exchange between pre-formed individuals posited by neoclassical theory and toward habits, organizations, and the distribution of coercive power that any transaction presupposes. Hale’s insight bears directly on antitrust: every market outcome reflects a distribution of rights and bargaining power that law has already set, so any judgment of efficiency presupposes that distribution.
Karl Polanyi pushed furthest, arguing that the self-regulating market of nineteenth-century liberalism was a deliberate, and in his account disastrous, political project. His evidence looked at legislation: laws by the British parliament promoted the enclosure of common grazing and farming land into private property, which overturned community relations and installed a landlord regime that accrued wealth to the aristocracy and wealthy. Meanwhile, the Poor Law Amendment Act of 1834 created a competitive labor market by leaving the poor no alternative to wage work. “Laissez-faire was planned, planning was not,” Polanyi famously wrote. Its central danger was “disembedding”: tearing the market loose from the social relations it was born from and meant to serve.
Ha-Joon Chang and Peter Evans drew the distinction that separates this tradition from both Chicago and the New Institutional Economics. Institutions can be read thinly as constraints that correct a market’s imperfections and move it toward an efficient benchmark. Or, institutions can be read thickly as enabling and constitutive structures that make entire forms of coordination possible and shape the interests and expectations of the actors operating inside them. On the thick reading, there is no single efficient design waiting to be copied, because which market exists, whom it serves, and how its gains are distributed are themselves outcomes of institutional choice. State capacity, accordingly, is a government’s ability to set goals, coordinate dispersed actors, and build the organizations needed to pursue them, a condition Evans called embedded autonomy: state institutions that operate close to economic actors without being captured by them. Laura Edelman and Robin Stryker add the methodological payoff: studying a market means studying the legal, organizational, and social processes that constitute it. Law, on this view, is part of the economic reality it regulates, both materially and cognitively.
Brazil’s digital marketcraft in action
Brazil’s digital markets grew inside a legal framework that presumed it best to restrain the role of the state. The 2014 Marco Civil da Internet shielded platforms from civil liability for user content unless they failed to comply with a specific court order, and it wrote net neutrality into statute. Those rules shaped what content platforms could profitably host and how fast they could scale. In June 2025, the
Brazilian Supreme Court declared that shield against liability partially unconstitutional and imposed a duty of care over the mass circulation of serious illegal content, such as hate speech, misogynistic opinions
or fake news. The Court invoked fundamental rights and democracy in its ruling, terms alien to the law & economics movement of the Chicago School. The effects on public debate usually attributed to platforms were, in part, effects of the liability regime the law had chosen.
Brazilian competition law did its part as well to limit its activity in digital markets. Under the 2011 Competition Act, CADE’s practice in unilateral conduct cases came to rest on effects-based analysis of individual firms, the approach the Chicago School made orthodox. This allowed markets to concentrate, for the proliferation of unfair practices and abuse by dominant firms, and the for corrosion of democratic ideals and institutions through inequality and misinformation. Where the Brazilian state has chosen instead to build market infrastructure, the result was a different market. For example, in 2020 the Central Bank launched Pix, a public, interoperable instant-payment system that opened payments to fintechs and other new entrants. The Central Bank made a “market shaping” decision to intervene at a layer of the financial system which would take away market power from, in particular, the dominant American payment networks operated by Visa and Mastercard. Not only does Pix encourage innovation and market entry in adjacent markets, but it also expands banking to Brazil’s least well off.
The Competition Act for Digital Markets turns these recent legal and policy shifts toward institutionalism into statutory design. Rather than create a freestanding regulator on the model of the European Union’s Digital Markets Act, it amends the existing competition law and builds a dedicated Digital Markets Superintendency inside CADE. The Superintendency would designate as “systemically relevant” platforms that cross revenue thresholds and meet qualitative criteria such as control over data, network effects, and vertical integration across an ecosystem.
Digital markets pose a problem that case-by-case enforcement alone cannot solve. Network effects and accumulated data mean that once most users and sellers converge on one platform, rivals cannot reach the scale needed to compete. Fines imposed years later rarely reverse that outcome. Contestability therefore has to be built into these markets before network effects and data advantages tip them into monopoly.
CADE’s own record shows why such a design is needed. For years the authority dealt with platforms case by case, one dominant firm and one alleged abuse at a time. In 2019, it closed the investigation into Google for favoring its own comparison-shopping service with a single, tie-breaking vote, two years after the European Commission had fined Google for the same conduct. A 2023 settlement addressed iFood’s exclusivity contracts with restaurants only after the company had become the dominant food-delivery platform. These proceedings gave CADE a granular, hard-won understanding of how network effects compound and how data advantages entrench incumbents. They also exposed the limits of the method the authority inherited, which requires proof of harm, firm by firm, only after the harm has occurred and shape a market.
Designation of “systemically relevant” is only the first step in the new regulatory procedure. Rather than apply a uniform list of obligations automatically, as the EU does through its Digital Markets Act, CADE would open an individual proceeding for each designated firm and negotiate tailored commitments on matters such as data portability, interoperability, self-preferencing, and alternative payment methods. The model draws more from Section 19 of the German Competition Act and the United Kingdom’s Digital Markets, Competition and Consumers Act.
These obligations matter for what they make possible. A user who can port her data can switch services, a developer who can steer customers to a cheaper payment method can compete on price, and a seller who can work across several marketplaces depends less on any one of them. Tailoring carries a risk of its own: it could slide back into case-by-case adjudication. Designation rests on structural criteria, and obligations follow from a firm’s position in the market without a prior finding of abuse.
Building, not finding
None of this credits law with a magical power to design social outcomes at will. Institutions are produced by multiple actors, constrained by prior trajectories, and capable of generating effects nobody intended, including the capture of the very superintendency the bill would create, as the theorists of the New Institutional Economics and other approaches feared.
The fight over the bill illustrates this point. The Chamber of Deputies approved an urgency motion in March 2026 to vote on the bill, but it has still not reached the floor. The rapporteur has proposed narrowing the designation criteria and capping CADE’s ability to expand its own jurisdiction. Industry lobbying has tried to influence the design of thresholds and the timeline, and several observers now expect a floor vote only in 2027, depending on the political realignment that follows the 2026 elections. Where the thresholds land and how much discretion CADE gets will decide which firms count as gatekeepers and what they owe. At the end of the day, the fight over Competition Act is a fight over how to structure the digital market.
However, the alternative to the fight to design institutions is not a political economy free from capture. Treating digital markets as pre-existing orders that any law would disturb is itself an institutional choice, one that has historically favored incumbents with the resources to exploit whatever ambiguity the law leaves in place. Brazil’s experiment, legislating later than the EU and the U.K. and borrowing from both, is worth following for a reason that goes beyond antitrust policy. It treats market architecture as something to be built deliberately and iteratively, matched to the state’s own administrative capacity, without waiting for the market to reveal itself or importing a best-practice rulebook wholesale. Whatever version of the Competition Act eventually passes, if any, the debate around it has already shown what the early and modern critical institutionalist tradition argued all along: markets, digital ones very much included, are made, not found. Who gets to do the making, and with what distributive effects, is a political question law cannot avoid answering.
Author’s Disclosure: The author reports no conflicts of interest. You can read our disclosure policy here.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
Subscribe here for ProMarket’s weekly newsletter, Special Interest, to stay up to date on ProMarket’s coverage of the political economy and other content from the Stigler Center.
