Regulatory independence: when formal autonomy is not enough

I went to Tbilisi prepared to talk about technology. I came back thinking largely about institutions.

On the first morning of the IX World Forum on Energy Regulation (WFER IX), held in Georgia in September 2026, breakfast at the hotel already gave a good sense of what the following days would be like. There were regulators and experts from many countries, speaking English in very different accents and representing power systems that had little in common with one another. Europe, Africa, Asia, the Americas. Mature markets, markets in transition, countries with a strong state presence, and others with structures more open to competition.

The forum’s program reflected that diversity. There were discussions on artificial intelligence, digitalization, networks, consumers, the energy transition, incentives, investment, resilience and new market models. Regulatory independence and accountability were also on the official agenda from the start. [1][2] I expected technology to be the subject that would follow me home. It was not.

Over the course of the sessions, a more basic question became hard to ignore: what is the point of giving the regulator more and more responsibilities if it does not have the real means to carry them out? This question is not limited to the power sector. Nor is it only a Brazilian one.

Independence in law

In Brazil, the legal starting point is clear. Law 13,848 of 2019 defines the special legal status of federal regulatory agencies by their freedom from ministerial control or hierarchical subordination, their functional, decision-making, administrative, and financial autonomy, and security of tenure for board members during their terms. [3] This is no minor detail.

Regulated sectors are characterized by long-term investments, significant information asymmetries, major economic interests, and decisions whose effects can last for decades. The logic of an independent agency is precisely to create some distance between regulatory decisions and the short-term interests of any single actor, whether the government, a regulated company, or an interest group. But there is an important difference between autonomy in law and autonomy in practice.

That distinction came up very directly in one of the sessions I attended at WFER, devoted to the role of independent regulators in the energy transition. The discussion covered budgets, staffing, technical capacity, relations with governments, expanding mandates and organizational autonomy. At one point, the idea was put almost in these words: formal independence does not necessarily mean real independence. [4] The sentence stayed with me because it shifts the discussion.

It is relatively simple to check whether the law gives an agency a mandate, legal authority, and safeguards for stable decision-making. It is harder to ask whether, in its day-to-day work, that institution can exercise its mandate without depending too heavily on resources, information, or decisions controlled by others.

An agency may have decision-making autonomy yet lack enough staff to perform the necessary analysis. It may have statutory oversight authority yet lack the capacity to conduct all the inspections it needs. It may receive new responsibilities without corresponding resources, or retain rulemaking authority while losing the ability to develop its own knowledge about a rapidly changing sector. In all these cases, autonomy still exists on paper. The question is how much of it survives in practice.

Independence is not isolation

Before going further, a clarification. Defending regulatory independence does not mean defending an agency closed in on itself, immune to public scrutiny or free to decide without being accountable. That would be a poor, and dangerous, understanding of independence.

Regulators are part of the state. They need to engage with the government, Congress, companies, consumers, associations, academia, and civil society. They need to justify decisions, publish data, receive input, explain their choices, and remain subject to the oversight mechanisms established by law.

The OECD treats independence and accountability as complementary dimensions of good regulatory governance. Its work on economic regulators starts precisely from the distinctive position these institutions occupy between public authorities, regulated entities, and users of regulated services. Independence protects against undue influence; accountability prevents autonomy from becoming freedom from responsibility. [5][6]

This distinction between legitimate and undue influence is central. It is normal for a ministry to present a public policy to the regulator. It is equally normal for a company to explain how a given rule affects its investments, and for consumers to advocate for lower rates. Regulatory work often begins precisely when these interests do not point in the same direction.

The regulator does not need to stop listening to anyone. It needs to keep the ability to decide without becoming an extension of any of the parties it listens to. The independence that matters here is independence of judgment.

There is another kind of dependence

It was during the forum that another dimension of this problem seemed even clearer to me: a regulator can be legally independent and intellectually dependent. This happens when the regulated sector possesses far greater expertise, data, and analytical capacity than the institution responsible for regulating it.

In the power sector, this gap can grow quickly. We are talking about systems with more distributed generation, storage, electric vehicles, demand response, aggregators, grid digitalization, artificial intelligence, cybersecurity, new rate designs, and consumers who are increasingly becoming active participants. None of these changes falls within a single discipline.

The engineer needs to understand economic incentives. The economist needs to grasp the physical limits of the grid. The lawyer needs to deal with algorithms, data and new contractual arrangements. At some point, all of them run into subjects such as data protection, consumer behavior, cybersecurity or artificial intelligence.

In the WFER session on regulatory independence, this issue arose in connection with the need for teams capable of integrating technical, economic, financial, and legal perspectives. There was also concern about institutions taking on more responsibilities without a corresponding increase in resources. [4]

The OECD has been drawing attention to something similar. In its work on the governance and resources of regulators, it notes that funding and staff management constraints can limit the autonomy and agility of these institutions. It also points out that difficulty recruiting IT professionals and data scientists can limit a regulator’s ability to deal with technological innovation. [7]

This changes the way I see institutional capacity. Staffing, budget, data, technology, and training are often treated as administrative matters, but they are part of the infrastructure of independence.

When an agency can no longer develop its own knowledge, it becomes more dependent on information supplied by the regulated industry itself. That does not, in itself, amount to regulatory capture. It is important not to trivialize the concept. But it does increase information asymmetry, making it harder to test assumptions, compare alternatives, and identify risks that may not be voluntarily brought to the regulator’s attention. There is a major difference between listening to a regulated company because it understands its business deeply and depending on that company in order to understand the business at all.

A problem WFER showed is not uniquely Brazilian

This was perhaps one of the most useful insights from Tbilisi: very different regulatory systems were asking similar questions. How do you preserve autonomy as responsibilities grow? How do you keep qualified teams and adequately fund institutions that must keep up with technologically sophisticated sectors? How do you respond to political and economic pressure without turning independence into isolation? And how do you remain technically relevant when the regulated sector changes faster than the internal processes of the state?

The OECD’s most recent data help put the issue in perspective. The 2025 edition of its indicators on the governance of sector regulators analyzed 149 economic regulators in 42 countries and organizes regulatory governance into three broad dimensions: independence, accountability, and scope of action. [8] In Government at a Glance 2025, the OECD observed little progress on average in independence mechanisms between 2018 and 2023, even as regulators’ responsibilities expanded with the green and digital transitions. [9]

This does not allow us to say there is a uniform global crisis of regulatory independence. The contexts are far too different for that. But it does support a more modest and, in my view, more interesting conclusion: the real capacity of independent institutions to keep up with growing responsibilities has become an international concern. That is what I recognized in several of the forum’s discussions. And it was impossible not to think about Brazil.

The Brazilian case: formal autonomy under capacity pressure

Here again, it is worth separating facts from interpretations. Brazilian law recognizes the agencies’ autonomy. [3] At the same time, there is recent evidence of material constraints that warrants attention.

In February 2026, the Federal Court of Accounts (TCU), Brazil’s external audit body, released the results of an audit of four federal regulators: ANATEL (telecommunications), ANEEL (electricity), ANP (oil, natural gas and biofuels) and ANM (mining). The TCU found that, despite the autonomy provided for in Law 13,848/2019, budget constraints in recent years had affected these agencies’ oversight and administrative activities, as well as investment in technology and innovation. [10]

Months later, the Committee of Federal Regulatory Agencies (COARF) expressed concern about a new budget freeze which, according to its joint statement, meant a flat cut of roughly 18% in the spending and commitment limits of federal bodies. The committee linked the measure to possible effects on the agencies’ oversight work and operational capacity. [11]

These facts do not, on their own, prove any attempt at political interference. That needs to be said. Fiscal constraints exist, governments have to make budget choices, and not every cut in resources should be read as an attack on an institution’s independence.

The relevant question is a different one: at what point does a lack of resources start to compromise the exercise of the mandate the law itself established? It is a less strident question, but a far more useful one. And it should apply regardless of which government is in office. If regulatory independence is only defended when we agree with the regulator’s decision, then we are not defending independence. We are defending the outcome that suits us.

Are we asking for more and providing enough?

There is a contradiction that seems increasingly hard to sustain. We ask regulators to be faster, more digital, and more data-driven. We expect them to understand artificial intelligence, prepare for the energy transition, stay alert to cybersecurity, remain close to consumers, improve the sophistication of their economic analysis, be open to innovation, and assess impacts more rigorously. At the same time, we often treat staffing, training, data, and technology as support costs.

They are not. A public institution’s capacity to exercise independent judgment depends on the quality of the information and expertise it can marshal.

It is true that no regulator will have every specialist it needs in-house. Nor should it. Cooperation with universities, consultants, other agencies and international bodies will remain necessary. But there is a core of knowledge that cannot be outsourced: the ability to frame the questions, critically assess the answers and take responsibility for the decision.

Perhaps this is the most important point. Being able to say “no” is part of independence, but independence depends on something that comes before it: having enough knowledge to know when, why and on what basis to say yes or no.

Why this matters to people who do not work at an agency

It is easy to turn regulatory independence into an insiders’ conversation. That would be a mistake. Consumers do not need to know an agency’s administrative structure to feel the effects of a bad regulatory decision. Investors do not need to follow the regulator’s budget to feel the consequences of unpredictable rules. And a company does not need to agree with every decision an agency makes to see the value of knowing that those decisions follow reasonably stable criteria.

In infrastructure sectors, predictability matters because the horizons are long. A network, a power plant, a storage system or a major digital transformation are not investments designed to last one presidential term.

It is natural for elected governments to set policies and priorities. That is part of democratic legitimacy. It is also natural for regulation to take those policies into account within the limits of the law. The problem arises when the distinction between setting public policy and deciding, on technical grounds, how to carry out a regulatory mandate begins to disappear.

An independent agency does not exist to oppose the government. Nor to oppose companies. Nor to please consumers. It exists because, in certain sectors, society decided to create a specialized institution able to arbitrate conflicting interests with some degree of stability, technical knowledge and distance from immediate pressures.

This places an enormous responsibility on the regulator itself. Independence without technical quality is not enough. Nor is it enough without transparency, or without accountability. But the reverse also holds: demanding quality, transparency, and results from an institution without ensuring that it has the practical capacity to function produces autonomy in name only.

At the end of those days in Tbilisi, I realized I had gone to a forum on the future of energy and was coming back with a concern that comes well before the next technology. Before asking how to regulate artificial intelligence, storage, flexibility or active consumers, there is an institutional question that cannot be forgotten: will those who regulate be able to understand what is changing and make independent decisions about it?

Technology may change quickly. Market models will certainly change, and some of today’s discussions will look dated a few years from now. The need for institutions able to decide with knowledge, autonomy and responsibility probably will not.

Perhaps that is exactly why regulatory independence needs to be defended at all times, as part of the institutional infrastructure of markets that we want to be predictable, reliable, and able to evolve, not as a privilege of the regulator.

References

1. INTERNATIONAL CONFEDERATION OF ENERGY REGULATORS (ICER). World Forum on Energy Regulation: WFER IX, Tbilisi, Georgia, September 21 to 24, 2026. Available at: https://wfertbilisi2026.com/en. Accessed on: Sept. 28, 2026.
2. WORLD FORUM ON ENERGY REGULATION (WFER IX). General Program. Tbilisi, Sept. 21 to 24, 2026. Available at: https://www.wfertbilisi2026.com/en/programa/general. Accessed on: Sept. 28, 2026.
3. BRAZIL. Lei nº 13.848, de 25 de junho de 2019 [Law 13,848 of June 25, 2019, on the management, organization, decision-making process, and public oversight of regulatory agencies]. Art. 3. Available at: https://www2.camara.leg.br/legin/fed/lei/2019/lei-13848-25-junho-2019-788523-publicacaooriginal-158408-pl.html. Accessed on: Sept. 28, 2026.
4. WORLD FORUM ON ENERGY REGULATION (WFER IX). Concurrent Session 1B: Independent regulators as key decision makers of the energy transition. Tbilisi, Sept. 22, 2026. Based on the official WFER IX program and the author’s personal notes taken during the session.
5. OECD. Being an Independent Regulator. Paris: OECD Publishing, 2016. DOI: 10.1787/9789264255401-en. Available at: https://www.oecd.org/en/publications/being-an-independent-regulator_9789264255401-en.html. Accessed on: Sept. 28, 2026.
6. OECD. Creating a Culture of Independence: Practical Guidance against Undue Influence. Paris: OECD Publishing, 2017. DOI: 10.1787/9789264274198-en. Available at: https://www.oecd.org/en/publications/creating-a-culture-of-independence_9789264274198-en.html. Accessed on: Sept. 28, 2026.
7. OECD. Equipping Agile and Autonomous Regulators. Paris: OECD Publishing, 2022. DOI: 10.1787/7dcb34c8-en. Available at: https://www.oecd.org/en/publications/equipping-agile-and-autonomous-regulators_7dcb34c8-en.html. Accessed on: Sept. 28, 2026.
8. VAN LANGEN, Vincent; PRIETO LA NOIRE, Valeria; BAXTER, Martha; DANITZ, Eszter. The 2023 Indicators on the Governance of Sector Regulators. OECD Regulatory Policy Working Papers, No. 22. Paris: OECD Publishing, 2025. DOI: 10.1787/dc22e402-en. Available at: https://www.oecd.org/en/publications/the-2023-indicators-on-the-governance-of-sector-regulators_dc22e402-en.html. Accessed on: Sept. 28, 2026.
9. OECD. Government at a Glance 2025: Governance of sector regulators. Paris: OECD Publishing, 2025. Available at: https://www.oecd.org/en/publications/government-at-a-glance-2025_0efd0bcd-en/full-report/governance-of-sector-regulators_48990352.html. Accessed on: Sept. 28, 2026.
10. TRIBUNAL DE CONTAS DA UNIÃO (TCU). TCU avalia autonomia financeira das agências reguladoras no processo orçamentário [TCU assesses the financial autonomy of regulatory agencies in the budget process]. Feb. 11, 2026. Available at: https://portal.tcu.gov.br/imprensa/noticias/secao-das-sessoes-2026-02-11. Accessed on: Sept. 28, 2026.
11. COMITÊ DAS AGÊNCIAS REGULADORAS FEDERAIS (COARF). Bloqueio no orçamento das agências reguladoras ameaça fiscalização e projetos estratégicos [Budget freeze threatens regulatory agencies’ oversight activities and strategic projects]. June 3, 2026. Available at: https://www.gov.br/anpd/pt-br/bloqueio%20de%20orcamento%20para%20agencias%20reguladoras. Accessed on: Sept. 28, 2026.

The opinions and analyses expressed in this article are personal and do not represent the positions, decisions or institutional views of the Brazilian Electricity Regulatory Agency (ANEEL).