The next big episode of distress in the financial system may not begin inside a bank. Instead, it may start with a margin call, a failed repo rollover, or an open-ended fund facing redemptions.Wherever it comes from, though, the fallout will affect banks through clearing, prime brokerage, custody, securities financing, derivatives, credit lines, and market-making.
This is the regulatory perimeter problem in modern finance. Post-financial crisis reforms made banks safer, but they did not make systemic risk disappear. In many cases, risk migrated to market-based areas where leverage, liquidity transformation, and maturity mismatches are less visible to traditional bank supervision. The resulting challenge is not only financial but also legal and institutional. Regulation remains organized largely around categories of institutions, while systemic risk increasingly travels through activities that cut across banks, funds, dealers, clearing houses and capital markets.
The distinction matters for securities regulators and banking supervisors alike. A bank may be well capitalized and still be exposed to stress created by non-bank intermediaries. A fund may not take deposits and still generate destabilizing liquidity pressure. A repo market may look like a technical funding source and still function as a site of systemic fragility. A collateral rule may appear to be private risk management and still amplify forced selling across markets.
Financial stability law must therefore follow the plumbing of risk, not only the charter of institutions.
The post-2008 regulatory framework was built around a clear diagnosis. Banks had entered the global financial crisis with excessive leverage, fragile funding structures, and opaque exposures. The policy response was necessary: higher capital requirements, liquidity rules, stress testing, resolution planning, and more robust supervision. These reforms reduced the probability that a bank failure would trigger a system-wide collapse similar to the 2008 financial crisis.
But the reform agenda also left an important boundary problem. As bank regulation became stricter, parts of credit intermediation and leverage moved into non-bank financial intermediation. This did not necessarily make the system worse; non-bank finance can diversify funding, deepen markets, and support investment. But it changed the channels through which stress is created and transmitted.
A traditional bank run is visible. Depositors withdraw funds, liquidity evaporates, and regulaors can identify the institution under pressure. A market-based run is harder to observe as it is happening. It occurs when secured funding is not rolled over, when collateral haircuts rise, when variation margin calls force investors to raise cash, or when open-ended funds sell assets to meet redemptions. The mechanics are different from a deposit run, but the effects can be similar: forced deleveraging, falling prices, declining liquidity, and contagion across markets.
This is where securities regulation and prudential regulation meet. The legal question is not simply whether a particular entity is a bank, broker-dealer, investment fund, clearing member, or asset manager. The more important systemic question is what activity the entity performs, how that activity is funded, how quickly it can generate liquidity, and which institutions absorb the shock when conditions change.
Repo and securities financing illustrate the problem. These markets are essential for liquidity and price discovery. They also create collateral chains and short-term funding dependencies. When haircuts are small and volatility is subdued, leverage can expand smoothly. When volatility rises, haircuts can increase, maturities can shorten, and funding can become harder to roll over. Leveraged investors may then be forced to sell assets into falling markets. If many investors do this simultaneously, private-risk management becomes a public stability problem.
Margin calls create a similar dynamic. Initial and variation margin requirements are essential tools for reducing counterparty risk, but they can also generate sudden cash needs. A margin call does not ask whether the investor’s long-term position is fundamentally sound; it requires liquidity now. If liquidity is unavailable, the investor sells what can be sold. This can transmit pressure from one market to another, including from advanced-economy funding markets to emerging-market bonds, equities, and currencies.
Investment funds add another layer. Open-ended funds may promise frequent liquidity to investors while holding assets that are less liquid under stress. If redemptions rise, funds may sell assets at depressed prices. Those sales can widen spreads, impair market functioning, and affect banks that provide credit, custody, derivatives, clearing, or market-making services. The risk is not identical to banking risk, but it can become bank-relevant.
The legal architecture has struggled to keep pace because responsibility is fragmented. Banking supervisors look at bank safety and soundness. Securities regulators focus on market integrity, disclosure, investor protection, and conduct. Central banks monitor liquidity and monetary transmission. Resolution authorities prepare for institutional failure. Each mandate is legitimate. The problem is that systemic risk often sits in the gaps between mandates.
The Financial Stability Oversight Council in the United States was created to address precisely this kind of cross-sector vulnerability. Yet the broader regulatory system still tends to treat bank supervision, securities regulation, and market-infrastructure oversight as separate disciplines. The next stage of reform should not be to collapse these disciplines into one. It should be to build a more explicit activity-based layer across them.
Activity-based oversight does not mean regulating every fund like a bank. That would be neither practical nor desirable. Banks are special because of deposits, payments, credit creation, and access to public backstops. But when non-bank activities generate bank-like systemic effects, supervisors should have the data, legal authority, and institutional ability to monitor them.
The core principle should be simple: Similar risk-creating activities should be visible to regulators, even when they occur in different legal forms. Leverage matters because it amplifies losses. Liquidity mismatch matters because it can force selling. Collateral rules matter because, in stressed markets, higher haircuts and collateral demands can force asset sales and transmit stress across balance sheets Cross-border portfolio rebalancing matters because it can move market pressure rapidly from one jurisdiction to another. These risks are not defined by the institutional label of the entity carrying them.
A practical regulatory agenda would begin with data. Supervisors need a clearer map of bank–non-bank interconnections: prime brokerage exposures, repo financing, derivatives positions, clearing relationships, credit lines, collateral reuse, and concentrated funding dependencies. They also need better information on margin liquidity, reducitions in loan values, fund flows, and market liquidity. This information should not be collected only after a crisis. It should be part of ordinary systemic surveillance.
Second, stress testing should cover more of the financial system. Bank stress tests are necessary, but they are not sufficient if the shock enters through market-based finance. Scenarios should examine how increases in margin calls, reduced repo rollover, and similar changes affect both non-bank intermediaries and the banks connected to them.
Third, regulators should develop clearer protocols for sharing information. A securities regulator may see fund outflows before a banking supervisor sees credit stress. A central bank may detect market liquidity deterioration before a prudential supervisor sees solvency concerns. A clearing house may observe margin pressure before public markets register disorderly selling. These signals need to be connected quickly.
Fourth, international coordination matters. Market-based finance crosses borders by design. Funding can be raised in one jurisdiction, collateral posted in another, risks hedged in a third, and assets sold in emerging markets when liquidity is needed. A purely domestic view of systemic risk is increasingly incomplete.
Emerging markets provide a useful stress test for this framework. They are often exposed to global dollar conditions, commodity prices, foreign portfolio flows, and external risk appetite. Their domestic banks may be sound, but their markets can still experience abrupt outflows when global investors deleverage. In these jurisdictions, the regulatory perimeter problem is not theoretical. It is part of the recurring experience of financial openness.
This does not imply that every market correction should be prevented. Markets must be allowed to reprice risk, and losses should not automatically become public liabilities. The objective is narrower and more important: to prevent ordinary repricing from becoming forced selling, liquidity spirals, and systemic stress.
The post-crisis era taught regulators to take bank resilience seriously. The next era will require them to take market plumbing just as seriously. Financial regulation cannot stop at the boundary of the bank charter when the sources of systemic risk run through collateral, margins, repo, derivatives, funds, and cross-border portfolios.
The regulatory perimeter should be defined not only by what institutions are called, but also by what activities do under stress.
Gustavo Pessoa is a professor of economics at Brazil’s Fundação Getulio Vargas, Escola de Administração de Empresas de São Paulo (FGV-EAESP) and holds a PhD in finance.