On the first of January 2025, the Capital Requirements Regulation III entered into force across the European Union, bringing with it the phased implementation of the Basel IV output floor: a mechanism which requires banks using internal risk models to hold capital calculated as if their risk-weighted assets were at least fifty per cent of what the standardised approach would produce, rising in annual increments to seventy-two and a half per cent by the first of January 2030 [1]. The stated purpose of the output floor is to constrain the discretion that internal-model banks have historically exercised over their own capital requirements, correcting for a decade of model-driven understatement that left the system visibly undercapitalised at the moment of the 2008 crisis. The purpose, in other words, is to make the regulated banking sector safer. The question that the available evidence now compels us to ask is whether it has instead made the financial system as a whole more dangerous.

The mechanism by which capital requirements produce shadow banking is well understood in the theoretical literature, and it was articulated with considerable force by critics of the Basel framework throughout the 2010s. When the cost of conducting an activity within the regulated perimeter rises sufficiently above the cost of conducting that same activity outside it, rational economic actors will migrate. This is not a failure of markets; it is markets functioning precisely as markets function. The question for regulators has always been whether they can raise the regulated sector's cost sufficiently to achieve their prudential objectives without simultaneously making the unregulated sector so attractive as to redirect the underlying risk rather than reduce it. The post-2008 Basel framework assumed the answer was yes. The data accumulated since 2015 suggests the assumption was optimistic.

The Parallel That Basel Chose Not to Learn

There is a precedent for this failure that is sufficiently exact to be instructive. In 1933, as part of the Banking Act that also established the Federal Deposit Insurance Corporation, Congress enacted Regulation Q, which prohibited commercial banks from paying interest on demand deposits and empowered the Federal Reserve to set maximum rates on savings and time deposits. The purpose was to protect bank solvency by preventing destructive competition for deposits. For two decades, the regulation operated within its intended parameters. In the late 1960s and early 1970s, as market interest rates rose above the ceilings imposed by Regulation Q, depositors discovered a compelling alternative: money market mutual funds, which were unregulated, offered market-rate returns, and provided liquidity characteristics sufficiently similar to bank deposits to serve as functional substitutes [2]. By the early 1980s, money market mutual funds held assets exceeding four hundred billion dollars, representing a transfer of credit intermediation of precisely the kind Regulation Q had been designed to prevent. The regulation had not reduced risk; it had relocated it.

The parallel to the present situation is structurally close. Basel IV's output floor raises the cost of credit intermediation within the regulated banking sector, particularly for the low-risk portfolios (prime residential mortgages, high-quality unrated corporates, certain trade finance structures) that benefit most from the internal-model approach and are most penalised by the risk-insensitive standardised approach that the output floor increasingly forces banks to approximate [3]. The institutions best positioned to absorb these credits at lower cost are those outside the regulatory perimeter: private credit funds, business development companies, direct lending platforms, real estate debt funds. These are, in aggregate, the contemporary equivalent of the money market fund industry of 1975.

The Scale of the Migration

The Financial Stability Board's December 2025 Global Monitoring Report on Non-Bank Financial Intermediation provides the most comprehensive accounting of where credit has gone. The non-bank financial intermediation sector grew 9.4 per cent in 2024, double the pace of the banking sector's 4.7 per cent growth, and now holds $256.8 trillion in global financial assets, representing fifty-one per cent of the total [4]. The narrow measure of NBFI, comprising entities specifically assessed as conducting credit intermediation activities that may pose bank-like financial stability risks, grew twelve per cent to reach $76.3 trillion. Non-bank sources now account for nearly twenty-three per cent of total credit to non-financial corporations globally. Within that figure, the private credit market alone reached approximately two trillion dollars in assets under management in 2024, up from an estimated five hundred billion dollars a decade earlier [5].

Data

The Migration of Credit: Private Markets Absorb What Basel IV Pushes Out

Global private credit assets under management, 2015–2024, USD billions

Source: Preqin, IMF Global Financial Stability Report, FSB Global Monitoring Report on NBFI 2025. Figures are approximate; private credit lacks a standard global definition. Compiled by The Bankers’ Magazine.

The FSB notes, with characteristic understatement, that there are "severe limitations in the availability of data for private credit in statistical and regulatory reports" [4], in part because there is no standard definition of private credit activities across jurisdictions. This observation is, in itself, the sharpest possible indictment of the regulatory architecture. The non-bank sector now intermediates a fifth of all corporate credit in the global economy, and the principal regulatory body responsible for financial stability monitoring acknowledges that it cannot measure that sector with precision. The regulated banking system, by contrast, reports its exposures on a standardised basis to national supervisors who report to the Basel Committee, the FSB, and the BIS. The comparison illuminates the core problem: Basel IV has shifted risk from the most transparent part of the financial system to the least transparent.

Interconnection as the Hidden Channel

The defenders of the Basel IV framework, including the Basel Committee itself and the European Banking Authority, advance two arguments in response to this critique. The first is that the migration of credit to non-bank entities is not inherently dangerous: direct lending is less procyclical than bank credit because it does not rely on short-term deposit funding and is therefore less subject to bank-run dynamics. Private credit funds hold their positions to maturity; they cannot be forced to sell by a deposit withdrawal. This argument has merit in isolation, but it neglects the interconnection between the bank and non-bank sectors that supervisors have consistently documented. The European Systemic Risk Board's Non-Bank Financial Intermediation Risk Monitor notes that the top ten banks supervised by the European Central Bank hold approximately seventy per cent of claims on NBFI entities and approximately sixty per cent of NBFI funding [6]. The non-bank sector is not independent of the banking sector; it is substantially funded by it. A stress event in the non-bank sector that impairs the collateral held by banks against their NBFI exposures does not remain in the non-bank sector.

The second argument is that the leverage and liquidity transformation in the private credit sector remain below the levels that characterised bank-funded structured credit before 2008. This is, as far as it can be verified, true. But it is also precisely what was said of money market mutual funds in 1975, and of the early-period mortgage-backed securities market, and of the asset-backed commercial paper conduits that proved so consequential in 2007. Each of these instruments appeared manageable at early scale and systemic at late scale. The FSB's observation that "a continuation of the rapid growth observed in recent years could lead to the sector becoming systemically relevant" applies directly to private credit, which has quadrupled in a decade [4].

What the Output Floor Does to European Banks

It is worth being precise about the mechanism by which CRR3 produces the migration. The output floor's effect is asymmetric across portfolio types. Banks with large exposures to high-quality, low-risk assets, the portfolios that internal models most consistently rate as requiring less capital than the standardised approach assumes, face the largest increases in required capital. The EBA's impact assessment identified Nordic banks, Dutch institutions, and certain German lenders as disproportionately affected, precisely because their portfolios are concentrated in prime residential mortgages and investment-grade corporate credit [3]. These are also the portfolios most attractive to private credit funds, which can price risk to the actual credit quality of the borrower rather than to a regulatory floor calibrated to the population of banks that used internal models to systematically understate risk in the period before 2008. The floor was designed to correct for the most aggressive model users; it penalises the most conservative ones in equal measure.

The transitional arrangements soften the immediate impact. CRR3 limits the year-on-year increase in risk-weighted assets attributable to the output floor to twenty-five per cent, and the floor itself reaches its final level of seventy-two and a half per cent only in 2030. European banks are managing a migration that will accelerate gradually rather than one that has already completed. But the incentive structure is visible, and the private credit market's extraordinary growth since 2021, coinciding with the finalisation of the Basel III reforms and the commencement of CRR3 implementation, is not a coincidence [5].

Three Variables to Watch

The question for 2026 and beyond is not whether credit has migrated to the non-bank sector, because it manifestly has, but whether the migration poses systemic risks that the existing regulatory perimeter cannot contain. Three variables are determinative. The first is the leverage embedded in private credit structures: as the market has grown, some funds have introduced credit facilities against their portfolios that introduce leverage in a form not reflected in AUM figures. The second is the bank-NBFI interconnection: if banks are both the funders and the counterparties of the private credit funds that have absorbed the credits they could no longer hold economically, the prudential separation between the sectors is largely illusory. The third is data: the FSB cannot monitor what it cannot measure, and the current definition of private credit for regulatory purposes is so loose as to be functionally meaningless. None of these variables is self-correcting. Each requires a supervisory response that the current framework does not provide.

There is a passage in the Federal Reserve History's account of Regulation Q that reads, in retrospect, as an epitaph for the entire episode: the interest rate controls failed to suppress underlying economic pressures; they instead channelled those pressures into shadow markets that eroded the intended stability by amplifying credit volatility during rate spikes [2]. The Basel Committee, to its credit, has spent fifteen years constructing a more sophisticated framework than Regulation Q. The outcome, measured in the migration of credit intermediation to entities whose leverage, liquidity profiles, and interconnections the supervisory system cannot fully see, is not entirely dissimilar.