The Credit Shift Hidden Inside a Banking Group
The subject of this series is the modern banking group: a parent organization that controls regulated bank subsidiaries alongside affiliated nonbank financial institutions, or NBFIs. The opening question is simple but consequential: when regulation tightens around banks, does corporate credit actually shrink, or can the group move the lending business to another legal entity? Understanding that distinction gives readers a practical way to judge both the effectiveness of regulation and the risks that may remain after a loan changes hands.
A bank and a nonbank are not opposites in the way the labels can suggest. A commercial bank takes deposits and operates under extensive prudential rules governing capital, liquidity and safety. An NBFI may be an investment bank, broker-dealer, investment fund, asset manager, insurer or pension fund. It generally does not take deposits in the same way, and in many jurisdictions it does not face the full set of bank-like requirements. Some NBFIs are independent firms. Others sit inside the same financial group as a bank. That ownership link is the hinge of the story: it can give a nonbank access to relationships, information and funding that an independent lender does not have.
The banking group is the unit that can move the loan
It is tempting to describe tighter regulation as a force that pushes lending out of the banking system. That description is incomplete when the bank belongs to a group with several lending entities. The parent can face a constraint at one subsidiary while retaining another subsidiary that is able to originate, arrange or hold the same kind of corporate credit. The borrower may still receive financing, but the institution recorded as the lender has changed.
This is why the relevant subject is not “nonbanks” in the abstract. The relevant subject is the parent’s allocation decision and the relationships that make that decision possible. A banking group may have a regulated bank arm, an affiliated investment bank or broker-dealer, and other financial businesses under common control. If a loan-supply rule binds the bank arm more directly than the nonbank arm, the parent has a substitution margin. It can reduce origination through the constrained entity and increase origination through the affiliate, at least where the affiliate has the expertise, balance-sheet capacity and market access to take the loan.
The distinction between affiliated and independent NBFIs makes the mechanism more concrete. An independent nonbank must obtain its own funding and build its own borrower and distribution relationships. An affiliate may draw on the broader group’s internal capital market or wholesale connections, although the available evidence does not directly observe every such transfer. That difference helps explain why the two types of nonbank can respond differently to the same regulatory shock. In the IMF study’s expanded sample, independent NBFIs reduce lending relative to banks after a tightening, while bank-owned NBFIs increase lending relative to them. The affiliated entities are fewer but account for a much larger share of the syndicated lending observed in the sample.
The term “macroprudential tightening” refers here to measures intended to restrain the credit cycle or strengthen the resilience of the financial system. Such a measure may limit loan supply, require more provisioning, alter reserve requirements or impose other prudential constraints. The important point is that the legal reach of the measure matters. A rule aimed at the lending capacity of an insured bank subsidiary does not necessarily have the same immediate effect on an affiliated broker-dealer or investment bank. That does not make the affiliate unregulated; it means that the same constraint may not bind it in the same way.
The Basel Committee’s report on bank–NBFI interconnections provides the broader reason this organizational distinction matters. Banks and NBFIs are connected through lending, deposits and repos, clearing, market-making, underwriting, guarantees, credit protection, asset management and ownership. A parent can therefore move the formal originator of a corporate loan without severing the economic links around that loan. The group may still provide funding, distribute risk, clear trades or support the affiliate. Credit can move across the regulatory perimeter while dependence remains inside the financial network.
The diagram is deliberately a map rather than a claim that every group has the same structure. It shows the minimum relationships needed to ask the right question. The bank-facing regulator constrains one node directly. The parent connects the bank and affiliated NBFI. Both may serve the corporate borrower. An independent NBFI can also lend to that borrower, but it does not share the parent’s ownership and internal funding links.
A lending decline may conceal a change of vehicle
The IMF working paper, Banking on Nonbanks, turns this conceptual possibility into an empirical question. It studies new syndicated loans to nonfinancial corporations, a market in which banks and nonbanks can both participate. Its central comparison is between bank and NBFI subsidiaries belonging to the same banking group, while also accounting for changes in borrower demand over time. That design matters because a group-level decline in lending could otherwise be confused with a borrower deciding to seek less credit or with an unrelated lender becoming more competitive.
The headline result is a measured shift rather than a disappearance of credit. For a one-standard-deviation tightening in the study’s baseline macroprudential shock, bank subsidiaries reduce lending by 1.0 percent. Affiliated NBFIs increase lending by 2.0 percent relative to bank subsidiaries and by 1.0 percent in absolute terms. Using the average share of NBFI subsidiaries in the sample, the authors calculate that the banking group offsets roughly 57 percent of the contraction that would have occurred through bank entities alone. That is a substantial cushion, but it is a result for the defined syndicated-loan market and the particular shock measure—not a claim that regulation has no effect on corporate finance.
The result also should not be read as proof that every parent explicitly transfers cash from its bank arm to its nonbank arm immediately after a rule changes. The study does not observe direct internal capital-market flows. Its estimates are consistent with a parent reallocating funding and lending capacity toward an affiliate, potentially on favorable terms, but that internal mechanism remains partly inferred. Other motives may coexist. A group may use nonbank affiliates to diversify funding, preserve relationships or manage liquidity risk, not only to reduce the force of a regulatory constraint.
The ownership relationship also changes the competitive landscape. In the IMF sample, independent NBFIs do not respond like bank-owned affiliates. Where banking groups already have a large market presence, the expansion of their affiliates is associated with weaker lending by other NBFIs, especially in domestic markets. The interpretation is not that independent lenders are incapable of making loans. It is that they may lack the parent’s funding channels and relationship advantages at the moment when credit is scarce. A group’s regulatory adaptation can therefore protect its own market share while making it harder for unaffiliated lenders to replace the bank.
This is the first practical judgment the subject supports: a lower volume of loans recorded by bank subsidiaries does not by itself demonstrate that credit has contracted by the same amount. To determine what happened, a reader must ask at least three questions. Which entity was directly constrained? Which affiliated or independent entity could take over the lending? And did the group’s total credit supply fall, or did the lender of record change? Without those questions, “lending moved to nonbanks” is only a label.
What the shift does—and does not—tell us
The reallocation creates a genuine trade-off. For a corporate borrower, continued access to a loan can be valuable when a bank-facing rule would otherwise reduce supply. At the group level, using an affiliate may preserve revenue, borrower relationships and some capital-allocation efficiency. From a public-policy perspective, however, a constraint that can be bypassed through an affiliated entity may have a smaller effect on credit growth than its designers intended.
At the same time, the lending shift is not automatically evidence of reckless risk-taking. The IMF study finds little evidence that the additional NBFI lending after tightening is disproportionately directed toward the risky borrower categories it tests. NBFIs already lend, on average, to riskier borrowers than banks, but the incremental response is not concentrated further in high-default-probability, highly leveraged or zombie firms in the study’s measures. That is a useful qualification, not a safety certificate. Those tests concern borrower characteristics in syndicated lending. They do not settle what happens through derivatives, collateral, short-term funding, private credit or common asset holdings.
The BCBS report shows why the wider system cannot be assessed by looking only at the borrower and the originating subsidiary. Banks may finance NBFIs, provide them with liquidity, clear their trades and make markets for them. They may also buy protection from NBFIs or own them. Each connection creates a different way for stress to return to the bank. A fund that needs cash to meet redemptions may withdraw a deposit or fail to meet a margin call. A bank that responds by raising collateral requirements may protect its own balance sheet while forcing asset sales that depress prices for many institutions. An insurer or fund that previously absorbed credit risk may stop doing so, leaving assets on the bank’s balance sheet.
The result is a change in the location and shape of risk, not a clean removal of risk. A group that is successful at preserving lending through an affiliate may become more interconnected with the nonbank sector. If that affiliate later faces a shock, the parent may confront funding pressure, reputational or conduct costs, counterparty losses, or a decision about whether to provide support. The Basel Committee treats these as plausible stress channels built from observed cases, not as a forecast that every affiliate will fail.
The evidence also has clear boundaries. The IMF paper examines a defined syndicated-loan market and does not directly measure the financial-stability consequences of the reallocation. Its ownership map is focused on entities active in that market, so it does not capture every affiliate a banking group may control. The BCBS report likewise emphasizes that supervisory data often miss cross-border, off-balance-sheet and private-market connections. The two sources therefore support a disciplined conclusion: tighter bank regulation can move some corporate lending within a banking group, while the consequences of that move depend on which rule applies, which affiliate is used and which economic links remain.
That conclusion motivates the next step. It is not enough to say that regulation became tighter. We need to follow the constraint through the parent’s allocation decision and ask what makes the affiliate route available, why some parent groups use it more than others, and where the route stops working. Once that mechanism is visible, the later questions about geography, competition and financial stability become questions about the same banking groups—not a separate story about an anonymous nonbank sector.