The First Half of the Map: Cushion or Evasion?

The banking group, the bank-facing rule and the affiliated NBFI are now visible as one relationship. The practical judgment is whether a move in lending represents a genuine group-level cushion, a change in legal labels, or both. Three questions provide the test: which entity is constrained, which affiliate is able to act, and what the evidence can actually observe. Applying those questions keeps “regulatory arbitrage” from becoming a conclusion in place of an explanation.

A conditional mechanism, not a sector-wide law

The central empirical finding is not that nonbanks universally replace banks. It is that, in the syndicated corporate-loan market studied by the IMF, bank subsidiaries and affiliated NBFIs within the same banking groups respond differently after the relevant macroprudential tightening. Bank lending falls, affiliated NBFI lending rises relative to the bank response and also rises in absolute terms. Weighted by the NBFI share of group lending in that sample, the group-level contraction is materially smaller than the decline that would be inferred from bank entities alone.

That is a group-level cushion. It tells us that the parent can preserve part of its lending capacity by changing the entity through which credit is supplied. It does not tell us that credit has been preserved everywhere, that the parent’s total balance sheet is unaffected or that the wider financial system has become safer. The estimate belongs to a defined market, a defined period and a defined shock. A reader who treats it as a property of “nonbanks” has already dropped the ownership relationship that makes the finding distinctive.

The comparison of regulatory instruments sharpens the point. When loan-supply-oriented measures constrain a bank subsidiary or a local lending margin, an affiliated NBFI may remain a usable alternative. When capital-based measures are calculated more broadly on consolidated group accounts, the parent may be unable to relieve the group-level pressure merely by changing the legal originator. In the IMF’s separate capital-based exercise, NBFI subsidiaries reduce lending relative to bank subsidiaries rather than showing the baseline expansion. The direction is consistent with the rule reaching more of the group.

This is not a contradiction between two studies of the same phenomenon. It is evidence that the phenomenon has conditions. “Tighter regulation” is not a single treatment. The binding level, the calculation base and the business activity covered by the rule determine whether a parent has room to reallocate. Jurisdictional differences matter as well; the IMF notes that consolidated capital treatment is not identical in every country.

The business model supplies a second condition. The observed response is concentrated in investment banks and broker-dealers, which are already active in the syndicated-loan market. Investment funds and asset managers do not show the same increase there.