Observatory · Systemic risk
In a sound financial system risk is spread. This is the principle that justifies portfolio diversification, loan syndication and securitisation itself: distributing an exposure among many parties each able to bear part of it reduces the likelihood that the failure of one propagates. The question worth asking is whether the system we are actually building does this, or the opposite.
Every tightening of bank capital requirements produces a predictable and repeated effect: part of the credit that becomes too costly to hold on a supervised balance sheet migrates towards entities not subject to those requirements. Credit funds, securitisation vehicles, direct lending platforms, intermediaries described in Anglo-Saxon debate as shadow banks, not because they operate unlawfully, but because they perform credit intermediation functions outside the perimeter of prudential banking supervision.
The critical point is not migration itself, which can also be efficient and has opened funding channels otherwise unavailable to companies. The point is that risk, on leaving the bank balance sheet, does not dissolve: it is repackaged, and often repackaged into instruments that end up held, directly or indirectly, by the same large financial institutions that transferred it. The chain lengthens, transparency falls, and assessing overall exposure becomes an exercise no single operator can perform in full.
The result is a situation in which the narrative and the facts diverge. The narrative is one of diversification: many vehicles, many investors, distributed risk. The facts describe instead a growing concentration on a few large, highly leveraged and tightly interconnected platforms, funded by the same narrow group of global institutions and holding exposures correlated with one another. In such a configuration the failure of one intermediary does not remain an isolated event, because transmission travels along shared funding lines before it travels through valuation contagion.
One point requires precision, because the argument must not become an indictment of structured finance as a technique. Securitisation, when it genuinely transfers risk to those able to price and bear it, does exactly what it promises. It becomes a problem when the transfer is merely formal, when risk returns by indirect routes to those who originated it, or when the multiplication of layers makes it impossible for anyone to know where it has settled. The difference between the two situations lies not in the instrument but in the transparency of the chain and the alignment of incentives along it.
The question regulators, banks and investors should ask is not whether systemic risk has been reduced compared with 2008, because in many respects it has. It is whether part of that reduction consists in moving risk below the waterline, into an area where measurement is harder and supervision thinner. As long as markets rise, the distinction between the two hypotheses remains academic; it ceases to be so at precisely the moment when one needs to know.
On this it is worth recalling the lesson of monetary circuit theory, and in particular the work of Augusto Graziani: credit is not the transfer of pre-existing savings but the creation of purchasing power, and the soundness of the system depends on the circuit's ability to close, that is on the financing generating the income needed to repay it. When financial engineering develops without that circuit closing in the real economy, what grows is not the capacity to fund production: it is merely the number of parties among whom the same obligation is passed around.
A first version of this contribution was published by the author on LinkedIn on 28 February 2026.