Basel’s Endgame Stalls as Regulators Wrestle with Bank Size

Banking regulators have been circling the same question since the 2008 financial crisis: how do you contain the risks posed by very large banks without throttling the system they underpin?

At the heart of this dilemma sit global systemically important banks, or G-SIBs. The choice is not new. Do you cut the largest down to size, or raise the rest to match them? Bismarck chose the latter, building systems that elevated the base. Marx argued for the opposite. Banking regulation now finds itself replaying that argument, with capital ratios instead of class.

Basel’s post-crisis reforms attempted to straddle both paths. First, by raising capital requirements, disproportionately affecting large banks, especially those using internal risk models. These banks must now hold at least 75% of the capital that would be required under standardised rules. Second, by bringing previously underplayed risks, including liquidity and interest rate risk in the banking book, or IRRBB, into the regulatory fold for all banks.

Yet the execution reveals a deeper discomfort. Supervisors have long worried that internal models allow large banks to understate risks and, by extension, capital needs. The logical response would have been to strengthen supervisory capacity: invest in skills, tighten oversight, improve model validation. Instead, regulators have opted to curb the use of such models altogether.

That has consequences. If sophistication no longer offers regulatory benefit, the incentive to invest in better risk modelling weakens.

Global Divide
Implementation has been anything but uniform.

Europe and Japan moved quickly on liquidity and IRRBB rules. But when it came to tightening capital via limits on internal models, the core of the Basel III “Endgame”, progress slowed. The hesitation was strategic. Moving ahead without clarity on US implementation risked putting their banks at a competitive disadvantage.

The United States, meanwhile, had already imposed stringent stress test-based capital regimes on its largest banks. Add to this the impact of provisioning standards such as CECL, which typically demand higher loan-loss recognition than IFRS-9, and US banks were arguably operating under a tougher regime even before the Endgame rules fully landed.

This asymmetry has fuelled resistance. Large American banks have pushed back, sometimes quietly, sometimes more visibly, arguing that a like-for-like increase in capital requirements could tilt the competitive balance against them. So the Endgame has been stuck in a loop of revisions and retreats.

Then came the failures of a few mid-sized US banks in early 2023. That episode did not accelerate reform; it complicated it. Regulators were forced to tighten parts of the system while simultaneously stepping back from anything that looked like overreach. Stability became the priority again, but so did optics.

Now, with a more accommodative regulatory tone emerging in Washington, there are signals that revised proposals may address some of the industry’s concerns. But clarity on implementation remains elusive, and timelines, even more so.

Lost in this regulatory back-and-forth is a structural reality: large banks dominate the system. In the US and Europe, these banks hold roughly 75% of system assets. Make them more expensive to run, and credit will not stay untouched.

Lending tightens, pricing shifts, and the cost quietly travels down the chain to businesses and households. In an already fragile global growth environment, that trade-off becomes harder to ignore.

Meanwhile, much of the rest of the world has moved on. Mid-sized banks in advanced economies, and entire banking systems in countries such as India, have largely adopted Basel III through standardised approaches. The shift, in effect, is towards uniformity, away from model-driven differentiation.

Which leaves Basel’s Endgame in an awkward place. In trying to decide whether to restrain the largest or elevate the rest, regulators have edged into doing both, without fully committing to either.

In practice, the system is already tilting one way. Closer, perhaps, to Marx than to Bismarck.