Basel to banks: we will stop you dressing your windows

If you are a bank then there are various dull but worthy regulatory boxes that you must check. One of those is the Basel III leverage ratio, which is your tier-1 capital divided by your on- and off-balance sheet exposures, and this calculation must spit out a ratio of at least 3%.

If you are a global systemically important bank (G-Sib), then you must add on an additional buffer. Outside the US, this is called the G-Sib leverage ratio surcharge and it was brought in at the start of 2023, having been agreed back in 2017.

The new surcharge is set at half of a G-Sib’s risk-weighted surcharge, which depends on which G-Sib 'bucket' the institution falls into, as determined by the Financial Stability Board’s annual assessments. Currently there are 29 G-Sibs, and the next list will be published in November.

In the US, meanwhile, all banks are subject to a basic leverage ratio requirement of 4% of on-balance sheet exposures. The US implementation of the Basel III leverage ratio, in 2013, resulted in the supplementary leverage ratio (SLR), which imposes the internationally agreed 3% ratio on both on- and off-balance sheet exposures.

The US had already been implementing its own additional leverage ratio buffer, which was set at a fixed additional 2% for the G-Sibs and was called the enhanced SLR (eSLR). There is debate rumbling now about whether the US should instead adopt something more like the Basel leverage ratio surcharge, particularly as it might ease the regular complaints about how capital regulation is harming banks’ ability to properly intermediate in the US Treasury market.

Leverage ratios are risk-insensitive by design, and that is what many people in banking hate about them, arguing that they are an excessively blunt instrument. From a regulatory perspective, of course, a few blunt instruments are not necessarily a bad thing.

But this is banking and so everything is gamed, even the blunt instrument of the leverage ratio. If, purely by coincidence, your exposures were to suddenly drop at around the time you would be reporting your leverage ratio calculations, usually at the end of a quarter and particularly at the end of a financial year, it is possible that the ratio you report might be a little higher than normal.

Obviously, it seems hard to believe that banks would deliberately do this, but apparently, they do – and the good folk of the Basel Committee on Banking Supervision (BCBS) have decided that enough is enough.

Or, more accurately, that enough is often not enough.

So, when they held a meeting in Madrid in late February to chat about various initiatives, they not only noted that something called commercial real estate was a headwind faced by the banking sector but they also resolved to tackle the shocking fact that some banks routinely manipulate their trading activity and balance-sheet deployment to allow them to report better leverage ratios.

This being a regulatory initiative, it has of course taken some considerable time to get to this point.

The committee noted in its virtual meetings in early December 2023 that “window-dressing by banks is unacceptable”. Apparently, a range of empirical studies had highlighted such behaviour “by some banks”. Imagine that!

The committee has at least been consistent over the years. In October 2018, it issued a statement specifically on the subject of leverage ratio window-dressing.

It was, the committee said, “unacceptable”.

Back then, it was surges in volatility in money markets and derivatives markets that had “alerted” the committee to “potential regulatory arbitrage” by banks.

The committee said that banks should “desist” from doing transactions with the sole purpose of reporting and disclosing higher leverage ratios.

Supervisors, the committee added, “might also consider” requiring more frequent reporting by banks and look at the whole reporting period, not just the reporting date.

You might have thought that this kind of tough-talking would have put an end to the issue once and for all. After all, the committee in 2018 had even tacked its catch-all threat onto the end of its statement, suggesting that it might “consider additional measures”, including Pillar 1.

To give the BCBS its due, that is actually meaningful, since no bank wants Basel to be doing anything at all with Pillar 1 if it can possibly help it.

Let no one accuse the committee of failing to act decisively. In its December 2023 meeting, it agreed “to consult in 2024 on potential policy options aimed at reducing window-dressing behaviour”, including collecting “higher frequency data items”.

What has given the committee the confidence to take practical steps at last is a new set of findings in a working paper that the Bank for International Settlements released on March 7 and that was written by Matthew Naylor of the Bank of England, Renzo Corrias of the BIS and Peter Welz of the European Central Bank.

In it, the authors find what they term as causal evidence that the G-Sib framework accounts for about half the year-end contractions in reported notional over-the-counter derivatives, a total of about €30 trillion and about 5% of total global annual activity.

The killer is what the authors call their “difference-in-differences” empirical strategy. This looks at the window-dressing behaviour that big banks displayed both before and after being designated as G-Sibs, and in particular at how banks’ behaviour changed according to how close they were to a particular G-Sib bucket.

By analysing a new, confidential dataset of 70 big banks, which the BCBS had collected for a study in 2023, the researchers discovered that year-end reported figures for especially OTC derivatives exposure and repo lending were lower than quarterly numbers, and that the behaviour was more pronounced for those firms that stood to gain most.

“We find that banks with greater G-Sib-specific window-dressing incentives began to window-dress notional OTC derivatives significantly more than peers after the implementation of the framework, having previously exhibited broadly similar behaviour,” the authors wrote.

With admirable self-scrutiny, one conclusion that the committee has drawn is that the way that G-Sibs are assessed may need to change, and it has published a consultation document asking the industry for its view on exactly that.

But banks have also been put on notice: Basel is coming for you. And this time it will have higher frequency data items.