The leaked tapes of conversations between two senior Anglo-Irish Bank officials in September 2008 highlights the problem that bankers can’t be trusted. Discussing the size of bailout they should request from Irish taxpayers, the Anglo-Irish bankers were recorded saying that the €7 billion figure was “picked out of my arse” (the actual amount needed would be five times that), while likelihood of repayment was “never”.
How did things look from the other side of the table? In September 2008, the Irish Finance Ministry was dependent on information compiled by the country’s financial regulator. And the framework used by that regulator came from Switzerland in the form of Basel II rules.
According to the standardised framework adopted by Anglo in 2007, the commercial property loans that dominated the bank’s balance sheet had a risk-weighting. That weighting fed into calculations approved by the regulator, which deemed the bank sufficiently capitalised.
That’s why regulators initially thought Anglo’s problem was about (temporary) liquidity rather than (permanent and fatal) insolvency – a perception that the tapes suggest the bankers exploited to the hilt.
In October 2009, while researching The Devil’s Derivatives, I visited the Irish finance ministry in Dublin. The bailout of Anglo was still relatively fresh in the civil servants’ minds. I asked them about the belief system that convinced them that Anglo’s Basel ratios were objectively accurate. In 2008, assessments of Irish banks were in the hands of country’s now-defunct Financial Regulator, an executive agency set up in 2003 by the finance ministry.
When Anglo came to the ministry asking for support, the civil servants wrote their cheques based on the regulator’s implementation of Basel II, which had been written into Irish banking law via the EU’s Capital Requirements Directive. It’s worth quoting their recollections at some length.
“There was the sense that here was this state-of-the-art approach to assessing risk and ensuring efficient use of capital in financial institutions”, one of the civil servants said. “So talking to people in the financial regulator, and I think this is relevant to any small country, the resources required to work with our institutions putting this system into place took an enormous amount of time”.
As the ministry officials put it, Basel II and its hundreds of pages of formulas led the regulator to stop seeing the forest for the trees. “I suppose they were focused in on the operational implementation, and got into the details of the project rather than standing back”.
That was an important point the official was making about standing back, and he expanded on it. “Obviously they talked about the Pillar II dimension, which is supposed to be the regulator stepping back and making an informed judgment on what is coming out of the models.”
What does ‘informed judgment’ mean? Assessing the likelihood of Anglo or its loan portfolio defaulting might start with what is called unconditional probability – perhaps using the default rate from a global historical sample of banks or real estate loans. Basel provided the objective framework for doing that. However, regulatory judgment should mean going beyond the unconditional probability to something conditioned on evidence gathered by inspecting a particular bank such as Anglo.
To make the judgment meaningful, the evidence gathering ought to consist of a test. In medicine, a patient is tested for a fatal disease such as cancer. The test isn’t infallible – there are ‘false positives’ when healthy patients fail the test and ‘false negatives’ when sick patients pass the test. So clinicians use something called Bayes’ theorem to combine background information about patients with test evidence.
Shouldn’t this be done by bank regulators too? Isn’t there a test that identifies sick banks (even if healthy banks occasionally fail the test too)? The problem with Basel rules was that institutions were responsible for testing themselves, and regulators were intimidated from questioning the reliability of the test. Expressed in Bayesian language, that crucial question would be: what is the probability of a bank being insolvent given that it has passed the test?
As the Irish finance ministry civil servants explained: “Maybe there’s another aspect to the CRD, talking about the principles-led approach to financial regulation. If they have their model and their responsibilities in terms of risk management using the information generated through that approach, then you’re devolving your responsibility to the institution to act on the basis of that analysis”.
This devolution – or abdication one might call it – put Ireland on the road to disaster. “One describes the regulatory approach as primarily being the board and senior management of the institutions having lead responsibility. We set down the principles and they have to respect those principles in spirit rather than having rules and box-ticking. Those distinctions can be a bit futile I think”.
Putting it another way, the trustworthiness of Basel models was hardwired into the Irish regulatory system, turning supervisors into the puppets of the bankers they regulated.
This is how the civil servant described it. “I think in terms of the behavioural factors, if you’re told that this regulatory approach is state-of-the-art, maybe you’re more inclined to believe what the institutions are telling you about their risks and the exposures they’re managing. It is human nature when someone presents you with a model to think that it is a good representation of reality.”
Perhaps this foreknowledge that their supervisors were cowed by Basel led the Anglo bankers to speak of their regulators with the contempt that emerges from the leaked tapes. Today, are bankers tested sufficiently so that they have learned to fear their regulators? Somehow, I doubt it.
Comments are closed.
How Japan is changing, in two charts
Playing MSCI’s quarterly rebalances
How the S&P 500 moved into hyperspace
iBoxx top movers
I agree with a lot of this, but the “somehow I doubt it” bit at the end is a bit too cynical. At the end of the day, fooling your regulators by putting unrealistic inputs into your Basel model is a bit like cheating at patience – it’s yourself you’re doing the damage to, unless you have very very good reason to believe that Basel was overstating economic capital. I don’t think that there actually are loads of European banks which are just as bad as Anglo-Irish; the EBA and Basel consistency exercises have repeatedly surprised me with how little variation and model-gaming they have found, even with a study design that’s set up to deliver all the variation that exists and a little bit more.
I think it’s more likely that the less exciting truth is that the majority of banks have been reasonably law-abiding and sensible, and that nearly all the miscreants were found out during the very large series of shocks that hit them in 2008-9. There’s a kind of adrenaline-junkie element to a lot of the press coverage that desperately wants to say that they’re all as bad as each other, and that all the banks which certainly look like they’ve doubled their capital and halved their risks since 08 are actually just like Anglos and about to be found out! any moment now!
But I do agree that there’s vastly too little use of Pillar 2 in general and much too much willingness to approve models and datasets, and the “blinded with science” factor is a really important aspect of this. It really wasn’t hard to work out from the accounts that Anglo was overleveraged and pumped up with CRE loans – I knew it at the time and said so repeatedly in writing and so did a lot of other analysts. The capture of the Financial Regulator was a really bad example of the things that were going wrong in Irish politics at the time.
Actually that just triggered a memory for me from 2006 when I was initiating coverage on the two big Irish banks. I was having a meeting with one of them (wish I could remember which, not that it matters), and he started talking about a CBoI data request that they had just received. I remember his words vividly (I’ll skip the phonetic stage-Irish!)
“They asked us for all the files on [a credit scoring issue we’d been talking about]. We said sure, but first you have to tell us which warehouse we’re going to deliver the documents to! That’s truckloads of paper – they have no idea of how the business of banking works!”
I remember thinking at the time – you have that data in electronic form, of course you do, and I’d guess that you deliver it regularly to the ratings agencies and your securitisation bankers. They were just literally trying to intimidate their regulators and apparently getting away with it.