The Bank’s ‘Stress’ Tests

My report on the Bank of England’s latest (November 2018) stress tests was published by the Adam Smith Institute on August 3rd.

The purpose of the stress tests is, in essence, to persuade us that the banking system is in good shape on the basis of a make-believe exercise which purports to show what might happen in the event of a supposed severe stress scenario as modelled by a central bank with a dodgy model and a vested interest in showing that the banking system is in great shape thanks to its own wise policies.

We are expected to believe that the central bank has managed to rebuild the banking system despite enormous pressure placed on it by the institutions it regulates, whose principal objective is to run down their capital ratios (or equivalently, maximise their leverage) in order to boost their returns on equity and resulting short-term profits, and never mind the systemic risks and associated costs imposed on everyone else or the damage their high leverage did in the Global Financial Crisis.

These latest Bank of England’s stress tests were published in the Bank’s November 2018 Financial Stability Report, the core message of which was that the UK banking system was doing just great, but that a No-Deal Brexit would be a disaster. Wrong on both counts.

I will focus here on the first issue, the state of the banking system.

In essence, the Bank paints a reassuring picture of bank resilience. The message is that the UK banking system is now so strong that it could sail through another crisis that is more severe than the last one and still be in good shape. How do we know this? Because the stress tests tell us, claims the Bank. However, the truth is that the Bank’s stress tests are useless at detecting bank fragility. And how do we know that? Because this fragility is apparent from banks’ leverage. So do you believe the results of the stress tests or do you believe the leverage numbers?

The centrepiece of the Bank of England’s oft-repeated narrative about the banking system is that bank capital ratios are high and still rising. The measure the Bank relies on is the ‘CET1 ratio’, the ratio of Common Equity Tier 1 (CET1) capital to Risk Weighted Assets (RWAs). As of 2018Q3, the average CET1 ratio across the big 5 UK banks was 14.7%. This number looks impressive until one realises that the denominator is meaningless, because it is highly gameable and the banks are expert at this game. 1 We should throw this ratio into the bin.

A better measure is bank leverage. Leverage is total assets divided by (CET1) equity.

What is bank leverage?

As of 2018Q3, the average leverage across the big 5 was 20.7. This number is book value, however, and a more accurate measure is market-value leverage, which is 34.4.

Thus, in market value terms, the big banks have issued over £34 of debt for every £1 in equity. That is an enormous level of leverage.

But weren’t we told that excessive leverage was a key driver of the severity of the financial crisis? Er, yes.

Bank leverage is also increasing. A year earlier, it was under 28.

In market-value terms, bank leverage is also higher than it was before the crisis.

The banks’ average price to book ratio (the ratio of share price to book share value) at the same date was 67%. These ratios suggest that the markets think that banks are still seriously impaired. Thus, the markets don’t believe the Bank of England’s narrative about the banks being fixed. And if the markets don’t believe the Bank, why should anyone else?

The combination, of high and increasing leverage and low price to book, is the sow’s ear that the Bank of England’s stress tests portray as a silk purse.

Like their predecessors, the central purpose of the stress tests is to persuade us that the banking system is strong when the evidence indicates it is not. The stress tests themselves are fatally flawed: they are undermined by conflicted objectives, inadequate scenarios, reliance on discredited metrics, low pass standards and woefully inadequate loss modelling. The stress test programme should be scrapped.

The banking system is weak nowbefore any ‘worse than financial crisis’ stress scenario, not strong after one.

All the stress tests provide is employment for modellers and false risk comfort for the public. False risk comfort is never good, however. It is about as useful as a cancer test that doesn’t work. If the banking system is weak, the public need to know.

The next downturn is only a matter of time and the UK banking system is nowhere near prepared for it.

The stress test programme is so severely compromised that it should be scrapped. Instead, the Bank of England and the Government should focus on the reforms really needed to get the UK banking system on its feet – raising capital standards, shutting down zombie banks, establishing tighter corporate governance and reforming accounting standards.


  1. The way RWAs work is simple. Every asset is given an arbitrary fixed ‘risk weight’ that is usually between 0% and 100% but in unusual cases more. The ‘risk-weighted’ asset is then equal to the risk weight times the size of the position. So, for example, Greek government debt would be given a zero risk weight on the grounds that it is deemed to be riskless.

Tags from the story
, ,
More from Prof Kevin Dowd
Lessons from the financial crisis: a libertarian perspective
This article is an expanded version of the Second Chris R. Tame...
Read More
0 replies on “The Bank’s ‘Stress’ Tests”