Bank of England fails its stress test again

On November 30th, the Bank of England released the results of its third publicly disclosed set of stress tests of the financial resilience of the UK banking system.  

The good news is that the news is good: our banking system is in good shape, but the bad news is that the good news is not credible.

In this post, I would like to put the stress test results through my own favourite stress test – a reality check.

By this most basic of tests the Bank scores an ‘F’.

Read on:

The spin is that the banking system passes with flying colours even though one of the seven banks involved failed (RBS) and two others (Barclays and Standard Chartered) were deemed as problematic.

Let’s pass over the slight glaring contradiction at the heart of that narrative and quote Governor Carney:

The resilience of the system during the past year in part reflects the consistent build-up of capital resources by banks since the global financial crisis. … the UK banking system is well placed to provide credit to households and businesses during periods of severe stress.

That conclusion is corroborated by the 2016 stress test [which is] broad, coherent and severe …  (Governor’s opening remarks, p. 3)

Every one of these claims is questionable, but let’s focus on the severity of the Bank’s ‘doomsday’ scenario.  

The scenario consists of a bunch of adverse events, including world and UK recessions (annual global GDP growth troughs at -1.9% and UK GDP falls by 4.3%), major falls (over 30% and over 40%) in the prices of houses and commercial real estate, unemployment rising by 4.5 percentage points and sundry other stuff. However, the key phrase is this: “overall, the UK stress is roughly equivalent to that experienced during the financial crisis, albeit with a shallower fall in domestic output” (Bank of England, 2016 stress test results report, p. 6).

Take-home message: the stress scenario was not quite as stressful as the Global Financial Crisis (GFC).


Now consider how this adverse scenario impacts the banks. To quote Carney, the adverse stress scenario led to “system-wide losses of £44 billion over the first two years of the stress – five times those incurred by the same banks over the two years at the height of the financial crisis.” (Governor’s remarks, p. 2)

This statement misled some commentators into thinking that the stress scenario was five times more severe than the GFC, but it isn’t.

Carney’ statement implies that the system-wide losses over the two height years of the crisis were less than £44 billion/5 = £8.8 billion.

This claim is misleading, however. In fact, the banks spread out their reported losses over the period since, because they were afraid of the adverse public reaction if they had disclosed them promptly. Consequently, the appropriate comparison is not with reported losses over these two years only but with cumulative losses post-2007. As James Ferguson of The MacroStrategy Group points out, the cumulative loan losses for the big 4 banks alone were nearly £200 billion and nearly double that if we include balance sheet reserves, securities losses, restructurings/goodwill write-downs and legal redress (Ferguson, 2016, p. 2). [1]

The £44 billion losses generated by the Bank’s stress model are not five times the losses incurred at the height of the crisis but 44¸ 450 or about a tenth of the gross losses banks experienced since 2007.


But how can a supposedly severe stress scenario lead to the moderate losses projected by the Bank’s stress model?

The most plausible answer is that the Bank’s model is wrong. The Bank model’s estimated losses merely indicate that the feedback link between the scenario and the simulated losses in the Bank’s model must (greatly) under-estimate the losses involved.

This point in and of itself is enough to discredit the entire exercise.


In the Q&A at the press conference, Carney makes a related claim:

the capital hit in this stress would have wiped out all of the capital that these same banks had prior to the crisis. So this is a big, big hit to capital. (Financial Stability Report Q&A, p. 19)

This claim is wrong. For the big four banks alone, their 2006 Annual Reports report that their capital going into 2007 was about £150 billion. The projected £44 billion loss from the Bank’s stress test model is barely 30% of this number, and would be even lower if we included the capital of the other banks in the exercise.

A loss of $44 billion is a fairly small hit anyway. With about £2.1 trillion in assets, a loss of $44 billion is equivalent to a loss rate of about 2 percent. Yet since the GFC, the accumulated loss rate so far has been about 10 percent, i.e., five times bigger rather than five times smaller than that simulated by the Bank.


Carney’s responses in the Q&A also alluded to another important issue that deserves more attention: the problem of the incurred loss accounting model by which losses are not recognised until they have occurred – implying that expected losses are not reported – and the inadequacy of the supposed solution to this problem, IFRS 9, which recognises expected losses so long as they are expected within 12 months. As he said in response to a question:

Now there is another issue which is not adjusted for in the stress tests which is coming which is IFRS 9, which not yet finalised and could have some impact. But I think you know the banks, the analyst community, ourselves, we all have equal line of sight to that and its timing. (Press conference Q&A, p. 18)

So Carney acknowledges that the Bank’s stress tests are based on accounting rules that ignore expected losses and some people might regard this omission as a bit of a worry.

But even when it is implemented, IFRS 9 won’t fix the hidden expected losses problem: it will merely encourage banks to adopt practices that make sure that losses are pushed out into the future so they are expected to occur more than 12 months hence. As the UK’s leading financial accountant Tim Bush wrote me: IFRS 9 “is going to be a disaster.”


There is one last issue that surfaced in the discussion last week: the setting of the Countercylical Capital Buffer (CCyB). Back in July, the FPC reduced the CCyB on banks’ UK exposures from 0.5% to 0%. To quote Carney again:

The FPC was concerned that banks could respond to these developments by hoarding capital and restricting lending. The reduction of the CCyB rate was intended to reinforce the FPC’s expectation that all elements of capital and liquidity buffers are able to be drawn on to support the real economy.

That position has not changed. In light of the continued uncertainty around the UK economic outlook and the resilience demonstrated in the 2016 stress test, the FPC agreed to maintain the CCyB rate at 0% and that it expects, absent any material change in the outlook, to maintain this rate until at least June 2017. (Governor’s opening remarks, pp. 4-5)

I don’t understand this passage. What does it mean to say that the FPC expects “that all elements of capital and liquidity buffers are able to be drawn on to support the real economy” and how is this relevant to the decisions first to reduce the CCyB to zero and then to keep it there?

And what is the point of relying on “continued uncertainty” to justify any CCyB decision, given that uncertainty, like the poor, will always be with us?

There is a deeper problem: the FPC seems to have it the wrong way round. The purpose of the CCyB is to counter the financial cycle: as aggregate credit builds, markets boom and risks build up; then the boom breaks, markets fall and the risks are realised and subsequently fall. [2] The CCyB should rise in the first phase to help slow the euphoria, and then fall in the second phase to ameliorate the distress. Over the last few years markets have been booming, so we are presumably still in Phase 1. If so, then the FPC should be to increase rates instead of to reduce them. The FPC’s ‘countercyclical’ policy is therefore procyclical: it is aggravating the problems it is meant to ameliorate!

On the other hand, it may be that the FPC’s and Carney’s thinking is that we are actually in Phase 2, the down phase of the cycle, in which case reducing the CCyB would make sense – if one accepts that assessment and buys into countercyclical financial policy in the first place, which I don’t. However, that is not the message that clearly comes across from their statements.

At the very least, the Bank should always base its decisions on CCyB settings on a clear statement about which phase of the cycle they think we are in: Phase 1 implying that the decision to be considered is to raise the buffer or Phase 2 implying the opposite. However, I suspect they don’t do that because they don’t know themselves or because they don’t wish to expose their views on the matter to criticism. Hence the Bank waffle around the issue. If this latter conjecture is correct, they would be better off abandoning CCyB policy altogether.

One gets the distinct impression that all that they see is uncertainty, so they think “uncertainty is bad so we should ease” and cut the buffer. However, if that is what they do, they are not even attempting to follow countercyclical policy: all they have is a policy bias towards ease, which is the same bias that we see with unconventional policy for most of the past decade.

But this might just be me. All I can say is that I don’t understand their CCyB policy and I suspect they don’t either.

There are further problems with the stress test results too, but I will come to those in my next posting.


Bank for International Settlements (2016) “Countercyclical capital buffer.” July 20. Available at

Bank of England, Financial Stability Report Press Conference Q&A, 30 November 2016. Available at

Bank of England, “Opening remarks by the Governor.” Financial Stability Report Press Conference, 30 November 2016. Available at

Bank of England (2016) “Stress testing the UK banking system: 2016 results.” Bank of England 30 November 2016. Available at

Ferguson, J. (2016) “UK bank ‘stress’ test.” The Macrostrategy Partnership, 2 December 2016.

Local Authority Pension Fund Forum (2011) UK and Irish Banks Capital Losses – Post Mortem. London: LAPFF.

Parliamentary Commission on Banking Standards (2013) ‘An accident waiting to happen’: The failure of HBOS. London: Stationery Office.

End Notes

* Kevin Dowd ( is professor of finance and economics at Durham University. I thank Tim Bush, James Ferguson, Martin Hutchinson and Basil Zafiriou for helpful comments.

[1] Let me also cite some other numbers give an idea of the true scale of losses. (i) Tim Bush in his LAPFF Post-Mortem report (LAPFF, 2011) cited bank losses over the 2007-2010 of over £89.4 billion. (ii) The Parliamentary Commission on Banking Standards (2013) reported that the losses for HBOS alone were £46.5 billion. (iii) The Asset Protection Scheme has estimated the losses to RBS at almost £60 billion.

[2] Consider this quote from a recent BIS statement (BIS, 2016): “The countercyclical capital buffer aims to ensure that banking sector capital requirements take account of the macro-financial environment in which banks operate. Its primary objective is to use a buffer of capital to achieve the broader macroprudential goal of protecting the banking sector from periods of excess aggregate credit growth that have often been associated with the build-up of system-wide risk. Due to its countercyclical nature, the countercyclical capital buffer regime may also help to lean against the build-up phase of the credit cycle in the first place. In downturns, the regime should help to reduce the risk that the supply of credit will be constrained by regulatory capital requirements that could undermine the performance of the real economy and result in additional credit losses in the banking system.”