Stress tests: everyone wins, everyone loses

The investigations into European banks were no ‘stress test’, but a pan-continental PR exercise. Nearly every bank passed. That means we all lose, says James Ferguson.

The European bank stress tests: a huge relief, or a complete whitewash? Officially, only seven of the 91 major banks tested (representing two-thirds of all Europe's banking assets) failed. That sounds pretty good too good to be true, in fact. The only banks that failed were ones that were already in the process of being rescued. And even under the most adverse assumptions, the amount needed to keep the sector well capitalised was put at an almost ludicrously tiny e3.5bn. This wasn't a stress test. It was a pan-continental PR announcement.

So can we ignore the tests as irrelevant? Bond markets showed their near-total disdain by doing practically nothing after the results were announced. The Libor-OIS spread, the benchmark cost of three-month interbank loans, worsened slightly. And bank sector CDS spreads widened, signalling that the risk of default had risen after the tests. But nothing dramatic happened.

However, history will show that this was anything but a non-event. By failing to take charge, European regulators have pushed back the resolution of this crisis far longer than necessary. Here's why...

Subscribe to MoneyWeek

Subscribe to MoneyWeek today and get your first six magazine issues absolutely FREE

Get 6 issues free

Sign up to Money Morning

Don't miss the latest investment and personal finances news, market analysis, plus money-saving tips with our free twice-daily newsletter

Don't miss the latest investment and personal finances news, market analysis, plus money-saving tips with our free twice-daily newsletter

Sign up

The US stress test

To get an idea of how flawed the European stress test was, let's look back at how the Americans did it. The US stress tested its banks in March/April 2009. The banks had already been forced to raise $400bn in new capital to cover mark-to-market' losses on securities. Changes to accounting rules meant that further big securities write-downs could be avoided. What really concerned the US Treasury was upcoming losses on the loan book.

However, Treasury Secretary Tim Geithner was running low on funds with which to underwrite any capital raisings. He had barely $100bn left in his TARP budget. So the US stress test concluded that the banks needed to raise surprise, surprise another $100bn. If there had been more money in Geithner's pot, no doubt the sum that banks needed to find would have risen to match. And during the first two quarters of 2009, just as they were raising that $100bn in new stress test' capital, banks also reported losses of $100bn. The only actual improvement in their capital positions came from retained earnings.

It's clear by now that the fresh capital didn't fix' the banks. In the three quarters since those capital raisings, US banks have only been able to tap the market for a further $9bn. Yet a further $100bn-plus of loan losses has rolled in over that period. Almost the entire commercial property book is being rolled over on an extend and pretend' basis. And the number of mortgages where homeowners are well behind on their payments is now five times the level of foreclosures.

So it's a fair bet that hidden US bank losses are still mounting. And this explains why bank lending is shrinking at an annual rate of more than 7%. This always happens after a bank crisis. Banks can't realise their losses all at once, because that would eat up their capital. They can only grind out the losses as quickly as they generate retained earnings to offset them. In the meantime, it makes no sense for them to extend new loans except to ultra-low-risk borrowers the government, in other words.

The lessons for Europe

Europe and the US have similar-sized banking systems. So it seems fair to assume that their problems may be on a similar scale too. The big difference is that the US went into recession earlier, and so had to start dealing with its bank crisis earlier too. Prior to the US stress test, American banks had already realised losses of $510bn, and raised $400bn in new capital. Ahead of Friday's stress test, European banks had realised some $530bn in losses, against which they'd raised $560bn in new capital. That sounds promising. Unfortunately, European banks were also more highly geared (ie, they had less capital relative to their assets) than their US peers. So the extra capital raised does not necessarily mean they're now on safer ground.

US banks went on to realise a further $200bn in losses between July 2009 and March 2010. Will European banks do the same between now and the end of March 2011? The bad news is that they probably won't. Why is that bad news? Because the bad loans will still be there the banks just won't have raised enough capital to allow them to recognise the fact. And that needs to happen before we can all return to economic normalcy.

The big difference between the US and Europe is that US banks were forced into raising that extra $100bn by US regulators because their stress test was more rigorous. So they could work through $200bn in losses, rather than the $100bn they might only have owned up to if they hadn't raised the extra capital. Sure, we don't know how many more unrealised losses are on US banks' balance sheets. And we won't know it's over until bank lending starts growing again of its own accord. But thanks to their tougher stress tests, the US is likely to be almost a year closer to the finish line than it otherwise would be.

Europe's stress-test disaster

So now we see why Europe's lenient stress tests were such a mistake. Their banks are being told to raise not $100bn, or even $10bn, but just $4.5bn. If you believe the banks are not sitting on any hidden losses, this looks like a good result. But no student of banking crises could possibly believe such a thing particularly as banks are clearly repairing their balance sheets. That's why bank lending right across the continent is shrinking. This never happens except after a bank crisis.

If you want more proof, you just have to look at how flimsy the stress tests were. The test's idea of an adverse' economic outlook was for eurozone GDP to fall by 0.6% in 2011 (hardly a worst-case scenario). And it looked only at haircuts on sovereign bond holdings, not at the possibility of a country actually defaulting, which is what spooked the market in the first place. The test also only took haircuts from the bonds held in the banks' trading books, but not their banking book. George Hay at Breakingviews reckons 23 banks would have failed if haircuts on banking book sovereign bonds had been included too, while Citigroup analysts put it at 24. The actual criteria for passing the test were also extremely low.

All of this shows just what a terrible lost opportunity the stress test has been. If the tests had been stricter, banks could have argued that they were being driven to raise capital, not because they were in dire straits, but by an overly prudent' regulator. This would have enabled them to speed up their loan-loss realisation. Instead, the European regulator has come dangerously close to losing all credibility by telling the market that a sector that has already felt obliged to raise $556bn in new capital in the past two and a half years only needs to raise another $4.5bn.

What does all of this mean?

Forget specific fears over subprime mortgages, or even sovereign bond defaults. This crisis has always been about one thing: the fact that the banking sector has insufficient capital to absorb losses from risk assets generally without becoming insolvent. It was this realisation by the net lenders in the interbank market that led to short-term wholesale funding drying up, and the failure of Northern Rock, for example.

This is also why the banks have to deal with each failing asset class separately and in turn. They just don't have the capital to deal with everything at once. Securities may push their way to the front of the queue now and then, but it's loans that will most likely make up the bulk of the final losses.

While sovereign bonds are top of the worry' pile just now, they're actually the least of banks' concerns. OECD sovereign bonds are officially classed as risk-free assets, at least within their local currency bloc. You may not consider this realistic. But it's fairly logical as a rule, the assets you should worry most about from a loss-generation point of view, are those classed as risk assets ie, loans to the private sector, rather than governments.

Unlike securities, losses on loans can usually be realised at a pace determined by the banks themselves. Loans in default must be provisioned against, of course, but not necessarily those that have renegotiated their repayment rate. Loans in breach of covenant should also technically be added to loan loss provisions but this isn't happening.

Anecdotal evidence suggests that British banks are refusing to carry out written valuations in order to disguise what proportion of their commercial loan book is in breach of covenant. In America, the accounting rules were changed last October so that all breaches of covenant, as long as they haven't actually defaulted on payments, can still be classified as performing loans. So maturing loans that can't be refinanced or repaid are having their maturity extended instead, with no write-down by the bank required. This is the extend and pretend' approach living in hope that a recovery will make the problem go away. This means that banks are already understating their loan losses. So if the European stress test based its expected future rate of losses on those generated by the 2008-2009 recession, it almost certainly understated them quite sizeably. So banks are likely to be even more short of capital than we'd believed.

The one thing that could have made a difference to bank balance sheets right now would have been a decent-sized capital injection. It may not have been enough on its own to get them to full resolution, but it would've helped. Banks can't easily raise capital through new rights issues when they actually get into trouble tanking share prices put paid to that idea so being forced to do so by an apparently paranoid regulator is a much better way to do things. In fact, it's just about the only way that banks can raise new loss-absorbing capital without having their solvency overtly questioned.

The European stress tests didn't do that. Politics won out. No single country wanted any of their banks exposed as failures, so the stress test assumptions were set just the right side of passable. But closer inspection of the numbers reveals frighteningly threadbare capital cushions from Britain to Bratislava. Everybody wins and so everybody loses.

Japanese banks refused to admit the scale of their hidden losses and tried instead to trade their way through, subjecting the country to a lost' decade of negative lending and zero GDP growth. By taking the politically easy way out, Europe has effectively voted for the same thing the long drawn-out resolution rather than an accelerated version. I fear we will rue this day for many years to come.

James Ferguson is head of strategy at Arbuthnot Securities Ltd.

Just how soft were the tests?

The capital ratio pass rate for the banks was set at 6%. To be specific, the 6% cut-off measures a bank's Tier 1 capital to risk-weighted assets. But Tier 1 capital isn't what you or I might hope. In a crisis, banks must draw on their capital to absorb losses. So you'd think the definition of capital would be a bank's equity. But in fact, that's what core' Tier 1, or more precisely tangible common equity' (TCE), consists of. Tier 1 capital, on the other hand, includes all sorts of items that are irrelevant in times of crisis (because they can't absorb losses), such as deferred tax assets and even some debt instruments.

And this makes a big difference. For British banks, Tier 1 capital is on average about 50% larger than their TCE. But for some European banks the gap can be much larger. Commerzbank, for example, one of the largest German banks, had Tier 1 capital of e29.5bn in the stress test. This gave it a comfortable 9.1% Tier 1 ratio under the adverse scenario as a result. But if you use the relatively tiny TCE measure of capital (e5.6bn), the ratio dives to 1.7%.

Then there's the use of risk-weighted assets (RWA) rather than tangible assets. You may think that a stress test capital ratio of 6% means there's enough capital for a bank to handle an average loss of 6% across the board. But you'd be wrong. RWA measures what the loss for any given asset might be, then uses only that figure. So the RWA of a mortgage loan might be less than half the size of the actual loan. Take Deutsche Bank. Its RWAs in the stress test totalled e273.5bn, yet total assets were actually five and a half times bigger than that, at e1.5trn. Deutsche Bank's stress test result showed a supposedly healthy 9.7% capital ratio. But had the test measured Tier 1 capital to total assets, it would have been just 1.8%. And loss-absorbing capital (TCE) is a mere 1.4% of total assets.

In case you think I'm picking on the Germans, let's look at the top performer among the European major banks: Barclays. It ended up with a 13.7% Tier 1 to RWA ratio under the adverse scenario. That's because the test assumes that 2009 pre-impairment income of £20.3bn will be maintained through 2010 and 2011.

However, Barclays sold BGI to Blackrock for $12.5bn (£7.7bn) in December 2009, booking a £6.3bn profit. Excluding this one-off profit, and a further £446m in profit from BGI's 2009 operations and £831m in insurance claims, likely pre-impairment earnings over the next two years will be some £15.2bn lower than advertised. On that basis, the published stress test ratio should have read 11.3%. If instead of Tier 1 vs RWA, we look at Barclays' loss-absorbing capital to total assets, the true capital ratio comes out at just 1.7%.

What does it mean for investors?

The basic point is that the longer it takes for banks to recognise their losses, the longer it will be before they start lending again. And if banks aren't lending, the economy will be prone to relapses and possibly even deflation. That means investors should play safe.

Firstly, stick with defensive, highly liquid blue-chip stocks, such as AstraZeneca (LSE: AZN), Vodafone (LSE: VOD), and National Grid (LSE: NG), for example. Secondly and I know I'm going against the broader MoneyWeek view on this I'd be happy to buy gilts against this deflationary backdrop. Thirdly, I'd avoid or sell banks they won't be able to afford decent dividends for years and may well need highly dilutive capital raisings in the future. Lastly, I'd be a buyer of gold on weakness (say around $1,050 an ounce) as a hedge against a run on sterling and the other Western currencies.

James Ferguson qualified with an MA (Hons) in economics from Edinburgh University in 1985. For the last 21 years he has had a high-powered career in institutional stock broking, specialising in equities, working for Nomura, Robert Fleming, SBC Warburg, Dresdner Kleinwort Wasserstein and Mitsubishi Securities.