Why can’t the ECB bring itself to stress test banks for deflation? I was thinking about this question when I remembered an episode from the past.

Just over ten years ago, I took a train from Frankfurt to the sleepy spa town of Wiesbaden. It was the sort of place where a lone taxi is parked outside the station and the non-English speaking driver continues smoking his cigarette after you get in his cab. I arrived for my meeting at R+V, a huge mutual life insurance company virtually unknown outside Germany, where I met the company’s chief investment officer.

He told me about the policyholder guarantees lurking on German insurer balance sheets, and their vulnerability to what he termed a ‘Japanese scenario’ of persistent low interest rates and a sluggish economy¹⁾The German policyholders were promised a minimum return of about 3.5 per cent, while ten-year bund yields in 2004 were well over 4 per cent. Because insurers can hold on to assets until they mature, yields would have to stay below guarantee levels and return-seeking assets would have to underperform for a long time before it caused solvency problems for the insurer.. Although the word wouldn’t be coined for another decade, it was the first time I’d heard about the risk of so-called ‘lowflation’.

On that sunny day in Wiesbaden the risk seemed impossibly remote. The Eurozone economy was healthy, markets were booming and the European Central Bank was obsessed about inflation. And I discovered that few in Germany’s opaque and politically influential insurance sector shared my CIO contact’s concerns²⁾There were a few insurers, such as Munich Re, that acted on the concern and bought swaption hedges..

A couple of weeks ago, bund yields hit 0.76 per cent ““ their lowest level since records began, while the ECB’s eurozone benchmark yield fell below 1 per cent. It turns out that the scenario the prescient CIO in Wiesbaden told me about a decade ago is actually with us. 

This week, the Bundesbank published a report warning that German life insurers might need more than €10.6 billion in additional capital to cover customer guarantees by 2023. This would happen under a ‘severe stress scenario’ in which bund yields followed the path of Japanese government bonds from 2003 onwards while other asset classes failed to deliver excess returns for the insurers.

This brings me back to the ECB’s stress tests and the adverse economic scenario used in the test. In particular, the bond yields of countries including Germany, as well as swap rates, were envisaged to increase in this scenario which entails a decrease in euro area inflation to near-zero levels.

There are a couple of problems with this scenario, which was devised in April by the European Systemic Risk Board. Firstly, since the financial crisis began, falling inflation has been accompanied by falling bond yields for core Eurozone countries (see chart). The mechanism behind that link is increasingly clear. Lack of growth, and deflationary pressure in euro periphery countries, is pushing the ECB towards outright quantitative easing and bond buying. That pushes government bond yields and swap rates down to record lows.

The Bundesbank paper explains the impact that falling rates have on life insurance companies, such as the temptation to ‘reach for yield’ in order to stave off default. Having being given a new systemic risk mandate, the Bundesbank worries about the contagion potential but luckily doesn’t have to regulate the insurers itself (that’s BaFin’s job).

Compare this to the ECB’s connection with banks. There’s no question that banks are vulnerable to lowflation. In a world of ultra-low short-term rates, falling longer-term yields hurt banks’ net interest margin, or the difference between what they earn on loans and pay out on deposits. Either banks try to hedge against falling rates ““ which is risky in itself—or pray for a time when rates rise. That’s right: they pray for the kind of government bond yields the ECB used in its adverse scenario.

Then there is the question of those banks that have structural derivatives exposures to falling rates and can’t raise the cash to pay margin calls, such as we saw with Dexia in 2011. These exposures can be opaque: in its 2010 annual report, Dexia claimed that it could lose money if long-term rates increased. In reality, rates decreased sharply and Dexia was bailed out.

Among the various flaws in the ECB’s stress tests, not addressing these risks is one of the most egregious, so much so that it needs a psychological explanation.

Perhaps one reason for the omission is that the ECB would have to model the impact of its own evolving policies on the banks it regulates. How else to explain the circularity in ECB vice-president Vitor Constancio’s remark that deflation wasn’t considered in the scenarios because “we don’t consider that deflation is going to happen”? (My translation: “We’ll do whatever it takes to make sure it doesn’t happen, even if we have to kill the banks we regulate”)

No wonder deflation (or lowflation) didn’t make it into the stress test.