Home
 
Choose Your City
Change City

Swedish banks - safe bet or risky business?

By Johan Ahlander

STOCKHOLM (Reuters) - Sweden's four largest banks are using a calculation of the risk to their loan portfolios that critics say is flawed and leaves them vulnerable to any correction in the booming housing market.

Nordea <NDA.ST>, Swedbank <SWEDa.ST>, SEB <SEBa.ST and Handelsbanken <SHBa.ST> are among the world's most well capitalized banks and made it through the 2008 financial crisis unscathed, relative to other European banks.

Swedish house prices took a small dip during the crisis but have more than tripled in the last 20 years, driven by low interest rates, tax breaks on debt and low building. Swedish households are now among the most indebted in Europe

RelatedArticles

Encouraged by the surging prices, banks have stocked up on property loans. They now account for close to 65 percent of the loan portfolios of the big four, according to the Swedish Financial Supervisory Authority (FSA).

While they have built up their home lending, the lenders' capital positions also appear to have improved.

But critics including Sweden's central bank and the International Monetary Fund say this is because since 2007 the banks were allowed to use a model where they self asses the risk of their portfolios based historic loan losses.

With the housing sector making up such a large chunk of Swedish loans, losses have been low for a long time. This means the amount of capital banks need to reserve against future shortfalls has also been set very low, so they have not built buffers big enough to compensate for bigger losses, they say.

"They haven't (increased their capital). They have reduced the risk-weights and it is quite different because the leverage ratio has been almost constant," Swedish central bank governor Stefan Ingves said in December.

Since 2010, Swedish banks have doubled their core tier 1 capital ratio, a measure of how much money a bank has in relation to risk-weighted assets, to around 24 percent, above the EU average of around 13 percent.

But the leverage ratio, the relationship between the banks capital and its total assets regardless of what risk you assign to them, have remained virtually unchanged at a average of around 4 percent.

Banks put a lower risk weighting on loans that are less likely to fail and set aside less money against losses of those loans. The idea is that this serves as an incentive to reduce the risks in the loan portfolios.

Spokesmen for Swedbank, SEB and Handelsbanken declined to comment on the risk weightings.

Nordea's head of investor relations, Rodney Alven, said: "What makes us favor risk-adjusted models is that it has taught us a lot about how to manage, minimize and properly price risk. We think that these models have genuinely reduced the risks in the banking system."

"We feel we are very well capitalized in every way you measure it," he added.

SELF ASSESSMENT IN THE SPOTLIGHT

Swedish banks often bemoan a heavier regulatory capital burden than counterparts in Europe. Sweden has tried to adopt a gold-standard approach to capital buffers to preserve the banks' reputations as solid institutions and because of the importance of the financial sector to the economy.

Only Switzerland and the Netherlands have a larger financial sector than Sweden in the European Union relative to the size of the economy.

Self-assessed risk-weightings are widely used in Europe but Sweden's are among the lowest, according to the IMF in a 2016 report.

Ingves, who was assigned to clean up the Swedish banking system after the 1990s crisis, says it is time to rein the banks in. He has called for a leverage ratio requirement to balance out creative internal models.

"It has been proved since the system was launched that banks have had too large degrees of freedom to decide their risk weights and that needs be to revised," Ingves said.

His sentiment is echoed by the IMF which wrote in its report that "available models may suffer from overreliance on recent historical experience, and have difficulties capturing unexpected losses occurring in extreme but plausible scenarios."

The internal models are also under scrutiny in other parts of the banking world as the Basel committee is trying to agree on a floor for how low banks risks can be set. The final levels are expected at the end of the month and could see Swedish banks having to set aside more capital against their lending.

The Swedish Bankers' Association says such models are still the best way to capture risk.

"If you try to have a single standard for all banks, it's the ones that work best that are most punished," said Johan Hansing, CEO of the association.

"All the Swedish banks have a very strong rating by rating agencies. If there were weaknesses and shortcomings in the Swedish banks' internal models professional actors like rating agencies would detect it."

The Riksbank says that international surveys have shown that there are major differences in banks' risk weights even for identical portfolios. It advocates a leverage ratio requirement of 4 percent now and a hike to 5 percent by 2018.

The FSA, responsible for financial stability in Sweden, agrees rules have been too lax.

"We agree partly to that criticism and that is why we have tightened rules," Swedish FSA general director Erik Thedeen said. "But it is to simplistic to say that just because the accumulated capital in relation to the assets have not gone up, the capital has not increased."

The FSA has taken a number of measures to shore up the system. A 25 percent risk-weight floor on mortgages has been introduced and banks are forced to count every fifth year as a bad one when calculating for loan losses.

The FSA hiked Nordea's capital requirements with 13.8 billion Swedish crowns ($1.5 billion) in October after a review found that the bank had underestimated the probability for loan losses in its internal models.

Asked if it is wise to let banks regulate themselves given that many crises have started with excessive risk-taking by banks, Ingves says: "In retrospect, when you look at it: no, not really."

(This analysis was refiled to fix typo in the first graph)

(Editing by Anna Willard)