Sweden's Banking Crisis: The 500% Rate and the 1992 Bank Guarantee

On the afternoon of 16 September 1992 the Riksbank told Sweden's banks what it would cost them to borrow overnight: 500 per cent. Bengt Dennis, governor since 1982, had already pushed the marginal lending rate from 24 per cent to 75 in the space of a single day the week before, and the market had gone on selling kronor anyway. Five hundred was not a monetary-policy setting in any ordinary sense. It was an announcement that the central bank would make a short position in the krona cost more, day by day, than any plausible devaluation could ever pay — for as long as anyone cared to test it.

Dennis called his memoir of that autumn simply 500%. The number is what Swedes remember, and it is close to the least important thing that happened. Behind the defence of the exchange rate sat a banking system that had already lost more money than it held in capital, and behind that sat six years of credit growth traceable to one decision taken quietly in a Stockholm boardroom in November 1985.

The November Revolution

Until that month, Swedish banks lent what the Riksbank allowed them to lend. Ceilings on bank lending, liquidity ratios, and requirements to hold government and housing bonds had rationed credit since the 1950s, and the rationing worked: households borrowed through queues, firms borrowed through relationships, and a large grey market in company-to-company credit grew up alongside.

On 21 November 1985 the Riksbank abolished the lending ceilings. Swedes call it novemberrevolutionen, the November Revolution, and the name is not much of an exaggeration. Exchange controls followed by 1989. Within five years, lending by banks and finance companies had risen roughly 170 per cent in nominal terms, and the ratio of bank credit to GDP had climbed by about half (Englund, 1999).

Nothing in the tax code discouraged it. Marginal income-tax rates reached into the seventies, interest paid was fully deductible against them, and inflation ran in the high single digits, so a Swedish borrower in 1988 was being paid to hold debt. Real after-tax interest rates on mortgage and property lending were negative. The 1990–91 tax reform, the largest in Swedish post-war history, cut the top marginal rate and capped the value of interest deductions at 30 per cent. It did so precisely when property was priced on the assumption that borrowing would stay cheap.

Real prices for commercial property roughly doubled between 1985 and 1990 and then fell by more than half by 1993. That arc is the crisis in one line. Everything else is a question of who was holding the paper when the second half happened.

The Finance Companies Go First

Sweden had around 300 finansbolag — finance companies outside the banking regulations, funding themselves in the market with short-term paper and lending it long against commercial real estate. Banks owned some of them and lent to most of them. They were, in the vocabulary of a later crisis, the shadow banking system, and they had the shadow banking system's characteristic weakness.

In September 1990 Nyckeln, one of the largest, could not roll over its maturing commercial paper and suspended payments. The market for finance-company paper closed within days, which meant that every other finansbolag faced the same wall at the same moment. Gamlestaden was rescued by its owners. Others were not. Their assets, being mostly property loans, went onto the market together, and the property market that had to absorb them was the one whose prices had just stopped rising.

Banks discovered that they had underwritten the risk twice: once in the credit lines they had extended to the finance companies, and again in the property lending they had done directly. Between 1990 and 1993, cumulative loan losses across Swedish banks came to roughly 12 per cent of GDP, peaking in 1992 at about 7.5 per cent of total lending in a single year (Englund, 1999). No banking system carries that.

Five Out of Six

Första Sparbanken was the first bank to need the state, receiving a guarantee in the autumn of 1991. Nordbanken came next. Already majority state-owned, it was recapitalised in 1991, and in early 1992 the government bought out its remaining private shareholders outright rather than hand them a free ride on a rescue. In September 1992 Gota Bank's parent company announced it would put no more money in; the government guaranteed Gota's obligations the same day, and in December bought the bank for one krona after the parent collapsed.

Six of Sweden's seven significant banking groups asked for or accepted state help. Handelsbanken, which had been through its own near-death experience in the early 1970s and had run a conservative credit policy ever since, did not. Its survival became the standing argument in Swedish banking that credit discipline is a cultural property of an institution rather than a regulatory one.

Defending a Number

Sweden had pegged the krona unilaterally to the ECU on 17 May 1991, a declaration of European intent by a country that had applied for membership of the Community two months later. The peg was not a treaty obligation and had no partners committed to defending it, which made it both cheaper to adopt and easier to attack.

Attack came from the east. Finland, whose banking crisis ran a year ahead of Sweden's, floated the markka on 8 September 1992, and traders drew the obvious inference about the country next door. What followed was the fastest sequence of interest-rate decisions in Riksbank history.

Rates at those levels do real damage to a banking system whose borrowers are already failing, which is what made the defence self-defeating in a way the Bank of England's simultaneous struggle was not. A British bank in September 1992 faced an expensive week. A Swedish bank faced an expensive week on top of a loan book already impaired by a fifth. Every hour the krona held was an hour in which the cost of holding it rose.

Politics moved faster than the market for once. Carl Bildt's centre-right government and the Social Democratic opposition under Ingvar Carlsson agreed two emergency packages of spending cuts, on 20 and 30 September, negotiated across the aisle in a matter of days. Finance minister Anne Wibble — who had already told Swedes that everyone should have en årslön på banken, a year's salary in the bank — signed off on cuts her own party had campaigned against. The rate came down from 500 to 50 within five days.

The Guarantee

On 24 September 1992, in the middle of the currency defence, the government announced something more consequential than any interest rate: the Swedish state would guarantee that banks and certain other credit institutions could meet their commitments on time. The guarantee was unlimited in amount, unlimited in duration, and covered every creditor except shareholders. The Riksdag passed it into law on 18 December with near-unanimous support, and a Bank Support Authority under Stefan Ingves began operating in May 1993 to administer it.

Four principles governed what followed, and they are the reason the Swedish episode is still taught. Losses had to be disclosed in full and written down immediately, with the authorities commissioning their own valuations rather than accepting the banks'. Support went to institutions, not to owners: any bank that took public capital gave up equity in proportion, which is why Nordbanken's minority shareholders were bought out first. Bad assets were separated from good ones, with Securum taking roughly SEK 67 billion of Nordbanken's property loans and Retriva doing the same for Gota. And the whole thing was run by an agency at arm's length from the ministry, with a mandate to sell.

Ingves and Lind, writing in the Riksbank's own quarterly three years later, described the aim as separating the question of whether a bank was viable from the question of what its assets were worth, so that neither answer could be used to postpone the other (Ingves and Lind, 1996). A bank judged solvent got liquidity. A bank judged insolvent got wound down. Nothing was allowed to stay ambiguous, because ambiguity is what a creditor runs from.

Letting Go

The defence ended on 19 November 1992. The Riksbank abandoned the fixed rate at 2.28 that afternoon, and the krona fell immediately, giving up around 20 per cent on a trade-weighted basis over the following months. Dennis cut the marginal rate to 12.5 per cent the next day.

Failure turned out to be the most useful policy of the whole period. Swedish exporters, whose costs were in kronor and whose revenues were not, got a devaluation they had not dared ask for; the current account swung from deficit to surplus within two years. On 15 January 1993 the Riksbank replaced the exchange-rate anchor with a different one, announcing an inflation target of 2 per cent to take effect in 1995 — one of the earliest adoptions of the framework that has governed Swedish monetary policy since. The country that had spent a decade defending a number for the krona spent the next three decades defending a number for the consumer price index, which proved considerably easier.

The contrast with the United Kingdom's exit from the Exchange Rate Mechanism two months earlier is instructive, and the two episodes are usually told as one story: see the 1992 speculative attack on sterling. Britain's crisis was a currency crisis with no banking crisis underneath it. Sweden's was a banking crisis wearing a currency crisis on the outside.

The Bill

Gross state outlays on the bank rescue reached around SEK 65 billion, roughly 4 per cent of 1993 GDP. That figure has been shrinking ever since, because the state got paid back. Securum wound itself up in 1997, years ahead of its planned life, having sold property into a recovering market. Nordbanken merged with Finland's Merita in 1997 and became Nordea; Sweden sold its last shares in 2013. Depending on how the state's equity holdings are valued, the net fiscal cost of the bank support ranges from about 2 per cent of GDP down to approximately nothing (Jonung, 2009).

The real bill was somewhere else. Output fell in three consecutive years, 1991 through 1993, for a cumulative decline of around 5 per cent. Open unemployment went from under 2 per cent in 1990 to above 8 per cent by 1993, and to nearly 13 per cent counting people in labour-market programmes — in a country where full employment had been a policy commitment since the 1930s. The general government deficit reached about 11 per cent of GDP in 1993, and gross public debt nearly doubled as a share of the economy. Laeven and Valencia's systemic-crisis database records the Swedish output loss as far larger than the fiscal one, which is the usual pattern and the one most easily forgotten when the fiscal number is the one in the headlines (Laeven and Valencia, 2013).

Göran Persson, who took the finance ministry in 1994 and inherited the consolidation, summarised the decade in a sentence that became the title of his own account of it: Den som är satt i skuld är icke fri — he who is in debt is not free. The budget consolidation that followed ran to some 8 per cent of GDP in cuts and tax rises, and it produced a surplus target and a spending-ceiling framework that outlasted every government since.

What Sweden Was Remembered For

Reinhart and Rogoff put Sweden among the "Big Five" post-war crises in advanced economies, alongside Spain, Norway, Finland and Japan, on the grounds that all five produced deep and durable output losses rather than the sharp reversals that textbook recessions promise (Reinhart and Rogoff, 2009). Four of the five clustered in a handful of years around 1990, which was not coincidence: Scandinavian deregulation, the Japanese property and equity bubble, and the American savings and loan collapse all ran on the same sequence of liberalised lending, collateral inflation, and a tightening that arrived after the collateral had already turned.

The blanket guarantee was withdrawn in July 1996, four years after it was given, without incident. Sixteen years later, when American and British officials were working out what to do about their own banks, the Swedish file came off the shelf: Bo Lundgren, the minister who had handled the 1992 rescue and by then ran the National Debt Office, spent the autumn of 2008 explaining to visitors how a government takes equity for its money. Some of what he described was adopted, some of it was not, and the parts that were not are broadly the parts that generated the most political damage afterwards, as the 2008 crisis worked through.

Stefan Ingves, who ran the Bank Support Authority out of a small office in 1993, became governor of the Riksbank in 2006 and chaired the Basel Committee through the writing of the post-2008 capital rules. The institution he eventually led had been founded in 1668 out of the wreckage of a private bank that issued more notes than it could redeem — see the collapse of Stockholms Banco — which gives the Riksbank a claim to more experience of this particular problem than any other central bank on earth.

What Sweden actually proved in 1992 is narrower than the version that circulates. It did not show that bank rescues are cheap; the output loss says otherwise. It showed that a rescue is cheaper when the government writes down the losses before it writes the cheques, and that the political capacity to do that is a resource a country either has in September or does not. Bildt and Carlsson, who agreed on very little else, found they had it.