Documente Academic
Documente Profesional
Documente Cultură
Introduction
The recovery
Conclusion
References
Introducti
on
In the late 1980s and early 1990s, the Scandinavia n count ries
under went their most severe financial crisis of the postwar period.
While such crises are not rare, what sets thes e count ries apart is
their freedom from the institutiona l weaknesse s that dogged less-
develope d count ries in the 1970s and emerging-ma rket economie s
in the 1990s (slow bank ruptc y procedures, lax law enforcement ,
regulato rs who lacked the power to close banks, and so forth).This
makes the Scandinavia n nations an interesting laboratory for
examining how crise s develop and are ultimate ly resol ved.
In this Poli cy Discussio n Paper , I focus on the expe rienc e of
Sweden , first reviewing the policy choices and exte rnal factors that
pushed this count ry’s financial system over the edge, and then
discussin g what steps the governmen t took to make its crisis
resolution proces s one of the most successfu l of the past 30 years.
The paper emphasizes the resolution proces s rather than the way in
which t h e financial c r i s i s unfolded , because any number of event s
can lead to crises, but the general pr in c ipl es for resolvin g crises
successful ly are uni versal.
The Road to
Crisis
In the mid-1970s , Sweden was a small, open econo my, whose growth
depended heavily on expo rts. Its economi c policie s were influence d
by the core tenet of its politica l system: The government ’s primary
focus should be keepin g unemploymen t low and promoting
economi c equalit y. At the same time, its policie s had to be aligned
with the central bank’s commitmen t to keepin g the exchan ge rate
fixed relative to a basket of Sweden ’s trading partne rs’ currencie s
(detail s later), all within the context of a closed-capital-account
envi ronment.
How did these constraints affect the economy? The succes s of the
expo rt-base d growth mode l is conditiona l on keepin g the tradable
goods indust ries competiti ve in global mar- kets ; this means that
productio n costs cannot grow faster than competito rs’ as long as
the exchan ge rate is fixed. Yet in the 1970s, wage inflation grew out
of cont rol. The seeds of inflation had been sown in the 1960s and
early 1970s, when policymake rs thought they faced only a
tempo rary slowdown in productivit y growth and in the development
of Sweden ’s primary export indust ries (such as forestry products
and ship building) , but in realit y, the slowdown had resulted from
tougher foreign competition . Although the natu ral rate of
employmen t was rising, the government would not back off from
the unsustain able policy of full employmen t to which it had
committe d itself. Instead, it decide d to overcom e economi c
weaknes s by increasin g its expenditu res on welfare pro- grams
(Bosworth and Lawrence , 1987) , a further sign that the governmen t
thought the decline i n productivit y growth w a s temporary. Workers
in export i n d u s t r i e s who lost
Their jobs were retrained a n d found employmen t in the white-
colla r service sector or
1
The public secto r. The ensuing growth of public-secto r employmen t T
was rema rkable: The sector ’s share of total employmen t grew from h
e
20 percent in 1965 to 30 percent in 1975 and 38 percent in 1985 .1
In hindsight , this policy seems to have been a costly alte rnati ve S
to higher unemployed- ment.As budget deficits grew, the holes t
a
were plugged by foreign borrowing and higher income taxes (the g
margina l income tax rate on full-time workers earning the e
average hour ly wage inc rease d from 35 percent in the second half
I
of the 1960s to 65 percent in s
1976) .The disincenti ve to invest created by higher income taxes
S
probably cost Sweden dear ly in future private-sector job creation .
e
And the exchan ge rate risk the government assumed by borrowing t
in inte rnationa l markets came back to haunt it in later years, when
High inflation
2
risin g labor costs and deficits necessitate d large devaluations.
made the fixed-
The growth in public-secto r and white-colla r employmen t also
had an unintended exchange-rate
policy
consequence : It strengthene d the unions that represente d these
unten able, so
workers, which, given the circumstances , played a signif icant role
the krona was
in breaking the link between wage growth and productivit y
devalued
growth.The Social Democrats who ran Sweden from 1932 to 1976 had
several times.
a very close relationshi p with the count ry’s most powerfu l union of
Foreseein g the
blue-colla r workers, the Landso rganisatione n (Weaver, 1987) .
unsustain abilit y
Both sides unde rstoo d that their interest lay in keepin g export
of the fixed
indust ries competiti ve by holding wage growth in check. The three
exchan ge rate,
white-colla r unions, which also participate d in the negotiatio n
the unions de-
process, were too small to dictate wage polic y. As public and white-
manded higher
colla r employmen t grew and these workers gained bargainin g
wages in
power, the unions began to compet e with each other for new mem-
anticipatio n of
bers. But public-secto r and white-colla r workers had no incenti ve to the krona’s
reduce their wage demands so that expo rt indust ries could remain dec rease d
competiti ve, so unions forsook the cen- trally dete rmine d wages and purchasin g
made side deals with emplo yers for higher wages (Flan agan, power.When
1987) . Soon the reafte r, the link between wage growth and consumer price
productivit y growth broke down . In addition , Sweden ’s gene rous inflation hit 15
unemploymen t benefits may have put the unions in a heads- we-win- percent in 1980,
tails-the- government-lose s situation .The unions won if they could ne- the
gotiat e higher wages for their membe rs, and if companie s failed, governmen t
their membe rs were still covered by government insu ranc e decide d to
(Calmfors, 1987). break the in-
flationary spiral
(i.e., inflation expectations ) and promote economi c growth through a
se- ries of financial reforms, made at intervals throughout the
1980s .The reforms have three majo r component s that are relevant 1. This growth is due, in
part, to the entry of
to our discussion. women into the labor
force and their taking
of jobs in the public
sector. So, in addition
to the switch from the
private
to public sector jobs,
the public sector also
grew internally.
2
First, Sweden lifted rest riction s on borrowing in the 1982–8 5
period. Until then, the governmen t had tried to control the quantity
and price of credit, favoring major export- ing companie s over smaller
enterp rises (such as those in property developmen t and real estat e
man agement ) and households . As might be expected , a gray 3. In a “gray” credit
3 market,
market develope d to circum vent these measures (Tranøy, 1999). noninstitutional lenders
Over time, the Riksbank (and the adminis- tration ) con clude d that and borrowers interact
directly, using banks as
these interest rate and credit ceiling s were unsustain able and brokers.
rest ricte d growth ; consequent ly, the credit market was fully
libe ralize d in 1985. Not sur- prising ly, the formerly credit-const rained
Swedish companie s and household s ran to their lenders, and a credi t
boom followed.After the capital accoun t was opened in late 1980s,
these companie s gained further freedom to borrow in inte rnationa l
markets.
Second, a series of devaluations in 1981 and 1982 had tempo rarily
revitalize d Sweden ’s expo rt indust ries. However, inflation
remaine d out of control despit e many ill-advised attempts to
restrain wage inflation through labor negotiation s and price and rent
freezes. When thos e attempts failed, the governmen t gradual ly
lifted restriction s on foreign capi- tal flows bet ween 1986 and 1989
as a commitmen t to a nonaccomodatio n polic y, thereby abdicatin g
its disc retio n over domesti c moneta ry policy to more-credible foreign
central banks.The krona was stil l anchored to a basket of currencies ,
primarily the German mark, based mostly on their trade weights .
However, the weigh t of the U.S. dollar in the basket was double what
it would have been if it had been based solely on the size of its trade
with Sweden .With the decline in the dolla r relative to the mark
during the 1980s, this weight gave an advantage on Swedish exports
to Germany (Bosworth and Lawrence , 1987b).
Third, the governmen t made a commitmen t to stop borrowing in
inte rnationa l mar- ket s and borrow only domestical ly in kronas.
Foreign borrowing had been seen as a way for the governmen t to
keep real krona borrowing rates low by staying out of the domes- tic
mon ey market and thus avoiding the crowdin g out of private firms
(Bosworth and Lawrence , 1987b) . An unintende d consequenc e of
this polic y, however, was that banks borrowed in foreign markets
and lent to the governmen t in kronas . In essence , this policy
transferred foreig n exchan ge risk from the government to the banks.
Unfortunate ly, these financial reforms were not enough to 3
Borrowing cheap abroad and lending at higher rates
domestical ly helped banks finance a consume r lending sur ge.Also,
foreign bank entry was allowed in 1986, one year afte r interest rate
liberalization , so locals may have wante d to move first in a
growing credi t market. It is possible that in a new financial
environment , bankers did not fully un- derstan d the risks they were
taking. A statemen t from the then-Riksbank- Vice-President Lars
Heikenste n (head of the Riksbank from 2003 to 2006) supports this
view:
220 160
Home prices
200 140
180 120
Price/rent (right axis)
1.5 1.2
1.3 1.0
0.9
100 0.8
1981 1983 1985 1987 1989 1991 1993 1995
4
ready-high domesti c rates even highe r.The currenc y basket ’s tilt in
favor of the U.S. dol- lar, which formerly had given Swedish exporters
a competiti ve edge, came back to haunt them whe n the krona
becam e overvalued . An attempt to fix the count ry’s fiscal imbal-
ances deal t a final blow to its weakene d economy:A change in tax
policy in the late 1980s abolished the tax-deducti ble status of interest
paid on consume r debt, effecti vely increas- ing households ’ interest
burden.
It is hard to maintain speculati ve asset price growth when the
credit feeding it be- come s costlie r. From 1990 to 1995, residentia l
real estate prices dropped 25 percent and comme rcial real estate
5. Once the fixed
prices dropped 42 percent (inflation-adjuste d rates). Banks’
exchange rate was
nonper forming loans mushroomed from 0.2 percent–0. 5 percent of abandoned, the
krona immediately
total loans in the dropped by
11.5 percent against
1980s to 5 percent in 1992. Then came George Soros’s infamous
the reference basket
1992 attack on the European exchange-rate mechanis m (ERM).After of currencies. A year
6. later, the decline
the krona was floated, domesti c banks’ losses increase d exceeded 30 percent.
significant ly.5 Nonper forming loans hit 11 percent of GDP in I borrow this
1993 .The institution s that suffered most (especial ly among the large description of events
from Calomiris,
banks) were those funded by foreign funds. Klingebiel, and
Laeven (2004) and
The first victims of the economi c downtu rn were two of Sweden ’s Bergstrom, Englund,
and Thorell (2003).
six largest banks, Första Sparbanke n and Nordbanken .6 After they
announce d in the fall of 1991 that they coul d no longer meet the
FIGURE 2 DEBT LOAD OF SWEDISH HOUSEHOLDS
Percent Percent
1.4 14
Debt/income
12
1.3
10
Interest expenditure
1.2 to disposable income 8
(right axis)
1.1 6
4
1
2
Real after-tax lending rate
0.9 (right axis) 0
–2
0.8
–4
0.7
–6
0.6 –8
1980 1984 1988 1992 1996
5
POLICY DISCUSSION PAPERS NUMBER 21, JUNE 2007
After acqui ring Nordbanken , the state owned 22 percent of all was the timing
banking-syste m assets. As a resolution strate gy, it electe d to divide of col- late ral
each bank into two separate entities , one with its good assets, the liquidations .
other with its bad ones.The entitie s holding the good assets Swedish banks
continued to ope rate under their old names and were later merged were required
under the name Nordbanken. The bad asset s were transferred to to dispose of
two asset man agemen t companie s (AMCs): Securum for any collate ral
Nordbanken ’s assets and Retrieva for Gota’s.7 within three
years. 8 Because
The the AMCs’
Rec overy
portfolio was so
In creating the AMCs, the legislatu re gave them a high degree of large,
independenc e from political and regulato ry const raints . The state policymake rs
never explicit ly assumed the role of an acti ve owner ; however, worried that
the chairman of Secu rum’s board resigne d when he could not selling all the
agree wit h the Ministry of Finance about compensatin g the AMC’s problem assets
leading staff, so the state seems to have had some say, at least on within a short
compensation- relate d issues. Still, the AMCs, which were adequate ly period would
overcapitalize d in relation to the expecte d costs and losses dur- ing put too much
their expecte d life-time, held their own purse strings. Securum alone pressu re on
was capitalized with SEK24 billion, an amount equal to Sweden ’s market prices.To
defense budget. This delibe rate over- capitalizatio n enabled the ease thos e
AMCs to carry out their salvage operation s autonomous ly; they did conce rns, the
not have to reques t fundin g from the legislatu re, which might have AMCs
tried to influence their decisions. emphasize d
The AMCs were also free of many regulation s and activity from the very
rest riction s that applied to banks , even though managing the loan beginnin g that
portfolio, collectin g payments , and possessin g col- late ral in case of cleanin g up the
bank ruptc y might be conside red banking activities .The AMCs bad assets
needed broad leeway becaus e they had to be more than woul d take a
banks .They injecte d equity into troubled borrowers to maintain and long time, more
restore their values as well as revive the real econo my. In many than a decade if
instances , this meant that they had to take over defaultin g necessa ry
companies .They became responsible for runnin g those companies , (Bergstrom,
much as a private owner would (including hiring and firing Englund, and
man agement , rehabilitatin g neglecte d property, acqui ring Thorell, 2003).
supplemen- tary property, changin g the focus of companies ’ In the end,the
ope rations , and so on) until liquidation. resolution
A second area in which they were exemp t from bank regulatio n occu rred much
more quickly
than initial ly anticipated ,thanks to the rapid growth of the Swedish
econo my, which parallele d the global economi c boom of the 1990s .
Liquidation s were complete d by 1997 at a smaller cost to the
taxp ayers than
was anticipated . Securum returned to the state nearly SEK14 billion
($1.8 billion in 1997
7. Part of the reason for
rehabilitat- ing an
existing insolvent bank
by stripping away its
bad assets stems is a
desire to protect (or
at least salvage) the
value of the bank’s
relationships,
particularly its lending
relationships (see
Diamond, 2001).
6
dollars) of its SEK24 billion ($4.5 billion in 1997 dollars) initial founded in 1989
capital—admitted ly, in dep reciate d kronas.9 That being said, the to handle the U.S.
savings and loan
cost of the crisis to Sweden was not limited to the capital spent by crisis.With no
the AMCs.There have been signif icant income and output losses as- internal source of
funding , it relied
sociate d with the crisis. In the early 1970s, Sweden had one of the on
highest income levels in Europe ; today, its lead has all but Congressional
disappea red. Cerra and Saxena (2005 ) found that the crisis caused appropriations,
which made long-
a permanen t decline in output that can explai n the enti re fall in term plannin g
Sweden ’s relative income . So, even well-man aged financial crises difficult. Between
March
don’t really have happy endings.
31, 1992, and
Lessons of 20/20 Decembe r 17,
Hindsight 1993, the RTC had
How much of Sweden ’s success in crisis resolution can be credited to reduce its
to its wise policies and how much to a strong global economic crisis-resolutio n
expansion in the years after the crisis? Did sensi ble policies pay off, activity for lack of
or did the rising tide lift all boats? I suspect it was a bit of both. As funds (Spade and
evidence for the rising-tide hypothesis, foreign demand for Thomas, 1998;
Swedish goods and ser- vices (measu red by exports) rose from 0.89 also see Kane,
percent of GDP in 1990 to 1.2 percent of GDP in 1995. Unfortunate ly, 1990).
there is no evidence that quantifies each factor’s contribu- tion. Still, The third trait of
Ergungor and Thomson (2006 ) point out a number of traits that is the
four: First, the process must be transpa rent ; expecte d losses of ma rket
the process. Second, crisis resolutio n is best handled by a pline .Withou t it,
political ly an d financiall y inde- penden t agen cy. An the stage is set for
independen t structu re is desirable becaus e it shields decisio n mak- future crises. For
ers from politica l pressu res, which mount as banks are closed and example ,
borrowers are pushed into liquidation .The decisio n to close a bank extendin g blanket
or a business must be an economic—no t a political—one. Financial guaran- tee s
independenc e is necessa ry to give credibilit y to politica l inde- during a crisis
pendence : If a governmen t agency holds the purse strings, it can weaken s the
dictate polic y. Financial independence is also impo rtant becaus e it market disciplin e
allows rapid respons e when funding needs emerge sudden ly (as exerted by
when new losses are disc overed in a financial institution ) and wait- uninsu red
ing for the legislatu re to app ropriate funds is imp ractical. 10 creditors in the
A case in point is the Resolution Trust Corpo ration (RTC), which was postc risis period.
Uninsured deposito rs and nondeposi t creditors have strong
incenti ves
9. Constant dollars are
calculated using the
GDP deflator.
10. As a
comparison, in its
financial crisis,
Finland decided
on a fixed and
limited amount of
funds for bank
resolution, and
when this amount
proved
inadequate later,
a new
Parliamentary
decision had to be
requested and
taken at a
significant cost to
the credibility of
the resolution
process.
7
to monitor and disciplin e financial institution s by raising their cost of
funding when their risk increases.At some point, as the probabilit y of
default increases, uninsu red claimants wil l threate n to liquidat e their
claims. If market disciplin e is to be effective, these inves- tors mus t
be credibly expose d to loss; that is, they must suffer the
consequence s of ignor- ing or failing to detect signs of trouble.An
explici t blanket gua rante e of all the liabilities of problem
institution s in the throes of a crisis reduce s the credibilit y of claims
that de facto gua rantee s will not be extende d in future bank
failu res.
Similar incenti ve problems arise when regulato rs and
policymake rs respond to the crisis by bailing out banks’ creditors
through a policy of capital forbea rance and unlim- ited liquidit y
suppo rt. Although bailouts tend to alleviat e pressures on the
financial sys- tem , forbea rance and unlimite d liquidit y support allow
uninsu red investors to take their money out of the bank, shieldin g
them from loss and reducin g their incenti ve to monitor in the future.
In essence , capital forbea rance and unlimite d liquidit y support
provide an implici t blanke t gua rante e and serve as a taxp ayer-
funde d rescue package for sophisti- cated investors wh o purpose ly
took on risk and were compensate d for bearing it.
The fourth—and final— requi remen t for successfu l crisis resolutio n
is a plan to jump- start credit flows in the financial system by
repairin g th e dama ged credit worthines s of th e rea l
econo my. Even if banks can be complete ly restored to health
through recapital- ization , borrowers may be in no position to repay
any new loan they may get.This was the case in the United States
during the Great Dep ression .The Reconst ructio n Finance
Corpo ratio n (RFC) was founded in 1932 to handle banking crises.
Noticing that loan growth was very slow, in 1934 Congress
authorized the RFC to make loans to business and indust ry.The RFC
invited local banke rs, who would have superior knowled ge about
borrowers, to participat e in the lending . The banks’ respons e was
unenthusiastic . Jesse Jones, head of the RFC unde r Presiden t Franklin
D. Roosevelt, reported (Jones and Angly,
195
1):
governmen t promised to increase its share of the subsidy by one mon ey is used
peso for every three pesos in new loans that the bank made. to save viable
This program incorpo rated some very intelli gent incenti ves. banks and
First, borrowers could get a chunk of their debt erased only if they borrowers, not
the well-
started to repay it. Because a borrower who is not hopeful about his
connecte d ones.
business ’s prospects is unlike ly to throw more money at it, this
Policies tha t
subsi dy mechanis m reveals which borrowers expec t to do well in
prese rve market
the future, solving the adverse-selectio n problem. Second, the
disciplin e are
program gave banks an incenti ve to provide new credit:They could
critical for
reduce the amount they had to charge off as a part of the subsidy
reducin g the
deal by making new loans .Third, the program got more bang for the
like- lihood and
buck by targetin g small enterp rises, the businesse s that depend
severity of
most heavily on banks for funding and are least like ly to have
futu re crises .
government connections .11
Finally, it may
I am not sug gestin g that this strategy would have been more
successfu l in Sweden be necessa ry to
rehabilitate the
than the AMCs’ qui ck-liquidatio n app roach, only that—in hindsight—
real sector
one can see alterna- tives to (or steps that might precede ) taking
because the
possessio n of a troubled compa ny.
financia l
Con clusi sector ’s troubles
on may cause a
We have analyzed the resolution of the financial crisis that hit Sweden credit crunch
during the 1990s in terms of four traits that are essentia l to a that damages
11. Although it
incorporated the right
incentives, the program
was impaired by
Mexican supervisors’
lack of enforcement
authority, an inefficient
bankruptcy system,
and the presence of
politically connected
lending (Calomiris,
Klingebiel, and Laeven,
2004). One would hope
that these problems
would be less prevalent
in a developed
economy.
10
Referen
ces
Bergström , Clas, Peter Englund, and Per Thorell, 2003 .“Secu rum and
the Way out of the Swedish Banking Crisis” (SNS—Center for Business
and Policy Studies, Stockholm) , trans- late d by Timothy Chamberlain.
11
POLICY DISCUSSION PAPERS NUMBER 21, JUNE 2007
Ingves, Stefan, and Göran Lind, 1996. “The Managemen t of the Bank
Crisis—in Retro- spect ,” Sweriges Riksbank, Qua rte rly Revi ew 1,
5–18.
Jones, Jesse H., and Edward Angly, 1951. Fift y Billio n Dolla rs: My
Thi rtee n Years wit h the
RFC (1932–1945 ) (The Macmillan Compa ny,
New York, NY).
12
Return Service
Requested: Please
send corrected
mailing label to the
Federal Reserve Bank
of Cleveland,
Research
Department, P.O. Box
6387, Cleveland, Ohio
44101.