Financial Stability 1/2002
3 Macroeconomic developments and credit risk
3.1 Macroeconomic developments
3.2 Credit growth
3.3 Credit risk associated with lending to households
3.4 Credit risk associated with lending to the enterprise
sector
Boxes:
- Household debt burden by category of household income
- How vulnerable are financial institutions to macroeconomic
changes?
3.1 Macroeconomic developments
The Norwegian mainland economy has grown at a relatively slow pace in recent
years. The sluggish growth rate is attributable to the supply side of the Norwegian
economy. In addition, the effects of the global economic slowdown have spilled
over to the Norwegian economy over the last year. World market prices for important
Norwegian exports have declined. Strong domestic demand pressures have resulted
in higher wage growth than among trading partners, which, in conjunction with
the appreciations of the Norwegian krone, has led to deteriorating competitiveness.
The internationally exposed sector will therefore be faced with considerable
challenges in the period ahead even if the world economy should pick up in line
with expectations.
Gross business fixed investment has exhibited weak growth in recent years.
As a result, growth in credit to the enterprise sector has slowed. Investment
is expected to pick up again as from 2003, which could imply higher growth in
credit to enterprises. However, there are wide sectoral differences.
Real wage growth in the household sector has been solid over several years
and is expected to remain high in the years ahead. Household expectations regarding
both the Norwegian economy and their own financial situation improved considerably
in the first quarter of this year. Household confidence in their own financial
prospects was particularly strong. Combined with high real income growth, this
may imply that pressures in the housing market and credit growth will remain
high.
3.2 Credit growth
Outstanding credit to the public in mainland Norway (households, non-financial
enterprises and municipalities) (C3) has increased in recent months after slowing
moderately in the third quarter of 2001. The slower rate of growth was ascribable
to a sharp reduction in foreign debt. At the end of 2001, total credit came
to close to 140% of GDP, while total mainland credit reached 167% of mainland
GDP (Chart 1.1).
Chart 1.1
Growth in total outstanding credit from domestic sources (C2) has been gradually
decreasing since November 2000, but is still relatively high. In March 2002,
twelve-month growth was 8.8% (9.5% adjusted for the effects of the state takeover
of hospitals).
Growth in credit to households and enterprises has moved on divergent paths
since the autumn of 2000 (Chart 3.1). Growth in credit to enterprises has slowed
during the period, while household debt has exhibited strong growth over a long
period. In the past few years, household debt has expanded at a rate of more
than 10%, which is markedly higher than in Sweden, Denmark and Finland, while
the rate for the UK reached the same level as in Norway in February 2002 (Chart
3.2).
Chart 3.1 Chart
3.2
The high rate of growth in household debt appears to be related to households’
increasing adaptation of loan levels to rising house values.
3.3 Credit risk associated with lending to households
Continued high growth in household debt
In Norway, growth in household debt has been markedly higher than nominal growth
in disposable income since 1999 (Chart 3.3). This has led to a sharp increase
in the debt burden. At the end of 2001, household debt accounted for about 135%
of disposable income. This level is still lower than that prevailing at the
end of the 1980s.
Chart 3.3
Growth in the debt burden for some income categories has been pronounced throughout
the 1990s. The removal of the credit restrictions in the 1980s and tax reforms
eliminating differences in interest deductibility based on income has led to
divergent financial behaviour across different income categories (see separate
box). The debt burden of high-income households is substantially lower today
than it was 12-15 years earlier. On the other hand, the debt burden of low and
middle-income households has exhibited a steady rise throughout the period.
Lower inflation and nominal interest rates today compared with 12-15 years
ago imply lower nominal interest expenses for households for a given level of
debt burden. This improves households’ debt-servicing capacity. On the
other hand, low inflation implies that the debt burden will remain high over
a longer period than earlier. Although the nominal interest rate level is lower,
the real rate of interest after tax is about as high as it was prior to the
banking crisis. An increase in interest rates implies a further increase in
real interest rates, as monetary policy is oriented towards an inflation target
of 2.5%. This would entail a marked rise in the real cost of credit.
Higher household real disposable income has led to an improvement in debt-servicing
capacity. Higher real disposable income and relatively cheaper necessities (excluding
housing) has led to a sustained decline in the share of spending on necessities.
This has freed up portions of household income for spending on other types of
consumption and investment in financial and real assets. An average household
will thus have more room in the household budget for servicing debt than was
the case 12-15 years ago.
International comparisons show that households in several countries are increasing
their debt at a faster rate than income growth. The debt burden level in Norway
is high compared with other countries. The debt burden of Norwegian households
was higher than the debt burden of Swedish, US and UK households throughout
the 1990s (Chart 3.4). In Japan however, households had a higher debt burden
at times during the period.
Chart 3.4
Varying traditions for home ownership structures may partly explain the differences.
A high proportion of Norwegian house holds are homeowners. The tax treatment
of interest expenses may also play a part.
The value of household real and financial wealth has risen in recent years.
Wealth functions as security both for households in a financial crisis situation
and for banks in that their loans are secured. House prices exhibited a steep
rise in the 1990s, but have increased at a more moderate pace in recent years
(Chart 3.5). House prices were about 7% higher at the end of April 2002 compared
with one year earlier. Higher house prices contribute to an increase in household
real wealth. As the predominant share of household debt is mortgage-backed (81%
at end-2001, up from 76% in 1997), high house prices provide solid collateral
for banks provided that the loan-to-asset value ratios remain well below 100.
Chart 3.5
At the same time, a high level of financial investments and positive valuation
changes have pushed up household gross financial wealth over several years.
However, at the end of last year, financial investments showed a decline (Chart
3.6). Bank deposits normally increase in periods of volatility in securities
markets. However, this effect was not very evident last autumn. This may be
because growth in bank deposits was already at a high level. In addition, households
may have shifted towards a more long-term saving approach to equity investment.
Chart 3.6
Net wealth has varied widely across the different household income categories.
A small share of high-income households have increased their share of financial
assets over time. Households in decile 10, i.e. the 10% of the households with
the highest income, had financial assets corresponding to just under 150% of
their debt in 1999 (Chart 3.7). The remaining households have shown falling
gross financial assets relative to debt throughout the 1990s. Against this background,
the financial asset position of most households has not improved over the past
10-15 years and cannot be said to provide enhanced collateral.
Chart 3.7
A high household saving ratio will function as a buffer because in a situation
with a loss of income or higher interest expenses households can choose to reduce
saving to maintain consumption. The saving ratio for high-income households
is higher than for low and middle-income households. The total household saving
ratio has increased over several years in Norway. The saving ratio was 8% in
2000. As a result of lower growth in real disposable income, combined with sustained
brisk consumption growth, the saving ratio edged down in 2001 according to preliminary
estimates (Chart 3.8). The decline was fully offset by a fall in financial investments.
Investments in real assets, which essentially comprise investments in new dwellings,
showed a moderate increase in 2001.
Chart 3.8
Housing starts exhibited a marked rise in the beginning of 2001, but levelled
off towards the end of the year. In Oslo, housing starts were 82% higher in
2001 than in the previous year. Rogaland, Vestfold and Nord-Trøndelag
showed the steepest decline at 8%, 10% and 30%, respectively. In spite of the
rise in housing starts over the last year, the continued sharp rise in house
prices indicates that housing demand remains strong, especially in city centres
where the population density is highest. It is therefore a reasonable assumption
that households will maintain the level of real investments, in the form of
housing investment, in the coming years. Political signals concerning wider
differences in the tax treatment of real and financial capital reinforces this
picture.
Household debt burden by category of household income
Growth in household debt has varied widely across the different household
income categories over the last 10-15 years. The proportion of households
with a heavy debt burden is markedly higher today than in the 1980s.
Statistics Norway’s income and wealth survey provides an indication
of the trend in household debt by household income category. Households are
divided into ten equally large groups (deciles) based on net household income.
Decile 10 consists of the 10% of households with the highest income (over
NOK 490 000 after tax), decile 9 consists of the next 10%, etc. The figures
as from 2000 are projections where debt and income in all the deciles are
assumed to grow in pace with overall household debt growth up to and including
2001, and subsequently in pace with Norges Bank’s projections for the
total household sector up to 2004. The figures from Statistics Norway’s
income and wealth survey deviate somewhat from Norges Bank’s financial
accounts because of the difference in sources used and data collection methods.
As a result, the levels of average household debt burden cannot be compared
directly, although developments over time are by and large convergent.
At the end of the 1980s, the debt/income ratio of households with the highest
income, in particular those in decile 10, was very high, while households
with lower income had a low ratio (Chart 1).
Chart 1
The wide differences in debt burden across income categories in the 1980s
can to a large extent be attributed to the tax system. Full deductibility
of debt interest, in conjunction with higher marginal taxes for households
with higher income, meant the higher the income the lower the real rate of
interest after tax.
The high-income categories adapted to this tax regime by taking up large
loans relative to income.
The removal of the credit restrictions in 1985 and a low after-tax interest
rate led to a sharp increase in credit demand. Credit growth was particularly
pronounced up to and including 1986. Credit growth was more or less equally
high for all income categories, which may have been an indication of pent-up
demand for credit across the different categories.
As from 1990, the overall household debt burden fell sharply, with a particularly
pronounced decline for the high-income categories, especially for decile 10,
but also for households in deciles 7-9. The debt burden of low and middle-income
households (deciles 1-6) remained either unchanged or exhibited higher growth
during the same period.
The tax reforms at the end of the 1980s and in 1992 were probably one of
the main reasons behind the sharp decline in the debt burden of high-income
households, and particularly for households in decile 10. The first reform
reduced the differences in interest deductibility based on income, and the
tax reform of 1992 eliminated the remaining differences.
The total household debt burden is lower today than it was at the end of
the 1980s, but this is primarily because the debt burden of households in
decile 10 is markedly lower. Given that the rise in the total debt burden
from 1999 to end-2001 was equally distributed across the different income
categories, the debt burden of households in decile 7-9 is now at about the
same level as the peak recorded at the end of the 1980s. The debt burden of
low and middle-income households (deciles 1-6) is a little more than 30% higher
today than at the end of the 1980s. This may imply that the credit risk for
large portions of bank loans to households is higher than that implied on
the basis of the average figures for the household sector as a whole.
Developments ahead
Our projections are based on the assumption that debt growth slows to the rate
of growth in disposable income in the period to 2004. This will increase the
debt burden somewhat in the period ahead, before it levels off towards the end
of 2004. The interest rate is assumed to remain unchanged during the projection
period. As a result of a higher debt burden, the interest burden (interest expenses
as a percentage of disposable income before deductions for interest expenses)
will increase somewhat in the period to 2004. The effects on the interest burden
of persistently high credit growth of 12% to the end of the projection period
will be moderate. The increase in vulnerability comes to light when we look
at the effects of an increased interest rate in the alternative scenario (Charts
3.9 and 3.10). According to this alternative scenario, the interest burden will
be 9.4%, i.e. 2.1 percentage points higher than in the baseline scenario.
Chart 3.9 Chart
3.10
The household sector’s financial position is so solid that the interest
burden does not constitute any imminent threat to debt-servicing capacity. This
is primarily because nominal interest rates are currently lower than they were
10-15 years ago. The increase in the debt burden over the past few years has,
however, led to greater financial vulnerability among large groups of households.
In the event of a rise in interest rates, the interest burden will increase
markedly. Households’ robustness in the face of a rise in unemployment
or an increase in interest rates has weakened, particularly for low and middle-income
households. On the whole, the credit risk associated with loans to households
is still considered to be relatively low. If the high growth of credit to households
continues, it could lead in due course to the development of imbalances in the
household sector, making it vulnerable to fluctuations in house prices.
3.4 Credit risk associated with loans to the enterprise
sector
Somewhat slower debt growth in the enterprise sector
Debt in non-financial enterprises increased substantially as from the mid-1990s
(Chart 3.11). A strong increase in fixed investment in the period to 1998 led
to a strong need for financing. Low investment growth in recent years has led
to slower debt growth. Growth in lending from private banks, which accounts
for ¾ of domestic credit to enterprises, has shown a comparable decline.
Chart 3.11
Debt growth rates vary widely across industries (Chart 3.12). Growth in credit
to internationally exposed enterprises is relatively low. The high level of
cost inflation over several years has weakened competitiveness, and an unfavourable
global economic environment appears to have had a dampening effect on activity
and earnings in this sector. This has contributed to slower debt growth. Loans
to enterprises in the manufacturing and mining/quarrying industries account
for only 12% of total bank loans to enterprises. Banks’ potential losses
on loans to these enterprises are therefore limited.
Chart 3.12
Banks’ highest potential losses are in the sheltered sector. Services
and property management enterprises account for about 40% of bank loans to enterprises.
Enterprises in these two industries and in the retail sector continue to record
high debt growth rates. Previous experience shows that changes may occur rapidly,
particularly in the property industry. After a long period of rising rent for
commercial property, the market has now stabilised. In some areas, prices have
declined. At the beginning of 2002, 425 000 square metres of commercial premises
were registered as vacant in Oslo, Akershus and Bærum, i.e. an increase
of 28% on the previous year, representing a vacancy rate of a little more than
5% of the stock of commercial buildings in this area. The number of buildings
that will be completed in 2002 is estimated to increase by 50% from the level
in 2001 to approximately 4% of total stock. This is substantially higher than
the levels recorded over the last three years. The increase in vacant premises
may push down rent and property prices over time. For large cities outside of
Oslo, rent for commercial premises has risen steadily since the trough-year
1993.
Weaker profitability, but the financial position of enterprises remains satisfactory
Debt growth in the enterprise sector is still higher than growth in value added.
This tendency was amplified through 2001 as result of weak operating profits.
According to preliminary data, operating profits for mainland Norway increased
by 0.7% between 2000 and 2001, compared with close to 2% in the previous year
(Chart 3.13). Preliminary accounts figures for listed companies show that overall
operating profits fell from 13% between 2000 and 2001. Companies in the IT,
media and retail industries recorded the steepest decline in profits, while
companies in the banking, health, shipping and telecommunications industries
posted a marked improvement.
Chart 3.13
Equity ratios are still at a high level in spite of weaker profits in 2001.
Most of the largest listed companies’ book capital came to about 30% or
more of total capital at the end of 2001, which indicates that they are still
in a solid financial position. However, enterprises are dependent on sufficient
earnings ahead in order to service their debt.
Several large enterprises declare bankruptcy
The number of bankruptcies among large enterprises rose markedly in 2001 compared
with the previous year. Total turnover among the bankrupt enterprises (measured
in terms of the turnover in the year preceding bankruptcy) increased by 26%
to a little more than NOK 10bn during the period. The number of employees affected
by the bankruptcies also showed a considerable increase.
Estimates from KMV Corporation based on stock prices and accounts data show
an increase in the bankruptcy probability for listed companies in the latter
half of 2001 and the beginning of 2002. The same applies to large private companies
(Chart 3.14). In March, bankruptcy probabilities for both categories declined.
Chart 3.14
Gross non-performing loans have edged up in recent months (Chart 3.15). The
magnitude of non-performing loans is low from a historical perspective, however.
At end-2001, non-performing loans accounted for 1.8% of total loan debt in the
enterprise sector.
Chart 3.15
Developments ahead
The underlying scenario for the real economy implies that corporate earnings
will remain relatively solid in the years ahead. The outlook appears to be brighter
for the sheltered sector than for the internationally exposed sector. During
the projection period, debt growth is assumed to slow to the rate of growth
in nominal GDP for the mainland economy. This is a somewhat lower rate than
the rate of debt growth over the last 10 years. Relatively low growth in mainland
business investment in the years ahead would also suggest a moderation in debt
growth in the enterprise sector. However, there is some uncertainty associated
with future business investment. Higher-than-assumed growth in investment may
result in higher credit growth. We assume that debt growth in internationally
exposed industries will continue to be low, while debt growth in sheltered industries
will remain high.
The estimates imply that the debt burden of enterprises will remain high in
the years ahead (Chart 3.16). Given unchanged money market rates, the interest
burden is expected to remain stable (Chart 3.17). Should debt growth turn out
to be 12% and the interest rate 2 percentage points higher, the debt burden
will be markedly higher than at any time since the 1980s. This scenario implies
that interest expenses would account for more than half of enterprises’
cash surplus in 2004.
Chart 3.16 Chart
3.17
Since the last report, corporate earnings have deteriorated slightly, but debt
has expanded at a slower pace. Moreover, equity ratios in the enterprise sector
remain solid. In addition, the uncertainty associated with developments in the
global and domestic economy has abated, and growth in domestic demand is expected
to be stronger than previously assumed. Since the last report, uncertainty in
stock markets has subsided and share prices have advanced. On balance, the credit
risk associated with loans to enterprises is now assessed as moderate, and somewhat
less than six months ago. However, a continued high debt burden implies that
enterprises will be vulnerable to pronounced changes in operating parameters
(see separate box).
How vulnerable are financial institutions to macroeconomic
changes?
Stress tests can show how vulnerable financial institutions are to macroeconomic
changes. A stress test analyses how much can be lost, not necessarily the
probability of how much will be lost. Stress tests are of growing importance
for financial institutions, and are also becoming an important tool for central
banks in their surveillance of financial stability.1)
1) See for example IMF (2001). “Finland:
Financial System Stability Assessment”, IMF Country Report No.
01/214, November 2001.
The macroeconomic projections in the March 2002 Inflation Report are used
as the baseline scenario. Norges Bank’s macroeconomic model RIMINI is
used to analyse the effects of two different changes in macroeconomic developments.2)
In the first alternative scenario, house prices gradually fall by about 25%
in relation to the baseline scenario up to 2004. In the second alternative
scenario, wage growth is 2 percentage points higher than in the baseline scenario
in 2002, while the interest rate is set 2 percentage points higher than in
the baseline scenario as from 2002.
2) A more in-depth analysis will appear
in Economic Bulletin 3/2002.
Losses on loans to households
Norges Bank has estimated an econometric model for financial institutions’
losses on loans to households using real economic variables as explanatory
factors. In the model, a higher household debt burden, rising unemployment
and increased bank lending rates result in higher losses. Lower real housing
wealth will also increase losses.
Chart 1 shows developments in losses on loans to households in the two alternative
scenarios. In the scenario with a fall in house prices, the effect will be
a substantial reduction in household consumption and housing wealth. Lower
demand leads to rising unemployment. In 2004, losses will increase by 0.1
percentage point or just under NOK 1bn (2001 prices), compared with the baseline
scenario. One reason why the losses are not higher than indicated above is
that households generally have solid collateral for their loans.
Chart 1
In the scenario with higher wage growth and higher interest rates, increased
wages will in isolation result in higher household disposable income, while
a higher interest rate will curb domestic demand. On balance, unemployment
will edge up. According to the model, financial institutions’ losses
on loans to households as a percentage of total loan debt in the household
sector will increase by about 0.1 percentage point, or a little less than
NOK 1bn (2001 prices), compared with the baseline scenario in 2004.
Losses on loans to enterprises
In addition, Norges Bank’s bankruptcy prediction model is used to estimate
the effects of the stress tests on enterprises. The model predicts the probability
of bankruptcy as a function of a selection of accounts variables, life spans,
size and industry characteristics.3)
To provide an indication of future developments beyond the predictions based
on historical accounts, the model’s explanatory variables must be projected
for each enterprise. This is has been done by assuming that key income and
cost items in the corporate accounts change in line with estimates for changes
in selected key macroeconomic variables. For example, operating income is
projected using growth in mainland GDP, labour costs using annual wage growth,
etc. Using the projected accounts, the model generates bankruptcy probabilities
and risk-exposed debt.4) Annual changes
in real house prices are used as an indicator of changes in the value of creditors’
collateral. Risk-exposed debt is used together with this indicator to simulate
banks’ loan losses.
3 See Eklund, Trond, Kai Larsen and
Eivind Bernhardsen (2001): “Model for analysing credit risk in the
enterprise sector”, Economic Bulletin 3/2001
4 That is to say the probability of
bankruptcy multiplied by long-term debt and overdraft debt for each enterprise.
Risk-exposed debt shows expected loan losses in the absence of collateral.
Applying these assumptions, the fall in house prices described above will
result in loan losses as a percentage of total loan debt in the enterprise
sector that are 1.1 percentage point, or about NOK 9bn (2001 prices) higher
than in the baseline scenario (Chart 2). The considerable increase is primarily
attributable to the sharp reduction in the value of financial institutions’
collateral and to a lesser extent to higher bankruptcy probabilities. Real
estate accounts for a substantial share of many enterprises’ assets.
In the scenario with higher wage growth and interest rates, loan losses will
be almost 0.2 percentage point (a little less than NOK 1.5bn (2001 prices)
higher than in the baseline scenario in 2004).
Chart 2
Conclusion
The stress tests show that financial institutions’ loan losses will
increase as a result of disturbances in the real economy. In the scenario
with a fall in house prices, financial institutions’ losses on loans
to households and enterprises will be about 0.6 percentage point higher than
in the baseline scenario in 2004, while loan losses will be 0.13 percentage
point higher (as a percentage of total household and enterprise loan debt)
in the scenario with higher wage growth and interest rates. The relatively
moderate losses indicate that the exposure of financial institutions’
loans to households and enterprises to the macroeconomic changes described
is fairly limited. This is partly because most enterprises and households
are robust today and in a position to absorb a deterioration in profitability/disposable
income. Generally solid collateral and moderate debt/collateral value ratios
also make a positive contribution. Although the results largely depend on
the models used and the underlying assumptions, our calculations provide an
indication of the potential degree of exposure of financial institutions to
changes in economic developments.
© 2002 Norges Bank webmaster@norges-bank.no