Journal Issue

Finland: Selected Issues and Analytical Notes

International Monetary Fund
Published Date:
August 2012
  • ShareShare
Show Summary Details

II. Analytical Note 2: Macro-Financial Linkages1

Despite the pronounced output contraction in 2009, the Finnish financial sector has weathered the 2008–09 crisis well. Recent evolution in a fairly comprehensive financial stress index points to an improvement of the financial sector situation in Finland. The financial stress index (FSI) is based on variables related to the banking sector, securities markets and foreign exchange market.2Compared with the stress in the 1991 crisis and the 2001 stock market drop, the recent financial crisis appears to have had minor repercussion for the Finnish financial sector.

1. However, uncertainty remains high as stress in other European countries has started to increase again and Finland tends to lag. Stress in the Finnish banking sector tends to co-move strongly with the stress in the Swedish banking sector. The latter has seen a renewed increase in recent months comparable to the increase in stress in Denmark. This indicates that uncertainty remains high and the recovery fragile.

Financial Stress Index 1/

Source: Cardarelli, R., S. Elekdag, and S. Lall (2009), “Financial Stress, Downturns, and Recoveries,” IMF Working Paper, WP/09/100 and Fund staff calculations.

1/ The financial stress index is a composite of the spread between commercial papers and sovereign bonds, the beta of the banking sector (from a CAPM), the term structure of interest rates, and volatilities in stock returns and the exchange rate. Large values imply higher distress. A value of zero indicates neutral financial conditions.

2. While starting from robust positions, several financial vulnerability indicators have deteriorated in Finland in recent years. Household indebtedness has risen from below 60 percent in the mid-90s to above 100 percent of disposable income in 2010 and is projected to increase further. Although still accounting for less than 1 percent of total loans, non-performing loans have increased from below 500 EUR million in 2007 to 1,250 EUR million in 2009 and have remained broadly unchanged since then. The share of households with debt exceeding 300 percent of disposable income has grown in recent years to 10 percent and accounts now for about 45 percent of total household debt. The debt accumulation has been facilitated by low interest rates and rising housing prices, which have boosted the collateral values.

3. Financial variables can affect the broader economy via multiple channels. Falling house prices and stock market indices worsen the balance sheet position of households and firms and are potential factors limiting consumption and investment growth. The near absence of fixed-rate lending, paired with the increased debt levels, bears an additional risk as debt servicing would become more difficult for the highly indebted households should interest rates increase. To analyze these channels in more detail and quantify the corresponding effects, we make use of three econometric approaches.

4. Various models are employed to assess the vulnerabilities of the economy to shocks transmitted via the financial sector. First, we construct an index that allows us to evaluate the impact of the change in financial conditions on GDP. A VAR analysis is then used to weigh the relevance of various financial sector variables for economic activity. Second, we analyze the potential existence of disequilibrium in the credit market, namely whether there is a buildup of excess demand or supply of credit. Finally, the implications of possible deleveraging and possible renewed contraction in housing prices for overall output are evaluated with the help of another VAR analysis, which links output and credit to financial sector variables.

A. Financial Conditions and their Effect on Output

5. A VAR analysis is used to decompose the contribution of various financial indicators to economic activity. The overall financial condition index (FCI) is the sum of the cumulative impulse responses of real GDP to each of the financial variables. The latter variables include the house price index, the short term interest rate (LIBOR), the stock price index, the banking sector risk (measured by the corresponding beta estimated in a CAPM), and the real effective exchange rate. The value of the overall FCI reflects the overall contribution of financial conditions to GDP. Additionally, the impulse responses are standardized such that a change in the index by one unit can be interpreted as an (annualized) change in GDP growth by 1 percentage point.

6. The evolution of the FCI implies a strong negative impact of financial conditions on GDP in 2009. The FCI’s deteriorating trend from 2007Q2 to 2009Q2 suggests a significant contribution to the cumulative reduction in GDP over the two years due to the deterioration in financial conditions. In 2009Q2 the index stood at -6 down from 4 in 2007Q2. However, the negative impact was short lived and financial conditions returned to a positive contribution to growth in 2010Q1.

Financial Conditions Index, 2000Q1–2011Q4

(Percentage points of y/y real GDP growth)

Sources: Cardarelli, R., S. Elekdag, and S. Lall (2009), “Financial Stress, Downturns, and Recoveries,” IMF Working Paper, WP/09/100 and IMF staff calculations.

7. The negative contribution to growth in 2009 was mainly due to falling housing prices, banking sector stress, and worsening competitiveness. The deterioration in financial conditions was initiated by a hike in interest rates, which was followed by a fall in equity prices. Higher interest rates and simultaneous declines in the prices of financial sector stocks fuelled the stress in the banking sector via lower asset values and increased costs of refinancing.3 While the decline in financial conditions was broad-based across the five contributors, the recovery was exclusively based on very low interest rates, which fuelled a renewed increase in housing prices and supported the recovery in the banking sector. Stock market prices remained below pre-crisis levels and the real effective exchange rate provided no significant support to growth.

B. Credit Market Imbalances

8. Adverse financial conditions can feed into a mismatch of demand and supply in the credit market. However, the policy implications are very different depending on whether the mismatch is driven by the supply side (credit crunch) or the demand side (credit contraction) of credit. In the case of a credit crunch, banks are constrained in their capacity to provide credit either because of liquidity problems or deleveraging. Thus, there is a case for policy to focus on restoring stability in the financial sector, possibly through direct support to financial institutions. In the event of a credit contraction, households’ and firms’ demand for credit is weak. In this case, policy should focus on fostering household and firm demand by improving the economic conditions for households and firms.

9. We estimate a system of equations for the demand for and the supply of credit to the private sector for the period from 2000Q1 to 2011Q3.4 The demand for credit is assumed to depend on economic activity, the lending rate, the stock market, and housing prices. The supply of credit is explained by economic activity, total private deposits (as a measure of available resources), and the lending and money market rates. The difference between the residuals of the supply and the residuals of the demand equation can be interpreted as disequilibrium in the credit market. Excess demand that coincides with a flat or falling volume of credit indicates the presence of a credit crunch.

10. The analysis of demand for and supply of credit is based on two alternative estimation methods: A two stage least square (2SLS) regression, and a maximum likelihood (ML) method. In both estimation methods, we apply a disequilibrium concept5 for the observed quantity (credit). Specifically, an excess demand for credit is expected to increase prices (lending rate) and have a positive effect on the supply of credit and vice versa. The difference between these arises as the ML method is based on a system of equations while the 2SLS method estimates the demand and supply equations separately. We apply the 2SLS estimates as initial values in the ML estimation.

Credit Market Disequilibrium Estimation Results
Two-stage least squareLog-likelihood
Lending rate−0.06***−0.08***
Interest margin0.11**−0.03
Stock market (t-6)0.24***0.020.23***−0.02
Housing prices (t-1)0.77***0.55*
Standard error0.
Significance level: *** 1 percent, ** 5 percent, and * 10 percent.
Significance level: *** 1 percent, ** 5 percent, and * 10 percent.

11. The estimation results suggest that the credit market is broadly in equilibrium, although the latest numbers suggest a move to tighter conditions. The excess supply of credit—which preceded the 2008–09 crisis—came to a halt in 2008Q4 as volatility increased, the interest margin shot up, and the growth of total deposits came to a halt. As the interest rate margin started easing, both demand and supply of credit started declining in tandem, leaving the credit market conditions relatively unchanged in 2009. A mild revival in deposit growth and a further decline in the interest margin led to an easing in the supply of credit. However, the worsening macro-financial outlook and the increase in inflation have contributed to a slight excess demand for credit in the first quarters of 2011.

Finland - Excess Supply/Demand for Credit

(Percentage by which demand exceeds supply)

Source: Fund staff calculations.

12. Survey data on corporate finance suggest that firms remain cautious in their investment plans. The recently published corporate finance survey describes the financing situation of about 1000 firms in 2010. The study indicates that the financial problems that firms faced during the financial crises gradually alleviated in 2010. Credit margins have been lifted somewhat during 2010. This has happened although the demand for credit is subdued as firms are still cautious with their investment plans.

C. Housing Sector Developments, Credit, and Growth

13. House prices have continued their rising trend at a moderate pace. From 2003Q1 to 2007Q3, nominal house prices grew at an average annual rate of 7 percent. Prices dropped a cumulative 5½ percent from 2008Q2 to 2009Q2 but exceeded the 2008Q2 high already two quarters later—in 2009Q4—by 2¾ percent. After this temporary acceleration, nominal house price growth appears to have leveled-off with an average q-o-q growth rate of ¾ percent in the last four quarters of 2011. Growth in house prices was flat in mid-2011.

14. Real house prices and the price-to-rent ratio have reached the pre-1991 crisis peak values. Real house prices are at similar levels as in the boom period in 1990, which was followed by a full-fledged housing and banking crisis. The price-to-rent ratio recovered quickly from the temporary low in 2008Q4 and stands now 16 percent above the temporary low and for the first time close to 5 percent above the 1989Q4 peak value. In comparison with international developments, real house prices in Finland are well in line with the secular trend to higher cost for housing relative to other goods.

15. However, a widely employed affordability measure suggests that the level of housing prices is not excessive. Measured by the price-to-income ratio, the increase in housing prices has exceeded income increases only marginally in recent years. Compared to its lowest level since 1970Q1—in 1993Q3 after the housing bust—the price-to-income ratio has increased by about 27 percent. Since 2000Q1, the price to income ratio has increased by 1½ percent.

16. Lending to households accounts traditionally for the largest share of credit to the private sector. Household—primarily mortgage—lending accounts for about 60 percent of total lending to the private sector and has been increasing in recent years. Banks are thus dependent on returns from the mortgage lending business and developments in the housing sector.

Finland: Housing Price Indicators

(index: 2005=100)

Sources: OECD and Fund staff calculations.

Nominal Housing Price Growth

(y/y percent change)

Sources: OECD and Fund staff calculations.

17. A high sensitivity of output to the housing market indicates that disruptions in the credit or housing market and deleveraging can be important sources of risk. Thus, it is important to assess the strength of the relationship between housing prices, credit, and output. Several VAR models are estimated to capture the transmission of shocks to economic activity.

Finland: House Prices and Output Developments


Sources: National authorities and Fund staff calculations

Finland: Credit to Households and Output Developments


Sources: National authorities, OECD, and Fund staff calculations

18. The VAR analysis points to a robust relationship between credit and output growth. We use quarterly data for the period from 2000Q1 to 2011Q3 to estimate four VAR models. The basic model (1) includes real GDP, real credit to the private sector, and housing prices. The framework is extended in model (2) to include also the interest rate as a relevant transmission channel. Model (3) controls additionally for the overall price level. Finally, model (4) contains, with the return on financial sector equity, a measure of banking sector health to assess the feedback loops between banks, households, and the overall economy.

19. A deleveraging implied by a reduction of credit by 5 percent could lower output by 1 to 2 percent within two years. A negative 5 percent shock to credit causes housing prices to fall initially by 1 percent and by 2½ to 3 percent after 2 years. The lowered credit availability and collateral value reduce output by 1 to 2 percent within the horizon of 2 years.

20. A shock to housing prices could affect output markedly in the short-run. Results suggest that a negative 5 percent shock to housing prices—comparable to the house price drop of the recent crisis in 2009—causes a contraction of credit by 2½ percent, and a fall in output by 3½ percent as a consequence, within one year for model 3.6 However, after 2 years, output exceeds initial output by 1½ percent as housing prices recover and credit stops declining. According to model 4, output recovers less quickly as housing prices and credit remain depressed for a longer period. Banking sector returns decline substantially in the short run, contributing to the decline in credit provision.

Impulse Responses from the VAR Analysis
Impact of a maximum drop in housing prices by 5 percent on:
GDPHousing pricesCreditBanks
− after 2 quarters−2.6−3.1−3.0−2.9−4.7−5.0−5.0−−0.1−2.1−16.6
− after 1 year−4.7−4.7−3.6−3.3−4.4−3.3−2.5−2.5−2.0−2.2−2.4−5.2−4.3
− after 2 years−2.1−0.61.4−0.3−−1.1−4.6−3.6−2.7−4.2−0.1
Impact of a maximum drop in credit by 5%
GDPHousing pricesCreditBanks
− after 2 quarters−0.1−0.1−0.3−0.1−1.1−0.9−0.8−0.5−4.9−4.9−4.8−2.63.3
− after 1 year−0.5−0.2−0.5−0.1−2.4−1.4−1.2−0.3−3.6−3.5−3.9−2.72.1
− after 2 years−1.4−1.1−1.7−1.5−2.7−2.6−2.5−2.2−3.4−2.9−2.9−2.0−12.3

D. Conclusion

21. The Finnish financial sector has weathered well the financial crisis. While indices of bank stocks and housing prices declined, and output dropped by a record close to 8½ percent, growth of credit to the private sector remained positive, non-performing loans remained well below international standards, and capital buffers of Finnish banks remained robust. There is a significant co-movement between financial variables and the real sector. While this has contributed to negative feedback loops between equity prices, output, and credit, the fall in interest rates has buffered the adverse consequences.

22. While fundamentals remain strong, risks have started to rise. Household debt has been rising continuously in the last decade, the share of highly indebted households has more than doubled, non-performing loans have hardly declined since they peaked in 2010, and real housing prices are at historic highs. These developments have made the private sector more exposed to a potential price shock in the housing market and—in conjunction with the high proliferation of variable rate loans—to a rise in the interest rate level.

E. References

    BalakrishnanRaviSelimElekdagStephanDanninger and IrinaTytell2009The Transmission of Financial Stress from Advanced to Emerging EconomiesIMF Working Paper 09/133Washington: International Monetary Fund.

    CardarelliRobertoSelimElekdag and SubirLall2009Financial Stress, Downturns, and RecoveriesIMF Working Paper 09/100Washington: International Monetary Fund.

    MaddalaG. S. and Forrest D.Nelson1974Maximum Likelihood Methods for Models of Markets in DisequilibriumEconometrica Vol. 42 No. 6 pp. 10131030.

    PazarbasiogluCeyla1997A Credit Crunch? Finland in the Aftermath of the Banking CrisisIMF Staff Papers No. 44 pp. 315327 (Washington: International Monetary Fund).

    RocaPaulo2010Macro-Financial LinkagesinFinland: 2010 Article IV Consultation IMF Country Report No. 10/273 (Washington: International Monetary Fund).

Prepared by Sebastian Weber and Mika Kortelainen.

The financial stress index (FSI) is a composite of the spread between commercial papers and sovereign bonds, the beta of the banking sector (from a CAPM), the term structure of interest rates, and volatilities in stock returns and the exchange rate. Large values imply higher distress. A value of zero indicates neutral financial conditions. See Cardarelli et al. (2009) and Balakrishnan et al. (2009).

To identify the relevance of the different factors in the VAR, a Choleski decomposition of the variance-covariance matrix is required. The Choleski decomposition is obtained by ordering output first, followed by the price level, the house price index, the real exchange rate, the banking sector risk, the interbank rate, and the stock market index. The conclusions are robust to changes in the ordering.

The model is described in Maddala and Nelson (1974) and a different variant has been applied to Finnish data earlier by Pazarbasioglu (1997). The credit data refers to the loans to the private sector.

In equilibrium, the model would be as follows:

where both quantity and price are the same in both the demand and supply equations. Instead of this, we estimate the following disequilibrium model with both 2SLS and ML methods:

Excess demand has an effect on the price changes in the disequilibrium model.

Model 1 implies a larger impact since the housing price response and the credit response are more persistent in this model. This suggests that it is necessary to control for the overall inflation and other variables.

Other Resources Citing This Publication