Denmark: Selected Issues

Selected Issues

Abstract

Selected Issues

Household Balance Sheet Structure in Denmark and Sensitivity to Rising Rates1

Households in Denmark have gotten considerably wealthier in recent decades. High household assets, in particular in the mandatory pension system and housing, provide stability by funding future consumption and protecting against shocks. The high, but mostly illiquid, assets have a counterpart, however, in the high household debt, as households often need to borrow to consume or buy property. The resulting combination of large assets and liabilities on household balance sheets make the Danish economy sensitive to interest-rate changes. Sudden increases in interest rates can create macroeconomic instability via their impact on the debt service of households and knock-on effects on consumption. Analysis of Danish microdata on household balance sheets shows a modest impact on consumption overall from a small rise in interest rates, but vulnerabilities are considerably larger for at-risk groups, such as households with high debt and adjustable-rate mortgages.

A. The Assets and Wealth of Households in Denmark

1. Danish households have amassed considerable wealth in past decades, and fare favorably among advanced economies. Having recovered from a temporary dip during the global financial crisis (when households had to use accumulated buffers to support consumption in the face of declining incomes), the net wealth of Danish households reaches close to 600 percent of gross disposable income (or about 300 percent of GDP) as of 2015, well above many other advanced economies (Figures 1 and 2).2 This is partly the result of sound long-term macroeconomic policies aiming at promoting greater pension coverage while increases in the prices of houses and other assets also play a role.

Figure 1.
Figure 1.

Household Net Wealth in Denmark and Elsewhere1

(Percent of gross disposable income)

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A001

Source: OECD1. Countries comprise Austria, Belgium, Czech Republic, Estonia, Finland, France, Germany, Hungary, Italy, Korea, Latvia, Netherlands, Norway, Poland, Portugal, Slovakia, Slovenia, Sweden, UK, and US.
Figure 2.
Figure 2.

Danish Household Assets and Net Wealth

(Percent of gross disposable income)

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A001

Source: OECDNote: Green bars are financial assets, and gray bars are nonfinancial assets. Net wealth is assets minus liabilities (not shown).

2. Danish household assets are concentrated in pensions and nonfinancial assets, in particular real estate. As collective pension schemes took shape (Box 1), the pension and insurance savings of Danish households have been increasing steadily in recent decades. They amounted to almost 300 percent of disposable income in 2015 and make up half of households’ financial assets and over a third of their total assets (Figure 2). Real estate property assets are another substantial component of wealth for households at nearly 200 percent of gross disposable income in 2015 (nearly 100 percent of GDP).

3. While Danish households have built large assets, this contrasts with a low measured household savings rate. Indeed, while household’s accumulated savings and assets in Denmark are large, the gross household savings rate has historically been low compared to peer economies. Box 2 discusses this discrepancy. The measurement issues and possible identification challenges relating to the savings of the nonfinancial corporate sector may be contributing to the discrepancies between the flow and stock indicators. Valuation effects may also play a role.

Figure 3.
Figure 3.

Financial Assets and Financial Net Wealth

(Percent of gross disposable income, latest available annual data)

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A001

Source: OECD.
Figure 4.
Figure 4.

Household Debt

(Percent, latest available annual data)

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A001

Source: OECD.

4. High household assets in Denmark have a counterpart in high household debt. Households in Denmark, and other northern European economies, tend to have a low share of their wealth in liquid instruments, such as deposits and portfolio investments (Figure 3). An unintended consequence of having a large share of illiquid assets is that it leads households to borrow more than they would otherwise to consume or acquire assets (such as property). The result is high household debt and balance sheet leverage (Figure 4). This is consistent with the life cycle hypothesis, which posits that household income and consumption patterns differ over the life cycle. Households raise debt early in their lives to finance consumption, then they accumulate net savings by paying down debt and paying into pensions, and finally they reduce their net wealth by drawing on their pension savings. Isaksen et. al. (2011a and 2011b) estimate that for every 100 kroner increase of pension wealth, household debt rises by 30–40 kroner.

The Pension System in Denmark

Denmark, as many other European countries, has adopted a three-pillar pension system. The first pillar—which is mandatory—is tax-funded and includes the basic state retirement scheme. The second, and biggest, pillar comprises collective occupational pension schemes (also mandatory for most labor-market plans), and the labor market supplementary pension (ATP).1 The third pillar comprises voluntary, private, savings-based pensions, typically managed by private insurance companies and banks. For more information, see Kramp et. al. (2012), and IMF (2014).

Contributions to labor market pensions outpace payouts consistently. During the catch-up phase of savings contributions were on a steady increasing path, climbing up to 6.7 percent of GDP before the financial crisis. Net contributions have been tempered somewhat in recent years due to technical reasons, such as early taxation incentives introduced in 2013, but are expected to resume and remain high for the next few decades (see Autrup et. al., 2015).

Figure 1.1.
Figure 1.1.

Pension Contributions and Payouts

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A001

Sources: ADAM; and Statistics Denmark.Note: Net contributions are adjusted for the early taxation of capital pensions after 2013.

1 There is some debate as to whether the ATP pension fund should be classified under pillar I or II, or be given its own classification, given its hybrid nature. Here we follow the designation of Kramp et. al. (2012).

The Puzzle of High Household Assets Accumulation and the Low Savings Rate

The gross household savings rate in Denmark, as measured in the national accounts, is low compared to other advanced economies (Figure 2.1). Yet Danish households have amassed high pension wealth at almost 150 percent of GDP (or 300 percent of gross disposable income; see Figure 2) as well as large housing and equity wealth.1 There are several potential explanations for the apparent discrepancy both between countries, and between the flow and stock indicators.

Cross-country comparisons of household savings rates are not straightforward. Even when the underlying consumption and saving behaviors of households are similar, various factors may affect the direct comparison of savings rates (see for example Rocher and Stierle, 2015). For instance, differences between tax systems can affect the comparability of household disposable incomes, and hence their savings.

Figure 2.1.
Figure 2.1.

Gross Household Savings Rate

Percent; average 2007–2015

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A001

Source: OECD.Note: Gross household savings rate = gross savings/(gross disposable income + change in net equity of pension fund reserves).

Similarly, the recordation of pension savings and social security contributions in the national accounts can differ from country to country and affect the comparability of disposable incomes and savings. Recent research by Statistics Denmark staff (Osterwald-Lenum, 2017) argues that the actual household savings rate could be as much as eight percentage points higher when adjustments are made for Denmark-specific effects related to pension contributions, taxes, and imputed rents in the calculation of household consumption.

In Denmark, it is also difficult to draw a clear line between household and corporate savings. There is evidence that part of the very high corporate savings (Figure 2.2) in Denmark may effectively reflect household savings (Autrup et al., 2015). Several of the biggest firms in Denmark are owned by households, often via foundations, which are classified as nonfinancial corporations for the purposes of national accounting. In addition, sole proprietorships and owners of small and medium enterprises, tend to invest their savings in their businesses. For tax reasons, owners of firms often opt to retain the firms’ profits as retained earnings, rather than distribute them as dividends. In other cases, firms can buy back their own stock which also has the effect of increasing the value of the firm, and creating capital gains to their owners. The Danish central bank estimated that on a national level the effect of capital gains for households, possibly reflecting such operations, reached over DKK 3 trillion through 2014.

The treatment of pension capital gains and tax incentives can also affect the savings rate. Capital gains on pension wealth are not included in the calculation of disposable income in Denmark, but pension-yield tax is (via current taxes). An additional complication is that Danish households were given the option in 2013 to pay taxes early on their future pension payouts and at a discount, which affected the measurement of net contributions.

Valuation effects may also be at play. The role of valuations can also make the comparison between household financial assets (stocks), and savings rate (flows) difficult.

Figure 2.2.
Figure 2.2.

Savings of Household and Nonfinancial Corporate Sectors in Denmark

Percent of GDP

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A001

Source: Statistics Denmark.

1 Household savings are a flow indicator from the national accounts, while pension wealth is a composite stock indicator constructed from the financial (flow of funds) accounts.

5. From the perspective of pension funds, household debt serves as an important component of their investment portfolio. Household’s pension contributions and insurance premiums are invested in mortgage-related debt. Pension funds and insurance firms held DKK 747 billion of covered mortgage-credit bonds in 2016 (about a quarter of the outstanding stock), equivalent to 13 percent of monetary financial institution (MFI) assets (Figure 5), from DKK 240 billion (ten percent of the outstanding stock) in 2010.

Figure 5.
Figure 5.

Holdings of Mortgage-Covered Bonds Issued by MFIs

(Percent of monetary financial institution (MFI) assets)

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A001

Source: Statistics Denmark.Note: Pension fund and insurers' holdings include indirect holdings through investment funds.

6. The large supply of savings and high demand for household credit is mirrored in the size of Denmark’s financial sector. Banks and mortgage credit institutions utilize the covered bond market to fund their assets. As there are no options to offload housing loans from their balance sheet, their assets have grown substantially to reach 380 percent of GDP in 2015 (Figure 6).

Figure 6.
Figure 6.

Unconsolidated Bank Assets

(Percent of GDP, 2015)

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A001

Sources: Danmarks Nationalbank; OECD; IM F International Financial Statistics; and IM F staff calculations.Note: The figure for Denmark includes banks and mortgage credit institutions.

B. Large Assets and Liabilities: Benefits and Drawbacks

7. On a national level, high household savings reduce fiscal and macroeconomic risks. Pension savings provide stability by funding future consumption, and buffer against shocks. Moreover, large household wealth contributes to fiscal sustainability by reducing the burden on the public sector, and lessen the need to borrow to fund future liabilities. In addition, because households fund future expenditures with current savings, there is less need for higher taxes in the future, which improves business confidence. Unlike in many other countries, on current projections, Danish public finances are robust to aging and it is expected that no major fiscal adjustment will be needed when aging peaks.

8. Investment can boost financial wealth. Borrowing to invest or purchase property can accelerate wealth creation, if the relative return on the debt-funded investment to the cost of borrowing is positive (such as during housing price rises). This is contingent on an eventual reduction of the debt load, which is important for Denmark because of the relatively low amortization rates of housing debt.

9. More types of assets can diversify household balance sheets. Holding various diverse types assets, such as real estate, equity, and pension claims, diversifies the net wealth of households, and reduces volatility of their balance sheets. Relative price stability of real estate properties over time, and their low long-term correlation to financial assets have been shown to reduce volatility of the financial wealth of households (see for example Table 3.1 of Chapter 3 of the Global Financial Stability Report (IMF, 2005)).

10. But the large financial assets and liabilities of households also make the economy more sensitive to interest-rate changes and can pose macroeconomic risks. High household leverage can amplify vulnerabilities in case of shocks. Shocks to household consumption may for instance arise from declines in property prices (which reduce wealth) and/or increases in interest rates (which raise debt servicing costs). During the global financial crisis, house prices in Denmark fell by more than 20 percent in aggregate real terms, and contributed to the significant hit to domestic demand and household consumption (Figure 7). Confidence effects can also become a drag on growth as investment is held back during deleveraging.

Figure 7.
Figure 7.

Household Leverage and Consumption

(Percent)

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A001

Source: Riksbank Financial Stability Report (H1 2015).Note: Consumption is real private consumption per working-age capita.

11. High household debt can also pose risks to the financial sector, although at present banks are sound and have considerable buffers. Danish financial institutions are liquid, well-capitalized, with regulatory capital amounting to nearly 20 percent of risk-weighted assets, and have set aside adequate buffers to withstand reasonable increases in nonperforming loans to the household sector. In recent stress tests, the Danish central bank concluded that risks to the financial system were small from even a severe recession, owing their large excess capital, and the households’ strong repayment capacity due to the generous social safety net (see Danmarks Nationalbank, 2016). Nevertheless, there are concerns about funding risk, particularly for financial institutions funded by short-term mortgage bonds that then finance adjustable-rate mortgages.

C. The Effect of Higher Rates on Household Consumption

12. A sustained rise in interest rates could have a significant impact on consumption. Rising interest rates are expected to have an immediate increase of the interest expense of households, as currently 60 percent of outstanding mortgages in Denmark are adjustable-rate (Figure 8). Mortgage payments could rise rapidly, as some floating-rate mortgages reset every three or six months, which has the potential to chip away a large portion of households’ disposable income and reduce consumption. This is particularly important in the current period of low interest rates that has afforded households the lowest debt-service ratio in the last two decades, and large purchasing power (Figure 9).

Figure 8.
Figure 8.

Stock of Residential Mortgages by Type

(Billions of Danish kroner)

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A001

Source: Danish Financial Supervisory Authority.
Figure 9.
Figure 9.

Household Interest Paid

(Percent of gross disposable income)

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A001

Source: Statistics Denmark.

13. We evaluate these risks with a rich household-level database containing detailed balance sheet information. The microdata gathered by Statistics Denmark include income, debt, and wealth information on individuals residing in Denmark, which are aggregated to the family level akin to Andersen (2016), and they allow for tracking individuals over time. Information on their age and geographic region of residence is also provided.

14. Vulnerabilities are not evenly distributed across households. The median net-debt-to-income level for households in the first 5 income deciles is around zero—a positive attribute. The households with higher debt relative to their income are generally wealthier, with the median debt-to-income ratio of the tenth income-decile households (that is, households with the highest 10 percent of incomes) at 256 percent (Figure 10). Of households in the first 3 deciles of income, that is, households with 30 percent of the lowest incomes, about half have no debt. The tail of households with high debt-to-income (75th to 90th percentiles) generally increases with the higher income cohorts, however there is a notable outlier in the households of the second decile with debt-to-income reaching 430 percent for the most indebted households in this income group. Netting financial assets from debt (Figure 11), reduces this outlier somewhat, but it continues to suggest a pocket of vulnerability. This may reflect proportionally larger borrowing by young households (first-time homebuyers), in anticipation of higher future income. Households with larger debt relative to the value of their home also have higher net debt-to-income (Figure 12). For instance, the median net debt of highly leveraged households (loan-to-value above 80 percent) is over 320 percent of their income, making them particularly vulnerable to house price declines.

Figure 10.
Figure 10.

Household Debt Distribution1

Source: Danmarks Nationalbank using registry data from Statistics Denmark.

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A001

1 Over approximately 2.2 million household observations.
Figure 11.
Figure 11.

Household Net Debt per Income Decile1

(Percent of income after tax, 2015)

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A001

Source: Danmarks Nationalbank using registry data from Statistics Denmark.1 Net debt is debt minus financial assets plus pension savings.
Figure 12.
Figure 12.

Household Net Debt per LTV Group1

(Percent of income after tax, 2014)

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A001

Source: Danmarks Nationalbank using registry data from Statistics Denmark.1 Net debt is debt minus financial assets plus pension savings.

15. Households with adjustable-rate mortgages also have more debt. Within households that have outstanding mortgages, the debt-to-income of households with adjustable-rate mortgages is considerably higher than the ones with fixed-rate mortgages with median levels of 225 and 290 percent respectively, for the most recent year reported (Figure 13). This creates a particular vulnerability to rising rates. In addition, Kuchler (2015) also reports a higher share of interest-only loans held by families with high loan-to-value ratios.

Figure 13.
Figure 13.

Household Net Debt per Mortgage Type1

(Percent of income after tax, 2014)

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A001

Source: Danmarks Nationalbank using registry data from Statistics Denmark.1 Net debt is debt minus financial assets plus pension savings.

Model Formulation

16. The effect of higher rates on household consumption is modeled at the individual household level. Following Andersen (2016), we estimate the annual consumption c for each household for a given year from the equation

c=ys,

where y is the observable annual disposable income, and s are the unobservable savings. The savings variable s is approximated by the annual changes of the values of the household’s holdings of assets (such as deposits, investments, pensions, and housing properties), and liabilities (such as mortgage loans).3 The simulation of the impact of higher rates on consumption is performed on both assets and liabilities, as well as on the disposable income to account for the difference in taxes (on the individual level before being aggregated to the household level). The DKK 50,000 threshold for interest-rate deductibility (DKK 100,000 for couples) is also applied on the household level to calculate the effect on disposable income.

17. The simple assumptions for the repricing of liabilities and interest income are meant to provide a range of possible outcomes for the period of higher rates. Since the individual household mortgage information is available, we calculate the change in mortgage payments, from the increase in the borrowing rate on the debt service as

paymentnewpaymentold=principalARM*Δrnew, ARM+α*principalFRM*Δrnew,FRM,

where principalFRM and principal are the outstanding principal of any adjustable- and fixed-rate mortgages of the household on the given year, respectively; rnew, ARM, rnew, FRM are the new effective interest rates on any adjustable- and fixed-rate mortgages, respectively; and α (α in [0, 1]) is the coefficient to model the effect of the higher borrowing rates on future refinancings into fixed-rate mortgages and new fixed-rate mortgages. The α coefficient models the effect of higher borrowing rates on the stock of fixed-rate mortgages, and can be thought of as a passthrough indicator. Reasonable values for α can vary considerably, depending on the time horizon of the analysis, and the longer the sensitivity period, the higher the passthrough. In what follows, we assume that the increase of rnew, FRM is equal to the increase of the overall borrowing rate (i.e., one-to-one effect), and the increase of rnew, ARM, α, is 0.25 times the increase of the overall borrowing rate, unless is otherwise noted. A 0.25 passthrough is admittedly a low figure for fixed-rate mortgages, and it offers a conservative estimate for the effects of higher rates. Interest income is also modeled via an increase to bank deposit rates.

18. Banks may be slow to raise rates on their deposits following many years at zero interest-rate bound. It is reasonable to assume that interest-rate increases would not be mirrored on deposit rates one-to-one initially, especially after the many years of deposit rates being stuck at the zero lower bound, despite negative policy rates. An exception would be deposits that are linked to loan account which carry the same rate as the loan, but these cannot be identified by the microdata. As a result, the two sets of results presented below are meant to serve as range of estimates for the true impact of rising rates on consumption. In its current formulation, the effect of interest rates is reflected both via the cashflow channel, with the income effect being modeled only through the savings account.

19. The effect of rising interest rates on consumption can be approximated via the reduction of the disposable income. It is straightforward to convert to consumption using one’s estimate for the households’ marginal propensity to consume. For example, a marginal propensity to consume equal to 1 would imply a one-to-one passthrough from disposable income to consumption, which would serve as an upper bound of the consumption response from changes in income. Reasonable estimates for the marginal propensity to consume range from 0.5 to 0.8, but given that the consumption data in the microdata refer to total consumption, which include autonomous consumption (such as fixed spending in utilities, housing expenditures, and more), higher values for the sensitivity of consumption to disposable income may be more realistic. For the sake of presentational simplicity, the results below are presented directly as the impact on consumption assuming 0.8 passthrough of income to consumption. Annual data end in 2014, and thus do not show the effect of the 7 Best Practices or the mortgage diamonds.

Results of Sensitivity Analysis

20. The median sensitivity of household consumption to rising rates is modest. A 100 basis points (bps) increase of the borrowing rate decreases consumption by 0.71 percent for all households when allowing for interest income effects from bank deposits, and assuming 0.25 passthrough to fixed-rate mortgages (Table 1). This implies a hit to consumption that would detract 0.3 percentage points from annual GDP,4 but when rates increase from the current very low rates, it is likely that banks will be slow to raise deposit rates given the long period of negative policy rates. In that case, ignoring interest income effects, the impact of the 100 bps increase of the borrowing rate decreases consumption by 1.03 percent (equivalent to 0.5 percentage points of GDP). When considering the effect of the rate rise only on households with debt, the sensitivity increases to 1 percent (including interest income effects) or 1.26 percent (excluding interest income effects; see Table 1). The mean reduction for all households is 0.62 percent (equivalent to 0.3 percentage points of GDP) factoring in the revaluation of bank deposits, and it increases to 0.98 percent when considering only households with outstanding debt. Similarly, when the effect of bank deposits is excluded, the mean decline of consumption for all households is 0.94 percent (thereby removing 0.4 percentage points of GDP), and 1.24 percent among households with debt. Given the large share of private sector consumption and investment in Denmark’s economic output, sudden and prolonged shocks to interest rates can become macro-critical if they are not accompanied by offsetting increases to income such as via higher wages. In other words, the preconditions for the interest rate increase can mean very different outcomes for growth.

21. The sensitivity of consumption to rate rises increases considerably for groups at risk. The simulation results presented in panels C–F of Figure 14 (which include the interest income channel and are therefore lower estimates for the impact on consumption) suggest that vulnerabilities are modest for most households, but substantial for specific groups. Higher leverage, as measured by loan-to-value (LTV) increases the sensitivity of consumption with respect to rates, with the median drop in consumption greater than 1 percent for households with LTV above 60 percent. While the majority of households has debt-to-income (DTI) below 100 percent, and thus smaller sensitivity of consumption to rising rates, the median consumption decline for households with debt-to-income (DTI) above 200 percent is over 1 percent, with the highly-indebted households (DTI greater than 400 percent). This makes highly-indebted households (accounting for 13 percent of households) very sensitive to rate increases with the median consumption drop over 2 percent. It is also worth noting the long right tail of households with a strong consumption response in the second decile of income in panel D, consistent with Figure 10 above. Households with their oldest family member in prime working age (30–59) display a higher sensitivity to interest rates, as expected, but older households (older than 65 years) have low sensitivity to rising rates, given their higher savings.

D. Conclusions and Policy Implications

22. Rising interest rates can impact consumption, especially of highly-indebted households. While the sensitivity of consumption to rising rates is modest for most households, there are pockets of vulnerability where the impact is substantial. The effect of rising rates is likely to affect liabilities faster than it will benefit households through returns on their savings, given the prolonged period of negative policy rates and banks’ decision to hold a zero-lower bound on deposits. Even including the offsetting benefit of interest income from bank deposits, the hit on consumption from a rise in rates can become substantial for households with large debt stocks, like ones with loan-to-value above 60 percent, or debt-to-income ratios of 300 percent and higher.

Table 1.

Denmark: Decline in Consumption from 100 bps Interest Rate Increase

(Including Interest Income Effects, Unless Otherwise Noted)

article image
Source: Authors' calculations.

23. The authorities have taken measures to keep housing prices and household debt in check. In particular, the authorities introduced a “supervisory diamond” for mortgage credit institutions in 2015, and lending guidance to institutions in areas with the fastest price growth via the Seven Best Practices. More recently, in line with staff advice, the DN and the Danish Financial Supervisory Authority (DFSA) have also begun examining additional macroprudential tools and the Systemic Risk Council (SRC) in April 2017, adopted a recommendation that the government caps variable-rate and interest-only loans in the Copenhagen and Aarhus areas to four times the borrowers' income.

24. Additional measures could contain help debt accumulation and prepare households for rising interest rates. Given the configuration of household balance sheets, comprising large assets and large debt, policies need to manage the risks from asymmetric hits from higher rates and lower housing prices. Specific policies may include:

  • Macroprudential policies: The proposed debt-to-income limit by the SRC is an important measure that can help provide a circuit breaker in case of unsustainable housing price increases and debt accumulation. However, a more general cap could apply to all loans, irrespective of the loan terms, possibly with tighter limits for interest only and variable rate instruments. Raising the recently introduced down payment requirement to at least ten percent, can help shield households from excessive indebtedness. Raising the recently introduced down payment requirement to at least ten percent, can help shield households from excessive indebtedness. Minimum amortization requirements could also be considered. For example, Sweden requires that borrowers make annual payments of at least 1 percent of the principal for mortgages with LTV over 50 percent and 2 percent for those with LTV above 70 percent. To further reduce risks from variable interest rates, the authorities should consider tightening the guidance in the mortgage diamond by further reducing the maximum share of such loans in banks' portfolios.

  • Tax policies: mortgage interest deductibility: The broad housing recovery and current low interest rates provide an opportune moment for reducing the tax deductibility of mortgage interest expenses and for further lowering, beyond what is currently planned, the value of the deduction for interest expenses from income taxes. Homeowners in Denmark are already exempt from capital gains taxes on the sale of their primary residence, and further lowering mortgage interest deductibility—or phasing it out entirely as in Ireland and Spain—can help reduce the debt bias in the tax system.

Figure 14.
Figure 14.

Distribution of the Decline in Consumption from 100 bps Interest Rate Increase

(Including Interest Income Effects, Unless Otherwise Noted)

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A001

The red dashed lines correspond to the level of 1 percent (100 bps) for ease of comparison with the size of the interest-rate increase.
Table 2.

Denmark: Current Mortgage Interest Deductibility from Personal Income Taxes

article image
Sources: National tax and other authorities; Bourassa et al. (2013); Smidova (2016).

References

  • Andersen, A. L., Duus, C., and Jensen, T. L. (2016). “Household debt and spending during the financial crisis: Evidence from Danish micro data,European Economic Review, 89, issue C, 96115.

    • Search Google Scholar
    • Export Citation
  • Autrup, S. L., Kramp, P. L., Pedersen, E. H., and Spange, M. (2015). “Balance of Payments, Net Foreign Assets and Foreign Exchange Reserve,Danmarks Nationalbank Monetary Review, 4th quarter, 3954.

    • Search Google Scholar
    • Export Citation
  • Bourassa, S. C., Haurin, D. R., Hendershott, P. H., and Hoesli, M. (2013). “Mortgage Interest Deductions and Homeownership: An International Survey,Swiss Finance Institute Research Paper, No. 12-06.

    • Search Google Scholar
    • Export Citation
  • Danish Economic Councils (2008). “Danish Economy” (in Danish), Chapter 2, Spring.

  • Danmarks National bank (2016). Financial Stability, 2nd Half, December.

  • IMF (2005). Global Financial Stability Report, Spring (Washington)

  • IMF, (2014). Denmark: Selected Issues, IMF Country Report No 13/332 (Washington)

  • Isaksen, J., Kramp, P. L., Sorensen, L. F., and Sorensen, S. V. (2011a). “Household Balance Sheets and Debt—an International Country Study,Danmarks Nationalbank Monetary Review, 4th quarter, part 1, 4758.

    • Search Google Scholar
    • Export Citation
  • Isaksen, J., Kramp, P. L., Sorensen, L. F., and Sorensen, S. V., (2011b). “Household Balance Sheets and Debt—an International Country Study,Danmarks Nationalbank Monetary Review, 4th quarter, part 2, 3981.

    • Search Google Scholar
    • Export Citation
  • Kramp, P. L., Lohff, J. L., and Maltbaek, J. P. (2012), “Pension Savings,Danmarks Nationalbank Monetary Review, 1st quarter, part 1, 101117.

    • Search Google Scholar
    • Export Citation
  • Kuchler, A. (2015). “Loan types, leverage, and savings behavior of Danish households,Danmarks Nationalbank Working Papers, no. 97, September.

    • Search Google Scholar
    • Export Citation
  • Osterwald-Lenum, M. (2017). “Value-added (at market prices) for a given set of locations, and for a given set of residents. From SNA1968 to SNA1993 to SNA2008,Working paper (code: eMOL09317; latest revision 15, as of May 10, 2017), Division of Economic Models, Statistics Denmark.

    • Search Google Scholar
    • Export Citation
  • Smidova, Z. (2016). “Betting the house in Denmark,OECD Economics Department Working Papers, No. 1337, OECD Publishing, Paris.

1

Prepared by Evan Papageorgiou (EUR) with contributions from Andreas Kuchler (Danmarks Nationalbank). This paper has benefitted from useful discussions with David Hofman (EUR), Paul Kramp (Danmarks Nationalbank), and Michael Osterwald-Lenum (Statistics Denmark).

2

Throughout this paper, households refer to the national accounts sectors households and non-profit institutions serving households.

3

Special considerations are given to households with real estate transactions, which are excluded for the year of the purchase or sale, but are included in the sample again the following year. For more information on the methodology and measurement issues we refer to Andersen (2016).

4

Assuming 47.5 percent private consumption as a share of GDP as in recent years.

Denmark: Selected Issues
Author: International Monetary Fund. European Dept.
  • View in gallery

    Household Net Wealth in Denmark and Elsewhere1

    (Percent of gross disposable income)

  • View in gallery

    Danish Household Assets and Net Wealth

    (Percent of gross disposable income)

  • View in gallery

    Financial Assets and Financial Net Wealth

    (Percent of gross disposable income, latest available annual data)

  • View in gallery

    Household Debt

    (Percent, latest available annual data)

  • View in gallery

    Pension Contributions and Payouts

  • View in gallery

    Gross Household Savings Rate

    Percent; average 2007–2015

  • View in gallery

    Savings of Household and Nonfinancial Corporate Sectors in Denmark

    Percent of GDP

  • View in gallery

    Holdings of Mortgage-Covered Bonds Issued by MFIs

    (Percent of monetary financial institution (MFI) assets)

  • View in gallery

    Unconsolidated Bank Assets

    (Percent of GDP, 2015)

  • View in gallery

    Household Leverage and Consumption

    (Percent)

  • View in gallery

    Stock of Residential Mortgages by Type

    (Billions of Danish kroner)

  • View in gallery

    Household Interest Paid

    (Percent of gross disposable income)

  • View in gallery

    Household Debt Distribution1

    Source: Danmarks Nationalbank using registry data from Statistics Denmark.

  • View in gallery

    Household Net Debt per Income Decile1

    (Percent of income after tax, 2015)

  • View in gallery

    Household Net Debt per LTV Group1

    (Percent of income after tax, 2014)

  • View in gallery

    Household Net Debt per Mortgage Type1

    (Percent of income after tax, 2014)

  • View in gallery

    Distribution of the Decline in Consumption from 100 bps Interest Rate Increase

    (Including Interest Income Effects, Unless Otherwise Noted)