Selected Issues

Abstract

Selected Issues

The Role of Asset Prices in Households, Pension Funds, and Credit Institutions1

Denmark is a financially deep economy where households have a central role in the propagation of credit and financial flows. As monetary policies across advanced economies may begin to normalize, this chapters explores the role of higher interest rates on the broader economy via the macrofinancial channel. The application of macroprudential policy measures may also need to be considered in light of potential vulnerabilities for some households-at-risk.

A. Introduction

1. The ongoing global economic recovery continues to raise the possibility of monetary policy normalization in the euro area and in other advanced economies. This chapter explores some implications to Denmark from a rise in interest rates. Danish households took advantage of low interest rates, and deleveraged, albeit only moderately, since the financial crisis. Nonetheless, household gross balance sheets continued to expand to considerable levels by international standards, suggesting potential vulnerabilities to shifts in interest rates and asset prices.

2. This chapter explores the impact of a shock to interest rates on financial assets held by households directly and indirectly via their pension investments. The large amount of financial assets reflects financial instruments both held (directly) as investments and (indirectly) as retirement savings. The exercise examines the impact of interest rates shocks on financial assets managed by pension funds and outright investments held by households. Given the robust and well capitalized nature of Denmark’s financial system, there is little evidence of systemic concerns arising from 100–200 basis points interest rate shocks. However, linkages between households and asset prices appear strong and capable to propagate shocks to the economy via reduced consumption.

3. Negative wealth effects from pension assets and financial assets could lower household consumption in a highly heterogenous way. The analysis finds some evidence that the sensitivity of consumption in response to shocks to financial wealth may be highly heterogenous across household groups. Households that share certain characteristics, such as low wealth levels and high debt-to-income ratios are impacted considerably harder than low-leverage households. If internalized, these asymmetries in household responses could aid the calibration of economic policies including macroprudential measures.

4. The structure of this chapter is as follows. Section B discusses the interaction of households with the rest of the economy, noting the important role of the financial system in allocating credit. Section C describes some basic features of Danish pension funds and the assets they hold. Section D explores the potential impact of higher interest rates on pension fund and life insurance assets. Section E extends the analysis to household consumption sensitivity to asset price shocks via wealth effects, discussing its potentially large role in propagating shocks to the economy. We also emphasize the likelihood of large heterogeneities across income groups. Finally, Section F suggests some policy recommendations to build resilience and some tentative conclusions.

B. Interaction of Households with the Rest of the Economy

5. Credit provision relies on the efficient transfer of capital between economic sectors. In a typical private-sector setup of a market economy, households save for future consumption or bequest via banks and pension funds, while also investing in financial and nonfinancial firms. Pensions and insurance firms also invest to generate income and pay out benefits. The financial sector facilitates lending and engages in investments with the rest of private sector (Figure 1).

Figure 1.
Figure 1.

Stylized Presentation of Private Sector Financial Transactions

Citation: IMF Staff Country Reports 2018, 178; 10.5089/9781484362556.002.A002

6. The role of households is central for credit creation and the structure of the private sector balance sheets. Half of domestic private sector credit is provided to households to finance housing purchases and consumption.2 The need for large borrowing is partly due to households’ relatively large savings via contributions to mandatory saving schemes. As a result, the fully-funded occupational pension system is a natural counterpart to the large household debt stock. The loop completes with pension funds investing heavily in covered bonds issued by credit institutions which use to fund their loan book. A balance sheet decomposition shows the magnitude of this funding/lending relationship of households, pension funds and credit institutions (Figure 2).

Figure 2.
Figure 2.
Figure 2.

Private Sector Financial Balance Sheets and Comparison

Citation: IMF Staff Country Reports 2018, 178; 10.5089/9781484362556.002.A002

uA02fig01

Monetary Financial Institutions Credit to the Private Sector

(DKK, billion)

Citation: IMF Staff Country Reports 2018, 178; 10.5089/9781484362556.002.A002

Source: Danmarks Nationalbank.Note: Includes credit from banks and mortgage credit institutions.

C. Household and Pension Fund Balance Sheets

7. The large assets and liabilities make households more sensitive to shocks. The sharp fall of house prices in the aftermath of the financial crisis resulted in a large consumption decline in Denmark, consistent with their high leverage. Further suggesting that high leverage tends to have an important role in households’ financial health, Andersen et al. (2016) find that the drop-in consumption was stronger among high-leverage Danish households.3 The 2017 Denmark, Selected Issues Chapter 1 examined the risk of higher rates to, primarily, household liabilities. Here the analysis looks further into the sensitivity of household assets to interest rates and asset prices, focusing on the effect of an interest shock on their pension assets, and the potential knock-on impact on household consumption.

uA02fig02

Household Leverage and Consumption

Percent

Citation: IMF Staff Country Reports 2018, 178; 10.5089/9781484362556.002.A002

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

8. The direct risks to the financial sector from household debt tend to be low. Despite highly indebted households the fallout from the decline in housing prices did not have significant first-order effects on the banking system. Housing loans in Denmark are full-recourse making it more difficult for households to default on their loans. In turn, households and their creditors are incentivized to seek out arrangements such as temporary modified repayment agreements to continue servicing their mortgage and staying in their properties. The strong social safety net and welfare programs can also soften the blow from economic hardship and smooth consumption. Danish credit institutions are liquid and well-capitalized, with regulatory capital amounting to nearly 20 percent of risk-weighted assets. In the past they have been able to withstand reasonable increases in nonperforming loans to the household sector. Ongoing stress tests by Danmarks Nationalbank show that risks to the financial system from household debt deterioration are generally manageable, owing to banks’ excess capital. Although some systemically important institutions may have small capital shortfalls under a severe recession scenario (DN, 2017), their effect arises from falling stock prices, changes in interest rates, and increasing credit spreads, rather than household debt specifically.

9. Credit conditions play a central role in the propagation of shocks in modern, financially deep economies, especially during downturns. Financial accelerator effects (BGG) are important in understanding the transmission of macro financial shocks. As noted by Bernanke, Gertler and Gilchrist (BGG, 1999) “endogenous developments in credit markets work to propagate and amplify shocks to the macroeconomy.” The key mechanism is the link of lending credit premia and the net worth of potential borrowers, which declines in value during downturns. These effects can play powerful roles in asset-rich and financially deep economies like Denmark. As noted by these authors, “this extra amplification is a step in resolving the puzzle of how relatively small shocks (such as modest changes in real rates induced by monetary policy, (…), can have nevertheless large real effects,” explaining how “deteriorating credit market conditions may materialize in the form of falling asset prices, rising real debt burdens or even sharp increases in insolvencies and bankruptcies”. The extension to bankruptcies and insolvencies does not apply directly to Denmark, where various institutional features such as variable mortgage and deferred amortization mean that households can offset shocks to incomes and interest rates, but the other channel, including via collateral effects remain valid.

10. Danish households hold a high amount of real (housing) and financial assets, even relative to the wealthiest economies. The combination of large asset holdings and leverage may explain why the Danish economy appears sensitive to adverse shifts in interest rates. Housing plays an important role in the financial system (Figure 3):

  • Housing is a major asset held by Danish households, together with pensions and financial assets.

  • High valuations and favorable tax treatment incentivize large house purchases, often funded via large mortgages. The factors above in conjunction with easy access to low interest rate borrowing helps to explain why Danish households’ debt to income ratios are among the highest among advanced economies.

  • Households are exposed to housing markets developments through their pension and life insurance investments. Mortgage credit institutions issue covered bonds to fund the mortgages provided to households, transferring part of mortgage risks to investors. These covered bonds are in turn purchased by other financial institutions (35 percent) and, crucially, pension funds and insurance companies which hold a substantial (28 percent) of the market.

11. Mortgage bonds link real estate developments back to households’ balance sheet and the broader financial system. Mortgage banks issue covered bonds to fund the mortgages provided to households with two main implications:

  • Mortgage credit institutions (MCIs) transfer part of mortgage risks to those investors who purchase their covered bonds, i.e. market risk. But MCIs are ultimately responsible to cover any shortfalls, should the value of (real estate) collateral fall sharply, for example.

  • Since the largest owners of covered bonds other than pension funds and insurance companies (28 percent) and domestic financial institutions (35 percent), most of the risks associated to lending to the real estate market remain in the domestic financial system.

Because of this additional channel of covered bond holdings, real estate developments not only affect household consumption via the usual wealth effects and collateral (financial accelerator) effects via housing, but also via potentially mutually reinforcing financial wealth effects through households’ large pension savings invested in financial assets, which are discussed next.

uA02fig03

Housing Cost Overburden by Income Group, 2016 1/

(Percentage)

Citation: IMF Staff Country Reports 2018, 178; 10.5089/9781484362556.002.A002

1/ 2017 data for DNK and LVA, 2015 data for SVK, ISL and TUR.Source: Eurostat.
uA02fig04

Apartments and Houses 1/

(Percent)

Citation: IMF Staff Country Reports 2018, 178; 10.5089/9781484362556.002.A002

1/ Weighted average for Copenhagen and Aarhus, simple average for all. Yellow dots denote Capital region.Sources: Danmarks Nationalbank, Finans Danmark and Fund staff calculations.

12. Understanding the effect of interest rate increases on household consumption is important for Denmark and other Nordic economies. Mian et al. (2013) examine the wealth shocks from the housing market collapse in the United States during the global financial crisis and their impact on consumption. Andersen et al. (2016) (and earlier versions of their work since 2013) make use of rich microdata for Danish households to study the relationship between household leverage and spending. Hviid and Kuchler (2017) analyze consumption and savings decision of Danish households and estimate the effect of rising housing prices on consumption formation. In other related studies, Floden et al. (2016) and Gustafsson et al. (2017) study the cashflow channel of repricing of liabilities and interest income for Swedish households using microdata.

Figure 3.
Figure 3.

Households, Credit and Leverage

Citation: IMF Staff Country Reports 2018, 178; 10.5089/9781484362556.002.A002

13. Two types of households appear particularly vulnerable to shocks: those on low incomes and those on high LTIs. The first vulnerable group are low-income households who spend a considerable share of their disposable income on housing (Figure 3). Higher interest rates will ultimately increase mortgage repayments, likely increasing the proportion of this vulnerable segment. The second group are households who have purchased in highly appreciating and potentially overvalued urban areas such as Copenhagen (Selected Issues), where LTI ratios and credit growth are noticeably higher than in the rest of the country. These vulnerabilities are compounded by the large proportion of variable-rate and interest-only mortgages in the system. There is a risk that high and still rising house prices (particularly in urban areas), relaxed credit standards (Figure 3), and extremely low mortgage rates, may increase the perception of affordability across a broad spectrum of income levels.

14. When households’ financial wealth falls, whether because the value of their financial asset prices decline, consumption falls (wealth effect). What makes Denmark vulnerable to shocks to asset prices is that its “households have very large assets and are among the households in the world with most financial assets” (Finance Denmark, 2017), combined with a large gross balance sheet and a high level of household debt. To get a sense of the potential magnitude of the ultimate impact on households’ balance sheets, we sketch the possible impact of a hypothetical 100 bps interest rate increase on the portfolio of Danish pension funds and life insurers,4 given they hold a significant share of household wealth in the form of pension savings. We then explore a series of possible implications for households.

15. Servicing mortgages and credit ultimately become more onerous for borrowers when interest rates rise. At the same time, asset prices decline when interest rates increase, this opposite move means that net asset position deteriorates when interest rates rise.

16. The pension fund system is among the largest as a share of the economy. Danish pension fund assets, at around $600 billion and 215 percent of GDP, are the largest relative to GDP, even relative to the most advanced and richest economies (Table 1). The large size of pension funds and life insurance portfolios reflects the high level of pension contributions and household savings.5

Table 1.

Denmark: Total Pension Assets in Major Advanced Economies

article image
Source: GDP: OECD (2016), except Denmark (2015). Old-age dependency ratio: ratio of population aged 65+ per 100 population 15–64: UN World Population Prospects (2015). Population: World Bank (2015). Total Pension assets (2016): Towers Watson Global Pension Assets Study, except Denmark: OECD (2015).Towers Watson Global Pension Assets Study 20159FT, “Foreign fund houses compete for $170 bln of Chilean retirement money”, 13 April 2014.

17. Households hold financial assets both (directly) as investments in their portfolios and (indirectly) by pension funds. The high debt of households is partly offset by large pension savings, but the resulting high gross balance sheets means that shifts to asset prices and changes in interest rates have the potential to shift the net asset position of households significantly because of the different effect on assets (which typically fall in value) and liabilities (which rise, even if with some desirable flexible features in Denmark, which tend to lessen the adverse impact). To better understand the potential type of vulnerabilities the first step of the analysis is on the balance sheets of pension funds and life insurers, as they hold a significant amount of households’ financial assets. The second step is to reconcile this with analysis on household liabilities and discuss the impact on households’ net asset position.

D. The Impact of Interest Rates on Asset Prices and Balance Sheets

18. The normalization of monetary policy by the ECB and other major central banks is likely to impact asset prices. This may have ramifications for Danish households and pension funds balance sheets. Interest rates are anticipated increase further across Europe, US and other advanced economies in the coming years, as the post crisis exceptional monetary policy measures are withdrawn, as warranted by improving economic conditions. Nonetheless, long-term interest rates could increase more than widely anticipated given that bond term premia remain near historically low levels across major advanced economies (see IMF GFSR, October 2017). Danmarks Nationalbank estimates that ECB’s “unconventional monetary policy in the beginning of 2015 reduced the term spread in Denmark by at least 50 bps” (see Danmarks Nationalbanken Working Paper). Even 50 bps may be a lower bound, as it does not encompass the effects of asset purchases of other major central banks6. Even assuming two 25 bps interest rates ECB hikes over the coming 2–5 years,7 in addition to the unwinding of the 50 bps fall of the term spread, Danish longer-term interest rates could easily rise by 100 basis points, which provides the baseline shock for the next exercise.

Exploring the Impact of Higher Interest Rates on Pension Fund Assets

19. Fluctuations of pension assets can affect consumption decisions. It is often assumed that pension assets have a relatively small impact on consumption and savings decisions, especially in the short to medium run, because they are tied until retirement, and paid out over many years, and not all households monitor the fluctuations of pension savings frequently, unless they are moving closer to retirement age (as populations age a larger proportion of households will get closer to entering retirement).8 Nonetheless, there is recent evidence that suggest otherwise, and so we explore some reasons why households may become more sensitive to movements in their pensions assets. With the increasing dominance of DC schemes, gains and losses on pensions assets are transferred back to households (net of eventual guarantees which are typically set conservatively low). For this reason, shifts in asset prices on pension funds’ investments of households’ future pensions may becoming a more important channel of transmission of revaluation effects, and by reflex to consumption and savings decisions.

20. As the provision of pension shifts more prominently to the private sector, future pension payouts may become more linked to financial market developments. Simultaneously, investment risk is increasingly being transferred from pension funds and life insurance companies back to households, as schemes are now predominantly DC. Consequently, it is plausible to speculate that household consumption and savings decisions may be becoming more sensitive to shifts in asset prices, and will continue to do so for the foreseeable future, particularly in wealthy and financially deep economies such as Denmark and the Nordics.

uA02fig05

Pension Funds and Life Insurance Companies: Assets

(Percent of total assets)

Citation: IMF Staff Country Reports 2018, 178; 10.5089/9781484362556.002.A002

Source: Danish FSA and Fund staff calculations

21. Danish pension funds invest heavily in equities and fixed income assets. About 35 percent of pension fund and insurance company (PFIC) portfolios are allocated to equities, a significant proportion of which are invested in US European and emerging markets. Another 30 percent are allocated across fixed income assets (spanning government (7 percent), corporate (9 percent) and mortgage bonds (14 percent)). The bond exposures are concentrated in European, Danish and other developed markets, while covered bonds are overwhelmingly domestic. The largest remaining category, “other assets” (35 percent), includes an array of instruments (including derivatives, trade credits, real estate and other illiquid assets whose returns we assume are moderately (positively) correlated with the rest of the portfolio).

22. Asset prices are sensitive to interest rates. The value of a financial asset, is determined by the future income investors expect to gain from it, discounted by an appropriate “risk free” interest rate (such as the yield on a long-term government bond). So, when interest rates rise, typically asset prices fall (all else equal). These estimates of the price impact of interest rate shocks were performed on the major asset categories held by pension funds (corporate, government and covered bonds and equities across the major developed and emerging economies). Given these characteristics, we then match a benchmark bond index to captures the essential features of each category: namely duration and convexity. Based on these characteristics we estimate9 the impact of interest rate changes on the price of that type of bond trading in that country, based on the key attributes that determine its interest rate sensitivity: duration and convexity. To analyze equity price valuations, we employ a multi-stage dividend discount model, calibrated separately on each major equity market, and then determine the impact of a given interest rate shock holding all other variables constant (see Annex 1 for more information).10

23. Assessing the impact of interest rate shocks on asset prices. Many factors determine the impact of an interest rate shock on asset prices, so these estimates are best viewed as tentative starting point, subject to large uncertainties. The exercise’s overall outcome is that a 100 bps (or one percentage point) interest rate shock would reduce the value of the portfolio assets held by pension funds and life insurance companies’ (from now on referred to just pension funds) by as much as 6–8 percent, rising to 9–15 percent if we double the shock to 200 bps. In terms of major asset classes our estimates suggest:

  • A one-off 100 basis points interest rate shock lowers the value of pension funds fixed income assets by approximately 4–7 percent. For the fixed income assets held by pension funds, we find that duration captures most of the negative impact from interest rates, although interest rate sensitivity varies significantly across the range of fixed income instruments.11 For example, Danish government bond benchmark indices,12 because of their higher duration (8.6 years), are more price-sensitive to interest rate increases than most US and European benchmark indices (in the 4-year and 7–8 year ranges, respectively). At the other extreme, Danish mortgage bonds, due to their short duration), are resilient to changes in interest rates.

  • The same 100 bps interest rate shock lowers the value of pension funds’ equities portfolios by around 8–11 percent. The equity markets analyzed included the US, Denmark, Germany and euro area ex-Germany, UK. The estimated impacts of the shock on stock prices ranged from under 9 percent to over 12 percent. The standard premise of our analysis was that the price of a stock captures the present value of its future income stream, and specifically future dividends, discounted by a long-term “risk free” interest rate. And a higher interest rates will lower the present value of that future income stream, resulting in a lower equity price (see Annex for details).13

24. Equity valuations are sensitive to shifts in interest rates. The current sensitivity of stock prices to interest rate shocks reflects two developments: i) Equity Risk Premia (ERP) in major advanced economies, Denmark included, have declined (according to some models14) as investors reached for yields in a global environment of accommodative monetary policies and low interest rates across, and, ii) corporate earnings growth has been slowing (or is expected to slow15) in several major advanced economies (it peaked around mid-2017 in Denmark). In principle, a low ERP signals reduced compensation for risk, and consequently vulnerability to shifts in risk perceptions. The combination of low ERP and slowing earnings growth reinforces the sensitivity of equity valuations to adverse shocks, whether economic or financial, domestic or external, because current prices do not offer much return for equity risk over and above the yield on domestic government bonds.

25. Small open and financially deep economies are vulnerable to global financial shocks. Denmark’s status as a safe-haven offers some protection in the face of adverse global financial shocks. Nonetheless, a marked decline in the global price of risk (reflecting, say, an increase in global risk aversion) could result in a significant downward adjustment in Danish risk assets, and notably equity prices, for the reasons discussed above. The large and increased share of foreign investors in Denmark’s financial markets also suggests a degree of exposure to external shocks, potentially increasing the sensitivity of domestic asset prices to external forces, and possibly to funding stress in the case of very large shocks (for example, if some foreign investors were to exit covered bond markets). Denmark’s financially-deep economy with some highly leveraged households, means that domestic macro-financial channels, such as the financial accelerator and collateral effects discussed earlier in the note could amplify the initial impact of asset price falls to the broader economy.

E. Asset Prices and Consumption: Who is More Vulnerable?

26. Adding the effect of rates on financial assets provides interesting insights into household consumption. The 2017 Denmark Article IV, Selected Issues Chapter 1 focused on the liabilities side of balance sheets, and found Danish household consumption to be overall modestly sensitive to interest rates, but with some groups exposed disproportionately. To examine the impact of interest rate shocks on the asset side of the balance sheet, the analysis focuses on the sensitivity of consumption, or marginal propensity to consume (MPC) out of financial wealth. The MPC out of wealth is often found to be relatively small when estimated from (aggregate) economy-wide data, and at times estimated to be slightly smaller than MPC out of housing wealth, especially in less recent studies. However, more recent studies have been attributing increasingly more importance to financial wealth effects, finding it to be larger than previously thought.

27. Recent evidence suggests financial wealth effects might be even larger than housing wealth effects. Wealth effects are typically believed to be important, particularly during downturns (see also Hiivid et al, 2017), and their magnitude can be amplified by high debt levels and credit constraints. These channels were an important amplifier of the downturn during the Great Recession, including in Denmark, where the drop of household consumption was among the sharpest in advanced economies. In addition, to the more widely explored wealth effects from housing, some recent studies based on extensive micro data find wealth effects on consumption from financial assets to be stronger than from housing for all but the wealthiest households. Although similar analytical studies based on Danish data are not available, it is plausible that given Danish households’ high levels of financial and housing wealth, these effects are significant and likely mutually reinforcing also in Denmark.

28. In addition, MPC out of financial wealth varies sharply across household types. Recent studies based on detailed micro data show how marginal propensity to consume out of financial wealth16 varies across income/wealth groups. Notably, ALS 2015 finds that French households’ MPC out of financial assets is considerably larger across the bottom half of the wealth distribution (see Table 2), even though the former group holds less (net and gross) wealth in the form of financial assets than their richer counterparts.

Table 2.

Denmark: Impact of Asset Price Shocks on Consumption, Across Different Wealth Groups

article image
Sources: Arrondel L., Lamarche P, and Savignac F, 20151 for details on the parameter estimates. Stars denote significance levels as follows: **: 5% *** 1%. IMF estimates for asset price changes resulting from interest rate shocks.

See Arrondel L., Lamarche P, and Savignac F, 2015; their study controls for a wide range of variables: income expectations, age, work status, education, household composition, credit constraints, unemployment spells, sick leave. Financial wealth includes voluntary pension and life insurance schemes.

29. The impact of asset prices on consumption likely varies significantly across wealth groups. In Table 2, we provide the results of the impact of shocks to financial asset prices on household consumption, across the wealth distribution. However, due to the lack of Danish estimates for the necessary MPC parameters (and wealth to consumption ratios), we apply ALS (2015) MPC estimates, which rely on a rich French dataset.17 The assumption is that the French parameters applied to calculate the impact of asset price shocks on consumption are broadly representative for the Danish case, or at least insightful on some potentially important heterogeneities. With this in mind, the results should be interpreted with caution, as there may be significant differences between the French and Danish results. Some insights however, are likely to be relevant for Denmark.

30. Pensions make up a considerable part of households’ financial assets and reflect Danish households investing and risk preferences. For this reason, to apply a representative shock to wealth it is assumed that household portfolios have the same structure as the portfolios of Danish pensions funds and life insurance companies. To then get a sense of the impact of the resulting change in asset prices on household consumption, we apply the change in the value of the asset price to the estimates of the elasticity of consumption to asset prices, as shown in Table 2. rather than assuming a single population-wide estimate of the consumption elasticity, we apply the relevant elasticity across each of the key wealth percentiles based. To obtain an estimate of the impact on consumption for a given shocks to financial wealth, we apply the same shocks we applied to the portfolios of Danish pension funds.18

31. Consumption of the least wealthy households is the most vulnerable to asset price shocks. As caveated upfront, the reference consumption parameter estimates used are based on micro data from French household survey (in absence of equivalent estimates on Danish households). With this limitation in mind, we use them as an indicative proxy for what Danish MPC and consumption elasticity estimates might look like across the wealth distribution. In addition, as mentioned above we make sure to apply the shock to a portfolio structure of Danish pension and life insurance companies, which capture Danish investor preferences, and are in this sense representative. The most important result is the heterogeneity of financial wealth effects across different income/wealth groups. A 5–8 percent drop in asset prices has a strong negative and statistically significant effect on consumption of the group with below median wealth (−2.7/-3.5 percent drop in consumption). Some appreciable impact is also found on the following wealth cluster (where consumption falls between 1.3 percent-and1.8 percent), there is virtually no effect on the wealthiest group (where the MPC parameter is not statistically different from zero, see last row of Table 2). While this discussion focuses on financial assets across the wealth distribution, it should be noted that wealth effects from housing could alter the estimated impact of a shock to households’ total wealth (i.e. both housing and financial assets). For instance, low income households often rent, and hence a housing price shock would likely impact middle income households comparatively more.

32. Less wealthy households are more consumption sensitive to shifts in asset prices. The higher sensitivity of the less well-off households is based on a French-specific dataset and may appear counterintuitive, but many members in this group are not home owners, and often hold some financial assets, even if in limited amounts (e.g. savings and pensions). Their increased sensitivity to shocks in asset prices may reflect limited buffers to absorb adverse shifts in their wealth (for example, gross wealth well in excess of net wealth, which may even be negative). This variation in consumption sensitivity across household groups dilutes the estimation of MPC when estimated from economy-wide (aggregate) data, hiding important heterogeneities that seem relevant for policy settings. For example, Jappelli and Pistaferri (2014) show that households with little liquid wealth and without high past income react particularly strongly to economic stimuli. Recent research has also shown that transitory shocks can have a very substantial impact on MPC and that such shocks can be quite large (Carrol, 2017).

33. The effect of interest rate increases on household liabilities continues to suggest modest overall consumption sensitivity but higher for certain groups-at-risk. Based on the methodology presented in earlier analysis (see Denmark 2017 Article IV Selected Issues, Chapter 1) household-level data on liabilities show that the median sensitivity of household consumption to rate increases is modest. One percentage point increase of the borrowing rate decreases consumption to the median household by less than 0.1 percentage points (Figure 4), and by 0.4 percentage points for the entire population on average, or some 0.2 percentage points of GDP. This is lower than the 2017 estimate (0.3 percentage points) based on the previous survey of household microdata. The rising share of fixed-rate mortgages owing to the Supervisory Diamond (see Annex [II]), and rising disposable income may explain the declining sensitivity. Nevertheless, the previously-identified groups-at-risk with high loan-to-value, and high loan-to-income characteristics maintain much higher consumption sensitivity with respect to rising rates (Figure 4).

Figure 4.
Figure 4.

Interest Rate Shock: Anatomy of the Impact on Consumption via Wealth Effects

(Percent; based on data from the 2015 annual registry of households)

Citation: IMF Staff Country Reports 2018, 178; 10.5089/9781484362556.002.A002

Note: The decline in consumption due to liabilities and interest income is based on the methodology described in the Denmark 2017 Article IV, Selected Issues Chapter 1.

34. Combining the microdata analysis on liabilities with the financial wealth effects mentioned above offers a more thorough view of the sensitivity of households to rates. Using the granular financial balance sheet data provided by the household-level registry and applying the framework for the effect of rate rises on financial assets generates estimates for the entire financial balance sheet. Like the analysis above, consumption elasticities for the marginal propensity to consume from assets are drawn from Arrondel et. al. (2015), and are applied to the asset composition of each cohort considered in the analysis of the liabilities.19 Figure 5 presents the results of the estimation, and for presentational simplicity only the effect on the median household in each cohort is shown.20 The analysis does not consider the value of nonfinancial assets, such as housing, and several offsetting factors such as rising wages and the effects of asset diversification from rising housing prices are not considered which may reduce the hit to consumption. Considering the effect of interest rate increases on real assets could change some conclusions.

Figure 5.
Figure 5.

Anatomy of Impact on Consumption, Introducing Liabilities Explicitly

(Percent; based on median household in each cohort from 2015 annual registry household data)

Citation: IMF Staff Country Reports 2018, 178; 10.5089/9781484362556.002.A002

Note: The decline in consumption due to liabilities and interest income is based on the methodology described in the Denmark 2017 Article IV, Selected Issues Chapter 1. The decline in consumption due to financial assets is based on elasticity assumptions for the marginal propensity to consume from financial wealth from Arrondel et. al. (2015) and simulated declines of 8 percent for equity and other investment and 5 percent for pension and insurance assets.

35. Consumption sensitivity varies considerably among households, and certain groups-at-risk are significantly more exposed to rising rates. Households with low incomes are exposed more to rising rates via their financial asset holdings (including via pension and insurance assets) than higher income households that have lower marginal propensity to consume from their financial wealth as shown earlier in this chapter. The combined effect on assets and liabilities shown in panel A of Figure 5 shows broadly rising impact with higher incomes (except the top-income households). Similarly, higher rates affect more households with higher debt relative to their value of their property rather than ones with low loan-to-value (LTV; panel B). The balance sheet impact appears to be increasing marginally more for households with LTV greater than 80 percent, which supports an ongoing staff recommendation to lower the LTV cap from the current 95 percent. Highly-leveraged households shown in panel C are also significantly more prone to reducing their consumption with rate increases. Like in last year’s assessment on liabilities, the marginal impact to consumption from the high loan-to-income (LTI) cohorts, such as above 400 percent, is found to be evident of larger vulnerabilities in high-leverage households. Equally, the consumption sensitivity of both financial assets and liabilities to rates declines faster within lower-LTI households. Unlike the LTV cap which is less effective as house prices continue to rise, LTI-based macroprudential measures are not subject to property price fluctuations and can be more effective in enhancing household resilience. Relatedly, with respect to the effect of higher rates on consumption by LTV and LTI cohorts the quantitative assessment is likely unchanged if nonfinancial assets (such as housing) were also considered, because of the positive relationship between indebtedness and housing values. Finally, the sensitivity of household consumption to rising rates does not decline for older households when considering their holdings of financial assets and the increased propensity of retired people to consume out of wealth rather than regular income. Considering only the liability structure of older households may be underestimating their consumption sensitivity.

F. Policy Recommendations and Conclusions

36. Macrofinancial linkages among households, assets prices and the financial system are strong. While Denmark has a robust economy, with a large net asset position and substantial buffers, risks from rising interest rates may exacerbate latent vulnerabilities, particularly for households and household consumption. The linkages between households, asset prices and the financial system are capable to propagate and amplify shocks to the broader economy.

37. Tightening existing macroprudential measures could help contain the formation of high risk debt, and build resilience. Some policy recommendations based on the analysis above are listed below. Their implementation via the existing frameworks of the Supervisory Diamond for Mortgage Credit Institutions and the Good Business Practice for Housing Credit can speed up their application.

  • Lower LTV limits from 95 to 90 percent or lower, to reduce the exposure of households to swings in housing prices, and help buffer the impact to consumption. Figure 5 suggests that consumption is affected marginally more for households with LTV greater than 80 percent, and limiting the effect of housing price fluctuations may boost resilience. Earlier staff analysis (IMF 2016 Article IV, Selected Issues, Chapter 3) suggests that an additional 5 percent down-payment lowers aggregate consumption by 1.5 percentage points one year after introduction, and increases it by 0.2 percentage points in a new steady-state because of lower debt-servicing burden.21 Therefore, the stability gains may be significant in the long-run.

  • Impose mandatory amortization for highly-leveraged households for households with loan-to-income above 400 percent. To encourage further reduction of interest-rate sensitivity the loan-to-income limit could be raised to 500 percent if financing is via fixed-rate mortgages.

  • Stricter loan-to-income restrictions for all loans irrespective of LTV considerations, possibly with tighter limits for interest-only and adjustable-rate mortgages should be evaluated to help contain household leverage, increase resilience, and limit the dampening effects on consumption. Staff analysis suggests that the consumption sensitivity of both financial assets and liabilities to rates declines faster for lower-LTI households, and removing the link to LTV, may remove the influence of rising property prices.

38. Mortgage interest deductibility (MID) could be reduced further than currently planned, as MID distorts investment incentives and incentivizes leverage (Gruber, 2017). That would bring Denmark in line with other European economies that removed the bias for such borrowing (see table below). During the transition period to a lower mortgage deductibility regime, the current low rate environment would mitigate the adverse impact on homeowners. Supply-side rigidities in urban areas should be addressed. There is room to simplify both the length and complexity of the planning process (EC, 2017b), particularly in urban areas. Simpler and more streamlined zoning and planning processes would allow housing supply to respond to increases in demand without steep price increases. Rent controls remain among the tightest in the EU and should be reduced. According to EC estimates, around 80 percent of private rental housing remains under rent control. Below-market rents limit the incentive to supply rental units, and incentivize the purchase of housing, adding upward pressure to property prices. This restricts labor mobility and migration toward cities, with adverse consequences for productivity and social mobility. There are emerging signs of pressures in the public transportation system contributing to housing pressures, especially around Copenhagen area, which could be alleviated by increasing public capital spending.

Current Mortgage Interest Deductibility from Personal Income Taxes

article image
Sources: National tax and other authorities; IMF staff calculations.

References

  • Andersen, A., Duus, C. and Jensen T.Household Debt and Spending During the Financial Crisis, Evidence from Danish Micro Data”, European Economic Review 2016

    • Search Google Scholar
    • Export Citation
  • Arrondel, L, Lamarche, P, and Savignac, F.,Wealth Effects on Consumption Across the Wealth Distribution: Empirical Evidence,” ECB Working Paper No 1817, 2015.

    • Search Google Scholar
    • Export Citation
  • Bernanke, B, Gertler, M, and Gilchrist, S., The Financial Accelerator in a Quantitative Business Cycle Framework, NBER Working Paper, 1999.

    • Search Google Scholar
    • Export Citation
  • Blinder, A.,Intergenerational transfers and Life Cycle Consumption,” American Economic Review, 1976.

  • Carroll, C. Slacalek, J. Tokukoka K. and White, M.,The Distribution of Wealth and the Marginal Propensity to Consume”, Quantitative Economics 2017

    • Search Google Scholar
    • Export Citation
  • Hviid, S.J, and Kuchler, A.,Consumption and Savings in a ow Interest Rate Environment”, Danmarks Nationalbank Working Paper, June 2017.

    • Search Google Scholar
    • Export Citation
  • IMF, Global Financial Stability Report, A Bumpy Road Ahead, April 2018.

  • Finance Denmark, Memo, Perspectives on Danish Household Debt, December 2017.

  • Danmarks Nationalbank, Several banks are stepping on the accelerator—No. 11, November 2017.

  • Danmarks Nationalbank, Risks are building up in the Financial Sector, Financial Stability, Second Half 2017.

  • Danmarks Nationalbank, Monetary and Financial Trends—No. 3, March 2018.

  • Fuller, R.J, and Hsia, C.C,A Simplified Common Stock Valuation Model, Financial Analyst Journal,” 1984.

  • Geng, N.Fundamental Drivers of House Prices in Advanced Economies,” IMF Working Paper, Forthcoming 2018.

  • King, M.Debt deflation Theory and Evidence,” European Economic Journal, 1994.

  • Jappelli, T, and Pistaferri, L,Fiscal Policy and MPC Heterogeneity,” American Economic Journal: Macroeconomics, 2014.

  • Inkinen, M, Stringa M, and Voutsinou, K,Interpreting equity price movements since the start of the financial crisis, Bank of England Quarterly Bulletin,” Q1 2010.

    • Search Google Scholar
    • Export Citation
  • Mian, Rao, and Sufi,Household Balance Sheets, Consumption and theEconomic Slump,” Quarterly Journal of Economics, 2013.

  • Panigirtzoglou, N, and Scammell, R., Analysts’ “Earnings forecasts and equity valuations, Bank of EnglandQuarterly Bulletin, Spring 2002.

    • Search Google Scholar
    • Export Citation
  • Slacalek, J.,What Drives Personal Consumption? The Role of Housing and Financial Wealth,” ECB Working Paper, 2013.

Annex I. Estimating the Impact of Interest Rate Shocks on Assets Prices—A Simple Approach

A standard approach is to express the current value of a stock p, at time t, in terms of the expected stream of dividends, D, from the next period (t+1) onwards:

pteq=Dt+1(1+rt+1)+Dt+2(1+rt+2)2+Dt+3(1+rt+3)3+

This stream of future dividends must then be discounted by a discount rate R(t), which contains a risk-free long-term interest rate (such as the US Treasury yield or the German Bund yield) and a risk spread known as the equity risk premium (ERP).

The ERP, captures the expected return on stocks in excess of the risk-free rate.1 The ERP is considered a measure of aggregate risk and a determinant of the cost of capital for corporations, and of savings decisions of individual households. In the simplest case, dividends are assumed to grow at a constant rate, g, over the life of the asset. In this case, we can re-write the equity price level as a ratio of the growth rate of dividends scaled by the ERP plus the difference between a constant2 risk free rate r and the steady state growth rate of earnings:

pteq=(1+g1)Dt(ERP+r)g1

The assumption of constant dividend growth can be relaxed. In this exercise, we assume a 3-stage process for earnings growth (specifically for earnings per share, EPS). That is, the growth rate of EPS ultimately converges to a steady state level (i.e. EPS growth is mean-reverting). While the equity price level (P(t)) is observable and so are the dividend and the current risk-free rate, the the growth rate of EPS g is not. So, analyst earnings forecasts (source: Reuters I/B/E/S) are used to provide an estimate for g=g(ibes) over the first dynamic stage of the model (see Figure A1). Thereafter, the growth rate transitions towards a steady-state rate g, which is pinned down by long-term restrictions.3 For a 3-stage dividend discount model, the valuation equation can be written as:

pteq=Dt(ERP+rt+h)gss[(1+gss)+γ(gfcts,t+hgss)]

The ERP affects the cost of capital, and is therefore a determinant of investment. For example, when the ERP is high, issuing equities is costly for corporates, and are not incentivized to finance in capital markets to invest.

Bonds and Fixed Income Assets

To estimate the effect of a given shock to interest rates on the price of a bond, we can approximate the impact by using the relevant bond’s duration and convexity properties.

pt,t+Tfi=Ct+1(1+rt+1)+Ct+2(1+rt+2)2+Ct+3(1+rt+3)3+...+Pt+T(1+rt+T)T

Specifically, the change in the price of the bond is a negative function of duration (d), plus a second order term which captures the effect of convexity:

py1p=1Φ[d]dy+122py2[dyΦ]2

The longer the duration of the bond (d) the more negative will be the impact of an interest rate increase on its price level. Higher convexity also results in a larger price impact of interest rate shocks.

Annex II. Summary of Macroprudential Policy Measures in Denmark

1. The global financial crisis and the ensuing credit crunch contributed to a housing market bust in Denmark. Housing prices declined 23 percent in real terms nationally from their 2006 peak to their eventual trough in 2009, and in some regions and housing market segments, such as the Copenhagen metropolitan area and owner-occupied flats, prices fell as much as 40 percent over the same period. The subsequent recovery was slow at first, led by the major metropolitan areas partly reversing their substantial decline, but it gained speed after 2013 and became more widespread as economic growth resumed.

2. Rising housing prices amid high household debt have beckoned the need to control the formation of new debt, and build resilience in households. The decline in housing prices and household incomes following the global financial crisis forced households to reduce consumption while deleveraging. Credit creation slowed down considerably compared to its pre-crisis trend, and has grown only marginally in nominal terms since 2009 (and declined as a share of GDP or disposable income). Nevertheless, household debt in Denmark remains among the highest in the OECD, and the recovery in housing prices has added to the propensity for risky borrowing. The rapid increase in housing prices in the metropolitan areas overlaps with the areas with strongest credit creation, inviting concerns that households are forced to borrow more than they can afford. This makes them sensitive to future increases in lending rates and tightening of financial conditions. As such, macroprudential and other housing policies in Denmark since 2012 centered around the need to control risks related with the high household debt, and rising housing prices.

3. Danish authorities have adopted many measures to address the increasing formation of risky debt. Measures are usually announced in combination of metrics, and the paragraphs below discuss the main macroprudential and other measures announced since 2012.

4. Nevertheless, ongoing housing price increases may be seeding more macroeconomic vulnerabilities. The rise in housing prices forces households to borrow more relative to their income. This can make them more sensitive to interest-rate shocks, as there is still a high share of adjustable-rate mortgages.

uA02fig06

Share of Households with LTI Greater than Four

(Percent of total new mortgage lending)

Citation: IMF Staff Country Reports 2018, 178; 10.5089/9781484362556.002.A002

Source: Dan mark* National bank, Financial Stability Report 2H2Q17.

5. Macroprudential policy in Denmark is performed with a broad input from several authorities. The designated macroprudential authority in Denmark is the Ministry of Industry, Business and Financial Affairs (formerly known as the Ministry of Business and Growth) and operates within an institutional framework. A key component of the framework is the Systemic Risk Council, which is tasked with identifying and monitoring systemic financial risks. The Systemic Risk Council meets and assesses risks periodically, and issues observations, warnings and recommendations to the Ministry of Industry, Business and Financial Affairs with a “comply or explain” principle. The Systemic Risk Council comprises the Danish central bank (head), the Danish Financial Supervisory Authority and other public entities and experts.

6. Many demand-side macroprudential measures have been put in place since 2014 to bolster the resilience of borrowers. Loan origination and amortization measures have been bundled in primarily four sets of measures:

  • Supervisory Diamond for Mortgage-Credit Institutions.1 (Proposed: December 2014; effective: January 2018, unless otherwise noted.) A regulatory framework consisting of 5 benchmarks for mortgage credit institutions was announced in 2014 and amended in 2016.

  • Mandatory Downpayment. (Proposed December 2014; effective November 2015). A consumer protection clause mandates at least 5 percent downpayment for residential real estate purchases, translating into an effective 95 percent maximum loan-to-value (LTV) limit. However, tighter single-loan restrictions apply, with 80 percent LTV per loan.2 The remaining 15 percent of the value of the property is financed with an additional loan having a secondary lien status.

  • Guidelines on Good Mortgage Lending in Growth Areas (Seven Best Practices). (Proposed September 2015; effective February 2016). On a recommendation by the Systemic Risk Council, the Danish FSA set seven guidelines in 2015 on good mortgage lending, particularly in areas with large price increases, dubbed Seven Best Practices. The first five practices apply only to borrowers in high housing price growth areas (Copenhagen district area and the municipality of Aarhus).

    i. Before granting floating-rate mortgages lenders must assess borrowers’ repayment ability under a scenario of interest rate hikes.

    ii. When granting loans to applicants for buying into cooperative housing associations lenders must assess the cooperative associations’ debt under a scenario of interest rate hikes.

    iii. Borrowers with negative equity position must amortize sufficiently with a maximum repayment period of 30 years, but exceptions may be granted if LTV is below 80 percent.

    iv. Borrowers with loan-to-income (LTI) between 4 and 5 must have positive net wealth in the event of a 10 percent decline in house prices, and borrowers with LTI above 5 are similarly stress-tested for a 25 percent decline in house prices. Exemptions may be given to borrowers with high job security if they have fixed-rate and amortizing mortgages.

    v. In cases where a borrower buys a new home before the old one is sold, the borrower must pay interest and principal payments on both homes until the lender expects the old home to be sold (6 months minimum). Similar provisions apply to borrowers with more than two residences.

    vi. Lenders must make individual considerations for assessing a borrower’s repayment capacity rather than only meeting institutional minimum requirements for their disposable income.

    vii. When financing purchases in cooperative housing associations, lenders must review the association’s underlying financial status (annual reports and budget) to ensure that the borrower is buying a healthy association.

  • Good Business Practice for Mortgage Lending. (Proposed March 2017; amended October 2017; effective January 2018.) Following recommendations by the Systemic Risk Council the government adopted lending restrictions for households with LTI greater than 4 times and LTV greater than 60 percent: (a) the interest-rate fixation of floating-rate mortgages needs to be at least 5 years, and (b) deferred amortization is only applicable on 30-year fixed-rate loans.

Demand measures tend to be tailored to fit the risks forming in households, rather than applying them as outright restrictions of lending. The summary table below shows the span of these measures along the traditional LTV, DTI, and amortization taxonomy.

uA02fig07

Supervisory Diamond for Mortgage Credit Institutions

Citation: IMF Staff Country Reports 2018, 178; 10.5089/9781484362556.002.A002

Source: Danish FSA.1/ Lending segments are private homeowners, rental property, agriculture and other corporate.2/ Applies only to lending to private homeowners and rental property.3/ Applies only to private homeowners.

Demand-Side Macroprudential Measures in Denmark and their Scope

article image
Source: IMF staff.

7. Supply-side measures have also increased in line with European legislation. The Danish authorities have introduced capital measures within the relevant European directives.

  • A capital conservation buffer applies to all Danish credit institutions since 2015, and will reach 2½ percent of risk-weighted assets by 2019.

  • A systemic risk buffer applies to six systemically-important institutions,3 and by 2019 it will range between 1 and 3 percent of risk-weighted assets, depending on the institution.

  • A countercyclical capital buffer for all credit institutions currently set at ½ percent of risk-weighted assets set to take effect on March 31, 2019. The Systemic Risk Council has recommended that this measure may have to rise to 1 percent next year if risks continue to build up, but the government warned that under current credit conditions this will not be approved.

The six systemically-important financial institutions will carry fully-loaded combined buffer requirements between 4 and 6 percent of risk-weighted assets by 2019, in excess of minimum capital and Pillar II add-ons.

8. New liquidity measures were introduced with the Capital Requirements Directive IV (CRD IV). The 2009 Danish Financial Business Act set requirements for banks to hold liquid and unencumbered assets as a liquidity buffer. Banks were subject to a funding ratio to prevent maturity mismatches. The 2010 Supervisory Diamond for Banks framework tightened some of these regulations with the excess liquidity coverage.4 Since 2015 under CRD IV a liquidity coverage ratio and net stable funding ratio (NSFR) have replaced the liquidity buffer and funding ratio. While most credit institutions observe liquidity requirements in line with the NSFR guidelines, institutions are not required to meet the NSFR requirements yet. The authorities are currently awaiting the legislative process in the EU with respect to the NSFR.

9. Some tax reforms have also been implemented. The deduction of mortgage interest expense against income taxes is being reduced through 2019. Borrowers can deduct up to 32.7 percent of their mortgage interest expense from their personal income taxes for expenses up to DKK 50,000. For expenses over DKK 50,000 the deduction is limited to 26 percent of interest expense in 2018 and 25 percent in 2019 and thereafter. A property and land tax reform was agreed in May 2017, and when it goes in effect in 2021 it aims to remove a property tax freeze and allow for more representative land tax increases.

1

Prepared by Evan Papageorgiou and Vladimir Pillonca (both EUR) with contributions from Andreas Kuchler (Danmarks Nationalbank). This chapter has benefitted from useful discussions with Miguel Segoviano (EUR).

2

Throughout this chapter “households” refer to the national accounts sectors households and non-profit institutions serving households.

3

To be sure, large diversified assets and savings also have significant benefits such as funding future consumption, contributing to fiscal sustainability and accelerating wealth creation (see 2017, Denmark Selected Issues, Chapter 1, Section B).

4

This exercise is not intended to be an accurate stress test, but merely an illustrative exploration of the potential impact of higher interest rates on asset prices and balance sheets, based on available information.

5

Selected Issues papers for the 2014 and 2017 Denmark Article IV consultations provide more information on the structure of the Danish pension system.

6

To evaluate the plausible increase in long-term interest rates as captured, say, by a 10-year bond, one would need to add to the term spread component (50 bps) an estimate of how much interest rates themselves are anticipated to rise over that period. Even assuming conservatively that interest rates are expected to be on average 50 bps higher over the coming ten years, would result in an increase of 10-year bond yield of 100 bps (assuming expectations theory).

7

This implicitly assumes a broadly unchanged interest rate monetary policy differential between the Danmarks Nationalbank and the ECB.

8

If life expectancy were to increase more than retirement age, retirees will have to rely on pensions savings for a higher number of years, although this is not a major issue in Denmark.

9

These estimates of the impact on fixed income assets were performed on assets by type and location e.g. bonds may be i) covered ii) corporate and iii) government) and may trade in a) Denmark, b) Germany, c) US, d) UK and so on). Each bond type was matched as closely as possible to a representative benchmark index (e.g. ICE-BofAML European Government index) on which the analysis was then performed, based on duration and convexity characteristics. See Annex for details.

10

For example, the equity risk premium and dividends are held constant, and earnings per share continue to grow steadily as the interest rate shock is applied.

11

See footnote 9.

12

ICE-BofAML Denmark Government index.

13

To analyze the impact of interest rates shocks on equity prices, we employ a dynamic multi stage dividend discount model see Annex 1 and Inkinen et al, 2010, Panigirtzoglou and Scammell, 2001, and Fuller and Hsia, 1984.

14

ERP are not directly observable and their estimates are inherently uncertain and dependent on the model specifications and the data series chosen. For example, the time horizon of EPS forecasts used (particularly where analyst forecasts are drawn from a small sample, as in Denmark). In addition, the calibration of the model’s settings can also have an impact (e.g. assumed transition path to steady state).

15

As captured by analyst projected earnings growth estimates 12- and 18-months ahead, source: Reuters I/B/E/S.

16

Financial assets in ALS (2015) are defined to include: deposits, mutual funds, shares, voluntary private pensions, whole life insurance and other financial assets (excluding business assets). The estimates control for many factors, including age (i.e. position in life-cycle), and also: income expectations, work status, education level, household composition, credit constraints, unemployment episodes and sick leaves.

17

This dataset combines the French Wealth Survey (INSEE) and the Household Budget Survey (INSEE-Eurostat).

18

That is, the 6–8 percent and 9–15 percent fall in financial assets resulting from, a 100 basis and 200 points interest rate shocks, respectively.

19

Consumption elasticities with respect to asset changes for the ten income deciles are {5, 4.5, 4, 3.5, 3, 2.2, 2.2, 2.4, 2.4, 0.9} percent for the 1st, 2nd,…, 10th deciles in line with Table 4, column 4 of Arrondel et. al. (2015). Consumption elasticities with respect to asset changes for the loan-to-value and age cohorts are taken constant at 3 percent because average income does not vary significantly among them. Finally, consumption elasticities for the 6 loan-to- income cohorts are {2, 2.5, 3, 4, 5, 6} percent (from the cohort with LTI < 100 percent to LTI ≥ 500 percent), based on calibration of the cohorts using income information.

20

Therefore the red dashed line of Figure 5, panel C is the same as the levels of the median (“50”) show in the box- plots of the right panel of Figure 4.

21

Borrowers’ mortgage debt is expected to fall by 7 percent.

1

Specifically, the ERP is the extra compensation that investors require to make them indifferent at the margin between holding the risky market portfolio and a risk-free bond. But because this compensation depends on the future performance of stocks, the ERP incorporates expectations of future stock market returns, which are not directly observable (but can be inferred using models).

2

During the life of the assets.

3

g which is pinned down by a long-term restriction which simply imposes the return on equity to be equal to the cost of equity. The duration of the transitional phase towards steady state may be up to 7–8 years, depending on the calibration chosen. This long- term equilibrium is pinned down by a LT restriction which simply imposes the return on equity to be equal to the cost of equity.

1

The Danish FSA characterizes the regulatory purpose of the supervisory diamond for MCIs as prudential due to its implementation, but in a broader context it functions as a macroprudential tool by applying to all systemic credit institutions.

2

Secondary homes and vacation properties have lower LTV limits. LTV limits were raised for vacation properties from 60 percent to 75 percent in 2017 to promote a broader national housing recovery following the housing bust.

3

The designated systemically-important institutions are Danske Bank, Nykredit, Jyske Bank, Sydbank, Nordea Kredit, and DLR Kredit.

4

The 2014 Denmark FSAP details the requirements under each regulation and the changes resulted from the transposition of the CRD IV framework.

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

    Stylized Presentation of Private Sector Financial Transactions

  • View in gallery View in gallery

    Private Sector Financial Balance Sheets and Comparison

  • View in gallery

    Monetary Financial Institutions Credit to the Private Sector

    (DKK, billion)

  • View in gallery

    Household Leverage and Consumption

    Percent

  • View in gallery

    Housing Cost Overburden by Income Group, 2016 1/

    (Percentage)

  • View in gallery

    Apartments and Houses 1/

    (Percent)

  • View in gallery

    Households, Credit and Leverage

  • View in gallery

    Pension Funds and Life Insurance Companies: Assets

    (Percent of total assets)

  • View in gallery

    Interest Rate Shock: Anatomy of the Impact on Consumption via Wealth Effects

    (Percent; based on data from the 2015 annual registry of households)

  • View in gallery

    Anatomy of Impact on Consumption, Introducing Liabilities Explicitly

    (Percent; based on median household in each cohort from 2015 annual registry household data)

  • View in gallery

    Share of Households with LTI Greater than Four

    (Percent of total new mortgage lending)

  • View in gallery

    Supervisory Diamond for Mortgage Credit Institutions