Luxembourg: Selected Issues
Author: International Monetary Fund. European Dept.
Publication Date:
03 Apr 2018
eISBN:
9781484350041
Language:
English
Keywords:
ISCR; CR; worker; participation rate; work; RRE price; latent demand; capital gain; land price; interest deductibility; Mortgages; Real estate prices; Income; Global
Download PDF (581.5 KB)

Selected Issues

Abstract

Selected Issues

Housing Market: Assessment and Policy Recommendations1

Luxembourg’s housing market has experienced continuously rising real estate prices which pose affordability problems and could lead to excessive indebtedness of some households as demand growth driven by demographic growth has outstripped new construction of dwellings. This affordability problem appears significant in a cross-country perspective. However, house prices remain broadly aligned with fundamentals, and the key issue to address is a lack of supply. Banks appear well capitalized and would be expected to withstand a partial equilibrium shock in the residential real estate market. However, household debt appears high, in particular among younger and middle-income households. Normalization of monetary policy could exacerbate the debt burden of some households and dent banks’ margins. To increase the housing supply, policies should reduce rigidities in land availability for construction, zoning restrictions, and increase means-testing for the various tax expenditures and subsidies that support demand.

A. Introduction

1. This chapter assesses the Luxembourg housing market from the point of view of affordability and suggests policy responses.2 The residential real estate market has experienced a growing affordability problem arising from a lack of supply in the context of a fast-growing economy and net demographic growth. High house prices also create financial stability risks, via the balance sheets of households or of banks if Luxembourg were to be affected by a severe negative shock.

2. The chapter is organized as follows. First, it provides a comparative stylized analysis of the housing market in Luxembourg and points at supply constraints; second, it presents the assessment of residential real estate (RRE) prices and concludes that they do not appear to deviate from fundamentals; third, it provides an analysis of financial stability risks; fourth, it elaborates on specificities in the Luxembourg housing market. The last section concludes.

B. Stylized Facts

3. In recent years, a significant affordability problem has become increasingly visible in the Luxembourg RRE market. In 2017:Q2, the price-to-income ratio stood at 45 percent above its 1990–2017 average, the second highest deviation from historical average among a group of OECD countries. Since the global financial crisis, Luxembourg’s rapid RRE price growth has ranked the 9th highest among 25 OECD countries. The affordability problem is illustrated by the striking divergence between income per capita and house prices since the early 2000s: while GNP/capita is only 1.1 percent higher than in 2002, house prices are on average 85 percent higher. This divergence is the consequence of both net demographic growth and a growing share of GDP accruing to cross-border commuters, in the context of a sustained growth of demand for labor. The rental market has not kept up with the increase in housing prices: in 2017:Q2, the price-to-rent ratio stood very significantly above its historical average, by some 45 percent. The cash flows on housing investments have declined, suggesting that such investments have become more and more driven by expected capital gains.3

Figure 1.
Figure 1.

Stylized Facts on Residential Real Estate Prices

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A002

4. The supply of dwellings has not kept up with a rapidly growing demand. Since 1990, resident population has increased by about 50 percent—or 200,000 inhabitants. In addition, 45 percent of the labor force (about 180,000 workers) are cross-border commuters. This creates a large latent demand which contributes to sustaining high residential real estate prices. Transactions have risen steadily since the global financial crisis, pushing up prices both for houses and apartments.

Figure 2.
Figure 2.

Demand Pressures and Transactions

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A002

C. Are Residential Real Estate Prices Aligned with Fundamentals?

5. We develop a dynamic empirical model of RRE prices that aims at capturing Luxembourg specificities (Box I). We perform a time series regression analysis of the dynamics of RRE prices over the period 1985:Q1-2016:Q1. We explain the growth rate of house prices by structural considerations, which include: (i) a measure of housing investment (the volume of housing in approved construction permits); (ii) a measure of structural demand related to the dynamics of population (net migration flows); and (iii) a measure of affordability (the price-to-income ratio). We also include as explanatory variables indicators of a more cyclical nature such as: (i) the growth rate of construction costs; (ii) real mortgage interest rates; (iii) mortgage credit growth, and (iv) the growth rate of residential permits (which captures the impact of non-linearities in the rate of new residential construction).

6. We find that structural supply and demand considerations as well as more cyclical factors help explain the dynamics of RRE prices. RRE prices respond positively to demand pressures related to demographic growth, and negatively to increases in the supply of dwellings. While the variable capturing affordability also plays a role in explaining RRE prices in a bivariate model, its sign becomes statistically insignificant once the two supply and demand variables are included in the regression. RRE prices are also influenced positively from the supply side by the costs of construction as expected, and negatively by mortgage interest rates which have a small cyclical effect on demand. Last, an increase in dwelling construction permits has a smaller negative impact on house prices when the flow of permits is already high (as shown by the positive coefficient on the growth rate of construction permits).

Determinants of Real House Price Growth (YoY)

article image
Note: Newey West estimator with 4 lags autocorrelated residuals, ***p<0.01, **p<0.05, *p<0.1Sample period: 1985:Q3-2016:Q1.

Deflated by CPI.

7. A simple quantification shows that the economic effects of structural and cyclical supply and demand considerations are significant. Based on the regression reported in the Table above, column (5), we find that, considering structural factors, a one standard deviation shock to residential permits (respectively net demographic growth) is associated with a 4.3 percentage point decline (respectively 3.2 percentage point increase) in real house price growth. These are large effects considering that the standard deviation of RRE price growth is 4.9 percentage points in our sample. Cyclical factors such as construction costs and real interest rates also have economically meaningful effects on house prices: one standard deviation increase in these two variables are associated respectively with a 3 percentage points increase and 1.5 percentage point decrease in RRE price growth, though the effect of interest rates is not always significant.

Quantification

article image

8. Our model suggests that real house prices are broadly in line with fundamentals. The predicted real house price index is computed as the index of house prices predicted by the regression coefficients under the assumption that the predicted index takes the same value as the historical average of the actual index during 1985–2016. At any point in time, the percent deviation between the actual house price index and the predicted house price index measures the overvaluation (if the deviation is positive) or the undervaluation (if the deviation is negative). We compute this deviation for the regression models 2–5 reported in the above table. We find evidence that real house prices were quite significantly overvalued (by about 15 percent) in the years leading to the global financial crisis and that since then prices have gradually converged to their expected levels consistent with fundamentals.

A02ufig1

Real House Price Valuation

(Percent deviation of actual prices from equilibrium levels)

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A002

Note: deviation normalized by assuming no misalignment on averageover 1990-2016.Source: IMF Staff calculations.

9. Illustrative scenario analysis suggests that changes in the supply of dwellings or in population flows would significantly affect the dynamics of RRE prices. We consider two scenarios: (i) an increase in the volume of approved residential construction permits of 10 percent relative to the 2012–16 average; and (ii) a 50 percent decline in average net migration flows relative to the 2012–16 average.4 In the first scenario, annual house prices would have grown at a rate about 1.8 percentage points lower relative to the 2012–16 average of close to 5 percent. In the second scenario, a return of net population flows to their 2002–06 average would result in a decline of annual house price growth by about 2 percentage points.

A02ufig2

Illustrative Impact of Housing Supply and Net Migration on Annual Real House Price Growth

(Percent)

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A002

1/ Net migration at about 2002–2006 levelsSource: IMF staff calculations.

A Model of Residential Real Estate Prices for Luxembourg

We estimate an empirical model linking house price growth to a set of structural and cyclical fundamentals estimated at a quarterly frequency over the period 1985:Q1–2016:Q1:

ΔRREt=φ+Σkβk⋅ΔXkt−1+Σνγν⋅Yνt−l+ɛt(1)

Where ΔRREt is the year-on-year growth rate of the real house price index, ΔXkt–1 is a vector of cyclical variables, Yvt–l is a vector of structural variables (lagged l times), εt is a residual with an autocorrelated lag structure with up to 4 lags. Cyclical determinants of real house prices include the growth rate of construction costs, the average real interest rate on [new] mortgages, and, in some specifications, the lagged growth rate of construction permits (which could for instance capture the speed at which administrative constraints are addressed) and the growth rate of mortgage credit. Structural factors include the log of the price-to-income ratio (a measure of affordability), lagged one year, a proxy for the growth of the supply of dwellings and a proxy for demographic determinants of housing demand. The proxy of residential investment is the log of residential permits issued, lagged one year, and the proxy for demographic factors is the log of net migration. The empirical model also includes a dummy variable equal to one in 1998/1999 to capture any potential breaks related the introduction of the euro.

The underlying “long-term” model of real house prices RREt is given by a relationship between real house prices, the stock of dwellings Kt, the stock of population POPt and an affordability indicator AFFt:

⁢RREt=α+β⋅Kt+γ⋅POPt+δ⋅AFFt+μt(2)

Expressed in log first difference and using building permits and net migration as proxy respectively for the net capital stock and population change, we obtain the following relation:

ΔRREt≈α+β⋅ΔKt+γ⋅ΔPOPt+δ⋅AFFt+μt(3)

The model builds on an earlier model developed in the 2017 Luxembourg FSAP Technical Note “Macroprudential Framework and Policies” by adding structural indicators of the stock of housing and of demand for housing, which, with the exception of affordability indicators, were often not included in the earlier literature (see for instance Loungani and Igan (2012)). Adding such indicators appears quite relevant in the case of Luxembourg where the stock of dwellings is relatively new and expanding, and where net migration has been very significant in recent decades. Several recent cross-country papers such as Caldera Sanchez and Johansson (2011) and Arestis and Gonzalez (2013) also consider similar supply and demand considerations and assess the relationship between housing investment and price responses among OECD countries. In the case of Luxembourg, and closest to our approach, Filipe (2017) develops a vector error correction model with co-integration equations between house prices, mortgage credit, building permits, construction costs, mortgage rates and real GDP per capita. In contrast to her paper, in our model migration flows and building permits are associated with the growth rate of house prices instead of their level.

D. Financial Stability Considerations

10. Mortgage credit has grown steadily in the past decade. Despite some volatility of new mortgage production, there is no evidence of an acceleration in the production of new mortgages which seems to have grown at a broadly steady rate since the early 2000s. The growth of mortgage credit seems to have been driven by lending from the domestically oriented banks to resident households, but other banks serving an international clientele have also increased their supply of mortgages for non-resident borrowers, often as part of packages of services related to private banking and wealth management activities.

Figure 3.
Figure 3.

Mortgage Credit in Luxembourg

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A002

11. The authorities have taken measures to strengthen the resilience of banks to potential shocks from their RRE exposures. Exposures of domestically oriented banks to residential real estate has risen very significantly since the global financial crisis. Since 2012, the Commission de Surveillance du Secteur Financier (CSSF) has put in place the following measures: adjusted risk-weights of 75 percent under the standardized approach for the part of new mortgage loans that exceeds a loan-to-value (LTV) ratio of 80 percent instead of the standard 35 percent risk-weight; stricter stress test requirements for IRB banks’ mortgage books, and Pillar II capital add-ons for banks with main exposures to real estate due to “concentration risk.” In addition, three credit institutions, operating in the residential real estate sector and maintaining major mortgage books, have been identified as systemically important with the application of additional capital requirements of 0.5 percent of RWAs. On August 31, 2016, and following a recommendation of the Systemic Risk Committee (CdRS), IRB banks have been required to apply a risk weight floor of 15 percent for exposures to Luxembourg residential real estate.5

A02ufig3

Mortgage Loans of Domestic Banks

(Percent of total assets)

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A002

Sources: Haver Analytics and IMF staff calculations.

12. Household debt has on average reached high levels in relation to disposable income, but financial assets are also high. At an aggregate level, household debt reached 175 percent of disposable income in 2016 placing Luxembourg above the average of about 150 percent among a group of OECD countries. However, at an aggregate level, this debt appears to be well covered by the liquid financial assets of households: according to the sectoral financial accounts of Luxembourg, total savings and deposit accounts of households reached €41.5 billion in 2017:Q3 compared with an aggregate stock of debt of €36.9 billion.6 Moreover, ability to repay appears good for the median household, at least until 2014: according to the 2014 Luxembourg Household Finance and Consumption Survey, the median debt service-to-income ratio stood at 14.8 percent and the median LTV ratio stood at 34.6 percent.

A02ufig4

Household Debt, 2016 or Latest Value

(Percent of disposable income)

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A002

Source: OECD.

13. However, medium-term vulnerabilities resulting from high indebtedness appear very clearly among younger household cohorts, and the share of high LTV mortgage loans has risen. According to the 2014 Luxembourg Household Finance and Consumption Survey, younger household cohorts (of age below 45) had very high debt-to-income ratios with a median approaching 200 percent.7 If households are classified by income, indebtedness appears to be larger among middle income households, while lower income and to some extent higher income households seem less indebted. The normalization of monetary policy in coming years would expose some households to interest rate risks and increase their debt burden, as about 50 percent of mortgages remain at adjustable rates.

A02ufig5

Median Debt-to-Income Ratio

(Percent)

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A002

Source: Giordana, G. and M. Ziegelmeyer (2017).

14. In addition, in recent years the share of potentially risky mortgages has become large in the new production. While the average LTV ratio on the stock of existing mortgages is below 50 percent, about 25 percent of mortgages have a LTV ratio equal to or above 80 percent, suggesting that some pockets of the market could be sensitive to price declines coupled with interest rate increases. In recent years, about 40 percent of new residential mortgage loans had an LTV ratio about 80 percent, well above the 25 percent of such loans among the outstanding stock of mortgages. Thus, in the past decade banks have increasingly offered loans that are less covered by the value of the property. Moreover, the growing share of fixed rate mortgages in the new loan production relative to historical practices has exposed banks to some compression of margins that would materialize when monetary policy normalizes. However, annual partial equilibrium stress tests by the CSSF suggests that banks thus far have generally solid buffers and could absorb a shock that would arise in the real estate sector.

E. Special Features of the Luxembourg Real Estate Market

15. The structure of real estate development highlights a preference for low-density housing. The residential real estate stock mainly consists of single-family houses and small-size apartment buildings. As of March 2017, half of the 233.7 thousand dwelling inventory consists of single-family houses and only 35 percent of the dwellings are multi-family buildings. Apartment buildings rarely encompass more than 10 apartments—more than half of multi-family units have three apartments or less.

A02ufig6

Apartment buildings by size, 2016

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A002

Source: STATEC, July 2017.

16. The gap in the supply of dwellings resulting from demand pressures is estimated to be significant and will continue to grow into the medium-term under current policies. The number of new dwellings completed averaged about 2,700 new units every year since 2000. Given the growth in resident population, this suggest an average gap of about 3,000–3,500 new dwellings per year—amounting to a cumulated gap in new housing estimated in the range of 54,000–63,000 since 2000. Projections by STATEC completed in 2010 found that about 6,500 dwellings are annually needed by 2030 to absorb population growth, given the demographic assumptions based on 2005–10 trends.8 Moreover, the large contingent of daily commuters from neighboring countries creates an additional potentially large latent demand.

17. Land available for new construction is sparse. Total land opened for new construction in Luxembourg is estimated at about 2,700 hectares (about 10.4 square miles or 1 percent of Luxembourg’s total territory). However, new construction since 2000 concentrated on a few areas—a quarter of all new housing was built in or in the immediate vicinity of Luxembourg-City. In those areas, land available for housing construction is now scarce, with vacancy rates for housing construction at a low 10–15 percent range.9 Outside these high-demand areas, vacant land is relatively abundant. Three-quarter of municipalities in Luxembourg used their available land sparingly, dedicating in average 2.5 hectares for new housing construction, and still have a large contingent of available land. This reflects tight distributional issues in land allocation, and municipalities’ diverse responses to higher demand for new housing.

A02ufig7

Luxembourg: New Dwellings and Population Growth — Estimating The Housing Gap

(In thousand units)

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A002

1/ Based on the assumption that the rate of occupancy per dwelling stays constant at its level of end-March 2017.Sources: STATEC data and IMF staff estimates.

18. Zoning is complex and fragmented, and municipalities have latitude to decide on new construction. The Luxembourg territory is divided in 102 municipalities, which have to establish a land-use plan (plan d’aménagement general, PAG) and so define the building perimeter of land for constructed areas. This land-use plan must be elaborated in respect of the Master Program for Spatial Planning (Programme Directeur d’Aménagement du Territoire, PDAT), and the sectoral plans. The PDAT encapsulates the government’s long-term general objectives and priorities. These are detailed and executed through sectoral plans. One sectoral plan will be dedicated to housing issues. However, sectoral plans only set general guidelines and broad principles, leaving wide scope for interpretation and implementation. Moreover, inter-municipal multi-level cooperation to foster regional growth does not create binding rules and government authorities cannot compel municipalities into action. 10

19. Construction relies on complex norms and requirements that tend to increase building costs. Construction norms are bound by municipality rules: The municipal land-use plan (PAG) identifies areas reserved for low density or high density housing, mixed areas (open to housing and/or complementary functions such as commercial, administrative, or trade), public facilities, areas for economic activity, and areas that are to remain free, dedicated for agricultural, recreational use and nature. Furthermore, the build-up areas are stated precisely with some density factors such as housing density, or norms for parking lots for example. The detailed urban development plan (plan d’aménagement particulier, PAP) specifies the land-use plan by precise urban development requirements such as the layout of the residential allotments, building heights or the implementation of the buildings. The cost of new housing developments has been increasing under cumulative requirements for the use of space and energy-efficiency. For example, each new dwelling must be supplied with one to two created parking spaces on private land—which often requires expensive digging for underground parking. Nevertheless, land-use and urban development regulations were reordered in 2011 to simplify urban planning and their adoption procedures were shortened in March 2017. The results of this administrative facilitation become only visible when the municipalities have reviewed their land-use plans (deadline August 2018).

20. Most of the available land is in the hands of private owners and difficult to mobilize. 92 percent of the 2,700 hectares available for construction belong to private owners; public developers can mobilize only 0.7 percent of it. While municipalities are allowed to discourage land hoarding by levying a tax surcharge on unused land, only a few municipalities have done this so far, suggesting insufficient tools to effectively mobilize new land for housing development. Conscious of the difficulties in mobilizing construction land the Ministry of Interior launched the so-called site development contract “Baulandvertrag” in a draft law. This contract is concluded between the owner of the land and the municipality. The objective is to guarantee a rapid development of the construction site after the land was converted from “reserve development land” into “active development land or building ground.” If the owner does not realize the servicing infrastructure in a time-lapse of 3 years, either the land will be returned to “reserve land” or the municipality could purchase the active development land.

A02ufig8

Dealing with Construction Permits (Doing Business Report)1/

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A002

1/ Data between square brackets indicate the average number of procedures.2/ The Gap-to-Frontier indicator is the distance to best performer.Source: World Bank (June 2017) and IMF Staff Calculations.

21. The process of obtaining building permit is not sufficiently fluid, although recent legislation would help streamline the process. Luxembourg ranks seventh worldwide for the ease in obtaining a construction permit. However, the number of procedures is comparatively high and requires on average 157 days to complete, reflecting complexity in obtaining the various required authorizations.

22. The share of social housing is very low compared to other European countries. Affordable housing is scarce, not least because of insufficient new supply in rental social housing, whose inventory is below 1 percent of total housing.

A02ufig9

Share of Social Rent in Total Housing

(Percent)

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A002

1/ Sweden: Share of public rent.Source: The State of Housing in the EU, Housing Europe Review, October 2017.

23. Public developers have a mixed mandate—supply both affordable real estate and low-cost rental social housing. Social housing is managed through waiting lists, attributed in the context of poverty reduction measures, and supplied mainly through three public developers. The largest one, the Fonds du Logement constructs, renovates, and manages about 2,000 dwellings. The Société Nationale des Habitations à Bon Marché (National Affordable Housing Company), formed as a joint stock company, has a broad mandate comprising new housing, new regional development plans, and completing public infrastructure. Its main activity is to provide real estate for sale below market prices to low-income qualifying households, but it also has a small inventory of rental dwellings. Municipalities—the third largest group of developers for social housing—may elect to provide social rentals, but tend to prefer increasing the share of owner-occupied housing and are reluctant to devote costly technical and administrative resources to manage social rental housing. The private supply of subsidized housing was sparse, as the requirement to develop 10 percent of the projected dwellings in any urban development plan with a surface of more than 1hectare was not effective enough. Therefore, this action has been reinforced in March of 2017 as 10 percent of the total square meters build in each urban development plan, counting more than 25 dwellings, must be affected by subsidized housing and the municipalities could even determine the price of these units. The effect of this recent adaptation is not visible for the moment.

Table 1.

Luxembourg: Rental Social Housing in Luxembourg—New Construction by Public Developers, 2015–18

article image
Sources: Ministry of Housing (Ministère du Logement); Annual Report SNHBM; and IMF staff calculations.

24. Innovative ownership forms have started to ease new supply and help improve affordability. The Ministry of Housing identified that about a third of the available land for housing (an aggregate 1,000 hectares or about 3.8 square miles) can be mobilized through emphyteutic leases (Baulückenprogramm).11 However, new housing supply associated with these regulations has so far been limited. Municipalities’ tax resources primarily stem from corporations operating on their territory, thereby creating a preference for commercial real estate at the local level. Businesses provide the lion’s share of local governments’ tax revenue (about 92 percent in 2016). The main tax is the local municipal tax on trade (rate at an average 7.5 percent on net profits), and the property tax (Impôt foncier).

A02ufig10

Luxembourg: Local Government Tax Revenue by Source

(Percent of GDP)

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A002

1/ Includes Taxes on income, profits and capital gains of corporates.Sources: OECD; and IMF staff calculations.

25. There has been demand to expand commercial real estate. The development of commercial real estate tends to yield more tax income for municipalities than residential real estate. It also tends to bear less stringent administrative and infrastructure constraints than housing. To promote Luxembourg as a commercial center in the Greater Region, the moratorium on large-scale retail areas was lifted in 2005, thereby creating large mall areas and shopping outlets.12 By 2010, more than 2,000 square hectares of large-scale retail areas had been approved. Moreover, municipalities are entitled to steer zoning rules on lots meant for mixed purposes (commercial and residential) mainly towards commercial development.13 Anecdotal evidence indicates that in certain cases dwellings in Luxembourg-City were converted to office space, given tight vacancy rates in commercial real estate.

26. In very rare cases, the reform of municipality funding at end-2016 could tend to reduce the bias in favor of commercial real estate. One of the main tax resource of municipalities is the Impôt commercial Communal, (municipal tax on trade, ICC) a local corporate income tax levied on businesses, which supplies a quarter of Luxembourg CIT revenue. The December 2016 reform of local finances mandates that municipalities keep a maximum of 35 percent of their ICC revenue, with the remainder redistributed among municipalities.14

27. Various tax mechanisms which are mostly not means-tested tend to stimulate demand for home-ownership. (Table 2) Home ownership in Luxembourg creates a strong tax shield and has benefited from direct subsidies, proportional to the taxable income and house value. Various fiscal support mechanisms are available to reduce closing costs for owner-occupied home purchases and interest payments are partly tax deductible. In the case of sale of real estate used as primary residence, the capital gain is not taxable, and there are reduced taxes on capital gains for other real estate. 15

Table 2.

Luxembourg: Taxation Measures Supporting Residential Real Estate Investment

article image
Sources: Observatoire de l’Habitat (2018)

28. Tax deductibility of mortgage debt payments is encouraging higher household leverage. The tax treatment of mortgage loans may encourage high household debt, as demonstrated by recent literature.16 Mortgage interest payments on the primary residence are partly tax deductible. Thus, it may provide a powerful incentive to use housing investment as a tool to buildup household assets through leverage. In comparison with European peers, where mortgage interest deductibility was reduced over the past decade, the Luxembourgish regime maintains a generous incentive.

29. Low property taxes associated with cadastral bases unrevised in the past 70 years point to the misallocation of land resources. Property taxes in Luxembourg are among the lowest in Europe. This tends to favor investment in housing over financial assets. In a context of rising land prices and scarce constructible land, a low tax rate on real estate may encourage owners to hold on to their land. Hence, sparse revision in tax bases combined with low rates may reduce the elasticity of housing supply to price developments. The temporary reduction in the tax on capital gains for real estate sales by end-2018 may nevertheless encourage sales in constructible land.

A02ufig11

Taxes on Land, Buildings, and Other Structures, 2016

(Percent)

Citation: IMF Staff Country Reports 2018, 097; 10.5089/9781484350041.002.A002

Source: Eurosat.

F. Conclusions and Policy Recommendations

30. Demand for housing has exceeded supply for many years. While house prices are in line with fundamentals, they have risen faster than disposable income for years, largely because of structural supply constraints in the context of strong demand, in part reflecting net demographic growth. The dynamics of house prices is also somewhat affected by cyclical factors such as the cost of construction and to some extent the low interest rate environment. Rigid zoning and administrative rules together with land hoarding prevent sufficient construction, while tax incentives and subsidies fuel demand. Reduced affordability has driven up household indebtedness, in particular among younger households.

31. Risks in the real estate market should continue to be closely monitored, and further actions taken as needed. Recent measures have appropriately built capital buffers in the banking system while discouraging riskier lending. However, household debt is relatively high and limits to debt-service-to-income ratios should be set if house prices continue to outpace disposable incomes. Going forward, the normalization of interest rates could add to the debt service of some households (who borrowed at variable rates) while banks’ margins on their stock of fixed rate mortgages would shrink.

32. Containing house price pressures and alleviating bottlenecks of housing require a strong effort to expand the stock of housing:

  • Excessive red tape in bringing additional land to construction should be pruned, and incentives strengthened. The initiatives of Baulücken for new construction are a step in the right direction;

  • Local zoning decisions should be better coordinated with a national spatial development plan and cooperation among municipalities should be encouraged;

  • Existing tools to mobilize vacant land and unoccupied dwellings could be strengthened. This includes implementing taxation on vacant lots. In this respect, the initiative of Baulandvertrag goes in the right direction; 17

  • In the PDAT and the municipal implementation, assigning “mixed construction” land in priority to residential real estate would widen the share of land eligible for housing development;

  • Tax biases at the municipality level against residential real estate should be reduced further. The reform of the distribution of municipal business taxes among municipalities is a step in the right direction as it reduces incentives favoring commercial over residential real estate zoning decisions. Going forward, policies should increase the share of the ICC redistributed in the equalization fund;

  • Increasing property taxes and revising cadastral values would help municipalities increase own resources.

33. The share of social and affordable housing in total housing could be increased:

  • To encourage social housing in the rental segment, public developers in the social sector (FSH, SNCHM, and municipalities) should be gradually steered only towards the development and management of social rentals. This would help clarify management roles and separate more clearly the rental activity from the construction-for-sale business.

34. Tax expenditures and subsidies should become more means-tested to help address affordability issues for young and lower-income households, and first-time homebuyers. Tightening eligibility for real-estate related tax incentives and subsidies would also help alleviate some of the demand pressures on prices.

References

  • Arestis, Philip and Ana González, 2013, Modeling the Housing Market in OECD countries, Levy Economics Institute of Bard College Working Paper No. 764.

    • Search Google Scholar
    • Export Citation

  • Caldera Sánchez, A. and Å. Johansson, The Price Responsiveness of Housing Supply in OECD Countries, OECD Economics Department Working Papers, No. 837, OECD Publishing, 2011.

    • Search Google Scholar
    • Export Citation

  • European Mortgage Federation (EMF), Hypostat 2016: A Review of Europe’s Mortgage and Housing Markets, November 2016.

  • Fatice S., Prammer D., Housing and the Tax System: How Large Are the Distortions in the Euro Area?, ECB Working Paper Series, N.2087, July 2017.

    • Search Google Scholar
    • Export Citation

  • Fereira Filipe S., Housing Prices and Mortgage Credit in Luxembourg, Revue de Stabilité financière, Banque centrale du Luxembourg, 2017.

    • Search Google Scholar
    • Export Citation

  • Geng N., Fundamental drivers of House Prices in Advanced Economies: The Role of Policies and Structural Factors, forthcoming, IMF Working Paper, 2018.

    • Search Google Scholar
    • Export Citation

  • IMF, Housing finance and financial stability—Back to Basics?, April 2011.

  • IMF, Housing Recoveries: Cluster Report on Denmark, Ireland, Kingdom of the Netherlands—the Netherlands, and Spain, 15/1, 2015.

  • Loungani, Prakash, and Deniz Igan, 2012, Global Housing Cycles, IMF Working Paper WP/12/217

  • Observatoire de l’Habitat, La Note de l’Observatoire de l’Habitat, Le Potentiel Foncier Constructible Théorique au Luxembourg en 2007, Septembre 2009

  • OECD 2011 (a), OECD, Housing markets and Structural Policies in OECD countries, January 2011.

  • OECD, The Evolution of Homeownership Rates in Selected OECD Countries: Demographic and Public Policy Influences, January 2011.

  • OECD Economic survey: Luxembourg, 2017

  • Poterba, James, and Todd Sinai. 2008. Tax Expenditures for Owner-Occupied Housing: Deductions for Property Taxes and Mortgage Interest and the Exclusion of Imputed Rental Income. American Economic Review, 98(2): 84–89.

    • Search Google Scholar
    • Export Citation

  • STATEC, Regards 15 Sur les Bâtiments Achevés en 2013, December 2015

  • STATEC, PIBen-Etre—The Report, October 2017

1

Prepared by Michelle Hassine and Thierry Tressel (EUR).

2

This paper builds on the analysis presented in “Luxembourg: Financial Sector Assessment Program: Technical Note – Macroprudential Framework and Policies”, published August 28, 2017, International Monetary Fund.

3

Rental income is capped at 5 percent of the purchase value of the dwelling but tax deductions stimulate the return on investment during the first five years after purchase. Rents can also be revalued when tenants change.

4

This broadly corresponds to a decline in average net migration back to the 2002–06 average.

5

Since 2014, all credit institutions in Luxembourg are subject to a (fully-loaded) capital conservation buffer of 2.5 percent.

6

BCL statistical table 05.08. The aggregate figures reported are an approximation based on the financial assets and liabilities of the aggregate accounts of households and non-profit organizations serving households.

7

Giordana, G. and M. Ziegelmeyer (2017), “Household debt burden and financial vulnerability in Luxembourg” Banque Centrale du Luxembourg Working Paper 113. However, low interest rates may have helped contain debt service.

9

For example, the percent of lots available for housing construction are 19 percent in Luxembourg-Ville, 9 percent in Esch-Sur-Alzette, 10 percent in Ettelbruck, 12 percent in Junglinster, 13 percent in Diekrich, and 15 percent in Mersch.

10

Municipalities predate the formation of the Luxembourg state in 1839 and have communal autonomy, which is enshrined in the Constitution. Municipalities have executive municipal councils (Schöffenräte) with large powers for local infrastructure and land use, including housing.

11

Baulücken were created in 2008 and are established on land belonging to municipalities which is constructible but not yet used. Under emphyteutic leases, municipalities retain the ownership of the land, while homeowners receive the right to use the land against an annual lease over 27 to 99 years and can sell it back.

12

The Greater Region has 11.2 million people and covers 25.2 thousand square miles and includes, besides the Grand Duchy of Luxembourg, Wallonia and the German-speaking community in Belgium, Saarland and Rhineland-Palatinate in Germany, and Lorraine in France.

13

See Affolderbach and Carr (2014), Blending Scales Of Governance: Land Use Policies And Practices In The Small State Of Luxembourg, https://orbilu.uni.lu/bitstream/10993/15357/1/Affolderbach_Carr_author%20pre-print_Regional%20Studies_2014.pdf

14

The distribution of the General Fund of Municipalities to municipalities observe the following key after distribution of a lump-sum to municipalities: adjusted population (82 percent), social demography (9–10 percent), adjusted territorial size (5 percent), number of wage-earners (3 percent), and number of social housing (1 percent). The Law is available from http://legilux.public.lu/eli/etat/leg/loi/2016/12/14/n1/jo.

15

The normal rate is half of the marginal personal income tax rate applicable; the reduced rate is a quarter of the relevant rate until end-2018. At the top marginal rate of 42 percent, the taxation rate on capital gains for real estate is 10.5 percent.

16

The seminal work of Poterba and Sinai (2008) highlights that household balance sheets are sensitive to the treatment of mortgage debt.

17

The Loi Omnibus in February 2017 created the Baulandvertrag contracts, which introduce a mandate for construction on certain vacant lots with a sanction on owners through a binding sale to municipalities.

Same Series

Other IMF Content

Other Publishers

Cover IMF Staff Country Reports
Luxembourg: Selected Issues
Author: International Monetary Fund. European Dept.
Volume/Issue:
Volume 2018: Issue 097
Publisher:
International Monetary Fund
ISBN:
9781484350041
ISSN:
1934-7685
Pages:
69
DOI:
https://doi.org/10.5089/9781484350041.002

Stylized Facts on Residential Real Estate Prices

Demand Pressures and Transactions

Real House Price Valuation

(Percent deviation of actual prices from equilibrium levels)

Illustrative Impact of Housing Supply and Net Migration on Annual Real House Price Growth

(Percent)

Mortgage Credit in Luxembourg

Mortgage Loans of Domestic Banks

(Percent of total assets)

Household Debt, 2016 or Latest Value

(Percent of disposable income)

Median Debt-to-Income Ratio

(Percent)

Apartment buildings by size, 2016

Luxembourg: New Dwellings and Population Growth — Estimating The Housing Gap

(In thousand units)

Dealing with Construction Permits (Doing Business Report)1/

Share of Social Rent in Total Housing

(Percent)

Luxembourg: Local Government Tax Revenue by Source

(Percent of GDP)

Taxes on Land, Buildings, and Other Structures, 2016

(Percent)