Haffner Winters Paper Optimal taxation of housing - Lirias - KU Leuven

Haffner Winters Paper Optimal taxation of housing - Lirias - KU Leuven

ENHR 2014 Conference ‘Beyond Globalisation: Remaking Housing Policy in a Complex World’ 30 June – 4 July 2014 Optimal housing taxation: Belgium compa...

179KB Sizes 0 Downloads 4 Views

ENHR 2014 Conference ‘Beyond Globalisation: Remaking Housing Policy in a Complex World’ 30 June – 4 July 2014

Optimal housing taxation: Belgium compared

Marietta Haffner1,2,4 and Sien Winters3,4 1

OTB – Research for the Built Environment, Faculty of Architecture and the Built Environment, Delft University of Technology, Jaffalaan 9, 2628 BX Delft, + 31-15-27-83005, [email protected] 2 Centre for Urban Studies, School of Global, Urban, and Social Studies, AHURI at RMIT University, Melbourne; and CCHPR, Department of Land Economy, Cambridge, UK 3 HIVA Research Institute for Work and Society, KU Leuven, Parkstraat 47, B-3000 Leuven, +31-16-323272, [email protected] 4 Steunpunt Wonen, Leuven

Abstract Belgium will devolve the responsibility and the budget for the favorable income tax treatment of owner-occupied dwellings to its administrative regions by 1 January 2015. This change would allow the regions (e.g. the Region of Flanders) to design their tax systems. This paper examines a reorientation according to the principles of the optimal tax theory based on the proposals of the Mirrlees Review. These principles imply that as little distortion in the behaviour of tax payers as possible should be achieved in order to minimize welfare losses to society. The principles are concerned with realizing a progressive and neutral tax system implying that the tax system should be considered to operate as a whole system; it should seek neutrality between types of activities and groups of people in order to avoid undesired behaviour; and it should achieve progressive taxation as efficiently as possible. This paper uses these principles as benchmark to determine the extent of subsidization of housing of private person owners of dwellings via the tax system in Belgium. The analyses of the tax system cover taxes at the point of acquisition (stamp duty, VAT), during the period of occupation/ownership (income and property taxes) and on sale of the dwelling (capital gains tax). This exercise is also carried out for four other countries (Denmark, Germany, the Netherlands and United Kingdom) in order to test for the use of theoretical principles in tax systems. Based on the empirics in the five countries it is concluded that none of countries come close to optimal taxation in housing. Based on theory and empirics it is analysed how the Belgian tax system and more specifically the future Flemish taxation could move closer to the optimal one. Keywords

Income tax, Mirrlees Review, Neutrality


Optimal housing taxation: Belgium compared Introduction The state reform of the 1980s turned Belgium from a centralized state into a federal state with three Administrative Regions: Flanders, the Walloon Provinces and Brussels. The regions are responsible for their own housing policies. From the tax policies only the transaction taxes and the tax on immovable property that allows the regions, provinces and municipalities to levy a surcharge was turned over to the regions. In the federal coalition agreement of December 2011 it was agreed that Belgium will devolve the responsibility and the budget for the favorable income tax treatment of owner-occupied dwellings to its regions. On 27 November 2013, the House and Senate approved the necessary changes to the Special Finance Act (governing the financial relations between the Federal State and the regions). On 22 April 2014 the changes to the Income Tax Code of 1992 were approved. These changes allow the regions to design their tax policies in relation to owner-occupied housing from 1 January 2015 onwards. This paper aims to examine the possibilities that the Administrative Regions have to design their new policy options. First, this paper examines a reorientation according to the principles of the optimal tax theory based on the proposals of the Mirrlees Review, a recent and esteemed review of the English tax system. These principles imply that as little distortion in the behaviour of tax payers as possible should be achieved in order to minimize welfare losses to society. Next, these theoretical principles are put forward as a benchmark to determine the deviation from the benchmark of the actual tax system that is relevant for housing in Belgium. The analyses of the tax system cover taxes at the point of acquisition (stamp duty, VAT), during the period of occupation/ownership (income and property taxes) and on sale of the dwelling (capital gains tax). Where relevant, it will be the Flemish situation that is described, as this is the region in Belgium that this contribution will make recommendations. It is the region of the three with the largest share of population. As a second ‘benchmark’, the comparison of the tax treatment of housing in four other countries with a similar level of welfare (Denmark, Germany, the Netherlands and United Kingdom) is also carried out. This exercises is to mirror the extent of the use of theoretical principles in relation to housing in tax systems. Based on these comparisons a final discussion will evaluate in a qualitative way options that are open to the Flemish government as of 1 January 2015. Before we turn to taxation, a brief introduction of the tenures and owners of dwellings of housing of the countries under study is presented, as taxation rules may differ accordingly.

Housing tenures and owners of dwellings The tenure distribution between the five countries and Flanders differ quite a lot, as Figure 1 shows. Flanders (75%) and Belgium (68%) have the largest owner-occupied sector. The Netherlands (51%) leads in the share of mortgagor followed by Denmark (42%), Belgium (31%) and the UK (27%). Germany has the biggest private rental sector (43%). In the latter case one also finds a statistic of 2

60% if the publicly-owned and non-profit renting is included as private renting (Haffner et al., 2009; Scanlon and Kochan (eds.) 2011). Denmark is the country where the rental sector is quite evenly spread across social (19%) and private (20%) dwellings, as well as other (14%; including cooperatives). The Netherlands is the country that has the biggest social rented sector (34%); the UK (20%) and Denmark (19%) follow. The smallest ones can be found in Belgium and Flanders (6% each). The social stock in Germany that can be owned by any landlord and is thus not to be classified in the statistics as a separate category, has also an estimated small size (4% in 2013; Kofner, 2014). Germany can be regarded as a unique case, as it according to the statistics social renting is not linked to a specific type of owner; contrary to the other countries listed. In the private rental sector personlandlords dominate in most countries as biggest single group of landlords: Flanders (95%), Belgium (86%), UK (75%), Germany (62%) and the Netherlands (44%) (Scanlon and Kochan (eds.) 2011). Generally, organization landlords are thus easily outnumbered by the person-landlords. As owneroccupiers are also private persons, the dominant owners of dwellings are private persons in most countries. Their taxation in relation to the dwelling ownership is the focus of this paper.

Figure 1

Housing market tenures in Flanders, Belgium, Denmark, Germany, the Netherlands and the UK, different years

Flanders (2013) Belgium (2009/2008) Denmark (2009/2008) Germany (2011, no distinction according to mortgage) Netherlands (2009) United Kingdom (2004/2008) 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% With mortgage


Without mortgage

Social renting

Private renting


Housing Statistics 2010 for tenures in Belgium, Denmark and UK and EU-SILC for share of mortgagees; Grote Woononderzoeks 2013 for Flanders (Heylen and Winters, 2014), WoON 2009 for the Netherlands (authors’ calculations) and for Germany the 2011 Zensus (Kofner (2014) interpretation which classifies public and cooperative ownership of rental dwellings as other).

Taxation of housing The extent of taxation of owners of dwellings will also be influenced by the consumption and investment characteristics of a dwelling (Poterba, 1984). From an investment point of view, the dwelling is an investment good that creates income by delivering housing services that can be consumed by the occupier. The occupier either pays rent when he is a tenant or pays rent fictitiously 3

to himself as owner in the case of owner-occupier. Therefore, from an investment point of view the owner of the dwelling will in theory have to pay (corporate or personal) income tax on the profits that the dwelling generates. This income from renting may be determined based on actual profits (rent income minus costs) or on imputed profits (Haffner et al., forthcoming). For owner-occupiers the income can only be determined as imputed income/rent, while costs can either be determined based on actual costs (big administration and control) or again based on imputed costs with the advantage of simplicity in administration. In practice, some countries choose the so-called consumption good option, where neither housing income nor housing costs are taken into account when taxable income is determined. Transaction taxes are the other type of tax that may be levied on in this case an ‘investment’ activity. It can be defined as rather a type of tax levied once-only on the transaction of a property. It exists in many member states of the European Union (European Commission, 2012a). According to Cnossen (2010: 83), “the rationale for these taxes, which have not been harmonized [in the European Union], is weak.” Such a tax is levied based on a transaction, which can be considered as an easy prey. The taxation of housing based on its consumption role is harmonized in the European Union with the Common VAT Directive of 2006 and which replaced the Sixth Council Directive of May 17, 1977. This Directive specifies the uniform basis for application of the Value Added Tax (VAT). In other words the goods and services that are to be consumed and on which VAT should be levied are designated. In principle, the VAT should be levied on consumption; in the case of housing it should be levied on the housing services (rent or impute rent) that are being consumed. As this exercise of having households pay a monthly VAT on their housing consumption would involve a huge amount of administration, most countries do not levy the VAT on housing services, but on the purchase of the dwelling: since the purchase price of a dwelling may be taken to represent the present discounted value of its future services, by extension the VAT on the purchase price may be considered a good proxy for the discounted value of the VAT that should have been levied on the flow of housing services. (Cnossen, 2010: 74; see also McDaniel and Surrey, 1985) Thus in theory this approach of the once-only VAT levy at acquisition can be regarded as a good approximation of the periodic (monthly) tax on consumption (rent or imputed rent). Cnossen (2010: 77) calls it no more than a “second-best substitute”, as revenues will be missed out on future changes in house value. Last, but not least, another popular tax on housing in the European Union is the periodic property tax which is connected to the location of the dwelling. It can be described as user charge or benefit charge (European Commission, 2012a, Wood and Ong, 2012). In theory, mobile consumers are able to choose the level of local services that they prefer. An alternative view is that property tax functions as a levy on capital or the consumption of capital embodied in the dwelling.


Optimal taxation according to the Mirrlees Review It has long been well-known among tax experts that all taxes other than a lump sum tax distort behaviour (Feldstein, 1976). The tax amount is not linked to ability to pay and can, therefore, not be changed by changing behaviour (Mirrlees et al., 2011a: 31). In all other cases, except the taxation of economic rent, taxes will influence behaviour. Therefore, the challenge for the design of an optimal tax system is to minimize inefficiency and to maximize welfare. This challenge was taken up by the Mirrlees Review for the UK and it was published in 2010 and 2011 (Mirrlees et al., 2010; Mirrlees et al., 2011a). The aim was to design a “good tax system” serving different masters at the same time (Mirrlees et al., 2011b: 332): The challenge for the review was to design a tax system that can raise the revenue that government needs to achieve its spending and distributional ambitions whilst minimizing economic and administrative inefficiency, keeping the system as simple and transparent as possible, and avoiding arbitrary tax differentiation across people and forms of economic activity (Mirrlees et al., 2011b: 332-333). The aim was in short to realize “a progressive, neutral tax system” implying that the tax system should be considered to operate as a whole system in which not all taxes need to address all objectives. An well-known example here is that the progressivity of the system would not imply that each tax has to be progressive. Furthermore, the tax system should seek neutrality between types of activities and groups of people in order to avoid undesired behaviour resulting in welfare losses. Last, but not least it should achieve progressive taxation as efficiently as possible (Mirrlees et al., 2011b: 334): This means having a rate schedule that reflects knowledge of the shape of the income distribution and the responsiveness of people to taxes and benefits at different income levels. It also implies taking decision over both whether to work (including when to retire) an how much work into account in addition to other responses such as tax avoidance and migration. As can be understood from the description above, there will not be one tax design that can be called optimal. The optimum will depend on the existing income distribution, the preferences of people, etc.

Optimal taxation and housing according to the Mirrlees Review In this section we explore what the Mirrlees Review has to say about the four types of taxation that are linked to housing and were identified earlier in this contribution: consumption tax (VAT), transaction tax (also known as registration tax or stamp duty), income tax and property tax. The focus is on private person owners of dwellings. The Mirrlees Review gives its conclusion specifically for the UK; therefore its recommendation may sound different for other starting situations. We will reflect on these possible deviations. The Mirrlees Review is clear about transaction taxes: there should be none, as taxes that are levied without the possibility to deduct costs or input costs are regarded as inefficient (Mirrlees et al., 5

2011b: 337). A well-known argument is related to the mobility of households which also impacts mobility on the labour market that will be discouraged by a high transaction tax (OECD, 2010a, b; Van Ewijk and Van Leuvensteijn (eds.) 2009). According to the Mirrlees Review the immobility will prevent people to choose housing according to their preferences. On the topic of indirect taxation, the Mirrlees Review furthermore recommends as part of a “good tax system” that the VAT should be: “A largely uniform VAT – with a small number of targeted exceptions on economic efficiency ground – and with equivalent taxes on financial services and housing” Agreement is wide in the literature on the non-suitability of VAT to realize redistribution of income of households via their final consumption expenditure, as Cnossen (2010: 36) formulates it: VAT can not be used to achieve vertical equity goals. Its main objective is to raise revenue as neutrally as possible. This requires the broadest possible base and a single rate. According to the Mirrlees Review, a broad tax base is preferred. It does, however, allow for a few targeted exception on economic efficiency grounds, such as to counter market failures (higher taxes for damaging goods). For the UK specifically, the Mirrlees Review recommends that the 0%-VAT for new construction, the stamp duty land tax (transaction tax), as well as the annual council (property) tax are traded in for an annual levy on housing which would be proportional to the value of the dwelling (Mirrlees et al., 2011b: 344). This type of taxation very much looks like a VAT-levy on the consumption of housing services. Presumably, in countries where the VAT-rate on housing construction is the normal rate and where no deduction on VAT tax base are allowed, the recommendation would not be to abolish it, but to maintain it. For the UK the introduction of a broad-based VAT (without the zero-rate and without all kinds of exemptions) could be linked with a compensation package for the most vulnerable households, if the package does not produce disincentives to work (Mirrlees et al., 2011b: 344). On income tax the Mirrlees Review proposes: Income from all sources should be taxed according to the same rate schedule… our approach would allow all costs of generating that income to be deducted… Saving and investment are [also] costs associated with generating future income… A deduction could be given each year for the opportunity cost of capital previously saved/invested. This is the rate-of-return allowance (RRA) treatment of saving… For assets where only the risk-free (‘normal’) rate of return is likely to be earned, this approach can be simplified, and returns on such assets can just be tax free. (Mirrlees et al., 2011b: 335-336). Clearly, the Mirrlees Review follows an investment good approach: income is taxed and costs that have to be made to generate that income are deductible from taxable income. The rate schedule that is referred to for income tax is described as: 6

When it comes to income taxation, there is a strong case for keeping things simple: single tax on income with an allowance and say two or three rates, combined with a single benefit to support those with low income and/or high needs. (Mirrlees et al., 2011b: 335). For housing these recommendations would imply no taxation on the normal rate of return. Only a tax on returns (capital income and capital gain) that are above the normal rate of return should be taxed in income tax. When no abnormal profits are made, no income tax will be paid. Then the question becomes what the recommendation would be for the property tax, as it is not mentioned in the general recommendations of the Mirrlees Review, except maybe as a periodic land value tax. But apparently the Mirrlees Review has large doubts that the practical difficulties in valuing the land can be overcome (p. 337). It is not listed in the final list of recommendations for consumers, only for business and agricultural land. On a property tax sec, possibly, the idea will be that if all other taxes are implemented (VAT, income tax on income from property, and a lifetime wealth transfer tax), a property tax may introduce welfare costs that outweigh its welfare benefits.After this presentation of the Mirrlees Review recommendations in relation to housing, the next sections are used to compare this benchmark to the actual taxation in Belgium/Flanders and four other countries. The other countries were chosen for different reasons, except that they all can be considered as having a comparable level of ‘Western-European’ welfare. As can be observed from Figure 1, Germany has the largest private rental sector. Maybe tax treatment plays a role here in maintaining its size. The UK has been included as the country that got the chance as of 2012 to implement the Mirrlees Review recommendations. Denmark was included as a country that levies a land tax. And the Netherlands was included as a country that implements the traditional investment good approach in income tax. From the four countries, two follow an investment good approach, the two other a consumption good approach.

Optimal and actual taxation of housing compared in five countries – indirect taxation No transaction tax and a broad-based uniform VAT are simply said the Mirrlees Review recommendations for an optimal tax design. As can be observed from Table 1 none of the countries chose to forego some kind of transaction tax. Some levy the tax on the transaction of existing dwellings and VAT on the added value in a newly-built dwelling (Belgium and Netherlands). Others levy the transaction on all transactions (Germany and UK). Belgium is the only country that allows for lower rates for affordable (and social) housing and allows also several exemptions, mainly confined to owner-occupied dwellings. A special measure is ‘portability’: transaction taxes paid on the purchase of a first dwelling can be credited against duties due on a second dwelling serving as principal residence.

On the VAT-levy, only the Netherlands seems to follow suit by levying the standard rate on added value in connection to new building of housing. Denmark and Germany seem to exempt certain transactions, while Belgium allows for a lower tax rate for social ownership and the UK allows for a zero-rate for all new construction. On repairs, the picture is mixed as well. Only Denmark and Germany do not seem to allow lower rates, while they seem to be applicable in the other countries under study.


For the UK, the Mirrlees Review formulated a special recommendation, as especially the abolition of the zero-rate VAT-levy on housing may be not such a sensible measure considering this starting position. But neither the VAT-levy has been abolished, nor has the stamp duty land tax been abolished, as suggested. Also the council tax has not been abolished and no new annual levy on housing has been introduced. However, the UK is not the only country that does not levy indirect taxes as recommended by the Mirrlees Review.

Table 1

Indirect taxation in Belgium/Flanders (2005), Denmark, Germany, the Netherlands (2009) and the UK and according to the Mirrlees Review

Mirrlees Review - general Actual taxation Belgium/Flanders

Denmark Germany Netherlands

Mirrlees Review - UK


Transaction tax on the acquisition o f

Broad-based, largely uniform VAT with equivalent taxes on housing

No transaction tax

Lower rate for social housing, social ownership and for repairs and renovation on dwellings of 5 years of older All added value; exemption possibly applicable to purchase of newly-built dwelling All added value; exemption for land and newly-built dwellings possible Lower tax rate for labour costs for repairs

Transaction with existing dwelling, but lower rate for affordable and social dwelling, several exemptions and portability Transaction with existing dwelling; possibly (?) also with newly-built dwelling All transactions

Transaction with existing dwelling

Replace the 0%-VAT for new construction, the stamp duty land tax (transaction tax) , as well as the council (property) tax with an annual levy on housing which would be proportional to the value of the dwelling

Actual taxation United Kingdom

Lower tax rates than standard: 0% for new construction; 5% for repair Source: Cnossen (2010); Haffner et al. (forthcoming)

All transactions

Optimal and actual taxation of housing compared in five countries – income taxation According to the Mirrlees Review income from all sources should be taxed the same way. Income taxation should allow for a deduction of all costs that are made in the process of generating the taxable income. It should be taxed using the standard income tax rates that also apply to other income. Table 2 gives an overview of income taxation for owner-occupiers and Table 3 for private person landlords in income tax. Table 2 shows clearly that none of the countries apply a rate-of-return allowance for the normal return to capital invested in the owner-occupied dwelling.


• •

The countries that apply a consumption good approach (Germany, UK) do not include the owneroccupied dwelling in income tax and can therefore be considered to apply a rate-of-return allowance in the situation where the mortgage loan has been repaid. However, other costs, e.g. for maintenance, are not taken into account then. In the countries that (still) use an investment good approach (Denmark, Netherlands), only a mortgage interest deduction is allowed (and possibly other costs), but no deduction of opportunity costs of equity investment. In Belgium which originally had an investment approach engrained in national income tax (with imputed rent taxation and cost deduction for debt interest, and additionally deductions for capital repayment and insurance fees), the imputed rent taxation has been exempted since 2005, while the cost deductions and other deductions have been remodelled into a new lump sum deduction called ‘Woonbonus’. This change in the tax relief system of 2005 nearly doubled the average benefit per household (Damen, Vastmans & Buyst, 2014; Heylen & Winters, 2012)

Table 2

Owner-occupiers’ income taxation in Belgium/Flanders, Denmark, Germany, the Netherlands and the UK and according to the Mirrlees Review

Mirrlees Review

Income tax rate

Imputed rent income taxed

Debt interest deductible (for simplicity no other costs considered)

Rate-ofreturn allowance

Capital gains tax

Same schedule for all income




Yes, if capital gains amounts to excess returns in the short term

Actual taxation Belgium/Flanders Yes

No in national Yes, included in No No income tax; ‘Woonbonus’*** Yes in regional/local property tax No Yes*, in Yes No No; size limit Denmark property tax applied No No** No No** No Germany As labour Yes*, but not Yes No No Netherlands income, not as economic income wealth value No No** No No** No UK *) Is imputed rent comparable to rent income – can it be considered a realistic income from investment? **) If mortgage loan has been repaid, the consumption good approach can be considered as amounting close to the effect as the rate-of-return allowance: no taxation of normal returns, if the returns are normal. However, other costs (maintenance, etc.) are ignored in this explanation ***) ‘Woonbonus’ is a lump sum deduction which includes interests and capital repayments for a mortgage loan, as well as premiums for life insurance Source: Haffner et al. (forthcoming)


It is not quite clear how excess return and normal return would be separated from each other according to the Mirrlees Review. If in the long-term the returns on capital invested in the housing market follow the inflation of construction costs, as seems to be the case in markets with little scarcity of building land (Boelhouwer, 2005), then there could not be any excess return achieved. But what would taxation look like, if that is not the case; e.g. when supply is inelastic? And what should be the levy in the short term (capital gains and losses?), if any? If taxation of accrued capital gains during occupation or realized capital gains at the moment of sale is the intention, these capital gains are not taxed in practice in the countries included in this study. Table 3 shows the tax treatment of the private person that operates as a landlord. General net income (income after the deduction of costs) is taxable, but again, no rate-of-return allowances are available, only a deduction for debt interest. Contrary to owner-occupiers the chances are that capital gains are somehow taxed for private person landlords, either when they are annually accrued or are realized at sale. Only in the Netherlands it is about an imputed total income, a concept that would include indirect return as well.

Table 3

Private person landlords’ personal income taxation in Belgium/Flanders, Denmark, Germany, the Netherlands and the UK and according to the Mirrlees Review Income tax rate

Rent income taxed

Debt interest deductible (for simplicity no other costs considered)*

Rate-ofreturn allowance

Capital gains tax

Same schedule for all income




Yes, if capital gains amounts to excess returns in the short term**




No No No

Yes; imputed rent + 40%, but imputed rent undervalued Yes Yes Yes

Yes Yes Yes

No No No

Yes Yes Yes



Yes; implicitly


Yes; implicit in total return that is imputed

No Yes Yes *) Usually other costs can be deducted **) Either accrued during normal business or realized at sale of dwelling Source: Haffner et al. (forthcoming)



Mirrlees Review

Actual taxation Belgium/Flanders Yes

Denmark Germany Netherlands – private person for a business Netherlands – private person for investment UK


Optimal and actual taxation of housing compared in five countries – property taxation As explained above, the Mirrlees Review does not recommend a property tax for owners of dwellings. The Mirrlees Review, however, recommends a land value tax only for business and agricultural land. The argumentation that a land value tax would be attractive for governments comes from the fact that land cannot be moved; thus behavior cannot be changed in reaction to the introduction of such a tax (Mirrlees, 2011a: 31). It would not be an annual tax, like property taxes generally are. It would be a tax on excess return that comes about because of the favorable location of the land. In practice, different types of annual taxes are levied on property, as Table 4 shows. Usually it is paid for by the owner (except for in the UK where it is paid for by the occupier) and covers land value plus the building value. However, in Denmark it came about as a replacement of the imputed rent taxation in income tax (European Commission, 2012a; Haffner, 2002). Denmark is the only country in this study that does also apply a land value only tax.

Table 4

Annual property taxation in Belgium/Flanders, Denmark, Germany, the Netherlands and the UK and according to the Mirrlees Review benchmark

Mirrlees Review Actual taxation Belgium/Flanders Denmark

Annual land value tax

Property tax

Only for business and agricultural land


No Yes; land value only

Yes Yes; instead of imputed rent taxation in income tax No


Yes; land plus building paid for by owner No Netherlands No UK Source: Haffner et al. (forthcoming)

Yes; owner Yes; occupier

Based on the empirics in the five countries under study, we conclude that none of the countries – if the taxonomy of taxation is evaluated – come close to optimal taxation in housing. The review also shows that none of the countries are treating their owner-occupied and rental dwelling the same way in the tax system. Tenure-neutral taxation cannot be observed. This also applies to Belgium/Flanders.

Considerations for evaluation The description about housing tax policies in the five countries has clarified that the guidelines based on the optimal tax theory which have been around much longer than the Mirrlees Review (although not always undisputed), do not seem to have made much headway. All countries have made their 11

own choices, though imputed rent taxation seems to taste defeat, as many European countries, including Belgium, have abolished this instrument (Haffner, 2002). Some kind of tenure neutral income tax treatment has therefore become largely impossible, even though it generally is promoted, also by the European Commission (2012a) that has formulated success criteria and points to consider in real estate taxation. Those points include not to favour debt finance above equity finance which would seem to support the rate-of-return allowance that is recommended by the Mirrlees Review. However, it is interesting to observe that the European Commission (2012a) in its recommendations does not always agree with the recommendations of the Mirrlees Review. The outstanding difference is concerned with the move towards taxation of real estate and away from taxation of employment in order to further economic growth, and within that to move to periodic taxation of residential real estate (instead of transaction tax) (see also European Commission, 2012b, 2013, and OECD, 2010a). Andrews et al. (2011) formulate in their OECD Economics Department Working Paper a number of general recommendations, again based on the premise of a tenure-neutral taxation of dwellings. First they pose the question that if imputed rent is taxed, will it be taxed based on market value of the real estate? Second, if mortgage interest is deductible from taxable income, are there limits in time or in tax base and has it been empirically proven that it is regressive? Third, a consideration is that if imputed rent is exempted from taxation, whether there are plans to go back to a balanced tax treatment of income and costs? Another question that these authors pose is whether the periodic property taxes can be regarded as a substitute for taxation of impute rent. Last, but not least there is the question of how capital gains are taxed. These questions seem to aim to stimulate the perception of the economically efficient tax treatment of dwellings.

Options for Flanders Now going back to the main question of this study: what should Belgium/Flanders consider when reexamining its housing taxation? Before we turn to this question, we have to explain more in detail, but briefly, the impact of the Belgian federal structure on housing taxation. In the past, the federal government has been entirely responsible for personal income tax and VAT. The competence for transaction taxes and property tax was transferred to the regions in 2002. In 2015, the Belgian regions will receive new fiscal responsibilities. The federal level will remain responsible for personal income tax in matters of determining the tax base and tax rates, but the regions will be allowed to levy a surplus on federal income tax. They can determine the level of the surplus and grant deductions (but no tax exemptions), as long as the progressivity of taxes is not eroded. Specific for housing is the transfer of the competence for tax deductions for owner-occupied dwellings1 to the regions. This transfer of competence is accompanied with a transfer of budgets amounting to the level of expenses in 20142. So for the year 2014, the budget will cover the expenses of the regions of the ‘Woonbonus’ and of the deductions that were in place before 2005. However, for the years ahead it is estimated that expenses will increase importantly. For Flanders in the case of ongoing


Limited to the single dwelling that is owned. Taxation of secondary residences remains federal. Exact figures will only be available by the end of 2016 and will be based on the expenses as calculated on the base of taxable income in 2014. 2


policy the necessary budget is estimated to increase with 40 billion euros each year because the expected entry of new beneficiaries (and because the benefit since 2005 is almost twice the amount under the former system), and as a consequence of indexation of amounts. The regions therefore will be obliged to make deductions less favourable or to cover the expenses by raising other taxes or cutting other expenses. For Flanders (and of course also for the other regions), this transfer of the favourable tax treatment of owner-occupiers from the federal to the regional level offers a policy window to evaluate, rethink and adapt the fiscal treatment of dwellings. If Flanders were to use the policy window in order to move towards an optimal tax design as recommended by the Mirrlees Review, what steps could be taken? We run through the four types of taxes that were distinguished earlier on. Transaction taxes in Belgium are among the highest of Europe (OECD, 2011). Optimal taxation theory would recommend abolition of transaction taxes. This measure is expected to increase mobility on the housing market and on the labour market (Isebaert, 2013) and to contribute to employment and economic growth (Hoi, 2009). Andrews et al. (2011: 72) stress that Transaction costs, including transaction taxes, have adverse effects in terms of hindering residential mobility and thereby labour mobility, which are likely to outweigh the benefits in terms of reducing excessive volatility in house prices. For income tax the Mirrlees Review recommends the same schedule for income from all sources. In Belgium, the same tax rate applies to income from dwellings and from other sources. However, since 2005 income from owner-occupied dwellings is exempt from taxation, while income from other dwellings remained subject to personal income tax. The application of optimal tax theory would imply to include owner-occupied dwellings in taxation again. Moreover, the tax base should reflect real market values of the dwellings. This would imply to replace the current tax base (the ’ cadastral income’), which strongly underestimates real market values, with a more accurate estimate of market rents. The principle of levying tax on imputed rent after deduction of costs would mean a reintroduction of the system before 2005, which is still in place for rented dwellings. It would also imply abolition of the current ‘Woonbonus’ (for owner-occupation) and of the deduction for ‘long term savings’ (for rented dwellings and second homes)3. Both deductions include capital repayments, which according to an investment approach cannot be considered as costs. According to Hoi (2009) such a reform could generate additional revenues in the order of 0.2% of GDP, while it would still leave Belgium with less reliance on immovable property taxes than other countries. Another outcome of such a reform would be a more neutral treatment of owner-occupied and rental dwellings. The recommendation of a deduction for the opportunity cost of capital would imply no taxation on the normal rate of return on capital invested in dwellings for owner-occupied as well as for rented houses. The only tax levied then would be a tax on capital gains, of which revenues are very unstable, depending on market conditions. If this recommendation is followed, tax revenue probably would decrease importantly and other kinds of income sources would have be found to finance government expenses. Under the current conditions, increasing taxes on labour income does not seem to be the right alternative. The overall tax burden on labour is considered to be particularly high in Belgium 3

As well as insurance fees.


(European Commission, 2013: 43). Furthermore, Decoster (2013) argues that property can be seen as a measure for the individual ability to earn money and that it therefore is an appropriate instrument for redistribution. Concerning property tax the Mirrlees Review does not propose it, possibly because it may be included in other taxes, such as income tax. The discussion for Flanders is complicated by the ambiguous position of property tax in Belgium. Two contrasting viewpoints on this are possible. Since property tax is levied only at the regional and local level, one can regard it as a tax for the use of regional and local services. On the other hand, it can be seen as a regional and local tax on income from owned property. That imputed rent is taken as tax base and that before 2005 (part of) the paid property tax was credited against (federal) personal income tax are arguments that point in the latter direction4. Depending on the point of view, different conclusions for property tax can be drawn then. • •

If property tax is seen as a fee for regional and local services, it can be levied on top of personal income tax. A relation with location of the dwelling is then evident; a relation with other aspects of the value of the dwelling less straightforward. If property tax is seen as income tax, the main principle is that income is not taxed twice. Then either the federal tax on income from dwellings is retained or the regional/local tax, or a justified mix of both. In the last case, several alternatives might be thinkable. - Property tax / income tax is levied once (federal) and then distributed among the different policy levels according to a formula. This would mean a step back in the fiscal autonomy of the Belgian regions. - Property tax / income tax on rental dwellings and second homes is levied at the federal level and a tax on owner-occupied dwellings at the regional level. Since the regions at this time can levy such a tax and determine autonomously the tax base, this would not imply a further transfer of competence between the state and the regions. The regions themselves can decide about the tax base and rates. So they can decide to replace the cadastral value (imputed rent) by a market value. A problem here is the difficult coordination between the policy levels with regard to the aim of tenure neutrality. - Income tax on rental dwellings and second homes is transferred to the regions. This would allow the regions to develop a tenure neutral housing policy, but would imply a next step in the reform of the state. In this scenario, the difficulty will be achieving tax neutrality (housing taxed in the same schedule as other forms of income).

Finally, for VAT the Mirrlees Review recommendations would imply a critical evaluation of the broadly used reduced VAT rate on renovation and repairs. Better targeted reductions can be considered. For example reductions for energy saving investments might be justified since these generate externalities to society. However, since VAT is a federal competence, Flanders is not allowed to change it.


This tax is still called ‘withholding tax’, referring to its origin as an advance tax payment for personal income tax. In 2005, the link between this ‘withholding tax’ and personal income tax was cut, but the term stayed in use.


Discussion – Flemish realism Based on theory and empirics it is analysed how the Belgian tax system and more specifically the future Flemish taxation could move closer to the optimal one. The transfer of the favourable tax treatment of owner-occupiers from the federal to the regional level offers a ‘window of opportunity’; especially, since decisions will be necessary to keep the budgets under control. From literature it becomes clear that it would be favourable to reshape housing taxation along the lines of optimal tax theory. However, at this time there are very few signs that this momentum will be used. On 25 May 2014, elections have been held for the European, federal and regional parliaments. During the political campaigns in the pre-election time, the reform of housing taxation has not been a main issue. In Flanders most political parties agreed upon continuation of the current system. Only the green party (‘Groen’) argued for a restriction of the benefits in time (15 years) (BIV, 2014). The political consensus is largely built on a very weak willingness of the population to pay more taxes in general and especially on their house. According to Peeters (2014) the legitimacy of the tax system in Belgium is under pressure, as a result of the intense issuing of new fiscal rules and the high tax burden: taxpayers increasingly have problems with understanding the logics and the values behind the complex tax system and (still according to Peeters) are less and less prepared to comply with the rules and to accept new taxes. This might apply even more to the owner-occupied dwelling, since home-ownership is so widespread in Flanders and the own house is of great significance for most people, as Meeus et al. (2013) show. For instance bringing the cadastral value closer to the market value will affect the majority of the households. Increasing the VAT rate for renovation to the standard rate is expected to provoke strong reactions from the building industry, that is arguing for broadening this reduction to new building in general. Also De Decker (2014) concludes that the window of opportunity will probably be missed. De Decker searches explanations in path dependency of housing policy and in a lack of planning culture in Flanders. A last reason for the resistance of the general public and politicians against a reform might lie in a lack of knowledge on the effects of tax benefits and on the distribution of benefits. The results of a recently finished study in Flanders might contribute to counterbalancing some opinions (Goeyvaerts et al., forthcoming). In conclusion, at this time there is probably too little public and political support for a change towards a tax design based on what optimal taxation theory would advise. A next question then can be which (small) steps in this direction might be feasible in the short term. Probably most within reach will be a reduction of transfer taxes. However, this will reduce government income and needs to be compensated by an increase of regular taxes on housing. A very gradual increase of property tax might be acceptable. The very unclear relation between the current cadastral value and the market value can provide an argument for actualizing the cadastral value, especially in regions where this value lags far behind the market. However, here again an institutional difficulty appears, since determining the cadastral value as tax base belongs to the competence of the federal state while most income from it accrues to the regions (and provinces and municipalities). So the federal government has no incentives to take this highly unpopular measure. Finally, a very slow cut-back of the ‘Woonbonus’ is not to be excluded. Under the current budgetary restraints, the non-indexation of the deductible amount might be acceptable for the coalition partners. At the time of writing, the first consultations for coalition formation are starting. Within a few weeks or months more can be told about the outcome. 15

References Andrews D., Caldera Sánchez A. & Johansson Å. (2011), ‘Housing Markets and Structural Policies in OECD Countries’, OECD Economics Department Working Papers, no. 836, OECD Publishing. BIV (2014), Verkiezingen 2014. De politieke partijen nemen een standpunt in. BIV-News, 2014, 1, BIV Beroepsinstituut voor Vastgoedmakelaars, Brussel. Boelhouwer Peter (2005), The incomplete privatization of the Dutch housing market: Exploding house prices versus falling house-building output, Journal of Housing and the Built Environment, 20, pp. 363–378. Cnossen S. (2010), Three VAT Studies, CPB Netherlands Bureau for Economic Policy Analysis, The Hague. Damen S., Vastmans, F. and Buyst, E. (2014), The long-run relationship between house prices and income re-examined: the role of mortgage interest deduction and mortgage product innovation, Discussion Paper Series DPS14.09, Faculty of Economics and Business, Leuven. Decoster A. (2013), ‘Vrijdenken over een belastinghervorming. Het hoofdstuk. Belastingdossier’, De gids op maatschappelijk gebied, 104, No. 4, pp. 5-21. Meeus, B., De Decker P. & B. Claessens (2013), De geest van suburbia. Garant Uitgevers, Antwerpen. De Decker P. (2014), Afbouw woonbonus: nu of nooit. Sampol, January 2014, p. 64-72. European Commission (2012a), European Economy. Possible reforms of real estate taxation. Criteria for successful policies, Directorate-General for Economic and Financial Affairs Occasional Papers, No. 119, European Union. European Commission (2012b), Tax reforms in EU Member States. Tax policy challengres for economic growth and fiscal sustainability. 2012 Report, Luxembourg, European Union. European Commission (2013), Tax reforms in EU Member States. Tax policy challenges for economic growth and fiscal sustainability. 2012 Report, Luxembourg, European Union. Feldstein M. (1976), On the theory of tax reform, Journal of Public Economics, 6(1-2), pp. 77-104. Goeyvaerts G., Haffner M., Heylen K., Van den Broeck K., Vastmans F. & Winters S. (forthcoming), Onderzoek naar de fiscaliteit in Vlaanderen. Deel 5. Samenvatting. CES-KULeuven, Leuven. Haffner Marietta E.A. (2002), ‘Dutch Personal Income Tax Reform 2001: An Exceptional Position for Owner-occupied Housing’, Housing Studies, 17, no. 3, p. 521-534. Haffner Marietta, Van den Broeck Katleen & Winters Sien (forthcoming) Onderzoek naar de woonfiscaliteit in Vlaanderen. Deel I Literatuur en Instrumenten, CES-KU Leuven. Haffner Marietta, Hoekstra Joris, Oxley Michael & Van der Heijden Harry (2009), Bridging the gap between market and social rented housing in six European countries, IOS Press BV, Amsterdam, http://repository.tudelft.nl/view/ir/uuid:d35c0ed4-7874-4413-8b90-25352ec8c980/. Heylen K. & Winters S. (2012), De verdeling van de subsidies op vlak van wonen in Vlaanderen, Steunpunt Wonen, Leuven. Heylen & Winters (2014), Eerste resultaten GWO2013 bevestigen zwakke positie van de private huurmarkt, Perstekst, Steunpunt Wonen, Leuven.

Hoj J. (2009), How to reform the Belgian tax system to enhance economic growth. OECD Economics Department working paper ° 741, OECD. Isebaert D. (2013), Housing and labour market performance. Doctoral Dissertation, UGent. Kofner Stefan (2014) The Private Rental Sector in Germany. OECD Research on private rental sector Consultancy report: Germany, Stefan Kofner, no place of publication given. McDaniel Paul R. & Surrey Stanley S. (eds.) (1985), International Aspects of Tax Expenditures: A Comparative Study, Kluwer Law and Taxation Publishers, Deventer, Series on International Taxation, no. 5. Mirrlees J., Adam S., Besley T., Blundell R., Bond S., Chote R., Gammie M., Johnson P., Myles G. & Poterba J. (eds) (2010), Dimensions of Tax Design: the Mirrlees Review, Oxford University Press, Oxford. Mirrlees J., Adam S., Besley T., Blundell R., Bond S., Chote R., Gammie M., Johnson P., Myles G. & Poterba J. (2011a), Tax by Design: the Mirrlees Review, Oxford University Press, Oxford. 16

Mirrlees J., Adam S., Besley T., Blundell R., Bond S., Chote R., Gammie M., Johnson P., Myles G. & Poterba J. (2011b), The Mirrlees Review: Conclusions and Recommendations for Reform, Fiscal Studies, 32, no. 3, p. 331-359. OECD (2010a), Tax Policy Reform and Economic Growth, OECD Tax Policy Studies, no. 20, Paris, OECD. OECD (2010b), OECD Economic Surveys: Netherlands 2010, OECD, geen plaats van publicatie. OECD (2011), ‘Housing and the Economy: Policies for Renovation’, Chapter from forthcoming Economic Policy Reforms 2011. Going for Growth. Peeters B. (2014), Is er behoefte aan een fundamentele hervorming van het Belgische belastingstelsel? Enkele bedenkingen over de personenbelasting, Christen-Democratische Reflecties, p. 63-72. Poterba J.M. (1984), Tax subsidies to owner-occupied housing: An asset-market approach, Quarterly Journal of Economics, 99, November, p. 729-752. Scanlon Kath & Kochan Ben (eds.) (2011), Towards a sustainable private rented sector. The lessons from other countries, LSE London, http://www2.lse.ac.uk/geographyAndEnvironment/research/london/events/HEIF/HEIF4b_1011%20-newlondonenv/prslaunch/Book.pdf. Van Ewijk Casper & Van Leuvensteijn Michiel (eds.) (2009) Homeownership & the Labour Market in Europe, Oxford University Press. Wood G. & Ong R. (2012), ‘Local Government Property Tax’, in S.J. Smith, M. Elsinga, L. Fox O’Mahony O. Seow Eng, S. Wachter & G. Wood (eds), International Encyclopedia of Housing and Home, 4, Elsevier, Oxford, p. 174-179.