Kingdom of the Netherlands—Netherlands: Selected Issues and Analytical Notes
Author:
International Monetary Fund
Search for other papers by International Monetary Fund in
Current site
Google Scholar Close
Publication Date:
21 Jun 2011
eISBN:
9781455286645
Language:
English
Keywords:
ISCR; CR; Netherlands; nondurable consumption; elasticity housing supply; mortgage interest subsidy; user cost; mortgage deductibility; Mortgages; Consumption; Income; interest deductibility
Download PDF (538.7 KB)

Important issues of the Netherlands are discussed. Openness to trade has benefited the Netherlands before the crisis and has supported the recent recovery process. However, both financial openness and trade linkages have also been a transmission channel for the financial crisis. Synchronized fiscal tightening across Europe has important spillover effects for GDP growth. The improvement on the supply side of credit has contributed to a normalization of the credit market. However, the recent increase in the financial stress index indicates that the situation is still fragile.

Abstract

Important issues of the Netherlands are discussed. Openness to trade has benefited the Netherlands before the crisis and has supported the recent recovery process. However, both financial openness and trade linkages have also been a transmission channel for the financial crisis. Synchronized fiscal tightening across Europe has important spillover effects for GDP growth. The improvement on the supply side of credit has contributed to a normalization of the credit market. However, the recent increase in the financial stress index indicates that the situation is still fragile.

Analytical Note 9: Costs and Benefits of Eliminating Mortgage Interest Deductibility1

A. Introduction

1. Mortgage interest deductibility represents a substantial subsidy to home owners in the Netherlands. Owner-occupiers can deduct all interest payments on mortgages from taxable income.2 The cost of servicing mortgages is deductible from taxable income, creating a budgetary cost of around 11 billion euro, or about 2.2 percent of GDP. The subsidy has been estimated to worth about 20 (van Ewijk et al. 2010) to 25 (Ter Rele and van Steen, 2001) percent of the implicit price of owner occupied housing.

2. Evidence for the Netherlands and other economies suggests that mortgage interest deductibility is inefficient with respect to social goals and very regressive. Mortgage interest deductions were originally promoted on the grounds that increased home ownership would generate wider benefits, such as an increased sense of community and a more stable a secure environment in which to raise children. However, there is little evidence that mortgage interest deductibility is effective at the margin of enabling greater home ownership. Instead, it encourages those who can already have access to housing finance to consume more housing services than they otherwise would. Hence, it is a quite regressive subsidy, benefiting mainly those on higher incomes.3 More generally, because the subsidies are tied to the consumption of housing, they distort the allocation of resources, creating a welfare loss.

3. The mortgage interest deductibility creates a clear incentive for households and financial intermediaries to expand balance sheets. To maximize the benefits of the deductibility, households have a clear incentive to maximize loans relative to the value of the house, and to not pay down the principle. Indeed, LTVs on new mortgages are now as high as 110-120 percent (depending on measurement and data sources), and many loans are zero amortization products. Financial intermediaries have a clear incentive to expand loan books and provide products that maximize benefits to households.

4. The pressures on fiscal balances as a result of the crisis raise the question as to whether the implicit subsidy could be better directed. Although most households can benefit from housing subsidies, they (or future generations) pay for it through taxes. The implicit subsidy from mortgage interest deductibility has been rising in real terms and now represents about 7 percent of total public revenues. The total cost of the tax treatment of housing is estimated to be in the region of 17 billion euro,4 which is more than the fall in revenues experienced as a result of the downturn from 2008 to 2009. The subsidy could be used instead to increase provision of government services or reduce taxes (both of which could be directed more progressively than the mortgage interest deduction if so wished) or reduced public debt.

5. However, eliminating mortgage interest deductibility could also risk depressing private demand and damaging the banking sector in the short run. A sudden elimination of mortgage interest deductibility could cause a severe house price downturn. Falling asset values could depress private demand at a time when growth is fragile and the economy is only just showing signs of recovery. Moreover, as discussed in the FSAP, financial institutions in the Netherlands are unusually exposed to the Dutch mortgage market. A fall in house prices, if large enough, could put balance sheets under stress and cause banks to tighten lending conditions, to the detriment of activity in the rest of the economy.

6. This note compares the potential long-run benefits of removing mortgage interest deductibility with potential short-run costs for private demand. Most papers on the mortgage interest deductibility in the Netherlands have focused on the welfare and distributional costs of the subsidy.5 By contrast, this note is in the spirit of Mankiw and Weinzierl (2005) in attempting to illustrate the macroeconomic dimensions of eliminating the deductibility. It is a complement to more detailed but partial equilibrium assessments of eliminating housing subsidies such as CPB (2010).

7. The optimal elimination of mortgage interest deductibility will balance the short-run costs with long-run benefits. Depending on the way by which the mortgage interest subsidy is recycled, potential supply side benefits could arise from increased labor participation and/or greater capital accumulation. However, house prices would likely fall with the elimination of the deductibility. In the short run, this could have a negative effect on private demand. Although the model used here is too crude to provide precise estimates of an optimal path for the elimination of mortgage interest deductibility, several plausible factors suggest that its elimination should be gradual.

8. The main factors affecting short-run costs include negative spillovers from house price falls on aggregate demand and rigidities in housing supply. Short-run costs will be greater to the extent that:

  • Falling house price falls produce spillover effects on current consumption; and

  • Housing supply is inelastic.

9. Long-run benefits will be greater to the extent that the way the revenue is recycled generates increased supply. In the model used here, the most efficient policy is to reduce capital taxation.

B. Key Features of the Model

10. The analysis uses a small dynamic general equilibrium model calibrated to the Dutch economy.6 For clarity and simplicity, a number of simplifying assumptions are made. First, the economy is assumed to be closed; there is no trade.7 Second, there is no nominal side; we abstract from nominal rigidities and monetary policy. Third, there are no real rigidities, such as from habits and investment adjustment costs.8

11. The economy contains households, a production sector, and a government.

  • On the demand side, households in a closed economy invest (building up capital for use in production), consume non-durable goods, and enjoy the services from owning a durable good (housing). More specifically, all households have identical preferences for a CES basket of nondurable consumption goods, C, and housing stock, D (for “durables” or “dwellings”). They are assumed to be infinitely lived and forward looking.9 Hence, consumption decisions follow conventional forward-looking Euler equations, given budget constraints. Expectations are a crucial element in the determination of the responses to the elimination of the deductibility, as households assess the potential benefits of the policy shock against the immediate effects.

  • On the supply side, capital and labor are combined in a Cobb-Douglas production function to generate output, with income returned to households in the form of rents paid on these factors. Rather than explicitly model production of houses in a second sector, a reduced-form house price equation is used that embeds the assumption that factor productivity for the production of houses is less than for the production of other goods (that is, given a demand shock, house prices change more than other prices, so the relative price of houses changes).10

  • A government taxes consumption, wage and capital income, while redistributing these revenues as transfers and making its own expenditures. In particular, the government subsidizes interest payments on housing, thereby affecting the user cost of housing, which households take into account when deciding how much to spend on nondurables and housing services.

12. The mortgage rate deductibility is modeled as a subsidy on mortgage interest rates and directly affects the user cost of housing. In essence, mortgage interest deductibility reduces the amount of labor income that is taxable. However, the deductibility is not quite that simple in the case of the Dutch economy.11 Rather than attempting to explicitly model all the tax distortions affecting housing decisions, one convenient way of modeling the deductibility is in the form of a subsidy on mortgage interest rates, τrd. In the model, mortgages carry the interest cost rd, which contains a risk premium over the riskless (1 − τrd)rd). Lowering τrd therefore increases the effective mortgage rate closer to the market rate.

13. The direct effect of eliminating the mortgage interest deduction will be to reduce income available to households for other expenditures. The size of the effect depends on the size of the subsidy, the mortgage rate, and the size of the mortgage stock, which in turn depends on the proportion of households holding mortgages and the average loan-to-value ratio.

14. A second effect of eliminating the mortgage interest deduction will be to cause households to substitute away from houses toward other consumption. The user cost of housing factors in depreciation, expected capital gains, the net mortgage interest rate, and the subsidy itself:

user cost of housing = f ⁡ ( δ d , E [ p t + 1 d p t d ] , r t d , τ t rd ) , + = − + −

where δd denotes the depreciation rate on dwellings and pd the relative price of dwellings. The decision about desired housing stock in any given period can be then related to current consumption by taking into account the user cost and preference parameters such as the elasticity of substitution. Although housing demand can be written in terms of a forward-looking Euler condition, it can be summarized neatly in terms of a simple intratemporal decision rule that describes how consumers allocate spending between (nondurable) consumption flow, C, and (durable) housing stock, D, given their choice about overall expenditures in any period:

D = f ( C , ucd ) , + −

where ucd denotes the user cost of housing. The direct effect of a reduction in the mortgage interest rate subsidy will therefore be to shift expenditures toward nondurable consumption. The net effect on aggregate demand will depend on the responses of desired capital investment, government consumption, and the extent to which nondurable consumption is reduced or raised.

15. Changes in the mortgage interest subsidy have effects via the government budget constraint. For the government, the subsidy is worth Trd = τrd. rd. M in foregone revenues, where M is the size of the mortgage stock. The flow government budget constraint in the model requires that tax revenues and changes in public debt have to balance debt servicing and government expenditures, net of the mortgage deductions. Consequently, any reduction in deductibility will afford a reduction in taxes or debt or an increase in expenditures.

C. Parameterizing the Model

16. The model is calibrated to match broadly the recent behavior of the Dutch economy. Table 9-1 lists parameter values. Table 9-2 lists steady-state expenditure shares of (domestic) demand; income shares; tax revenues, debt, and spending (all as ratios to domestic demand) and associated implied average tax rates. The household discount rate is assumed to be 0.99, as standard, producing a riskless interest rate of 4 percent per annum. The share of national income returning to capital is 0.4, which is higher than in many other economies, and will make responses to changes in capital accumulation larger (such as in the case of a reduction in capital taxation, as below). Fixed capital is assumed to depreciate at a rate of 6 percent each year, while the housing stock depreciates at nearly 1.3 percent (this value approximates maintenance, insurance, and transactions costs; see van Ewijk et al. 2010).

Table 9-1.

Parameter Values in the Model16

article image
Table 9-2.

Steady-State Values of the Model

article image

17. The effective interest rate that homeowners pay on mortgages reflects the implicit subsidy. This parameter is calibrated so that the foregone tax revenue matches the estimate of 2.2 percent of GDP. The average loan-to-value ratio is imposed at the current value for the Netherlands at 50 percent of GDP.12 Given the assumptions about parameters that affect the mortgage interest rate, the required value is 0.7.13

18. Effective tax rates are calculated to match fiscal revenues and expenditures. Total tax revenues are allocated to four tax types: labor income taxes, capital income taxes, consumption taxes, and lump-sum (non-distortionary) taxes. Lump sum taxes are treated as the residual after allocating to the first three to ensure total tax revenue is accounted for. Actual lump-sum taxes net out transfers (social security and benefits).

19. Housing supply is assumed to be relatively inelastic. Whereas estimates of housing supply estimates are as high as 4 in other advanced economies, estimates for the Netherlands are much lower, probably reflecting tight zoning regulation of the Dutch housing market.14

The shares of capital and labor in housing production imply a relatively high importance of land as a fixed factor; these parameters are calibrated so that the elasticity housing supply is 0.65, the same as the long-run elasticity in CPB (2010).15

20. In the baseline specification, housing services and nondurable consumption are not highly substitutable. The elasticity of substitution of housing services and nondurable goods has an important bearing on the impact of eliminating the MID: a low elasticity will generate larger spillovers onto nondurable consumption. Evidence from the U.S. suggests that housing services and nondurable goods are consumed virtually in fixed proportions and are not willingly substituted.17 However, CPB (2010) uses a Cobb-Douglas specification. The assumption in this model is that the elasticity is 0.5.

D. Long-run Benefits from Eliminating MID

21. The long-run benefits from removing mortgage interest deductibility depend on the efficiency with which the implied subsidy is redirected. Table 9-3 shows, from left to right, the steady-state percentage changes in output, nondurable consumption, housing investment, and house prices in response to complete elimination of the deductions. In addition, the final column shows the percentage change in the consumption aggregator, which has no equivalent in the national accounts data, but is a measure of household utility. There are three variations, depending on how the subsidy is recycled: (i) increased government consumption; (ii) increased transfers; and (iii) reduced capital income taxes.

Table 9-3.

Long-run Effects of Eliminating Mortgage Interest Deductibility

(Percentage deviation from starting values)

article image

22. In this model, and given the baseline calibration, the first-best policy is to recycle the mortgage interest subsidy back as lower capital taxes. Lowering capital income taxes raises saving and capital investment, boosting output in the long run. The fiscal policy encourages a relative shift away from putting resources into housing capital to putting them into productive capital instead. With higher permanent income and the relative shift in capital allocation, nondurable consumption increases even more than output. In absolute terms, housing investment nonetheless also increases, given the scale of the increase in output. Concomitantly, house prices also rise. The results are sensitive to changes in parameter values, but nonetheless illustrate some implications that might not be clear from a partial equilibrium analysis.

23. Other policies do not boost welfare in the simulations by as much because they have no long-run effects on output. The fiscal choices shown here are deliberately starkly different, and more for illustration rather than prescriptive purposes. Government expenditure is assumed to be entirely nonproductive (a proportion could be assumed to be invested in ways that are known to generate positive externalities, such as infrastructure and R&D), so this policy is not very efficient. Without any increase in permanent income, the result is substitution from housing investment to nondurable consumption. Returning the subsidy to households as a lump-sum transfer is better in welfare terms (on the assumption that government expenditures do not enter households’ utility.)18

24. Hence, based on the assumptions in the model, eliminating mortgage interest deductibility could raise productive potential by as much as 2½ percent. The efficiency implications of redirecting the mortgage interest subsidy depend crucially on assumptions made about the type and amount of distortions from spending, transfers, and other taxes. Used completely inefficiently, long-run benefits would be zero.

E. Short-run Costs of Eliminating Mortgage Interest Deductibility

25. The short-run costs on aggregate consumption from removing mortgage interest deductibility will depend crucially on how households react to falling house prices. In the case of a fiscal strategy that raises long-run potential output, households will increase overall consumption due to the rise in permanent income.19 But, in the short run, the direct effect of the withdrawal of the subsidy is to reduce disposable income. Aside from that, there might be extra short-run negative “spillover” effects from reduced house prices on consumption. In this model, these effects are proxied by assumptions about the elasticity of substitution between housing services and nondurable consumption. If the elasticity is high, the spillover effects on nondurable consumption are small. If the expenditures are complements, as is the case in the calibration here, households sacrifice nondurable consumption so as to preserve housing utility.

26. Short-run costs also depend on supply conditions. The same factors that imply that house prices rise relatively quickly given increases in demand—land as a fixed factor, zoning restrictions, and other adjustment costs—also imply that house prices would fall relatively severely given an increase in the user cost of housing. In essence, the impact of the policy change would be less if the capital stored in the form of housing could be easily unbundled and converted into capital stock for production. On the assumption that such conversion costs are very large, the price fall will be greater. These rigidities are approximated by the low elasticity of housing supply.

27. In the baseline calibration, a complete elimination of mortgage interest deductibility in one step results in a sharp decrease in the value of the housing stock. Figures 9-1 to 9-4 show the results of a complete, unanticipated, immediate elimination of the deductibility, on the assumptions that the revenue is used to reduce capital income taxes.

  • The direct effect of the elimination is to raise the user cost of housing. The expected depreciation of house prices adds to the effect of eliminating the subsidy. In addition, market real interest rates rise with the expectation of increased permanent income, although they will return to original levels eventually (Figure 9-1). House prices fall by nearly 15 percent on impact, but will eventually recover to the level shown in Table 7-3 (Figure 9-2). Similarly, residential investment and nondurable consumption are below the starting level over the short and medium run, before eventually recovering (Figure 9-3). Wealth is reallocated from housing stock to fixed capital (Figure 9-4).

  • These results are merely suggestive, and depend critically on a number of parameters. A lower housing supply elasticity would result in a more severe fall in house prices. Smaller shares of housing in the consumption basket would reduce the spillovers onto nondurable consumption. A lower elasticity of substitution between nondurable consumption and housing would raise the spillovers. The long-run supply response of output would be smaller if capital’s share of income were lower (in the case of the reduction of capital taxes).

  • Further, the model does not include financial mechanisms that might amplify the effects of the shock. In Aoki et al. (2004) and Kannan et al. (2009), when house prices fall, “banks” face pressures on balance sheets and raise financing rates to protect against defaults. In such models, a shock that causes mortgage interest rate premia to rise will raise the user cost of housing, raising downward pressure on house prices further, and thereby “accelerating” the downturn in house prices. Here, spillovers are proxied by the substitution elasticity of demand, but a more detailed specification of credit mechanisms could illustrate other macroeconomic implications, such as stress on the banking system.20

Figure 9-1.
Figure 9-1.

User Cost of Housing and Interest Rate Responses

(Permanent, immediate elimination of MID, recycled as lower capital taxes)

Citation: IMF Staff Country Reports 2011, 143; 10.5089/9781455286645.002.A009

Source: IMF staff estimates.
Figure 9-2.
Figure 9-2.

House Price Responses

(Permanent, immediate elimination of MID, recycled as lower capital taxes)

Citation: IMF Staff Country Reports 2011, 143; 10.5089/9781455286645.002.A009

Source: IMF staff estimates.
Figure 9-3.
Figure 9-3.

Nondurable Consumption and Residential Investment Responses

(Permanent, immediate elimination of MID, recycled as lower capital taxes)

Citation: IMF Staff Country Reports 2011, 143; 10.5089/9781455286645.002.A009

Source: IMF staff estimates.
Figure 9-4.
Figure 9-4.

Housing and Fixed Capital Responses

(Permanent, immediate elimination of MID, recycled as lower capital taxes)

Citation: IMF Staff Country Reports 2011, 143; 10.5089/9781455286645.002.A009

Source: IMF staff estimates.

28. Long-run supply responses are similar to those in other studies on the potential effects of eliminating MID, but the effects on prices are somewhat less.

  • The long-run responses on supply are comparable to those in van Ewijk et al. (2010) for a labor supply shock.

  • Estimates of the impact on house prices are generally somewhat more benign than other studies. Ter Rele and van Steen (2001) estimate that the effect of government subsidies is to raise house prices by as much as 25 percent for higher income cohorts. Mechanically, their estimates would imply a much larger fall in house prices were the subsidy to be eliminated. CPB (2010) estimates the initial impact to be in the order of -15 percent, troughing at -25 percent, and rising with time to -20 percent.

  • It can be difficult to directly compare the results here with those in other studies.21

  • Some of the differences are likely explained by different assumptions about elasticities. In particular, if the short-run elasticity of supply is close to zero, as argued in Vermeulen and Rouwendal (2007) and used in CPB (2010), then the price response could be much larger.

  • More generally, different conclusions are unsurprising when comparing partial with general equilibrium analyses. Crucially, it is important to be clear how the subsidies are to be re-used—in the scenario used here, households benefit is not simply from lower taxes, but the higher incomes generated from lower taxes. As in Mankiw and Weinzierl (2005), this “second-round” effect is not small. What households lose from having subsidies taken away is at least made up for by higher permanent incomes—although not proven here, it seems plausible that this should play a role in supporting house prices, compared with the results from partial equilibrium analyses. Hence, simply looking at the size of the subsidy is a potentially misleading guide to the potential effects on the housing market of removing the deductibility.

F. Timing the Elimination of Mortgage Interest Deductibility

29. A sharp reduction in house prices poses risks. Although the elimination of mortgage interest deductibility could bring significant fiscal and supply benefits, a legitimate concern is that an immediate, unanticipated and complete elimination of the mortgage interest subsidy could have severe effects on house prices, housing wealth, and aggregate consumption. In particular, analysis of Dutch banks in the FSAP indicates that this could result in an increase in non-performing mortgages, which could damage the banking sector and further affect real activity. This suggests that the elimination of the mortgage deductibility should be gradual.

30. “Grandfathering” the elimination is desirable. A credible preannouncement of policy changes would help to avoid a sharp fall in housing values (and, in reality, reduce costs of adjustment, given nominal contracts). Figure 9-5 illustrates the case. Instead of an immediate and completely unanticipated change as presented previously, consider reactions when the change is announced to come into effect five years into the future. Even though credit-constrained agents are not themselves forward looking, their consumption falls by less, because the immediate fall in housing values is not as severe. Hence, aggregate consumption falls by less. The trade-off is that consumption does not rise as quickly. These results assume that the policy announcement is completely credible and understood, but nonetheless illustrate that preannouncement could have important short-run benefits.

Figure 9-5.
Figure 9-5.

Responses to Anticipated Shock

(Anticipated elimination of MID at t + 20, recycled as lower capital taxes)

Citation: IMF Staff Country Reports 2011, 143; 10.5089/9781455286645.002.A009

Source: IMF staff estimates.

31. Gradual elimination would lower risks. The model used here is too crude to be used to precisely calculate an optimal path for the elimination of the subsidy. Nonetheless, a simple exercise illustrates the trade-offs between eliminating the deduction swiftly and a more gradual reduction. Instead of the immediate and complete elimination shown in Figures 9-1 to 7-4, consider gradual elimination in constant steps over the course of 5 years (reducing once per year) to zero deductibility. Each of these reductions in deductibility is anticipated and perceived to be permanent. Figure 9-6 shows that this staggered reduction would avoid sharp falls in consumption and prices. On the other hand, as above, the potential benefits are not realized as quickly.22

Figure 9-6.
Figure 9-6.

Consumption Responses

(Staggered vs immediate elimination of MID, recycled as lower capital taxes)

Citation: IMF Staff Country Reports 2011, 143; 10.5089/9781455286645.002.A009

Source: IMF staff estimates.
Figure 9-7.
Figure 9-7.

House Price Responses

(Staggered vs immediate elimination of MID, recycled as lower capital taxes)

Citation: IMF Staff Country Reports 2011, 143; 10.5089/9781455286645.002.A009

Source: IMF staff estimates.

G. Conclusions

32. Mortgage interest deductibility is costly. Mortgage interest deductibility is a highly regressive subsidy that lowers the user cost of housing.

33. Redirecting the mortgage subsidy could boost output. The implicit revenues are large and could be potentially be otherwise used to boost potential output.

34. Immediate unanticipated elimination brings risks. Empirical evidence for the Netherlands and studies of other economies indicates that changes in housing wealth have potentially important spillover effects onto aggregate consumption. An immediate unanticipated elimination of the deductibility would likely induce a sudden fall in house prices and aggregate demand.

35. Elimination should be gradual and preannounced—but not too slow. A credible preannouncement of changes to the deductibility would allow households time to make plans. Gradual elimination would reduce the impact on other components of demand. However, the longer the elimination is put off, so too are the potential benefits from redirecting the subsidy.

Equations for the model

The model is coded in nonlinear levels. In general, upper case denotes a flow or a stock, while lower case denotes a relative price or rate. (All superscripts are in lower case.) Because the model is a simultaneous system, there is technically no such thing as a “consumption equation”. However, an attempt is made below to attribute endogenous variables to equations to aid interpretation. For example, in the case of the first equation, the household flow budget constraint can be seen as “determining” capital stock, conditional on households’ decisions about consumption (saving).

Flow budget constraint (K):

K t = ( 1 + ( 1 − τ k ) . r t − 1 k − δ k ) . K t − 1 + ( 1 − δ d ) . p t d . D t − 1 − p t d . D t − τ t − 1 rd . r t − 1 d . M t − 1 + ( 1 − τ w ) . W t . ( 1 − λ ) . L t − C t + Π t − T t l

Mortgage stock (M):

M t = θ . μ . p t d . D t

Consumption Euler equation (C):

[ ɸ . ( ι . C t ) ρ + ( 1 − ɸ ) . ( ( 1 − ι ) . D t ) ρ ] ( 1 ρ 1 ) . C t ρ − 1 = [ ɸ . ( ι . C t + 1 ) ρ + ( 1 − ɸ ) . ( ( 1 − ι ) . D t + 1 ) ρ ] ( 1 ρ 1 ) . β . ( 1 + r t ) . C t + 1 ρ − 1

Intratemporal decision rule for housing for forward-looking agents (D):

D t ρ − 1 = [ ɸ . ι ρ ( 1 − ɸ ) . ( 1 − ι ) ρ ] 1 1 − ρ . [ 1 p t d . χ t d ] 1 1 − ρ C t ρ − 1

Residential investment by forward-looking households (ID):

IDt = Dt − (1 − δd).Dt-1

Output (Y):

Y t = Z t K t − 1 α L t ( 1 − α )

Profits (Π):

Π t = C t + IK t + ID t + G t − W t L t − r t − 1 k . K t − 1

Real wages (W):

W t = ( 1 − α ) Z t K t − 1 α L t ( 1 − α ) L t

Real rental rate of capital (rk):

r t k = α Z t + 1 K t α L t + 1 ( 1 − α ) K t

Fixed capital investment (IK):

IKt = Yt − Ct − IDt − Gt

Real interest rate (r):

r t = ( 1 − τ t + 1 k ) . r t k − δ k

Mortgage interest rate (rd):

r t d = r t + k

User cost of housing (χd):

χ d = ( 1 − 1 − δ d 1 + r t . p t + 1 d p t d ) + μ . ( 1 + ( 1 − τ t rd ) . r t d 1 + r t − 1 )

Supply schedule for housing (pd):

p t d = ( Z t d ) ( − 1 γ 1 + γ 2 ) . ( ID t ) ( 1 − γ 1 − γ 2 γ 1 + γ 2 ) . ( W t γ 2 Z t ) ( γ 2 γ 1 + γ 2 ) . ( r t − 1 k γ 1 Z t ) ( γ 2 γ 1 + γ 2 )

Housing sector productivity (Zd):

Z d = ( 1 − γ 1 − γ 2 ) . Y t Y t − 1

Government spending (G):

G t = G t ¯ . GDP t

Government debt (B):

B t = B t ¯ . GDP t

Mortgage interest subsidy (Trd):

T t rd = τ t − 1 rd . r t − 1 d . M t − 1

Labor income tax revenue (Tw):

T t w = τ t w . W t . L t

Capital income tax revenue (Tk):

T t k = τ t k . r t − 1 k . K t − 1

Flow government budget constraint (Tl):

T t l + T t w + T t k + B t = G t + ( 1 + r t − 1 ) . B t − 1 + T t rd

Gross Domestic Product (GDP):

GDP t = C t + p t d . ID t + IK t + G t

Parameters

article image

Exogenous variables

(depending on shock configuration)

article image

Exogenous variables

article image
article image

References

  • Aoki, Kosuke, James Proudman, and Gertjan Vlieghe, 2004, “House prices, consumption, and monetary policy: a financial accelerator approach,” Journal of Financial Intermediation 13, pp. 414-35.

    • Search Google Scholar
    • Export Citation

  • Ewijk, Casper van, Martin Koning, Marcel Lever, and Ruud de Mooij, 2006, “Economische effecten van aanpassing fiscale behandeling eigen woning” CPB Paper No. 62.

    • Search Google Scholar
    • Export Citation

  • Ewijk, Casper van, Bas Jacobs, and Ruud de Mooij, 2010, “Welfare effects of fiscal subsidies on home ownership in the Netherlands,” manuscript.

    • Search Google Scholar
    • Export Citation

  • Dijk, Machiel van, and Gerbert Romijn, 2010, “Hervorming van het Nederlandse woonbeleid,” CPB Special Publication 84.

  • Glaeser, Edward L., and Jesse M. Shapiro, 2002, “The benefits of the home mortgage interest deduction,” NBER Working Paper 9284.

  • International Monetary Fund, 2006, “House prices in the Netherlands: determinants, concerns and considerations related to phasing out the tax deductibility of mortgage interest payments,” in Kingdom of the Netherlands—Selected Issues.

    • Search Google Scholar
    • Export Citation

  • Kannan, Prakash, Pau Rabanal, and Alasdair Scott, 2009, “Monetary and macroprudential policy rules in a model with house price booms,” International Monetary Fund Working Paper WP09/251.

    • Search Google Scholar
    • Export Citation

  • Mankiw, N. Gregory, and Matthew Weinzierl, 2005, “Dynamic scoring: a back-of-the-envelope guide,” manuscript.

  • OECD (2004), OECD Economic Survey of the Netherlands, 2004: Housing Policies.

  • Poterba, James, 1984, “Tax subsidies to owner-occupied housing: an asset market approach,” Quarterly Journal of Economics 99(4).

  • Rouwendal, Jan, 2007, “Mortgage interest deductibility and homeownership in the Netherlands,” Department of Spatial Economics, VU University Discussion Paper 2007-050.

    • Search Google Scholar
    • Export Citation

  • Verbruggen, Johan, Henk Kranendonk, Michiel van Leuvensteijn en Michel Toet, 2005, Welke factoren bepalen de ontwikkeling van de huizenprijs in Nederland, CPB Paper No. 81.

    • Search Google Scholar
    • Export Citation

  • ter Rele, Harry, and Guido van Steen, 2001, “Housing subsidization in the Netherlands: measuring its distortionary and distributional impacts,” CPB Discussion Paper No. 002.

    • Search Google Scholar
    • Export Citation

  • Vermeulen, Wouter, and Jan Rouwendal, 2007, “Housing supply in the Netherlands,” CPB Discussion Paper No 87.

1

Prepared by Alasdair Scott.

2

Some imputed rent is added to taxable income, which reduces the implicit subsidy, but the current levels of imputed rent are generally considered to be far below the values that would be required to make the interest deductibility neutral from a consumption choice perspective. Other policies affecting household decisions about housing include (i) a 6 percent transactions tax, (ii) exemptions of net home equity (the value of a house less the value of the mortgage) from wealth tax, (iii) rental subsidies, and (iv) zoning regulations on land available for housing construction.

6

More details are provided in a companion technical paper. Simulation code in the Troll programming language is available on request.

7

This reflects an assumption that the openness of the economy does not substantially affect the results below: housing is a nontradeable good. Changes in the real exchange rate would affect the relative price of nondurable goods, to the extent that some nondurables are imported, but the assumption is that this is of second order importance. Some of the benefits of reduced taxes or higher public spending might leak out in an open economy setting.

8

This is mainly a modeling choice to preserve simplicity and transparency—consumption habits and investment adjustment costs could be added easily.

9

This has the implication that public debt is not net wealth. Hence, changes in debt will have no effect on the economy, so debt reduction scenarios are not considered in this paper.

10

This short-cut implicitly assumes that the shock will not have a material impact on the relative prices of factor inputs across production sectors.

11

For example, there are limitations to the mortgage rate interest deductibility (deductibility lowered for capital gains made on the previously-owned house, no interest deductibility for part of the mortgage related to consumption spending). Also, in conjunction with the interest rate deductibility, houses themselves are taxed.

12

The average is steadily rising, as the loan-to-value ratio of new mortgages is in the region of 110-120 percent of GDP (depending on measurement and sources).

13

This looks like a very large value, but note that the model is quarterly. Hence, the equivalent mortgage interest subsidy rate at annual frequency to compare with those from other studies is 0.7/4 = 17.5 percent.

14

See also OECD (2004).

15

Vermeulen and Rouwendal (2007) find that residential investment is almost fully inelastic with respect to house prices in the short run. Hence, in CPB (2010), the short-run elasticity is zero. For simplicity, this model has no short-run rigidities; these could be added, but a view would need to be taken on all relevant short-run rigidities, not only those that applied to housing supply, which would increase the complexity of the model and reduce the transparency of the analysis.

16

NB: the model is calibrated for quarterly frequency.

17

For example, Iacoviello (2004) estimates the elasticity of substitution to be as low as 0.1.

18

In the current calibration, the elasticity of labor supply is zero, so that the effects of a decrease in labor income tax are the same as decreasing lump-sum taxes net of transfers.

19

In what follows, the focus is on the responses of consumption. A practical reason is the model implies immediate increases in capital investment, which might be considered unrealistic. This is because supply is fixed for the first period—capital is predetermined and labor supply is inelastic. Production clearing implies that if consumption and residential investment fall, the extra output has to be accommodated by increased capital investment demand. Instead, we might expect that demand falls would lead to reduced output, with output accumulated as inventories and/or reduced capacity utilization resulting in a negative output gap, as currently exists.

20

Empirical work with Dutch data (not reported) shows that nondurable consumption is strongly associated with house prices and real mortgage credit. Estimating reduced-form Euler equations implied by Iacoviello’s (2004) model indicates that the data seem to be consistent with the existence of credit-constrained behavior in Dutch consumption choices.

21

For example, the scenario in CPB (2010) is a combination of shocks, including deregulation of the rental market and abolition of the transactions tax, rather than only the elimination of the MID, as here.

22

Two examples from history also suggest that gradualism is desirable. Sweden reduced the tax deductibility over a short span of time, from 1985 to 1991. The marginal effects of this policy change are very difficult to isolate, as many other tax changes were made during this period. Further, the second reduction in mortgage deductibility was enacted at the same time as the U.S. entered recession. Nonetheless, the example suggests that large step changes over a short period of time might compound existing financial vulnerabilities. By contrast, the U.K. completely eliminated the deduction, from 1974 to 1999, by implementing a nominal cap on the size of mortgage loans that qualified for tax deductibility. (See IMF (2006) for more details.) By phasing out the deductibility over many years, the policy change was less vulnerable to business cycle interactions and allowed time for households to fully factor the future path of deductibility into their decisions. Moreover, the nominal cap (as compared to cuts in real deductibility) lessens the risk of house price falls.

Same Series

Other IMF Content

Other Publishers

  • Collapse
  • Expand
Cover IMF Staff Country Reports
Kingdom of Netherlands: Netherlands: Selected Issues and Analytical Notes
Author:
International Monetary Fund
Volume/Issue:
Volume 2011: Issue 143
Publisher:
International Monetary Fund
ISBN:
9781455286645
ISSN:
1934-7685
Pages:
144
DOI:
https://doi.org/10.5089/9781455286645.002
  • Figure 9-1.
    View in gallery
    Figure 9-1.

    User Cost of Housing and Interest Rate Responses

    (Permanent, immediate elimination of MID, recycled as lower capital taxes)

  • Figure 9-2.
    View in gallery
    Figure 9-2.

    House Price Responses

    (Permanent, immediate elimination of MID, recycled as lower capital taxes)

  • Figure 9-3.
    View in gallery
    Figure 9-3.

    Nondurable Consumption and Residential Investment Responses

    (Permanent, immediate elimination of MID, recycled as lower capital taxes)

  • Figure 9-4.
    View in gallery
    Figure 9-4.

    Housing and Fixed Capital Responses

    (Permanent, immediate elimination of MID, recycled as lower capital taxes)

  • Figure 9-5.
    View in gallery
    Figure 9-5.

    Responses to Anticipated Shock

    (Anticipated elimination of MID at t + 20, recycled as lower capital taxes)

  • Figure 9-6.
    View in gallery
    Figure 9-6.

    Consumption Responses

    (Staggered vs immediate elimination of MID, recycled as lower capital taxes)

  • Figure 9-7.
    View in gallery
    Figure 9-7.

    House Price Responses

    (Staggered vs immediate elimination of MID, recycled as lower capital taxes)

Figure 9-1.

User Cost of Housing and Interest Rate Responses

(Permanent, immediate elimination of MID, recycled as lower capital taxes)
Figure 9-2.

House Price Responses

(Permanent, immediate elimination of MID, recycled as lower capital taxes)
Figure 9-3.

Nondurable Consumption and Residential Investment Responses

(Permanent, immediate elimination of MID, recycled as lower capital taxes)
Figure 9-4.

Housing and Fixed Capital Responses

(Permanent, immediate elimination of MID, recycled as lower capital taxes)
Figure 9-5.

Responses to Anticipated Shock

(Anticipated elimination of MID at t + 20, recycled as lower capital taxes)
Figure 9-6.

Consumption Responses

(Staggered vs immediate elimination of MID, recycled as lower capital taxes)
Figure 9-7.

House Price Responses

(Staggered vs immediate elimination of MID, recycled as lower capital taxes)