## Julian S. Alworth and Giampaolo Arachi

Print publication date: 2012

Print ISBN-13: 9780199698165

Published to Oxford Scholarship Online: May 2012

DOI: 10.1093/acprof:oso/9780199698165.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy).date: 16 July 2018

# The Role of Housing Tax Provisions

Chapter:
(p.61) 3 The Role of Housing Tax Provisions
Source:
Taxation and the Financial Crisis
Publisher:
Oxford University Press
DOI:10.1093/acprof:oso/9780199698165.003.0003

# Abstract and Keywords

The 2008 financial crisis has been the worst economic crisis since the Great Depression of 1929. It has been characterized by a housing bubble in a context of rapid credit expansion, high risk-taking, and exacerbated financial leverage, ending in deleveraging and a credit crunch when the bubble burst. This chapter discusses the interactions between housing tax provisions and the financial crisis. In particular, it reviews the existing evidence on the links between capital gains taxation of houses, interest mortgage deductibility, and characteristics of the crisis.

# 3.1 Introduction

The 2008 financial crisis hit the world economy severely. While taxes are not considered a proximate cause of the crisis, some aspects of tax policy may have led to increased risk taking and indebtedness (see Chapter 2). Tax incentives may indeed have exacerbated the behaviour of economic agents, leading them to wrong economic decisions. Among the tax factors, the widespread tax-induced bias to homeownership has attracted considerable attention, because excessive demand in the housing market, combined with lax lending practices, may have contributed to the speculative bubble in real estate prices.

This chapter proposes a detailed account of the manner in which tax provisions relating to the housing market may have led to the banking crisis. Monetary and regulatory policies opened up the possibility for a housing bubble that eventually burst and created a credit crunch because of a lack of confidence between actors on financial markets. Governments reacted by a combination of capital and liquidity injections, regulatory measures, and fiscal stimulus.

In most narratives of the financial crisis, the dynamics of the US housing market play a decisive role: in fact, the problems started with the housing market and the financial structure that was built on it. Not surprisingly, many commentators have found fault with some tax provisions that may have contributed to an overheated housing market. In particular, attention has (p.62) been focused on the tax treatment of residential housing capital gains and on the deductibility of interest expenses on mortgages.

Some commentators argue that the quasi repeal of residential housing capital gains taxation in 1997 may have fuelled the housing bubble. On the other hand, neither the OECD nor the IMF believe that this factor has played a significant role;1 moreover, the academic research that has analysed the dynamics of the US housing market tends to have reached the same conclusion (see also Chapter 2).

The role of the mortgage interest deductibility in the crisis is also controversial. There was no relevant change in the US tax rules on this tax break in the 2000s; the housing boom did not take place evenly across the country, although the federal tax system has a nationwide coverage. Housing prices went up both in countries where interest on mortgages was deductible and in countries where it was not or where it was deductible only within limits. Nevertheless, this tax break can be thought of as a catalyst in a chemical reaction: the deductibility did not cause the bubble, but it may have accelerated the run-up in prices. It remains true that the US regime is one of the most generous in an international comparison; while all other countries allow interest deductibility only for acquisition or renovation of residential buildings, the US tax code extends this allowance to other purposes (‘home equity loan’); moreover, the relatively generous limits to the benefit are capped on the amount of the mortgage, not on the amount of interest payments (as in all other countries). Since it is proportional to debt, the tax break is more relevant for riskier mortgages with higher interest rates and may have contributed to trigger ‘gambles’ on housing, especially in the context of ‘exuberant’ price expectations.

A meaningful comparison can be made between the 2008 financial crisis and the 1990s Scandinavian banking crises. In the late 1980s Norway, Sweden, and Finland experienced large credit and asset upswings, followed by severe downturns after the burst of asset prices. As in the recent US and world financial crisis, so in the Scandinavian banking crises: deregulatory measures, expansionary monetary policy, and lax risk analysis interacted and paved the way for a rapid credit expansion and increases in asset prices;2 an important role was played by the housing market dynamics, which in turn were probably affected by housing tax provisions; and the asset bubbles burst when interest rates started to increase.3 An interesting case in regard to the relationship between housing tax rules and financial crisis is Sweden: here the housing tax rules may have contributed indirectly to the price upswings (p.63) (the ‘catalyst’ argument) and directly to the bursting of the price bubble, through the 1990–1 tax reform, which substantially reduced the benefit stemming from the deductibility of interest payments, increasing the real cost of borrowing.4 With respect to the Scandinavian banking crises, the US financial crisis seems to have been even more related to the housing market developments; another distinctive and important feature of the current crisis regards the role of securitization (see below).

This chapter is organized as follows. Section 3.2 provides an introduction to developments of the 2008 financial crisis. Section 3.3 offers a reflection on whether specific tax provisions may have aggravated the crisis by encouraging homeownership and risky behaviour. Section 3.4 contains some final remarks.

# 3.2 The build-up to the 2008 financial crisis

## 3.2.1 General economic conditions before the crisis

The events leading to the financial and economic crisis that began in 2008 are heavily debated and the dust has not yet settled on the real causes of the crisis. The arguments set out in this chapter are, therefore, somewhat speculative and subject to debate, and they will eventually be judged by history. Yet, a majority of commentators point to several elements that have facilitated an easing of credit and an increase in risk taking.

Figure 3.1. Nasdaq Composite Index, 1993–2004

Source: Yahoo! Finance.

(p.64)

Figure 3.2. US Federal Reserve discount rate, 2000–9

Source: Federal Reserve.

The economic conditions in the early 2000s were characterized by the bursting of the dot-com bubble, which peaked in March 2000 and led to a pronounced decline in world stock-market indices in the following years (see Figure 3.1). The reaction of the Federal Reserve to this stock-market decline and the steady worsening of economic conditions were to reduce interest rates. Accordingly, the US Primary Credit Discount Rate was progressively lowered from 6.5 per cent at the peak of the bubble in mid-2000 to 1 per cent by mid-2003 (see Figure 3.2).5

A second characteristic of the world economy in the early 2000s was massive inflows of capital on international financial markets. The US Capital and Financial Account is illustrative of this phenomenon (see Figure 3.3).6 Between 1995 and 2000 it increased from 1.54 per cent to 4.25 per cent of GDP and continued to rise in the first half of the 2000s to peak at 6.10 per cent of GDP in 2006. The main driver of this expansion was net portfolio investment, which grew from $42.7 billion in 1998 to over$807 billion in 2007—a twentyfold increase over nine years (see Figure 3.4). As a result, in the first half of the 2000s the US economy was characterized by a rapid recovery in a low-interest-rate environment, despite a high degree of risk aversion in stock markets, following the tech bubble burst.

## (p.65) 3.2.2 Promotion of homeownership, deregulation, and subprime credits

Figure 3.3. US Capital and Financial Account

Source: US Bureau of Economic Analysis.

Figure 3.4. US Capital and Financial Account, components

Source: US Bureau of Economic Analysis.

In their search for new places in which to invest, many economic agents saw property as a safe and more profitable haven. The conditions were consequently slowly put in place for a housing bubble. Between 2001 and 2005 in the United States the number of houses sold increased by 41.3 per cent and the average Case–Shiller price index rose by 57.8. per cent and 43.1 per cent in (p.66)

Figure 3.5. Case–Shiller house price index

Source: Shiller (2005). See 〈http://www.irrationalexuberance.com〉.

nominal and real terms, respectively (see Figure 3.5).7 In addition to favourable economic conditions (low interest rates, large inflow of capital that needed to be recycled in the economy, and cold feet of investors towards stock markets), several regulatory measures also created incentives towards homeownership.

First, politicians wanted to expand homeownership, especially for poorer families. Two institutions played a particular role in this policy: Fannie Mae and Freddie Mac. The Federal National Mortgage Association (Fannie Mae) was created in 1938 under the Roosevelt administration to purchase and securitize mortgages in order to ensure enough liquidity for lending institutions. It became an independent body—albeit with implicit government guarantee—in 1968 and was complemented in 1970 by a competitor, the Federal Home Loan Mortgage Corporation (Freddie Mac), which achieved similar functions on this secondary mortgage market. The role of Fannie Mae and Freddie Mac was to purchase loans from mortgage sellers such as banks and other financial institutions, securitize them into mortgage-backed bonds, and resell them on the secondary market, guaranteeing the principal and interest of the loan in exchange for a fee. This mechanism proved to be a powerful instrument to refinance lending institutions with fresh cash and subsequently allow them to engage in additional lending (p.67) activities. The US administrations also used these agencies to expand housing credit to middle- and low-income families as well as in distressed areas.8

Second, the US tax system contained several incentives for homeowners to increase their use of mortgages. For example, the 1986 Tax Reform Act disallowed consumers to deduct interest payments on consumer loans (car loans, credit card loans, and so on). This created a perverse incentive for homeowners to use or refinance their home mortgages—whose interest payments were tax deductible—to pay off their other debts or to extract cash for personal expenses. This incentive became increasingly larger because of the wealth effect of ever-rising home values. In addition, the 1997 Taxpayer Relief Act simplified the tax treatment of housing capital gains and increased in many cases the tax exemption for these incomes—giving further incentives to buy houses. The effects of the Low Income Housing Tax Credit and the 2004 American Dream Downpayment Act provided further fiscal and support measures in favour of homeownership.9

In this context, financial institutions reacted by opening the credit tap, helped by more lax regulations. The 1999 Gramm–Leach-Bliley Act repealed some of the provision of the 1933 Glass–Steagall Act, which disallowed financial institutions to combine commercial, insurance, and investment activities and this might have led to more risk-prone attitudes from the part of commercial banks.10 Risk taking was also encouraged by relaxed rules on capital adequacy and new accounting standards. The decision on 28 April 2004 by the Securities and Exchange Commission to loosen the capital rules for large financial institutions (following their request) and to let computer models of those investment companies determine the level of risk of investment (that is, de facto self-monitoring) may have led to a sharp increase in the leverage of the main US financial institutions.11 This trend was also facilitated by the (p.68) Basel-II agreements, which entered into force in 2008 and gave more scope for financial institutions to assess their risks, as well as by the introduction of the International Accounting Standards in 2005, which forced companies to register gains and losses on financial assets immediately, possibly leading to more stock volatility.

In this context, the proportion of subprime mortgages12 soared from 7.2 per cent of the total in 2001 to over 20 per cent in 2005 and 2006 (see Figure 3.6). Gambling was also at play, as some studies pointed out that over a third of the houses bought were for investment or second-residence purposes, and those specific acquisitions were made with the hope that continued price increases would allow buyers to resell with profit. Accordingly, a third of the loans made in 2002 were either interest only (where only interest is repaid) or negative amortization loans (where less than the interest is paid during a first period and the accrued unpaid interest is added to the outstanding amount of the loan).13 Moreover, an increasing number of loans were granted as adjustable rate mortgages (ARMs)14 between 2001 and 2004—mostly for the two pre-cited types of loans—and this despite stabilizing interest rates, which possibly indicates an increasing number of credit-constrained borrowers (see Figure 3.7).

## (p.69) 3.2.3 The securitization of mortgages

Figure 3.6. US prime and subprime mortgages

Source: Joint Center for Housing Studies at Harvard University (2008).

Source: AMECO and Joint Center for Housing Studies at Harvard University (2008, 2009).

Note: Central government benchmark bond of ten years.

The spread of mortgages, in particular subprime loans, was largely helped by the development of new financial instruments, in particular the technique of securitization, which consists of pooling loans into an investment vehicle and then selling securities backed by payments for these loans. In the case of mortgages, those financial instruments are mortgage-backed securities (p.70) (MBS). Typically, the financial institution will buy the claims of thousands of mortgages and pool them into a so-called special purpose vehicle (SPV), a legal entity outside the balance sheet of a financial institution and hence not subject to capital requirements. The securities are separated in several tranches—senior, mezzanine (or junior), and equity (non-investment grade)—with a sequential preference for the claims (that is, the senior tranche has preferred claim on the proceeds over the other two and the mezzanine tranche has preference over the equity tranche). By doing so, financial institutions are able to rearrange the risk of the pool and to redistribute it across investors with different preferences.15 This, in turn, lowers the cost of lending and extends credit to borrowers with a lower credit quality.

An important development has been the issue of collaterized debt obligations (CDOs), a family of asset-based securities that is backed by diversified debt obligations such as mortgage-backed securities, corporate bonds, bank loans, credit-card debt, and so on.16 While an MBS is backed by mortgage payments, a CDO is backed by mortgage-based securities within a portfolio and therefore represents a resecuritization (Baily et al. 2008). The advantage of a CDO is that it allows financial institutions to rearrange the securities into new compartments within the CDO and to transform low-rated mortgage-based securities into high-rated CDOs. According to Baily et al. (2008), CDO issuances went from virtually zero in 1995 to over $500 billion in 2006 and virtually all CDOs issued after 2006 were backed by low-rated subprime MBS. This securitization process was itself helped by the emergence of a new class of derivatives that allowed the transfer of credit risk to a third party: the credit default swaps (CDS). These are common instruments, representing 73 per cent of the$2.3 trillion credit derivative products in 2002 (O’Kane 2005). The principle is that a third party agrees to assume the default risk of a specific asset in exchange for an income. This process allows the CDO issuer to shield from the risk and to increase the rating of its bonds. The CDS market has developed mainly outside organized markets (that is, they were over-the-counter operations) and grew exponentially from virtually zero in 2001 to about $15 trillion in 2005 and over$60 trillion in 2007 (Baily et al. 2008).

## 3.2.4 The bubble burst

With US inflation rising from 1.6 per cent in 2002 to 2.3 per cent in 2003, 2.7 per cent in 2004, 3.4 per cent in 2005, peaking at 4.3 per cent in June 2006,17 the Federal Reserve gradually raised interest rates from 1 per cent to (p.71) 5.25 per cent (see Figure 3.2) and the first cracks appeared in the housing market. Some borrowers, especially those with ARMs, started feeling the pain and eventually could not repay their mortgages. The number of foreclosures increased exponentially from 885,000 in 2005 to 1,259,118 in 2006, 2,203,295 in 2007, and 3,157,806 in 2008.18 The number of houses sold declined, and prices levelled off before plunging.

Financial institutions started to be hit, as they were heavily indebted and exposed via mortgage-backed securities, whose value are based on mortgage payments and house values. HSBC announced in February 2007 that it was writing down $10.5 billion of subprime MBS. This event was followed in April by the bankruptcy of New Century Financial, the largest US subprime lender. In July, the collapse of two hedge funds run by Bear Stearns because of subprime losses was another alarming sign of deterioration, as were the announcements of heavy losses in other financial institutions, which put some of them on the verge of bankruptcy (for example, Bear Stearn)—with in some cases bank runs (for example, Northern Rock in the UK19). The near collapse of the banking system happened in September 2008. On 7 September, ailing government-sponsored enterprises Fannie Mae and Freddy Mac were urgently nationalized. On 14 September, Merrill Lynch saw itself close to illiquidity and was sold to the Bank of America. The next day, Lehman Brothers filed for bankruptcy, and the day after, American International Group (AIG), one of the largest CDS20 providers, avoided bankruptcy only thanks to a$85 billion loan from the Federal Reserve.21 The uncertainty about external positions and liquidity or solvency of financial institutions led to a sharp drop in confidence among financial market participants. This led in turn to a sharp increase in the TED Spread—an indicator of perceived credit risk—which went over 300 basis points on 17 September 200722 and to a sharp fall in the interbank lending activities (see Figure 3.8). These financial problems spread into the real economy via a credit crunch, creating a drop in available funds for private investment.

The banking crisis also quickly spread to stock markets. The S&P-500 index started to decline from its peak of 1565.15 points on 9 October 2007 to (p.72)

Figure 3.8. US interbank loans

Source: Board of Governors of the Federal Reserve System (2011).

1251.70 points on 12 September 2008, a decline of more than 20 per cent in less than a year. On 15 and 17 September, amid turmoil in large financial institutions, it lost twice an additional 4.71 per cent. The descent into hell was not over yet, with stock crashes of 8.8, 7.6, 9.0, and 8.9 per cent on 29 September, 9 October, 15 October, and 1 December, respectively. On 9 March 2009, the S&P-500 index reached its lowest point: 676.53 points, only 43.2 per cent of its value fifteen months earlier.

# 3.3 Did taxes on housing contribute to the crisis?

The end of the speculative price bubble in the US housing market has been identified as an important trigger for the financial crisis. In Europe, Ireland and Spain faced similar price bubbles, and, when the international crisis hit, this led to a severe downturn in these two countries, which had formerly shown some of the best economic performance in the eurozone. Other European countries such as the UK, France, Sweden, and the Netherlands had experienced similar increases in house prices since the mid-1990s, albeit to a lesser extent.23

While real house prices rose in many countries, the same was true for the price-to-rent ratio, which is defined as the nominal housing price index divided by the rent component of the consumer price index. This is an indicator of the relative attractiveness to own a house versus renting it; it can be interpreted also as an indicator of the over- or undervaluation of the (p.73)

Figure 3.9. Price-to-rent ratios

Source: OECD (2010b).

housing market. As seen in Figure 3.9, since the mid-1990s the price-to-rent ratio has increased significantly in many countries, especially Ireland and Spain. Ireland also faced the most dramatic decrease after the peak had been reached in 2005. These dynamics point to an overvaluation of the housing market in the years prior to 2008 in several countries.

Given these observations, the question arises how taxes might influence house prices and price-to-rent ratios. To answer this question, we first recall the main results stemming from a very simple economic model of the housing market, which provides a basis for discussion of the effects of some basic housing tax provisions. Second, we compare the tax systems of different countries with respect to the US system. Finally, we discuss the possible role of housing tax provisions in the financial crisis.

## 3.3.1 The economic analysis of the housing sector

The decision to buy a house entails two economic dimensions: a consumption decision and an investment–production decision.24 The first facet is related to the decision of households to consume housing services, which is mainly related to the quality of the house. The household decides what type of house and in which location it would like to consume. The investment–production decision is related to the potential value increase of the property, as households take into account that owning a house is also an investment. (p.74) Housing is a durable good that can potentially be sold at a higher price, even after years of use. This makes the decision to buy a house more complex than consumption decisions for other goods that are based mainly on the price and on the consumer’s budget constraint.

A simple and quite general demand-supply model of the housing market (see Box 3.1) predicts that: (p.75)

• since housing supply is basically fixed in the short run, the housing market is subject to price overshooting in the face of demand shocks; the housing market is therefore intrinsically volatile;

• the deductibility of mortgage interests, by reducing the user cost of ownership, decreases the demand elasticity; it therefore increases the volatility of the housing market;25

• the increase in volatility could have negative effects, especially if agents form expectations (also partially) in an extrapolative manner, inducing prolonged price upswings or downswings not linked to ‘fundamentals’;26 in the best-case scenario, the choices of households and firms could be temporarily distorted; in the worst-case scenario, a price bubble may form;

• rise in expectations on housing price appreciation and more generous tax breaks on housing (for example, lower capital gains tax rates) may generate, in principle and under some conditions, unsustainable dynamics in the housing market.27

## 3.3.2 The taxation of housing in Europe and the USA

There is a great diversity of housing tax regimes across countries.28 International comparisons are difficult to carry out because of the complexity of tax codes (in terms of deductions, exceptions, threshold limits, and so on).

Table 3.1 summarizes the information concerning the tax treatment of mortgage interest expenses and imputed income for owner-occupied housing, capital gains on first-home selling, and property taxes on owner-occupied dwellings for a set of countries comprising Belgium, France, Germany, Ireland, (p.76) (p.77)

Table 3.1. The taxation of owner-occupied housing in Europe and the USA, 2009

Country

Taxation of imputed rents

Mortgage interest tax relief

Capital gains taxation

Property tax

Belgium

Yesa

Tax deductibility with a limitb

No

Noa

France

No

Tax credit for the first five years with a limit

No

Taxe d’habitation and Taxe foncièrec

Germany

No

No

No

Grundsteuer

Ireland

No

Tax credit for the first seven years with a limitd

No

No

Italy

No

Tax credit with a limit

No

No

The Netherlands

Yese

Tax deductibility without limit

No

Onroerende-zaakbelasting (OZB)

Spain

No

Tax credit with a limit on the amount of housing costs

Nof

Impuesto sobre bienes inmuebles

UK

No

No

No

Council Tax

USA

No

Tax deductibility

No

Yesg

with a limit on the amount of mortgage principal ($1 million) (if CG 〈$500,000)

a In Belgium the imputed rent is a ‘cadastral income’, which was last reviewed in 1975 and has been indexed to inflation since 1990. In the case of owner-occupation, the deemed income (after the deduction of some deemed expenses) is not subject to income tax, but only to an ‘immovable withholding tax’ (precompte immobilier), with a rate that depends on the region where the property is located (see Haulotte et al. 2010). Taking into account the municipal surcharges, the effective tax rate of the immovable withholding tax ranges from 18% to 50% (IBFD 2009). In the computations of the effective tax rate in Figure 3.10 we assume an immovable withholding tax rate of 34% (the average between 18% and 50%).

b The deduction pour habitation propre et unique refers to mortgage interest, mortgage capital, and to particular insurance premiums regarding the loan (Haulotte et al. 2010: 29). The limit to the deduction is €2,770 for the first ten years and €2,080 thereafter.

c France also levies a net wealth tax with specific rules for owner-occupied houses. More precisely, the impôt de solidarité sur la fortune is an annual tax on the excess of the overall value of assets over a certain threshold, with a 30% reduction of the value of the owner-occupied house (see Borselli et al. 2010).

d Since 1 May 2009 in Ireland the interest relief has been restricted to the first seven years of the mortgage. In 2010 Ireland started to phase out mortgage interest relief: the tax break will be abolished from 2018; for new loans the relief will be gradually reduced (see IBDF 2010).

e In the Netherlands the imputed income is calculated as a percentage (until 2008 up to 0.55%) of the market value of the property (since 2009 there has been no maximum imputed income (see IBFD 2009)).

f In Spain full rollover relief is available for the sale of the primary residence. Moreover, housing capital gains are exempt when realized by a taxpayer aged 65 or more.

g The state and local real estate taxes are deductible against the federal personal income tax.

Sources: IBDF (2009); Borselli et al. (2010).

Italy, the Netherlands, Spain, the USA and the UK (an overview of the property tax regimes can be found in the appendix to this chapter).29

From a theoretical point of view, under a comprehensive income tax, a fully neutral taxation of owner-occupation requires the taxation of imputed rents and capital gains of housing and the deductibility of mortgage interests.30 Real-world tax systems are anything but neutral. In fact, owner-occupancy is tax favoured with respect to renting in many countries, and with respect to most forms of return on personal savings: with only a few exceptions, imputed rents and capital gains on owner-occupied housing are not taxed; the tax relief on mortgages’ interest further reinforces the tax bias towards housing.

Table 3.1 shows that only Belgium and the Netherlands tax the imputed rent on owner-occupancy. Mortgage interest costs attract tax relief in all countries except Germany and the UK. In the Netherlands, Belgium, and the United States interest expense is deductible from the tax base (but in Belgium the deduction is capped at a given amount of interest payments, while in the USA the cap refers to the amount of mortgage principal), so the tax advantage depends on the marginal tax rate of the owner. In the other countries the tax relief for financing costs mainly takes the form of a tax credit, often with limited duration.31 Basically, no country in our set taxes capital gains on owner-occupied housing. Finally, with regard to property taxes, only Belgium, Ireland, and Italy do not tax the owner-occupied property; notice that in the USA it is possible to deduct state and local real estate taxes against the federal personal income tax.

To get an idea of the quantitative effects of the personal income-tax rules regarding imputed income, mortgage interests and capital gains, and the fiscal burden of the property taxes, we compute the effective average tax rate on owner-occupied housing using a simplified version of the IMF methodology, which does not consider transaction taxes32 (see Box 3.2 for a description of the methodology). The results of these computations are shown in Figure 3.10 (see the appendix to this chapter for the assumptions regarding property taxation).33

(p.78)

Figure 3.10. Effective average taxation of owner-occupation, Europe and the USA (%)

Source: our calculations. Data: IBFD (2009). Methodology: IMF (2009).

In all the countries the personal tax system provides incentives to owner-occupation: since the effective average personal tax rates are negative in all countries save Germany and the UK, housing investment is subsidized by the personal income-tax system. Owner-occupancy is tax favoured de facto by the (p.79) personal tax system also in Germany and the UK, since the effective average tax rate (equal to zero) is generally lower than the tax rates on alternative investments.

The implicit tax subsidy stemming from the personal income-taxation rules is particularly strong in the Netherlands and in the USA, where mortgage interest is deductible basically with no limits.34 The subsidy is also strong in Ireland; but notice that Ireland started to phase out this tax relief in 2009.

As regards property taxation, the fiscal burden is comparatively higher in the USA, Spain, Germany, and France than in the UK and the Netherlands (as noted above, Belgium, Ireland, and Italy do not tax owner-occupied property).

By considering the overall effective tax rate, the tax subsidy for owner-occupied housing is found to be the largest in the Netherlands, the USA, and Ireland.

When the international comparison refers, not only to interest tax relief, taxation of imputed rents and capital gains, and taxes on property, but also to transaction taxes,35 as in the IMF’s analysis, the results show large differences across countries.36 However, as regards the United States, the results are consistent with the IMF’s: overall, the USA stands out as having low taxes on owner-occupied housing (see Section 3.3.3.2 for a brief discussion on the cross-country correlation between measures of effective taxation and house prices).

## 3.3.3 Housing taxation and the financial crisis

The main direct cause of the financial crisis lies in the bursting of the US housing bubble, so, in assessing responsibility for the financial crisis, it is natural to begin by examining the structure and the dynamics of the US housing market, particularly its demand side. The focus here is on the possible role of the ‘tax factor’ in the US housing market dynamics.37

The IMF and the OECD do not consider tax rules as the main reason for the housing bubble: housing prices increased in countries with different tax systems, and there were no tax breaks clear and big enough to explain the (p.80) price dynamics that were observed.38 At the same time, many commentators have found fault especially with the tax treatment of housing capital gains and mortgage interest deductibility. Let us consider each of them in turn.

### 3.3.3.1 Capital gains taxation

Since the bursting of the US housing bubble, it has been asserted that the housing policies pursued in the USA since the mid-1990s are partly to blame for the financial crisis, particularly the policies aimed at increasing homeownership through access to mortgage loans for first-time buyers with low and variable incomes.39 A tax measure in the same vein is the repeal of capital gains taxation on home selling with the Tax Relief Act of 1997 (henceforth TRA97).

TRA97 generated a change in the tax treatment of housing capital gains. Previously, housing capital gains had been taxed when homeowners sold their houses, unless they resorted to the ‘rollover rule’ or to the ‘55-age rule’. The rollover rule allowed homeowners to postpone the taxation, provided they bought a house of equal or higher value within two years. The 55-age rule allowed sellers aged 55 or more to claim a one-time exclusion of $125,000 against the capital gains tax. TRA97 abolished both rules and allowed homeowners to claim$500,000 exclusion ($250,000 for singles) against the capital gains tax as often as every two years. Since the ownership and use tests to claim the exclusion are not very strict, it was often possible to get the tax benefit for a second home (Shan 2008). The repeal of capital gains taxation may have had an important impact on the effective taxation on housing: using the same IMF methodology and the same assumptions as above,40 it can be demonstrated that, following TRA97, the effective average tax rate on housing decreased from –6.37% to –17.12%.41 However, the overall effects of TRA97 on the US housing market are theoretically ambiguous,42 and the existing empirical evidence does not offer clear-cut answers.43 (p.81) Some commentators observe a structural break in the time series of US house prices between 1997 and 1998 (see Figure 3.5) and associate it with the repeal of capital gains taxation. In subsequent years, other factors became important: the rise in house prices, drawing investors’ attention; the end of the stock-market boom following the peak in March 2000; the attempt by the Federal Reserve to avoid a severe recession in 2001 by pumping liquidity into the system; the public policies aimed at increasing the homeownership rate. However, the new provisions on capital gains taxation of 1997 may have contributed to the house-price boom, ‘fuelling the mother of all housing bubbles’,44 playing the role of a precipitating factor. Their effects were then amplified by mechanisms involving investor confidence and expectations of market performance (besides the other factors mentioned above); adaptive or extrapolative expectations may have played a role in these amplification mechanisms.45 Other commentators hold that the repeal of the capital gains taxation did not play a significant role in the genesis of the financial crisis. The IMF considers the role of the 1997 measures unclear, since the elimination of rollover relief may have resulted in worse tax treatment for some taxpayers, and since house prices did not increase everywhere, which implied that other factors were at work.46 More importantly, perhaps, many scholars assign no significant role in the price boom to the 1997 break.47 ### 3.3.3.2 Mortgage interests deductibility In the USA it is possible to deduct interest costs on mortgages taken to buy, build, or improve a house (so-called home acquisition debt), up to$1 million.48

In general, mortgage interest deductibility, particularly when unlimited (as in the Netherlands) or with mostly non-binding limits (as in the USA), decreases the cost of ownership and tends to tilt households’ decisions whether to rent or buy a house towards ownership; it also encourages people to spend too much on housing, and it may actually end up subsidizing wealthier homeowners, who have higher marginal tax rates, and the construction (p.82) industry.49 Moreover, since the benefit is proportional to debt, the deduction is basically a subsidy to ‘gambles on housing’,50 and this could lead to excessive risk taking.51

As regards risk taking, since the second half of the 1990s, credit-scoring methods have been widely used in the USA to price lending. This has probably facilitated the access to credit for many high-risk borrowers. Obviously, given the amount of debt, the riskier the borrower, the higher the interest rate charged, and the greater the tax benefits from deduction. Unlike countries that cap the deduction at a given amount of interest, the United States establishes the cap to the mortgage principal, and this implies a tax favour to riskier borrowers.52

In contrast with other countries,53 in the USA it is possible to claim a deduction for interest on mortgage loans taken out for purposes other than house purchase—for example, to buy a car or pay for college tuition, up to $100,000 (so-called home equity loan). Home equity—the difference between the market value of the house and the loans secured by the value of the house—can be used as collateral. Home equity loans may have played an indirect role in leading up to the financial crisis. In fact, the run-up in US housing prices, along with the rise in home acquisition debt fuelled by ‘bull’ price expectations, may have directly fed the growth of home equity loans, thereby providing part of the mortgage raw materials for the strong growth of the securitization industry (see Section 2.3). This process may have been magnified by the deductibility of mortgage interests. The deductibility of interest on home equity loans clearly creates a bias to personal debt, encouraging people to prefer borrowing against home equity to other forms of borrowing, and to extract the equity from their home for personal and business reasons. The lowering over time of personal income-tax rates in the USA has reduced the size of the tax benefit stemming from the mortgage interest deductibility,54 which nevertheless remains substantial by international standards (see Figure 3.10). (p.83) However, the role of tax deductibility alone with respect to the recent bubble is unclear because of conflicting evidence. If we consider our set of countries (see Section 3.3.2), it is true that the Netherlands, the other country with strong interest deductibility and providing substantial tax benefits according to our computations, belongs to the ‘fast-lane’ set of countries, according to the IMF ranking based on the house price increases in the 1990s and 2000s.55 On the other hand, the UK was also a ‘fast-lane’ country, basically without having provision for any interest tax relief for most of the recent boom period.56 More importantly, as far as the USA is concerned, there was no break in the tax relief for interest expense to explain the housing boom. Moreover, the price dynamics in the USA differed across states and regions, although there are no interstate differences in interest deductibility. A possible indirect role of the interest deductibility for the US housing market dynamics may be related to the large increase in low- and no-downpayment mortgages during the second part of the price boom period,57 which was probably facilitated by the housing policies enacted in 2004 and subsequent years.58 Given the asymmetric treatment of personal debt and equity, the decrease of mortgage downpayments may have given rise indirectly to a tax break: since the cost of personal housing debt is deductible, unlike the opportunity cost of housing equity, the consequence of the increase of lower- or no-downpayment mortgages may have been an abrupt fall in the user cost of housing.59 Despite the inconclusive evidence based on simple time-series and cross-section comparisons, it is very likely that the interest tax relief may have somehow contributed to housing price inflation in the USA,60 along the (p.84) lines of a catalyst in a chemical reaction. The simple and very general economic model sketched above predicts that tax relief of the kind provided in the USA can contribute to the volatility of the housing market; and that mortgage interest tax relief can be a contributing factor of instability if it is coupled with low financing costs and/or ‘exuberant’ housing price expectations. In conclusion, tax incentives may have played a role in the development of the housing bubble, but the size of this role is difficult to assess, although the odds are that this role has been secondary to monetary policy and credit markets developments. # 3.4 Conclusions The 2008 financial crisis has already proved to be the worst economic crisis since the Second World War. The burst of a housing bubble in the United States led to a stop in confidence of investors towards all mortgage-based assets that had flourished in previous years and to uncertainties with regards to the financial exposure and liquidity of world major financial institutions. This banking crisis eventually spread to a stock-market crash and to a credit crunch in the real economy. The rapid expansion of credit and the increasing degree of indebtedness and risk-taking behaviour of financial institutions was a striking characteristic of the build-up to the crisis. In this context, one important policy question is whether tax systems may have created negative incentives, favouring risk. Several tax provisions in favour of homeownership may have led to increased purchases of houses in several countries. However, the available evidence is mixed when it comes to assessing whether different tax treatments have led to different price developments, suggesting that lax monetary policy and increased risk taking by lenders are more powerful explanations for the housing bubble. In turn, this risk-taking behaviour may have been exacerbated by tax provisions on the treatment of executive compensation and by tax arbitrage possibilities across different types of investors. Chapter 4 discusses these issues in depth. Countries have implemented strong policy responses to the crisis. In particular, many countries have taken tax measures as part of broader fiscal stimulus packages. They have, however, come short of changing tax systems. Two issues have attracted some attention and are evaluated in Chapter 5: the idea of a financial transaction tax and/or a financial activities tax to collect more revenues from the financial sector and to correct for negative externalities. (p.85) Appendix Property taxation on owner-occupied housing # Belgium There is no real estate taxation. # France The property tax (taxe foncière) is a local tax due by the legal owner of the house. It is based on the notional rental value of the property (valeur locative cadastrale). The tax is calculated by multiplying half of the latter value by coefficients determined by the local councils. In 2008 the average municipal rate was 17.76% (see Borselli et al. 2010). The dwelling tax (taxe d’habitation) is a local tax owed by the person who occupies a dwelling. As the property tax, it is assessed on the deemed rental value of the property. In 2008 the average municipal rate was equal to 13.85%; in special cases, state government adds a surtax (0.2% for principal residence) (see Borselli et al. 2010). Tax reductions are generally granted to taxpayers with dependants, for the principal residence. The last general revision of the valeur locative cadastrale was made in 1974 (and based on data for 1970) and updated in 1980. Every year the cadastral values are revised to reflect the changes of the characteristics of the properties; to some extent the annual revisions should reflect also the changes of the market values (Lefebvre 2010a: 626). # Germany The real estate tax (Grundsteuer) is levied annually by the municipalities on immovable properties. The tax is calculated by applying a federal tax rate of 0.35% to the fiscal value of the property (Einheitswert). The result is then multiplied by a municipal coefficient (Hebesatz). On average the final effective tax rate is equal to 1.9% (see IBFD 2011). The real estate tax is not deductible for income-tax purposes in case of owner-occupation. The fiscal value of the property is generally much lower than the market value (about 20–50% of the market value; see Mayer 2006). # (p.86) Ireland There is no real estate taxation. # Italy There is no real estate taxation on owner-occupation (since 2008). # The Netherlands The real estate tax (Onroerendezaakbelasting) is levied annually by the municipalities on immovable properties. The tax base is the fair market value for fiscal purposes (Wet waardering onroerende zaken (WOZ)) issued every year by the municipalities. On average the tax rate is equal to 0.1% (see 〈http://www.cijfernieuws.nl/ozbd.html〉). The tax is not deductible for income-tax purposes in case of owner-occupation. # Spain The real estate tax (impuesto sobre bienes inmuebles (IBI)) is levied annually by the municipal authorities on the possession of immovable property. The tax base is the cadastral value of the property (valor cadastral). The tax rate ranges between 0.4% and 1.1% (see Borselli et al. 2010). The tax is not deductible for income-tax purposes in case of owner-occupation. The cadastral value is revised with reference to the market values according to several procedures. The general revision is made at least every ten years (see Lefebvre 2010b: 1598; see also 〈http://www.catastro.meh.es/esp/procedimientos/Renovaciones〉). # UK The council tax is an annual tax on domestic property levied by local authorities. It is primarily a property tax; it is also a charge on local services users, since the payment is related to household size. Each property is assigned to one of eight bands (from A to H), according to its assessed capital value in 1991. Every year, each local council decides the bill for band D; the bills for the other bands are charged at a fixed proportion of the band D bill. On average, in 2008–9 the council tax per dwelling was equal to £1,145; the tax bill increased at an annual rate of 5% over the period 2004–9 (see 〈http://www.communities.gov.uk〉). # USA Property taxes are imposed by local municipalities and counties in each US state. The tax liability is generally calculated by multiplying a tax rate by an assessment ratio (the part of the value of the property that is taxed) by the market value of the property. On average, in 2009 the tax bill was 1.04% of home value (see 〈http://www. (p.87) taxfoundation.org/research/topic/89.html〉). The state and local real estate taxes are deductible for the purposes of income taxation. # Assumptions used in the computations of the EATR for property taxes It is well known that cadastral values are usually lower than market values. Hence, in countries where property taxes are based on cadastral values (as in France, Spain, and Germany), the burden of these taxes is usually lower than in countries where these taxes are more closely related to market values (as in the Netherlands and the USA). To consider these differences, for the countries where property taxes are based on cadastral values, we assume that at the beginning of the holding period the fiscal value of the property is lower than the market value by a given percentage; then we keep the property taxes constant in real terms over the holding period (we assume a 2 per cent inflation rate). More precisely, we assume an initial 65 per cent cut of the market value in Germany (see Mayer 2006) and a 33 per cent cut in France (for which we do not have more precise information). For Spain we assume an initial fiscal value equal to the market value (since the cadastral values are more updated than in other countries; see 〈http://www.catastro.meh.es/esp/procedimientos/Renovaciones〉). For the UK, we use actual data on the council tax. More precisely, we use the average council tax bill per dwelling for 2008–9 and we let it grow at a 5 per cent rate over the holding period (source: 〈http://www.communities.gov.uk〉). For the Netherlands and the USA, property taxes are computed with reference to the market value of the properties. ## Notes: The authors thank Jean-Pierre De Laet, Vieri Ceriani, Stefano Manestra, David Pitaro, Giacomo Ricotti and Alessandra Sanelli for useful comments. This chapter was written by Mr Gaetan Nicodeme (European Commission, Université Libre de Bruxelles, CEPR and CESifo), Mr Thomas Hemmelgarn (European Commission), and Mr Ernesto Zangari (Banca d’Italia). The views expressed in this chapter are those of the authors and do not necessarily reflect the official positions of the respective institutions with which the authors are affiliated. (1) See IMF (2009a) and OECD (2009a). (2) See Englund (1999: 80). (3) The increase of interest rates was due not only to a change in the stance of monetary policy, but also to the effects of German reunification (Englund 1999: 89). (4) See Agell et al. (1995) and Englund et al. (1995). (5) Note that the Federal Reserve most certainly also tried to combat the economic consequences of the 11 Sept. 2001 terrorist attacks. The US economy was also in a context of low inflation, if not of deflation risk, which facilitated an ease in monetary policy. (6) The Capital and Financial Account is composed of the net capital transfers, the change in the domestically owned assets abroad, and the change in foreign-owned assets at home. It mirrors the current account (which is composed of the trade balance and the net unilateral current transfers). (7) See 〈http://www.census.gov〉 for data about house sales. For a description of the Case–Shiller index, see 〈http://www.standardandpoors.com/indices/sp-case-shiller-home-price-indices〉. (8) See the 1977 Community Reinvestment Act (CRA), which was extended by the 1992 Federal Housing Enterprises Financial Safety and Soundness Act and scrutinized by the 1995 New Community Reinvestment Act, or the decision of the Department of Housing and Urban Development in 2000 to order Fannie Mae to devote half of its business to poorer families, which was increased to a 56% goal in 2004. (9) The Low Income Housing Tax Credit was part of the 1986 Tax Reform Act and provides developers with tax credit for equity investment when investing in low-income units in a housing project. The 2004 American Dream Downpayment Act provided downpayment grants of a maximum of$10,000 or 6% of the purchase price of the house (whichever amount was larger) to first-time homebuyers with annual incomes that do not exceeded 80% of the area median income. See also Gale et al. (2007).

(10) See Lloyd in Chapter 8.

(11) See 〈http://securities.stanford.edu/news-archive/2004/20040428_Headline08_Drawbaugh.htm〉; 〈http://www.nytimes.com/2008/10/03/business/03sec.html〉. Between 2003 and 2007 the leverage of the top five US financial institutions evolved as follow: Lehman Brothers from 22.7% to 29.7%, Bear Stearns from 27.4% to 32.5%, Merrill Lynch from 15.6% to 30.9%, Goldman Sachs from 17.7% to 25.2%, and Morgan Stanley from 23.2% to 32.4%. In 2007, their total debt amounted to $4.1 trillion, a third of US GDP (sources: Wikipedia using annual reports 〈http://www.lehman.com/annual/2007/fin_highlights〉; 〈http://www.bearstearns.com/sitewide/investor_relations/sec_filings/proxy/index.htm〉; 〈http://ir.ml.com/sec.cfm?DocType=Annual&Year=2008〉; 〈http://www2.goldmansachs.com/our-firm/investors/financials/current/annual-reports/revised-financial-section-2007.pdf〉; 〈http://www.morganstanley.com/about/ir/shareholder/10k2007/10k11302007.pdf〉). (12) By definition, a subprime loan is a loan that does not meet the ‘prime’ standards and is consequently risky. There may be various elements that make the loan fail the ‘prime’ test (e.g., length, structure, etc.). In this context, it is the profile of the borrower and/or the difference between the loan and the value of the house or the collateral. Loans are usually classified based on the guidelines of government-sponsored enterprises (GSEs). When a credit fulfils the GSE’s criteria, it is labelled ‘conventional’. When the loan fulfils all guidelines but the amount of the credit (usually loans above$300,000), it is labelled as ‘jumbo’. In those two cases, the creditworthiness of the borrower is not questioned and both loans are ‘prime’ loans. Non-prime loans can be Alternative A, when, for instance, the borrower has income that is difficult to assess (e.g., self-employed), a high debt-to-income ratio, little documentation, or several mortgaged houses. In this case, the creditworthiness is not questioned, but there is a higher risk. Non-prime loans can also be home-equity loans, which is a heterogeneous category of second- and first-lien mortgages with high loan-to-value ratios, home improvement loans, and revolving home-equity lines of credits. Finally, non-prime loans also includes the subprime loans with low-credit-quality borrowers (Fabozzi 2005).

(13) The Economist (2005).

(14) For subprime mortgages, the proportion of fixed-rate mortgages dropped from 33.2% in 2001 to 18.6% in 2005, while the bulk of the loans were of hybrid nature (i.e., with a fixed rate during an initial period of 2–3 years and then adjustable based on a reference rate) and not pure ARM. From 2005 the share of balloon mortgages in subprime mortgages jumped to reach 25–30%. Those mortgages require a large final payment. Note also that 55–60% of subprime mortgages were originated to extract cash, while only 30–40% of the loans were to buy a house (Demyanyk and Van Hemert 2009).

(15) See Fabozzi (2005) for a description of these instruments and Baily et al. (2008) for an account description of the processes.

(16) See Fabozzi (2005: chs 30 and 31).

(18) See 〈http://www.realtytrac.com〉 for data sources.

(19) Northern Rock asked and received liquidity support from the Bank of England in September 2007 and was eventually nationalized in Feb. 2008.

(20) Credit default swaps (CDSs) are ‘insurance’ contracts through which the buyer of the CDS hedges the risk of default of an investment in a financial asset (the reference instrument). The buyer of the CDS provides the seller with payments in the form of fees or premiums, and, in exchange, the seller provides the buyer with a payoff if the reference instrument suffers a credit event (e.g., default of the counter-party).

(21) See Wibaut (2008) for an excellent description of the events.

(22) The TED spread is the difference in basis points between the short-term interbank rate (i.e., the LIBOR) and the three-month US treasury rate. Its historical fluctuation is between 10 and 50 basis points. On 10 October 2008 it reached a record 465 basis points 〈http://www.tedspread.com〉.

(23) See Hilbers et al. (2008).

(24) A detailed analysis of the functioning of the housing market can be found in Pozdena (1988).

(25) In general, the price sensitivity of demand for housing tends to fall with the extent of preferential tax treatment for housing and with the expected rate of housing price appreciation (see Van den Noord 2005).

(26) For the USA, see Case and Shiller (1988). See also the general discussion in Poterba (1991).

(27) The model sketched in Box 3.1 can easily account for disequilibria dynamics. Suppose, for example, that, for whatever reason, the user cost of ownership becomes equal to zero. This can happen because of: either a decrease in the (net) mortgage interest rate (for instance, because of more generous interest deductibility and/or lower monetary policy interest rates), given expected housing price appreciation; or a sudden increase of the expectations of housing price appreciation; or a decrease of taxes on housing capital gains; or a combination of the previous factors. The right-hand side of equation (1.1) in Box 3.1 becomes equal to zero. The left-hand side can be equal to zero only when the demand for housing is infinite. With very strong demand for housing there will pressure on prices in the short run (given the low short-run supply elasticity). Regardless of how expectations are formed, agents will anticipate higher prices, and this would push the user cost into negative territory, with a further increase in demand, and so on. Here we have a vicious cycle—a price bubble process—which can be rationalized even by a very simple model, with very general assumptions.

(28) For a review of housing tax regimes in Europe, see Hilbers et al. (2008). See also ECB (2003) and Van den Noord (2005). By comparing the information in these papers with ours, it is possible to get a picture of how housing taxation changed in the decade prior to 2007.

(29) Tax information refers to the 2009 tax codes reported in IBFD (2009). See also Borselli et al. (2010).

(30) See Van den Noord and Heady (2001: 30) and IMF (2009a: 17).

(31) For example: in Spain, taxpayers are allowed to set off 15% of the costs incurred for acquisition or renovation of a primary residence against their income-tax liability, up to €9,015 (i.e. the maximum credit is €1,352); in Ireland, for first-time buyers, the relief—given at source with the effect of reducing the borrower’s interest payments—takes the form of a tax credit at a rate of 25% for years 1 and 2, 22.5% for years 3, 4, and 5, 20% for years 6 and 7 (the interest relief is restricted to an interest payment of €20,000 for a couple); in France, interest on loans for purchase or the construction of the principal residence entitles the taxpayer to a 20% tax credit for the initial five-year period of the loan (40% for the first twelve months), up to €7,500 per year for a couple (i.e., the maximum credit is €3,000 in the first year and €1,500 for the remaining four years); in Italy, interest on mortgage loans taken to build or buy the principal residence entitles the taxpayer to a 19% tax credit up to €4,000 (i.e., maximum credit equal to €760).

(32) Which can arguably be substantial for some countries (see IMF 2009a).

(33) For Belgium, as Van den Noord (2005: 36), we assume that the imputed income is the same fraction of the value of the unit of housing as in the Netherlands (0.55%). To take into account the difference between the two tax systems, in contrast with Van den Noord’s analysis, we consider a ‘fiscal value’ of the house lower than the market value. We assume a 33% cut of the market value of the house. Moreover, we assume that the Belgium ‘fiscal’ imputed income grows at the rate of 2% (rather than at 5%, as in the Netherlands). For the sake of comparison, assuming no cut of the market value and a 5% growth rate for the ‘fiscal’ imputed income, the effective tax rate on housing would be –0.4% (rather than –1.3% of Figure 3.10); assuming a 50% cut of the market value and a 2% growth rate for the imputed income, the effective tax rate would instead be –1.6%.

(34) In the countries where the tax allowance related to mortgage interest is limited, the EATR of the personal income-tax system depends on the value of the house: once the limit of the tax break is reached, the EATR increases with the value of the house, and it has an upper bound that depends on the overall housing (personal) tax regime. For instance, if the imputed income is not taxed, the EATR is negative and will tend to zero as the housing income increases. This happens because the tax break becomes less relevant as (the present value of) the total housing income gets higher. If we assume an initial house price equal to € 250,000, and we keep all the remaining assumptions used in the main text, the EATR of the personal income-tax system would be –0.8% for Belgium (against –1.3% of Figure 3.10), –3.7% for France (against –1.9%), –7.3% for Ireland (against –6.1%), –2.7% for Italy (against –1.4%), and –4.9% for Spain (against –2.4%). Notice that the ranking of the countries is basically unaffected by the change of the assumption regarding the house price.

(35) Which can arguably be substantial for some countries (see IMF 2009a).

(36) See IMF (2009a: 20–1a).

(37) For in-depth analyses of the dynamics of the US housing market, see Case and Shiller (2004); Glaeser et al. (2005); Himmelberg et al. (2005); and Shiller (2005).

(38) See IMF (2009a) and OECD (2009a).

(39) See Katz (2009).

(40) In the computations we use the highest marginal tax rate in 1996 and 1997, which was equal to 39.6%. The capital gains tax rate applied for 1996 is 28% (see Shan 2008).

(41) The average tax rate in 1996, –6.37%, is the average between the tax rate for homeowners aged 55 years or more at the time of selling (–8.23%) and the tax rate for homeowners aged less than 55 years at the time of selling who decide to buy after ten years (–4.51%).

(42) For example, TRA97 also lowered all long-term capital gains tax rates, which were further reduced in 2001 and 2003 (see Shan 2008). The reduction of the capital gains tax rate with the TRA97 may have had effects on the market for rental properties (where the rents are determined), then indirectly affecting the market for owner-occupied houses (see Box 3.1, equation (1.)). The lower long-term capital gains tax rates may have allowed the building of rental projects in which landlords could also break even with lower rents. And lower rents could have eased the demand pressure on the market for owner-occupied houses, contrasting the possible demand pressure coming from the repeal of capital gains taxation on first-home selling.

(43) See Bier et al. (2000); Cunningham and Engelhardt (2008); Biehl and Hoyt (2008); Shan (2008); and Quayes (2010).

(44) See Smith (2007); Bajaj and Leonhardt (2008); and Gjerstad and Smith (2009).

(45) See Shiller (2005: 69).

(46) See IMF (2009a).

(47) See Case and Shiller (2004); Glaeser et al. (2005); Himmelberg et al. (2005); and Shiller (2005). Burman (2008) argues that the new capital gains tax rules were unimportant with respect to the bubble, stressing that the previous rules raised little revenue.

(48) Mortgage interest is an itemized deduction. In the US system, most taxpayers can choose to deduct either the total amount of itemized deductions or a standard deduction; the latter depends on the filing status of the taxpayer. In general, itemized deductions are chosen by wealthier taxpayers (Glaeser and Shapiro 2002).

(49) See Glaeser (2009).

(50) See Glaeser (2009) and Sullivan (2005).

(51) One could argue that there could be positive externalities associated with homeownership and housing consumption that might be worth subsidizing through mortgage interest deductibility. These externalities are, however, very difficult to measure, and, moreover, interest deductibility appears to have been ineffective in promoting homeownership in the USA (see Glaeser and Shapiro 2002 and the references therein).

(52) The same holds true in countries, such as the Netherlands, where there are no limits to interest deductibility.

(53) In the Netherlands it was possible to claim interest deductions on equity withdrawals until 1996.

(54) See Poterba (1992).

(55) See Hilbers et al. (2008).

(56) The UK phased out interest deductibility over the period 1974–99. First, a ceiling on the mortgage principal eligible for deduction was introduced. Then, the rate at which it was possible to claim the deduction was gradually lowered to zero (OECD 2000: 151).

(57) According to the surveys conducted by the National Association of Realtors, in 2003 the median downpayment for first-time homebuyers was equal to 6%, a figure that fell to 2% in the period 2004–7. The median downpayment for repeat homebuyers also declined starting in 2004, although to a lesser extent (see 〈www.realtor.org〉).

(58) On 16 Dec. 2003 the American Dream Downpayment Act was signed into law, with a view to assisting low-income first-time homebuyers by providing downpayment. Among other things, the Act expanded the supply of no-downpayment mortgages for first-time homebuyers (US Department of Housing and Urban Development 2005).

(59) As an example, using the highest marginal tax rate of the federal income tax in 2003 and 2004 (35%), and assuming a downpayment to buy the house equal to 5% in 2003 and 0% in 2004, the effective average tax rate computed with the IMF methodology (see Section 3.2) would decrease from –17.97% to –18.92%. Since the reduction of downpayments referred especially to low-income first-time buyers, it is reasonable to compute the change of the effective taxation also with the lowest income-tax rate (10%); in this case the effective average tax rate would decrease from –5.13% to –5.41%. Of course, it is hard to say whether and to what extent these changes in the economic advantageousness of the housing investment may have been statistically significant at the margin for housing market dynamics.

(60) See Sullivan (2008) and Surowiecki (2009).