Printer Friendly

Fiscal, distributional and efficiency impacts of land and property taxes.

1. Introduction

Many developed countries, including New Zealand, face the prospect of significant structural central government budget deficits (IMF, 2009; New Zealand Government, 2009). These pressures make it sensible to revisit both expenditure decisions and revenue-raising options. This paper addresses aspects of the latter, focusing on land and property taxes. Some economies raise a material proportion of tax revenues by way of land and/or property taxes. (1) In New Zealand, local government raises approximately 60% of its revenue requirements through local authority "rates" variously levied on land values or capital values of properties (McCluskey et al., 2006).

We analyse fiscal, distributional and efficiency effects of land taxes and/or property taxes. We model the effects of a fiscally-neutral change to the tax structure, whereby the additional revenues raised through such taxes are used to reduce other tax revenues, with expenditures being left unchanged. Our analysis covers the effects both of a 'land tax' (defined as a tax on land value) and of a 'property tax' (defined as a tax on the capital value of property, being the sum of improvements plus land value). (2) For much of the analysis we will be specific about whether we are dealing with a tax solely on land value (a land tax) or on total capital values (property tax). In some cases, where we wish to be more general, we refer to land/property taxes. We utilise existing valuations (rateable values) performed by Quotable Value New Zealand (QVNZ) for all New Zealand properties. These valuations, which split capital values into land and improvement values, have an existing statutory basis and are used as the base for local authority rates revenues.

We summarise prior treatments of land and property taxes, focusing particularly on efficiency and distributional (equity) considerations. We follow this summary with a partial equilibrium treatment of the effects of a land/property tax on individual land/property values. While the partial equilibrium analysis cements key concepts, it ignores system-wide effects that may produce different results in aggregate. We adopt a general equilibrium approach to gauge economy-wide impacts of a property tax. The general equilibrium model, designed to reflect key properties of the New Zealand economy, produces insights that are not apparent from a partial equilibrium approach.

The paper's empirical work assesses fiscal and distributional impacts within New Zealand that might occur following a shift towards land/property taxes. Unless otherwise stated, property values and other data refer to 2006. Fiscal implications depend on the breadth of the tax base and the specifics of the tax. We estimate the quantum of tax revenue that may be raised by certain land and property tax variants. In order to examine distributional impacts, we utilise data at differing levels of aggregation. We use census and QVNZ data pertaining to area units to examine relationships between household incomes, land taxes and property taxes. Separately, we use household level data from Statistics New Zealand's Survey of Family Income and Expenditure (SoFIE) to examine relationships between property values, household incomes, household wealth and other variables that are relevant for distributional considerations. Coleman and Grimes (2009), henceforth CG, present additional empirical results regarding distributional outcomes.

2. Context

Taxation reduces the disposable incomes of those paying the tax in order to provide sufficient revenue to meet government expenditures. As well as reducing disposable incomes, the design of the tax system has distributional impacts and generally distorts economic activity relative to activities under a tax-free environment. (3) In light of these effects, several properties of "good' tax systems are commonly postulated, including: allocative efficiency, dynamic efficiency, administrative efficiency and transparency, minimising avoidance/evasion, horizontal equity and vertical equity.

Land taxes are an ancient form of taxation (Dye and England, 2009) and have commonly been recognised as meeting at least some of the tests for a good tax system. Mill (1865, Book 5, Chapter 2, [section] 5) supported adoption of a special kind of land tax, levied on the increment to land values over and above those at a fixed point in time. His reasoning was that the increment in land values was due to general societal influences and that this increment should form the basis for government revenues required for the upkeep of society. George (1880) favoured a land tax as a form of taxation that does not diminish effort or investment while at the same time taxing private value earned from community efforts. The analytical basis for this approach was rooted in the insight of Ricardo (1817) that land values impound the rents available to land-owners arising from location-specific factors. Modern spatial economics analyses of urban development and infrastructure investment embody a related analytical approach (Roback, 1982; Haughwout, 2002).

In efficiency terms, a first-best tax (e.g. a lump-sum tax (4)) does not alter the structure of production, consumption or investment relative to the untaxed economy (Ramsey, 1927). Land can be treated as an item that has (virtually) a completely inelastic supply, with the quantity being given by 'nature' and so a land tax fulfils this criterion. (5) The allocative efficiency properties of a land tax do not automatically flow through to a property tax since improvements are subject to tax under a property tax system, whereas they are not taxed under a land tax. Thus, the supply of improvements is affected by a property tax, resulting in distorted resource allocation (McLeod et al., 2001, p. 31). (6)

Given the existence of the local authority rating system, a central government land/property tax would perform well in terms of administrative efficiency. All valuations required to provide the tax base are already performed and taxes (rates) are already levied comprehensively on property owners by Territorial Local Authorities and Regional Councils. A central government tax could be added as an adjunct to the current system with little additional administrative cost. Furthermore, the ability to avoid (or evade) the tax is virtually non-existent since the land/property is valued by an independent agency and is available as collateral in cases of non-payment of tax.

Taxation of land/property would extend the central government tax base, not just by taxing an asset that is currently not taxed at central government level, but also by taxing non-New Zealanders. Foreign owners would be obliged to pay the tax (as they currently do for rates). In practice, existing foreign owners of land/property at the time of the tax's establishment would bear the present discounted value of the future tax stream.

One complication of land/property taxes (and of local authority rates) is that some households are "property rich but income poor': this may particularly be the case for retired households. In these cases, systems already exist within some local authorities whereby rates payments can be accrued against the value of the property, to be met when the property is sold or from the estate. (7) In these cases, government would accrue tax owed to it and debt-fund the lost cash-flow, backed by the accrued tax asset.

Distributional impacts of land taxes depend on the direct impact of the tax, the impacts of other accompanying fiscal changes, and on general equilibrium reactions of asset and other prices to the package of tax changes. Plummer (2009) reviews international evidence on distributional impacts of a switch between land and property taxes, finding that area-specific features are important in determining both who gains/loses, and the overall progressivity/regressivity of such a change. The lack of consensus concerning distributional impacts of land/property taxes makes a New Zealand-specific analysis of the effects important if such a tax were to be considered here. (8)

3. Partial equilibrium effects of a land tax

3.1. Outline

Initial effects of a change in land/property taxes can be ascertained through partial equilibrium models of land/property valuations. (9) Partial equilibrium calculations leave out broader economy-wide effects that may affect the market in question. We provide one example to show the potential impact of such feedbacks prior to a more comprehensive general equilibrium analysis in the next section. Our main focus in the examples that follow concerns the effects of a land tax rather than a property tax for the efficiency reasons discussed above. We analyse one case with a property tax for comparison and also examine variants of a land tax, including that proposed by Mill. Additional variants, including a 'betterment tax' designed to capture increases in land values following investment in local infrastructure, are analysed by CG.

3.2. Simple land tax

Consider the purchase price of a plot of land at the end of year i = 0 that pays the owner an after-income-tax rental stream of [Y.sub.i] = [(1 + [pi]).sup.i] Y[r,k,t] in years i = 1 ..., [infinity], where [pi] is the annual rate of land rent inflation. For simplicity, [pi] is set equal to the general rate of inflation unless otherwise specified. Rents may reflect the imputed value of the property to the owner-occupier, or may be the explicit contractual amount paid by a tenant to a landlord. In general, market rents are likely to be a function of the real interest rate, r, the land tax rate, t, and other costs or benefits associated with the land, k, which may include local authority rates payments, maintenance charges and any expected rate of real capital gain/loss on the land. The parameter k is expressed as a ratio of the land value. For simplicity, [pi], r, and k are treated as known, fixed rates. The nominal discount rate is given as [(1 + r)(1 + [pi]) - 1], for which a close approximation is r + [pi]. When the tax rate is t, the value of the property at the end of year i is denoted [V.sup.t.sub.i].

The purchase price of the property at the end of year 0 is given by the discounted value of future rents less tax and other payments:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)

From the solution of a discounted infinite sum:

[V.sup.t.sub.0] = Y[r,k,t]/r - k[V.sup.t.sub.0]/r - t[V.sup.t.sub.0]/r (2)

Equating terms and solving for [V.sup.t.sub.0] gives the purchase price:

[V.sup.t.sub.0] = Y[r,k,t]/r + k + t (3)

This expression specifies a general relationship between rents and property values, without stating how the level of either is determined. Consider an example where Y is constant (i.e. independent of r, k or t), there is a land plot with initial rent of $10,000 p.a., an annual real discount rate of 0.05, k = 0.00 (expected annual costs equal expected annual gains), and the initial tax rate is zero (t = 0.00). Denoting [V.sup.0.sub.0] as the tax-free land value, [V.sup.0.sub.0] = $200,000, while if a 1% p.a. land tax (t = 0.01) is imposed [V.sup.t.sub.0] = $166,667. In general, manipulation of equation (3) demonstrates that if real rents are constant, the proportionate change in land value is:

[V.sup.t.sub.0] - [V.sup.0.sub.0]/[V.sup.0.sub.0] = -t/r + k + t (4)

Equation (4) indicates that, given the imposition of a land tax, the proportionate change in property values will be more marked: (a) the lower is the discount rate, and (b) the lower is k. The latter case may occur where real capital gains expectations are 'high'. These results indicate that the imposition of a land tax reduces the wealth of existing land owners if rents do not change. In the absence of borrowing constraints, the effect on a prospective home-buyer in year 0 is neutral (abstracting from the positive effects of any recycled tax revenues). The reason is that the purchase price is reduced by the exact amount of the (discounted) tax payments due on the property. If the prospective purchaser faces a borrowing constraint, the situation may change. With a debt servicing constraint (relative to income), and with unchanged income, the imposition of the tax will not alter the severity of the constraint if tax payments are included in the servicing constraint. Where the borrower is constrained by having insufficient equity for a deposit, house purchase will become easier with the tax since the house purchase price will be reduced. However, the general relaxation of this constraint may act to raise house prices relative to the level shown for the unconstrained case. The general equilibrium analysis of the following section analyses these issues in greater depth.

3.3. Endogenous rents

In general, rents are a function of the tax rate and may change in a step fashion upon imposition of a new land tax. For instance, the revenue from a land tax may be recycled by way of an income tax reduction, in which case disposable incomes (prior to property taxes) would rise. A general equilibrium model is required to determine the impact of a change in land taxes on the level of rents and property prices, as well as the ratio of the two. We can take account of the general equilibrium effects analytically by denoting the new initial rental level consequent on a shift from a tax-free environment to one with land tax at rate t as Y[r,k,t] = Y[r,k]Z[t]. Manipulation of equation (3) in this case produces the expression in equation (5) for the change in land value consequent on the introduction of the tax:

[V.sup.t.sub.0] - [V.sup.0.sub.0]/[V.sup.0.sub.0] = (Z - 1)(r + k) - t/r + k + t (5)

Using our previous example (Y = $10,000; r = 0.05; k - 0.00; t = 0.01), equation (5) indicates that land prices would not fall if Z were at least 1.2 (a 20% upward shift in land rents).

3.4. Property tax

An analysis of the impacts of a land tax on land prices can be extended to an analysis of the impacts of a property tax on the property's capital value. In order to model this generally, we allow the tax on land value (t) to differ from the tax on improvements (s), and allow other costs or benefits to differ, with rate k for land and j for improvements. The initial annual rental stream from improvements is denoted M. We denote the year 0 value of the property as H0. Since [H.sub.o] equals the value of land plus the value of improvements, we can use the same methods as in equations (1)-(3) to derive:

[H.sub.0] = [Y/r + k + t] + [M/r + j + s] (6)

Assuming no changes in Y or M after introduction of a tax, and assuming t = s and k = j (for simplicity), the impact of a change in property taxes from zero to t (= s) is again given by equation (4). If j > k, the proportionate effect on property prices of the introduction of a property tax (with t = s) will be less than that indicated by equation (4). In practice, we expect j > k, firstly because maintenance costs on improvements are likely to exceed those on land and, secondly, because land availability constraints mean that urban land may be less elastically supplied than improvements, in which case real capital gains expectations for land may exceed those for improvements. We can examine the impacts of the imposition of a land tax only (i.e. s = 0) on the overall property price. Using the same terminology as before, we can manipulate equation (6) to show the proportionate change as:

[H.sup.t.sub.0] - [H.sup.0.sub.0]/[H.sup.0.sub.0] = -t(r + j)/(r + k + t)[(r + j) + (r + k)(M/Y)] (7)

If k = j, the proportionate change in property value from a land tax is given by:

[H.sup.t.sub.0] - [H.sup.0.sub.0]/[H.sup.0.sub.0] = -t/(r + k + t)[1 + (M/Y)] (8)

Equation (8) has consequences for the distributional effects of a property tax relative to a land tax amongst existing property owners. The proportionate price drop that results from a land tax is relatively large for properties with a relatively small value of improvements relative to land. Consequently, land-extensive productive activities bear a proportionately larger fall in property values than land-intensive productive activities.

3.5. Incremental tax base

One approach to alleviating initial cash-flow impacts of the tax is to levy the tax only on increments of land value over and above some initial level, as suggested by Mill. By taxing only the increment in property values, the tax becomes a form of capital gains tax on property. This tax is more complicated to model because taxation of incremental value still causes the initial land value to fall by the discounted amount of the future tax stream. Thus, the increment above initial (pre-tax) values will disappear for some years until such time as capital gains are sufficient to raise land value beyond the pre-tax level. We assume that there is some threshold level of land value, X, that is not taxed, and that all increments above this level are taxed. We set X such that X [less than or equal to] [V.sup.t.sub.0]): thus, land values do not fall below the threshold value even after the imposition of the (incremental) tax. The expression determining land value (with rents fixed at Y) becomes:

[V.sup.t.sub.0] = Y/r + k + t + rtX/(r + [pi])(r + k + t) (9)

The second term in equation (9) is the "rebate" on the tax-free value of land, X. One can solve for the threshold value that equals the new level of land prices by equating X with [V.sub.0]:

[V.sup.t.sub.0] = X = (r + [pi])Y/(r + [pi])(r + k + t) - rt (10)

For our standard example (Y = $10,000; r = 0.05; k = 0.00; t = 0.01) with [pi] = 0.02, the threshold value becomes $I 89,189. Thus, even though only increments above the new land value are taxed, land value still falls by 5.4% given these parameter choices. If [pi] = 0.04, the threshold value falls further to $183,673, an 8.2% fall. Thus, as inflation rises, the post-tax market value of land falls. The reason for this result is that the incremental tax considered here is one that taxes nominal capital gains, rather than real capital gains. If only real capital gains were taxed (i.e. if the threshold was set so that X = [V.sup.0.sub.0] and indexed at rate (1 + [pi])), no tax would be collected on the inflation component.

4. General equilibrium effects of a property tax

4.1. Outline

The above equations describe how the ratio of rents to property prices depends on the land tax rate and other factors. However, they do not indicate how taxes affect the level of either rents or property prices, except under special assumptions. To do so, we adopt a general equilibrium model in which the level of rents and land prices are determined endogenously as a function of other factors such as incomes and interest rates. The model is an overlapping generations housing model, as described in Coleman (2009), based on that of Ortalo-Magne and Rady (2006). The basic structure of the model is a set of equations describing the demand for rental and owner-occupied housing, the supply of rental housing, and the total supply of housing.

The demand for rental and owner-occupied housing is based on an intertemporal utility maximisation model of consumer demand applied to a large number of agents who differ by age, income and wealth. Each agent passes through four stages (two young stages, middle-age, and retirement). The agents have different exogenously determined life-cycle patterns of labour income, they pay tax, save for retirement, and choose different types of houses and housing tenure at each stage of their lives. Agents choose their preferred housing options given their age, wealth and after-tax incomes, the cost of renting or buying different houses, their ability to raise a mortgage, and any property taxes. They are forward looking, so their choices take into account not only their current income and current housing prices, but their remaining length of life, future house prices, their future income stream, and their desired future housing patterns.

Agents have budget constraints that ensure lifetime expenditure on housing and goods is equal to their after-tax lifetime income, adjusted for inheritance. The agents can borrow or lend at exogenously determined interest rates, although young agents face bank imposed credit constraints limiting the amount they can borrow. The credit limits reflect both a minimum deposit requirement on house purchase and a maximum debt servicing constraint in relation to current income. All mortgages are table mortgages; thus, with a rising income profile over a person's lifetime, the credit constraint is most likely to bind in the earlier years of adulthood, and induce young agents to rent rather than purchase.

The agents choose between four housing options: they can (a) share a rental apartment with another agent; (b) rent an apartment without sharing; (c) live in an owner occupied apartment; or (d) live in an owner-occupied house. Utility increases as the agent moves up the quality ladder from (a) to (d). The credit constraints mean that agents typically ascend a property ladder as they age, although they may trade down in retirement. When agents die, they leave any house they have to the younger generation, but otherwise are assumed to run their financial wealth to zero by the time of their deaths. Agents may receive a bequest in addition to leaving one. We assume that in middle age each agent receives a bequest from someone at the corresponding point in the income scale.

In the baseline model when the property tax rate is zero, agents face a GST rate of 12.5% and a two-step income tax regime with a lower marginal tax rate of 20% and an upper marginal tax rate of 33% on current incomes. As the property tax rate is varied in the simulations, these rates are adjusted so that the fiscal position remains balanced. We assume a real interest rate of 5% and focus on the zero inflation case. We do not allow for any production changes as a result of taxation changes. Our simulations will therefore understate net benefits of such a tax shift if productivity improvements would eventuate.

Rental accommodation is supplied by (wealthier) agents who become landlords. They bid for houses and set rents at levels that leave them indifferent between the after-tax returns from lending money and the after-tax returns from investing in residential property. These returns include rents adjusted for costs like rates and property tax, upon which income tax is paid; and capital appreciation, which is exempt from income tax. (10) Since the marginal landlord is assumed to be a middle aged agent who is on the top marginal income tax rate, equality between the after-tax returns from residential property and lending means:

([Y.sub.t] - [k.sup.*][V.sub.t] - t[V.sub.t])(1 - [t.sup.u]) + [V.sub.t](1 + [[pi].sup.h.sub.t]) = [V.sub.t] (1 + ([r.sub.t] + [[pi].sub.t])(1 - [t.sup.u]))

where [k.sup.*] are the costs associated with leasing an apartment, [[pi].sup.h] is the rate of property price appreciation, and [t.sup.u] is the top marginal tax rate. Rearranging,

[V.sub.t] = [Y.sub.t](1 - [t.sup.u])/([r.sub.t] + [[pi].sub.t] + [k.sup.*] + t)(1 - [t.sup.u]) - [[pi].sup.h.sub.t] (11)

Equation (11) has a form similar to equation (3), although it is more complicated to take into account the different income tax treatment of interest income and capital gains. This equation determines the relationship between rents, property prices and the rate of property price appreciation, and can be used to derive analytic results about the way that taxes affects the ratio of rents to property prices. The rest of the model determines the level of rents, property prices, and property price appreciation rates, allowing the distributional and welfare effects of different taxes to be calculated.

We simulate the model with two contrasting housing supply elasticity assumptions. First, we assume that the housing supply is perfectly inelastic. This assumption may approximate the case of a land tax that is levied on all unimproved land in the economy. In that case, while the supply of land to housing may not be perfectly inelastic, all land will be affected by the land tax and so there is no wedge created between housing land and agricultural land, and hence no reason to expect land to be reallocated from one use to another. Second, we assume that housing supply has an elasticity of unity at equilibrium. (11) This may be appropriate for cases either where a tax on improvements is being considered, since new improvements will be forthcoming as long as the value of improvements exceeds the cost of building them (Grimes & Aitken, 2010), or where a land or property tax is imposed on housing, with agricultural land taxed at a lower rate and where there is some ability to rezone land between alternative uses. The model does not include land separately from improvements, which is the reason that we describe the new tax as a property tax rather than as a land tax. The inelastic supply case is closer to a land tax in which all land in the economy is subject to the same tax rate. We focus on these results here, but also discuss how the elastic supply results differ.

Under all sets of assumptions, prices, rents, and price appreciation rates adjust to ensure there is a steady state equilibrium in the housing market across all types of houses and tenure. For a given set of parameters, the equilibrium prices are found numerically and are used to calculate the welfare and distributional effects of different tax systems. First, we run the model with a baseline set of parameters in which GST and income tax rates are set as outlined above, with no property tax. We then impose a property tax of 0.5% p.a. on all owner-occupied and investor-owned houses. (12) In order to ensure fiscal neutrality, we offset the resulting income to government by: (a) reducing GST, or (b) reducing income tax (with all agents' income tax payments reduced proportionately). We have also investigated a property tax with an exemption for owner-occupiers. This resulted in the virtually complete collapse of the rental market, as the long-term costs of renting exceeded the long term costs of home ownership. We do not consider this result to be just an artefact of the model; in reality, a distortion of this nature could severely impact on renters and could have major welfare consequences.

4.2. Results

Initially, we examine the general equilibrium effects of a shift in taxes under the assumption that housing (and apartment) supply is perfectly inelastic. We simulate two tax changes relative to the baseline tax structure. Letting ([t.sup.l], [t.sup.u], [t.sup.g], [t.sup.p]) be the vector of tax rates pertaining to the lower marginal income tax rate, the upper marginal income tax rate, the GST rate and the property tax rate respectively, our baseline model adopts the tax vector (20%, 33%, 12.2%, 0%). (13) A 0.5% property tax is then introduced, with either the GST rate or the income tax rate reduced.

The first column of Table 1 shows the outcomes for key variables. The annual rent of $12,550 represents a rental yield of just over 5% on the apartment price, which equals the return on financial assets (5%) plus a small allowance for rates and other housing costs. In the model, 18% of agents rent, although 88% of houses are owner-occupied since some agents share apartments. Home-owning households initially borrow to purchase their property, resulting in a household gross debt ratio of 69% of GDP. Many middle-aged agents hold financial assets, so net household financial assets are 28% of GDP.

Column 2 shows the steady-state outcomes for the case where all property is taxed at a rate of 0.5% p.a. and GST is reduced to 8.8% to achieve fiscal balance. Equilibrium rents are almost unchanged: in contrast, apartment and house prices both fall by approximately 10% as the present discounted value of the tax is impounded in the property price. Rents do not change because they are subject to two offsetting forces: first, the annual tax is passed on to renters, raising rents: but secondly, property prices fall, as competition amongst landlords to fill vacant houses reduces rents and house prices by an amount that almost exactly offsets the first effect. The decline in house prices means gross household debt declines and net household financial assets increase, as less money needs to be borrowed to purchase a property. The rise in net financial assets allows an increase in consumption. While the inelastic supply means that the aggregate number of dwellings does not change, the homeownership rate increases slightly as a few low income middle aged agents decide to buy rather than rent.

Column 3 shows the steady state outcomes when income tax rates are reduced when the property tax is introduced, and GST is left approximately unchanged. Prices of properties again impound the bulk of the present discounted value of the property tax, although some substitution between houses, apartments and goods consumption leaves house prices and to a lesser extent, apartment prices a little above their level in the GST case. (14) The homeownership rate again rises slightly relative to the baseline case, debt levels decline and net financial assets of households increase.

These results are broadly as expected from our partial equilibrium analysis given the assumption of completely inelastic supply. The debt and net financial asset results are particularly interesting from a macroeconomic perspective. The household balance sheet, inter alia, comprises property as an asset and mortgage debt as a liability. At the macroeconomic level, debt, at the margin, is financed from offshore. If a tax is introduced that lowers the value of property assets, the steady state effect is a corresponding reduction in households' gross debt and thus an increase in their net financial assets; at the national level, New Zealand's steady state gross and net offshore debt falls and there is a rise in the country's net international investment position (NIIP). (15) Debt servicing costs are therefore reduced, resulting in a sustained rise in the current account balance. Put simply, high domestic property prices raise the portion of the country's production that is paid annually to foreigners.

While macroeconomic benefits might accrue from imposition of a land/property tax, it is important also to analyse welfare changes at the level of the individual family. Since the model is one in which agents maximise lifetime utility subject to constraints, we can use the utility outcome for each agent to measure whether utility rises or falls in each case. Furthermore, we can compare the degree of utility changes across the income spectrum to see how welfare changes according to lifetime earnings. Figure 1 divides agents into deciles according to lifetime incomes. It charts the average steady state change in utility for households in that decile, firstly for the GST-financed land tax (labelled G) and secondly for the income tax-financed land tax (labelled I). The actual levels of the utility changes need not concern us, but the overall utility changes are important, as are the differing patterns when GST rather than income tax is changed.

Every decile experiences a substantial improvement in welfare under both financing options. For most deciles, there is little to choose between the income tax or GST options, although low income agents are slightly better off if the revenue raised from a property tax is refunded through a cut in the GST rate rather than a cut in income taxes, while high income agents are better off with a cut in income tax rates. The first decile records the greatest increase in welfare, although not too much emphasis should be placed on this result as much of the difference between the welfare improvement for the first and all other deciles reflects a change in inheritance patterns. Welfare improves primarily because house prices fall and agents spend less of their income on taxes and housing, and thus can afford more consumption. While rents and the user cost of housing (interest rate on the purchase price, plus land tax) change by a little when a property tax is introduced, part of this expenditure is collected by government, allowing alternative taxes to be reduced. In essence, some of the expenditure on housing is diverted from overseas lenders to government, resulting in a reduction of the country's gross debt servicing bill.

How is it that all deciles can benefit from the change when, by assumption, aggregate GDP is unchanged? One reason is that the tax base has been widened to include imputed rentals, thus enabling a broader base, lower rate tax regime to emerge with a reduced overall excess burden caused by taxation. A separate cause is that initial holders of property suffer a reduction in their property wealth as the initial property price falls.16 This once-only welfare cost on a particular generation is reflected in a permanent welfare gain for all future generations who pay lower servicing costs to foreigners.

[FIGURE 1 OMITTED]

When we repeat the same three tax options for the elastic supply case, several key results change. Apartment and house prices change by much less when the property tax is introduced compared with the inelastic case. Accordingly, rents rise to maintain required rental yields for landlords. This results in annual rents increasing to incorporate virtually the entire annual property tax payment. The increased tax on housing results in a decline in the number of properties, with more people sharing a flat than under the baseline case. Furthermore, there is a change in the mix of dwellings, with a substantial decline in the number of houses and an increase in the number of apartments. The homeownership rate falls slightly as the total user cost of property ownership increases once a property tax is introduced. The household debt ratio again falls, driven principally by a change in the mix of dwellings and also by a small decline in the total number of properties in the economy. Accordingly, net financial assets increase. These effects are in the same direction, but are not as strong, as the inelastic supply case. Overall, welfare gains are still positive but the gains are less than in the inelastic case. Seven of the ten decries show a net welfare gain under each tax financing option.

The elastic supply case indicates the direction of changes that might occur relative to the inelastic case as the elasticity of residential land supply increases. This may be relevant if a fully comprehensive tax were not adopted and/or if a property tax rather than a land tax were adopted. The elasticity in the property tax case will also vary according to the time period considered; the supply may be highly inelastic in the short-medium term but elastic in the long term.

5. Fiscal impacts of a land/property tax

We use 2006 QVNZ rateable values to estimate the base upon which a land or property tax could be levied. These valuations form the basis for local authority rates and thus provide a statutory basis for a central government land/property tax. Valuations are updated every three years for most local authorities. Using 2006 data, we have valuations for 2004, 2005 and 2006; we set these onto a 2006 basis by updating the 2004 and 2005 valuations using movements in QVNZ's national house price index through to 2006.

Table 2 presents valuation data across several categories. First, we present data for the total of all properties. Included in this total are public buildings, public land and conservation forestry. It is unlikely that such properties would be subject to a land/property tax; thus, we include a second total ('Total--ECFO') that excludes the conservation forestry estate and 'other' properties (mainly public buildings and public land). This total is decomposed into four groups: Residential; Commercial Forestry; Agriculture; and Industrial/Commercial/Mining. We follow QVNZ's categories in this decomposition except that we allocate "lifestyle' properties (both vacant and improved) to residential rather than to agriculture.

For each of the major categories listed above, we present data for total land value, improvements, capital value, the number of assessments (i.e. number of properties), average land value and capital value per assessment, and the ratio of land value to capital value. The latter calculation is useful in judging how different sectors would be affected by a land tax versus a property tax. We provide additional information according to ownership definitions. We use central government accounts to obtain figures for central-government-owned land and property, and Statistics New Zealand data for local-government-owned property. The residual is attributed to private ownership. (17)

Central and local government property may be exempted from any land/property tax, which is why we list their values here. We cannot deduct their totals directly from 'Total-ECFO' since the latter has already deducted the value of public buildings/land and conservation forestry from 'Total--All properties'. The deductions in that case totalled $24.9 billion (land value) and $84.1 billion (capital value); these deductions compare with estimated total central and local government holdings of $20 billion (land value) and $51.8 billion (capital value). Estimated central and local government holdings are therefore less than the deductions already applied to the Total category. We surmise that the bulk of these deductions pertain to government holdings and so do not make further deductions for government holdings from the Total-ECFO category.

Several key results are apparent from Table 2 (all ratios are specified relative to Total-ECFO unless otherwise noted). First, Residential comprises 65% of all land values and 69% of all capital values. Second, Agriculture and Commercial Forestry together comprise just 24% of all land value, and an even smaller percentage of capital values. Third, for both Residential and Industrial/Commercial/Mining, land values comprise around half of capital value; by contrast, the ratios for Agriculture and Commercial Forestry are around four-fifths. Thus, a single rate proportional land value tax (with no exemptions) would fall more heavily on existing property owners within land-based industries. Fourth, the average Agriculture land value (per assessment) is over five times that for Residential, while the capital value ratio is 3 1/2 times as great. Thus, if each property was owned by a single occupying household, a proportionate tax would hit agricultural-based households considerably harder than it would hit residential households. These results suggest that consideration may be given (under either a land value or a capital value tax) to a differential rate applied to land classified as (and used as) Agriculture or Commercial Forestry. (18) Alternatively, and more simply, consideration may be given to a fiat rate 'deductible' per hectare.

Table 3 presents potential revenue figures from a hypothetical 1% p.a. land tax based on the values in Table 2. The third column presents estimates of the initial year land tax revenues both for Total-ECFO and for each sector. The fourth column presents estimates of the initial year revenues from a property tax (i.e. on capital values) that raises the same aggregate revenue as a 1% land tax (ceteris paribus). The resulting capital value tax is set at 0.549%. Initial year tax revenue would be $4.6 billion provided the value of the tax base remained at its 2006 value of $461 billion. (19)

The two revenue columns demonstrate that a land value tax would raise more from the commercial forestry and agriculture sectors than would a capital value tax, while a capital value tax would raise more from the industrial/commercial/mining sectors. The residential sector, in aggregate, would pay similar amounts of tax in either case. The complete exclusion of agriculture and forestry from either tax base would lose between 17% and 24% of total revenue. We note that any exemption creates incentives to reclassify lands into exempted sectors; an alternative approach is to consider a per hectare deductible that could effectively exempt lower value land.

6. Distributional impacts

We examine distributional impacts initially at the Statistics NZ area unit (AU) level (i.e. 'suburb' level). For each AU, we obtain QVNZ 2006 valuation data (20) for median residential dwelling (RD) land value (MLV), median RD improvements value (MIV), median RD capital value--i.e. land plus improvements (MCV), and the ratio of median RD land value to median RD capital value (MRAT). We also obtain Statistics NZ 2006 census AU data for median household income (HHY).

The top half of Table 4 presents the results of OLS regressions of each of lnMLV, lnMIV, lnMCV and MRAT against lnHHY plus a constant. In each of the first three regressions, if the coefficient on lnHHY is significantly greater than unity we can conclude that a land/property tax is, on average, progressive (i.e. areas with higher incomes pay proportionately higher land/property taxes than do lower income areas). (21)

The fourth regression tests whether a land tax is more progressive (across householders) than is a property tax (McCluskey et al., 2006): this would be the case if the coefficient on lnHHY is significantly positive. (Note that these regressions are used solely to identify systematic associations between the variables: they cannot be used to attribute causality.)

Each of the value variables is positively related to median household income. Furthermore, the coefficient on lnHHY in the land value and capital value equations is significantly greater than unity (using a 95% confidence interval) indicating that each of a land tax and a property tax is progressive. The coefficient on lnHHY in the MRAT equation is positive, implying that higher income households tend to live in areas where houses have relatively high ratios of land to capital value: consequently, a land tax is more progressive (for households) than is a property tax.

These results reflect nationwide relationships; however, they may be driven by differences that exist chiefly across territorial local authorities (TLAs), e.g. by city versus rural differences, rather than within-TLA differences. We can extract the two influences by estimating the equations with the inclusion of a fixed effect for each of the 73 TLAs. The reported slope coefficients then measure the relationship between the relevant housing variable and household income solely within TLAs. This set of estimates is provided in the lower portion of the table. Coefficients on the income variable in the first three equations remain significantly positive, but are now significantly less than unity. In addition, lnHHY is no longer significant in the MRAT equation. Thus, the observed relationship between the ratio of land to capital values and household incomes at a national level is explained by a cross-TLA, rather than a within-TLA, relationship. Furthermore, the progressivity of a land/ property tax at a national level results from higher income households on average residing in higher value TLAs (e.g. cities): within a given TLA, land/property values still rise with income but not proportionately.

We examine these issues further at a household level using data from Statistics New Zealand's 2006 wave of the Survey of Family Income and Expenditure (SoFIE) that includes a wealth survey component. SoFIE is an official longitudinal survey that is designed to be representative of the New Zealand population. The SoFIE wealth survey includes data for the capital value of the household's owner-occupied house in cases where someone in the household owns the house. Also included is the date of valuation: we update all capital values to 2006. Land values are not available so we analyse the relationship between property values and other variables.

We analyse the relationship between household income and the household's tenancy status (renter (22) or owner-occupier) and, for the latter, the household's owner-occupied property value. (23) Table 5 presents a matrix of household incomes by house value (the first category relates solely to renters); household incomes are presented by quintile. Each cell in the table (other than the final row) represents the proportion of households in that income quintile that owns a house within the relevant capital value category (or, if in the first column, rents a house). The final row presents the proportion of the population that is within that capital value category. Thus, 43.2% of households do not live in an owner-occupied home. Of the 56.8% of households that are owner-occupiers, most have properties with a capital value between $150,000 and $500,000, i.e. categories (iii)-(v).

For the top income quintile, the proportion of homeowners in each category rises as house values rise, whereas for the lowest income quintile, the proportion of homes that are owned is most heavily weighted to houses below $250,000. These observations, consistent with the remaining data in the table and with the AU results, indicate a positive relationship between property values and household incomes. Thus, the initial wealth effect and subsequent direct cash-flow impact of a property tax will tend to be greater for higher income households than for lower income households. (24)

Another way of assessing the distributional effects of a property tax is to examine the relationship between ownership of property and net worth. Over 96% of households in the lowest net worth quintile do not own a home. For the top net worth quintile, the proportion of homeowners increases in line with the capital value category. Over 38% of that quintile own a property worth more than $500,000. By contrast, only 6% of the second highest net worth quintile owns a property of at least that value. These figures indicate that a property tax would tend to be progressive, not just according to income, but also according to wealth.

Across the sample, 14.2% of households own one or more 'investment' properties (i.e. a house that is not their primary residence). Ownership of a second house is most concentrated in those who reside in a category (vi) owner-occupied house (i.e. one worth more than $500,000); 31.3% of this group own additional property. This reduces to 19.9% and 16.5%, 10.5% and 9.2% for categories (v), (iv) (iii) and (ii) residents respectively. For owner-occupiers, the impact of a property tax therefore tends to rise according to the value of their primary residence, since owners-occupiers who reside in more expensive homes also, on average, have higher exposures to other property assets.

We have also examined how a property tax might impact on groups of people with differing characteristics. 'Retired' households are more likely than other households to own property and are therefore more prone to suffering an initial capital value loss than non-retired households upon introduction of a property tax. Furthermore, retired households would gain less from any offsetting income tax reduction than employed households. The impact of a property tax across age-groups is, however, complicated. Most retired people fully own their property (with no debt) whereas younger owners are likely to have a mortgage. If a tax were introduced that reduced all property values by 10%, those who own a house without debt would lose 10% of their housing equity, whereas those who initially had a 90% loan to value ratio would lose their entire housing equity. In this sense, young homeowners would face greater losses, on average, than older homeowners. (25) Conversely, they would have more to gain from an offsetting reduction in income taxes.

Households with three or more children are under-represented in the higher capital value categories, and are over-represented in the rental category. Similarly, Maori and Pacific populations are under-represented in the high value ownership categories and are strongly over-represented in the rental category. These groups will therefore suffer significantly less initial capital value loss than the rest of the population upon introduction of a land/property tax. Furthermore, if their rental houses are generally in low capital value areas, they will face smaller rental rises than renters in higher value suburbs.

7. Conclusions

A land tax has favourable efficiency properties relative to other taxes. To a first-order approximation, the economy's supply of land is fixed and a tax does not distort the allocation of this resource. A tax on land could have non-trivial effects on fiscal revenues. Using 2006 figures, a 1% p.a. tax on all non-government land could raise approximately $4.6 billion annually. To place this figure into perspective, $4.6 billion represents 20% of all income tax revenue forecast for 2009/10. The top personal tax rate of 38% applies above an income threshold of $70,000 p.a. Total income tax revenue raised on earnings above this figure is forecast to be $9.8 billion for 2009/10. If the top personal tax rate were reduced to 33%, the direct loss in income tax would be $491 million, which represents just 11% of the revenue from a 1% p.a. land tax. The impact of a land tax would fall more heavily on some segments of society than on others. One currently untaxed segment that it would fall on is foreign-domiciled owners of New Zealand property. The tax paid by non-New Zealanders contributes a net benefit to the country.

The overall effects of a switch to land/property taxes would depend on what other tax changes are made at the same time and on the structure of the economy. Our partial and general equilibrium analyses demonstrate that certain key parameters and the exact nature of the tax will lead to different housing market outcomes. One consistent result across our general equilibrium simulations is that aggregate indebtedness of the economy declines with the introduction of a land/ property tax. Owners of existing property would incur a wealth loss following introduction of a land/property tax unless there were perfectly elastic supply. Even in this case, owners would face the present discounted value of the future land/property tax flow. With a flat tax, the wealth loss would be proportionate to the existing value of land/property. Owners of land-extensive residential properties, farms and forests would be liable for the largest losses in proportion to their property holdings. A land tax appears more progressive than a property tax; however, overall progressivity depends on the nature of offsetting taxation changes.

Variants of a land tax could be envisaged if different outcomes were sought. For instance, a land tax could be levied only on the increment of land value above some base level or there could be a gradual increase in the tax rate over time. A per hectare rebate could reduce the impact on land-extensive activities, such as farming and forestry, since land values per hectare are relatively low for these land uses. This approach would also assist the progressivity of a land tax and would reduce the impost of the tax on Maori land.

We have not modelled the productivity impacts of a switch away from income taxes to a land (or property) tax. Results from the wider tax literature (26) indicate that the reduction in distortions to labour supply and investment resulting from lower personal and company tax rates could lead to improvements in labour productivity and hence per capita living standards. Fiscal revenues can therefore be enhanced at the same time as reducing excess burdens caused by distortionary taxation. While these aspects favour a land tax, distributive aspects of such a tax will remain a central issue in considering its merits, and we make no claims as to which distributive concerns should prevail.

DOI: 10.1080/00779954.2010.492576

Acknowledgements

The authors thank the New Zealand Treasury for partial funding of the study, together with support from the Marsden Fund of the Royal Society of New Zealand (grant 07-MEP-003, Home Ownership and Neighbourhood Wellbeing) and the Foundation for Research, Science and Technology (programme MOTU0601 Infrastructure) for funding aspects relevant to each programme. Chris Young (Motu) and Trinh Le (NZIER) provided expert assistance with data extraction. The authors also thank the editor, officials at the Treasury and IRD and participants at the 2009 New Zealand Association of Economists conference for helpful comments on an earlier draft; we also thank both Quotable Value New Zealand and Statistics New Zealand for access to data. Access to the SoFIE data used in this study was provided by Statistics New Zealand in a secure environment designed to give effect to the confidentiality provisions of the Statistics Act, 1975. The results in this study and any errors contained therein are those of the authors, not Statistics New Zealand or of any other organisation.

References

Blaug, M. (1980). Review of Robert V. Andelson: critics of Henry George: a centenary appraisal of their strictures on progress and poverty. Economica, 47(188), 471-472.

Bleaney, M., Gemmell, N., & Kneller, R. (2001). testing the endogenous growth model: public expenditure, taxation, and growth over the long run. Canadian Journal of Economics, 34(1), 36-57.

Coleman, A. (2009). The long term effects of capital gains taxes in New Zealand. Motu Working Paper 09-13. Wellington: Motu.

Coleman, A., & Grimes, A. (2009). Fiseal, distributional and effeciency impacts of land and property taxes. Motu Working Paper 09-14. Wellington: Motu.

Dowse, G., & Hargreaves, R. (1999). Rating systems in New Zealand. In W.J. McCluskey (Ed.), Property tax: an international comparative review. UK: Ashgate.

Dye, R.F., & England, R.W. (2009). The principles and promises of land value taxation. In R.F. Dye & R.W. England (Eds.), Land value taxation: theory, evidence and practice (pp. 3-10). Cambridge, MA: Lincoln Institute of Land Policy.

Franzsen, R. (2009). International experience. In R.F. Dye & R.W. England (Eds.), Land value taxation: theory, evidence and practice (pp. 28-47). Cambridge, MA: Lincoln Institute of Land Policy.

George, H. (1880). Progress and poverty. New York: D. Appleton and Co.

Goldsmith, P. (2008). We won, you lost. Eat that! Auckland: David Ling Publishing.

Grimes, A. (2003). Economic growth and the size and structure of government: implications for New Zealand. New Zealand Economic Papers, 37(1), 151-174.

Grimes, A., & Aitken, A. (2010). Housing supply, land costs and price adjustment. Real Estate Economies, forthcoming.

Hargreaves, R. (1991). Is site value still an appropriate basis for property taxation? In International Conference on Property Taxation and its Interaction with Land Policy. Cambridge, MA: Lincoln Institute of Land Policy.

Haughwout, A. (2002). Public infrastructure investments, productivity and welfare in fixed geographic areas. Journal of Public Economics, 83(3), 405-428.

Hong Kong Democratic Foundation (1996). 'Land tax' and high land prices in Hong Kong. HKDF Policy Paper, www.hkdf.org/pr.asp?func=show&pr=24

Hong Kong Government (2006). Hong Kong in brief 2006. Updated with information from: www.govt.hk/en/residents/taxes/ taxfiling/taxrates/propertyrates.htm

IMF (2009). World economic outlook: crisis and recovery. Washington, DC: International Monetary Fund.

Kneller, R., Bleaney, M., & Gemmell, N. (1999). Fiscal policy and growth: evidence from OECD countries. Journal of Public Economics, 74(2), 171-190.

Local Government Rates Inquiry Panel (2007). Funding local government. Wellington: New Zealand Government.

McCluskey, W., Grimes, A., Aitken, A., Kerr, S., & Timmins, J. (2006). Rating systems in New Zealand: an empirical investigation into local choice. Journal of Real Estate Literature, 14(3), 381-397.

McLeod, R., Patterson, D., & Jones, S. (2001). Tax review 2001 final report. Wellington: New Zealand Government.

Mill, J.S. (1865). Principles of political economy with some of their applications to social philosophy. London: Longmans, Green, & Co.

New Zealand Government (2009). 2009 budget. Wellington: New Zealand Government.

New Zealand Government Railways Department (1927). New principle in state railway construction. Building track out of betterment values. The New Zealand Railways Magazine, 2(3), 6-8.

Oates, W.E., & Schwab, R.M. (2009). The simple analytics of land value taxation. In R.F. Dye & R.W. England (Eds.), Land value taxation: theory, evidence and practice (pp. 51-72). Cambridge, MA: Lincoln Institute of Land Policy.

Ortalo-Magne, F., & Rady, S. (2006). Housing market dynamics: on the contribution of income shocks and credit constraints. Review of Economic Studies, 73(2), 459-485.

Plummet, E. (2009). Fairness and distributional issues. In R.F. Dye & R.W. England (Eds.), Land value taxation: theory, evidence and practice (pp. 73-98). Cambridge, MA: Lincoln Institute of Land Policy.

Ramsey, F.P. (1927). A contribution to the theory of taxation. The Economic Journal, 37, 47-61.

Ricardo, D. (1817). On the principles of political economy and taxation. London: John Murray.

Roback, J. (1982). Wages, rents and the quality of life. Journal of Political Economy, 90, 1257-1278. SGS Economics and Planning (2007). Value capture mechanisms: international models and their relevance to New Zealand. Wellington: Ministry for the Environment.

Woods, N.S. (1935). The birth of our railways--the great public works policy of 1870. Part II. The New Zealand Railways Magazine, 10(2), 43-44.

Andrew Coleman (a) and Arthur Grimes (a,b) *

(a) Motu Economic and Public Policy Research, Wellington, New Zealand; (b) University of Waikato, New Zealand

* Corresponding author. Email: Arthur.grimes@motu.org.nz

Notes

(1.) Hong Kong raises over 35% of government revenue from its property base (Hong Kong Democratic Foundation, 1996; Hong Kong Government, 2006).

(2.) 'Land value' is best thought of as unimproved land value. In practice, land incorporates some improvements leading to taxation of 'land improvements'; see Franzsen (2009) for further discussion.

(3.) Exceptions may occur in cases where the imposition of a tax corrects for a non-existent (or insufficient) market price for a good that has real resource costs; e.g. a carbon tax.

(4.) With international migration, even a poll tax cannot be regarded as a lump-sum tax.

(5.) Accordingly, Milton Friedman considered that 'the least bad tax is the property tax on the unimproved value of land' (Blaug, 1980).

(6.) One consequence is that urban development is likely to be relatively more land-extensive (i.e. sprawling) under a property tax system than under a land tax system (Oates and Schwab, 2009).

(7.) See: Local Government Rates Inquiry Panel (2007, p. 13) and McLeod et al. (2001, p. 28).

(8.) For related studies on land/property taxes in New Zealand, see: New Zealand Government Railways Department (1927), Woods (1935), Hargreaves (1991), Dowse and Hargreaves (1999), Grimes (2003), McCluskey et al. (2006), SGS Economics and Planning (2007), Goldsmith (2008), Franzsen (2009).

(9.) The analysis here draws on, and extends, that in Oates and Schwab (2009).

(10.) Coleman (2009) examines the effect of applying different types of capital gains taxes in this model.

(11.) This parameter choice reflects the long-run elasticity of dwellings with respect to population.

(12.) In Section 5, we show that a 0.5% property tax is approximately fiscally equivalent to a 1% land tax.

(13.) GST is set as a residual to ensure that the fiscal position is balanced. This leads to slight variations in the model's GST rate from the statutory rate in some simulations: but the differences are immaterial.

(14.) Note that the reduction in GST leads to a less favourable valuation of housing relative to goods consumption than in the income tax reduction case.

(15.) To the extent that foreigners own some of the property directly, the reduction in the value of their property equity in New Zealand will also lead to an initial rise in the NIIP.

(16.) Given that we are using a steady-state model, this cost is not incorporated into our welfare figures.

(17.) CG use Te Puni Kokiri data to show that the value of land under the aegis of the Maori trustee was $0.7 billion, or 0.15% of total private sector land value. This land could therefore be omitted from the coverage of the tax with little fiscal impact.

(18.) Another reason for doing so is that some of the value of agricultural land is due to 'land improvements' (e.g. drainage); a lower rate could therefore be justified to reduce the taxation of such improvements.

(19.) Since 2006, rises in the tax base have exceeded falls, leading to a higher tax base than $461 billion: however, from Sections 3 and 4, the tax base would be expected to fall consequent on the introduction of a land tax. Thus, actual revenue (with a 1% land tax) could be greater or less than the $4.6 billion estimate.

(20.) Valuation data relating to 2004 and 2005 are updated to 2006 using the national house price index.

(21.) Conclusions from these results about progressivity at the individual or household level would require the assumption of homogeneity of outcomes within each AU.

(22.) All non-owner-occupiers are termed 'renters', including those 'renting' from a family trust.

(23.) Additional (investment or holiday) properties are considered separately.

(24.) CG demonstrate that the observed positive relationship between capital value of owned homes and household incomes is mirrored in the rental market.

(25.) This raises an issue that the introduction of a land/property tax would reduce the collateral that borrowers have in a house, potentially creating riskier exposures for lenders on existing loans.

(26.) For instance, see Kneller et al. (1999) and Bleaney et al. (2001).
Table 1. General equilibrium property tax effects: inelastic supply.

                                        Taxation Regime

                                                         0.5% property
                            Baseline     0.5% property     tax with
                            with no        tax with         reduced
Variable                  property tax    reduced GST     income tax

Lower marginal income        20.0%           20.0%           18.2%
  tax rate
Upper marginal income        33.0%           33.0%           30.1%
  tax rate
GST rate                     12.2%           8.8%            12.5%
Property tax rate             0.0%           0.5%            0.5%
Apartment price             $238,600       $215,100        $215,700
House price                 $394,900       $355,600        $363,400
Rent (annual)               $12,550         $12,350         $12,350
Number of                    93.2%           93.2%           93.2%
  dwellings/family
Fraction small                54%             54%             54%
  properties
Fraction young agents         42%             42%             42%
  renting
Fraction agents renting      18.0%           16.6%           16.3%
Fraction houses owned        88.0%           89.5%           89.7%
  by occupiers
Gross debt/GDP               69.0%           56.3%           53.3%
Net financial                28.3%           41.8%           45.3%
  assets/GDP

Table 2. Estimated land and property values (New Zealand--2006).

                                                              Capital
                               Land value    Improvements      value
Variable                       ($ billion)   ($ billion)    ($ billion)

Total--All properties             486.0         438.0          924.0
Total--ECFO *                     461.1         378.9          839.9
Residential                       298.0         280.0          578.0
Commercial forestry                 4.3           0.8            5.1
Agriculture                       105.0          29.0          134.0
Industrial/commercial/mining       53.5          69.5          123.0
Memo Items
Central government-owned           13.8          21.9           35.7
Local government-owned              6.2           9.8           16.1
Privately-owned                   466.0         406.3          872.3

                               Assessments      LV/CV
Variable                         ('000)        (Ratio)

Total--All properties            1826.8         0.526
Total--ECFO *                    1755.7         0.549
Residential                      1541.5         0.516
Commercial forestry                 5.0         0.841
Agriculture                       102.0         0.784
Industrial/commercial/mining      107.2         0.435
Memo Items
Central government-owned           n.a          0.387
Local government-owned             n.a          0.387
Privately-owned                    n.a          0.534

                                 LV per         CV per
                               assessment     assessment
Variable                       ($ million)   ($ million)

Total--All properties             0.266         0.506
Total--ECFO *                     0.263         0.478
Residential                       0.193         0.375
Commercial forestry               0.860         1.022
Agriculture                       1.030         1.314
Industrial/commercial/mining      0.499         1.148
Memo Items
Central government-owned
Local government-owned
Privately-owned

* ECFO = Excluding conservation forestry & 'other'

Sources:

1. Data in top portion of table are QVNZ-sourced rateable values;
2004 and 2005 values are rated upwards to 2006 by the national
house price index.

a. Lifestyle properties are attributed to residential rather
than agriculture. (Total capital value of lifestyle properties is
$75.2 billion.)

b. 'Other' in ECFO includes public buildings, land etc; so overlaps
with central & local government-owned memo items.

c. Conservation forestry in ECFO has capital value of $2.1 billion.

2. Central government-owned data are sourced from 'Financial
Statements of the Government of New Zealand (2007).

3. Local government-owned capital values are sourced from
'Individual local authority statistics, balance sheet items and
capital transactions, year ended June' (Statistics NZ); land values
are calculated using the same ratio as for central government.

Table 3. Initial year revenue figures for land tax & property
tax--2006 values. *

                                  Tax base           Tax revenue
                                 ($ billion)       ($ billion p.a.)

                                                                CV
                                                    LV       flat tax
                               Land    Capital   flat tax    @ 0.549%
Source                         value    value    @ 1% p.a.     p.a.

Residential                    298.0    578.0      2.980       3.173
Commercial forestry              4.3      5.1      0.043       0.028
Agriculture                    105.0    134.0      1.050       0.736
Industrial/commercial/mining    53.5    123.0      0.535       0.675
Total--ECFO                    461.1    839.9      4.611       4.611

* No account is taken here of potential drops in the value of the
tax base caused by the imposition of a land  or property tax, or of
potential increases in values between 2006 and year of introduction.

Table 4. Relationship of household income to housing variables *.

                                Explanatory variable is 1nHHY

                                             AU
Spatial level:
Dependent variable **      lnMLV      lnMIV      lnMCV       MRAT

Coefficient                1.650      0.856      1.115       0.147
Standard error             0.076      0.026      0.037       0.012
p-value ***                0.0000     0.0000     0.0000      0.0000
N                          1733       1733       1733        1733
[R.sup.2]                  0.2072     0.3768     0.3400      0.0834

With TLA fixed effects
Coefficient                0.757      0.697      0.705      -0.002
Standard error             0.066      0.027      0.033       0.010
p-value                    0.0000     0.0000     0.0000      0.8164
N                          1733       1733       1733        1733
[R.sup.2]                  0.9965     0.9995     0.9993      0.9367

* Definitions:

AU = Area Unit.

TLA = Territorial Local Authority.

HHY = median household income of the area unit (Current 2006 $s).

MLV = median land value of residential dwellings in the area unit
(Current 2006 $s).

MIV = median improvement value of residential dwellings in the area
unit (Current 2006 $s).

MCV = median capital value of residential dwellings in the area
unit (Current 2006 $s).

MRAT = median land value capital value ratio for residential
dwellings in the area unit.

** For the top portion of the table, a constant is included in each
equation but not reported. TLA fixed effects are included in the
lower portion.

*** The p-value tests the null hypothesis that the reported
coefficient equals zero.

Table 5. Household incomes (HY) and household tenancy status/
owner-occupied capital values (CVs) *.

                             Capital value category

HY quintile     (i)     (ii)     (iii)    (iv)      (v)     (vi)

(i)            0.564    0.130    0.143    0.080    0.056    0.029
(ii)           0.486    0.119    0.152    0.109    0.083    0.050
(iii)          0.436    0.088    0.172    0.147    0.092    0.065
(iv)           0.344    0.065    0.159    0.176    0.165    0.091
(v)            0.331    0.028    0.075    0.140    0.210    0.215
Total          0.432    0.086    0.140    0.131    0.121    0.090

* Weight shown in each cell (other than the last row) is proportion
of that income quintile falling in the capital value category shown.

For example, 15.2% of households in quintile 2 (of household income)
own a house valued between $150,000 and $250,000.

Weight shown in final row is proportion of population within that
capital value category (thus 43.2% of the sample do not live in an
owner-occupied house).

HY quintiles given by:

(i) <$25,030: mean: $14,835;
(ii) ($25,030, $43,737); mean: $33,717;
(iii) ($43,737, $66,782); mean: $54,799;
(iv) ($66,782, $100.850); mean: $82,364:
(v) >$100,850; mean: $177,159.

Capital Value categories given by:

(i) $0 [i.e. renter];
(ii) ($0, $150,000];
(iii) ($150,000, $250,000];
(iv) ($250,000, $350,000];
(v) ($350,000, $500.000];
(vi) > $500,000.
COPYRIGHT 2010 New Zealand Association of Economists
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2010 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:SYMPOSIUM ON PROPOSED CHANGES TO THE NEW ZEALAND TAX SYSTEM
Author:Coleman, Andrew; Grimes, Arthur
Publication:New Zealand Economic Papers
Article Type:Report
Geographic Code:8NEWZ
Date:Aug 1, 2010
Words:11300
Previous Article:The long-term effects of capital gains taxes in New Zealand.
Next Article:Taxation of land: a comment on Coleman and Grimes.
Topics:

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters