Australia's Future Tax System

Final Report: Detailed Analysis

Chapter C: Land and resources taxes

C2. Land tax and conveyance stamp duty

C2–4 Directions for reform

Recommendation 51:

Ideally, there would be no role for any stamp duties, including conveyancing stamp duties, in a modern Australian tax system. Recognising the revenue needs of the States, the removal of stamp duty should be achieved through a switch to more efficient taxes, such as those levied on broad consumption or land bases. Increasing land tax at the same time as reducing stamp duty has the additional benefit of some offsetting impacts on asset prices.

Recommendation 52:

Given the efficiency benefits of a broad land tax, it should be levied on as broad a base as possible. In order to tax more valuable land at higher rates, consideration should be given to levying land tax using an increasing marginal rate schedule, with the lowest rate being zero, with thresholds determined by the per-square-metre value.

Recommendation 53:

In the long run, the land tax base should be broadened to eventually include all land. If this occurs, low-value land, such as most agricultural land, would not face a land tax liability where its value per-square-metre is below the lowest rate threshold.

Recommendation 54:

There are a number of incremental reforms that could potentially improve the operation of land tax, including:

  1. ensuring that land tax applies per land holding, not on an entity's total holding, in order to promote investment in land development;
  2. eliminating stamp duties on commercial and industrial properties in return for a broad land tax on those properties; and
  3. investigating various transitional arrangements necessary to achieve a broader land tax.

Stamp duty

Ideally, there is no place for stamp duty in a modern Australian tax system. Stamp duties generate large efficiency costs, as they discourage turnover in property and tax improvements as well as land. The tax also imposes a higher burden on people who need to move, which is not equitable. The only positive feature of stamp duty — its relative simplicity — has long since ceased to justify its continued use in the face of the costs it imposes on Australian society (see Recommendation 51).

While removing stamp duty would lead to more equitable and efficient outcomes, it would create a substantial hole in State revenues. As discussed in greater detail in Section G2 State tax reform, this shortfall should be met though increased reliance on more efficient State taxes. The Australian government should consider facilitating a transition away from stamp duties, reflecting the national benefit of reforms to State taxes and the quality of the Australian government tax bases. Another option is to reduce stamp duties incrementally, including capping the maximum rate, possibly as part of an intergovernmental agreement.

There is a case to link the reform of stamp duty to that of land tax to reduce the impact on prices and wealth caused by tax reform. Some of the reduction in stamp duty would lead to higher property prices,10 whereas increases in land tax would lead to lower land prices. The overall impact on property prices and investment is uncertain and depends on a range of policies affecting land use, but there is likely to be two effects of note. First, (depending in part on future policies affecting land use) property prices might increase because a more efficient tax system increases economic growth, some of which is captured in land rent — what was a 'deadweight' loss from stamp duty is captured in higher economic returns to the land owner. Second, land is a complement to property investment, so moving to a zero tax rate on capital investment (as stamp duty rates reduce) would increase the demand for land. International empirical evidence on the impact on building activity from switching an improved property tax for land tax is inconclusive (Oates & Schwab 1997) or mildly positive (Plassman & Tideman 2000).11

Land tax

The future Australian tax system should increasingly rely on land values as a tax base.

Along with natural resources (see Section C1 Charging for non-renewable resources), land tax is the only major tax that can be levied directly on economic rent. Shifting taxes away from mobile bases toward an immobile base, increases efficiency and potentially leads to higher long-term economic growth. Further, as land values tend to be correlated with growth in the economy and population, land tax is well-suited to future demographic pressures.

Current land taxes should be reformed to make them more efficient and equitable.

Reform the assessment mechanism

Land tax should no longer be based on aggregate land holdings. As well as discouraging large-scale investment in the rental property market, this approach does not appropriately target the economic rent from land.

The simplest approach would be to levy the tax at a flat rate on the unimproved land value, irrespective of total value. This would avoid arbitrary distinctions between tax burdens based on land parcel size or the landholder's characteristics. A flat rate would also avoid the problem of 'bracket creep', which, because of existing thresholds, has raised the real effective tax rate over time. However, a flat rate would reduce the top marginal tax on many properties relative to the current land tax. Some of these are likely to be land of high value, leading to windfall gains to some landowners. One approach would be to adopt a slow transition to the new rate structure, such as only slowly reducing existing land tax rates.

Alternatively, increasing marginal rates of tax could be applied to the economic rent in land. That is, stepped rates could be based on the value of the property per square metre, starting with a zero rate on low-value land. Higher valued land with more significant economic rents would pay a higher rate of tax (see Recommendation 52). Higher rates of tax on economic rents do not distort economic decision-making, as higher rates on labour or capital would. Targeting higher rates in this way would allow higher rates to be levied in areas of high demand for land.

An increasing marginal rate of tax may be justified the more certainty there is that the land valuation accurately reflects economic rent, rather than returns to other factors such as capital. This is particularly important should site (rather than unimproved) value be used to administer land tax. While easier to administer, site value has some merged capital costs (such as land clearing) included in the base (see Box C2–4). This approach is justifiable in areas, such as major urban centres where merged capital represents only a small portion of their value, but may be more problematic in low-value areas.

This approach would levy a higher rate on areas with greater economic rent, which would be more accurate than by using zoning as a proxy for economic rent. As reflected in Chart C2–8, while commercial land tends to have a higher value than the residential land in the same area, this does not hold across a whole State. Commercial land in country areas is often of lower value than residential land in capital cities. Basing the tax on land value per-square-metre would also ensure more timely changes in tax in response to changes in value of the land than if the assessment were based on changes in zoning, which can occur after the land has increased in value.

Chart C2–8: Land values by zoning in selected local government areas in Queensland

Chart C2–8: Land values by zoning in selected local government areas in Queensland

Source: Queensland Department of Environment and Resource Management (unpublished).

Broaden the base

Land used for owner-occupied housing should not be exempt from the tax base. The current exemption is inequitable, as it is likely that it contributes to renters bearing some or all of the tax. Excluding owner-occupied land also reduces efficiency of the tax, by distorting land use.

Broadening the tax base to include land used for owner-occupied housing would add significant revenue raising capacity to the tax base. This would improve the overall efficiency of the tax system, by reducing the reliance on alternative, less efficient taxes (see Recommendation 53).

Land used for primary production

Uniform application of the marginal rate scale on a per-square-metre basis with a low minimum threshold is likely to result in no tax paid by most land likely to be used for primary production. However, as it is based on value, this would significantly reduce the administration and compliance burden of land tax compared to the current use-based exemption. Further, a land tax would be inefficient if it affected land use. The scope for any inaccuracy in land valuation to affect land use is likely to be greater for lower-value land or for where it is difficult to separate the value from improvements to land from its inherent value. Targeting land use toward higher-value land above a minimum per-square-metre reduces the potential for the tax to affect land use.

However, primary production land on the fringes of urban areas (such as market gardens) may find its value increasing as demand for residential or industrial development increases. The value of primary production land in this situation could increase to the point where it becomes taxable even before it is zoned for more intensive development. This outcome reflects the increase in economic rent to the owner.

By basing the eligibility for tax based on value, rather than use, primary production land would not become taxable merely because it is converted to a different use, such as from primary production to biodiversity conservation.

Income-poor, asset-rich owner-occupiers

Some taxpayers may have difficulty in finding the cash to pay their land tax every year. For example, many low-income earners may live in valuable properties but not have cash readily available to pay their land tax liability.

For low-income earners who lack the cash flow to pay land tax every year the land tax could be deferred. The amount could accrue as a debt attaching to the property, with an appropriate caveat registered at the Land Title Office and a non-concessional rate of interest applied (in line with the standard variable mortgage rate). Asset-rich, income-poor persons could then allow the debt to accumulate until they move. The debt would be acquitted at the next transfer. Deferral arrangements already apply for local government rates in South Australia. To protect people in areas of declining value, the value of any debt should be non-recourse — that is, capped at the land value realised upon sale.

Land used for commercial and industrial use

A large share of land tax is currently raised from land subject to commercial and industrial use. However, large thresholds may mean that the full incidence of land tax is not borne in lower property values and fall instead on those who use land for business. Taxes on business inputs are a particularly high-cost source of tax revenue. In combination with stamp duty, levying increasing rates on a base with large thresholds means that the taxes borne by businesses are likely to be variable and, in some case, high. This affects efficient land use, as well as increasing the complexity and uncertainty for business.

A potential reform priority could be to remove the thresholds for land used for commercial and industrial purposes in return for rationalising the rate scale and for abolishing stamp duty on those properties (see Recommendation 54).

Valuation methodology

A redesigned land tax system could be simply administered by aligning local government rates with the land tax. Ideally, landowners should receive just one bill per year covering both and have a single point of contact for enquires, debt management and compliance. More significant simplification could be achieved if all local government rates had the same base as State land tax. This would reduce administration and compliance costs for individuals and businesses that pay rates across different councils in the same State and lower the cost of valuation, which is a significant part of the cost of collecting land tax and rates.

To be efficient, land tax valuations need to reflect the 'highest and best' use of the land — that is, its current market value — rather than its value in actual use (Oates & Schwab 2009). So long as the tax liability reflects its best use, then the tax does not affect the decisions of the owner. If some types of land (such as agriculture) are exempt or in other ways preferenced by valuation methods or land tax, then use-value assessment can delay development (England & Mohr 2003).

In major urban centres the administration of unimproved valuations has become increasingly difficult, with most States instead using site valuations. Very little unimproved land actually remains. There is declining knowledge of what land was like in its original state, and the historical information regarding fill and other improvements is increasingly difficult to determine. Consequently unimproved values continue to be regularly challenged by landowners in the courts with escalating costs for both land owners and the State (Hefferan & Boyd 2008).

To instil confidence in a system where greater revenue is raised from taxes on land values, greater investment in valuation and information collection methodologies would be warranted. This should include moving to a standard land or site value basis, using transparent and nationally consistent valuation methodologies and the updating of valuations on a consistently frequent basis to maintain alignment with movements in values.

Ensuring a smooth transition

This Review is not the first to consider a shift in the tax mix from inefficient transaction taxes towards a broader land tax base (for example IPART 2008, Productivity Commission 2004, Harvey 2001). While this would deliver substantial long-term benefits to the Australian community, the transition is clearly challenging. Transitional arrangements are important to build community acceptance and to minimise potential disruption (see Recommendation 54).

Successful transitional arrangements are likely to have a number of key design features.

First, any special transitional arrangements to a broader land tax regime should be limited to existing owners. Land tax is borne by existing owners of land when the tax is introduced. Future owners who are required to remit land tax are effectively 'compensated' by paying a lower price for the land. Future owners who remit tax payments only bear land tax on any unexpected capital growth in their land. Since this is associated with an unexpected windfall, there is no case for compensating future owners.

Second, the clearest need for a transition mechanism is for owner-occupied land. Existing owner-occupied landholders are likely to have bought their homes with the expectation that they would continue to be exempt from land tax. Additionally, a shift to land tax might generate perceptions of unfairness for people who purchased their property recently and paid stamp duty. Compared to longstanding holders of land, recent buyers would not have benefited from the land tax exemption and would face higher effective tax rates on their property over the time of ownership (see Chart C2–4). Therefore, for new land tax payers, transitional mechanisms may have to take into account the time at which properties were purchased. These concerns are ameliorated somewhat by the fact that reducing or abolishing stamp duty is likely to improve the property values of all owners.

Third, transitional mechanisms need to be designed to minimise harmful unintended consequences. If transitional arrangements exempted existing landholders from a tax until they sell, they would create lock-in effects that discourage sales. These should be minimised, recognising that lock-in caused by stamp duty is an important reason for removal of that tax. Further, during the time between announcement and introduction of a significant reform to taxation, there is the potential for significant market disruption. For example, if it were announced that land tax would replace stamp duty from a specific date in the future, people might defer the purchase of property pending the abolition of stamp duty.

Fourth, transitional arrangements that reduce tax burdens to facilitate reform also reduce revenue collections. These lower revenues mean that higher rates of tax must be applied to other tax bases or spending reduced. Some of the revenue cost could potentially be met by reductions in spending that may be less effective at improving housing affordability than tax reform. The overall revenue cost should be balanced, particularly where transitional arrangements over long time periods are concerned.

Transitional mechanisms are most likely to be effective when they reflect agreement between the Australian government and all the State governments. This recognises that the Australian government has access to larger and more efficient tax bases with which to finance revenue shortfalls, and that the reform would deliver significant benefits across the country.

In deciding on an acceptable transition mechanism it would be necessary to strike a balance between revenue cost, complexity of design and the extent of shift in policy. The balance of these different considerations is best made by government at the time any reform is undertaken. Several potential transition approaches are flagged in Box C2–5.

Box C2–5: Potential transition mechanisms for land used for owner-occupied housing

A simple option for facilitating the introduction of land tax on owner-occupied housing would be to levy the tax only on land that had been acquired after a given date, while continuing the exemption for all land held before that time. However, this complete grandfathering approach retains the lock-in effect of stamp duty for existing owners — they would begin to pay land tax only if they move — and would also come at a significant revenue cost.

A more flexible way of managing the transition would be to give purchasers of owner-occupied housing a choice between paying stamp duty or paying land tax, while grandfathering existing landholders. Once a property became liable for land tax it would remain liable. Purchasers who intended to move again soon would probably choose to pay land tax while purchasers who intended to live in the house for many years would probably choose to pay stamp duty. This option would have advantages and disadvantages. It would give purchasers more options. Since home buyers could avoid paying stamp duty up-front, access to housing would be immediately improved. Existing concessions and exemptions from stamp duty could be retained. Where people opt to pay stamp duty, this would reduce the revenue shortfall from the transition to land tax. On the downside, the transition could be very protracted unless some end date were specified.

An alternative approach may involve providing a credit to be used against any future land tax liability. A credit could be based on previous stamp duty paid or on the land tax expected to be paid over a set period of ownership. A full credit could be provided to people who buy between the announcement and introduction of the tax, to prevent people deferring purchases to avoid the tax. The credit would offset their annual land tax liability until it was exhausted. A partial credit — possibly on a sliding scale based on years held — could be provided to people who had paid stamp duty in a specified period before the announcement. A sliding scale would reflect revenue considerations and the fact that the effective tax rate from stamp duty declines with length of holding period. Alternatively, a flat credit irrespective of the length of time owned or amount of previously paid stamp duty could be provided to all existing holders of land for owner-occupied housing. This approach would be simpler to administer and allow longer deferral of land tax liabilities for holders of lower value land. Compared to permanent grandfathering of existing landholders, the use of a credit scheme would bring owner-occupied housing into the tax base sooner and lead to smaller revenue shortfalls.

Finally, a phase-in arrangement could be adopted. For example, the level of stamp duty could annually step down by one-tenth of its current level and the level of land tax could step up by one-tenth of its ultimate level. Under this arrangement, for example, a house sold in the third year would pay 70 per cent of the full stamp duty on the transaction and 30 per cent of the assessed land tax each year for a specified period. This would result in some stamp duty collections occurring in the phase-in period, reducing the fiscal cost compared to complete grandfathering. Limiting the period over which discounted land tax applies, perhaps to 10 years, reflects the fact that the discount will have lock-in effects eventually. After this period, the percentage paid in land tax could gradually phase up to the full rate. Similarly, people who never transact could remain fully exempt for a period, say 15 years, with the tax then gradually phased in, in line with the time periods applied to others. This would provide a measured phase-in over a predictable period and would avoid sudden jumps in liability.

10 Leigh (2009) suggests that the incidence of the tax falls on sellers.

11 For a survey of empirical studies see Anderson (2009).