Land Markets and Business Cycles
in the United Kingdom and Australia

David Richards



Estimation of Australian Land Values

Since 1944 the valuation of real property for property tax purposes has been the responsibility of the Valuer-General of each State. The "unimproved value" of land, the tax base of the States' Land Tax, has been distinguished from improvement values in such valuations for over a century. During the last quarter century its definition has been modified, and its label changed - "unimproved capital value" (UCV) has virtually given away to "site value" - but the only difference is that ancient, invisible man-made improvements are included in the valuation. This relieves the valuer of the task of imagining the bare site restored to its primitive state, and thus improves the valuation process without blurring the essential distinction between the part of real property which the owner can alter (the improvements) and the part which the owner cannot alter (its location and inherent qualities).

Thus, the land of Australia is valued officially. However, to estimate the aggregate market value of all Australian land in each year it is necessary to know the following facts about the valuations:

1. The dates to which they apply;

2. The degree to which they under-assess values at those dates;

3. The categories of land use which are exempted from valuation.

These facts are fairly accurately known, so official valuations may be adjusted accordingly to approximate true market values at any date. What is less well known is the degree to which valuation standards other than "market value" alter the relation of the official figures to market values. "Existing use" values are commonly the valuation standard. They are used for concessionary reasons, because they are below market values. Therefore, the estimates based on them must be regarded as conservative.

Another source of under-estimation throughout the following calculations is the treatment of the best modern assessments as representing true market value. In fact, to minimise challenges in the appeals courts, valuers do not seek to assess at more than 90% of market value (Hutchinson, c.1967: 7; 1981: 34-35).

The sources for the calculations for each year are as follows:


The Land Values Research Group (LVRG), based in Melbourne, Victoria, has been collecting and analysing data from official sources throughout the post-war period. Its 1961 booklet Public Charges Upon Land Values reprinted a table from the 1945 edition of the booklet which showed the ratios of direct charges on land rents to "assessed ground rent" in each state (p.5). Assessed ground rent was taken as equal to 5% of the assessed unimproved capital value of each state. For the relevant direct charges on ground rent we have to turn to the Group's booklet published circa. 1967, Land Rent as Public Revenue in Australia, written by Allan R. Hutchinson, Director of Research. Pages 17 and 18 list the State Land Taxes, local government site value rates and leasehold rents received by the States. From these two tables the assessed capital values of land in the States can be calculated.

The rates and land rent revenue of $542,000 for the Australian Capital Territory (ACT) are divided equally, as Brennan indicates that they were both charged at approximately 4% of assessed UCV.[28]

Assessed UCV was therefore $6,775,000 (271,000 x 100/4). The local rates would have reduced the bids for land leases which determined the assessed UCVs. The leasehold rents, however, would not have done so. The rents were effectively what was bidded, their corresponding UCVs were taken as a fixed multiple of the rent bids (20, at the time of the original lease auctions in the 1920s. In fact, UCVs were bidded, but in the knowledge that the actual payments would be one-twentieth of UCV each year).

This raises a question that was not tackled by the Melbourne research group. Do rents paid for public leaseholds reduce the assessed values (assumed to be related to market values) of the sites? Theory suggests not. Only expected taxes do so, as they reduce the net rents affordable. The net rents themselves are the direct basis of the market valuation. The Melbourne Group, however, treated leasehold rents in the same way as taxes, that is as subtracting from the net rents of sites. They therefore double-counted leasehold rents. These rents appeared in their calculations both as rents accruing to the public sector and as rents enjoyed by the private sector ("5% of the assessed unimproved capital value...is the...rent remaining in private hands" - 1961: 6). This error is avoided in the following calculations.

Far larger adjustments have to be made to the Group's calculations to compensate for land rents not included at all, however (see Appendix 2).

The three general adjustments to the valuation data before arriving at the figure in Column B of Table 1 are as follows:

1. No adjustment is made for time periods elapsed between each individual valuation date and the 1st July 1937. Land values were fairly stable during the 1930s, so assessments up to a decade out of date would not have been under-estimates.

2. LVRG (c.1967: 15) noted for the 1950s that "valuation methods were inferior and properties were valued much below market prices." This appears to be the reason why they considered doubling the official figures for 1st July 1957 (1961: 6), but adding only one-third to official figures for 1st July 1964. The one-third addition was to cover adjustment 1. above, and would probably apply equally to both dates. That leaves a one-half addition to the 1957 figure for under-assessment. The position would probably have been even worse in the 1930s. However, for the 1st July 1937 the official UCV figures are adjusted upwards by one-half.

3. Commonwealth and State government properties, churches, hospitals, charities and other properties are exempt from property taxation and are not included in official valuations. Official evidence suggests that between one-eighth and one-quarter might be added to aggregate valuations were these exemptions removed (LVRG, c.1967: 8; 1981: 9). However, in line with the estimates below for the 1980s, a uniform addition of one-tenth is made to official land valuations throughout this study.


LVRG's 1961 booklet lists the assessed UCVs for the six States (p.7). The three general adjustments are as follows:

1. & 2. As noted above, the LVRG considered that doubling the assessed UCVs was reasonable to cover these adjustments.

3. The doubled UCVs are then increased by 10%, as noted above.

ACT's rates were about $1.1m according to Hutchinson's update of Land Rent as Public Revenue in Australia, London: Economic and Social Science Research Association, 1981, p.23. Leasehold land rents were about $0.4m.[29] The latter were falling far behind market rents by this time, as land values were rising and leases were only re-appraised at 20-year intervals. For 1964 the taxable UCV of the ACT was 45 times the leasehold rents received (Hutchinson, c.1967: 7), so this multiple is taken as applying to 1957 as well.

1964/65 The data for this financial year is taken from Hutchinson (c.1967). The assessed UCV of land included the ACT (p.7). Land value data was provided for the Northern Territory for the first time (publicly owned perpetual leasehold tenures capitalised at 5%) though it only amounted to 0.1% of the total.

The three adjustments to official UCVs are these:

1. Hutchinson noted that the official valuations "lag behind the market by an average of three years. The true figure will be at least one third more than that shown..." (p.14). So one-third is added to the UCV total.

2. No adjustment is made for assessment ratio, which is assumed to be 90%, as Hutchinson indicated (pp. 7, 15).

3. 10% is added for the land exempted from valuation. 1976/77

Hutchinson (1981: 10) provided the UCV valuation for 1st July 1976. The adjustments are as follows:

1. The official UCV aggregate is raised by 42.2% to counter the effect of valuation intervals, as calculated by Hutchinson.

2. No adjustment is made.

3. 10% is added for exempt land.

1980/81 - 1989/90

Bryan Kavanagh of the LVRG has produced a fully adjusted series of UCV aggregates for the period from 1st July 1980 to 1st July 1991. His primary source is the Commonwealth Grants Commission, specifically the calculations made by the Commission's consultant Doug Herps using raw data from the States' Valuer-Generals. The adjustments are consistent with those made above.

Interpolations: 1970/71 - 1975/76 and 1977/78 - 1979/80

The missing values for 1st July 1977, 1978 and 1979 are interpolated roughly in line with the movement of dwelling rents in the Australian National Accounts (Canberra: ABS, 1982, p.21). The effect of rising property market yields during this period is taken into account. Thus, 1st July 1979 rents would have been 8% higher due to yields rising from 6.5% to 7.0% alone. Land prices, therefore, had to rise by only 10% to ensure an 18% rent increase.

For the years back to 1970 there is sufficient evidence, without undertaking a study of all the primary data, to provide a reasonably accurate picture of the trend in aggregate land values. Hutchinson (1981: 7), referring to the period before 1st July 1976, wrote that "land prices have been increasing for many years at rates varying among the different states but averaging at least 20% annually for metropolitan areas in every state..." An official report, Urban Land Prices, 1968-1974, [30] provided data on the average levels of new homesite prices in Australian capital cities, which showed that they rose by considerably less than 10% a year in the early years, but by over 20% in 1972 and over 30% in 1973. The rate of increase of dwelling rents estimated in the national accounts rose from 15% a year in 1972 to 24% in 1975, and then declined to 12% in 1980. The path of land rents is always smoother than the path of land prices. Ross King recorded the growth of the median price of dwellings in Melbourne as dipping below the rate of inflation (which was 17%) in 1985, and of city centre offices doing so even more vigorously.[31] There was a property slump in the mid-1970s, though housing recovered strongly in 1976.

This evidence suggests that the capital value of land may reasonably be interpolated by assuming the following annual growth rates to 1st July each year: 1971, 10%; 1972, 20%; 1973, 30%; 1974, 25%; 1975, 15%; 1976, 20%.

The increase in the resulting value for 1st July 1972 over that for 1964 in Table 1 (238%) ties in with an increase of about 220% for industrial land prices in the central area of Sydney over the same period. The increase in the values between 1st July 1957 and 1970 (318%) is considerably lower than the 350% increase in residential land prices in Sydney, 1959-1969 indicates,[32] but the latter might be expected to grow faster than national average land prices.

Mineral and Forest rents

These natural resource rents are not included in the calculations, partly because of the difficulty of valuing them, and partly because they are not captured for public revenue by the traditional real property taxes which are the subject of this study. However, a true assessment of the rental income of the nation for the purpose of comparison with other types of income and tax base should include them.

Main Paper * Appendix 2 * End Notes