The superiority of Site-Value rating

If we tax buildings, there will be fewer buildings. But if we tax land, there will not be less land. If we tax values of buildings, we will get not only fewer buildings, but also inferior buildings. But if we tax the value of land (exclusive of buildings and other artificial "improvements"), we will get neither less land nor less valuable land, because land is not the product of private effort or enterprise, while its value is conferred by the surrounding community, not by the owner who pays the tax.

The argument for imposing municipal rates on land values alone consists in the elaboration of these simple facts. The argument for any other rating system consists in the concealment or obfuscation of them.

1.1 No production without accommodation

At the risk of stating the obvious, we note that there can be no job creation and no associated wealth creation unless:

The first condition implies that if commercial accommodation is unaffordable, employment and economic growth will be directly frustrated. The second implies that if residential accommodation is unaffordable, employment and economic growth will be indirectly frustrated because

The implication is that lack of affordable accommodation not only represents inequity in the distribution of wealth, but also impedes the creation of wealth.

1.2 No accommodation without construction

Again at the risk of stating the obvious, we note that affordability of accommodation requires an adequate supply, hence an adequate rate of construction. Slower construction means tightening supply, causing rents and prices to rise faster than capacity to pay, so that a rising fraction of potential renters and buyers are forced out of the market. Again this offends not only against equity but also against efficiency, because it impedes the economic activities that the excluded renters and buyers were planning to engage in — to say nothing of the fact that construction itself is an economic activity, or that it stimulates further activity both upstream (e.g. in production of building materials) and downstream (e.g. in production and retailing of furniture and appliances).

1.3 Consequences of taxing construction

In view of the above, if values of buildings are included in the base of municipal rates — as they are in almost all Victorian municipalities — the following consequences are to be expected:

1.4 Corresponding (non-)consequences of taxing site values

A site is a piece of ground or airspace, whose value includes the value of any attached rights to construct buildings on that ground or into that airspace or to use the ground or airspace for particular purposes, but excludes the value of any actual buildings or other works. Taxpayers cannot create or eliminate sites. By increasing permitted building heights or rezoning land for more intensive uses, governments can effectively create sites or add value to sites; but these actions are not within the power of taxpayers.

From the viewpoint of taxpayers, then, the supply of sites is fixed. Therefore a recurrent tax on the value of a site, payable by the owner, cannot reduce the supply of sites, but can only induce the owner to use the site more productively in order to generate income to cover the tax — or sell the site to someone who will. Selling the site does not eliminate the site or the need to use it productively, but merely transfers both to the buyer.

So, compared with the above consequences of including building values in the rating base, the effects of rating site values alone are, respectively, as follows:

1.5 Spurious counter-arguments

It might be said that to exclude values of buildings from the rating base is to "narrow" the tax base. We note in passing that if some things are more fit for taxation than others, narrowing the tax base is not always a bad idea. But more importantly, if rates are imposed on sites but not on buildings, in what sense is the tax base "narrowed"? Rates, by definition, are imposed on property, and every rateable property includes a site. (Even the value of a unit in a strata-titled complex includes a site value, as is clearly shown by the fact that units of similar size and quality have different values in different suburbs.) So, if only the sites are rated, there is no reduction in the number of taxable properties or property owners. But there is, as we have seen, a considerable reduction of perverse incentives.

It might be said that if building values are included in the rating base, the incentive to build can be restored by differential rating, with a higher tax rate on vacant lots. But if it is admitted that building should not be discouraged, why not eliminate the problem at the source by rating the site alone? A higher rate for vacant lots creates a new problem: exactly how much does one need to build before the lot ceases to be vacant? If that problem is satisfactorily solved, it remains the case that the owner who builds or preserves more accommodation will pay more tax than the owner who builds or preserves less, as long as the latter builds or preserves something. Thus the "advantage" of such arrangements is that land owners, instead of having an incentive to build nothing and demolish everything, have an incentive to build or preserve as little as possible!

It might be said that if building values are included in the rating base, exemptions can be given for renovations carried out for the benefit of disabled persons. But then what qualifies as a disability? And what building modifications are to be recognized as responses to that disability? What loopholes will be opened up, and what collateral damage will be done when they are closed? If buildings are excluded from the rating base, these problems are completely and automatically avoided.

It might be said that to include building values in the rating base is "progressive" because opulent buildings indicate capacity to pay. But this ignores the fact that the fractions of total property values attributable to buildings (as opposed to sites) tend to be higher in suburbs where the property owners are less wealthy [2]. The familiar sight of "McMansions" on small lots in outer suburbs is a case in point. Besides, more valuable buildings tend to be newer, and newer buildings tend to be more heavily encumbered with mortgages that offset the owners' "capacity to pay". But we do not need to debate this point further, because:

Descending from the fallacious to the ridiculous, it might be said (nay, has often been said) that because the market value of a site is built up by developments on surrounding sites, taxing the said value deters the said developments. That the developments are generally done by parties other than the one paying the tax, and therefore cannot be discouraged by the tax, is not mentioned. "Ah," says the objector, "but if the same party develops a large cluster of sites, some of the uplift in the value of each site is due to work by the same party on the other sites in the cluster!" Oh. So because it is so iniquitous that site value taxation sometimes falls partly and accidentally on productive activities, we must expand the tax base in ways that always fall entirely and deliberately on productive activities! But such are the depths to which one must often descend if one first chooses the conclusion and then looks for arguments to justify it.

1.6 De facto building taxes

The construction or extension of buildings is deterred not only by the rating of building values, but also by any other property taxes that tend to increase with the quantity of accommodation. In Victoria, at the local level of government, such taxes may include "municipal charges", water access charges, and sewer access charges (all of which are levied on a per-property basis, which presumably means a per-unit basis in a multi-unit complex), and garbage and recycling charges (which appear to be levied on bins according to capacity, and which therefore increase with the number of dwellings). They do not include water consumption tariffs or sewer discharge tariffs, because such tariffs, if they are imposed at all, are not simply holding costs on property.

Per-dwelling taxes are defended on the ground that it costs the local government more to service a multi-unit complex than a single dwelling. But this is already taken into account by site-value rating, in that the mere permission to build more units on a site increases the site value. Defenders of per-dwelling taxes might further argue that it is not fair to charge a property owner for building rights that are not used. But it is fair in the sense that a site owner who fails to build the maximum permitted number of dwellings is needlessly restricting the supply of accommodation, and thereby needlessly causing serious economic damage (recall subsections 1.1 & 1.2). If owners are charged for the right to build more dwellings whether they use it or not, they have some encouragement to use it — and that is healthy.

(Of course, because multi-unit complexes tend to be more valuable than single dwellings, the arguments offered for per-dwelling taxes are also offered for including buildings in the rating base — and are invalid in the latter case as in the former.)

We therefore suggest that all local de facto building taxes be rolled into site-value rates — as described under the next heading.

2. Implementation with no losers

If all existing rates and local property charges were replaced by site-value rates in the simplest possible manner, some property owners would expect their bills to decrease while others would expect their bills to increase; and the latter would punish the offending legislators by all available means, including disinformation concerning the effects of the reform on other ratepayers and the wider economy. To minimize the risk of organized opposition and the risk that any such disinformation will be believed, one must ensure that there are no losers in the transition from the old system to the new.

The arguments for site-value rating (subsections 1.3 & 1.4) remain valid in the presence of thresholds; subjecting the rateable site value to a threshold does not introduce any of the claimed disadvantages of rating buildings and, as long as some tax is paid, does not negate any of the claimed advantages of rating sites. The same applies if the threshold is allowed to vary from one site to another (as long as such variation will not be influenced by any future activities of taxpayers and therefore cannot act as an incentive or disincentive). Hence we can obtain the advantages of site rating, with no losers, if each local council proceeds as follows:

Thresholds could be negative if necessary. The initial tax rate is theoretically arbitrary; but obviously a higher rate would reduce the number of negative thresholds, and one could choose the initial rate so that the lowest threshold is zero. The purpose of adjusting thresholds in line with the CPI is to minimize the effect of inflation on rate bills, especially for ratepayers with high thresholds.

As the present Victorian Local Government Act, at face value, does not permit thresholds, the above plan would require enabling legislation at State level; local councils cannot implement it by themselves.

Even if the basic "rates" are currently levied on site values alone, as in Queensland and New South Wales [3], the above implementation method can be used to eliminate other recurrent property taxes (or "charges" or "levies") which are imposed in addition to the basic "rates", and which are not based on site values. (Indeed, such charges exist in Queensland and NSW.)

3. The empirical evidence

In any other debate about taxation, it would be common ground that taxing a product reduces its production and all the effects that follow therefrom. The proposition is so simple, so reasonable, so commonplace, so confirmed by experience, and so often assumed as a foundation of policy, that to question it would be regarded not as skepticism but as folly or knavery. Applied to local rates, the proposition implies that the inclusion of building values in the rating base reduces the supply of buildings and all the economic activity that depends thereon. And yet, through most of the world (with notable exceptions including Queensland and NSW), local property taxes on the combined values of land and buildings are the norm, and it is the advocates of site-value rating who, if they are not simply ignored (as they usually are), are called upon to prove their case with hard data — and to prove it again and again.

The so-called evidence against site-value rating can be quickly disposed of. If (say) Smithville adopts site-value rating, and if this causes a burst of construction, yielding a generation of new buildings, there may be a subsequent dip in construction as those buildings serve out their economic lives, followed by a second, larger, broader "hump" in construction as those buildings become ripe for replacement with larger and more valuable buildings. Meanwhile, if Jonesville continues to tax buildings, and if this delays construction of new buildings, the mere passage of time will increase the optimal sizes and values of those buildings when they are eventually built. So if the researchers choose a "long-term" timeline that includes some delayed construction in Jonesville but excludes the second "hump" in Smithville, they may be able to claim that Smithville got an early payoff followed by a "long-term" slowdown, because what they call the "long term" is only a fraction (and an unrepresentative one) of the building life cycle. It seems that three studies of this sort were done in the early 1980s [4, p.22].

In contrast, the evidence in favour of site-value rating has grown so voluminous that only a small sample can be reported here.

3.1 Names of rating systems

At present, local councils in Victoria must levy rates on one of three bases: (i) capital-improved value (CIV), which is the total market value of the land plus buildings and other works; (ii) net annual value (NAV), which is the net annual rent of the property (defined by statute as at least 5% of the CIV for a commercial property and exactly 5% for residential property); or (iii) site value (SV). Thus CIV and NAV include buildings in the rating base, while SV does not.

From 1968 until the Kennett government imposed the present regime, Victorian councils also had the option of a composite rate, commonly called a "shandy", which raised half the revenue from SV and half from NAV (other proportions were permitted but never used). Some foreign jurisdictions have used a weighted sum or average of SV and CIV (also called a composite tax) or, equivalently, a lower rate on the building and a higher rate on the site (a two-rate or split-rate tax). These various hybrid systems also tax buildings, albeit less severely than CIV or NAV.

3.2 Camberwell, 1920s

In A History of Camberwell (Lothian Publishers, 1980, p.86), Geoffrey Blainey wrote:

In the five years prior to 1923, the total number of dwelling permits issued in Camberwell was 2051. In the following five years the number increased to 4373. Camberwell tops suburban building statistics until 1946 [4, pp. 8,12].

3.3 Hutchinson's zonal studies

Before the Kennett reforms, the Melbourne metropolitan area included a large number of municipalities with a variety of rating systems. Hence it was in Melbourne that A.R. Hutchinson, founding director of the Land Values Research Group [5], pioneered the technique of grouping the municipalities into zones and correlating tax policy with economic activity within each zone. His results were published in Melbourne in 1945, but also attracted attention in the USA [6].

Hutchinson divided the municipalities of greater Melbourne into six zones according to distance from the CBD. For the purpose of comparing rating systems, zones 1 to 3 were useless because they contained no SV jurisdictions. But each of zones 4 to 6 (the outer zones) contained a mixture of rating systems. In the years 1928–1942, the number of dwellings constructed per available acre was 50% higher in the SV-rating areas than in other areas for zone 4, and more than twice as high for each of zones 5 and 6. Extrapolating these figures, Hutchinson concluded that if all municipalities in greater Melbourne had used SV rating, the additional construction would have eliminated Victoria's housing shortage (then estimated at 40,000 dwellings).

In 1949 the Land Values Research Group published "Rising Municipal Costs: A Comparison of Relative Abilities of Alternative Rating Systems to Provide Increased Rate Yield". This was a survey of the per-acre increases in rate bases and yields for seven (then) outer suburban areas using SV rating (site only) and ten which used NAV rating (site and buildings), over a 20-year period. The SV councils were Brunswick, Coburg, Camberwell, Caulfield, Essendon, Oakleigh, and Sandringham. The NAV councils were Brighton, Footscray, Hawthorn, Kew, Malvern, Northcote, Moorabbin, Preston, Williamstown, and Heidelberg. The per acre increase in total NAV was £85 for the SV councils, and £57 for the NAV councils [4, p.12]. In other words, the use of site-value rating was associated with greater gains in property values.

3.4 Other studies on Melbourne

Melbourne also provided opportunities for longitudinal studies as the voters of individual municipalities switched to or from SV rating ("to" being more common).

The American Institute of Economic Research conducted a study of building activity in Victoria from 1927 to 1951. All councils that had changed from NAV to SV in the 1940s were shown to have experienced marked increases in building activity immediately after the rating change. In suburban areas, construction occurred preferentially in SV-rating municipalities [4, p.13].

Notwithstanding South Melbourne's experience, the nearby municipality of Caulfield switched from SV rating to composite rating in 1969-70. In Caulfield, the number of building permits issued for 1969–72 was 66% below the number for 1966–69 [8].

3.5 Comparisons between Australian States

Taking advantage of the fact that some Australian States mandated SV rating while others did not, Hutchinson made interstate comparisons of construction rates (for 1921–33) and ratios of improvement values to land values (for 1939–40), noting that both measures were superior in the SV States. He also noted that there was a net interstate migration into SV States between 1929 and 1938, indicating that prospects in those States were perceived as superior [6].

In 1963 the Land Values Research Group published a "A Study of the effects of local government rating systems upon the social and economic development of the Australian states". From that study comes the following table [4, pp. 14–15], showing the change in acreage under crops in mostly SV-rating States and mostly NAV-rating States during two time intervals, one before the war and one after (during the war years agriculture was under government direction).

Notice that those states which had greater use of site value had greater increases in the acreage being used for crops. In other words, where site-value rating applied, farmers were encouraged to put the land to use.

1929–1938	1946–1958
SV rating	NAV rating	SV rating	NAV rating
`Qld +66% NSW +22% WA +3% Total +21%`	`SA -5% Vic -10% Tas -8% Total -8%`	`Qld +76% NSW +5% WA +71% Total +35%`	`SA +7% Vic -6% Tas -6% Total -1%`

3.6 South African experience

The following table shows the growth in the value of improvements on land between 1974 and 1984 (right column) for various rating bases (left column) for 48 South African towns [9]. The middle column shows the number of towns with each rating base.

The two "towns" with CIV rating were Cape Town and Port Elizabeth. Both had the advantage of being ports, and Cape Town had the added advantage of being the national and provincial capital. But, as the table shows, their ten-year growth in the value of improvements was less than half that of the SV towns.

3.7 North American experience

Prof. Cord then gives some more details on 22 of those studies. His references to "two-rate" taxes are by comparison with "one-rate" taxes on the combined values of land and buildings, not with pure site-value rating (which is very rare in the USA). Again we quote him:

3.8 The people have spoken (and been ignored)

In the years 1920 to 1986, there were 90 plebiscites in which the ratepayers of various Victorian municipalities had the opportunity to choose between two rating systems. Some municipalities had more than one plebiscite over that period, and many had none. In 70 out of 90 plebiscites, the voters chose site value [4, pp. 33–36]. By 1993, the breakdown of rating systems was as shown in the following table [4, p.4].

Region	SV	NAV	CIV	shandy	Total
Metropolitan:	26	27	1	1	55
Provincial:	21	23	–	–	44
Rural:	8	92	11	–	111
Total:	55	142	12	1	210

Nevertheless, Victoria's Office of Local Government, in the same report in which the above table first appeared [Rates: Proposals To Improve [sic] Victoria's Municipal Rating System, Office of Local Government, 1993], recommended that all newly amalgamated councils adopt CIV. Under legislation subsequently passed by the Kennett government, councils may apply an arbitrary number of different rates for different classes of land if, and only if, they assess property according to CIV [Local Government Act, s.161]. Inevitably, council after council introduced CIV rating in order to be free to buy the votes of particular classes of land users. In effect, the Kennett government imposed CIV rating from the top under the guise of giving choice to local councils (not to be confused with the ratepayers).

Today, the only Victorian municipality that retains SV rating is Monash. In Prosperity for the Next Generation (2001), the Monash City Council boasted:

We submit that if it was good enough for the Kennett government to impose by stealth a rating system that was overwhelmingly rejected by voters for two thirds of a century, it is good enough for the present Victorian Parliament to impose the system that the voters tended to endorse when given the chance: site value. That is the only system for which the Parliament could claim a mandate. And if the system were introduced as we propose — that is, with no losers — the need for the mandate would be minimized.

4. Conclusion

The adoption of site-value rating, especially when combined with the removal of supplementary local property taxes, would increase the supply of buildings, making accommodation more affordable and raising the speed limits on job creation and economic growth. In Victoria, as in all Australian States, such a reform could be imposed top-down by the State Parliament.

In the past, State legislators might have hesitated to impose site-value rating for fear of a backlash from the losers. But now we have shown how site-value rating can be implemented with no losers, leaving the politicians with no excuse for further inaction.

Notes

[1] Reducing the rights of workers relative to employers tends to replace cause (b) by cause (a), with the result that job seekers settle in jobless locations because they cannot afford the rents in locations where job prospects are better. This does not solve the problem, but does make it easier to blame the job seekers!

[2] See e.g. M. Gaffney, "A Cannan Hits the Mark", American Journal of Economics and Sociology, vol.63, no.2, pp. 273–290 (April 2004), reprinted as ch.21 of Robert V. Andelson (ed.), Critics of Henry George, 2nd Ed. (Blackwell Publishing, 2003), vol.2.

[3] The site value (as it is called in Victoria) or land value (as it is called in NSW) includes the value added by merged improvements, i.e. improvements that could be mistaken for natural features; such improvements may include historical clearing or grading. Technically, Queensland uses the unimproved value, which attempts to exclude merged improvements (with debatable success). But the difference between unimproved value and site value is usually insignificant — whereas the difference between SV and CIV (subsection 3.1) is crucial.

[4] P. Anderson, Victoria's Municipal Rating System (Melbourne: Australian Institute of Urban Studies [Victoria Division], 1996).

[5] See http://lvrg.org.au.

[6] Harry Gunnison Brown, "The Challenge of Australian Tax Policy", American Journal of Economics and Sociology, vol.8, no.4, pp. 377–400 (July 1949), Part I, reprinted in Selected Articles by Harry Gunnison Brown (New York: Robert Schalkenbach Foundation, 1980), pp. 158–163.

[7] Harry Gunnison Brown (with E.R. Brown), "Incentive Taxation in Australia", American Journal of Economics and Sociology, vol.26, no.4, p.416 (October 1967), reprinted in Selected Articles by Harry Gunnison Brown (New York: Robert Schalkenbach Foundation, 1980), pp. 227–228.

[8] Fred Harrison, The Power in the Land (New York: Universe Books, 1983), pp. 195–196.

[9] Godfrey Dunkley, That All May Live (Roosevelt Park, RSA: A. Whyte, 1990), p.124, quoted in M. Gaffney, F. Harrison, and K. Feder, The Corruption of Economics (London: Shepheard-Walwyn, 1994), p.239.

[10] Steven B. Cord, The Golden Key To Continuous Prosperity (AuthorHouse, 2005), ch.3.

Revision history

Version 2007.02.14 (considered a draft) was the original. For some of the history of Victorian rating systems and for some of the empirical evidence on their effects, the author is indebted to Lev Lafayette's Site Revenue for Public Finance (submitted to the Executive of Prosper Australia on October 26, 2006).

In Version 2007.05.25, the author corrected the opening paragraph of subsection 3.8 (on local plebiscites), added an introduction to Section 1, expanded subsection 1.5, extended the introduction to Section 3 (to deal with the "early payoff" of SV rating), added dates to subsection 3.5, amended the Conclusion (explaining why a paper about local rating is addressed to State legislators), and added reference [4].

In Version 2007.05.31, the author added a paragraph to Section 2, corrected reference [8], and added minor details (including hyperlinks) to other references.

In Version 2008.04.14, the table in subsection 3.8 was corrected.

Version numbers are dates in the form yyyy.mm.dd. Minor textual corrections and adjustments are not necessarily recorded.

The superiority of Site-Value rating — and how to implement it with no losers

Contents

1. The rational arguments