Housing affordability

In a speech delivered in Edinburgh on July 17, 1909, a certain British MP declared:

The MP was Winston S. Churchill [1]. In view of his observations, what assurance do we have that "rent assistance" for social security recipients will not be competed away in higher rents, or that grants for first home buyers will not be competed away in higher prices? Indeed, what assurance do we have that rent assistance and home buyers' grants are not covert attempts to enrich property owners (landlords and potential sellers) at taxpayers' expense, on the pretext of helping renters and potential buyers? If the political alignments are such that nothing can be done for renters and potential buyers unless there is a side-benefit for landlords and potential sellers, can anything be done at all? Or is housing affordability a zero-sum game in which renters and potential buyers, being the weaker side, are condemned to lose? To answer these questions, we must consider the ingredients of affordability and unaffordability, the balance between them, and the affect of each on established property owners.

1.2 Mostly not capacity to pay

According to Churchill, if workers in a particular area get more spending power (e.g. because they save on bridge tolls or bread), rents in that area will rise to cancel the advantage. This claim was nothing new, but was simply a statement of Ricardo's law as applied to wages and residential rents. It is also an obvious consequence of locational arbitrage: if workers in a particular location have a particular advantage, more workers are induced to move there, and the consequent demand for accommodation pushes up rents in that location until the rent premium cancels the initial advantage and stops the migration. The locational component of rent is the land rent or, more generally, the site rent — that is, a payment for the use of space. It does not include the so-called "rent" of buildings, because the value of a building is limited by its depreciated construction cost, and because the value of a site reflects the value of its location even if no buildings yet occupy the site.

Similar comments apply to the prices (i.e. lump-sum values) of sites, because these are capitalized rental values (and may be re-annualized as mortgage interest).

Even if the increase in spending power is not restricted to particular locations, the overall supply of land is still fixed. More importantly, from the viewpoint of private economic agents, the supply of land with a particular zoning, or within a particular distance of particular supplies or markets or infrastructure, is also fixed. So any increase in effective demand for land is not offset by a market-driven increase in supply. And of course no economic agent — worker or employer — can function without occupying space. Consequently, any increase in capacity to pay for that space is converted by the market into higher rents.

If the increase in capacity to pay is confined to part of the population, the increase in rent will be some sort of average of the increases in capacity to pay. If those who get the increase in spending power are competing for the same locations as those who do not, the former will be winners at the expense of the latter; for the winners, the increase in spending power will be only partly competed away. This is probably the case with rent assistance that is available to welfare recipients but not to low-paid workers, and with the First Home Owners' Grant (except that the latter affects prices rather than rents). But if those who get the increase in spending power are competing only among themselves, the increase will be competed away. For example, it would not be surprising if a compression in income relativities were mostly converted into a corresponding compression in rent relativities, with higher minimum incomes being mostly cancelled by higher rents and prices at the bottom of the market.

1.3 Partly (de-)skewing of supply

As the housing market is stratified, it may be oversupplied at one price point and undersupplied at another. The most severe form of housing stress — homelessness — represents undersupply at the bottom of the market, and may therefore be alleviated to some extent by encouraging investors to build more accommodation aimed at the bottom. An investor with a limited budget may choose to build a larger number of cheaper dwellings or a smaller number of more expensive dwellings. The former choice can be encouraged by a bounty that takes the form of a fixed sum per dwelling. This could be funded by abolishing the present First Home Owners' Grant, which, as implied above, is partly self-defeating in that it is partly converted into higher prices. Thus we have our first specific proposal:

The overall supply of housing also matters, especially for those at the bottom of the market. If there is an overall shortage, people who would otherwise take more up-market housing will be forced into the cheapest housing, and people of lesser means will be forced onto the streets. Furthermore, the stratification of the market is not completely rigid; an oversupply at one level of the market will attract occupants from other levels, causing a downward influence on demand (hence prices and rents) at other levels.

But the proposed NHBG also addresses overall supply: the NHBG is contingent on the construction of new dwellings, whereas the FHOG is available for the mere acquisition of existing dwellings.

1.4 Not lack of utility

If the infrastructure in a particular location is run down or overloaded, rents and prices of land in that location may fall, but only because the utility of living on that land is reduced. This may mean that it is easier to afford housing in that location, but does not mean that it is easier to afford accommodation of given utility — which, we submit, is what greater affordability ought to mean. Similarly, if land values rise in a particular location because of better infrastructure, this does not of itself make it harder to afford housing of given utility. It does not even imply that renters in that location are forced to move to a cheaper location, because better infrastructure often translates into greater capacity to pay rent. For example, new freeways or cheaper public transport make it possible to save money on commuting, while better coverage by public transport may make it possible to live without a car; in either case, the savings may be spent on accommodation. But even if renters indeed must move because their location goes up-market, that is better than not being able to afford accommodation at all.

1.5 Not simply low rents or prices

All else being equal, a reduction in prices or rents means an improvement in affordability. But all else need not be equal. If a reduction in prices or rents is offset by, or caused by, a loss of utility or (which may be the same thing) a loss of capacity to pay, that does not constitute an improvement in affordability. Similarly, if rents or prices somehow remain constant while utility improves or spending power increases, that is an improvement in affordability.

1.6 Not necessarily a cost for present owners

As a improvement in affordability does not necessarily mean a reduction in rents or prices, it does not necessarily come at the expense of incumbent property owners. If the economy grows, and if the benefit of that growth is shared between property owners, tenants, and prospective buyers, then the benefit to tenants and prospective buyers amounts to an improvement in affordability; yet the gain for tenants and prospective buyers is not a loss for current owners, but rather a share of the overall gain, of which current owners also receive a share. Thus, if the policy that improves housing affordability also promotes overall economic growth, housing affordability is not a zero-sum game.

1.7 A prerequisite for full employment

Indeed, housing affordability by itself promotes economic growth. To prove this, suffice it to note that:

The second condition implies that if housing is unaffordable, employment (and hence economic growth) will be restricted either because workers, on the wages offered, cannot afford housing where the jobs are available, or because employers, in view of the wages needed to pay for housing, cannot afford to offer jobs. (The fact that job seekers still outnumber job vacancies indicates that the latter cause is dominant.)

1.8 Answer: A matter of competition

In summary, affordability is an increasing function of utility and capacity to pay (which are sometimes the same thing), and a decreasing function of rent or price. But the combined effect cannot be determined by separately evaluating the changes in utility, in capacity to pay, and in rent or price, because these variables are not independent; an increase in utility or in capacity to pay, which by itself increases affordability, also tends to cause an increase in rent or price, which by itself reduces affordability. To see which effect is dominant, we must focus on the mechanism by which a change in utility or in capacity to pay is converted into a change in rent or price. That mechanism is competition. If the vacancy rate is low, prospective tenants and owner-occupants face much competition while prospective lessors and sellers do not; therefore the successful bidders will need to try hard, and will consequently pay high rents or prices relative to the utility of the housing and their spending power. If the vacancy rate is high, prospective lessors and sellers must compete for tenants and buyers while prospective tenants and buyers face little competition; therefore the successful bidders need not try very hard, and will consequently pay low rents or prices relative to the utility of the housing and their spending power. Thus we arrive at a short definition:

This definition may be applied to the housing market as a whole or to particular regions or segments or strata of the market.

In a context of overall economic growth, greater affordability can coexist with absolute gains for lessors and sellers provided that the gains of renters and buyers do not absorb all of the growth.

1.9 The problem defined

A policy promoting affordability of housing must strengthen the competitive positions of renters and buyers relative to lessors and sellers. If this policy is not to encounter opposition from prospective lessors and sellers (who obviously have the means to campaign against it), it must also promote overall economic growth and ensure that not only renters and buyers, but also lessors and sellers, share the benefit.

2. Four crucial distinctions

2.1 Buildings vs. sites

A site is a piece of ground or airspace, including any attached rights to construct buildings on that ground or into that airspace, but excluding any actual buildings or other works. Buildings can be produced by private entities (individuals or firms). Sites cannot. Increasing the permitted building height on a residential lot or permitting finer subdivision may effectively create additional housing sites; but that is the prerogative of governments, not private entities. From the viewpoint of private entities, then, the supply of sites is fixed.

Taxation can reduce the supply of buildings by deterring their construction. But taxation cannot reduce the supply of sites, because taxpayers cannot create sites or destroy them or move them into or out of the taxing jurisdiction. Similarly, subsidies can increase the supply of buildings, but not the supply of sites.

As every home must include a habitable building or part thereof, any increase in the supply of residential buildings tends to strengthen the competitive positions of renters and buyers — that is, to improve affordability. So the preceding paragraph implies that for maximum affordability:

Our proposal to replace the First Home Owners' Grant (FHOG) with a New Home Builders' Grant (NHBG) is consistent with the latter point; the NHBG is purely for the construction of a building, while the FHOG is partly for the purchase of a site.

This emphasis on buildings arises not because a building can be worth more than its depreciated construction cost, thus eclipsing the significance of site rents, but rather because, for those who lack the means to add to the supply of buildings, a shortage of buildings is equivalent to a shortage of usable sites.

2.2 Improved land vs. unimproved land

Buildings, fences, walls and other recognizable artificial structures on or under land are called improvements. The construction of all such items, like the construction of buildings, can be deterred by taxation. The improved value of a block of land includes the value added by improvements within the block. The unimproved value of the block excludes the value added by improvements within the block, but includes the locational value added by improvements on surrounding land — including roads, power lines and other services that pass by just outside the block.

For present purposes we may take the site value as being synonymous with the unimproved value [2]. In particular, the site value includes the locational value, because the party buying the block usually cannot afford to add locational value by "building up" the surrounding area.

But the scope of the "block" is relative. For the consumer buying a single lot in a new estate, the block is the single lot, and the adjacent roads and other services provided by the developer add to the locational value, hence the site value. But for the developer, the block is the whole estate, and the services provided by the developer are improvements to that block. That is, roads, footpaths, drains, sewers, underground power lines, water mains, and gas mains provided by a developer within an estate are, from the viewpoint of the developer, improvements.

2.3 Developed land vs. raw land

What a home builder or home buyer calls "land" is not raw land, but developed land, i.e. land that has been serviced by roads, footpaths, drains, sewers, power, water, etc., and which has been subjected to various government charges. To the extent that these services and charges are costs borne by developers as a condition of development, they must be recovered through the resale prices of developed land, or else development will become uneconomic. The implication is that, while taxes cannot reduce the supply of land in general, they can reduce the supply of developed land; this is possible because a developed estate is substantially improved from the developer's viewpoint, although a vacant lot in that estate is unimproved from the buyer's viewpoint.

These observations, however, tell us only that taxes can reduce the supply of developed land; they do not tell us which taxes actually do so.

2.4 Transaction taxes; holding taxes; transfer taxes

A transaction tax is one for which the tax liability is attached to an avoidable economic exchange (the "transaction"). A holding tax, on the contrary, is a periodic tax payable by the owner of an asset regardless of any transactions (e.g. rent payments) that occur during the period of ownership. In this paper, a transfer tax is one which is payable at the time of transfer (sale or purchase) of an asset, but which may or may not be apportioned to the transfer price; in form it resembles a transaction tax, but in substance it may be a transaction tax or a holding tax, depending on how it is calculated.

Obvious examples of holding taxes on property are municipal rates and State land taxes. If municipal rates are levied on improved values, they deter construction and extension of buildings and thereby restrict the supply of housing and damage affordability. But if they apply to site values only, they have no such effect; on the contrary, they encourage owners of vacant sites to build on the sites and seek tenants in order to cover the rate bill, thereby adding to the supply of accommodation, strengthening the bargaining position of tenants, and improving affordability. All this is consistent with the principle that any taxes on residential property should be on sites rather than buildings.

An obvious example of a transaction tax on property is a conveyancing stamp duty based on the sale price. This is also a transfer tax, as the tax liability is realized on transfer of the title. If the stamp duty is payable by the seller, the seller will try to add it to the price. If it is payable by the buyer, the buyer will try to subtract it from the price. In the end, the cost will be shared between the buyer and the seller in inverse proportion to their bargaining power, regardless of who actually "pays" the tax to the revenue office. But, as the tax is partly borne by the buyer, it damages affordability.

Only by resisting or delaying the sale can the seller shift any part of the burden onto the buyer (if the seller remits the stamp duty) or prevent the buyer from shifting the whole burden onto the seller (if the buyer remits the duty). In either case, it is by discouraging sales that the stamp duty affects affordability.

The same logic would apply even if stamp duty were levied on the site value only, because the stamp duty would still impede sales of sites; the fixed stock of sites is immaterial to the argument because it does not imply fixed turnover of sites.

But the same logic does not apply to holding taxes on sites, which do not target turnover, but rather encourage site owners to build accommodation and seek tenants (or sell the sites). Thus we have shown that for maximum affordability:

Now suppose that the stamp duty payable on each transfer is some fraction of the real increase in the site value since the last transfer. (In practice, such a duty would be payable by the seller because the seller knows, or should know, the taxable site value at the time of acquisition, and is therefore able to work out the tax bill in advance, without relying on anyone else's honesty.) Let us call this arrangement a site windfall tax (SWT). The SWT, being payable on each transfer of title, is a transfer tax. But it has the substance of a holding tax in that the total tax paid on an asset over the long term is not proportional to the frequency with which the asset changes hands. With a true transaction tax, a higher frequency of transfers over the same time frame means a larger total tax bill, and each transfer creates an additional liability; but with the SWT, a higher frequency of transfers merely divides the taxable capital gain into a larger number of smaller steps, and each transaction merely realizes an already accumulated tax liability. Moreover the SWT, by definition, cannot cause a property investor to make a capital loss, but merely reduces capital gains (and, if allowed to be negative, reduces capital losses), whereas a true transaction tax can cause or increase a capital loss. Thus the SWT does not discourage sales in the ways that a transaction tax does, and therefore does not restrict supply and damage affordability to the same degree. This leads to our next proposal:

A more obscure transaction tax on property is income tax on rental income. This reduces the return on constructing accommodation and/or offering it "to let", and on both counts reduces the supply of housing and damages affordability. If the assessable income were the imputed (potential) rent rather than the rent paid, the tax would then be a holding tax and would no longer discourage letting to tenants (the presently taxed transaction); indeed, it would force the owner to seek tenants in order to cover the tax bill, or to sell the property, and would therefore strengthen the bargaining position of tenants and buyers and improve affordability. If, in addition, the tax assessment were confined to the site rent, it would no longer discourage construction; indeed, it would encourage the owner of a vacant site to build accommodation in order to attract tenants and cover the tax bill, or to sell the site, and would therefore further improve affordability. Hence our proposal:

3. Taming those development levies

3.1 The spin

As building values are limited by depreciated construction costs, wide fluctuations in residential property values are mostly fluctuations in site values. This elementary fact is traditionally obfuscated by describing the state of the market in terms of "house prices" rather than "home prices" or "land prices". But in recent years there has been an outbreak of glasnost as the major property developers, through their mouthpieces such as the Housing Industry Association and the Property Council of Australia, have campaigned against the increasingly prevalent lump-sum infrastructure levies imposed on developers by State and local governments: the developers blame the levies for allegedly driving up prices of developed land. As the Labor Party is in power in all States and Territories, the "conservative" Federal government has found it convenient to join the developers' chorus, so that now even the Prime Minister hammers the point that unaffordable housing is really unaffordable land [3], having previously boasted that "I haven't found anybody in 7½ years shake their fist at me and say, Howard, I'm angry with you for letting the value of my house [sic] increase."

3.2 The facts

While the developer of a new estate may provide considerable infrastructure within the estate — e.g. roads, footpaths, drains, sewers, power lines, water mains, and gas mains — governments or their corporatized utilities still need to provide the headworks that connect the internal infrastructure to the outside systems, plus any incidental works to ensure that the systems have sufficient capacity to handle the new estate. The headworks and incidental works are the "infrastructure" that the development levies are intended to pay for.

In opposing these levies, the developers fail to tell us that the provision of infrastructure, by itself, increases the value of the serviced land. So even if the infrastructure were funded out of general taxes, first home buyers would still pay for it in the prices of housing lots and house-land packages. The difference is that in the absence of the development levies, more of the uplift in land values would accrue to the developers — or, if the developers do not yet own the land, to those who are waiting to sell the land to developers (which may tell us where the rest of the opposition to the levies is coming from).

3.3 The SWT as an infrastructure levy

If development levies raise land prices, they do so only by causing a bottleneck in the supply of developed land — by delaying development (or at least the resale of developed lots) until such time as the developers can recover the levies. This sort of bottleneck cannot be prevented by funding infrastructure out of general taxes, because that would attach a fiscal cost to every development application so that governments would be less likely to approve development. But it can be loosened by replacing the development levies with an SWT on the residential lots resold by the developer, provided that the cost of providing the internal infrastructure and acquiring the land between the lots (if that land is to be transferred to the public domain as a condition of development) is allowed as a deduction against the SWT base; the SWT will then tax the unearned gains in site values due to the headworks, incidental works, and approval of development, but will not penalize the actual work of development, and cannot cause an otherwise profitable development to become unprofitable. Admittedly, a component of the uplift in lot values will be due to the infrastructure provided by the developer between the lots, and the SWT will unfortunately take a fraction of that component; but this is a minor imperfection by comparison with existing levies which bear absolutely no relation to uplifts in site values, including those due to public works, and which can cause otherwise profitable developments to become unprofitable.

The market cannot value the benefit of infrastructure except through the price of access to the infrastructure: market value equals price of access. But the price of access has two components: the obvious one, namely the charges (fares, tolls, etc.) payable for actual use of the infrastructure; and the hidden one, namely the price of living or doing business in a location when the service provided by the infrastructure is available, as opposed to a location where it is not available. "Location, location..."

The "hidden" component of the price of access to infrastructure is none other than the uplift in site values caused by provision of the infrastructure. Moreover, the benefit of the infrastructure to the public (as opposed to the provider) is net of charges for actual use; that is, it is equal to the "hidden" component of the price of access:

It follows that the cost/benefit ratio of an infrastructure project is simply the cost/uplift ratio, which in turn is the fraction of the uplift that must be recovered through the tax system in order to pay for the project. If the project passes a cost-benefit test, this fraction is less than 100%. If the necessary fraction is reclaimed through the SWT, the project is funded.

More generally, if a certain fraction of every uplift is reclaimed through the tax system (e.g. via the SWT), infrastructure projects whose cost/benefit ratios are equal to that fraction will be self-funding, while projects with lower cost/benefit ratios will be more than self-funding, yielding a net contribution to revenue which may be used for, e.g., tax cuts, while the remainder of the uplifts accrues to the property owners, ensuring that they gain in absolute terms in spite of any associated improvement in housing affordability.

In short, a tax based on unearned uplifts in site values can finance any infrastructure that is worth building. We have already seen that such a tax has the substance of a holding tax even if it is payable on transfers, and that holding taxes are preferable to transaction taxes for the purpose of affordability. We may therefore state as a guiding principle that:

The SWT taxes all unearned increases in site values howsoever caused. These include not only increases due to market trends or infrastructure, but also increases due to rezoning, which some jurisdictions (e.g. the ACT in Australia) tax by means of so-called betterment levies. So we modify our SWT proposal as follows:

As the taxes to be replaced by the SWT are imposed by local and State governments, both of which have some responsibility for infrastructure, it would be appropriate to:

3.4 Avoiding "retrospectivity"

If every property sold after a certain day ("D-day") pays SWT on the real increase in the site value since acquisition, even if acquisition occurred before D-day, then the continuing turnover in the property market will cause a steady stream of SWT revenue to start immediately, allowing the immediate abolition of existing property-transfer taxes. That turnover will also ensure that uplifts due to infrastructure projects are immediately reflected as increases in revenue. But vendors who have received large uplifts before D-day will pay more tax on those uplifts than they would have paid under the old system, and might therefore allege that the SWT is retrospective. Accordingly we propose as follows:

Under this option, any property vendor who pays more tax than would have been payable under a continuation of the old system does so solely because the site has increased in value after D-day. The exercise of this option does not affect the financing of infrastructure, because uplifts caused by infrastructure after D-day are still immediately reflected in higher SWT receipts through normal market turnover.

4. Taming the NIMBYists

Not-In-My-Back-Yard campaigns by established property owners against proposed housing developments restrict the supply of housing and inflate prices and rents — and may indeed be intended to do precisely that. To the extent that such campaigns are motivated by fear of property devaluations caused by noise, traffic, loss of views, increased supply of accommodation, or (in the case of more affordable housing) prejudice against the kinds of neighbours that the new housing will attract, the motive can be neutralized by promising compensation for any devaluations. If the SWT is allowed to be negative, it gives partial compensation. Because the overall trend in site values is upward, a 100% SWT rate on negative windfalls could be funded by a less-than-100% rate on positive windfalls, giving complete compensation for devaluations. But even without that ultimate refinement, the SWT would reduce the ferocity of NIMBY campaigns.

5. De facto building taxes

5.1 Untaxing buildings

As noted above, municipal rates levied on improved values discourage building and damage affordability. The same is true of all other recurrent taxes that tend to increase with the quantity or quality of accommodation, whether such taxes are called fire levies, ambulance levies, garbage collection levies, water access levies (as distinct from consumption charges), sewerage access levies, toilet pedestal taxes, or by any other names. Some such taxes are imposed by local governments, and some by State governments. All amount to de facto holding taxes on buildings.

The supply of buildings, and therefore the affordability of accommodation, would be increased if all these taxes were shifted onto site values. Hence the following proposal:

The method of calculating the threshold ensures that there are no losers in the transition from the old system to the new. This is obviously crucial for political feasibility.

If the ILT were allowed to be negative, it would provide partial compensation (in the form of a periodic payment) to owners of devalued sites even while they still hold those sites. This would be a further defence against NIMBYists.

5.2 No more rate bills for home owners?

For most taxpayers, income taxes are administered on a pay-as-you-earn basis, and consumption taxes on a pay-as-you-spend basis. In contrast, recurrent property taxes tend to appear as periodic bills, so the cash with which to pay them is not automatically available. The need to set aside cash for the tax bill is especially likely to annoy residential owner occupants, for whom the taxable property value is not realized as an obvious cash flow.

To avoid this political liability in the case of the ILT, we propose as follows:

The deferred ILT, together with the SWT, would greatly improve the affordability of homes for first-time buyers. If owner-occupants can sell their old homes without paying any tax, they can spend the entire proceeds — including unearned capital gains — on new homes, and thereby outbid first-time buyers who have no capital gains to spend. A transfer-tax liability for sellers would automatically reduce the imbalance without the use of special grants or concessions for first-time buyers.

6. Empirical evidence

6.1 Methodology

The effects of taxation on housing affordability cannot be demonstrated simply by correlating nominal prices or rents with tax policy, because affordability does not simply mean low prices or rents. Indeed, if the same policies that enhance the competitive positions of renters and buyers also encourage the provision of infrastructure, which raises prices and rents, affordability will be paradoxically correlated with higher nominal prices and rents. Even if one could devise a reliable measure of affordability, correlating that measure with tax policy would still be problematic because of all the other influences on affordability, such as geography (is the growth of the city constrained by natural barriers?), planning policies (is land readily released for housing or rezoned for higher-density housing?), demographics (is the growth of a particular industry attracting a flood of high-income immigrants?), etc.

There are two obvious methods of "controlling" for some of the non-tax influences on affordability. One is to study a single locality over a long period during which tax policy has varied. If a large number of such studies yield qualitatively similar correlations, a firm conclusion can be drawn. The other method is to study one city within which otherwise comparable localities have different tax policies at the same time. In a large city with many local governments, it may be possible to divide the local government areas into bands and correlate affordability with tax policy within each band. One may be able to repeat the analysis for different cities and/or for different historical snapshots of the same city. Again, if the results of the various studies are qualitatively consistent, a firm conclusion can be drawn.

But what statistic should be used as a proxy for affordability? We have already eliminated raw prices and rents, capacity to pay, utility, and returns to owners. As some of these concepts are vague, a combination of them would be correspondingly vague.

The incidence of homelessness is unsuitable because it is especially prone to non-tax influences. If one jurisdiction criminalizes street-dwellers as "vagrants" (or worse), it will drive its homeless people into other jurisdictions and thereby make its own rate of homelessness look good, without doing anything about affordability of housing. If another jurisdiction (or some institution within it) treats the homeless more kindly, it may attract homeless people from other places and thereby make its own housing policies look worse than they are.

The rate of construction of housing is a more promising indicator. Construction of new housing adds to the supply of housing and therefore, of itself, improves affordability. Statistics on housing construction are readily available. A problem with this indicator is that land prices are subject to speculative bubbles, during which construction also peaks as speculators try (in vain) to get reasonable return on the exorbitant prices paid for sites. So, before drawing any conclusions from variations in building activity, one must be sure that variations over time are not fully explainable in terms of the boom-bust cycle, and that variations from place to place are not explainable as locational variations in the timing of the cycle. The latter condition is almost automatically satisfied by studies confined to a single city.

In most debates over taxes and subsidies, it would be considered self-evident that production of a commodity can be increased by taxing it less or subsidizing it more. The SWT and the ILT would shift a tax burden off buildings, while the New Home Builders' Grant would shift a subsidy onto buildings. It should therefore be considered self-evident that these reforms would encourage building and therefore enhance affordability of housing. That the same reforms would shift a tax burden onto sites and a subsidy off sites does not invalidate the conclusion, because sites cannot be produced by taxpayers but can be pushed onto the market by holding taxes. In short, the shifting of a tax burden — especially a holding tax — off buildings and onto sites should be expected to encourage construction.

This conclusion is supported by numerous studies into the effects of changes in municipal rating systems. Local property taxes ("rates") may be levied on the site value (SV), or the combined value of the site and improvements (capital-improved value or CIV), or the annual equivalent of that combined value (net annual value or NAV). Also possible is a weighted sum or average of SV and CIV or of SV and NAV (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). Theory suggests, and the studies confirm, that SV rating leads to more building than a composite or two-rate system, which in turn leads to more building than CIV or NAV rating.

6.2 Australian experience

One convenient laboratory for such studies was the metropolitan area of Melbourne, where, until the system was changed by the State government during the premiership of Jeff Kennett, each municipality chose its rating system by a popular vote. It was in Melbourne that A.R. Hutchinson, founding director of the Land Values Research Group [4], pioneered the technique of grouping the municipalities into zones and correlating tax policy with building activity within each zone. His results were published in Melbourne in 1945, but also attracted attention in the USA [5].

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).

Taking advantage of the fact that some Australian States mandated SV rating while others did not, Hutchinson also made interstate comparisons of construction rates and ratios of improvement values to land values, noting that both measures were superior in the SV States. He also noted that there was a net interstate migration into SV States, indicating that prospects in those States were perceived as superior. These interstate comparisons are more susceptible to "extraneous" influences than intra-zone comparisons in a single city; but, for what they are worth, they further support the thesis that SV rating enhances housing affordability.

Melbourne also provided opportunities for longitudinal studies as the voters of individual municipalities switched to or from SV rating ("to" being more common). The following may be cited as an example:

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 [7].

6.3 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 [8]. 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.

Conclusion

Because of the correlation between utility or spending power on one hand, and rents or prices on the other, affordability of housing is to be understood not in terms of any of these variables, but rather in terms of the competitive advantage of renters and buyers relative to lessors and sellers.

Affordable housing is a prerequisite not only for reducing homelessness and housing stress, but also for reducing unemployment.

But if a policy on affordable housing is to be politically feasible, it must avoid creating a class of losers.

Applying these principles to Australian conditions leads to the following prescriptions:

Notes

[1] Speech reported by The Times and reprinted in Liberalism and the Social Problem (Hodder & Stoughton, 1909; http://pge.rastko.net/dirs/1/8/4/1/18419/18419-h/18419-h.htm) and The People's Rights (Hodder & Stoughton, 1910).

[2] Technically, 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. The unimproved value, which is used for rating purposes in Queensland, attempts to exclude merged improvements. But the difference between the site value and the unimproved value is usually small.

[3] Doorstop interview at Granville, NSW, August 21, 2006; http://www.patv.com.au/Polityka-Bajkowski%20Eugeniusz-10.htm.

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

[5] Harry Gunnison Brown, "The Challenge of Australian Tax Policy", American Journal of Economics and Sociology, vol.8, pp. 377–400 (July 1949), Part I, reprinted in Selected Articles by Harry Gunnison Brown (New York: Robert Schalkenbach Foundation, 1980), pp. 158–163.

[6] Harry Gunnison Brown (with E.R. Brown), "Incentive Taxation in Australia", American Journal of Economics and Sociology, vol.26, p.416 (October 1967), reprinted in Selected Articles by Harry Gunnison Brown (New York: Robert Schalkenbach Foundation, 1980), pp. 227–228.

[7] A.R. Hutchinson et al., Victorian Local Government Rating Study (Land Values Research Group, 1979), quoted in Fred Harrison, The Power in the Land (New York: Universe Books, 1983), pp. 195–196.

[8] 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.

Revision history

Version 2006.09.12 (considered an incomplete draft) was the original.

Version 2006.11.09:

Added subsections 3.4 & 5.2 and section 6;

Expanded the conclusion;

Made numerous additions and clarifications elsewhere in the text.

Version 2006.11.24 corrected an inconsistency concerning the interest rate for deferment of the ILT. The rate is "nominal".

Version 2007.05.27 corrected a serious single-word textual error: In the bold paragraph in subsection 5.1, "site value tax exceeds" was replaced by "site value exceeds".

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

Housing affordability:
What it is, what it isn't, and how to achieve it

Contents

1. What is "affordability"?

1.1 An opportunity for mischief

1.2 Mostly not capacity to pay