How to Pay for Public Transport

By Dr. Gavin R. Putland (revised June 12, 2003)

In response to A Vision of Sustainable Transport in Queensland (Queensland Greens, February 2003)

Summary

Any new public transport service increases property values (or, more precisely, land values) along the route. The total increase in land values is consistently much larger than the fixed cost of providing the service. If the tax system were reformed so as to claw back an appropriate fraction of all increases in land values, every desirable public transport development would automatically pay for itself in fiscal terms — while still delivering a windfall to property owners. Under the proposed reforms, all sensible public transport proposals that are currently stalled for want of funding would be able to proceed, delivering benefits to property owners, the travelling public, and the environment.

1. The problem

The marginal cost of public transport is the cost of carrying an extra passenger. If the price of a ticket covers the marginal cost plus a reasonable profit margin, then the service provider will find it attractive to carry more passengers. However, the marginal cost is not the only cost to be covered.

It costs money to build a railway even if you don’t provide any trains. And it costs money to provide trains even if they never run. And it costs money to run a train even if it carries no passengers. These are examples of fixed costs. It is usually impossible to cover all the fixed costs from ticket sales. If you raise the ticket price above the marginal cost, each ticket defrays some of the fixed costs, but you sell fewer tickets. If you keep raising the ticket price, you eventually reach a point where the loss of revenue caused by decreased sales outweighs the gain in revenue caused by higher prices; and this point is usually reached long before the fixed costs are fully covered.

Hence, if public transport is to be viable in microeconomic terms (i.e. from the viewpoint of the provider), it needs another source of income as a supplement to ticket prices. If the supplement comes from conventional taxation, it has two problems:

It is inequitable; taxpayers who have no access to public transport still have to pay for it, while taxpayers in poor areas pay for transport services in rich areas.
It is inefficient; conventional taxes stifle the economy by penalizing work, employment, saving, investment and consumption.

These problems discourage governments from providing public transport and restrict the subsidies available to service providers, forcing them to raise ticket prices above the marginal cost to the level that minimizes losses, which level leads to low ticket sales. The result is a public transport system that is too thinly spread, too expensive and too infrequent to attract the travelling public away from their cars. The public then demand more expenditure on roads, which further squeezes the funding available for public transport — completing the vicious cycle.

(Note: It is arguable that the cost of running an extra train or bus should be regarded as a marginal cost rather than a fixed cost. But, wherever we draw the line between fixed and marginal costs, the essential conclusion is the same: we cannot cover all the fixed costs from ticket prices.)

2. Snake oil: Public-Private Partnerships (PPPs)

It is fashionable nowadays to claim that the problem of financing public infrastructure, such as transport, can be solved by handing it over to the private sector.

Of course the private sector has always been involved to some extent. Traditionally, when a government department wants to build something big, it calls tenders for the major components of the project, and the successful tenderers are usually from the private sector; but the department retains overall responsibility for the project and finances it from borrowed or saved funds. The new fashion, however, is to hand over both the overall management and the financing to a principal private contractor. This arrangement is called a public-private partnership (PPP), where “public” apparently describes the source of subsidies and other special favours conferred on the private partner.

The basic problem of public transport is that the fixed costs of provision cannot be recovered through ticket prices. This problem cannot be solved by privatizing it unless the private contractor can build the system at lower cost and/or inspire the public to pay higher ticket prices. Even if this condition is met, the advantage of the private contractor is only one of degree, not one of essence. But in fact the condition is not met. Here are some of the reasons:

1. PPPs increase costs, including:

Interest: As governments can levy taxes to repay their debts, it is safer to lend to the public sector than to the private sector. So the public sector can borrow at lower interest. To the extent that PPPs replace public borrowing by private borrowing, they increase interest costs.
Dividends: Private firms, unlike government departments, need to make profits and pay dividends to their shareholders.
Set-up costs: Under PPPs, the big accounting firms rake in huge sums in consultancy fees, both from governments seeking advice on the feasibility of PPP projects and from corporations bidding for principal contracts (and can anyone not see the conflict of interest?!).
Social costs: Experience shows that the main “savings” made by private principal contractors come in the form of inferior wages and conditions for their own employees and inferior contractual protection of wages and conditions of subcontractors’ employees. This expedient does not reduce costs, but merely shifts costs onto the employees and their families and the social security system (i.e. the taxpayer).

2. A PPP-provided transport system is likely to be inferior in design and build, and hence less attractive to passengers, than a publicly provided system, because of:

Moral hazard: For a government department, providing substandard public transport is bad politics; but for a private firm, it may be a safe way to cut costs and boost shareholder value — especially if the taxpayers have promised to bail you out if you can’t sell enough tickets to stay viable.
Technical errors: Because the PPP concept is new, private principal contractors have less experience in infrastructure provision than the responsible government departments and are therefore more likely to make mistakes. (Of course, some of these mistakes could also be listed under higher costs.)
Market constraints: While a private firm need not worry about a political backlash against high ticket prices, it is no more immune to market forces than governments are: if tickets are too expensive, passengers may use other modes of public transport, or use cars, or stay home. (And even if a private firm can raise ticket prices without delivering a better service, this does not represent a gain to the travelling public.)
Short time horizons: Under the best-known form of PPP, namely the build-own-operate-transfer (BOOT) contract, the principal contractor has a limited time (e.g. 25 years) in which to recover its costs and make its profits before handing over the system to the public sector (the “transfer”). This handicap increases the required cost savings and/or increases the required ticket prices. Moreover, the transfer requirement is an admission that it is desirable to have the system publicly owned and operated!
Commercial errors: Because the history of PPPs is short by comparison with the service lives of PPP projects, and because innovation in PPPs continues, the financial details of PPPs are not adequately tested and may fail in unexpected ways, causing cost blowouts. At best, such failures will be caused by honest mistakes; at worst, they will be caused by misleading and deceptive conduct.

6. Alternative competitive models: While it is true that competition in the private sector encourages efficiency and reduces costs, the benefit of that competition can be obtained through the traditional tendering process without resorting to PPPs.

For governments, one attraction of PPPs is the avoidance of public debt. However, when a government privatizes an infrastructure project, it forgoes the income (e.g. from ticket sales) that it could earn by keeping the project public, and must subsidize the project if it is not microeconomically viable. The expenses involved in the subsidy (if any) and forgone income are equivalent to servicing a debt; and for the reasons given above, the equivalent debt for the PPP option is greater than the debt for the traditional public financing option. But this equivalence is not recognized by accrual-based accounting methods, in which a debt is counted as a negative contribution to net worth, while the present value of a future series of outgoing payments is not (see Appendix A on the present value of an annuity). So the technical avoidance of debt allows the government to pretend that its financial position is better than it is, and that the taxpayers are getting something for nothing when in fact they are losing.

Governments need to be consistent in their attitude to debt. If the limits on government borrowing are too low, they should be increased; if not, they should not be circumvented by financial arrangements that are equivalent to borrowing but called by other names.

Of course, the fact that PPPs increase costs means that private firms — including contractors, consultants and banks — stand to gain at public expense if PPPs are more widely adopted. In view of the dependence of political parties on corporate donations, this increases the pressure on governments to adopt a “PPP or nothing” policy. When this policy trickles down to the public servants who must make recommendations concerning infrastructure projects, they realize that they will not get any project approved unless they “adjust the assumptions” (the non-criminal method of cooking figures) to make the PPP financing option come out on top. So the myth is perpetuated.

3. The problem in terms of benefits

When we say that the fixed costs of public transport cannot be recovered through ticket prices, we are stating the problem purely in terms of costs. To understand why the costs cannot be recovered from ticket buyers, we have to consider not only costs but also benefits. The obvious benefits of public transport include:

reduced road congestion and associated delays,
reduced road trauma,
reduced pollution,
reduced energy consumption,
reduced road construction/maintenance costs,
reduced car acquisition/maintenance costs, and
reduced space consumption by carports, driveways and car parks.

If these benefits were enjoyed fully and exclusively by the users of public transport in proportion to their use, then the users would be willing to pay the full market value of these benefits in ticket prices. But in fact, as we shall see, some benefit accrue to non-users, so that users cannot be expected to pay for them at all, while other benefits accrue to occasional users out of proportion to their use, so that regular users cannot be expected to pay through ticket purchases (which are proportional to use).

In economic jargon, the problem is that public transport produces positive externalities, i.e. benefits enjoyed by people who don’t pay. The solution, then, is to find a way to make the beneficiaries pay, i.e. to internalize the externalities.

Two recent policy documents produced by the Queensland Greens (references [ 2] and [ 3]) acknowledge many economic benefits of public transport, but signally fail to offer a coherent strategy for internalizing those benefits — i.e. for turning them into cash flows to pay for more public transport. The main funding proposal in these documents, namely the reallocation of most of the road budget to public transport, addresses only one of the seven benefits on the above list. The suggestion that reduced ticket prices will not lead to reduced revenue (because of increased patronage) is budget-neutral. The more recent document [ 3] suggests that when residential developments are approved outside public transport spine areas, the approvals should stipulate that the developers provide linking public transport at their own cost. This proposal is financially sound, being based on the unstated premise that approvals for development increase the values of developers’ land holdings. But it would tend to increase the numbers of transport providers and interchange points, leading to difficulties of coordination, and would do nothing to improve services in areas that are already developed. It also requires a separate deal to be struck for every service so provided, raising the possibility of inconsistency or even corruption. What is needed is a system that is equally financially sound, but which automatically covers the cost of any sensible extension or improvement of public transport.

4. Solution: Land Value Taxation (LVT)

Reducing road funding to pay for public transport reclaims one of the benefits on the above list (“reduced road construction/maintenance costs”). Let us consider the other benefits:

If public transport reduces the congestion and delays on the roads servicing a particular area, that area becomes a more desirable place to live, and people will pay more to live there — even if they commute by car.
Because travel by public transport is safer than travel by car, safety-conscious commuters will pay more to live in an area with public transport than in an area without, other things being equal.
If a city produces less pollution, it becomes a more desirable place to live and increases the price premium that people will pay to live in that city, as opposed to a rural area. Again, this benefit is shared by those who don’t actually use the public transport. (Reduced energy consumption is of course correlated with reduced pollution.)
If public transport allows residents in a particular area to get by without cars, or with one car per household instead of two, the incidental costs of living in that area are reduced, so that people will pay more for just being there. These benefits do not accrue to households in strict proportion to their use of public transport; for example, even occasional use may eliminate the need for a second car.
If public transport allows commercial property owners in a particular area to devote less space to car parks and associated driveways, the owners can devote a higher percentage of their space to business, and investors will be willing to pay more for that space. This logic applies regardless of whether the investors themselves use the public transport.

Now, to live or do business in a particular area, you need access to a building (or a share thereof) plus a site, i.e. a piece of land (or a share thereof). Because of competition between new and existing buildings, the price of a building cannot rise much above the depreciated replacement cost, while the rent of a building cannot rise much above the interest on that cost. Nor can the price and rent of a building fall far below these respective levels (unless the building is so badly located as to impair its intended use). So, if public transport drives people to pay more for the privilege of living or doing business in a particular area, they pay more for access to the land, not for access to buildings. That is:

* Most benefits of public transport are reflected in increased land values.

The increases in land values obviously cannot be captured in ticket prices, because people do not buy land with their tickets. But increased land values are certainly realized in cash whenever land is sold or let or used for business. These cash flows can be captured for public purposes through the mechanism of land value taxation (LVT).

Under an LVT régime, the owner of each site annually pays a certain (small) percentage of the land value into the public treasury. That percentage is called the LVT rate. The gross rental value of a site is the market rent that a landlord could charge for the site. In order to receive this rent, the landlord must own the site and pay the LVT, so that the LVT is an expense incurred in obtaining the rent. Hence we subtract the LVT from the gross rental value to obtain the net rental value of the site; in this essay, the word net means “net of LVT”. Using these definitions and a bit of high-school mathematics (see Appendix A for the details), we can establish the following:

* If the value of a site increases by a certain fraction while the LVT rate remains constant, then

the annual LVT liability increases by the same fraction, and
the gross rental value increases by the same fraction, and
the net rental value increases by the same fraction.

In other words, if the value of a site increases while the LVT rate remains constant, the owner always gets an unearned windfall in spite of the tax! But the public treasury also gets a contribution. This is reasonable because the unearned windfall increases the owner’s capacity to pay tax. It is even more reasonable if the unearned windfall is due to a public project funded by the treasury: “Render to Caesar the things that are Caesar’s.”

The total increase in land values caused by a desirable public transport project is consistently greater than the fixed cost of the project; indeed, because the benefits of public transport are largely realized as increased land values, we may regard the ratio of the total increase in land values to the fixed cost as the figure of merit of the project. So, by capturing a sufficient fraction of the increased land values through LVT, we can obtain enough revenue to cover the fixed cost. One minor complication is that the revenue from LVT is an annuity, whereas most of the fixed cost of the project usually appears up-front as a capital cost. But the capital cost is easily converted to an annuity by borrowing. Contrary to the current PPP dogma, financing an investment by borrowing is perfectly rational and responsible when the investment yields more than enough revenue to repay the loan.

Furthermore, ignoring the connection between public infrastructure and land values, and considering LVT simply as a general-purpose tax, we must conclude that LVT is an exceptionally convenient, efficient and equitable method of raising revenue. It is convenient because it has negligible compliance cost. It is efficient because the value of land is not due to any activity of the owner, so that taxing the value in the hands of the owner cannot discourage any economic activity. It is equitable because the value of land owned by a household tends to increase much more than proportionally with household income — so that, for example, a flat-rate LVT is more progressive than any practical “progressive” income tax — and concessions are easily devised to deal with exceptional households that are asset-rich but income-poor.

The method by which land values are calculated for the purpose of taxation is described in Appendix B. Some minor details of definition are relegated to Appendix C.

5. Funding of other infrastructure

While the present policy discussion is concerned with public transport, we should note that public transport is not the only kind of infrastructure that can pay for itself through increases in land values.

For example, the value of land in a new suburb is enhanced when the suburb is provided with a state school. Similarly, any improvement to an existing state school enhances land values in its student catchment area. A British study published in 2001 found that being in the catchment area of a desirable state school could increase the value of a home by up to 19 percent, and that the additional mortgage cost of moving into such a catchment area — i.e. the price charged by the incumbent landowners for admission to the desired “free” state school — was typically 700 to 1400 pounds a year [ 1, p.14].

6. Friendly fire: Property owners’ associations and LVT

People of privilege will always risk their complete destruction rather than surrender any material part of their advantage.

— John Kenneth Galbraith.

Property owners and their associations are vociferous in complaining about land value taxes, opposing any suggestion that LVT rates be increased, and applauding any political party (e.g. the Liberal Party in Queensland) that professes a long-term goal of doing away with LVT. Superficially their position makes sense; it seems obvious that the owners of any asset class stand to lose if that asset class is taxed. It is only when one considered the connection between land values and infrastructure (such as public transport) that the folly of their position is exposed.

If LVT rates are sufficiently high, the increased annual taxes collected on the increased land values caused by desirable public infrastructure projects are enough to pay for the projects. So the projects go ahead, and the landowners reap unearned windfalls in the form of increased sale values and increased after-tax rental values (see Appendix A again — and note that the landowners gain from increased land values no matter how high the tax rate may be). The only owners who might suffer from the increased annual tax are those who do not realize the rental value of their land as cash flows (e.g. residential owner-occupants). Even these owners have the opportunity to be winners, because they can sell up and move out and pocket the capital gains. Nevertheless, for political reasons it is inevitable that such owners would be offered special concessions, such as the option of allowing unpaid LVT to accumulate as a debt against the property. But this is a concession to sentiment; from a purely financial viewpoint, all landowners whose LVT bills increase are winners, because the rental value of their land after tax also increases.

But if LVT rates are not high enough to recoup the costs of public infrastructure projects, those projects face fiscal barriers and may therefore not proceed, in which case landowners miss out on the associated windfalls. Admittedly, in the presence of LVT, the landowners only get a fraction of the increases in gross rental values; but a fraction of something is better than 100 percent of nothing!

Nor is this all. While a desirable public project increases the values of most affected sites, there will usually be a few sites that decline in value — e.g. business sites that are bypassed, or residential sites that lose a desirable view or are affected by noise. The owners of such properties are said to suffer wipeouts instead of windfalls. Under an LVT system, these owners receive partial compensation in the form of reduced tax assessments. The higher the LVT rate, the greater the the compensation for wipeouts, and the lower the risk of a successful NIMBY (Not In My Back Yard!) campaign against the project.

NIMBY campaigns would be even less likely if landowners had a legal right to full compensation (not merely partial compensation) for wipeouts. Under the present tax system, which underutilizes LVT, infrastructure projects do not generate enough new revenue to cover fixed costs, so that, if full compensation for wipeouts were made mandatory, it would constitute yet another fiscal barrier to development. But if LVT rates were sufficiently high, desirable projects would generate enough additional revenue, via increases in land values, to cover fixed costs plus full compensation for the few landowners who suffered wipeouts.

In summary, if LVT rates are too low, landowners suffer because projects that would enrich landowners languish on the shelf for want of funding and for fear of a NIMBYist backlash.

The above arguments apply to changing land values at a constant (high or low) LVT rate. But even if the LVT rate is increased, landowners do not suffer provided that the increase in LVT is balanced by cuts in transaction taxes, because those tax cuts increase the rental value of land (see Appendix A again). So not only do landowners have nothing to fear from high LVT rates; they also have nothing to fear from the transition to those rates.

7. LVT and rail freight

The Queensland Greens policy documents [ 2, 3] complain that rail freight lines have been neglected in favour of environmentally destructive road freight, and suggest that rail freight carriage could be encouraged by allowing door-to-door access via private rail sidings [ 3].

If rail freight charges were kept sufficiently low, the revival of freight lines passing through industrial and commercial land would increase the value of that land. The values of land holdings adjacent to the lines would be further increased if permission were given to build private sidings. These windfalls could be partially recaptured through LVT to cover the fixed costs of the rail freight services, so that freight charges would only need to cover marginal costs.

8. LVT and urban sprawl

The Queensland Greens policy documents [ 2, 3] seem to imply that the car is the only cause of urban sprawl. Reference [ 2] goes so far as to say “The emphasis on the car in the town planning process since the 1960s has been solely responsible for the massive sprawl that all Australian cities are now facing.”

Certainly the car is a major contributor to sprawl, with its demands for storage space, parking space and moving space. But to blame the car alone is a gross exaggeration, because sprawl is caused not only by non-optimal use of land, but also by complete non-use of large amounts of land inside urban areas. In 2000, two researchers from the Brookings Institution studied land waste in 70 American cities with populations over 100,000. They found that, on average, about 15 percent of urban land was vacant, while typically there were about 3 abandoned structures per 1000 inhabitants [ 1, pp.83-4]. Some cities were spectacularly worse than average; e.g., 43 percent of land in Phoenix was vacant, while Philadelphia reported 36.5 abandoned structures per 1000 inhabitants. Considering that we will always need roads and parking spaces (even if only for buses!), one is inclined to think that optimal use of presently unused urban land would reduce sprawl as much as any feasible reduction in car use.

(Note: While it may seem that unused land is environmentally desirable, unused urban land is not likely to be in its natural state — indeed, as the above figures show, it may even be littered with abandoned buildings — and even if it is in its natural state, the environmental benefit is vitiated by isolation from other virgin land. Occasionally a piece of land may be of such unique ecological value that it must be preserved even if it is surrounded by urban development; but usually it is better to keep the city more compact and save more virgin land outside.)

Unused urban land is a symptom of inadequate LVT rates. If the annual interest and LVT payable after the purchase of a piece of land are together less than the annual real increase in the land value, then a speculator can buy the land and sell it for a capital gain without having to earn any income from the land in the mean time. By keeping the land vacant, or failing to use any existing structure on the land, the speculator can retain maximum freedom to sell the land for any desired use at the most opportune time. Land speculation constitutes an artificial demand for land and pushes up land prices, forcing genuine land users to seek cheaper land further out of town. This scattering of development is itself a form of sprawl, and causes further sprawl as space-consuming roads must be built to service the remote suburbs.

If the LVT rate is slightly higher, the speculator may not be able to cover the LVT liability from the capital gain alone, but may be able to cover it by some temporary cheap development such as a car park, warehouse, or self-storage centre. Such developments, known in the trade as “cash crops” or “tax payers”, are designed to cover the property tax liability while minimizing capital expenditure.

The higher the LVT rate, the higher the annual cost of holding land, and the greater the pressure on landowners to cover their tax liabilities by putting land to its best possible use — instead of wasting it and causing sprawl. LVT does not discourage governments from reserving land of unique ecological value or setting aside a reasonable amount of land for parks and gardens, because these things make a city a more desirable place to live and consequently increase LVT receipts from other land in the city. But LVT does militate against purely speculative holding of land. In so doing, it improves the competitive position of tenants and intending buyers (see Appendix A).

A secondary cause of sprawl is property taxes levied on the values of land plus buildings, instead of land values alone. Including values of buildings in the tax base is obviously a disincentive to the construction of new buildings and the renovation or extension of existing buildings. While the additional annual tax may not seem very high when expressed as a percentage of the capital investment, it is many times higher when expressed as a percentage of the annualized value of the investment and can easily have a decisive effect on viability.

Minimizing sprawl is therefore a matter of maximizing the taxation of land values and minimizing (preferably eliminating) the taxation of building values.

9. Implementation

In principle, all we have to do is reduce or abolish transaction taxes and make up the lost revenue by increasing LVT rates, until the LVT rates are sufficient to recover the fixed costs of desirable public infrastructure, including public transport, through the resulting increases in land values. In practice, the problem is complicated by the existence of three levels of government: Federal, State and Local. But the complication is minor; for any project, if the total of the LVT rates levied by the respective levels of government is more than sufficient to cover the fixed cost, then the governments will have an incentive to fund the project jointly, with each level of government contributing in proportion to its expected LVT windfall. So let us consider the three level of government separately and ask which taxes could be replaced by LVT.

(a) Local:

Local councils already impose LVT in the form of “rates”. In Queensland and NSW, the value of the land alone is rated. In other States, some councils rate the total value of the land and building(s); these councils should confine the tax to land values alone, and compensate for the lost revenue by increasing the LVT rate.

Local councils also raise revenue from various service charges in addition to “rates”. If these charges are not easily avoidable and not proportional to use of the services, they too should be abolished in favour of higher LVT rates — especially if the availability of the services enhances the value of land.

(b) State:

The States impose some of the worst taxes in the present system, such as payroll tax, stamp duties and debit taxes. These taxes feed into the prices of goods and services and impede transactions, including employment. Thus they increase the cost of living for poor households while reducing their opportunities to help themselves. The States also impose LVT in the form of State land taxes. But these taxes are riddled with multiple thresholds and politically-motivated exemptions which reduce their effectiveness for capturing increases in land value caused by State-funded projects.

What all these taxes have in common, however, is that they are collected from or through businesses that operate on commercial and industrial land. So, given that the current mess of State taxes is to be replaced by a single flat-rate State LVT, it would be logical for that LVT to apply only to commercial and industrial land. The landlords and owner-occupants of such land would pay the LVT, while the tenant businesses would simply forget about State taxes.

What about gambling taxes? At first sight it may seem that these taxes discourage gambling. However, the effect of taxation on the odds is not sufficient to deter problem gamblers, while the availability of gambling is limited by licensing, not taxation. Moreover, while a gambling licence increases the value of the land to which the licence applies, it reduces the total value of surrounding land by a greater margin. (Why? Because gambling is an unproductive business that draws customers away from more productive businesses, reducing overall productivity.) So if the States were totally reliant on land values for revenue, they would have an overall incentive to reduce the number of gambling licences — whereas at present they have an incentive to grant more licences so that they can collect more gambling taxes.

If commercial and industrial land is to be taxed at State level, that leaves residential and agricultural land to be taxed at Federal level. A Federal “tax” on residential and agricultural land is not such bad politics as one might think, because it need not look or feel like a tax.

When we think of Federal taxes, we tend to think of income tax, company tax, GST and excises. However, the most stultifying Federal taxes — and hence the ones most fit to be replaced — are not officially called taxes. They are officially called means tests on social security benefits. But means tests are fully equivalent to taxes, both in terms of their fiscal effects and in terms of the economic disincentives that they cause.

On your income, your effective marginal tax rate (EMTR) is the number of cents that the government claws back from each additional dollar that you earn. You don’t care how many of those cents are clawed back by the Tax Office and how many by Centrelink, and neither does the government; the combined number is all that counts. So an income-tested benefit, abated at a certain marginal rate within a certain income band, is equivalent to a non-income-tested benefit of the same amount plus an income tax surcharge at the same marginal rate within the same income band, payable by recipients of the benefit. Newstart allowance, for example, is clawed back at a maximum rate of 70 cents in the dollar. The result is that a job seeker who finds a bit of casual work faces a higher EMTR than an executive on a six-figure salary.

Similarly, on your assets, your effective marginal rate of taxation is the number of dollars per fortnight that the government claws back for each additional $1000 worth of assets that you own. Again, how the clawback is divided between the Tax Office and Centrelink makes no difference to you or to the government. So an assets-tested benefit, abated at a certain marginal rate within a certain band of asset values, is equivalent to a non-assets-tested benefit of the same amount plus a wealth tax at the same marginal rate within the same range of asset values. A pensioner subject to the assets test loses $3 per fortnight per additional $1000 of assets, which is equivalent to paying interest at a marginal rate of almost 8 percent per year on one’s own money — not exactly an incentive to save for one’s retirement!

LVT is free of such disincentives (see Appendix A). But what all these disincentives have in common is that they affect individuals and families living on residential and agricultural land. So if the present tangle of means tests is to be replaced by a single flat-rate Federal LVT, it would be logical for that LVT to apply only to residential and agricultural land. In the case of owner-occupants, the LVT could be deducted from the total of welfare payments otherwise due to the occupants, and any LVT over and above the total of benefits could be written off; that is, the LVT could become a single means test (SMT) replacing all current means tests. Absentee landlords, of course, should simply receive a bill for the Federal LVT.

(A technicality: It is arguable that if your equity in your home is worth less than the land under it, then your SMT should be based on the equity, not the land value. Otherwise first home buyers would be deterred by sudden large decreases in welfare entitlements.)

10. Instead of a conclusion

Both ground-rents and the ordinary rent of land are a species of revenue which the owner, in many cases, enjoys without any care or attention of his own. Though a part of this revenue should be taken from him in order to defray the expenses of the state, no discouragement will thereby be given to any sort of industry. The annual produce of the land and labour of the society, the real wealth and revenue of the great body of the people, might be the same after such a tax as before. Ground-rents and the ordinary rent of land are, therefore, perhaps the species of revenue which can best bear to have a peculiar tax imposed upon them.

… Ground-rents, so far as they exceed the ordinary rent of land, are altogether owing to the good government of the sovereign, which, by protecting the industry either of the whole people or of the inhabitants of some particular place, enables them to pay so much more than its real value for the ground which they build their houses upon, or to make to its owner so much more than compensation for the loss which he might sustain by this use of it. Nothing can be more reasonable than that a fund which owes its existence to the good government of the state should be taxed peculiarly, or should contribute something more than the greater part of other funds, towards the support of that government.

— Adam Smith (1723-1790),
The Wealth of Nations, V.ii.75,76.

Appendix A: Mathematics of LVT

Using the distributive law, we can easily verify that
(1 – x)(1 + x + x2 + … + xn-1) = 1 – xn.

Rearranging this gives a formula for the sum of a geometric progression:
1 + x + x2 + … + xn-1 = (1 – xn) / (1 – x) . (1)

Using this result, we can find a formula for the present value of an annuity. Consider n annual payments, each of an amount A, with the first payment due at the end of the first year. Let the annual discounting rate, or “interest rate”, be i (e.g. if i=0.05, the discounting rate is 5 percent per annum). Then the first payment is discounted by one year, so its present value is A / (1+i); and the second payment is discounted by two years, so its present value is A / (1+i)2; etc. So the present value of the series of n payments is
P = A/(1+i) + A/(1+i)2 + … + A/(1+i)n
= [A/(1+i)] {1 + 1/(1+i) + … + 1/(1+i)n-1} .

The factor in curly braces is a geometric progression which may be evaluated using Eq.(1), with x = 1/(1+i). Doing this and simplifying, we get
P = (A / i) [1 – 1/(1+i)n ] .

In the case of a perpetuity (i.e. a perpetual annuity), we let n –> infinity, so that the term 1/(1+i)n approaches zero, and we have simply
P = A / i . (2)

This is the present value of a payment of A per year in perpetuity, with the first payment due after 1 year.

Now let’s apply this to land rent and land value taxation (LVT). In Eq.(2), let A be the annual rental value of a piece of land, net of LVT. This is the rent (net of LVT) that a landlord could get for the land, or the advantage (net of LVT) that would accrue to an owner-occupant making optimal use of the land. Then P in Eq.(2) is the present value of the perpetual rent stream, which is the value (selling price) of the land. If we rename this value as V, Eq.(2) becomes
V = A / i . (3)

This V is also called the unimproved capitalized value (UCV) of the land.

Now let E be the gross annual rental value of the land, including LVT. This is the market rent that a tenant would pay for the use of the land, knowing that the actual remittance of LVT is the landlord’s responsibility. Let the annual LVT rate, expressed as a fraction of the UCV, be t. Then the annual tax payable is tV, and the net annual rental value (i.e. net of LVT) is
A = E – tV. (4)

Substituting this into Eq.(3) and solving for V gives the sale value of the land as
V = E / (i+t) . (5)

Substituting for V from Eq.(3) and rearranging gives the net rental value of the land as
A = iE / (i+t) . (6)

The annual LVT payable is T = tV. From Eq.(5), this is
T = tE / (i+t) . (7)

In Eqs. (5) and (6), notice that the sale value V and the net rental value A are decreasing functions of t ; that is, as the LVT rate increases, V and A decrease. However, no value of t can reduce V or A to zero. Moreover, for a given LVT rate, V and A are proportional to the gross rental value E. This leads to our first important conclusion:

* For a given LVT rate — however high it may be — any development that increases the gross rental value of land also increases its sale value and net rental value in the same proportion.

We may therefore speak simply of “developments that increase the value of land” without having to specify which measure of “value” we mean.

Now consider what happens when the value of a piece of land increases because of some fortunate “development” (e.g. a public transport project). Because of the proportionality between the various measures of value, a landowner who can realize the increase in A as a cash flow (e.g. a landlord who charges market rents or a business owner-occupier who uses the land optimally) wins in every way. Another landowner who cannot realize an increase in A (e.g. a residential owner-occupant) does not have any additional cash flow to cover the increased LVT, but can avoid the increase in LVT and pocket the increase in V by selling out and moving to another location. Alternatively, (s)he can allow the increased LVT to accumulate as a debt against the increased sale value until the property is eventually sold or bequeathed. (But if the LVT for this class of owner is collected as a deduction from social security, the owner might not be affected at all, because his/her social security entitlement might already be zero. And if the owner is a charitable concern, such as a hospital, one might reasonably grant an exemption.) All this is summarized in our second important conclusion:

* For a given LVT rate — however high it may be — all property owners gain or have the opportunity to gain from any development that increases the value of their land.

Comment: Because of the proportionality between V and A in Eqs. (5) and (6), the fraction by which the LVT reduces the sale value of the land is the same as the fraction of the gross rental value taken in tax. That fraction is T/E, which may be found from Eq.(7):
T/E = t / (i+t) . (8)

For example, if the LVT rate is equal to the discounting rate (t=i), then the sale value of the land is reduced by half. But no value of t can reduce the sale value by 100 percent. This peculiarity arises because the annual LVT is expressed as a fraction of the “reduced” sale value, i.e. the sale value in the presence of LVT, not what the sale value would be in the absence of LVT.

There is one more complication to be considered. At first, we took the gross rental value E as “given”; for example, Eqs. (5) to (7) express the sale value, net rental value and annual LVT in terms of E. Then we considered “developments” (e.g. public transport projects) as causing increases in E. But another cause of such increases is tax cuts.

Because business transactions take place on land, all taxes on transactions, such as income tax, company tax, GST, payroll tax, stamp duties and debit taxes, are effectively charges for various uses of land. Similarly, because government services are consumed on land, charges for such services are effectively charges for various uses of land. If such taxes and charges were reduced, the respective uses of land would become more profitable, so tenants would offer more rent for such uses, so the gross rental value of the land (E) would increase — by the amount of the reduction of taxes and charges. Hence, if the lost revenue from these taxes and charges were replaced by LVT, the increase in gross rent would cancel the increase in LVT, so that, on average, net rental values would be unchanged and landowners would be no worse off. (Landowners who cannot realize the increased gross rent would receive special treatment as above.)

Moreover, transaction taxes render some transactions unprofitable that would otherwise be profitable, so that the volume of transactions is reduced; this is called the deadweight effect of taxation. Reductions in transaction taxes would reduce the deadweight effect and increase the volume of business that could be done on each piece of land, further increasing its gross rental value. But the LVT replacing the lost revenue from transaction taxes would have no deadweight effect, because LVT is not a transaction tax and cannot be avoided by avoiding transactions; LVT is a tax on the value of the land, which value does not depend on any activity of the owner (taxpayer). So the removal of transaction taxes would cause a net reduction in deadweight, which would fund an increase in the net rental value of the land (over and above the additional LVT replacing the lost transaction taxes). This represents a net gain to landowners.

Thus we have two more conclusions:

LVT can generate sufficient revenue to compensate for any abolitions or reductions of transaction taxes.
Property owners have nothing to fear from any LVT rate increase that simply replaces the revenue lost due to reductions or abolitions of transaction taxes.

In the category of “transaction taxes” we can also include taxes on the values of buildings — including the component of municipal rates that falls on buildings in States other than Queensland and NSW. Property taxes on buildings are transaction taxes in that (a) a tax liability is created by every decision to build, extend or renovate a building, which decision involves a transaction or set of transactions, and (b) the tax liability therefore discourages the decision to build, extend or renovate, and consequently has a deadweight cost. But the same cannot be said of LVT, because (a) the decision to buy land does not create a tax liability, but merely transfers the existing tax liability (reflected in the purchase price) from seller to buyer, and (b) even if the tax were to discourage the purchase, this would not amount to deadweight, because the purchase of land (unlike the purchase of buildings or other products) cannot create a new asset, but merely transfers ownership of an existing one.

N.B.: A reduction in transaction taxes, by itself, enables tenants to pay more rent, and competition between tenants ensures that they do; but an increase in LVT, by itself, does not enable landowners to charge higher rents or prices. Taxes on transactions are passed on in prices because if producers cannot cover these taxes, supply decreases and prices rise. But this mechanism does not work with land, because land has no producers and its supply is fixed. So landowners cannot shift the burden of LVT onto tenants or other customers, but tenants can shift the benefit of transaction tax cuts onto landlords.

Here is some more evidence that the burden of LVT cannot be shifted:

It is said that for every two economists you get three opinions. But virtually all economists agree that LVT cannot be shifted.
Because LVT is a fixed overhead cost, it does not change the prices at which traders maximize their profits (allowing for the effect of prices on sales); if traders could raise profits by raising prices, they would do so with or without the tax. “Traders” include landlords, for whom “prices” mean rents and “sales” mean occupancy rates. So neither can LVT be passed on in rents.
If it were more profitable to do business on one site than on another, the additional demand for the first site would raise its gross rental value so as to cancel the difference in profitability. So the effect of land rents is to equalize the profitability of all unimproved sites, reducing profit margins to those obtainable on marginal sites, i.e. sites with zero rental value. More desirable sites are called supermarginal (or viable) and yield positive rent. Less desirable sites are called submarginal (or unviable). LVT cannot cause supermarginal sites to become submarginal, because marginal sites pay no tax, while supermarginal sites pay only part of their gross rental values in tax. Because traders who own higher-taxed sites must compete with those who own lower-taxed sites, and because the tax does not reduce competition by putting any sites out of business, differences in LVT liabilities cannot be passed on in prices. Again, this argument applies to rents as to other prices.

Warning: Landowners resent LVT because they cannot pass it on in rents and prices. So they campaign against it — by claiming that they can pass it on! But in so doing, they campaign against a tax which would pay for projects that would increase the value of their land while allowing other tax cuts that would further increase the value of their land. So they pay the just penalty for their unenlightened self-interest. The tragedy is that everyone else also pays.

Although this essay emphasizes the benefits of LVT to landowners, the reduction or elimination of the speculative motive (discussed above in connection with urban sprawl) confers a substantial benefit on intending buyers and on tenants. Those who wish to use land face a choice between buying and renting. If one option becomes more expensive, demand shifts away from it and re-balances the prices. Similarly, landowners have a choice between selling their land and letting it to tenants. If one option becomes more attractive, owners become more willing to offer that option, and the shift in supply re-balances the prices. Thus there is a nexus between prices and rents. If LVT discourages speculation, land becomes more affordable to those who wish to buy it for productive purposes (because they no longer have to compete with speculators), and hence more affordable for tenants. However, because cuts in transaction taxes tend to increase land rents, “more affordable” does not mean cheaper in dollar terms; it means cheaper in relation to capacity to pay.

At this point it may be helpful to summarize the effects on land values of replacing transaction taxes with LVT:

(i) The tax liability on the land reduces its after-tax rental value, hence sale value.

(ii) The reduction/elimination of the speculative motive reduces sale values, hence rental values. But, because land speculation is inflationary, the reduction of speculation allows lower interest rates, increasing the sale value for a given rental value while maintaining affordability for intending buyers.

(iii) The abolition/reduction of transaction taxes makes the use of land more profitable, increasing rental values, hence sale values.

(iv) The reduction of deadweight increases the variety of profitable uses of land, increasing rental values, hence sale values.

(v) Public projects that increase land values become self-funding, so that more such projects are completed and more land values are enhanced by this mechanism.

The author claims that (i) and (iii) roughly cancel out, while (iv) and (v) represent substantial gains to property owners. Point (ii) reduces the net gain to landowners in terms of rental values, but not necessarily in terms of sale values, and represents a gain for tenants and intending buyers. Most importantly, (iv) and (v) represent overall gains to the community (city, state or country), so that nobody’s gain needs to be anybody else’s loss: if you make a bigger cake, everyone can have a bigger slice.

Appendix B: Computing Unimproved Capitalized Values (UCVs)

First a note on terminology: improvements to a block of land include buildings, fences and other structures on the land, but not roads or power lines or other services that pass by just outside the land; the latter are regarded as contributing to the locational value of the land, which is part of the unimproved value. Improvements are so-called because they normally add to the economic value of the land as measured by the market.

Now, every time a property is sold or let, we can obtain a spot value for V (the UCV or sale price of the land) as follows:

(a) If the property is sold without improvements, the sale price is V.

(b) If the property is let without improvements, the annual rent is E, and V is found from Eq.(5).

(c) If the property is sold with improvements which are promptly demolished by the purchaser, then V is the total price that the purchaser was willing to pay for the bare site, i.e. the sale price plus the anticipated cost of demolishing the improvements.

(d) If the property is sold with improvements which are used by the purchaser, the depreciated replacement value of the improvements is subtracted from the sale price to obtain V.

(e) If the property is let with improvements, the depreciated replacement value of the improvements is multiplied by the applicable interest rate to produce an annualized value, which is then subtracted from the annual rent to produce E. Then V is found from Eq.(5).

Recent spot values are brought up to date with reference to local trends. Then their accuracy is checked, and unknown values are filled in, by exploiting the requirement of spatial continuity: in the absence of significant boundaries (see note iii below), the unimproved rental value per unit area must be a smoothly-varying function of position.

For agricultural land, it may be impossible to obtain nearby recent spot values. However, such land can be valued from records of stock-carrying capacity, crop yields and operating costs; and the influence of public infrastructure on operating costs is readily estimated.

Notes:

1. Methods (a), (b) and (c) are preferred to (d) and (e), because if the location of a building interferes with its intended use, the value that the building adds to the site can be less than the depreciated replacement value of the building. An extreme example of this is found in method (c), in which the useless “improvements” have a negative effect on the sale price.

2. The equations in Appendix A assume that rent is paid annually in arrears. So in points (d) and (e) above, the “annual rent” must be defined as the future value, after one year, of the full year’s rental payments.

3. For the purpose of spatial continuity, significant boundaries obviously include visible barriers to movement, such as waterways, freeways (between entry/exit points) and railways (between stations). The fact that railways can mark sudden changes in land value per unit area is well known, as attested by the expression “the wrong side of the tracks”. But significant boundaries also include invisible lines such as boundaries of suburbs (some suburbs being more prestigious than others) and boundaries of catchment areas for state schools (some schools having better reputations than others).

Opponents of LVT — e.g. speculators who want to maximize their investment in (appreciating) land and minimize their exposure to (depreciating) improvements — often claim that it is impossible to value land separately from buildings. On this basis, they argue that if land is taxed at all, it should be taxed on the improved capitalized value (ICV), i.e. the value of the land plus improvements. This broadening of the tax base results in a reduction of the rate on unimproved land, which just happens to suit the plans of the speculators.

But consider this: How do you calculate the ICV of a home that has not recently been sold or let? If a comparable home has recently been sold or let nearby, you can use that for a benchmark. If not, you have to estimate the depreciated replacement value of the improvements and then add the value of the land, which is found by interpolation between known values using spatial continuity; in other words, there are cases in which you can’t value the house and land together unless you can value them separately!

The valuation of land, as distinct from buildings and other improvements erected thereon, is a well developed science. For more than a century, local governments in Australia, especially Queensland and NSW, have been charging rates on land values alone. The claim that it can’t be done is self-serving and not to be believed.

Appendix C: UCV vs. Site Value (SV)

In Queensland, rates are levied on the unimproved capital value (UCV). In Victoria, a few councils levy rates on the site value (SV). The SV includes human works that have become merged with the land (e.g. old draining, grading, tree-felling and tree-planting) whereas the UCV excludes them. The SV is considered more objective because, as the years pass and the land changes hands, it may become difficult to distinguish merged improvements from natural features. The “land value” as defined in NSW is closer to the SV than to the UCV.

The capital-improved value (CIV) is synonymous with the improved capitalized value (ICV).

Because both the UCV and the SV exclude the most visible and valuable improvements, such as buildings, the difference between them is usually minor and of interest only to professional valuers. For the purposes of this essay and of wider economic policy, the two terms may be used interchangeably.

References

[1] Don Riley, Taken for a ride: Trains, taxpayers and the treasury (Teddington, England: Centre for Land Policy Studies, 2001). Note: Numerous sources are cited in the footnotes. In particular, the idea that landowners stand to gain from high rates of LVT is apparently due to Prof. William Vickrey (1914-1996).

[2] Rod Milne et al., Breaking Brisbane’s car culture: A public transport vision for Brisbane (Queensland Greens, January 2002), downloadable from www.qld.greens.org.au/info/misc/pub_trans_report.html.

[3] Rod Milne et al., A vision of sustainable transport in Queensland (Queensland Greens, February 2003), from www.qld.greens.org.au/info/misc/pub_trans_report.html.

For the arguments against PPPs, the author is indebted to various speakers at the Queensland Council of Unions Government Revenues and Services Forum, Customs House, Brisbane, April 2, 2003.

The single means test (SMT) is the author’s invention.

Disclaimer / declaration

The opinions expressed in this essay are those of the author and not necessarily those of any organization with which the author is affiliated. The tax/welfare reforms proposed in this essay are aimed specifically at funding of public transport and are not necessarily the reforms that the author would prefer if the terms of reference were wider.

At the time of writing, the author is not a property owner, nor a shareholder, nor a welfare recipient, nor a member of any political party. These circumstances may change.

Permission is given to transmit this essay by email, to republish it on WWW sites, and to print it in non-profit newspapers and periodicals, always provided that the essay remains intact and that this copyright notice is included.