|Wealth and Want
|... because democracy alone is not enough to produce widely shared prosperity.
Case for Taxing Land
note: I've not included the figures here yet.
II. What is Land?
III. The simple efficiency argument for taxing land
IV. Taxing Land is Better Than Neutral
V. Measuring the Economic Gains from Shifting Taxes to Land
VI. The Ethical Case for Taxing Land
VII. Answer to Arguments against Taxing Land
I. Taxing Land as Ethics and Efficiency
There is a case for taxing land based on ethical principles and a case for taxing land based on efficiency principles. As a matter of logic, these two cases are separate. Ethical conclusions follow from ethical premises and efficiency conclusions from efficiency principles. However, it is natural for human minds to conflate the two cases. It is easier to believe that something is good if one knows that it is efficient, and it is easier to see that something is efficient if one believes that it is good. Therefore it is important for a discussion of land taxation to address both question of efficiency and questions of ethics.
This monograph will first address the efficiency case for taxing land, because that is the less controversial case. The efficiency case for taxing land has two main parts. The first and more traditional part is that taxing land permits the reduction of other taxes. This improves efficiency because it reduces the economic distortions that cause people to work less and save less. Because land is fixed in supply, there is no reduction in economic efficiency from taxing it.
The second and less customary part of the efficiency argument for taxing land is that, apart from the opportunity it offers to reduce harmful taxes, taxing land actually improves economic efficiency, in at least three ways.
The ethical case for taxing land is based on two ethical premises:
1) every person has a right to himself or herself, andThe first premise leads to the conclusion that taxing people according to the products of their efforts or the products of their saving can only be just if people voluntarily agree, individually, to be subject to such taxes. Taxing land, on the other hand, does not involve such an intrusion on individual rights. In fact, taxing land is a way equalizing the advantages of access to land, as required by the second premise.
2) all persons have equal rights to the natural opportunities that are not embodied in persons.
The ethical case for taxing land ends with a discussion of the reasons why recognition of the equal rights of all to land may be essential for world peace.
After developing the efficiency argument and the ethical argument for taxing land, I consider a variety of counter-arguments that have been offered against taxing land. For a given level of other taxes, a rise in the rate at which land is taxed causes a fall in the selling price of land. It is sometimes argued that only modest taxes on land are therefore feasible, because as the rate of taxation on land increases and the selling price of land falls, market transactions become increasingly less reliable as indicators of the value of land. The answer to this argument is that collecting a large percentage of the rental value of land through taxes will require new assessment methods based on observing the rental price of land rather than the selling price. I describe how such methods can be devised.
Another basis on which it is argued that greatly increased taxes on land are infeasible is that if land values were to fall precipitously, the financial system would collapse. It is true that many properties have mortgages that would exceed the value of the property if land taxes were increased significantly. This makes it necessary to think carefully about who should absorb the decline in aggregate asset value that would accompany a significant shift toward taxing land. Nevertheless, it is possible to plan for a restructured financial system that would have shed its dependence on land as collateral.
Apart from questions of feasibility, it is sometimes argued that erosion of land values from taxing land would harm economic efficiency, because it would reduce opportunities for entrepreneurs to use land as collateral for loans to finance their ideas. The answer to this argument is that the use of land as collateral encourages banks to lend mindlessly to anyone who has collateral, rather than work to find good lending opportunities. Thus the disappearance of land as collateral would actually improve the efficiency with which financial resources are invested.
Another ethical argument that is made against taxing land is that the return to unusual ability is “rent” just as the return to land is rent. This argument represents a refusal to make a distinction that begs to be made. The first principle of economic justice is that people have rights to themselves. While some scholars have asserted that people have rights to themselves but not to their talents, this is nonsensical. Without talents, there is no self. Talents are fundamentally different from land. The equal rights of all to land can consistently be asserted while still asserting that every person has right to the use of his or her talents.
But before developing any of these arguments, I must discuss what land is.
II. What is Land?
David Ricardo defined land, memorably, as ‘the original and indestructible powers of the soil.’ This definition is overly narrow. It would exclude from land such valuable and destructible things as minerals and topsoil. Ricardo’s definition even suggests that the value of land arising from urban locations might not be included in ‘land,’ since it would not necessarily qualify as a ‘power of the soil.’
The definition of land that is most useful in economic theory is that land is all scarce factors of production other than people and the products of human effort. Land is the gifts of nature. Thus land includes both rural and urban territory, mineral resources, water, fish in oceans and rivers, virgin forests, geosynchronous orbits and the frequency spectrum.
While there is some tendency to think of measuring land in ‘stock’ terms (‘What is the value of that piece of land?’), the flow of services from land is more fundamental. While capital goods have selling prices related to their costs of production, such a calculation does not apply to land since, by definition, land is not produced. The selling price of land is conceived in economic theory as the present value of the net return after taxes to the future flow of land services, when the land is used in the way that maximizes that present value. Following David Ricardo, the value of the flow of services from land is sometimes described as the residual after paying other factors their opportunity costs, when land is used efficiently. But in equilibrium, a similar statement can be made about any other factor.
With some components of land, such as river water flowing into an ocean, the extent to which the land is used in one period has no influence on the extent to which it can be used in other periods. In other cases, such as mineral deposits, use in one period comes at the expense of use in other periods. Agricultural land has a capacity to be ‘mined’ of the nutrients that make it productive, making it somewhat similar to mineral deposits.
With urban land there is a different type of interdependency between use at one time and the potential flow of services at another. To be used most productively, urban land must be combined with durable, immobile capital. Therefore an increase in the intensity of urban land use often requires the destruction of previous capital investments. This intertemporal dependence means that, in principle, a forecast of all future economic conditions is needed to know what use of land is best today. Since people make different forecasts, they reach different conclusions about what use of land is best today. Only with the passage of time, if at all, is it possible to know what use of land would have been best.
The fact that structures are durable and immobile also means that care must be taken in defining the value of the flow of land services. There is a tendency to think of “the rent of land” as the amount of money that land yields to those who have exclusive use of it. However, this formulation is not always useful for defining the rent of land over a particular interval of time. If an investor spends a year and £20 million erecting a building that is expected to last for 30 years, what was the rent of the land under the building during the year of construction? It is not sensible to say that, if the best possible use of the land produces a negative cash flow over a given interval of time, then land has no rental value over that interval. If markets were perfect and the decision to construct the structure was optimal, the finished building would have a value that was greater than its cost of construction by the rent of the land it occupied and the accumulated interest on construction costs and land rent. But if a non-optimal construction decision is made, that does not reduce the rent of land.
To give a meaning to ‘the rent of land’ that does not depend on when construction happens to occur, it is useful to define ‘the rent of land’ as the opportunity cost of leaving vacant land vacant. Thus in the case of the year of construction discussed above, one would ask, ‘Suppose that the construction had been postponed for a year. Perhaps by that time it would be appropriate to invest in a £21 million building rather than a £20 million building. If one developed the most profitable possible plan for the land, subject to the constraint that the land be left vacant for the first year, how much lower would the present value of net returns be?’ The answer to this question, the loss of the present value of net returns from postponing use of land for a year, would be the rent of the land for the year. This is the amount of money that one would need to get in net returns from some pre-existing use to justify postponing redevelopment for a year. Thus the rent of land for any developed site, for any year, is defined to be the answer to the question, ‘If this site were undeveloped, what would be the cost of leaving undeveloped and unused for a year?’
Land also differs from labor and capital in the origin of claims to own it.
Ownership of land is thus a form of privilege. The word
‘privilege’ comes from the Latin prive + legis, meaning ‘private
law’. A private law is a law that has someone’s name it, that is,
a law that authorizes one person to do what others are not permitted to
do. In a just world, there would be no privilege. (Thus no
one is underprivileged.) In a just world, land would not be
‘owned’, but rather ‘held’ or ‘possessed’,
subject to a payment that reflected the value given up by others in allowing
one person to have
exclusive use of a site.
To summarize: If land is subject to a tax or combination of taxes with a present value that is independent of how the land is used, and if potential investors are confident that the magnitude of the tax will not exceed the value of using the land over the life of potential investments, then a tax on land has no excess burden; it is neutral.
IV. Taxing Land is Better Than Neutral
The analysis above explained how a tax on land can be ‘neutral’, that is, how such a tax can have no excess burden. In fact, taxing land is better than neutral; it improves economic efficiency compared to the no-tax situation. This section explains how.
First, a land tax can be used to take account of positive and negative externalities associated with land use. The typical example of a negative externality is pollution. If labor and capital were perfectly mobile, then the effect of pollution on near-by locations would be reflected entirely in land values. Thus one might say that a polluter ‘uses’ surrounding land and should pay for that use in the form of a ‘land tax’ equal to the reduction in the rental value of land that results from his activities. Such a land tax would not only raise government revenue, but would also improve efficiency by ensuring that people would engage in activities that produced pollution only when the benefits of those activities exceeded the costs.
Because labor and capital are not perfectly mobile, pollution affects not only land values, but also the value of structures and the net well-being that people experience from the opportunity to continue to live in accustomed places. Efficient management of pollution would charge polluters for these costs as well.
When there is traffic congestion on city streets or on highways, bridges, or tunnels, one might say that there is a problem caused by people not being charged adequately for the use of particularly valuable space. If there is a congestion charge (land tax) that reflects the extent to which any one user of streets and highways imposes costs on others users in the form of increased travel times, then that congestion charge improves efficiency while also raising government revenue.
What applies to negative externalities applies also to positive externalities, although now some extra steps in the analysis are needed. If there are private activities that raise the rental value of surrounding land, such as the provision of a car park in a city centre, then to motivate private users of land to undertake these activities at efficient levels, one must compensate them for the value of the positive externalities they generate. Again, if labour and capital were perfectly mobile, these externalities would be reflected entirely in changes (this time, positive changes) in land values.
With compensation for the externalities now generating an outflow from the public treasury rather than an inflow, it is not immediately clear that the government can afford to provide efficient incentives. However, if the tax regime is one in which the government collects all of the rental value of land in taxes, then efficient incentives for activities with positive externalities that are reflected in land values will be self-financing. What is spent in rewarding activities with positive externalities will be recaptured from taxes on the land that rises in value. In fact, there is bound to be a net surplus, because the improved coordination of land uses that would be induced by the rebates for positive externalities would make land worth more in the aggregate with the rebates than without them.
If the government collects only a fraction of the rental value of land, then rewards for positive externalities would not be self-financing, and one would need to ask whether this particular use of scarce government funds was more important than other potential uses. Still, it would be possible to have ‘betterment districts’ that would levy special-purpose taxes on the land that benefited from activities like private parking garages. With such special-purpose taxes there is no competition for scarce public funds. And while the system of rewards for positive externalities would not generate a net surplus for the Treasury as it would under a system of full taxation of the rental value of land, it would generate a net improvement in the efficiency of the economy.
As with negative externalities, the fact that labor and capital are not perfectly mobile means that not all of the effects of activities with positive externalities are reflected in land values. And to the extent that labor and capital are affected, negative effects are likely to be more prominent than positive effects. Consider structures. If an existing structure is efficient for the current pattern of surrounding land uses, and if new use of surrounding land raises the value of the land under the structure, then the structure can be expected to thereby become suboptimal. The pre-existing structure is generally not as intensive a use of land as would have been chosen with knowledge of the new activity. Therefore the economic life of the pre-existing structure (the time until it is efficient to demolish it) is shortened by the new activity that increases land value. Even though rent per square foot of built space rises, the rise is not enough to offset the increase in land value, which presumes the newly efficient size for a structure, thereby reducing the amount of the current return that is attributable to the existing structure and pushing the structure toward obsolescence.
A similar phenomenon occurs with respect to the personal value that people attach to locations, what might be called ‘location-specific human capital’. If all persons had equal demands for all locations, there would be no location-specific human capital. But incomes, tastes and personal histories differ, leading to differing personal values for locations. Familiarity with a location tends to lead people to attach a special value to it, a value that tends to be lost if the characteristics of the location change. Thus a change in land use that raises surrounding land values will generally reduce the location-specific human capital of the residents of that area. One example of this phenomenon occurs when a neighborhood is ‘gentrified’, and the prior residents of modest means can no longer afford to live there and must seek other housing options that are less attractive to them than the opportunities they lost. This represents a reduction in their location-specific human capital.
Because increases in land value generally cause reductions in both the value of location-specific human capital and the value of pre-existing structures, an efficient system of incentives for activities that increase land values would offer incentives reflecting only the net benefits, after the value of the negative consequences for physical and human capital had been subtracted from the increase in land values. An ideal system would collect all of the extra increase in land value and use it to provide compensation for the negative consequences for physical and human capital.
Compensation for private activities that increase land values is more like a rebate of a land tax than a tax itself, but it nicely complements a system of taxes for activities that reduce land values. They belong together in a discussion of how a land tax can be better than neutral.
Another way in which a tax on land can be better than neutral is that it can ameliorate imperfections in lending markets. Because capital markets are not perfect, people vary with respect to the discount rate that they apply to future taxes. The lower the discount rate, the higher will be the present value that a person assigns to future land tax payments. This means that an increase in taxes on land will lower the bid prices of those with low discount rates by more than it lowers the bid prices of those with high discount rates. Thus an increase in taxes on land will tend to shift land from people with low discount rates to people with high discount rates. Since people with high discount rates get greater returns from their assets, this increases economic output.
The last way in which a tax on land can be better than neutral is that it can reduce inefficiency associated with land speculation. If everyone had the same view as to what the future held, then there would be no possibility of making speculative gains. But people differ in their views about the future, creating opportunities for people to engage in speculations that they believe will be profitable. For speculations in widely traded standardized commodities, it is not necessary to own the commodity to speculate; one can speculate through futures contracts. With land however, while there are some rudimentary futures contracts, these are available in only a few places and do not reflect the individual variations in land values that arise because of land’s locational fixity. Therefore most of the speculations that people might wish to make with respect to land require that the speculators own land for the duration of the speculation.
In theory, land speculation can either improve or worsen economic efficiency. The way that land speculation can improve efficiency is by preventing the premature development of land at suboptimal intensity. If land in a particular region is destined to rise in value at a rapid rate and few people realize this, then in the absence of speculation in land, people would tend to build structures that would turn out not to have been worth building once the higher value of land became known. In retrospect it would be seen to have been more efficient to leave land undeveloped until its value rose, and then develop it at a higher intensity. The possibility of profitable speculation in land would induce those who could foresee the future higher value of land to buy land and hold it unused until its high value became generally known. In the process, the inefficient premature development of land at a sub-optimal intensity would be prevented.
The way that land speculation can worsen economic efficiency is by giving land the greatest value to those who make the most foolish overestimates of the rate at which it will rise in value. The result is that much land is owned by people who have paid more for it than can ever be recovered from its use. This is an example of a phenomenon that economists call ‘the winner’s curse’. The general description of the winner’s curse is that when a thing of uncertain value is sold to the highest bidder (or a contract is let to the lowest bidder), the person who wins the thing by bidding highest (or lowest) will suffer the curse of having paid too much for it (or bid too low). The winner’s curse arises commonly in such arenas as bidding for oil leases and construction contracts.
The effect of the winner’s curse in the land market is that those who attach the greatest value to land are persons who believe that it would be unprofitable to invest in improvements to the land, because it will soon rise rapidly in value, and any improvements constructed now would be sacrificed when the higher value of land was realized. Thus, in land speculation, the winner’s curse produces the social cost of an artificial scarcity of land for current use.
A rise in the rate of taxation of land reduces the potential profit from speculating in land because it reduces the selling price of land. Thus a rise in land taxes reduces the speculative demand for land (the number of pounds per month that a speculator is willing to pay in interest in taxes to hold a parcel of land with given prospects in his mind), but a rise in land taxes does not reduce the number of pounds per month that a current user of land is willing to pay in interest in taxes for current use of land. Thus a tax on land shifts land from speculators to current users, thereby reducing the tendency toward an artificial scarcity of land from speculation. This is shown diagrammatically in Figure 3.
The width of Figure 3 represents the fixed quantity of land that is available for current or speculative use. The demand curve that goes down from the left, Du, represents the demand for land for current use. Speculative demand for land is measured from right to left and is represented by the demand curve Ds. The intersection of these two curves specifies the price, P0, at which the demands for land for use and for speculation sum to the available quantity of land. The amount of land that is devoted to current use is Q0. An increase in the tax on land leaves the demand schedule for land for current use unchanged, but lowers the demand schedule for land for speculation to D¢s. With the new demand schedule for land for speculation, the price for land for current use falls to P1 and the quantity of land devoted to current use rises to Q1.
Since land speculation can, in theory, either improve or worsen the efficiency with which land is used, it is an empirical question as to which effect predominates. While no conclusive evidence is available, there are reasons for believing that the winner’s curse is more powerful than the beneficial speculative withholding of land. The first reason is casual empiricism. As one looks around cities, it is easy to find places where vast amounts of barely improved land surround islands of intense development. It commonly happens that land near very tall buildings has been occupied for many decades by one- or two-storey buildings, or even by nothing more than a paved car park. This cannot be efficient. There would have been time to fully depreciate substantial buildings while owners have been waiting for the right time to develop land. The inefficiency of leaving such land underused is the social cost of the winner’s curse.
There is little evidence of inefficiency from premature land development. If there were such evidence, it would take the form of recently constructed buildings being demolished for more extensive development, or possibly, in less extreme and less obvious cases, new development surrounded by recent, substantially less extensive development that could have been just as extensive as the new development if only the developers had waited a little longer. Examples of the first type of evidence are quite rare. Examples of the second type would be easy to miss, but still there seems to be no reason to suppose that they represent a problem of efficiency as widespread as the problem of indefinitely deferred development induced by the winner’s curse.
The second reason for believing that deferred development induced by the winner’s curse is more of a problem than premature development induced by a lack of foresight is that the latter problem can more easily be resolved by new markets. If land should come to be taxed so heavily that people stopped speculating in land, and if it then transpired that people often lost money on their investments in developing land because they underestimated the rate at which land values would rise, then there would be a potential private market for good predictions of the rates at which land values would rise. A person who could foresee future rises in land values (a good speculator in today’s markets) would have a comparative advantage in selling insurance against a rise in land values, and hence in land taxes, that would make current development unprofitable. Thus, in making land speculation unprofitable, land taxation would redirect the energies of those who believed they could predict rises in land values, away from buying and holding land that they believed would rise in value, and toward selling insurance against rises in land values where they believed that there would be none. People would be discouraged from building where land values were foreseen to rise rapidly, by the unavailability of reasonably priced insurance against rises in land taxes. Thus the potential development of a market for insurance against rises in land values makes it unnecessary to maintain land speculation as a mechanism to secure efficient deferral of development. The benefits of eliminating the inefficient deferral of development that is caused by the winner’s curse can be secured without foregoing the benefits of redirecting development, based on private knowledge of where land values will rise rapidly.
V. Measuring the Economic Gains from Shifting Taxes to Land
There are three principal sources of economic gains from shifting taxes away from capital and labor and onto land:
1) The removal of taxes from labor motivates people to make more efficient decisions about how much to work;In this section I report estimates of the magnitudes of these effects, derived from a general equilibrium model of the U.S. economy.
2) The removal of taxes from capital motivates people to make more efficient decisions about how much to save; and
3) increasing taxes on land motivates people to speculate less in land, increasing the amount of land available for current use.
A model of an economy is necessarily oversimplified compared to a real economy. Nevertheless, model builders hope that their models will have enough resemblance to actual economies that results derived from models will be relevant to actual economies. Whether this is true of any particular model must be a matter of conjecture until the model is tested against reality. The model presented here has many features in common with models that economists have found useful, but its novel features are untested.
In the model, the total output in year is a function of the quantities of three inputs used, land, labor and capital. The output that can be attained from given inputs increases each year as a result of technological advance. The relatively novel element in the production function is the distinction between land and capital as two separate factors of production; many economists include land with capital, making no distinction between the two.
Economists give the name ‘elasticity of substitution’ to the ratio of a percentage change in factor proportions to the percentage change in relative factor prices that induced it. If a 1% change in the ratio of the price of land to the price of labor induces a 2% change in the ratio of the quantity of labor used to the quantity of land used, then the corresponding production function is said to have an elasticity of substitution of 2 between land and labor. The greater the elasticity of substitution of a production function, the more an economy benefits from increases in the availability of individual factors of production.
Many economists have studied the elasticity of substitution between labor and capital. After considering such studies, I decided that 0.8 is a plausible figure for the elasticity of substitution between labor and capital, and I assumed that the elasticity of substitution between any two of land, labor and capital is 0.8.
Like many other economists, I employ a model with a very limited range of goods: There is one private consumption good in the model, one public consumption good and one capital good. Consumers are also concerned with leisure, which they consume whenever they are not sleeping (8 hours per day) or working.
I assume that the resources that are capable of producing a unit of the private good are also capable of producing a unit of the public good. Capital goods are treated somewhat differently, in being subject to rising costs. For the level of output where average cost is minimized, a unit of capital goods has the same cost as a unit of consumption goods. For other levels of output, the average cost of capital goods is a quadratic function of quantity. The level of output of capital goods at which average cost is minimized in any given year is a weighted average of the level at which average cost would have been minimized the previous year and the actual level of output of capital goods the previous year.
The quantity of land that is employed in any year reflects the response of land speculation to taxes on land. The idea that the quantity of land employed in production might vary positively with the level of taxes on land is an innovation in the model.
Households choose the quantity of labor that is employed in production, taking account of the wage after taxes, their total resources, and their preferences. (The wage before taxes is the marginal product of labor.) The quantity of capital that is employed in production is the quantity that households have chosen to accumulate, taking account of the return to saving after taxes, their total resources and their preferences. (The return to capital before taxes is the marginal product of capital.)
The statements about labor and capital in the previous paragraph incorporate an assumption that the economy that is being modeled is ‘closed’. That is, capital and labor cannot go elsewhere or come from elsewhere. That is a reasonable approximation for labor, but it is increasingly implausible for capital. Therefore, in exploring the consequences of changes in taxes, I investigate the implications of both the assumption that the economy is closed with respect to capital and the assumption that it is open with respect to capital and small with respect to the world, so that it can import as much capital at it wishes without affecting the interest rate. It is not plausible to suppose that the U.S. economy is small enough compared to the rest of the world that its actions would have no effect on interest rates. Therefore one should regard the results for open and closed economies as bounds on what is plausible with respect to saving and investing.
The model treats an economy as composed of a large number of households with identical tastes and incomes. The members of these households are assumed to make spending and saving decisions under an assumption that they will live forever. They are assumed to derive utility from consumption of three goods: a private good, a public good, and leisure.
Just as with elasticity of substitution in the production function, the propensity of people to change their consumption patterns in response to changing relative prices can be described by elasticities of substitution. If people increase the ratio of private goods to leisure that they consume by 2% when the relative price of goods compared to leisure falls by 1%, then they are said to have an elasticity of substitution between goods and leisure of 2. Based on studies by others, I assume that the elasticity of substitution between goods and leisure in any period of time is 0.8.
The decisions people make about saving depend, among other things, on the elasticity of substitution between consumption in different periods. Based again on a review of the work of others, I assume that the ‘intertemporal elasticity of substitution’ is 0.375. This means that if the price of consumption in one period relative to another falls by 1%, consumers will want to increase their consumption in that period relative to the other by 0.375%.
The National Income and Product Accounts (NIPA) for the U.S. economy list ten sources of public revenue, as shown in Table I. The first eight of these sources are taxes; the last two are treated in the model as payments for privately produced goods.
Based on data published by the Internal Revenue Service, the revenue from the individual income tax was divided into revenue from asset income and revenue from labor income, and marginal and average tax rates were estimated for both tax bases. Based on social security legislation, social insurance payments were divided equally between taxes on employers and taxes on employees. Based on figures developed by Martin Feldstein, social insurance payments were also divided into a fraction that is taxes (h) and a fraction that is individual savings.
Taking account of the upper limit on Social Security contributions, I estimated average and marginal social insurance tax rates. I assumed that excise taxes internalise externalities, so that they generate no excess burden. On this basis, they were assigned a marginal tax rate of zero, since the marginal rate of a tax reflects the excess burden that it generates. (Average tax rates determine the revenue that they generate.) Tax rates for other taxes were computed as the ratio of tax revenue to the tax base. This produced the tax rates shown below:
Next, the eight taxes were integrated into three ‘standardized taxes’: a tax on asset income, a tax on labor income and a tax on consumption. The rates of these taxes are shown in Table III.
For the first year of the model, the economic facts are determined by history. (2000 is taken as the initial year of the model because that is the most recent year for which it might reasonably be supposed that people would make economic decisions based on an assumption that tax rates in the current year would prevail in future years. In subsequent years, proposals for tax rate changes have continually been under public discussion.) These facts permit one to identify the parameters of the production function and the utility function.
Once the parameters of the production function and the utility function have been identified, one can “grow” the model over time, based on the assumption that the decisions made in the initial year were part of a rational long-term plan. People make work and saving decisions based on the maximization of their utility function, given the prices of goods and leisure, now and in the future. Output is a function of inputs. Tax revenues are functions of tax rates and the magnitudes of tax bases. The addition to the capital stock is determined by how much people save. The discount that people attach to future goods must be such that following their long-term plan neither bankrupts them nor leaves them eventually saving all their income.
If one assumes that at some subsequent time tax rates will change, then this produces new prices of consumption and leisure. As soon as people are able to take account of the new tax rates, they change their behavior, embarking on a new plan that maximizes their utility, given the new prices of goods and leisure, now and in the future, that are implied by the new tax rates. One can thus evaluate the alternative futures implied by alternative tax regimes.
Figures 1 through 10 show the effects over 25 years of shifting as much tax as possible to land in the U.S., treating the U.S. economy as a closed and as an open economy with respect to capital. I interpret ‘shifting as much tax as possible to land’ as collecting 90% of the rental value of land. This permits elimination of the corporation income tax, all sales taxes, all tariffs, and all property taxes (rates) that fall on structures. The rate of the income tax is varied from year to year to insure that the same revenue is collected under the new tax system as under the old system. For the U.S. as a closed economy, the income tax rate starts at 8.21% and falls to 6.190% after 25 years. For the U.S. as an open economy, it starts at 5.37%, falls to 4.16% after 18 years, and then rises to 4.38% after 25 years.
The reason that the tax rates needed to match the revenues of the existing system are so much lower for an open economy is that, with an open economy, interest rates do not fall when people decide to save more, and this induces them to save so much more, and to work so much more so that they can save without cutting back too much on their consumption, so that the tax base is much larger in the open economy.
Figure 1 shows the relative prices of current and future goods in the three scenarios.
Figure 1: Relative Prices of Present and Future Goods, Under Current Taxes and after Shifting as Much Tax as Possible to Land, in Closed and Open Economies
In both the closed economy and the open economy, the prices of future goods are generally lower after shifting taxes to land. The primary cause of this is the removal of taxes from savings, which increases the interest rate for savers. In the closed economy this is offset to some extent by the fact that as people save more, the rate of return to capital declines. (In the open economy, the interest rate is determined by the global market for capital.) For the first few years, there is an additional factor that reduces the net interest rate in the closed economy, and causes the price of future goods to be higher after the tax shift than before. This is the continuing fall in the price of capital (producing ‘negative capital gains’), as the economy becomes adapted to producing capital at the higher rate that is induced by the tax shift.
Figure 2 shows the impact of the tax shift on the relative price of leisure, immediately and over time. The price of leisure rises in the short run, because of the reduction in taxes on labor income. There is also an increase in the rate at which the price of leisure falls over time, because of the higher returns to saving that result from the removal of taxes from saving. For the closed economy, the price falls slowly initially, because of the impact of the falling price of capital on the return to savings. In the long run, the price of leisure falls most rapidly in the open economy, because added saving does not reduce the interest rate in this economy.
Figure 2: Relative Prices of Present and Future Leisure, Under Current Taxes and after Shifting as Much Tax as Possible to Land, in Closed and Open Economies
Figure 3 shows the impacts of the tax shift on wages. In the short run, after-tax wages rise by about 20% in the closed economy and about 18% in the open economy. The smaller rise in the open economy arises from the fact that the greater savings opportunities of the open economy induce people to want to work more, causing wages to be lower. After 25 years, wages are about 36% greater in both the closed economy and the open economy.
Figure 3: Percent Change in Real Wage After Tax from Shifting as Much Tax as Possible to Land, in Closed and Open Economies
Figure 4 shows that work effort in the closed economy rises initially by about 12%, moving up to an increase of about 17% after nine years, sliding back to an increase of about 14% after 25 years. For the open economy, work effort increases initially by about 27%, rises to an increase of about 27% after 10 years, and then falls back to an increase of about 22% after 25 years. The reason that the increase in work effort is so much greater for the open economy than the closed economy is that the open economy provides an opportunity to expand saving without causing interest rates to fall. People respond to this opportunity by working more, so that they can save more without cutting back significantly on their current consumption.
Figure 4: Percent Change in Effort per Worker from Shifting as Much Tax as Possible to Land, in Closed and Open Economies
Figure 5 shows the impact on the return to saving. For the closed economy, there is an initial fall in the return to saving of about 34%. This is caused by the fact that the price of capital falls steadily, after an initial increase of 22% in response to the strong increase in the demand for capital. But the return to capital then increases for ten years, when it is 22% higher than with current taxes. After that it falls steadily, reaching a rate that is up only about 6% after 25 years. In the open economy, people will save abroad rather than accept lower rates of return. But the open economy is able to absorb more capital without a reduction in returns, because the opportunity to save abroad generates such a great increase in the supply of labor. In the open economy, the increase in the return to saving is determined by the fall in taxes. The return to saving is up 19% initially, rising to an increase of 23% after 21 years, and falling only fractionally by 25 years.
Figure 5: Percent Change in Return to Capital from Shifting as Much Tax as Possible to Land, in Closed and Open Economies
Figure 6 shows increases in saving and investment. In the closed economy, saving and investment rise initially by about 86%. They continue to rise for ten years, reaching an increase of 182%, then fall back gradually to an increase of 139% after 25 years. In the open economy, saving can differ from investment. Investment rises initially by 74%, increases for nine years to an increase 189%, then falls to an increase of 43% after 25 years. This fall in investment comes from the fact that people in the open economy are becoming so rich that they are much less interested in working, so less capital is needed. Saving in the open economy is astounding, rising by 287% initially and then continuing to grow, reaching an increase of 739% after 25 years. What is happening is that the untaxed citizens of this economy have a comparative advantage in saving, and are moving in the direction of doing all of the world’s saving. The idea that this economy has no effect of world interest rates will eventually become untenable. Some economists have suggested that such great increases in saving are evidence against the hypothesis that people choose saving to maximize the kind of utility function that is used in the model. The suggested alternative hypothesis is that saving is undertaken to deal with uncertainty. This leads to significantly lower increases in savings in response to tax changes.
Figure 6: Percent Change in Saving and Investment from Shifting as Much Tax as Possible to Land, in Closed and Open Economies
Figure 7 shows the increase in the stock of capital that results from shifting as much tax as possible to land. In the closed economy, the increase in the stock of capital increases steadily for the full 25 years that are shown, reaching an increase of 57% after 25 years. In the open economy, the capital stock initially rises at the same rate, but then it levels off, reaching an increase of 48% after 25 years, as people become so rich that they want to work less, and therefore have less need for capital in the domestic economy.
Figure 7: Percent Change in Capital Stock from Shifting as Much Tax as Possible to Land, in Closed and Open Economies
Figure 8 shows the change in the consumption of private goods that results from shifting as much tax as possible to land. In the closed economy, there is an initial increase in private goods per worker of about 23%. This slows down fractionally for two years as the production of capital gears up, then rise to an increase of 35% after 25 years. In the open economy, the initial increase in private goods per worker is smaller, only 14%, because people are devoting so much of their income to foreign investment. But the increase is steady after that, reaching 30% after 25 years.
Figure 8: Percent Change in Private Goods per Worker from Shifting as Much Tax as Possible to Land, in Closed and Open Economies
Figure 9 shows the increase in the consumption index that results from shifting as much tax as possible to land. The consumption index combines changes in goods and changes in leisure, along with an unchanging quantity of public goods. In the closed economy, there is an increase in the consumption index that starts at 4% then declines over four years to about 3% as work effort expands and consumption of goods falls, to accommodate the growing productivity of capital investments. After 25 years, the increase in the consumption index is just under 8%. In the open economy, people work so hard to take advantage of the opportunity for foreign investment that the consumption index falls initially by about 2%. It then rises gradually, reaching the current level after 11 years and an increase of 4% after 25 years.
Figure 9: Percent Change in Consumption Index from Shifting as Much Tax as Possible to Land, in Closed and Open Economies
The index of consumption is not an index of well-being, because it does not take account of the additional saving that people do. To obtain a measure of the annual improvement in the well-being of citizens that comes from shifting as much tax as possible to land, I take the increase in saving and add or subtract the amount necessary to compensate for the change in the consumption index. The result is a concept known in economics as the “Hicks equivalent variation.” It reveals the amount of money that people would need to be given to make them as well off as under a contemplated change. This is shown if Figure 10. In the closed economy, the net gain per worker from shifting as much tax as possible to land starts out at about $10,200 per year and rises to about $19,000 per year after 25 years. In the open economy it starts out at about $12,500 per year and rises to $40,400 per year after 25 years.
Figure 10: Net Annual Gain per Worker from Shifting as Much Tax as Possible to Land, in Closed and Open Economies
VI. The Ethical Case for Taxing Land
The ethical case for taxing land is based on two premises. The first is that people have rights to themselves. This has not been controversial since the end of slavery, so I will simply assume that this is agreed. The second premise is that all people have equal rights to natural opportunities. This is not so widely agreed.
Natural opportunities include not only land, but also water, fish in oceans and rivers, the frequency spectrum, minerals, virgin forests, and geosynchronous orbits. Some natural opportunities, such as the opportunity to use the oceans for transport, are most valuable to people when all are allowed to use them as they wish. (This does not imply that their value is greatest when all can pollute as they wish.) Other natural opportunities, such as most plots of land, are most valuable when one person has exclusive use of them.
The processes that humans employ to determine who shall have exclusive use of natural opportunities are complex. To some extent, opportunities are assigned to those who first make use of them. However, another important component of the natural-opportunity-assignment process is the ability and willingness to use deadly force to exclude others. Americans from Europe undertook some negotiations with the native American Indians, but primarily they threatened to kill the Indians if they did not agree to move into smaller territories. All over the world, nations emerged when war-minded leaders imposed their rule where they could. We have built a relatively humane world on this violent foundation, but the origins of the assignment of natural opportunities cannot be characterized as just.
Nor would have been just (or efficient) to adhere to a rule of initial assignment based on first use. It would not be just because a person who arrives later than another is not inherently less deserving. (It would not be efficient because a rule of assignment based on first use promotes inefficient, excessive investment in being first. Still, to motivate efficient discovery, it pays to provide some reward for discoverers.)
Justice requires that we acknowledge the equal rights of all persons to the gifts of nature. At the level of relations among nations, this requires every nation to determine whether it is using more than its share of natural opportunities, and if it is using more than its share, to compensate other nations that therefore have less than their shares.
Within a nation, the requirements of justice are not so rigorous, as long as anyone who wishes can migrate to another nation and take with him a claim to an equal share of natural opportunities. But the competitive equilibrium among nations, in a world that recognized an obligation to share natural opportunities among nations in proportion to population would be for each nation to grant each citizen an equal share of natural opportunities. Any nation that declined to do so could expect to lose citizens to nations that made such offers. One way of providing for the equal sharing of the value of natural opportunities within a nation is to collect the rental value of natural opportunities from each person who is granted exclusive access to them, and use the proceeds to provide a guaranteed income for everyone.
An additional ethical reason for recognizing equal rights to natural opportunities is that it may be necessary to secure world peace. Nations have arisen through violence. While the world condemns violence among nations, it has persistently acquiesced to regimes established by violence. The greater the natural resources of a nation, the greater is the attraction to potential tyrants of the possibility of taking over the nation. If the world is able to establish an understanding, backed up by the threat of economic boycotts, that nations have an obligation to share the value of natural opportunities in proportion to population, and that people are free to leave nations that they find unacceptable, then the return to violent appropriation of power will be removed. As long as we accept the continued exercise of disproportionate power over natural opportunities by those who acquired that power through violence, we will have difficulty persuading potential usurpers of power that we will not accept their conquests.
VII. Answer to Arguments against Taxing Land
One argument that is sometimes made against high taxes on land is that as land taxes rise, the selling price of land falls, eliminating the base of the tax, so that only modest taxes on land are feasible. As a theoretical matter, this argument is completely false. This is easier to see if you recognize that a tax on land incorporates a time dimension that is arbitrary. The tax is prescribed as some percentage of the value of land per year, but it could just as well be described as a percentage of the value per month, per day or per hour. For land that is not rising in value, the percentage of the rental value of the land that is taken by an accurately assessed tax of t % per unit of time is t/(t + i), where i is the interest rate per unit of time. As the unit of time approaches zero, the interest rate per unit of time approaches zero, so that a tax of t % of value per shrinking unit of time approaches a tax that collects all the rent of land, in the limit as the unit of time approaches 0.
While it is theoretically possible to capture fractions of rent approaching 100% by taxes on the value of land, there are practical difficulties in doing so. The selling price of unimproved land will be the present value of the part of rent that buyers and sellers expect to be left after taxes. Therefore, if taxes collect fractions of rent that approach 100%, the selling price of land will be dominated by the errors that are expected in the assessment process. Therefore a tax system that seeks to collect almost all of the rental value of land must use some assessment system other than observing market prices of land. There are several techniques that may be useful.
First, if nearly all of the rental value of land is collected in taxes, the selling price of land will be nearly zero. Assessors can purchase parcels of land with obsolescent improvements, demolish the improvements, and offer the land for sale at auction, under a rule that the bid will be the tax per year for, say, the first three years, and after that the tax will be determined by the assessor’s estimate of the rental value of the land, as determined by similar auctions and other processes.
If assessors were conducting such auctions regularly, they could hold assessment contests in which the contestants competed by offering land value functions that would be evaluated by the accuracy with which they predicted auction results. The contestant who provided the function with the smallest average error would be given a prize, and the winning function would be used to assess the value of land that had not been auctioned.
Another thing that assessors can do is to develop options markets in land. That is, they can enter into contracts with potential users of land to supply land with specified characteristics for specified tax rates. For example, someone who was interested in opening a restaurant might offer £3,000 per month for a parcel of 2,000 square feet within a quarter of a mile of the center of town. Such offers would set lower limits on the rental value of land. With such devices, land can be assessed for tax purposes even if the selling price of land is close to zero.
It is sometimes suggested that collecting all of the rental value of land is not feasible because it would bankrupt the financial system, which uses land as collateral for loans. The use of land as collateral for loans is a fact that would need to be weighed carefully in working out a feasible transition to a system of collecting nearly all rent publicly, but it does not make public collection of rent impossible. Public collection of the rental value of land, and the consequent elimination of other taxes, would represent redistribution from the old to the young and unborn. It is equivalent to saying that each person enters the world with a more complete right to himself or herself (to the extent the income and value added taxes are reduced) and with a right to an equal share of the gifts of nature. If we are to acknowledge that every newborn person does have such rights, we must reduce the wealth of previously born persons, corresponding to the expectation of collecting the shares of rent of the newborns. The question that then arises is, Who should pay?
This can be interpreted as a question of who should pay the cost of an accident. We have had a moral accident. We have been living under the delusion that it is possible for a person to have a respectable claim to a disproportionate share of the gifts of nature. This is akin to the delusion under which humanity long lived, that it was possible for one human being to own another. In recovering from such a delusion, we need to disappoint some people. If not the slave owners, then the people who compensate the slave owners. If not the land owners, then the people who compensate the land owners. I propose that the cost be divided among land owners, shareholders in financial institutions that made loans with land as collateral, and owners of wealth in general.
As a first approximation, people would continue to hold title to the land to which they now hold title, and would continue to owe whatever money they now owe. But compensation could be sought on a case-by-case basis, by individuals who stood to bear the costs of the moral accident disproportionately and did not have substantial assets. Any financial institutional whose continued existence was threatened by the transition would be bailed out in exchange for a significant fraction of its equity. The costs of the compensation would be paid by a capital levy.
Morality requires that we acknowledge the equal rights of all to the gifts of nature. Doing so is feasible and promotes efficiency as well.
In the model, the total output in year t, Qt, is a ‘constant elasticity of substitution’ (CES) function of the quantities of three inputs used in year t, land (Tt), labor (Lt) and capital (Kt):
Qt = c (t – 2000) (a1T ta + a2Lta + a3Kta)(1/a). (1)
The factor c (t – 2000) represents the impact of technological
change. It is 1.0 in the year 2000 and increases by the factor c
each year after 2000. Based on studies by other scholars, I chose
a value of 1.01 for c in this model, implying that technological
innovation increases the output that can be obtained from given inputs
by 1% per year. The relatively novel element in this production
function is the distinction between land and capital as two separate
factors of production.
The statements about labor and capital in the previous paragraph
incorporate an assumption that the economy that is being modeled
is ‘closed’. That is, capital and labor cannot go elsewhere or
come from elsewhere. That is a reasonable approximation for
labor, but it is increasingly implausible for capital. Therefore,
in exploring the consequences of changes in taxes, I investigate the
implications of both the assumption that the economy is closed with
respect to capital and the assumption that it is open with respect to
capital and small with respect to the world, so that it can import as
much capital at it wishes without affecting the interest rate.
The nine taxes mentioned in Table II in the text are assigned symbol in
Table I A
These nine taxes are combined into three ‘standardized taxes’, a tax on asset income ( j), a tax on labor income (l ) and a tax on consumption ( p), using the following equations:
j = 1 – (1 – n)(1 – f )(1 – w)(1 – m) (5)
l = 1 – (1 – (g + r + y)/(1 + r/h)(1 – m) (6)
p = (1 + s)(1 + x) – 1 (7)
These equations reflect the ways in which payments of some taxes are deducted from the bases of other taxes. The consistency of the equations is established by the fact that total tax revenue is the same for the combination of the standardized taxes as for the combination of the actual taxes.
For the first year of the model, economic facts are determined by history. (2000 is taken as the initial year of the model because that is the most recent year for which it might reasonably be supposed that people would make economic decisions based on an assumption that tax rates in the current year would prevail in future years. In subsequent years, proposals for tax rate changes have continually been under public discussion.) These facts permit one to identify the coefficients of the production function and the utility function.
An additional parameter that emerges from the first-year data is l, the ratio of marginal utility to price for private goods and for leisure. Utility maximization requires that, for a utility-maximizing individual’s planned consumption and leisure in the current period and in all future periods, this ratio will be constant. This permits one to grow the model over time. People make work and saving decisions based on the maximization of their utility function. Output is a function of inputs. Tax revenues are functions of tax rates and the magnitudes of tax bases. The addition to the capital stock is determined by how much people save. If the discount factor, d, is chosen properly, the saving rate as a function of time will be asymptotically horizontal. (With too high a discount factor, the saving rate will rise over time to 100%, while with too low a discount factor, the saving rate will become negative, and the capital stock will vanish.)
If one assumes that at some time after the initial time tax rates will change, then this produces new prices of consumption and leisure. As soon as people are able to take account of the new tax rates, they change their behavior. It would be improper to restore equilibrium by changing d. This parameter of the utility function should remain fixed once it is estimated. Instead, equilibrium is restored by identifying l¢, a revised value of l, the ratio of marginal utility to price for consumption goods and leisure in all periods. For the correct value of l¢, the savings rate will again be asymptotically horizontal.
to email this page to a friend: right click, choose "send"
Wealth and Want
... because democracy alone hasn't yet led to a society in which all can prosper