By Nicolaus Tideman
“The Henry George Theorem” is the name that economists have given to the postulate that, under certain circumstances, a public service will increase land rent enough that, if the increase in land rent is collected as public revenue, this revenue will sufficient to pay for the service. This idea is important because it maintains that public services can be financed not by sales taxes, income taxes or other taxes that burden the economy, but simply by collecting the increase in land rents that result from the public services provided.
The main premise of the Henry George Theorem is simple. For many public services, such as parks and libraries, their benefit is greater to the people who are closer to the place where the service itself is provided. Therefore, people bid up the rental value of land that is closer to provision of such desirable services. If all of the benefits from proximity to a new public service are reflected in increased rents, and if the service is worth at least as much as it costs to provide, then public collection of the increase in rents will suffice to pay for the service.
Despite its simplicity, there are half-hidden conditions that must be satisfied for this premise to apply in practice. First, the public service must be worthwhile. If a public service does not have benefits that are at least as great as its costs, there would be no reason to expect provision of the service to raise land rents by enough to pay for the service.
Second, access to the benefits of the public service must be spatially limited. If, for example, the public service is something like the Wikipedia encyclopedia that is available on the Internet, then people can use it wherever they use their computers and thus do not need to be close to its location to reap the associated benefits. It is possible that a service like Wikipedia raises land rents by increasing everyone’s income, by for example, giving easy access to information that can speed and inform decision making. There is no reason expect, however, that a service for which physical proximity does not increase the attraction, efficacy, and/or use of that service will raise land rents to a degree equal to or greater than the cost of benefits from the service.
Third, there must be a market process by which rents are determined. If people live in rent-controlled apartments and landlords are not allowed to raise rents or tear buildings down, then there will be no increase in observed rents, despite the benefits provided by the public service.
Fourth, there must be enough people who are both a) mobile, and b) able to benefit from the service in full, to bid up rents to a degree equal to or greater than the value of the full benefit of the service. The measure of demand for this condition is not dollars per person, but rather dollars of benefit from the service per acre of land occupied. Thus, there is room for some variation among people, as long as, for any given distance from a public service, the ratio of benefit from the public service to acres occupied would be the same for all persons. The simplest way to ensure that this condition is satisfied is to assume that all persons within the locality where the service is provided have the same income and the same tastes.
This kind of assumption is attractive to economists who want to reach purely mathematical conclusions, but, practically applied, it is not realistic. If instead one makes the realistic assumption that people differ in their incomes and tastes, and that, for each type of land, the ratio of benefit from the service to acres occupied varies among persons, then it is no longer possible to reach the conclusion that the full benefit from the service will be reflected in increases in land rents. Still, the more uniform the ratio of benefit to amount of land occupied at any given distance, the more fully the benefit of a public service will be reflected in increased land rent.
In some cases, increases in the rental value of land will be more than enough to pay for a public service, even though the service has costs that are greater than its benefits. This can happen if the service has negative consequences that are not charged to its account. Two examples are given here.
First, consider the building of a new subway line. Land rents in the vicinity of the subway stops will rise greatly. But to take efficient advantage of the subway stops, it will be necessary to tear down existing structures and replace them with taller ones. The coming of the subway line turns the existing structures into trash. A proper accounting of the costs and benefits of the subway line would include a charge for the reduction in the value of the structures. If this loss of value is not charged to the subway line, then it would be possible for the benefits to be greater than the calculated costs, even though the benefits were actually less than the full social cost. Any change in public services can be expected to change the value of fixed improvements, and, to the extent that the existing improvements were appropriate for the site, a change in public services will generally reduce the value of the improvements.
Second, consider public parking lots opened in a neighborhood with narrow streets, old apartment buildings, and a previous severe shortage of parking. The neighborhood was previously well suited to people who did not own cars, but the public parking lots make it attractive to people who own cars. Rents rise, and the people who do not own cars can no longer afford to live there. If these people were perfectly mobile, so that they could move at zero cost to some other place that provided the same level of satisfaction for them, then the need for them to move would not be economically consequential. However, in the more realistic case, in which those who move are worse off for the combination of moving costs and higher rents that are not quite high enough to make it worth moving, these “gentrification costs” are part of the cost of opening the parking lots. If these costs are not included in the calculation of the costs and benefits of the parking lot, then it would be possible for the increase in land rents to be greater than the calculated cost of the parking lots, even though that increase in land rents was less than the full social cost of the parking lots.
To summarize, the benefits from public services tend to be reflected in increased land rents in the areas where the services are accessible. Under certain conditions, the benefits from public services will exactly equal the increase in land rents. For this equality to occur, the benefits must be received only by those in a limited area, rents must be determined by a free market process, and all persons in the affected area must have the same ratio of benefits to acres occupied, at any given distance from the service. When people differ in their incomes and tastes, some small fraction of benefits will not be reflected in increased land rents.
It is also possible for the financial cost of a public service to be less than the resulting increase in land rents––even though the service is not worthwhile, if the public service has negative consequences that are not charged to its account. Two costs that must be included in a full cost accounting for a public service are the resulting reduction in the value of fixed improvements and the dislocation costs for people who have lived in the area and do not value the public service as highly as it is valued by newcomers. When these costs are charged to the account of a public service that has benefits within a limited radius, a comparison of the increase in rent, with the combination of the ordinary costs of the service and these external costs provides a good test of whether the services are worthwhile.
 Rent, as I use the term in this blogpost, is a payment for the use of land or some other natural opportunity. When money is paid for the use of a building, the land rent is the part of that payment that is attributable to the land under the building. Because the best use of land often involves constructing buildings that last a long time, the meaning of the rent of land under an existing building is not obvious. The conceptual resolution of this difficulty is that the rental value of a plot of land for the coming year is how much more valuable it would be to have the use of the plot of land, beginning with vacant land, into the indefinite future beginning now, than it would be to have the use of the plot of land, beginning with vacant land, into the indefinite future beginning one year from now.
This is a superbly lucid statement of the theorem, Nic. Thanks