Comments

Robert M. Schwab:

Arnott's current paper and related recent research are very valuable contributions to the property tax literature. Together, they are excellent examples of the way economic theory can inform policy debates. I am certain that this line of research will have a profound effect on the way people think about the property tax.

This paper sets out a model where a landowner needs to make two decisions: when and how intensely to develop land. Equations (2) and (3) in the paper characterize the solution to this problem in the absence of any taxes. Equation (2) shows that the landowner should wait to develop until forgone rents just equal the cost of capital. Equation (3) shows that the optimal density requires the present value of the increase in rents from the last unit of capital to just equal the cost of purchasing that unit of capital. That is, along both the timing and density dimensions, optimality requires the landowner to balance costs and benefits at the margin. The paper then looks at how various forms of property taxation alter the landowner's timing and density decisions and investigates the necessary ingredients of a neutral tax system. The paper also examines two interesting policy questions: the taxation of undeveloped land and two-rate taxation.

At several points in the paper, Arnott argues that there are not many examples of two-rate taxes (or the most extreme case, a pure land tax) because of the difficulty of assessing land and structures separately. He suggests that some recent advances in assessment techniques may offer solutions to these assessment problems. For example, he points to Dye and McMillen's (2005) work on teardowns as an important step forward. I argue that the assessment problem is not really the key reason we see so few deviations from the traditional property tax where land and structures are taxed at the same rate. In particular, I argue that the two-rate tax and land tax will remain unpopular (even if the assessment problems were solved) for at least two reasons: they raise troubling equity concerns and are inconsistent with the prevailing view of property rights in the United States.

Land taxation has a long intellectual history. Henry George's 1879 Progress and Poverty is certainly an important contribution to that history. In that book, George argued that one should tax away the returns from land. Such a tax, George said, would penalize speculation and thus encourage people to develop their land. George further argued that the land tax would be just; returns to land are an "unearned increment."¹ That is, those returns are the result of public policies and the public should enjoy the benefits of those policies.

While George was very popular and influential at one time (he nearly won the New York mayoral election in 1886), it is difficult to see any current examples of his influence in the United States. Two U.S. towns—Fairhope, Alabama, and Arden, Delaware—were founded on Georgist principles. A number of small towns in Pennsylvania have a long history of two-rate taxes. But Pittsburgh is clearly the most prominent example in the United States of a graded tax. Pittsburgh instituted a two-rate tax in 1913. The tax on land was raised several times. By 1980 the tax rate on land in Pittsburgh was roughly six times as high as the tax on structures. But Pittsburgh unceremoniously abolished the two-rate tax in 2001. This was clearly a blow to Georgism; as one observer put it, "How could the shining light be doused so easily?"²

What can one make of all of this? It is difficult to argue that it is impossible to implement a two-rate tax. After all, a graded tax survived in Pittsburgh for nearly a century. Admittedly, the assessment of land values is undoubtedly far from perfect. But as Arnott explains in this paper, it is important to recall that from an economic perspective assessment practices do not need to be perfect. The neutrality result requires only that tax liability be independent of the way land is actually...