Assessing Land Accurately

Nicolaus Tideman

[A response to issues raised during discussions
sponsored by the World Bank, December 1998]

Richard Green raises important questions with respect to accurate land assessment. Answers to some of these questions will be empirical, and will be conditioned by the appropriateness and sophistication of the empirical techniques that are used.

Florenz Plassmann and I are working on the problem of land assessment now. We expect to have results by May 1, and sooner if we are lucky. We have data on the 290 property sales within half a mile of the center of downtown Portland, Oregon in the last 10 years. Perhaps a couple dozen are vacant land, and some undetermined number are property sold on the eve of demolition, where all value can be attributed to land. We have a model that divides value between land and buildings in a way that we believe has better theoretical foundations than others that we are aware of. And we plan to use maximum likelihood estimation techniques, which we believe are more suitable than least-squares approaches.

It is important to remember that the neutrality of a land tax does not require that assessments be accurate, but only non-contingent, as long as the tax is less than the rent of unimproved land. This suggests that in highly built-up areas, we should think of land value as the product of a suitable estimating equation. We should require anyone who wishes to appeal an assessment to specify an alternative assessment function, which would then be applied to all property if the function were judged to be better.

Florenz and I do not expect to explain sales values perfectly. We will be satisfied if we can achieve a lower coefficient of variation for sales of vacant land and land on the eve of demolition than the assessor achieves for total property value. That should establish that land taxation can work with at least as much administrative accuracy as the present property tax. If someone would like to bet that we cannot achieve that hoped-for accuracy, I will bet $100, even money, that we can do it, with a formula to be delivered on May 1, for downtown Portland sales for the subsequent year.

As for the elasticity of the supply of land, that is the wrong question. The question should be, "Is it possible to devise a non-contingent tax, a tax under which people do not increase their tax liabilities by being more productive?" I see no reason for supposing that we cannot.

To address the question of the extent to which land values and GDP could be expected to increase if taxes were shifted from labor and capital to land, Florenz and I are working on a computational general equilibirum model of the economies of the 48 continental state of the U.S. plus the District of Columbia. The model has four factors of production (land, labor, machines and structures), 20 industries and 8 consumer groups. The model will predict the effects of tax changes in individual states on economic activity in all states and on aggregate excess burden.

Incidentally, Florenz, having finished his Ph.D. under my supervision a year and a half ago, is now on the job market. He is an outstanding programmer and practical econometrician, a specialist in public finance and urban economics, and a very good teacher as well. If you would like to interview him in New York, you can reach him at florenz@vt.edu.