Why landscape tree prices are hard to predict

Lag, lumpy demand, and private quotes make the landscape tree market resist forecasting. The honest response is to measure it anyway and report the uncertainty.

Ask three nurseries for the price of a zelkova with a 12-centimeter trunk and you will get three numbers, none of them wrong. This is not noise to be averaged away. It is the structure of the market.

Three things make landscape tree prices resist prediction. The first is lag. A tree sold this year was planted five to ten years ago, so today’s supply answers a question asked a decade ago, and today’s planting decisions answer a question about 2035. The second is lumpiness. Demand arrives in projects: a park, an apartment complex, a road. One large order clears a region’s stock of a species and moves the quoted price for a season. The third is opacity. Transactions happen bilaterally, quotes are private, and the standard measurements of trunk diameter, height, and crown width hide most of what determines value, such as form, root condition, and transplanting history.

The patented approach we work from takes the only honest route through this. Assemble per-species records of price, sales volume, cultivated area, production volume, and weather, and let a model learn which of them move together. The interesting result is usually not the forecast itself but its confidence interval. When the interval is wide, that is the market telling you that the data does not yet exist.