Abstract Aggregate timber supply by ownership was investigated for a small region by applying stand-level harvest choice models to a representative sample of stands and then aggregating to regional totals using the area-frame of the forest survey. Timber harvest choices were estimated as probit models for three ownership categories in coastal plain southern pine stands of North Carolina using individual permanent and remeasured stand-level data from last two available USDA Forest Service Forest Inventory and Analysis (FIA) surveys. The timber harvest decision was modeled as a function of timber values, a cost factor, and stand volume as a proxy for nontimber values. Probit models were statistically significant at 1% for all ownerships. Area expansion factors (the portion of forest area in the region represented by the sampled stand) were then combined with harvest probabilities to model the aggregate effects of price changes on timber supply, given a fixed forest area. Implied price elasticities were estimated using this modeling of aggregate effects, and a bootstrapping procedure was applied to estimate confidence limits for supply elasticities with respect to price. Our results showed that NIPF and industry were elastically responsive in the aggregate when price increases are perceived as temporary but much less elastically and usually negatively responsive when increases are perceived as permanent. Results are consistent with theory of optimal rotations and highlight the critical influence of both existing inventory structure and expectations on aggregate timber supply. FOR. SCI. 46(3): 377–389.