tion of joint costs. For example, forest roads are a costly input to the production of many forest products, including timber, minor forest products, and recreation. Yet standard accounting practices, especially for the national forests, attribute the full cost of these roads to the timber program. As currently constructed, the NIPA include the costs of road construction, but exclude the benefits produced by the road.

To the extent that near-market forest products, such as fuel wood, berries, mushrooms, and Christmas trees, are produced by households but not purchased through markets, they would be included in the forest accounts.

The accounts would include the value of household production of activities such as hiking, hunting, and fishing. However, if there is uncongested, open access to the forest-based inputs needed for household production, the contribution of these inputs to household value on the margin is zero. Current practice often uses average rather than marginal values, so care must be taken, particularly for open-access forests, to ensure consistent valuation in order to prevent overvaluation of nonmarket activities.

Some of the impacts of forests are already included in the NIPA. For example, if forests moderate water flows and reduce the cost of agricultural production, this benefit is fully incorporated in the NIPA. Ascribing the benefit to the forest sector, while a difficult task, would be required for a full accounting.

At present, the only public goods that have been the subject of widespread attempts at valuation are those associated with carbon sequestration (Brown, 1996). While quantitative data on carbon sequestration are available, valuation is still highly uncertain. Moreover, because valuation of carbon sequestration is based on global benefits, the issue of how such benefits would be incorporated in a single nation's accounts is unresolved.

There are few comprehensive studies of the total value of forest products. Recent work on goods and services produced on public lands managed by the U.S. Forest Service indicates that more forestland value is due to recreational and wildlife services than to timber, mineral, and range goods (U.S. Department of Agriculture Forest Service, 1995). For example, of the estimated total $9 billion value of forest goods and services in 1993 (valued at market prices), recreational and wildlife services accounted for 80 percent, whereas the production of minerals and timber and grazing range services accounted for just 20 percent.

While the above estimates illustrate the importance of nonmarket