Nordhaus discussed efforts to model climate change response at the macro economic level. Economists estimate the economic effects of climate change by measuring impacts on consumption goods. They evaluate differences anticipated to take place in consumption over time and then compare costs and benefits, taking into account maximization of social welfare by including some social welfare function.

Impact analysis involves two distinct tasks: measuring the physical changes and valuing them. The really controversial part of impacts analysis is valuation. Impact analysis in practice is based on the value of market goods and services and the value of environmental (non-market) goods and services. Economists try to determine what one would have to compensate people for, in terms of market goods and services, for deterioration of the environment. Because this is a matter of individual preference, they aggregate over people and time to come up with an average response.

First, economists look for market or other experiments in which consumers are faced with different prices or quasi-prices and observe their choices. Estimates are made of what the changes will be and prices are assigned to value those changes. A price line is then fit to the observed choice curve. In cases where the factors are appropriately estimated, Nordhaus says, this technique offers a good approximation of the true loss in real income according to the preferences of the consumer.

How is the indifference curve or price line measured? First, major distinctions are made between market and non-market goods and services. Market goods and services include food and energy, while non-market goods include such things as public schools and parks. A major distinction is also made between private and public goods. In the case of private goods, individuals can appropriate all the benefits and others can be excluded, whereas in the case of public goods, others are not excluded from joint consumption and benefits accrue to a variety of people, for example, radio signals, coastal light houses, or defense measures. In general, it is possible to use individual behavior to infer preferences about private goods, but there is no reliable behavioral information about genuine public goods (such as culture, biodiversity or wet lands).

Direct measures are used by economists for marketed goods and services estimates. For farming, for example, economists estimate loss of output and use market prices to determine the loss. Alternatively, they use the direct measure of effects on land values. In measuring economic impacts of sea level rise, they can estimate the change in land in different categories or value by market land prices. When adaptation responses are included, Nordhaus says, economic losses due to sea level rise are reduced by two to three orders of magnitude, according to the work of Yohe.

Three techniques for estimating impacts of climate change are: production function approaches, which do not treat adaptation; time series approaches, such as standard studies of agriculture where trends in temperature, farm output and U. S. GNP are tracked; and hedonic approaches, where only impacts to the value of the underlying asset are considered.

There is a bias inherent in production function studies, Nordhaus says, illustrating with an example about how farmers might adapt to changes in climate by altering the crop varieties they plant or other techniques they could employ to increase their yields. Nordhaus explains that by leaving out such expected adaptation, production function studies assume "dumb farmer" scenarios, in which farmers do nothing to adapt to climate change and thus, losses due to climate change are overestimated. In a hedonic approach based on land values, only the impact to the land value is considered and so there is a disconnect with biological and other factors. In sum, Nordhaus says, most economic evaluations lead to the conclusion that the net economic effect of climate change is basically zero. There has been a study by Adams et. al., however, which shows significant losses, Nordhaus noted.

There are indirect measures for non-marketed private goods. For publicly-provided private goods (recreation, schools), we can examine effects of "quasi-prices" such as travel costs on behavior. To assess "amenity effects," we assume that labor is mobile and requires a "compensating differential" for unpleasant locations or tasks. Then we look at wage differentials across regions, jobs or climates and infer the value people place on amenities. Such an analysis has not yet been done for climate change, but Nordhaus expects that amenity effects will dominate all other effects (Viscusi's analysis on the value of human lives is based largely on wage differentials).

There are problems with determining values placed on public goods. Because of free riding, we cannot rely on private behavior to reveal preferences. One alternative is to examine voting behavior by legislatures to reveal some function of voter preferences. Another is to use preference-revelation mechanisms, but these seem to work better in theory and in the lab than in practice. A third alternative is to use "contingent valuation" which offers a direct statement of preferences, but is subject to the severe criticism that there is no budget constraint or cost of not revealing genuine preferences. Nordhaus does not believe contingent valuation gives reliable estimates for these reasons.

Nordhaus offered a few suggestions as to why nobody believes the economist's approach to impact analysis. At the core, there is a disagreement on what the impacts of climate change, or other environmental problems, might be. In principle, he says, this is resolvable by further research. However, there are also deep-seated differences in outlook on mechanisms such as adaptation, use of technology, and the price system. For example, it is often said that economists don't understand the complexity of ecosystems or how vulnerable human systems are to changes in natural systems.

Nordhaus pointed out basic disagreements concerning values. Whose values should be adopted: individuals, George Bush's, Al Gore's, Joe Lumberjack's? And how should we treat distribution of income? Should we accept the current market's distribution? Should we try to change it? How should the claims of future generations versus those of the present, or North versus South claims be dealt with? Should certain items be sacrosanct and fall outside the market place? Where do we draw the line around the market? Should environmental quality be absolute or balanced against costs? Who is included in the social welfare function? Do non-human life forms count?

There are a number of especially difficult issues that arise when attempting to assess economic impacts of climate change. One of these involves the fact that people's tastes are endogenous, and are subject to change over time, and therefore, we may overestimate (or underestimate) future impacts based on current tastes. Similarly, technologies are endogenous, but we generally think of the future in terms of current technologies as we can not predict future technologies. Other difficult issues involve an endogenous population and shifting national boundaries, as well as whether we view nature as dumb or benign (trees) versus smart or competitive (bugs).

Nordhaus says there is a good deal of confidence among economists that climate change will not have large effects on the market economies of the developed world. He estimates these impacts between 0 and 1/2% of GDP, plus or minus 1%. But with regard to non-market goods in the developed world and both market and non-market goods in the less developed world, these are all areas in question where no significant research has been performed.