Nordhaus finds that the optimal policy, given the current state of knowledge and ignoring
uncertainty, is to do very little in the short term (a $6/ton tax in 2005) and only
marginally more in the long term. While his mild long-term policy suggestion helps to
inform the basis of his short term recommendations, Nordhaus only intends to propose a
short-term strategy and emphasizes that policies should change as we learn more.

Recognizing wide-ranging uncertainties and the possibility of calamity, Nordhaus investigates optimal policy choices over a reasonable distribution of possible states of the world. He shows that while the median optimal carbon tax is on the order of $6/ton in 2005, the mean optimal carbon tax is somewhat larger, at $18/ton in 2005.

His conclusions are viewed with skepticism by some critics, who believe some or all of the following: he overstates control costs, he understates control benefits, and his choice of a positive pure rate of time preference is indefensible. Others find his results and methodology more robust.

Introduction

In a 1994 book', Nordhaus calculates an optimal policy response to the problem of climate change. A distinguishing feature of this work is that it aims to uncover a truly optimal policy—that is, one that maximizes net benefit as opposed to one that merely minimizes the cost of achieving an ad hoc CO2 concentration target.

Basic Results

Ignoring uncertainty and with a best guess of costs and benefits (based on a wide purview of existing studies) Nordhaus finds that the optimal policy is to do close to nothing. With the optimal policy, carbon taxes rise from $6 per ton in 2005 to $20 per ton (1989 dollars) in 2100. Global average temperature under this policy increases 3.2°C-only 0.2°C less than in the baseline case. On net, the world economy under this scenario enjoys a small net benefit relative to the baseline case. See Table 1.

Restricting emissions of greenhouse gases to 1990 levels, on the other hand, inflicts relatively large net costs. Carbon taxes rise to $400/ton in 2100, and the increase in global average

"Managing the Global Commons," which presents analysis based on Nordhaus's DICE model. It is the culmination of arguments made in earlier papers.

temperature is limited to only 2.4°C — a full degree less than in the baseline.

Given that the optimal policy is much closer to doing nothing than to restricting to 1990 levels, it appears that literally doing nothing in the short term is superior to current proposals oriented on 1990 emissions levels.

It is worth emphasizing that Nordhaus's result is not simply a matter of delaying controls. While the typical cost-minimization exercise also shows that carbon emissions should be allowed to follow a near-baseline path at first, Nordhaus's result goes further: he finds that relatively little should be done even over the long term. In his optimal scenario, carbon concentrations are allowed to grow without apparent bound.

Incorporating uncertainty, Nordhaus's results change though the basic message of modest shortterm measures does not.

Nordhaus examines how the optimal policy changes when key parameter values vary from his best guess and finds that the mean optimal policy calls for higher carbon taxes than he finds for the best-guess case. The mean optimal policy calls for a carbon tax of $18/ton in 2005 and $53/ton in 2045.

The "mean" result differs from the "best guess” result because the distribution of outcomes is skewed so that favorable climate change scenarios are similar to the bestguess case and occur with high probability while extremely unfavorable scenarios occur with low probability.

Nordhaus finds that only 2% of possible scenarios -- the least favorable -- would call for a carbon tax in 1995 of over $100/ton.

2Source for Tables 1 and 2: Nordhaus, “Managing the Global Commons”, 1994. These data come from the PICE model.

"Net benefits are measured relative to the base case of taking no action.

Basis for Results

Nordhaus's results derive from three key assumptions:

that the costs of reducing carbon emissions are relatively high;

that the benefits of carbon control (in the form of averted damages) are relatively low; and

that the social rate of discount is greater than zero.

Critics have attacked Nordhaus on all three points, and it is worth examining each point in greater detail.

1. Costs

The cost of reducing carbon emissions is determined by the responsiveness of the economy to a given level of carbon-control effort. For example, if a $10 carbon fee elicits a large response in terms of lower carbon emissions, the cost of carbon control will be low. Conversely, if a $10 carbon fee elicits a small response, program cost will be high.

The impact on the economy of a carbon-control program in Nordhaus's model is determined by two key assumptions one concerning the "static" response, given current technologies and carbon use, the other concerning the “dynamic" response, reflecting induced changes in carbon use and advances in technology.

• Nordhaus bases his assumption about static responses on existing empirical studies. His assumptions here are not particularly contentious.

Nordhaus attempts to capture dynamic responses by assuming that technology will improve at about the same rate in the future as it has done in the past. Importantly, the trend rate of improvement is not allowed to vary depending on the price of carbon permits or the level of a carbon tax. Critics find fault with this approach, since his model thus assumes that carbon taxes will in no way accelerate the development of non-carbon technologies — in stark contrast to the "technologist" view.

Nordhaus notes that his results are sensitive to assumptions about technology, but he does not document the extent of that sensitivity in isolation from other uncertainties.

2. Benefits

The benefits of carbon control are determined by the magnitude of damages averted -- that is, damages that would have occurred in the absence of a control program. Nordhaus calculates that these damages would be relatively modest. For example, he estimates that a 3°C temperature increase results in only,a 1.3 percent loss of output per year.

Some critics feel that damages averted will be greater than Nordhaus assumes. They note that SO2 and particulate emissions will be reduced, biodiversity will be better maintained, and negative health effects will be avoided.

In a survey of experts, Nordhaus finds that the median cost prediction for a 3°C rise in temperature is 1.9 percent of output annually -- 40% higher than his own. The mean prediction is much higher, at 3.6 percent.

3. The social rate of discount

Given that costs are borne up front and benefits accrue mainly in the distant future, assumptions about the social rate of discount are crucial. Nordhaus assumes that the social rate of discount is 3 percent per year, this rate is sufficiently large as to significantly reduce the present value of the distant and already modest benefits.

Many critics (mainly non-economists) feel that a social rate of discount greater than zero is indefensible on ethical grounds. However, mainstream economists strongly endorse the approach to discounting which Nordhaus uses and find it difficult to understand very low rates of time preference in the context of observed behavior.

Nordhaus re-evaluates the optimal policy when the pure rate of time preference is only 1 percent and finds that the optimal carbon tax in this case is considerably greater (see Table 2).

Nordhaus's analysis implies that little should be done either in the short term or in the long term. His results hinge on assumptions about technology improvements, appropriate discount rates, and the size of potential damages. Critics feel that Nordhaus may be:

overestimating mitigation costs because he ignores the possible influence of higher
energy prices in spurring the development of new technology; and

underestimating damages from climate change.

While both would tend to cause Nordhaus to be too lax in his policy recommendations, it is not clear that such criticisms can be sustained far enough to overturn Nordhaus's basic point that strict action is inappropriate in the short term. The discount rate issue, while crucial, is difficult to settle.