The likelihood that the Kyoto Protocol will achieve significant real reductions in greenhouse gas emissions is very low. The United States withdrew from the protocol in March 2001, a move which was angrily denounced by surprised commentators in Europe and around the world. It was described as arrogant, isolationist, and a "betrayal [by the Bush Administration] of their responsibilities as global citizens". Yet the announcement was really nothing more than a blunt public acknowledgment of a fact that was well known within the policy community: the Kyoto Protocol was already dead in the United States. The U.S. Senate, which must ratify all international treaties by a two-thirds majority, overwhelmingly opposed the protocol and had voted 95-0 against U.S. participation as early as July 1997, five months before the protocol was signed. Opposition was so great that the Clinton Administration, which negotiated and signed the protocol, never bothered to submit it to the Senate for ratification. Even if the Bush Administration had enthusiastically supported the treaty—which it did not—there was little it could have done.

What doomed the protocol in the Senate is a critical flaw in its design: it requires each participating industrialized country to agree to achieve a specified emissions target regardless of the cost of doing so. The focus on rigid targets also makes the treaty impractical as a long-term climate policy for the rest of the world as well. Because the costs of reducing emissions are unknown and could be very large, countries with substantial emissions have insisted on increasingly lax targets as a condition for their continued participation. Japan, Canada and Russia, for example, were able to negotiate large increases in their "sink" allowances during COP6bis, held in Bonn, and COP7, held in Marrakesh. Between the U.S. withdrawal and the increase in sink allowances, the protocol has been relaxed substantially. The effect on estimated emissions permit prices in the 2008-2012 period is dramatic. Relative to the original Kyoto agreement, permit prices are likely to be reduced by 14 percent (Bohringer, 2001) to 85 percent (Kemfert, 2001).

In this paper we update our earlier estimates of the cost of the Kyoto Protocol using the G-Cubed model, taking into account the new sink allowances from recent negotiations as well as allowing for multiple gases and new land clearing estimates. We then move on to the paper's main topic, which is an examination of the protocol's core flaw – the sensitivity of compliance costs to unexpected changes in future economic conditions. To explore this point we consider two plausible alternative assumptions about a single aspect of the future world economy and evaluate the protocol under each assumption. The two cases we examine are: (1) moderate productivity growth in Russia sufficient to produce a 3.24% average rate of economic growth from 2000 to 2012; and (2) higher productivity growth leading to a 4.24% average rate of economic growth over 2000 to 2012. Comparing the two sets of results shows the sensitivity of greenhouse gas abatement costs to variables that are very difficult to predict. The cost of the protocol in foregone GNP, for example, varies by up to 50 percent. The comparison thus illustrates the risks created by the protocol's focus on rigid emissions targets.

We also evaluate an alternative climate change policy described in McKibbin and Wilcoxen (1997a,b and 2002a,b) under both Russian growth scenarios. The alternative policy, hereafter called the "Blueprint", differs sharply from the protocol because it does not impose fixed emissions targets. Its design, which will be discussed in more detail below, takes explicit account of the vast uncertainties surrounding climate change. As a result, the policy's costs are predictable and are affected very little by the change in Russian economic growth.9 Predictability of costs is essential because a climate agreement will have to remain in force indefinitely: unexpected future increases in compliance costs would place enormous pressure on governments to abrogate the agreement. At the same time, the stability of costs does not necessarily compromise the policy's effect on emissions: because costs are limited and predictable, participation in the policy would be likely to be broader than under the Kyoto Protocol, and reductions would be undertaken sooner. To illustrate this effect we compare cumulative emissions under both policies through the year 2015. Although the Blueprint is less stringent than the protocol, with wider participation (by the United States, in particular), it can actually have a substantially larger effect on emissions. Overall, the Blueprint is a promising alternative to the Kyoto Protocol.

The remainder of the paper is structured as follows. In section 2, we give a brief overview of the Kyoto Protocol and highlight the quantitative changes that have occurred to the effective targets facing countries from COP3 to COP7. In section 3, we present a brief critique of the protocol and describe the alternative Blueprint proposal. Section 4 gives an overview of GCubed, the multi-country, multi-sector general equilibrium model we use to evaluate each of the policies. Section 5 discusses the key elements of our baseline simulation and presents our results. The baseline is particularly important because the rate at which emissions grow in the absence of a climate change policy is a major determinant of the cost of the Kyoto Protocol: the more emissions rise in the baseline, the more they must be reduced to hit the protocol's targets. Major world economic events that have occurred since our previous analysis of the Kyoto Protocol (McKibbin and Wilcoxen, 1999)—including the Asian financial crisis and the global recession of 2001-2002—are reflected in the baseline. Following discussion of the baseline, the remainder of section 5 evaluates and compares the Kyoto Protocol and the Blueprint. Our conclusions are presented in section 6.