Most of the work in the environmental economics literature on the dynamic effects of policy instruments on technological change has been theoretical, rather than empirical, and the theoretical literature is considered first. The predominant theoretical framework involves what could be called the discrete technology choice model. In this, firms contemplate the use of a certain technology that reduces the marginal costs of pollution abatement and that has a known fixed cost (Downing and White, 1986; Jung et al., 1996; Malueg, 1989; Milliman and Prince, 1989; Zerbe, 1970).
While some authors present this approach as a model of innovation, it is perhaps more useful as a model of adoption.107 The adoption decision is one in which firms face a given technology with a known fixed cost and certain consequences, and must decide whether or not to use it; this corresponds precisely to the discrete technology choice model. Innovation, on the other hand, involves choices about research and development expenditures, with some uncertainty over the technology that will result and the costs of developing it. Models of innovation allow firms to choose their research and development expenditures, as in Magat (1978, 1979), or incorporate uncertainty over the outcome of research (Biglaiser and Horowitz, 1995; Biglaiser et al., 1995).
Several researchers have found that the incentive to adopt new technologies is greater under market-based instruments than under direct regulation (Downing and White, 1986; Jung et al., 1996; Milliman and Prince, 1989; Zerbe, 1970). This view is tempered by Malueg (1989), who points out that the adoption incentive under a freely allocated tradable permits system depends on whether a firm is a buyer or seller of permits. For permit buyers, the incentive is larger under a performance standard than under tradable permits.
Comparisons among market-based instruments are less consistent. Downing and White (1986), who consider the case of a single (sole) polluter, argue that taxes and tradable permit systems are essentially equivalent. On the other hand, Milliman and Prince (1989) find that auctioned permits provide the largest adoption incentive of any instrument, with emissions taxes and subsidies second, and freely allocated permits and direct controls last. Jung et al. (1996) consider heterogeneous firms, and model the market-level incentive created by various instruments. This measure is simply the aggregate cost savings to the industry as a whole from adopting the technology. Their rankings echo those of Milliman and Prince (1989).
On the basis of an analytical and numerical comparison of the welfare impacts of alternative policy instruments in the presence of endogenous technological change, Fischer et al. (1998) argue that the relative ranking of policy instruments depends critically on firms ability to imitate innovations, innovation costs, environmental benefit functions, and the number of firms that produce emissions.108 Finally, the study includes an explicit model of the final output market, and finds that it depends upon empirical values of the relevant parameters whether (auctioned) permits or taxes provide a stronger incentive to adopt an improved technology.
Finally, recent research investigates the combined effect of the pollution externality and the positive externality that results from learning-by-doing with mitigation technologies. Since the benefit from learning occurs after the learning has taken place, a dynamic analysis is needed. Some analyses shown that dynamic efficiency (discounted least cost, aggregated over time) requires that the incentive for emissions-mitigating innovations be set higher than the penalty on emissions, especially if account is taken of leakage. This is in contrast with the conclusions of comparative static analysis upon which most environmental policy analysis is grounded (e.g., Baumol and Oates, 1988), under which the two incentives should be equal in all time periods (for a formal analysis, see Read (1999, 2000)).
Other reports in this collection