تاثیر ابزار سیاست های زیست محیطی بر زمان استفاده از فن آوری های صرفه جویی در انرژی

One of the main considerations in designing environmental policy is the impact of policy instruments on the timing of firms’ investment decisions with respect to energy-saving technologies. This paper analyzes the impact of environmental taxes and quotas on the timing of adoption when (i) the rate at which new, improved energy-efficient technologies become available, is uncertain, and (ii) the investment decision is (at least partially) irreversible. We find that neither policy instrument is unambiguously preferred to the other when it comes to stimulating early adoption of new technologies.

Much attention has been paid to the potential of public policies to reduce environmental pollution, such as emissions of carbon dioxide. Conventional wisdom is that, from an efficiency perspective, market-based instruments are preferred over command-and-control instruments. Taxes are an example of such market-based instruments, and their alleged superiority is based on the notion that they equalize marginal abatement costs across firms and hence yield statically efficient outcomes (Baumol and Oates, 1988). In addition, taxes are believed to be more effective in inducing technological change than command-and-control instruments as they offer a permanent incentive to use less of the environmental commodity and thus save on tax expenditures (Downing and White, 1986 and Newell et al., 1999). Kneese and Schulze (1975) have argued that “over the long haul, perhaps the most important single criterion on which to judge environmental policies is the extent to which they spur new technology toward the efficient conservation of environmental quality”. The case for taxes thus appears strong, both from a static and a dynamic efficiency viewpoint. Whereas to date the literature has focused on calculating the willingness to pay for one particular new technology under different policy regimes (e.g., Milliman and Prince, 1989, Jung et al., 1996 and Requate and Unold, 2003), we analyze the impact of the choice of policy instruments on the timing of the investment. We assume that new generations of energy-saving technologies become available at unknown future dates, and define the “adoption lag” as the expected number of periods that elapses until a firm purchases a new technology. We analyze whether taxes or command-and-control instruments (specifically, non-tradable quotas) are more conducive to early adoption by comparing the length of the adoption lag under the two policy regimes. Governments may care about early adoption for at least two reasons. First, governments may be bound by international agreements (think of the Kyoto protocol) that have specific time horizons for compliance, and therefore be interested in immediate reductions in the energy intensity of production. Second, governments may have a short planning horizon, as predicted by political economy models ( Mueller, 1993). Our model is simple in many respects. We assume that the rate of invention (the arrival of new technologies) is exogenous.1 Technologies produce output, using two variable factors of production, labor and energy. New technologies are of the energy-augmenting type but, upon equating total use of energy to total emissions, we can also say that technical progress reduces the pollution intensity of production. Firms may decide to invest and apply the new technology henceforth, or they may decide to forego the option to invest and await better technologies expected to become available in the future. Due to the fixed costs associated with purchasing and installing new machinery, irreversibility plays an important role in the investment decision as the firm will regret having adopted a specific new technology if an even better technology arrives shortly after adopting (Dixit and Pindyck, 1994). We take an extreme view on irreversibility by assuming that firms can purchase an energy-efficient technology only once, but the qualitative results of the model spill over to other cases where investments are at best partially reversible (for example when there are scrap markets for obsolete technologies) or when firms can invest more than once. This paper aims to fill a gap in the environmental economics literature by focusing on the impact of instrument choice on the adoption of new technologies, if technological progress is expected to be ongoing. The effectiveness of policy instruments in inducing adoption have been addressed before, but always with respect to a single, specific new technology (e.g., Milliman and Prince, 1989, Jung et al., 1996, Malueg, 1989, Requate, 1995 and Requate and Unold, 2003). The importance of irreversibility and uncertainty in these adoption decisions has been emphasized in several papers. Irreversibilities considered are with respect to irreparable environmental damage (e.g., Arrow and Fisher, 1974 and Pindyck, 2000), with respect to sunk investment costs (Requate, 1995, Requate and Unold, 2003 and Pindyck, 2000), or both (Kolstad, 1996 and Pindyck, 2002). The types of uncertainties analyzed are with respect to environmental damage or with respect to the net benefits of adoption (e.g., Dosi and Moretto, 1997 and Phaneuf and Requate, 2002). To our knowledge, however, this paper is the first to address the consequences of policy instrument choice for investment if new technologies arrive stochastically over time. This paper has in common with the cited environmental economics literature that it abstracts from many of the adoption barriers that may induce managers to postpone or even abstain from investment. We assume that the decision whether or not to adopt a specific technology depends only on the comparison of the benefits and costs of adoption, taking into account the possibility of regretting having adopted if a new and better technology arrives soon after having made the decision. In reality, investment behavior also depends on the attitudes of the managers involved (Riemenschneider et al., 2003), the characteristics of the technology in question (especially its flexibility with respect to adapting to – unforeseen – changes in market circumstances; Krishnan and Bhattacharya, 2002 and Li et al., 2003), and the key characteristics of the firm including firm size, organizational structure, and age (cf. Patterson et al., 2003 and McGee and Sawyerr, 2003). As our paper abstracts from these factors, its approach is similar to the management literature on the adoption of new technologies when even better technologies are expected to arrive stochastically over time.2 Here, papers also focus on the bare bones of the investment decision problem by leaving out many of the complicating factors mentioned above; see for example Hinomoto (1965), Li et al. (2003), Rajagopalan (1999), Schwartz and Zozaya-Gorostiza (2003). Rather than using scenario analysis or assuming that technological progress can be modeled as a Markov process – as some of these papers do – we employ the model developed by Farzin et al. (1998) in which technological progress is represented by a Poisson process. The advantages of this approach are increased realism and its mathematical tractability. Thus, we abstract from many of the adoption barriers that may affect the investment decisions of firms because of tractability considerations. By abstracting from them, we implicitly assume that whereas these factors undoubtedly affect the actual timing of investment, the ordering of the instruments’ impact on the investment timing is unaffected. Whether indeed this assumption holds true in practice, is left for future research. The instruments we focus on are taxes and non-tradable quotas, and hence we do not consider the dynamic incentives associated with a tradable permit system. The reason for this is that in terms of the speed of adoption, tradable permits are strictly dominated by taxes. The argument is as follows. If the tax rate and permit price are equal, the costs of using an additional unit of energy are the same in the two policy regimes, and hence also the optimal amount of energy used is the same for a given technology (cf. Milliman and Prince, 1989 and Malueg, 1989). That means that the benefits of adopting new technologies are also equal under taxes and tradable permits, and hence firms end up adopting the same new technology in these two regimes. 3 However, as argued by Requate and Unold (2003), the assumption of equal permit prices and tax rates may hold at some point of time, but is unlikely to hold in all future periods. Whereas the tax rate may be kept constant, the permit price decreases over time as more firms invest in new technologies, and hence reduces the incentives for other firms to also invest. Determining the optimal adoption lag under tradable permits is difficult because of the strategic problems associated with solving the adoption strategies of a set of (heterogeneous) firms. 4 However, combining the fact that (i) taxes and tradable permits result in the same technology being adopted in case the permit price equals the tax rate and (ii) the permit price gives less strong incentives for investment than taxes as it is likely to decrease over time, we can infer that the adoption lags under taxes are a lower-bound estimate of the adoption lags in case of tradable permits. So, to avoid the difficulties of analyzing strategic interaction in the permit market, we contend that tradable permits are strictly less effective in inducing adoption of new technologies than taxes, and hence limit our analysis to the comparison of the dynamic incentives of taxes and non-tradable quotas. The set-up of the paper is as follows. In the second section we introduce the model, and the optimal investment lag under taxes and non-tradable quotas is determined in the third. In the fourth we analyze how these adoption lags are affected if environmental stringency increases. The fifth section concludes. Our results can be summarized as follows. Although it is impossible to obtain closed-form solutions for the adoption lags and hence have to rely on numerical simulations, we find that the ranking in policy instruments in terms of the speed of adoption is typically ambiguous. Depending on whether the required level of environmental stringency is high or low, taxes outperform non-tradable quotas, or vice versa. More specifically, if the government wishes to induce early adoption and the required level of environmental stringency is fairly low, quotas are the preferred instrument, whereas taxes should be applied if more stringent environmental policies are called for.

The choice of environmental policy instruments has received a substantial amount of attention in the literature, focusing on differences in both static and dynamic efficiency. This paper contributes to the literature by analyzing the impact of taxes and quotas on the time firms spent waiting for the arrival of new energy-saving technologies. Two main conclusions are that (i) increased environmental stringency does not necessarily induce early adoption—in fact, our numerical analysis suggests that the opposite is more likely to hold, and (ii) that there is no unambiguous ranking of policy instruments in terms of the length of the adoption lag. Our finding that the adoption lag is increasing in environmental stringency for a very wide range of parameters is interesting because it indicates that, contrary to conventional wisdom, higher taxes will not unambiguously promote the use of new technologies by firms. We treat invention as exogenous (a simplification, no doubt) but find that firms will postpone their decision to adopt a new technology and thereby forego the option to lower the energy intensity of production for a longer period. This may be undesirable for governments interested in producing environmental benefits in the short run, be it for electoral purposes of because they are bound by international agreements. The finding that the adoption lag for the case of taxes may be greater or smaller than the adoption lag under a quota regime casts new light on the conventional wisdom that market-based instruments are to be preferred over command-and-control measures on dynamic efficiency grounds. Jung et al. (1996, p. 95), for example, write that “encouraging … market penetration of advanced pollution abatement technology is an important environmental policy objective”. Depending on the initial conditions, we find that the adoption under taxes may be faster or slower than under quotas. Specifically, for relatively lax environmental policies, quotas yield shorter adoption lags than taxes. Hence, if governments wish to speed up the implementation of new technologies under such conditions, they may prefer quotas to taxes. If, on the other hand, a more stringent environmental policy is considered and the government worries about the implementation of new technologies, taxes are to be preferred on these grounds. Our theoretical model shows that the ordering of policy instruments in terms of inducing adoption is not unique, and identifies the relevant parameters needed to determine the critical level of stringency at which the instrument ordering switches. Actually identifying this critical level is an empirical question, especially because it is likely to be sector- or even technology-specific. The conclusions of this paper do not have direct welfare implications in the sense that either taxes or quotas are the preferred instrument given the level of environmental stringency required. One reason is that, from a welfare perspective, a longer adoption lag is not necessarily inferior to a shorter adoption lag. Although the firm will continue to use the “old” (energy-inefficient) technology for a longer period of time, the technology eventually adopted will also be better and have a lower pollution intensity (see also Malueg, 1989). This means that the repercussions in terms of the total (discounted) amount of energy used, is an empirical question as well. To obtain tractable analytical results, our model is simple as it abstracts from many of the factors that affect the adoption behavior of firms such as the lack of managerial skill, imperfect capital markets, the fact that technologies can be evaluated along various dimensions (including not only energy efficiency but also, for example, quality of output or ease of use), etc. By abstracting from these, we ignore possible interactions between these other factors and the impact of uncertainty and irreversibility. To what extent such interactions exist and whether they affect our policy conclusion with respect to the ambiguous ranking of policy instruments, is left for future research.