توضیح نفوذ آرام فن آوری های صرفه جویی در انرژی ; مدل محصول همراه با تبادل در تنوع ویادگیری به وسیله استفاده

This paper studies the adoption and diffusion of energy-saving technologies in a vintage model. An important characteristic of the model is that vintages are complementary: there are returns to diversity of using a mix of vintages. We analyse how diffusion patterns and adoption behaviour are affected by complementarity and learning-by-using. It is shown that the stronger the complementarity between different vintages and the stronger the learning-by-using, the longer it takes before firms scrap old vintages. We argue that this is a relevant part of the explanation for the observed slow diffusion of energy-saving technologies. Finally, we show that an energy price tax reduces energy consumption, because it speeds up the diffusion of new energy-saving technologies and induces substitution from capital to labour.

Concerns about global climate change associated with the combustion of fossil fuels urge a call for the development and widespread adoption of energy-saving technologies. It is beyond doubt that the development of new energy-saving technologies—often labelled with the subsequent phases of invention and innovation—plays an important role in meeting policy targets with respect to the stabilisation or reduction of greenhouse gas emissions. However, the diffusion of existing technologies is at least equally important, costly and difficult, as the development of new technologies ( Jovanovic, 1997). It has indeed been shown that the widespread adoption of existing energy-saving technologies could enable a significant reduction in energy use, especially in the short and medium run ( de Beer, 1998 and IWG, 1997). At the same time, however, it is known that diffusion of new technologies is a lengthy process, that adoption of new technologies is costly and that many firms continue to invest in old technologies. The latter phenomenon is known as the energy efficiency paradox: the existing gap between the most energy-efficient technologies available at some point in time and those that are actually in use ( Jaffe and Stavins, 1994 and Jaffe et al., 1999). The aim of this paper is to contribute to our understanding of adoption behaviour of firms and of diffusion processes of new energy-saving technologies in order to improve our understanding of the observed slow diffusion of energy-saving technologies. The question why firms do not exclusively invest in the newest technologies has already achieved much attention in the literature. We can distinguish four major explanations for the relatively slow diffusion of new technologies. The first explanation is that the combination of uncertainty and some degree of irreversibility in investment creates an option-value of waiting (Balcer and Lippman, 1984, Dixit and Pindyck, 1994 and Farzin et al., 1998). The second explanation stresses strategic issues: in a world characterised by spillovers and limited appropriability, the presence of (expected) rival innovation and imitation creates an incentive for firms to postpone innovation or adoption (Kamien and Schwartz, 1972 and Reinganum, 1981). The third explanation highlights the fact that over time the performance of existing technologies improves and their price reduces due to learning-by-doing and spillover effects (Jovanovic and Lach, 1989 and OECD/IEA, 2000). A final explanation emphasises the role of vested interests. As switching to new technologies (temporarily) reduces expertise and hence destroys rents associated with working with relatively old technologies for particular subgroups in the economy, these groups may engage in efforts aimed at keeping the old technologies in place (Canton et al., in press, Krusell and Rios-Rull, 1996 and Mokyr, 1992). In this paper, we offer two additional explanations for the slow diffusion of energy-saving technologies. The first explanation is rooted in a complementarity effect and the second in a learning-by-using effect associated with the use of the technology. We argue that complementarity (or, alternatively, imperfect substitutability) is not so much a by-product of past investment decisions, but an essential ingredient of the process of technological change. It is evident that at the macro level there is continuous investment in both old and new technologies. Similar patterns exist at the sector or even at the firm level, depending on the technology and the type of production process. Many new technologies pass through a life cycle, in which they initially complement older technologies, and only subsequently (and often slowly) substitute for older technologies. A number of historical examples, like the replacement of the waterwheel by the steam engine or the diffusion of new types of processes in the iron and steel industry, illustrates the relevance of complementarities in this ‘life cycle view’ of technological change (Rosenberg, 1976, Rosenberg, 1982 and Young, 1993b). One can argue that modern production processes consist of even more interrelated and mutually reinforcing technologies than the documented historical examples. Hence, the production process may be seen as a puzzle of a large number of interdependent and thus complementary technology pieces, making it profitable for firms to continually invest in improvement of distinct pieces instead of replacing the whole puzzle at once (Antonelli, 1993, Antonelli et al., 1990, Berg and Friedman, 1977, Colombo and Mosconi, 1995, Jovanovic and Stolyarov, 2000 and Milgrom et al., 1991). In sum, the complementarity effect emphasises the multi-dimensional character of quality. Technologies differ not only in terms of their productivity (the vertical dimension), but also with respect to other qualities as a result of which firms face returns to diversity (the horizontal dimension). A good example of the latter can be found in the electricity sector where hydropower, unlike thermal power, is characterised by relatively high fixed costs (system of water-inflow and reservoir capacity) and low variable costs (expanding output capacity). Von der Fehr and Sandsbråten (1997) argue that because of this quality (and given the fact that output can only be stored at high costs or not at all), efficiency requires the use of a mix of hydro and thermal technologies, of which hydro is used in particular to meet the need for peaking-load capacity. Other benefits from using complementary technologies arise, for example, from flexibility with respect to inputs (different technologies use different types of fuels or raw materials like, for example, iron ore and scrap in steel production), plant location, plant size, or required managerial and organisational skills ( Rosenberg, 1982). The model that we develop in this paper captures these phenomena in a stylised way. The slow diffusion of technologies that is associated with the taste for technological diversity can be further intensified by learning-by-using effects; our second explanation for the slow diffusion of energy-saving technologies. In accordance with broad historical evidence (Mokyr, 1990, Rosenberg, 1982 and Young, 1993a) we allow for new technologies to be initially inferior to more mature technologies (along the vertical dimension). Learning-by-using improves the productivity of the new technology over time. Therefore, a switch of technologies temporarily reduces expertise and thereby the productivity level of the capital stock. The prospect of such a temporarily productivity drop prevents agents from immediate and ‘total’ switching, but rather induces a gradual adoption of new technologies resulting in co-existence of old and new technologies. We explore the above mentioned ideas by developing a simple two-sector macroeconomic model including a final goods sector, producing a homogenous consumption good, and a capital production sector producing heterogeneous vintages. The distinctive features of our model are as follows.1 First, technology is embodied in physical capital. Newer vintages need less energy and labour to produce the same output than older vintages. This reflects the vertical dimension of quality. Second, different vintages are imperfect substitutes in production. The economy exhibits a ‘taste for diversity’ of vintages, creating an incentive to invest in both new and older technologies. This reflects the horizontal dimension of quality. Third, our model allows for the endogenous determination of the number of vintages used in the final goods sector, so we offer an economically motivated approach for the scrapping of vintages. Fourth, the representative firm in the final goods sector gains expertise in a technology by using the technologies in its production process. In other words, we include learning-by-using.2 There are a number of related articles in which issues of learning and technological innovation and diffusion are analysed. Without extensive discussion, we refer to, for example, Aghion and Howitt (1996), Aghion et al., 1997 and Aghion et al., 1999, Arrow (1962), Chari and Hopenhayn (1991), Jovanovic and Nyarko (1996), Parente (1994), Stokey (1988) and Young, 1993a and Young, 1993b. We refer to Jaffe et al. (2000) for an extensive survey of the literature on environmental or energy-saving technological change. The main differences between these articles and ours lie in our focus on energy-saving technological change, the specifications we make with respect to the complementarity of vintages, the interpretation of intermediate goods as technologies, the emphasis on the complementarity effect in diffusion processes instead of innovation processes (cf. Young, 1993b) and the endogenous scrapping mechanism. In focusing on diffusion and adoption, our paper differs from most literature on induced or endogenous technological change in which the focus is almost solely on innovation (Goulder and Schneider, 1999 and Goulder and Mathai, 2000). The paper is organised as follows. In Section 2 we set up the basic model. In Section 3 we derive the solution of the model. Section 4 illustrates the properties of the model by providing some comparative statics on the complementarity effect and the learning-by-using effect. Section 5 brings together the gained insights to address energy-saving technological change in more detail, including an analysis of the effects of imposing an energy price tax. Section 6 concludes.

The widespread adoption of energy-efficient technologies is a lengthy and costly process. In this paper, we developed a vintage model to study the diffusion of energy-saving technologies and to explain why diffusion is gradual and why firms continue to invest in old technologies. An important characteristic of our model is that vintages are complementary; there are returns to diversity of using different vintages. We have argued that this is a potentially relevant part of the explanation of the energy efficiency paradox. Furthermore, we showed that this effect is intensified when we take learning-by-using effects into account. A firm faces loss of expertise on a particular vintage, gained by virtue of experience, when it switches to a newer vintage and this provides an extra argument for firms to invest in older vintages. We show that an increase in energy efficiency speeds up scrapping of older technologies and induces an increase in the demand for vintage capital, since it makes the capital–energy composite cheaper. The former reduces the demand for energy whereas the latter increases the demand for energy since we assume no substitution possibilities between capital and energy. Finally, we show that an energy price tax speeds up the diffusion of new energy-saving technologies and induces substitution from capital to labour, both decreasing the demand for energy. Therefore, in our model the development in total energy consumption depends on the relative magnitude of the energy price tax and the exogenous energy efficiency improvement.