مدلسازی پذیرش تکنولوژی به عنوان یک سرمایه گذاری غیر قابل برگشت تحت عدم قطعیت: مورد ،صنعت عرضه برق ترکیه

This paper studies energy conversion technology adoption in the electricity supply sector from the perspective of irreversible investments under uncertainty and with a particular interest in environmental sustainability. We develop a dynamic technology adoption model that is firmly rooted in economic theory and that takes important determinants of optimal investment in available technologies (e.g., life cycle capital and operation cost) explicitly into account. Uncertainty is introduced for the demand for peak-load capacity, unit generation costs, and for the average electricity price. We test the model empirically by applying it to data for the Turkish power supply industry. The model-guided optimal investment schedule based on net present value considerations exhibits significant deviations from the actual investment outcome. We find that the increased adoption of natural-gas-fired power generation technologies in Turkey in recent years, while contributing to environmental sustainability, has had doubtful merits from an investor's perspective.

Increasing concern about the adverse socioeconomic and environmental impacts of current energy use patterns, in many cases coupled with staggering levels of fossil fuel import dependence, call for substantial changes in the energy technology and fuel mix towards a more sustainable energy supply system. Technology adoption and diffusion models (e.g., see Thirtle and Ruttan, 1987 and Sarkar, 1998), both at the microeconomic and the aggregate level, can provide valuable insights for a better understanding of actual and required transitions in the energy-converting capital stock composition, related fuel consumption patterns, underlying investment decisions, and technological trajectories followed. In this paper, we study energy conversion technology adoption in the electricity sector from the perspectives of irreversible investment under uncertainty and environmentally sustainable development. Particularly, we develop a dynamic technology adoption model that features elements from real options theory (Dixit and Pindyck, 1994) and apply it to detailed industry level time series cross-sectional data for Turkey (e.g., for installed capacities, unit generation costs, and input fuel and electricity consumption and prices). The model developed is firmly rooted in economic theory and rests on important determinants of investment in available technology options, such as (expected) capital and operation costs over the lifetime of a certain vintage of a specific technology. Investment decisions in liberalized markets, in contrast to noncompetitive markets, are based on market-driven value maximization criteria. Because the profitability of investment projects is contingent upon input and output price variations, project values evolve dynamically over time. Therefore, it is optimal to invest in some physical asset (‘real option’) when the present value of the expected cash flow exceeds the cost of investment by a (strictly) positive amount that is at least equal to the compensation for the loss of forfeiting the real option. Two alternative approaches are discussed in the literature to derive the optimal investment rule and the value of the optimal investment in a real asset. While contingent claims analysis is essentially rooted in the finance literature, dynamic programming starts from a given discount rate and considers the maximization problem of the expected value of discounted cash flows. The two methods are linked through the equivalent risk-neutral valuation principle, and although they make different assumptions about financial markets and the rates firms use to discount future cash flows, they yield identical results in many applications. In contingent claims analysis, one attempts to find some combination or portfolio of traded assets that will be an exact replication of the return and risk pattern pertaining to the investment project studied. In this paper, a dynamic programming approach is adopted and the timing of the irreversible investment formulated as an optimal stopping problem (e.g., see Karatzas and Shreve, 1991). In particular, we use a model that accommodates plant availability, load duration curves, and irreversibility of investment similar to those of Moreira et al. (2004) and Chaton and Doucet (2003). This allows us to analyze the investment decisions taken for different vintages of power generating technologies based on different energy resources. The Turkish electricity supply industry provides the subject of our empirical analysis. Power plant expansion planning in Turkey has so far been based on the two main models MAED2 and WASP3, whose shortcomings have been discussed in various studies. For example, an investigation of historical MAED/WASP projections indicates that the model results have persistently overestimated electricity demand, as documented in Ediger and Tatlıdil (2002). Arıkan and Kumbaroğlu (2000) highlight the importance of the energy–economy feedback link that is missing in the MAED/WASP approach. We discuss our model's characteristics and compare its predictions for the Turkish electricity sector with actual developments. We also assess the differences between model and actual outcome in terms of environmental sustainability indicators (i.e., greenhouse gas and pollutant emissions). The remainder of the paper is organized as follows. Section 2 contains some general considerations regarding the adoption of electricity generating technologies. Section 3 introduces the literature and theoretical approaches considered; Section 4 describes and discusses the proposed model formulation; and Section 5 presents the empirical analysis and results from applying our model to the Turkish electricity-generating sector. Section 6 summarizes and concludes.

In this paper, we have applied a dynamic technology adoption model for the evaluation of irreversible investment options for electricity generating technologies, taking into account uncertainty, and vintage-specific life-cycle capital and operation costs. The consequences of electricity conversion technology choices for environmental sustainability have been a particular focus of our investigation. In the case of Turkey, we find that historical investments strongly diverge from the prediction of the model, indicating that the actual choices are far off from what a net present value-based optimization model would suggest. In particular, there is an accelerated adoption of natural-gas-fired technologies in reality that limits the increase in pollutant emissions, which would otherwise occur from the utilization of domestic fossil fuel sources. Indeed, according to the model, investments in lignite-fired technologies should dominate in view of the lower volatility in lignite prices. Conversely, high volatility of natural gas prices (in domestic currency)—essentially due to Turkey's economic instability and limited foreign exchange availability—reduces the attractiveness of technologies using this imported fuel. The explicit inclusion of macroeconomic uncertainty, expected technological change and learning, regulatory change, capital depreciation, and construction lead times into the formulation of the model are promising avenues for further improvement and testing.