| 摘要 | 1. Background and objective of this study
In 2005, a ��study on a national vision of the hydrogen economy and the action plan�� was conducted to set a roadmap on how hydrogen economy can be realized in Korea by 2040. According to the study, 207 trillion won should be invested to increase the portion of hydrogen energy in the final energy demand upto 15% by 2040. Based on the investment demand, the study analyzed the economic impact of the increase of hydrogen energy using input-output analysis. However, input-output analysis is based on simple assumptions which do not consider opportunity costs of investment, optimization behavior, or inter-sectoral equilibrium conditions. Therefore, input-output analysis is inclined to over-estimate the expected benefit from the investment on hydrogen technology.
The purpose of this study is to estimate the economic impact of hydrogen economy by 2040 using a dynamic CGE (Computable General Equilibrium) model. The dynamic CGE model is expected to overcome the limit of input-output analysis, and reflect specific features of hydrogen technology such as learning effect, economy of scale (imperfect competition), and complementarity of various energy sources.
Aggregation of industry in the model considers to split the industry into transportation, power generation, and all other industries, since hydrogen economy would affect transportation and power generation sectors through the application of HFCV(hydrogen fuel cell vehicle) and fuel cell for the power generation of household, commercial, and industry. As major features of this model, final energy consumption estimated by KEEI(Korea Energy Economics Institute) was applied and hydrogen supply predicted by the same approach was used as baseline scenario. Hydrogen was separated from new and renewable energy sectors in the production function, and hydrogen was assumed as substitute for fossil and new and renewable fuels in the transportation and power generation sectors.
Based on the above assumption, the impact of hydrogen economy on the energy mix, energy supply and demand, household consumption, investment, GNP, and output of transportation and other industries.
2. Major Results
Main economic features of hydrogen technology as a frontier technology are learning effect, energy complementarity, and economy of scale. Learning effect is observed in the introduction of new technology, which means that each stage of production including R&D, labor, capital, management has improvement through the accumulation of experience. Energy complementarity implies that a firm prefers to mix a spectrum of technology to prepare for future risk. Economy of scale occurs in the monopoly or imperfect competition. These factors are considered in the model.
According to previous studies on the learning effect and complementarity between energy sources, learning effect reduces production cost of new technology, but the positive externality spreads over other companies without charge. Therefore it is not possible to obtain socially optimal level of promotion of new technology. Rivers and Jaccard (2006) claimed that government should intervene in the penetration of new technology to internalize the positive externality generated from learning effect. Their results show that economic incentive policy is always dominant, but RPS(Renewable Portfolio Standard) could be more effective. In addition, political acceptability and other principle might be more important than efficiency, which leads to make command and control policy more attractive in promoting new technology.
Mulder et al. (2003) proposed 'complementarity' between energy sources and learning effect as major factors of 'paradox of energy efficiency'. More specifically, they argue that firms prefer diversified energy mix to monotonous one in order to hedge external shock. Also, firms introduce a new technology gradually for the purpose of minimizing low productivity due to the offset of learning effect of conventional technology as conventional technology is substituted by new technology.
Isoard and Soria (2001) showed that economy of scale and learning effect simultaneously determine the speed of entrance of a new technology to the market. Economy of scale moves along the long term prodcution curve, while learning effect shift down the long term production curve. Based on these assumptions, they estimated the learning effects of solar photovoltaic and wind power. According to the result, diseconomy of scale offsets learning effect, which leads to slow promotion of new technology. Therefore, they insist that government should induce the development of new technology by subsidizing it.
Next, a dynamic CGE model integrating learning effect and energy complementarity was constructed to analyze the economic impact of hydrogen economy. The dynamic CGE model is based on the endogenous growth model with imperfect competition and economic growth driven by investment on technology. In the process of the growth, energy is used as production factor and various kinds of energy are consumed to produce goods. With these features, vintage model is applied to the model so that the period of energy supply is determined endogenously. Productivity is enhanced when various energy sources are used at the same time (energy complementarity). Learning effect is supposed to increase productivity as the period of using a technology is longer.
As a main data, social accounting matrix(SAM) is created from 2000 Korean input-output table. Final energy demand and BAU scenarios of hydrogen supply and demand by sectors (residential, commerce, transportation, and industry) are imported from external source (KEEI, 2006). Time horizon was set between 2005 and 2040. Energy sectors are aggregated into coal, oil, town gas, heat, new and renewable energy and industry disaggregated into transportation and other industry. The dynamic CGE model relies on Ramsey type which implies perfect forecast on the future and no uncertainty. As a policy scenario, subsidy on the price of hydrogen and fuelcell industries is set as 10%, 20%, and 30%.
Results of the dynamic CGE model are as follow;
First, supply period of hydrogen energy increased by 40%, 120%, and 350% compared to reference case.
Second, proportion of hydrogen supply on the final energy consumption reached 9.2%, 15.2%, and 37.7% for the 10%, 20%, and 30% subsidy on the price (the proportion of hydrogen in BAU case was 6.5%).
Third, proportion of oil and electricity supply on the final energy consumption was constant but the proportion of oil declined by 0.6~3.4%, and electricity declined by 0.3~1.5% in 2040.
Fourth, household consumption decreased by 0.05~0.18% in 2040 since the necessary fund for the subsidy was assumed to come from household income.
Fifth, the overall investment increased by 0.1%~0.45% in 2040, and export increased by 0.1~0.8% in 2040.
Sixth, even if final consumption contracted, increases in the production, investment, and export leaded to the increase of GDP by 0.03~0.13% and small decrease of GDP in 2040.
Seventh, production of transportation increased by 0.3% in 2040 for the 30% subsidy.
Eighth, production of coal, heat and new and renewable energy was constant, since main production input of hydrogen energy was assumed as labor and capital. To be more strict, the model should consider other important inputs of production of hydrogen energy such as coal, natural gas, nuclear, off-gas, and new and renewable energy. If the other inputs are considered in the model, production of coal, heat, and new and renewable energy would change significantly. This is the main limit of this study.
Ninth, output of other industries declined since overall consumption by household was reduced.
3. Conclusion and Future Study
To sum up the analysis, the price subsidy policy for promoting hydrogen energy leaded to the small increase of hydrogen production in 2015 and enormous increase in 2040. Besides, final consumption was affected negatively, while production, investment, export, and GDP was influenced positively. Price subsidy should be funded from the increase of tax income or reduction to household transfer, which results in the negative impact on the household income. It would worsen the household consumption. Thus, if government raises the fund from the general tax revenue, it would affect social welfare negatively. It would be more desirable if government raises the fund from the charge of carbon tax on the emission of carbon or environmental tax on the fossil fuels.
Since the increase of price subsidy will impact negatively on the household consumption, appropriate level of hydrogen production and subsidy size should be calculated for the policy implementation in the future's study.
Future topics of this study would include intermediate input scenarios and uncertainty on the production cost of hydrogen energy. In this study, there was no intermediate inputs in the production of hydrogen, but coal, natural gas, nuclear power, and new and renewable energy are the main input of hydrogen production. Therefore, there was no change in the production of coal, heat, and new and renewable energy. But if the model assume the use of intermediate inputs in the production of hydrogen, there should be some changes in those intermediate inputs.
On the other hand, if there is uncertainty in the production of hydrogen, feed-in-tariff (FIT) and renewable portfolio standard (RPS) would yield different economic impact based on the theory provided by Baumol and Oats. Accordingly, future study should be launched to evaluate the different effect between FIT and RPS when there exist uncertainty in the production cost. |
