Economic impacts of solar thermal electricity technology deployment on Andalusian productive activities: a CGE approach

Ann Reg Sci (2013) 50:25–47 DOI 10.1007/s00168-011-0471-3 ORIGINAL PAPER Economic impacts of solar thermal electricity technology deployment on Andalusian productive activities: a CGE approach J. M. Cansino ·M. A. Cardenete · J. M. Gonzalez ·M. del P. Pablo-Romero Received: 19 November 2010 / Accepted: 25 July 2011 / Published online: 11 September 2011 © Springer-Verlag 2011 Abstract Solar thermal electricity is a type of renewable energy technology of special interest for Andalusia (southern Spain) because of the large number of annual sunshine hours. This paper estimates the impact on productive activities of increasing the production capacity of the installed solar thermal plants in Andalusia. Using a computable general equilibrium (CGE) approach, estimates of the changes in the economic sectors’ activity under two different scenarios are obtained: i) based on two types of solar thermal electricity plants currently in operation and ii) based on an increase from 11MW in 2007 to 800MW installed capacity by 2013 to comply with the ‘Plan Andaluz de Sostenibilidad Energética (PASENER)—Sustainable Energy Plan for Andalusia’. For the case of a parabolic trough solar collector power plant, results show that compliance with the PASENER goal would increase the level of the productive activities by around 30%. For the alternative technology—a solar tower power plant—results show that activities would increase by around 5% for 30 years, the estimated lifetime of this type of plant. Thus, the impact of the solar thermal J. M. Cansino ·J. M. Gonzalez ·M. del P. Pablo-Romero (B ) Department of Economic Analysis and Political Economy, Seville University, Avda. Ramon y Cajal, 1, 41018 Seville, Spain e-mail: [email protected] J. M. Cansino e-mail: [email protected] J. M. Gonzalez e-mail: [email protected] M. A. Cardenete Department of Economics, University Pablo of Olavide, Ctra. Utrera, km. 1, 41013 Seville, Spain e-mail: [email protected] M. A. Cardenete European Commission (IPTS-JRC), Inca Garcilaso, 3, 41092 Seville, Spain 123 26 J. M. Cansino et al. electricity plants on the productive activities of the PASENER compliant production goal would be remarkable. JEL Classification Q28 ·C63 ·Q43 1 Introduction The view that renewable energy can be an engine of regional growth and development has been firming slowly, with its recognition in EU law since 2001. Directive 2001/77/EC (EP & C 2001) establishes, as one of the objectives of the Community, the policy on renewable electricity, its contribution to employment generation, especially local employment, and to the promotion of regional development. On the other hand, the European Commission Communication of 2005 related to renewable energy stated as one of its objectives the improvement in economic and social prospects, especially for rural regions (European Comission 2005). Recently, the Renewable Energy Policy (EP & C 2009) emphasized the benefits of new renewable energy plants to rural development because they can generate, among other benefits, opportunities for the development of small- and medium-sized enterprises. Thus, promotion of new renewable energy technologies in Andalusia (southern Spain), and therefore solar thermal electricity technology deployment, can be seen as an opportunity for growth and regional development. This paper estimates the changes in productive activities in Andalusia due to the solar thermal electricity deployment.1Using a computable general equilibrium (CGE) approach, the estimation is carried out under two different scenarios. In the first scenario, the paper estimates the impact on productive activities from the deployment of two types of solar thermal electricity plants.2Impacts of each type of plant are obtained by comparing results from the model with the benchmark. For this scenario, the paper follows technological assumptions made in Caldés et al. (2009). In the second scenario, to estimate the impact on the same productive activities, we consider compliance with the ‘Plan Andaluz de Sostenibilidad Energética (PASENER) 2007–2013’ (Consejería de Innovación, Ciencia y Empresa 2007) requiring an installed capacity of 800MW by 2013 starting from 11MW in 2007. Impacts of this increase are obtained by comparing results from the model in case of PASENER compliance with the same benchmark. The economic activity values for the Andalusian economy updated to 2008 define the benchmark. In case of renewable energy sources such as solar energy, their economic impact has been usually estimated using Input–Output (I–O) models. For the USA, I–O analysis has been reported in Cook (1998); US DOE (1992) and Ciorba et al. (2004); for the European area in Kulisic et al. (2007), Madlener and Koller (2007) and Allan et al. 1The Andalusian electricity sector is deeply integrated with the rest of the Spanish sector. Spain has a regulated electricity market. It is a free market for the power plants but with wholesale and retail customers. Prices are fixed by authorities. 2Solar energy is an intermittent energy source, making energy storage an important issue in order to provide energy continuously. However, solar thermal power plants generate heat using lenses and reflectors to concentrate the sun’s energy. As the heat can be stored, these plants are unique because they can generate power when it is needed, i.e. during day or night. We assume that solar thermal energy is part of the Andalusian mix. 123 Economic impacts of solar thermal electricity technology 27 (2008) developed a similar approach. Caldés et al. (2009), Calzada et al. (2009) and the European Commission (MITRE) (2009) recently used an I–O model to estimate the economic impact of renewable energy sources in Spain. Besides the estimation of the economic impacts of solar thermal electricity deployment, this paper aims to expand the literature using a computable general equilibrium (CGE) approach based on a social accounting matrix (SAM) instead of the I–O models used in similar works. In the last 25 years, the CGE has been widely used to analyse government economic policies, both in developed and developing countries (Shoven and Whalley 1992;Arndt et al. 2009). In general terms, these models translate the theoretical Walrasian general equilibrium system into fully operational tools, including an endogenous output and price system, substitutability in production and demand, and the optimization behaviour of individual agents. A CGE analysis allows one to study the changes in the spheres of production and consumption as well as in income distribution, in response to changes in a given economic policy, as these models explicitly include a representation of the framework of interdependencies among all markets in an economy. Section 2explains the CGE methodology. Section 3describes the economic database used. Results are contained in Sect. 4. Section 5gives conclusions and discussion. 2 Computable general equilibrium methodology In this section, we give a general overview of the CGE methodology and the associated SAM used for the modelling and computation. 2.1 The social accounting matrix (SAM) Table 1shows a standard framework of a social accounting matrix (SAM). According to Stone (1962), a SAM model is a representation of all the transactions made in the setting of an economy in a certain period of time. The Input–Output Tables define the relation between the final demand and production, whereas the SAM describes how the production process influences and determines the demand. Thus, they extend the Leontief model and the relations shown by the Input–Output Tables because they describe the flows between the value added and the final demand and, therefore, represent the circular flow of the income. The SAM rows show the incomes for different accounts, whereas the columns show the expenses. The cells show monetary values. Following the early work on SAM in Stone (1962) and Pyatt and Round (1979) among others, its first applications in Spain3can be found in works such as Kehoe et al. (1988). Recently, SAM at a regional level has been developed,4and this paper follows this trend. 3To read more about SAM applied to the Spanish economy, see Polo and Sancho (1993), Uriel et al. (1998, 2005), Fernandez and Polo (2001), Cardenete and Sancho (2003)andSanchez-Choliz et al. (2007). 4See, for example, Manresa and Sancho (2004), and Llop and Manresa (1999) for Catalonia; Mainar and Flores (2005) for Aragon; Crámara and Marcos (2009) for Madrid; De Miguel and Manresa (2004)for Extremadura; and Cardenete (1998), and Cardenete and Moniche (2001) for Andalusia. 123 28 J. M. Cansino et al. Table 1 Social accounting matrix (SAM) for Andalusia, 2008 Production Productive factors Institutional sectors Capital Foreign sector Production Intermediate consumption Public sector and domestic consumption Gross capital formation Export Productive factors Payment of VA to the factors Institutional sectors Taxes on activities and goods & services Factors income allowance to institutional sectors Common transfers among institutional sectors Taxes on capital goods Transfers from the rest of the world, taxes on export Capital Fixed capital consumption Institutional sectors saving Foreign saving Foreign sector Import Transfers to the rest of the world Source Symmetrical matrix IOFAN95 (IEA) and Cardenete and Moniche (2001) The SAM describes an economy in great detail as in Fernandez and Gonzalez (2004). Therefore, SAM becomes a very useful instrument for public policy evaluation, allowing an investigator to overcome some of the statistical limitations. Starting with some hypothesis related to the behaviour of the economic agents, their economic environment and the structure of the economy, SAM are used as databases which allow one to develop a range of multisectoral models such as the computable general equilibrium models (CGE) (Shoven and Whalley 1992) described in the following subsection. 2.2 Computable general equilibrium model A static CGE model has been developed for the Andalusian economy, following the structure of Andre et al. (2005), who presented the first CGE applied to environmental issues for Andalusia. This model allows us to determine the effects on resource allocation under the two scenarios considered. In our case, we have not used the environmental taxes and we have focused on the purely economic CGE model as detailed in the following. This model involves a set of equations that reflect equilibrium conditions and the behaviour of the different economic agents. For that reason, the producers, the households, the public sector and the foreign sector are considered in general terms. In this subsection, a detailed analysis of each sector or agent is given (Sects. 2.2.1 to 2.2.4), including some observations in relation to the labour market (Sect. 2.2.5) and the notion of equilibrium used (Sect. 2.2.6).5 2.2.1 Activities The model for the Andalusian economy incorporates 27 productive activities. It is assumed that each productive activity generates a homogeneous product, according to 5The main equations of the model are shown here. The full listing of equations is available upon request. 123 Economic impacts of solar thermal electricity technology 29 a nested production function. At the first nested level, following the Armington (1969) hypothesis, the total production of each activity (Qj)is obtained as a Cobb–Douglas aggregate of domestic output (Qdj)and imports (Qmj). At the second level, the domestic production for each activity is obtained with a fixed-coefficients technology between intermediate inputs (Xij) and value added (VAj). Finally, at the third nested level, the value added of each activity is obtained by combining the primary factors of capital (Kj) and labour (Lj), according to a Cobb–Douglas technology function. The expressions used at these three levels are given in Eqs. (1), (2) and (3), respectively: Qj=βAjQdδdj jQm1−δdj j(1) Qdj=min X1ja1j,X2ja2j,...,X27 ja27 j,VAjvj(2) VAj=βjKαj jL1−αj j,j=1,2,...,27 (3) In these expressions, βAj and βjare scale parameters; δdjare parameters which reflect the share of domestic output of jin j’s total production; parameters azj express the minimum amount of zneeded to obtain a unit of j;vjis the technical coefficient of value added; and, finally, αjand (1−αj) are parameters which represent the participation of the primary factors, namely capital and labour, with regard to value added.6We have chosen these two types of methodology—Leontief and Cobb–Douglas—because we have used a calibration method to give values to overall parameters and coefficients in the model, avoiding the need to estimate any value following econometric estimations, as CES (Constant Elasticity of Substitution) technology requires. Finally, it is assumed that firms obtain their demand functions for inputs and supplies of outputs by maximizing profits under the specified technological constraints. 2.2.2 Consumption The model assumes only one consumer. The following Cobb–Douglas utility function (U), defined in terms of saving and consumption, is considered: maximize UhCjh,Sjh=⎛ ⎝ 27  j=1 Cφjh jh ⎞ ⎠S(1−φ)h h(4) such that pjCjh +invpSh=YD h(5) In Eq. (4), the parameters φand (1 −φ) represent the share coefficients corresponding to consumption goods and savings, respectively. Srepresents the saving, C expresses the private consumption of commodity, and jand his the representative consumer. Recall that the economy is divided into 27 activities. Equation (5)shows 6For the simulations considered in the paper, a sensitivity analysis for functional forms has been done. Specifically, a Cobb–Douglas function between intermediate inputs and value added has been introduced instead of the Leontief function of Eq. (2). The results obtained in both cases are very similar—qualitatively and quantitatively—and, therefore, those from the Cobb–Douglas specification have not been included in the paper. 123 30 J. M. Cansino et al. the budget constraint for this representative household account, where pjare present consumption prices and invp is the price index of savings/investment goods. Thus, the disposable income of the only household group7is given by Eq. (6): YD=(1−τ)[rK +wL(1−u)+cpi T PS +rowpT FS−ess wL(1−u)].(6) The representative consumer derives the consumption demand functions by maximizing the utility function subject to the budget constraint shown in (4) and (5). The right-hand side of Eq. (6) shows disposable income, YD. This income comes from the sale of its endowments of capital (K) and labour (L), at the prices rand w, respectively, which are our numerarie in the model. In addition, households receive transfers from the public sector (TPS), indexed by the consumer price index (cpi), and receive transfers from the foreign sector (TFS) indexed by rest-of-the-world price index (rowp), although their total quantitative importance is minimal. Finally, households have to pay employees’ social contributions and income tax, whose rates are ess and τ, respectively. The unemployment rate is represented by u. 2.2.3 Government The activity of the government consists, on the one hand, of producing public services, using the technology of “Non-sales oriented services” ( j27), while, on the other hand, of demanding public services (public consumption, CG j27) and investment goods (CG i). In this sense, this agent can be considered to maximize a Leontief utility function (UG), defined by Eq. (7): UG=min CG j27,γ GCG i(7) where γGis an economic policy parameter reflecting the existence of a fixed proportion between public consumption and public investment. The budget constraint that the government confronts can be expressed by inequality (8): pj27CG j27 +piCG i≤RG+piwG i−cpi T PS.(8) The left-hand side of this inequality reflects government spending on consumption and investment. On the right-hand side, tax revenues are (RG), from which transfers paid to households have to be subtracted. wG irepresents the stock of debt that the government issues when it is in budgetary deficit. The rest of the activities could buy this debt at the same price as saving/investment, pi. With respect to the total tax revenues RG, the model includes net taxes on production, employers’ social security contributions, import taxes and the previously 7As will be discussed later, uis an endogenous variable that reflects the unemployment rate. 123 Economic impacts of solar thermal electricity technology 31 mentioned value added tax as indirect taxes. As direct taxes, employees’ social security contributions and income tax are considered. The tax revenue components (a) to (f) are specified in Eqs. (9)to(14), respectively. a) Taxes on production (Rt): Rt = 27  j=1 tj27  z=1 pzXzj +w(1+escj)Lj+rKj.(9) That is, the domestic output of each activity is subject to a tax at a rate tj.The production price for activity zis pz. Finally, escjstands for the employers’ social contributions rate. b) Employers’ social security contributions (resc): resc = 27  j=1 escjwLj.(10) c) Import taxes (Rtarif): Rtarif = 27  j=1 tarif jpmQmj(11) where tarifjis the import tariff rate for activity j, while pmis the weighted price index of imported products. d) Value Added Tax (Rvat): Rvat = 27  j=1 vat jpjCj.(12) e) Employees’ social security contributions (ress): ress =esswL(1−u). (13) f) Income tax (Rτ): Rτ=τ[rK +wL(1−u)+cpi T PS +TFS−esswL(1−u)].(14) 2.2.4 Foreign sector Following the Armington (1969) hypothesis, since our analysis is based on the Andalusian regional economy, the foreign sector is modelled in a simple, aggregated way, 123 32 J. M. Cansino et al. as a single sector that includes the rest of Spain, the European Union and the rest of the world. ROWD = 27  j=1 rowpIMPj− 1  h=1 TFS h− 27  j=1 rowpEXPj(15) where IMPjrepresents the import activity j,EXPjthe export activity j, and TFS h the transfers which come from the foreign sector to the representative consumer h. The foreign deficit or surplus is represented by ROWD. 2.2.5 Labour market Capital and labour demands are obtained from conditional factor demand functions, thus minimizing the cost of obtaining value added. For the capital factor, we assume perfectly inelastic supply and therefore, this factor is always fully employed. However, the model allows possible rigidities in the labour market, so the unemployment rate may be positive. More precisely, we consider the relationship in Eq. (16) between the real wage and the unemployment rate: w cpi =1−u 1−u01/βd .(16) This formulation of the labour market in CGE modelling is due to Kehoe et al. (1988), following the precepts established in Oswald (1982). The variable (w/cpi) represents the real wage, uis the unemployment rate, u0is a parameter that reflects the unemployment rate at the benchmark equilibrium and βdis a parameter that expresses the sensitivity of the real wage to the unemployment rate. This last parameter can have values between zero and infinity. If βd=0, the real wage will adjust sufficiently so that the unemployment rate remains constant and equal to the benchmark equilibrium rate. If βd→∞, the situation is exactly the opposite, that is to say, the real wage remains constant and the unemployment rate varies. For intermediate values, higher values of this parameter represent greater wage rigidity. In other words, the sensitivity of the real wage to the unemployment rate diminishes. In the simulations presented below, calculations are carried out for different values of this parameter. Specifically, the extreme values βd=0 and βd→∞are used, as well as the value from econometric literature, βd=1.25 (see Andrés et al. 1990). 2.2.6 Equilibrium The notion of equilibrium that is used in the model is that of the Walrasian competitive equilibrium, extended to include not only producers and households, but also the government and foreign sectors (see, for instance, Shoven and Whalley 1992). Specifically, economic equilibrium is determined by a price vector, an activity-level vector and a set of macro variables such that supply equals demand in all markets, with the sole exception of the labour market, as previously discussed. Further, each 123 Economic impacts of solar thermal electricity technology 33 one of the economic agents included in the model attains its corresponding optimal choices under the respective budget constraint, i.e., the agents implement their optimal equilibrium solutions. We assume full employment for capital and unemployment in labour factor in the benchmark equilibrium. Additionally, the level of activity of government and the foreign sectors will be fixed, allowing relative prices, activity levels, public deficit and foreign deficit work as exogenous variables as mentioned before. Therefore, the equilibrium will be an economic state in which the representative consumer will maximize his utility, the productive activities will maximize their profits after taxes and the public revenues will be equal to the payments to the different economic agents. In this equilibrium, total sales will be equal to total demands in every market. Formally, the model achieves the equilibrium state of the Andalusian economy where the supply and demand functions for every good and service will be obtained as the solution of maximization of utility and profit problems. The result will be a price vector of goods and factors, utility level and tax revenues which satisfy the given conditions. Following these specifications, we reproduce the data contained in the SAM asabenchmark equilibrium in which all prices (endogenous and exogenous) are normalized to be equal to 1 at the initial time. From this initial condition, we introduce the increase in demand associated with the solar thermal power activity implied in the PASENER goal provoking an exogenous shock. This will allow us to evaluate the changes by comparing benchmark equilibrium with the simulation equilibrium. The model has been implemented using GAMS software (Brooke et al. 1988) with MINOS as the solver. The CGE model has been calibrated using the SAM for Andalusia, avoiding econometric estimations for coefficient and parameters included in the technological functions developed in the model. For more information about SAM and CGE structures, see Cardenete et al. (2010) and Cardenete and Sancho (2003), respectively. Similar to Caldés et al. (2009), the model used in this paper assumes an absence of production capacity limitations for produced goods. This assumption implies that there are no limits to the amount that a certain activity can produce in response to an increased demand (either in a direct or indirect way) generated by a new investment. It is also assumed that during the period under consideration, operation and investment costs remain constant. But unlike in Caldés et al. (2009), primary factors act as a disciplining constraint since their levels, being elastic within limits, are subject to the labour market constraints as specified in Eq. (16). The model allowed us to analyse two types of productive impacts: (i) Direct impacts caused by the expansion of production in other productive activities that need intermediate inputs of the manufacturing process from another branch of activity. In this case, the construction, operation and dismantling (which is ignored in this analysis as stated in footnote 8) of a solar thermal power plant requires inputs of other activities and this requirement causes effects on production.8 8Due to the lack of actual data regarding the dismantling phase of the project this last phase will not be considered in the paper. 123 40 J. M. Cansino et al. prices. However, this does not usually happen, as the cost of renewable energy is high. The activities that have the biggest weight on total production variation (column nine), in compliance with PASENER (i.e. the effect increased production of each activity has on the increase in the total economy), are transport and communications, other services and services oriented to sales. Together, they add up to around 73.44% of the production increase in the whole economy, due to these technologies. We now assess the role of direct and indirect equilibrium results after the first shock is introduced and absorbed. Comparing weights on total production variation (column nine of Table 4) with data from the third column of Table 2, we observe that activities with the greatest weights on total production variation are the same as those shown in Table 2. This shows the importance of direct effects. Nevertheless, percentage values in the two Tables 2and 4are different. This is due to the multiplier effects in the economy. Weights on total production variation of activities 24, 25 and 26 are substantially higher than the initial increases in demand shares. In particular, that occurs in activity 25 (other services). These three activities considered together account for 45% of the initial increase in demand (sum of transport and communications, other services and services oriented to sales in third column of Table 2), while the share of total production variation is 73.44% (sum of transport and communications, other services and services oriented to sales in column nine of Table 4). As the other activities included in Table 2have declined in value in Table 4,the multiplier effects work mainly on these three services activities. On the other hand, weights on total production variation of activities not present in Table 2represent the indirect effects of the initial demand on these activities. In general, these values are small, indicating that they have small multiplier effects. It is worth noting the impacts of activities 4, 6 related to the energy sector and 27 non-oriented services sales. Table 5shows the productive effects of solar tower power plants. Columns four to six show again one plant effects (Scenario 1) and columns seven to nine the effects obtained in compliance with the PASENER goal (20%) (Scenario 2). Column three represents the benchmark. For one plant, the increase in total production is shown in column five and amounts to 0.68%. Activities with the largest increases are productive services oriented to sales (2.03%), electricity (1.45%), metal products (1.11%), construction materials (1.05%), and transport and communications (1.07%). In terms of their share in total increases in production, column six shows that service-oriented sales activity accounts for 27.07% of the increase in total production and construction activity 16.04%. Also noteworthy is the weight of total variation of machinery’s activity, rest of commerce, transport and communications, and other services. Overall, as in case of the technology discussed above, it is worth noting that activities 23, 24, 25 and 26 represent together 49.98% of the total production variation caused by the plant. Table 5also shows the effects associated with PASENER compliance. The total production increase due to solar tower power plant is 4.57%, with the most notable activity increases, compared to its benchmark production level, in service-oriented sales (14.20%). The three activities with the biggest weight are transport and communications, other services and facilities for sales. The total increase in these is 47.78%. 123 Economic impacts of solar thermal electricity technology 41 Table 5 Productive impacts of a solar tower power plant technology No Activity Benchmark production (kC =) Scenario 1 plant effects Scenario 2 compliance with PASENER goal effects, 20% Simulation production (kC =) Variation (%) Weight of total variation Simulation production (kC =) Variation (%) Weight of total variation 1 Agriculture 9,360.93 9,347.85 −0.14 −0.56 9,372.90 0.13 0.08 2 Animal products and forestry 2,240.88 2,245.59 0.21 0.20 2,277.89 1.65 0.24 3 Fishing 961.21 965.81 0.48 0.20 990.83 3.08 0.19 4Coal 2,855.65 2,874.48 0.66 0.81 2,998.34 5.00 0.92 5 Rest of extracts 2,590.45 2,607.61 0.66 0.74 2,690.56 3.86 0.64 6 Oil and natural gas 12,271.30 12,338.71 0.55 2.90 12,766.40 4.03 3.18 7 Electricity 4,706.44 4,774.46 1.45 2.92 5,278.12 12.15 3.68 8 Gas and water steam 506.28 509.69 0.67 0.15 531.02 4.89 0.16 9 Water 1,082.83 1,088.75 0.55 0.25 1,124.43 3.84 0.27 10 Food industry 31,299.09 31,377.91 0.25 3.39 3,209.16 2.27 4.57 11 Textile and leather 7,265.36 7,303.74 0.53 1.65 7,492.63 3.13 1.46 12 Wood products 4,460.81 4,485.74 0.56 1.07 4,599.49 3.11 0.89 13 Chemicals 13,347.20 13,406.48 0.44 2.55 13,820.39 3.55 3.04 14 Mining, iron and steel industry 5,580.48 5,608.71 0.51 1.21 5,819.78 4.29 1.54 15 Metal products 4,419.72 4,468.99 1.11 2.12 4,797.50 8.55 2.43 16 Machinery 16,080.16 16,25.73 0.91 6.26 17,340.16 7.84 8.11 17 Vehicles 7,059.81 7,088.10 0.40 1.22 7,743.66 9.69 4.40 18 Construction materials 7,208.83 7,284.62 1.05 3.26 7,707.82 6.92 3.21 19 Other transport elements 2,322.26 2,325.98 0.16 0.16 2,352.58 1.31 0.20 20 Other manufacturing 9,306.62 9,371.75 0.70 2.80 9,806.40 5.37 3.21 123 42 J. M. Cansino et al. Table 5 continued No Activity Benchmark production (kC =) Scenario 1 plant effects Scenario 2 compliance with PASENER goal effects, 20% Simulation production (kC =) Variation (%) Weight of total variation Simulation production (kC =) Variation (%) Weight of total variation 21 Construction 48,650.64 49,023.91 0.77 16.04 49,753.89 2.27 7.10 22 Commerce and repairing 2,147.20 2,156.62 0.44 0.40 2,224.97 3.62 0.50 23 Rest of commerce 39,031.29 39,166.61 0.35 5.82 39,331.33 0.77 1.93 24 Transport and communications 19,593.29 19,803.09 1.07 9.02 21,255.22 8.48 10.69 25 Other services 28,924.04 29,111.92 0.65 8.07 30,275.34 4.67 8.69 26 Services oriented to sales 31,082.10 31,711.97 2.03 27.07 35,497.15 14.20 28.40 27 Services not oriented to sales 26,174.92 26,181.70 0.03 0.29 2,617.61 0.16 0.27 Total 340,529.82 342,856.55 0.68 100.00 356,075.59 4.57 100.00 Source Own elaboration 123 Economic impacts of solar thermal electricity technology 43 The composition of the induced equilibrium shifts associated with PASENER is now discussed. When we compare weights on total production variation with data from the fourth column of Table 2, we observe that activities shown in Table 2have the greatest weight of total variation. However, there are other activities that have significant weights of total production variation. These include activities 6, 10, 11, 17 and 23. This shows that there are significant indirect effects. On the other hand, it is worth noting that weights on total production variation of activities 25 and 26 are far higher compared to their shares in initial demand increases. This shows that these activities are also significantly affected by indirect effects in the economy. Adding the effects of both technologies used in the assumed proportion (80% & 20%) shows that the increase in production is 35.37%. Most significant increases are associated with three activities: transport and communications, other services and services oriented to sale, which together total just over 70% of the variation of production, and basically are associated with the maintenance and repair of technical plants, which requires the employment of well-trained workforce. In rural areas where these plants are most likely to be located, this may be an opportunity to promote new economic activities and new local companies. Finally, the impact on regional GDP would be an increase of 7.99% for solar tower power plants and of 41.77% for parabolic trough power plants. In both cases, the figures refer to the total lifetime of the plants in Scenario 2. 5 Remarks The gradual penetration of renewable energy sources (RES) like solar thermal technology in electrical power production is shown to directly contribute to two of the four objectives of the EU’s energy strategy: to reduce the energy dependence on primary energy consumption (solar energy is an indigenous source) and to reduce the demand stress on the primary energy resources. It also contributes to the climate change policy due to lower greenhouse gas emissions associated with the use of RES. These objectives are also considered in the PASENER. The total increase in the economic activity due to a parabolic trough power plant amounts to 0.75%. The largest increases are linked to activities of services oriented to sales, electricity, transport and communications, mining, and the iron and steel industry and metal products. The total increase in the economic activity due to parabolic trough power plants in compliance with the PASENER goal is 30.81%, with remarkable increases in electricity. The activities with the biggest weight of total production variation are transport and communications, other services and services oriented to sales. The increase in the economic activity due to the solar tower power plant is 0.68%. The activities experiencing the largest increases are productive services oriented to sales, electricity, metal products, construction materials, and transport and communications. The total increase in the economic activity due to solar tower power plants is 4.57%. The transport and communication activities have the biggest weight. Adding the effects of both technologies used in the assumed proportion, the increase due to compliance with the PASENER goal is 35.37%. The variation of production is 123 44 J. M. Cansino et al. basically associated with the maintenance and repair of technical plants located in rural areas (70%), so it can be an opportunity for the development of these areas in Andalusia. Finally, as the regional GDP variation is positive, the net benefit of introducing the technologies considered in the paper is also positive. Acknowledgments The authors are grateful for financial support provided by the Andalusian Agency of Energy. The second author thanks Research Excellence Project SEJ-01252, SEJ-246—both from Andalusian Regional Government and the Spanish Ministry of Science and Technology (SEC2003-05112/ECO). The first, third and fourth authors are grateful for the financial support received from the project SEJ-132 from Andalusian Regional Government. The authors thank suggestions for improvements in the draught paper made by the participants in the Workshop on Public Economics and Renewable Energy, University of Seville, April 2009. We thank Prof. Ferran Sancho for his useful comments too. We wish to extend our gratitude to the anonymous reviewers of the Annals of Regional Science for their considerations and important suggestions. The usual disclaimer applies. Appendix See Tables 6,7,8and 9. Table 6 Investment costs of the parabolic trough power plant Item I II (I/Total) III IV (100%–III) Investment (kC =) Investment (%) Imports (%) Domestic (%) Domestic investment (kC =) Solar field 123,487 46.45 32.46 67.54 83,403.12 Power block 55,690 20.95 40.15 59.85 33,330.46 Land 1,211 0.46 0 100.00 1,211.00 Storage 33,187 12.48 40.22 59.78 19,839.18 Construction 26,584 10.00 0 100.00 26,584.00 Engineering 12,839 4.83 0 100.00 12,839.00 Contingencies 12,839 4.83 0 100.00 12,839.00 Total 265,837 100.00 28.51 71.49 190,046.87 Source Own elaboration from Caldés et al. (2009) Table 7 Operation and maintenance (O&M) costs of the parabolic trough power plant Source Own elaboration from Caldés et al. (2009) O& M Cost Annual Cost (%) Total cost over item cost (kC =) 20years (kC =) Fixed costs 1,292 10.50 25,250 Maintenance 2,761 22.45 53,958 Financing 5,432 44.16 106,158 Natural gas 1,563 12.71 30,546 Electricity 1,252 10.18 24,468 Total 12,300 100.00 240,380 123 Economic impacts of solar thermal electricity technology 45 Table 8 Investment costs of the solar tower power plant Item I II (I/Total) III IV (100%–III) Investment (kC =) Investment (%) Imports (%) Domestic (%) Domestic (kC =) Solar field 62,384 42.43 30.89 69.11 43,113.58 Tower 23,753 16.16 0 100.00 23,753.00 Power block 29,686 20.19 37.75 62.25 18,479.54 Land 1,423 0.97 0 100.00 1,423.00 Storage 9,412 6.40 34.45 65.55 6,169.56 Construction 9,414 6.40 0 100.00 9,414.00 Engineering 5,472 3.72 0 100.00 5,472.00 Contingencies 5,472 3.72 0 100.00 5,472.00 Total 147,016 100.00 23.00 77.00 113,296.68 Source Own elaboration from Caldés et al. (2009) Table 9 Operation and maintenance (O&M) costs of the solar tower power plant Source Own elaboration from Caldés et al. (2009) O&M Costs Annual Costs (%) Total costs over item costs (kC =) 20years (kC =) Fixed costs 1,292 18.06 25,250 Maintenance 1,455 20.34 28,435 Financing 2,812 39.31 54,955 Natural Gas 771 10.78 15,068 Electricity 824 11.52 16,103 Total 7,154 100.00 139,811 References Allan GJ, Bryden I, McGregor PG, Stallard T, Swales JK, Turner K, Wallace R (2008) Concurrent and legacy economic and environmental impacts from establishing a marine energy sector in Scotland. Energy Policy 36:2734–2753 Andre FJ, Cardenete MA, Velázquez E (2005) Performing an environmental tax reform in a regional economy. Ann Reg Sci 39:375–392 Andrés J, Dolado JJ, Molinas C, Sebastián M, Zabalza A (1990) The influence of demand and capital constraints on Spanish unemployment. In: Drèze J, Bean C (eds) Europe’s unemployment problem. MIT Press, Cambridge Armington (1969) A theory of demand for products distinguished by place of production. Int Monet Fund Staff Papers 16:159–178 Arndt C, Benfica R, Tarp F, Thurlow J, Uaiene R (2009) Biofuels, Poverty, and Growth: A Computable General Equilibrium Analysis of Mozambique. Contributed Paper prepared for presentation at the International Association of Agricultural Economists Conference, Beijing (China), August 16–22, 2009 http://ageconsearch.umn.edu/bitstream/52004/2/52004.pdf. Accessed 24 February 2010 Brooke A, Kendrick D, Meeraus A (1988) GAMS, a user’s guide. The Scientific Press, San Francisco Caldés N, Varela M, Santamaria M, Saez R (2009) Economic impact of solar thermal electricity deployment in Spain. Energy Policy 37(5):1628–1636. doi:10.1016/j.emplo.2008.12.022 123 46 J. M. Cansino et al. Calzada G, Merino R, Rallo JR (2009). Study of the effects on employment of public aid to renewable energy sources. Working paper, Universidad Rey Juan Carlos. http://www.worldcoal. org/bin/pdf/original_pdf_file/wci_briefing_note_report_effects_employment_public_aid_renewables (08_07_2009).pdf Accessed 24 February 2010 Cardenete MA (1998) Una matriz de contabilidad social para la economría andaluza: 1990. Revista de Estudios Regionales 52:137–153 Cardenete MA, Moniche L (2001) El nuevo marco input–output y la SAM de Andalucía para 1995. Cuadernos de Ciencias Económicas y Empresariales 41:13–31 Cardenete MA, Sancho F (2003) An applied general equilibrium model to assess the impact of national tax changes on a regional economy. Rev Urban Reg Dev Stud 15:55–65 Cardenete MA, Fuentes P, Polo C (2010) Sectores claves a partir de la matriz de contabilidad social de Andalucía para el año 2000. Revista de Estudios Regionales 88:15–44 Ciorba U, Pauli F, Menna P (2004) Technical and economical analysis of an induced demand in the photovoltaic sector. Energy Policy 32:949–960 Consejería de Innovación, Ciencia y Empresa (2007) Plan Andaluz de Sostenibilidad Energética. (PASENER). Agencia Andaluza de Energía. Junta de Andalucía. Sevilla. http://www. agenciaandaluzadelaenergia.es/agenciadelaenergia/nav/com/contenido.jsp?pag=/contenidos/doc_ estrategicos/pasener. Accessed 18 January 2011 Cook C (1998) The Maryland solar roofs program: state and industry partnership for PV residential commercial viability using the state procurement process. In Proceedings of the second world conference on photovoltaic solar energy conversion, Vienna, 6–10 July, pp 3425–3428 Crámara A, Marcos MA (2009) Análisis del impacto de los Fondos Europeos 2000–2006 en la Comunidad de Madrid a partir de la matriz de contabilidad social del año 2000. Investigaciones Regionales 16:71–92 De Miguel FJ, Manresa A (2004) Modelos SAM lineales y distribución de la renta: una aplicación para la economía extremeña. Estudios de Economía Aplicada 22:577–603 European Commision (2005) Communication from the Commission on the Support for Electricity from Renewable Energy Sources, SEC(2005)1571. COM(2005)627 final European Comission (2009) Monitoring and modelling initiative on the targets for renewable energies (MITRE). Country report, Spain. http://mitre/energyprojets.net. Accessed 19 February 2010 European Parliament and the Council (EP&C) 2001 Directive 2001/77/EC of the European Parliament and of the Council of 27 September 2001 on the promotion of electricity produced from renewable energy sources in the internal electricity market, Brussels, Belgium European Parliament and the Council (EP&C) 2009 Directive 2009/28/EC of the European Parliament and of the Council of 23 April 2009 on the promotion of the use of energy from renewable sources and amending and subsequently repealing Directives 2001/77/EC and 2003/30/EC, Brussels, Belgium Fernandez J, Gonzalez P (2004) Matrices de contabilidad social: Una panorámica. Ekonomiaz 57(III Quarterly):133–158 Fernandez M, Polo C (2001) Una nueva Matriz de Contabilidad Social para España: la SAM-90. Estadística Española 43(148):281–311 Kehoe TJ, Manresa A, Polo C, Sancho F (1988) Una matriz de contabilidad social de la economía española. Estadística Española 30:5–33 Kulisic B, Loizou , Rozakis S, Segon V (2007) Impacts of biodiesel production on Croatian economy. Energy Policy 35(12):6036–6045 Llop M, Manresa A (1999) Análisis de la economía de Cataluña (1994) a través de una matriz de contabilidad social. Estadística Española 41:241–268 Madlener R, Koller M (2007) Economic and CO2mitigation impacts of promoting biomass heating systems: an input–output study for Vorarlberg, Austria. Energy Policy 35:6021–6035 Mainar A, Flores M (2005) Aproximación a la estructura de la economía aragonesa: Matriz de Contabilidad Social 1999 y Landscape. In: Proceedings of the Actas de la XXXI Reunión de Estudios Regionales. Alcalá de Henares Manresa A, Sancho F (2004) Energy intensities and CO2emissions in Catalonia: a SAM analysis. Int J Environ Workplace Employ 1:91–106 Oswald AJ (1982) The microeconomic theory of the trade union. Econ J 92:576–595 Polo C, Sancho F (1993) An analysis of Spain’s integration in the EEC. J Policy Model 15(2):157–178 Pyatt G, Round J (1979) Accounting and fixed price multipliers in a social accounting framework. Econ J 89:850–873 123 Economic impacts of solar thermal electricity technology 47 Robinson S, Cattaneo A, El-Said M (2001) Updating and estimating a social accounting matrix using cross entropy methods. Econ Syst Res 13(1):47–64 Sanchez-Choliz J, Duarte R, Mainar A (2007) Environmental impact of household activity in Spain. Ecol Econ 62:308–318 Shoven JB, Whalley J (1992) Applying general equilibrium. Cambridge University Press, New York Stone R (1962) A social accounting matrix for 1960, a programme for growth edit. Chapman and Hall, London Uriel E, Beneito P, Ferri J, Moltró ML (1998) Matriz de Contabilidad Social de España (MCS-1990). Instituto Nacional de Estadrística, Madrid Uriel E, Ferri J, Moltró ML (2005) Matriz de Contabilidad Social de 1995 para España (MCS-95). Estadrística Española 47(158):5–54 US DOE (1992) Economic impact of a photovoltaic module manufacturing facility 123