The Scottish Government

« Previous | Contents | Next »

Listen

Technical Appendix A6. Simulation results: Renewable Energy Supply 1 - CGE Analysis

A6.1 Introduction

Our final set of simulations involves modelling the economic and environmental impact of increases in the share of electricity generated in Scotland from renewable energy sources. This appendix sets out some illustrative results from running the AMOSENVI model to simulate such an outcome. A number of practical problems have been encountered during the simulating of this outcome and we set these out in this note as well, before discussing the simulation strategy employed and the results of such policies. We have sought to model the effects of increases in the amount (and share) of renewable electricity generation in Scotland from the base year levels in the AMOSENVI model (1999). In the base year of the model (1999), we begin with a situation where renewable electricity generation provides 10.4% of all electricity generated in Scotland. In total, 42,482 GWh of electricity was generated in the base year of the analysis. In the core simulation which we present below we have sought to increase the share of electricity coming from renewable technologies, while maintaining the total amount of electricity generated in Scotland at levels as close as possible to the original figures for 1999. It would be possible, of course, to explore alternative assumptions about future consumption and production, so the present analysis should be regarded as indicative.

We carry out and report sensitivity analysis where in order to increase the proportion of the renewables, we relax the assumption that generation levels remain close to base year levels. There are a number of issues about the simulations which we report. Firstly, we assume that the underlying technology used to create the output of the renewable electricity generation remains unchanged. There is an extensive literature on learning rates, and the reduction in the costs of electricity generation from increased development and deployment of renewable energy technologies ( e.g. Winskell et al, 2007). These simulations do not incorporate such developments. Secondly, we seek to make changes to the sectoral output of the renewable and non-renewable electricity generation sectors, so that total output of the electricity sectors remains close to existing levels. Results for electricity consumption relate to total electricity consumption by industries and final demand categories in Scotland, and as such, include imports of electricity. Thirdly, the database used for these simulations is that using an experimental disaggregation of the Electricity sector in the original IO table for Scotland. In summary, features of the AMOSENVI model make the results presented no more than illustrative of the type of results which can be obtained from CGE analysis.

In Section A6.2, we set out some of the practical issues encountered in running the AMOSENVI model to capture the effects of increased penetration of renewable electricity generation in Scotland. In Section A6.3, we briefly describe the simulation strategy employed, while in Section A6.4 we report the economic, environmental and energy results from a "central" scenario in which we significantly increase the level and share of renewable electricity generation. In sensitivity analysis in Section A6.4, we further increase the share of renewables, but this can only be accommodated in the model with an associated further rise in the price of electricity. Therefore allowing total electricity generated in Scotland to be lower than the base year, we report the results from scenarios where the level and share of renewable electricity generation is significantly greater. These scenarios are also associated with a relative decrease in Scottish CO2 emissions, against the "core" scenario, but also larger declines in Scottish GDP.

A6.2 Practical issues in modelling increased penetration of renewable generation

Our initial plans to model the impact of renewable energy supply were to focus on the electricity sectors in the AMOSENVI model (sectors 24 - renewable electricity; and sector 25 - non-renewable electricity) and introduce shocks to the efficiency of production at various points on the sectors' production function. With the greater penetration of renewable energy supply, there would be a requirement for additional grid enhancement, requiring greater capital intensity, as well as additional capacity or capital-intensive electricity storage being required to accommodate the greater intermittency of electricity production. The CGE model could then be used to identify the substitution and output effects of the movement towards greater capital intensity of production in both the electricity renewables and non-renewable generation sector. The levels of negative shocks to capital efficiency which would be introduced into the model, would be calibrated on existing estimates of direct cost changes for greater levels of renewable electricity generation.

This simulation strategy ran into a number of problems, most notably that we were unable to calibrate the necessary shock and we were not able to enter the capital efficiency changes correctly. These problems appear to suggest that additional programming development work is necessary before the AMOSENVI model can accommodate simulations using this route. An alternative approach was necessary which we use for the "central" simulation and the sensitivity analysis that follows.

A6.3 Simulation strategy

Our alternative simulation strategy involves introducing subsidies to renewable electricity generation and taxes on non-renewable electricity generation. The intention is to choose the appropriate tax and subsidy rates such that the outputs of these two sectors adjust so that the combined "physical" electrical output of these two sectors remains approximately constant, but that the share of electricity produced by renewable electricity increases from its base year value. When we hold "physical" electricity output constant this is not equal to the combined real value of the output of the two electricity sectors being kept constant.

Ideally, the tax and subsidy raised should be revenue-neutral to the Government exchequer. We ensure this by allowing government expenditure to adjust so as to maintain the ratio of government deficit to GDP at its base year level. In all the simulations that follow, government expenditure is lower than in the base year, indicating that increased tax revenues in the non-renewable sector are not large enough to offset the subsidies required to stimulate the renewable electricity sector. The increases in tax necessary for the non-renewable sector to get the relative prices of renewable output to non-renewable output to shift, will have the effect of reducing the real wage, and in principle might increase government revenues. In the simulations which we report, however, the competitiveness effect of high prices is larger than the demand stimulus, and, in fact, government expenditure, and GDP, fall. The tax take is lower

Results in Section A6.4 consider the economic implications of a Government policy package designed to increase the share of renewable electricity generation as a proportion of total electricity production. This is intended to explore the potential system-wide consequences of the Scottish Government's stated objective for 31% of total energy generation to be sourced from renewable energy technologies by 2011. For reasons explained above, we analyse alternative subsidy and taxation combinations that are applied to the renewable/non-renewable electricity generation sectors, respectively. Various model constraints, however, are such that we are not able to replicate exactly the magnitude of renewable electricity generation penetration that is implied by the Scottish Government's objective. In total, we ran approximately 500 simulations, with different levels of taxes and subsidies such that we held total electrical output approximately constant, and increased the share of electricity from renewable sources.

As in our previous CGE modelling analyses, we examine the effects of the policy change subject to our benchmark equilibrium time period; that is, our results refer to percentage changes in variables compared to base. In this model framework, wages are determined according to our bargaining set-up, and we allow for migration of the labour supply to and from the rest of the UK. We report long-run results, where this represents a conceptual time period over which labour and capital stocks fully adjust to new equilibrium values. In the current model set-up, this corresponds to a timeframe whereby real wages and unemployment are restored to initial equilibrium values, and the capital rental rate is equalized across all sectors.

A6.5 Central results and sensitivity analysis

A6.4.1 Central aggregate and sectoral results

Our central scenario involves a subsidy package equivalent to 94.1% of value added for the renewable electricity generation sector, and a tax equivalent to 36.9% of value added to the non-renewable electricity generation sector. Table A6.1 reports the long-run impacts on key aggregate economic, energy and environmental variables. This policy change has the effect of reducing long-run GDP by 1.15%. The key factor underlying the negative impact on output are the price effects associated with the policy change. The extent of taxation in the non-renewable electricity sector is such that the price of output in this sector increases significantly (by 28.54%). This leads to a relative increase in the cost of the electricity composite, which combines with other energy inputs to form an overall energy composite. Increases in the price of the energy composite will serve to raise the cost of intermediate inputs, which will have negative implications for economic activity across the economy as a whole.

Table A6.1: Long-run aggregate economic, energy and environmental impact from "central" increase in renewable electricity generation in Scotland, bargaining labour market, % changes from base expect where indicated

Figure A6.1 illustrates the long-run changes in output and employment across all sectors. It shows that those industries which are heavily dependent on the activity of the non-renewable electricity generation sectors (such as the coal extraction and gas sectors), are most negatively affected by the fall in output in that sector

Figure A6.1: Long-run impact on sectoral output and employment, % changes from base year

Significantly higher production costs mean that output contracts relative to base in the coal sector by 17.85% (higher even than the fall in output in the non-renewable electricity sector of 10.82%), and in the gas sector by 3.56%. The only sector to experience an increase in output is, as expected, the renewable electricity sector. In this sector the subsidy leads to a reduction in the price of outputs (by 49.79%), and is associated with an increase in sectoral output of 92.8%.

The effect of this reduction in the price of renewable electricity as an intermediate input, and the overall boost in activity in this sector is, however, insufficient to outweigh the negative effects in the non-renewable energy sector. The relative dominance of non-renewable electricity generation in the supply chain is such that all other sectors experience an overall increase in input prices. As noted above, we hold "physical" electricity output constant, but the real value of output of the electricity sectors decreases as the increases in the price of the electricity composite is greater than the increase in the value of output. This leads to an economy-wide increase in prices: CPI increases by 0.51% relative to base. In the long-run, real wages return to their pre-shock level, but there is a lasting effect on nominal wages. Nominal wages increase by 0.51%, reflecting the increase in CPI, and a reduction in external competitiveness means that exports fall by 0.33%. Government expenditure falls by 1.42% in total, as the subsidies required to bring forward renewable electricity generation are greater than the taxes raised from non-renewable electricity generation, requiring government expenditure to contract to maintain the ratio of Government deficit to GDP, as described above.

The implications for the labour market are clear. In line with changes in output, employment falls across all sectors, except for the non-renewable electricity sector, and the highest relative reductions occur in the most energy-dependent sectors. Across all sectors, the percentage change in employment is closely comparable with changes in output, with the exceptions of the renewable and non-renewable electricity sectors, which reflects the fact that the tax and subsidy are effected on capital, and so incentivise a substitution towards/from capital in the renewable and non-renewable electricity generation sectors respectively. The overall fall in aggregate employment leads to outward migration, and a fall in Scottish population relative to base.

The environmental consequences of this policy are lower CO2 emissions. The fall in CO2 emissions outweighs the reduction in GDP, partly due to the shift in the composition of electricity generation from non-renewable to renewable sources. In the long-run, the share of electricity generation sourced from renewable technologies is 20.06%, compared to a share of 10.4% before the policy shock. This means that the CO2 intensity of Scottish production falls, along with total CO2 generation.

A6.4.2 Sensitivity analysis

Table A6.2 shows the long-run impact on aggregate economic, energy and environmental indicators for alternative combinations of subsidy/taxation rates on the renewable and non-renewable sectors respectively. Our aim of this analysis was to analyse a subsidy/taxation mix that would lead to an increase in the penetration of renewable electricity generation in order to match the Scottish Government's objective of a 31% share of total electricity output, whilst at the same time keeping total "physical" ( i.e. kWh) electricity output constant. However, modelling constraints mean that to achieve a 31% share would require significant alterations to the current model framework, which is outwith the scope of this study. Although we are able to determine subsidy/taxation ratios that achieve a higher market share than in our central scenario, this is at the expense of keeping total physical electricity output close to its initial level. As such, we carry out a number of simulations, which are feasible within our current model framework, for alternative subsidy/taxation mixes. These consider (i) a subsidy/taxation mix that is designed to achieve the highest possible penetration of renewable electricity outputs, whilst keeping total electricity output fixed at its base year value (the central case scenario), and (ii) subsidy/taxation packages that increase the percentage share of renewable electricity output to as close to 31% as possible, whilst allowing the total level of electricity generation to fluctuate away from the base year value (scenarios 1, 2 and 3).

The findings of these simulations are broadly as anticipated. In scenario 1, the higher subsidy/taxation rates have a corresponding impact on the wider economy. The same effects on input prices and export competitiveness are evident as in the central case, but to a greater extent. The magnitude of the effects on aggregate output and employment are therefore much increased compared to the central scenario: output and employment fall 4.24% and 4.51% respectively, relative to base, compared with falls of 1.51% and 1.35% in the central case. This translates to more significant improvements in the environmental indicators than in our central scenario: CO ₂ generation falls by 10.05%, compared with a fall of 4.15% in the previous case. This is partly due to the fact that renewable electricity output as a share of total electricity is higher (24.21% in this scenario compared to 20.06% in the central case), but is also because total electricity output has fallen compared to base (by 13.38% in this scenario, compared with a fall of 0.04% in the central scenario).

For the higher subsidy/taxation mixes (scenarios 2 and 3), we find that we are able to achieve a higher penetration of renewables electricity (market shares of 25.46% and 26.08% respectively), though for each of these scenarios, total "physical" electrical output moves significantly away from the base year values. These two scenarios involve higher taxation of the non-renewable electricity sector compared to both the central case and scenario 1. As expected, this is associated with a higher increase in the price of non-renewable electricity inputs to production, and therefore a greater increase in CPI, and underlies significant reductions in export competitiveness and GDP. The fall in the competitiveness is greater than the demand stimulus from an increased government revenue; in fact in these scenarios the tax take is reduced with a consequent fall in government expenditure. The increase in the market share of renewable electricity, combined with reduction in total electricity output (of 12.13% and 12.98% respectively), leads to a notable improvement in environmental indicators, with CO ₂ generation falling by almost 13% for scenario 3.

Overall, the results suggest that policy intervention to increase the market share of renewable electricity generation, in the form of a subsidy/taxation package such as those described above, could lead to a deterioration in overall economic performance in the long-run. The scenario analyses suggest that the higher subsidy/taxation rates outlined above are associated with greater downturns in economic activity. One important caveat is that we assume that the cost of the required net subsidy is decreased government expenditure. An alternative specification could be possible, where government budget remains in balance by changing the average tax rate. This could have important consequences for the scale of the aggregate economic impact, as the tax rate would affect the labour supply decisions facing households.

These results, however, are subject to there being no other policy measures or economic influences at work to offset or reinforce these effects. We do not consider, for example, the consequences of significant skilled labour shortages in the renewable electricity industry, and consequent wage increases. This may be a feasible outcome of an intense subsidisation policy, and such effects would exacerbate any economy-wide price pressures. Nor do we consider scenarios reflecting the growth in dominance of renewable electricity inputs in the supply chain over time, which could alleviate, to some degree, the increase in prices of non-renewable electricity inputs. These effects, and other policy measures designed to complement the subsidy/taxation mix, could have important implications for the overall outcome of increased penetration of renewable electricity generation.

« Previous | Contents | Next »