The Scottish Government

« Previous | Contents | Next »

Listen

2. Simulation results: impacts of increased energy efficiency

In the recent AEA report to the Scottish Government (Mitigating Against Climate Change in Scotland: Identification and Initial Assessment of Policy Options), one policy option suggested to reduce GHG emissions is to reduce demand for energy through efficiency improvements. However, as some of our own recent work (Allan et al, 2006, 2007a; Hanley et al, 2008; Turner, 2008a) and recent developments in the policy and academic literature relating to the possibility of 'rebound' effects demonstrates, the relationship between efficiency in energy use and demand for energy may not be so straightforward.

There are five distinct types of effects that occur throughout the economy in response to an energy efficiency improvement. These comprise (i) a need to use less physical energy inputs to produce any given level of output (the pure engineering or efficiency effect); (ii) an incentive to use more energy inputs since their effective price - the cost of energy to produce one unit of output - has fallen (the substitution effect); (iii) a compositional effect in output choice, since relatively energy-intensive products benefit more from this fall in the effective price; (iv) an output effect, since supply prices fall and competitiveness increases; and (v) an income effect as real household incomes rise. While (i) will reduce energy demand, (ii)-(v) will increase it. However, in an economy, such as Scotland, where energy is produced locally, the actual price of energy will also fall as production becomes more efficient (if the energy supply sectors themselves are targeted with the efficiency improvement) and as demand contracts (due to the efficiency effect, (i) above). Falling actual energy prices will give further impetus to effects (ii)-(v) above, which are the drivers of rebound effect. However, falling prices will also lead to reduced revenue and profitability in the energy supply sectors, which, if not countered by increased demand as competitiveness improves, will cause a drop in the return to all factors of production, particularly capital in what are relatively capital intensive sectors. The fall in the return in capital will trigger a contraction in the capital stock in the energy supply sectors (what we refer to as (vi), a 'disinvestment' effect) and prices will begin rising again, dampening rebound effects over time.

In the simulations reported in this section, we introduce a 5% increase in energy-augmenting technological progress to each of the 25 sectors identified in the AMOSEVNI model in turn ( i.e. 25 separate simulations are carried out, with the efficiency shock directed at one sector at a time). We abstract from how the efficiency improvement is actually made, or any associated costs, in order to identify the main drivers of any rebound effects that occur. In each case, we find evidence of the six effects identified above to varying degrees, depending on the demand and supply characteristics of the sector targeted with the improvement in energy efficiency. However, as the results reported in Tables 1 and 2 show, some degree of rebound effect occurs in all cases.

If we have a rebound effect, this means that there is a fall in energy consumption in response to an increase in energy efficiency, but this is less than proportionate. For example, where energy efficiency increases by 5%, we would expect the direct (engineering) efficiency effect - effect (i) above - to be a 5% decrease in energy consumption. However, if the change in the effective and/or actual price of energy triggers substitution, output/competitiveness, composition and/or income effects (effects (ii)-(v) above, which all act to increase energy consumption) we would expect to see a decrease in energy consumption that is less than 5%. If, for example, energy consumption only falls by 2.5%, we have 50% rebound. However, if there is sufficient price responsiveness in the system (through direct and indirect, or derived, internal and external - or local and export - demands for energy) coupled with features such as the direct and/or indirect energy intensity of the sector targeted with shock, the increase in energy consumption may act to more than fully offset any pure efficiency gains. This would give us backfire effects (rebound effects of more than 100%), with a consequent increase in energy-related emissions generation at the economy-wide level. As noted above, the strength of rebound effects is governed by the direct and indirect/derived elasticities of demand for energy throughout the economic system, as well as features such as direct and indirect energy intensities, openness to trade, elasticity of supply of factors of production etc.

Tables 2.1 below shows the short and long run (equating, respectively, to the first year after the shock is introduced and the point at which the economy is full adjusted) impacts on GDP, economy-wide CO2 and, the CO2 intensity of Scottish production. Table 2.2 indicates the presence of rebound and disinvestment effects.

Table 2.1- Short and Long Run Impacts on GDP and CO2 from a 5% Increase in Energy Efficiency in Each Sector of the Scottish Economy

Table 2.2 - Short and Long Run Rebound Effects from a 5% Energy Efficiency Improvement Targeted at Each Sector of the Economy

An increase in any type of efficiency generally manifests as a positive supply shock, which will lower the unemployment rate, increase wages and have a positive impact on Scottish economic activity (represented at the aggregate level by GDP) that is greater in the long run (with boosted activity in all sectors of the economy). However, as reflected in Table 2.1, the extent of the positive economics effect, and the nature, magnitude and direction of effects on environmental indicator variables depends on the type of activity targeted with the efficiency improvement. Generally, the more energy-intensive the sector, the greater the more important the improvement in energy efficiency will be. However, the extent to which even a very energy-intensive sector will be boosted, and the strength of the consequent ripple (multiplier) effects throughout the economy, will be determined by the responsiveness of the system (including external, or export demands) to the improvement in its productivity and competitiveness.

If any one of the 23 non-electricity sectors ( i.e. Sectors 1-23 identified in Table 1.1 in Section 1 above) is the recipient of the 5% improvement in energy efficiency, the long run result is an increase in GDP over and above the baseline, with an accompanying reduction in CO2 emissions as energy consumption contracts to some extent (the universal presence of rebound effects in Table 2.2 shows that the pure efficiency effect is offset to some degree in all cases). Therefore the CO2 intensity of Scottish production falls if any one of these 23 sectors is targeted with the shock. However, it is important to examine the component changes in GDP and CO2 in determining where efficiency improvements would best be targeted. It is also important to note that in all 23 cases, the positive impacts on GDP and CO2 generation are accompanied by disinvestment effects leading to a contraction in capacity in some or all of the Scottish energy supply sectors.

Net increases in energy consumption and backfire effects are observed in three cases: where the increase in energy efficiency is directed at Sea Fishing (only in the case of electricity consumption), Renewable Electricity or Non-Renewable Electricity. These cases are likely to be of particular interest because the increased energy consumption in response to increased energy efficiency will lead to increases in the level of Scottish CO2 emissions (though, in the case of Sea Fishing, this is offset by reductions in CO2 generation from non-electricity energy consumption). It is useful to look more closely at the cases where backfire is observed as the nature of this effect is quite different in each of the three cases.

Backfire in the Non-renewable Electricity sector follows the patterns expected in the existing literature. This is the most directly energy-intensive production sector and, in our 1999 database, accounts for around 25% of total electricity use and around 20% of total non-electricity energy use in the Scottish economy. It is also a relatively heavily traded sector and we assume here that export demand is highly responsive to the drop in price for what is a relatively homogenous commodity. That is, there are strong competitiveness effects when energy efficiency improves in this sector. Table 2.1 above and Figures 2.1 and 2.2 below show that, while there is a significant positive impact on GDP (0.5% over the long run - the largest of all of the 25 sectoral cases), the proportionate increases in all types of energy consumption at the economy-wide level are much bigger (2.1% for electricity and 1.6% for non-electricity energy consumption), with a resulting negative impact on all the key 'sustainability' indicators reported.

Figure 2.1 Impact of a 5% increase in energy efficiency in the Non- Renewable Electricity Sector on Environmental Indicators

Figure 2.2 Impact of a 5% increase in energy efficiency in the Non-Renewable Electricity Sector on Environmental Indicators

In the case of Renewable Electricity, on the other hand, while the backfire results in Table 2.1 are also very large, while this sector is as open to trade as the Non-renewable sector, it is much less energy intensive, with the implication that the impact on its output price and the consequent positive competitiveness effects will be smaller. Figures 2.3 and 2.4 show that, while both types of energy consumption rise over the long run - by 0.2% for electricity and 0.07% for non-electricity - these increases are much smaller than in the case of Non-Renewable Electricity, with the large backfire effect driven by the fact that such a small share of energy use is directly affected by the shock. ² However, Table 2.1 shows that the impacts on GDP, CO2 and the CO2 intensity of Scottish production are much smaller when the shock is targeted at the Renewable sector. Thus, there is a trade-off to be considered - both positive economic and negative environmental effects are smaller. However, there are a wide range of variables to be taken into account; for example when the Renewable Electricity sector is targeted, there is a fairly rapid and significant increase in the share of electricity generated from renewable sources (see Figure 2.3), but this is assuming no constraints on the growth of this sector in response to the positive supply stimulus.

Figure 2.3-Impact of a 5% increase in energy efficiency in the Renewbale Electricity Sector on Key Indicators

The Sea Fishing sector is an interesting case. This is the least electricity-intensive sector in the Scottish economy. Table 2.2 shows that here we observe the smallest electricity rebound effect in the short-run, but the biggest long-run backfire effect (bigger even than in the electricity sectors). Again, the changes in energy consumption underlying this dramatic result are very small in the case examined here. Electricity consumption in the Sea Fishing sector itself falls in response to the increase in efficiency, but there is a small increase in aggregate electricity consumption of 0.0008% (over the long-run). This is mainly driven by increases in imported and domestic electricity used by the 'Transport' and 'Textiles and Clothing' sectors, both of which are direct intermediate suppliers of inputs to the 'Sea Fishing' sector. The increase in aggregate energy consumption is small but the efficiency shock is applied to a very small share of total energy use. This means that there is in fact a sizeable backfire effect in terms of electricity consumption (323.8% over the long run) even though the shock is limited to the least (directly) electricity intensive production sector in the economy. This demonstrates why a general equilibrium framework is essential in assessing the nature and scope of rebound effects, even when improvements in energy efficiency are focussed in a single sector/activity.

In summary, the results of the simulations in this section suggest that improvements in energy efficiency will always give rise to some extent of rebound effects (this will be the case if there is any degree of direct and/or indirect price responsiveness in the system to falling energy prices) but that in most cases there will be a reduction in the level of CO2 emissions at the economy-wide level and a reduction in the CO2 intensity of GDP. However, the more directly or indirectly energy-intensive the sector targeted with efficiency improvements is, the more its competitiveness will increase, and greater the degree of price responsiveness to this, the more likely we are to observe backfire effects and increases in CO2 emissions. A crucial point, though, is to be aware of the assumptions underlying the analysis reported here. In the absence of econometric evidence to inform specification of key parameters in our model, we have adopted a common assumption from the CGE literature that elasticities of substitution in production ( i.e. the ease with which producers can switch between different types of inputs in response to a change in relative prices) are relatively inelastic ( i.e. for an X% change in relative prices, intermediate demand will change by less than X%). These are set at 0/3 in most, but not all, cases (see Hanley et al, 2008 and Turner, 2008a, for details). However, as Turner (2008a) shows, as price responsiveness increases in any part of the system, rebound and other economic and environmental effects will also increase.

It is out with the scope of the current project to carry out a systematic sensitivity analysis of the results reported in Tables 2.1 and 2.2 to the specification of key parameter values. However, initial (as yet unpublished) results for another project have involved extending the simulation work for the commercial Transport sector (the base simulation for which is reported in Tables 2.1 and 2.2). ³ This analysis focuses on rebound effects for the key energy input of oil, and suggests that if only one parameter, representing the substitutability of energy and non-energy intermediate inputs to production in the Transport sector is raised from the current value of 0.3 to 1 ( i.e. unitary elasticity of demand - the relative price of energy falls by X%, intermediate demand rises by X%), we get rebound effects of around 100% (and all disinvestment effects disappear). If we raise it any further we get backfire (energy consumption and emissions rise). Similar changes would be expected if we were to increase the responsiveness of different elements of direct and derived energy demands to changes in prices in any one of the sectors for which results are reported here. This conclusion emphasises the need to improve the modelling infrastructure for Scotland, with attention to, and availability of appropriate data for the econometric estimation of key energy demand relationships.

Labour productivity

Another stream of ongoing work by the energy modelling team (under the ESRC First Grants Initiative project) involves examining the impacts of increasing labour rather than energy efficiency. We are currently at a very early stage in running and analysing simulation results. Nonetheless, the initial results may be of some interest. As with the energy efficiency simulations, we introduce a 5% increase in productivity, but this time the technological progress is labour augmenting. Table 2.3 shows results for the short and long run changes in GDP, CO2 and the CO2 intensity of production at the economy-wide level when the labour efficiency improvement is introduced to each sector. These results are comparable with those in Table 2.2

Table 2.3 Short and Long Run Impacts on GDP and CO2 from a 5% Increase in Labour Efficiency in Each Sector of the Scottish Economy

The first point to note is that the GDP effects are significantly bigger in Table 2.3, with the exception of the cases where efficiency improvements are introduced to the Electricity sectors. This is largely explained by the fact that labour is a more important input to production than energy in most sectors. In most cases, the absolute level of CO2 emissions increases. However, in a number of cases the greater growth in GDP in the labour efficiency shocks brings with it a bigger long run decrease, or smaller increase, in the CO2 intensity of Scottish production. For example, if the efficiency improvement is directed at the Communications, Business and Finance sector (which contains a number of the key sectors identified in the Scottish Government Economic Strategy), the long run decline in the CO2 intensity of Scottish production is 0.24% with the labour efficiency improvement, compared with 0.13% in Table 2.1 (energy efficiency). When efficiency improvements are directed at the Non Renewable Electricity sector, the increase in the CO2 intensity of Scottish production is 0.68% when this takes the form of an increase in labour productivity compared with 1.29% for energy efficiency.

However, it is important to bear in mind that improved labour productivity does increase CO2 emissions in most cases. The exceptions are where the efficiency improvement is aimed at Sea Fishing and Public and Other Services sectors (at least over the long run). Generally, over the long run, if all sectors experience a 5% improvement in either labour or energy efficiency, our initial results suggest that improved labour productivity gives better aggregate results in terms of GDP and the CO2 intensity of Scottish production, but not levels of CO2 production. However, if we focus the shock only on energy use sectors ( i.e. omit the five energy supply sectors), the results are mixed in terms of the CO2 intensity of production and the larger increases in GDP from improving labour productivity need to be set against larger increases in Scottish CO2 production. However, again some initial sensitivity analyses suggest that if we make it easer to substitute between different types of input in production (including labour and energy), the results in terms of the CO2 intensity of production become more favourable for labour productivity and less so for energy efficiency. Therefore, further research is required. Nonetheless, the initial results presented here will hopefully stimulate discussion and consideration of potential positive and negative spillover effects of existing labour productivity policies and objectives to addressing the problem of climate change.

« Previous | Contents | Next »