Home / Renewable Grid Power – Index / Renewable Grid Power – Application

Renewable Grid Power – Application

Case Study on Renewable Grid-Power Electricity

Previous - Content - Next

5. Economic and Policy Analysis – Application to Canada

5.1 Introduction

This section presents the results of the economic modelling of each of the fiscal instruments. The discussion is organized and presented as follows:

• Overview of the fiscal instruments that are assessed
• Overview of the RFF model used to assess the instruments
• Summary of results (including a road map for understanding the results)
• Detailed discussion of the base case and each fiscal instrument
• Sensitivity analysis results
• Conclusion.

5.2 Overview of Fiscal Instruments Assessed

In collaboration with the NRTEE, a base case and five fiscal instruments were selected and modelled. They were:

• An emissions price, which is analogous to an emissions trading permit system or a carbon tax. Under this scenario, a shadow price is placed on carbon equivalent to $10/tonne of CO2. This shadow price is equivalent to the cost of an emission trading permit or the tax rate on carbon. The emissions price is applied uniformly across all fossil fuel generation in Canada in 2010. 
• A renewable portfolio standard (RPS), which requires that green certificates, or the equivalent, be purchased by utilities so that renewables generation increases relative to fossil fuel generation. Thus, the model compares the uptake of renewables attributable to an RPS relative to generation from fossil fuels, not from the entire electrical generation. Constraints are not placed on technologies or regional shares of the total RPS. Instead, prices are used as the determinant for the type of technology that is supplied at the prevailing electricity price under the RPS. 
• A renewable generation subsidy, which is modelled as a direct subsidy from government to grid-power RET producers on a per-kWh basis. In practice, a subsidy could include any fiscal instrument that lowers the cost of production for producers such as a direct production subsidy or a capital cost allowance. 
• A combination of RPS and generation subsidy, applied in tandem in both the first and second stages. We let the RPS be the dominant policy, since the standard is meaningless if the subsidy encourages more renewable generation than required. A notable feature of this combination is that relative to the case when the instruments are implemented in isolation, the price of the green certificates is offset in part by the subsidy. This outcome will therefore trigger distributional shifts in terms of cost imposition. 
• An R&D subsidy, which is a program targeted at reducing the future cost of renewables generation. As such, the instrument can be anticipated to have a greater impact in future periods. In the model, the increase above current renewables R&D required to achieve the emission reduction target is identified.

An important point to keep in mind is that the levels of the instruments, such as an RPS target (e.g., 10% of generation from renewables) or a subsidy level (e.g., $0.01 per kWh), are solved endogenously in the model. Each instrument is required to achieve a common emission reduction (or policy target) and then the model solves for the policy level that would achieve the carbon target. Included in the predicted outcomes are the instrument levels such as the RPS target, the subsidy level or the investment in R&D.

5.3 Overview of RFF Renewables Uptake Model

The RFF unified analytical model was employed to assess the impacts of the fiscal instruments on the reduction of greenhouse gas emissions as well as the development and diffusion of renewable energy. This model was developed and tested for the U.S. Environmental Protection Agency to assess the preferred fiscal instruments for promoting renewable energy technologies. The analytical model is deliberately kept simple to highlight key features. It includes two sectors, one emitting and one non-emitting, and both are assumed to be perfectly competitive and supplying an identical product, electricity. Fossil fuel production is assumed to be the marginal technology, setting the overall market price; thus, to the extent that renewable energy is competitive, it displaces fossil fuel generation in future policy periods. The model has two stages (i.e., the short-term covering 2010 to 2015 and the longer-term covering 2015 to 2030). Electricity generation, consumption and emissions occur in both, while investment in knowledge takes place in the first stage and through technological change and innovation lowers the cost of renewable generation in the second. Thus, the model incorporates technological change and innovation when assessing the relative merits of the fiscal instruments. An important assumption is that firms take not only current prices as given, but they also take prices in the second stage as given, having rational expectations about those prices. This is a plausible assumption given that utilities routinely forecast prices in future periods, especially when contemplating large capital investments. As well, long-term power contracts are prevalent, thus adding certainty to the price in future periods. Appendix B provides further details on the analytical model. The carbon-emitting sector of the electrical generation industry relies on fossil fuels. As it is the mature technology, the productivity improvements available through new R&D are assumed to be negligible.33 Its marginal production costs are assumed to be constant with respect to output and increasing with reductions in emissions intensity. The representative firm chooses emissions intensity to equate the additional costs of abatement to the price of emissions. The full marginal costs of generation then include both the marginal production costs, given the emissions intensity choice, and any effective tax, such as the price of the emissions embodied in an extra unit of output, or the cost of green certificates under an RPS. As long as fossil fuel generation occurs, the competitive market price must equal the sum of these marginal costs. Another sector of the industry generates without emissions by using renewable resources. Unlike the fossil supply curve, which is flat and set at the long-term marginal cost of electricity, the renewable supply curve slopes upward, reflecting marginal production costs that increase with output (see Section 3). As the young technology, the costs of renewable power shift down over time as the knowledge stock increases. There are two ways to increase the knowledge stock: investments in R&D and “learning by doing,” which is a function of total output during the first stage in the model. The representative renewable energy firm chooses output in each stage and R&D investment to maximize profits. In the first stage, it produces until the marginal cost of production equals the value it receives from additional output, including the competitive market price, any production subsidy, and the contribution of such output to future cost reduction through learning by doing. The firm also invests in research until the discounted returns from R&D equal investment costs on the margin. In the second stage, output does not generate a learning benefit, so it depends only on the effective price for renewables. Grid-RET electricity generation and fossil fuel electricity generation are assumed to be perfect substitutes from the consumer’s point of view. Consumer demand for electricity is a declining function of the price, and in equilibrium, total consumption must equal total supply, where total supply is the sum of fossil fuel and grid-RET electricity generation. In this model, fossil fuels are the marginal technology (by the assumption of flat marginal costs), and their generation costs determine the price of electricity. Fossil fuel output satisfies remaining demand after profitable renewable energy is produced, so any increase in renewable energy production “crowds out” fossil fuel production. Since all of the fossil sector costs are passed on to consumers, their producer surplus is by definition always zero (i.e., the fossil sector is not impacted directly by the fiscal instruments in terms of price changes, but does experience changes in the demand for fossils, which could lead to indirect costs associated with fixed variable costs for reduced fossil capacity). The renewable sector has positive producer surplus: the area between the price received and the marginal cost curve. That is, some firms in the sector can produce power below the subsidized market price and therefore may reap some profit under the fiscal instrument. This indicator is assessed in the model. Consumer surplus is the area under the demand curve above the price of electricity. The change in consumer surplus due to the renewable energy policy in this partial equilibrium model is roughly the retail price change attributable to the fiscal instrument multiplied by the average of the quantities demanded. The change in the consumer surplus provides a proxy for the cost impact on consumers of each fiscal instrument – that is, faced with higher prices, what is the cost increase to the consumer sector? Policies also have implications for government revenues, which we denote as transfers, assuming that these revenues are raised or returned in a lump-sum fashion. For example, a tax will raise government revenue, whereas an R&D program is a disbursement. The change in these transfers equals the tax revenues net of the cost of the subsidies or programs. The change in societal welfare due to a policy is the sum of the changes in consumer and producer surplus, net of the change in revenue transfers from the subsidy or tax. Since we target equivalent emissions reductions for each of the fiscal instruments, we hold the environmental effects constant across the policy scenarios. That said, the distribution of emission reductions in the two stages is altered by each of the instruments. Of course, welfare is unlikely to be the only metric for evaluating policy. Other indicators may be total emissions, consumer surplus, renewable energy market share, and so on. “General equilibrium” factors – like interactions with tax distortions, leakage or other market failures – can also be important for determining welfare impacts. Political economy constraints may also be important for determining policy goals. To the extent that these unmodelled issues are present, this partial equilibrium presentation of welfare within the sector will not reflect the full social impacts; still, it represents a useful baseline metric. While we calculate the costs of achieving emissions targets in this case study, the benefits of the fiscal instruments are not estimated. The fiscal instruments through their displacement of fossil fuel can be expected to trigger a number of environmental and economic benefits, including:

• Improved ambient air quality and reduced carbon in the atmosphere
• Avoided ambient air quality impacts on sensitive ecosystem and health receptors and the associated economic value of the avoided damages
• Climate change mitigation benefits such as avoided ecosystem, health and economic damages stemming from extreme weather events, temperature changes and sea-level rise and the associated economic value of the avoided damages.

While important to assessing the desirability of the fiscal instruments from a social perspective, the benefits are in a sense fixed in the case study by stipulating a common emission target that all instruments achieve. Thus, benefits are assumed to be constant across the fiscal instruments and each instrument is then assessed on a welfare cost basis. The following section presents and discusses the modelling results with respect to changes from the baseline as well as absolute indicators.

5.4 Summary of Results

When reviewing the summary results it is important to remember that the outcomes are a function of how the energy market is influenced by each instrument. In the model, this means that outcomes differ due to changes in three “decarbonization drivers”:

• Renewables penetration
• The carbon intensity of fossil fuel generation
• Total electricity demand.

Outcomes listed in Exhibits 5.1 and 5.2 can therefore be traced back to an instrument’s ability to impact one or all of these three decarbonization drivers. Exhibit 5.1 provides the estimated values for each scenario; Exhibit 5.2 presents the percentage change in the outcomes relative to a base. Each numbered item in the first column, which is a modelling output, is defined as follows:

1. No policy base case: Our model predicts that with no policy, renewable energy generation will increase from 13% to 17% of included generation in the second stage, which corresponds to a 5% emissions reduction. Subsequent policy scenarios will target a 12% reduction overall from the combined emissions in the two stages of the no policy case. 
2. Policy level for 12% emissions reduction: This row provides an estimate of the size of the fiscal instrument that is required to achieve the carbon reduction target:
   • For the emissions price, a tax of $10/tonne CO2 would achieve the 12% reduction in carbon emissions from the base case carbon emissions of 106 megatonnes (MT).
   • For the RPS, a portfolio standard of 24% would achieve the 12% carbon reduction. This 24% is the final share of renewables generation in generation covered under this study, which consists of both renewables and fossil fuel generation but excludes major hydro and nuclear.
   • For the renewable generation subsidy, a value of about 0.6 cents per kWh achieves the policy objective of a 12% carbon reduction.
   • When combined with a subsidy of 0.2 cents, the RPS needs to be set at a slightly higher target of 24.2%.
   • For the R&D subsidy, a program that increases R&D spending by 61% above the base case R&D levels would achieve the target.
3. Electricity price ($/kWh): This row indicates the impact of the fiscal measure on the annual price of electricity in the first and second stages (2015 and 2030, respectively). In Exhibit 5.1, a price above the base case indicates an electricity price change attributable to the instrument. Exhibit 5.2 provides an indication of the percentage change in price relative to the base case. 
4. Carbon emissions (MT): Carbon emissions are presented as annual estimates in megatonnes of CO2 for the last years in the first and second stages. Carbon reductions are influenced by the three drivers in the following ways:
   • Renewable penetration displaces fossil generation when an instrument reduces renewable production costs relative to fossil generation costs.
   • The carbon intensity of fossil fuel generation is reduced when carbon is priced in the fossil sector (i.e., abatement from natural gas generation that displaces coal).
   • An increase in electricity prices reduces total electricity demand, which displaces output from fossil fuels.For each scenario, carbon emissions are estimated by multiplying the “on margin” emissions intensity of fossil fuel by the quantity of fossil fuel supplied. There are no emissions associated with renewable output.
5. Renewable output (MWh 10^11): This row indicates the output of renewable generation in the two stages. Renewable output is a function of production cost differentials between renewables and fossil fuels. Instruments impact the cost differential through subsidizing renewable generation, inducing renewable production cost decreases through innovation, and/or taxing fossil fuel production. There is also a temporal aspect here, where instruments that promote innovation reduce renewables costs in the second stage. 
6. Fossil output (MWh 10^11): As with renewable output, fossil fuel output is altered by the instruments through price changes in production costs. Fossil output is also altered by total demand reductions, which occur when an instrument increases the price of electricity. 
7. Total electricity output (MWh 10^11): Total generation includes fossil and renewable output; changes indicate that the instrument influences final demand through electricity price increases. 
8. Renewable R&D ($M): Expenditures are expressed in millions of dollars annually in total R&D spending by the public and private sectors. 
9. Additional renewables cost reduction: This row indicates the percent reduction in the cost of renewables electricity below the base case. 
10. Consumer surplus ($M): This is the net consumer cost of the instrument measured as the change in the present value of the total cost to consumers for both stages. The consumer surplus is negative and is present when the instrument increases the price of electricity. 
11. Producer surplus ($M): This is the change in the measure of total profit in the renewable sector for both stages. Renewable sector profits increase when the instrument raises the price received by renewable generation, by either a subsidy or a tax on fossil generation. When this occurs, profits can be made if some renewable production costs are below the instrument electricity price in the scenario. 
12. Transfers ($M): This is the change in government revenues, where a positive number is revenue and a negative is a disbursement. Again, the estimate is a total cost for both stages. 
13. Welfare (excluding environmental benefits) ($M): This is the change in social welfare, and is a proxy for the societal cost of the instrument. It is the sum of consumer and producer surpluses and transfers. It is an important metric, since all scenarios achieve the same carbon reduction target, yet have differing social costs. 
14. Welfare relative to emissions price: This is simply a ratio that indicates the welfare costs of each scenario compared with the emissions price scenario. The emissions price is selected as the basis for comparison since it has the lowest welfare cost.

Exhibit 5.1 Summary of Modelling Results for Fiscal Instruments

	Base Case	Emissions Price	Renewable Portfolio Standard	Renewable Generation Subsidy	Combination RPS and RGS	Renewable Research Subsidy
1. Policy level for 12% emissions reduction		10 $/tCO2	24%	$0.006	RPS = 24.21%, RGS = $0.002	61%
2. Electricity price (in $/kWh)
1st stage	$0.092	$0.097	$0.095	$0.092	$0.095	$0.092
2nd stage	$0.092	$0.097	$0.093	$0.092	$0.092	$0.092
3. Carbon emissions (MT CO2)
1st stage	106	98.10	91.00	98.97	91.08	104.00
2nd stage	101	84.40	91.90	83.50	91.95	77.40
4. Renewable output (MWh 10^11)
1st stage	0.29	0.40	0.54	0.42	0.55	0.31
2nd stage	0.38	0.66	0.55	0.72	0.55	0.83
5. Fossil output (MWh 10^11)
1st stage	2.00	1.85	1.71	1.87	1.72	1.98
2nd stage	1.91	1.59	1.73	1.57	1.73	1.46
6. Total electricity output (MWh 10^11)
1st stage	2.29	2.25	2.26	2.29	2.27	2.29
2nd stage	2.29	2.25	2.28	2.29	2.29	2.29
7. Renewable R&D ($M)	$129	$450	$320	$533	$325	$1,576
8. Additional renewables cost reduction	0%	15%	13%	16%	13%	26%
9. Consumer surplus ($M)	$0	($11,690)	($4,521)	$0	($3,533)	$0
10. Producer surplus ($M)	$0	$2,215	$3,480	$2,846	$3,547	$1,590
11. Transfers ($M)	$0	$8,896	$0	($3,557)	($1,072)	($3,890)
12. Welfare – no benefits measured ($M) (9+10+11=12)	$0	($579)	($1,041)	($711)	($1,058)	($2,300)
13. Welfare relative to emissions price	-	1.00	1.80	1.23	$1.83	3.97

* Figures may not add due to rounding. Source: Marbek and RFF. Exhibit 5.2 Summary of Modelling Results Comparison of Scenarios from Base Case or Emissions Price

	Base Case	Emissions Price	Renewable Portfolio Standard	Renewable Generation Subsidy	Combination RPS and RGS	Renewable Research Subsidy
2. Electricity price (in $/kWh)
1st stage	From Base	5%	4%	0%	3%	0%
2nd stage	From Base	5%	1%	0%	0%	0%
3. Carbon emissions (MT)
1st stage	From Base	-7%	-14%	-7%	-14%	-2%
2nd stage	From Base	-16%	-9%	-17%	-9%	-23%
4. Renewable output (MWh 10^11)
1st stage	From Base	38%	88%	46%	89%	8%
2nd stage	From Base	74%	45%	89%	46%	118%
5. Fossil output (MWh 10^11)
1st stage	From Base	-8%	-14%	-7%	-14%	-1%
2nd stage	From Base	-17%	-9%	-18%	-9%	-23%
6. Total electricity output (MWh 10^11)
1st stage	From Base	-2%	-1%	0%	-1%	0%
2nd stage	From Base	-2%	0%	0%	0%	0%
7. Renewable R&D ($M)	From Base	248%	148%	313%	151.9%	1121%
8. Additional renewables cost reduction	From Base	8%	6%	9%	6%	19%
9. Consumer surplus ($M)		From Emission	39%	0%	30%	0.0%
10. Producer surplus ($M)		From Emission	157%	129%	160%	72%
11. Transfers ($M)		From Emission	0%	-40%	-12%	-44%
12. Welfare – no benefits measured ($M) (9+10+11=12)		From Emission	180%	123%	183%	397%

  *Figures may not add due to rounding. Source: Marbek and RFF.

5.5 Discussion of the Base Case and Fiscal Instruments

A discussion of the base case and each of the five fiscal instruments (emissions price, RPS, generation subsidy, combination and the R&D subsidy) is presented below.

5.5.1 Base Case

The base case provides the reference from which the percentage changes are estimated in Exhibit 5.2. Renewables penetration is forecast based on the relative costs of fossil fuel and renewables production. The baseline penetration of renewables increases over time, reflecting decreasing renewables production costs due to innovation. Total electricity output remains fixed in both periods in the base case, and thus increased renewables penetration decreases the carbon intensity of overall generation. This is captured as a decrease in carbon emissions over time, from an annual level of 106 MT in the first stage to 101 MT in the second stage. It is recognized that electrical production is increasing over time, but total electricity output in the model is fixed in both stages so that the demand and supply responses of the policies can be better understood.

5.5.2 Emissions Price

An emissions price works to reduce emissions by reflecting their cost, either in terms of environmental damages (as with an environmental levy) or in terms of opportunity cost elsewhere in the economy (as with an emissions cap-and-trade system). This price sends a signal to everyone in the energy market to conserve carbon. Fossil energy producers can reduce costs by boosting efficiency or switching to lower-carbon fuels and processes. Since the price of fossil energy will then incorporate the cost of the carbon associated with that form of generation, the price of electricity will also rise, creating two effects. First, it signals consumers to conserve and take advantage of opportunities to reduce their demand, like adopting more energy-efficient appliances. Second, it increases the price received by renewable energy producers, encouraging production and investment in non-emitting generation technologies. In the absence of major market failures, this three-pronged decarbonizing effect is the most cost-effective way to reduce emissions. In other words, with fossil generators, renewables producers and consumers all responding, fewer resources are expended to meet emissions reduction targets compared with policies that focus only on one or two of the “decarbonizing drivers.” In the results, a modest emissions price of $10/tCO2 causes emissions to fall by 10% from the base case. Compared with the other policies that meet that target, the emissions price induces the greatest rise in electricity prices (5%) and the least amount of renewables. Those results go hand in hand, since the reduction in demand due to the price increase and the efforts by fossil producers to reduce their emissions intensity in response to the carbon price reduces the need to rely on more costly renewables to reduce carbon. This three-pronged approach is reflected in the change in welfare ($579 million), where the emissions price has the lowest welfare impact of the instruments under consideration (see Exhibit 5.1). From a distributional perspective:

• Consumers incur the highest electricity price increase and consumer surplus loss under the emissions price. Since consumers are also taxpayers, the use of the revenues (i.e., transfers) is important in assessing the net effect on households. 
• From a renewables producer perspective, the emissions price has a modest but significant impact on renewables output, production cost decreases and producer surplus. The impact is also relatively consistent across stages, not targeting one more heavily than the other. 
• For fossil fuel electricity generators, the emissions price is the only policy with an incentive to reduce emissions intensity. Although profits for the fossil sector are not modelled – rather, they are assumed to be driven to zero in the long run by the market – the potential costs to the fossil sector under an emissions price would depend on their ability to pass along the production costs increases due to carbon abatement (i.e., coal to gas) to consumers, as well as any windfall gains from permit allocation. 
• For government, significant transfers or revenue could be raised under the emissions price either through a tax-based system that collects revenue or through the allocation or auctioning of carbon permits under an emissions trading system. This is the only modelled scenario where significant government revenue potential exists. It also represents the value of the emissions rents, which are available to be allocated to consumers, generators and their shareholders, funds for transition assistance, or taxpayers more generally. 
• From society’s perspective, the welfare costs are lowest with the emissions price, making it the preferred option. One negative consequence of this scenario, not incorporated into this single-sector analysis, is that the increase in electricity prices could lead to economy-wide competitiveness impacts such as reduced exports, to the extent that industrial users of electricity are affected. Reserving some permits for allocation to trade-exposed sectors that are electricity intensive could mitigate these impacts. 
• An advantage of a cap-and-trade system is certainty in reaching the carbon target; however, uncertainty will then manifest itself in the price. All the other policies face challenges in setting a policy level that would achieve the emissions target with certainty.

5.5.3 Renewable Portfolio Standard

The renewable portfolio standard (RPS) requires total electricity generation to comprise a minimum share of renewable sources. Although such a market share requirement can be implemented in several ways – quota obligations for retailers, green certificates for fossil generators – the general effect is the same. As long as the market would not meet the requirement on its own, renewables producers receive a price premium (the value of the green certificates they generate), while fossil energy producers receive a negative one (the cost of the green certificates they must buy in proportion to their generation). Moreover, the total subsidy to renewables producers is equal to the total effective tax paid by fossil generators, so no net revenues are raised or lost by the government (i.e., no transfers to or from government). Since the RPS does not distinguish among fossil generation technologies, there is no incentive to reduce emissions intensity in that sector. Consumer prices rise due to the effective tax on fossil energy to fund the renewables subsidy (i.e., buy green certificates), but not as much as with the emissions price instrument. Although under the RPS more renewable energy is generated than under the emissions price, the timing of that generation is changed. Normally, when prices are fixed, as costs fall over time, renewable generation expands. However, the RPS fixes the renewables share in both periods, and over time this becomes easier to meet; hence, the effective tax and subsidy fall (i.e., the price of green certificates falls), while total electricity generation increases with the reduced price (recall that the market price is equal to the price of electricity plus the price of green certificates, which fall due to innovation over time, and therefore electricity prices fall and final demand increases). Renewables then get a bigger boost in the first period and less in the second. The larger current subsidy may enable more learning by doing, but recognizing that the support will fall in the future, investment in cost-reducing R&D may be smaller (this result is borne out in our scenarios). Compared with the emissions price, the welfare costs of the RPS are higher by about 1.8 times. That said, the cost imposition on consumers is lower and the profits to the renewables sector is higher. Transfers to or from government do not occur. As there is less of a price impact on electricity, competitiveness and trade issues are less of a factor under an RPS than with an emissions price. Therefore, although the RPS may be less desirable from an economic efficiency perspective (i.e., higher welfare costs), the stakeholder acceptability of an RPS may in fact be greater. An exception may be the fossil generators, who must pay for the RPS, lose market share, and do not have the option to seek more cost-effective ways to reduce their carbon emissions. Furthermore, fixing an RPS has an important effect on the time path of renewable generation and thereby on the present value of the costs. The level of the RPS that achieves the carbon reduction target is 24.1% of total generation, which requires much greater renewables penetration in the early stage and less in the second stage, relative to the emissions price. Since more renewable production occurs when it is more costly, and less expansion occurs after knowledge accumulation brings costs down, this front-loading of effort under a fixed RPS raises overall welfare costs. From a distributional perspective:

• Consumers experience some electricity price increase and consumer surplus loss under the RPS. This effect is about 80% as large as with the emissions price in the first stage, and nearly negligible in the second. The electricity price rise is due to the purchase of renewable power in the form of green certificates (or the equivalent) by the fossil sector. Since renewables become cheaper with technical innovation, the cost of green certificates (and thereby consumer prices) is higher in the first stage but lower in the second as the cost of renewables supply decreases. 
• For renewables producers, the RPS induces a high uniform penetration through both periods, which is not surprising since the RPS fixes the share of renewables in both periods. Producer profits are also high, indicating the potential for the sector to benefit under an RPS. While there is certainty in terms of market share for the renewables sector, there is less stability in terms of prices, and less flexibility in terms of the timing of renewable generation. Furthermore, the fact that the implicit subsidy falls over time with cost decreases means that incentives for innovation may be muted – indeed, our model predicts less R&D spending than under the emissions price. Although more renewable generation is needed overall, so much is done in the first stage that the return to lowering costs in the second stage is lower, both because of the lower second-stage output (relative to the other policy scenarios) and also possibly because of greater learning by doing in the first stage, which can substitute for R&D. 
• For fossil fuel generators, output shares remain steady in the two periods, with the lower output in the first stage and higher in the second, compared with other scenarios. In other words, cost reductions in renewables allow for fossil sector expansion. Still, short-term transitional costs could be expected to be greater under the RPS than in other scenarios. Actual potential costs to the fossil sector under an RPS will be higher if they are not fully able to pass along the costs of green certificate costs to consumers. 
• For government, the RPS has a neutral impact, with no revenue and no program disbursements. The implicit subsidy to renewables producers is fully funded by the implicit tax on fossil producers (and consumers) in the form of the green certificates. 
• From society’s perspective, the welfare costs in our estimates are greater than the emissions price and generation subsidy, but lower than the combination and R&D subsidy. This ranking does not necessarily hold under all circumstances, but rather depends on the particular trade-off between the extra costs of encouraging more effort up front and the inefficiencies of not giving consumers incentives to conserve. Indeed, if one coped with the former problem by optimally designing the RPS requirement to increase over time, the RPS could be made to dominate the subsidy always, due to the presence of the modest conservation incentive. 
• Looking beyond the electricity sector, the increase in electricity prices risks causing some economy-wide competitiveness impacts such as decreased productivity or reduced exports, but these effects will be less severe than with the emissions price, particularly in the second stage. The RPS strikes a certain balance in coping with cost uncertainties over time, with less emission reduction uncertainty than a fixed emissions price, and less price uncertainty than a fixed cap.

5.5.4 Renewable Generation Subsidy

This fiscal instrument includes a range of possible policies that subsidize renewable generation (e.g., tax credits, direct subsidies) to encourage the expansion of carbon-free generation; however, they do nothing to encourage conservation or reduce the emissions intensity of fossil generators. As well, there is no impact on the price of electricity and thus consumers are not encouraged to reduce demand and therefore carbon emissions. Hence, much more effort must be expended on higher-priced renewables to displace fossil generation and meet the carbon reduction target. The simulations show that consumer prices remain stable, while renewable generation expands more than with the emissions price, particularly in the second stage, bringing with it greater investment in R&D to reduce costs. This extra expansion is necessary to make up for the lack of incentives to abate directly or to conserve. The effective subsidy that achieves the emission reduction target is $0.006 per kWh. In this scenario, we estimate the generation subsidy to cost 23% more in terms of welfare than the emissions price. This welfare cost seems quite modest, and we note some important caveats. First, the target is relatively modest, given the near competitiveness of renewable supply; the need for more substantial reductions could magnify the limitations of the single-lever approach. Second, consumer demand is assumed to be quite inelastic; if demand were more sensitive to price changes, conservation incentives would be much more important. Third, it indicates that abatement opportunities are not as cost effective at low prices. Finally, we do not account for any inefficiencies in raising the substantial public revenues needed to fund subsidies to the renewables sector. The first and third issues are explored to some extent in Section 5.6. Surprisingly, the subsidy also costs less from a welfare perspective than the RPS. This result was somewhat unexpected, since the RPS does include some incentives for demand reduction. However, as we noted, the magnitude of this inefficiency loss is dwarfed by the costs of changing the timing of the renewables expansion under a fixed RPS. Both policies require greater expansion than the emissions price, but the path of the renewables subsidy more closely follows that of the optimal one. From a distributional perspective:

• Consumer prices are not impacted in the subsidy scenario, since all of the reductions are supplied through lower renewables costs, which do not affect the fossil fuel sector directly. Consumers would be indirectly impacted since it is their tax revenue that funds some portion of the subsidy transferred to the renewables sector. 
• For renewables producers, generation subsidies have the largest impact on profits, since they must be encouraged to displace more fossil output than the preceding scenarios. Ongoing innovation is stimulated by the greater scope to reduce production costs at the higher output levels induced by the price premium. 
• For fossil fuel generators, the generation subsidy has a similar impact on fossil output as the emissions price, since the additional renewable generation is partly offset by additional demand. The decline is slightly larger in the second stage, due to the more dramatic increase in the competitiveness of renewables from innovation. That fossil output may be lower with the subsidy than with the emissions price may seem surprising, since the electricity price increase is absent, but since the fossil sector lacks an opportunity to adjust its own emissions, the full burden of reductions falls on renewables to displace fossil output. 
• For government, the subsidy required to achieve the emission reduction target is a significant disbursement. 
• From society’s perspective, the welfare costs are greater than the emissions price. With respect to reaching the emissions target, the renewables subsidy is likely to suffer from greater uncertainty than the preceding policies. Although this is not modelled, the reasoning is twofold:
  • First, the uncertainty over the scope and speed for cost reductions in renewables is likely to be greater than the uncertainty surrounding the costs of abatement in the fossil sector or conservation by consumers.
  • Second, even if all cost uncertainties were similar, the reliance on only one method of emissions reductions raises overall uncertainty. Otherwise, if innovation does not lower renewable production costs significantly, one could engage in relatively more emissions abatement or conservation, whichever turns out to have the lower costs.

Consequently, the renewables subsidy alone has more uncertainty regarding the emissions that will be reduced. It also has more uncertainty about the revenue requirement. If costs fall more than expected, a high subsidy would induce an oversupply relative to the carbon target, reflecting additional efficiency loss, as well as lost public funds. If costs do not fall as expected, either emissions targets will not be met, and some public funds will be saved, or the subsidy must be increased even more to meet them, requiring greater than expected outlays.

5.5.5 A Combination of RPS and Generation Subsidy

Particularly in renewable energy, a combination of policies is often implemented, partly out of overlapping jurisdictions of the federal, provincial and local governments, and perhaps out of a diversification motive. In response to a request by the Scoping Group, we have estimated the effects of placing a portfolio standard and a renewable production subsidy in place simultaneously. The key result is that the subsidy weakens the effect of the portfolio standard and raises costs slightly. With both policies, the fossil fuel producer must still purchase “green certificates” for every unit of electricity generated. For the renewables producer, there are now two subsidies – the value of a green certificate, and the direct subsidy. Since the direct subsidy boosts renewable supply, the equilibrium price of a green certificate does not need to be as high to reach the portfolio standard (as compared to the RPS implemented in isolation). Consequently, when the policy target is a portfolio share, a direct subsidy to renewables primarily offsets the burden to fossil producers and consumers instead. Another way to think of this problem is to recall that the value of the green certificate represents the price differential between the market price of electricity and the price received by renewables; since the direct subsidy raises the price received, that differential is lowered accordingly. However, the effective RPS subsidy is not completely crowded out by the direct subsidy. The lower certificate price means lower electricity prices and less conservation, requiring a bit more renewable production to meet its share of the extra electricity demand; in our case, with a $0.002 subsidy, we estimate the RPS must rise slightly to 24.2%. In other combinations, it is also possible – particularly in the second stage after costs fall – for the subsidy to be strong enough to ensure that the portfolio standard is more than met, recognizing that the quantity of renewables supplied is a function of prices received and knowledge in current and future periods (which lowers renewables cost and increases uptake). For example, if the RPS were set at 18% instead and combined with the $0.006 subsidy, the subsidy would be the driving policy, and green certificates would have no value. Since we assume the RPS is the driving policy instrument in our combination scenario, the distributional effects are quite similar to the RPS alone. The slight differences are as follows:

• Consumer prices are slightly lower. Despite the additional electricity demand, emissions are also lower in the first stage. This results from the fact that the standard must be raised to offset the loss of conservation incentive, leading to even more reductions in the first stage and less in the second.
• Renewable production is 0.5% higher and R&D spending is 1.5% higher.
• For fossil fuel generators, between the lower certificates cost and the additional renewable generation, output is nearly unchanged relative to the portfolio standard alone.
• Perhaps the most telling effect is that the government in this combination scenario spends just over $1 billion on a subsidy that has little or no effect on behaviour, given the presence of the RPS.
• From society’s perspective, to the extent the subsidy does affect behaviour, it tends to lower prices and raise overall welfare costs. The weaker conservation incentive and the additional front-loading of emissions reduction efforts by increasing the RPS are the cause of the increase in welfare costs, from 1.8 to 1.83 times that of the emissions price.

5.5.6 Renewable Research Subsidy

The renewable research subsidy uses current investments in reducing costs to increase future renewable production. Since it does not change any price incentives for demand or production, nor change current costs, all the burden of emissions reduction is placed on future displacement of fossil by renewable generation. Furthermore, given the lack of future production incentives, the required cost reductions are large, and the required investments even larger. The ability for an R&D subsidy alone to deliver all of this is clearly an area of uncertainty. In the simulations, to achieve the same emissions reductions as the $10/tC emissions price, a 61% subsidy per year above forecast 2010 levels of $129 million to R&D must be offered. A small increase in first-stage renewable generation may reflect some complementarity between learning by doing and R&D. However, nearly the entire increase is in the second period. The cost of this delay and lack of incentives for other actors in the electricity market is a fourfold increase in the welfare costs above the emissions price instrument. That said, the uptake and cost decreases of renewable are ultimately maximized under the R&D subsidy – because it is targeted to meet emissions goals. This instrument clearly supports innovation in the sector, but that is because it relies on no other method for reducing emissions. The price of this uptake is borne entirely by the government sector through large transfers to the renewables sector. Electricity prices remain unaffected, thus reducing any direct impacts on consumers. Of course, the tax revenue is an indirect cost borne in part by the consumers. From a distributional perspective:

• Consumers do not experience electricity price increases and consumer surplus losses under the R&D subsidy. As with the generation subsidy, they indirectly contribute to the renewables sector through tax contribution to fund the R&D subsidy. 
• For renewables producers, the R&D subsidy induces the highest penetration in the second stage. This penetration is driven exclusively by innovation and cost decreases from renewable production. An important caveat is the degree to which Canadian learning by doing and R&D can drive cost decreases in renewables. While such production cost decreases are observed in Canada and internationally, it is not certain that price decreases can occur through Canadian R&D alone that are sufficient to achieve the high levels of renewables penetration predicted in this scenario. This is particularly questionable since, as a general observation, innovation in renewable production occurs internationally and is imported into Canada. This uncertainty in the ability of domestic R&D subsidies to achieve the penetration predicted in the model only reinforces the result that this policy is a much more costly method for achieving emissions reductions. 
• For fossil fuel generators, the R&D subsidy does not impact electricity price, but does significantly reduce fossil output in the second stage. Although not modelled, costs associated with stranded assets or variable costs due to lower capacity utilization could occur. But transaction costs associated with decreased fossil demand are likely lower in this scenario, since a majority of reductions occur in the second stage. Thus, the transition period for the fossil sector to adjust to decreased demand is long and has the potential for costs to be minimized. 
• For government, the R&D subsidy requires the largest disbursement of the instruments. That said, promoting innovation is a government policy and therefore R&D programs are generally part of a desirable policy approach to decarbonization. However, given the longer-term nature of the reductions associated with R&D, a government faced with a carbon reduction target would likely not achieve significant reductions in the short term under an R&D program. 
• From society’s perspective, the welfare costs are greatest under the R&D subsidy. Another negative consequence of this scenario is uncertainty. For similar reasons to the renewable generation subsidy, the uncertainty of renewable cost reductions makes this a relatively risky policy for promoting carbon reductions – all the more so, since in the absence of cost reductions, there is no incentive for additional renewables uptake, in either stage. Given the uncertainty about innovation success more generally, and the impact of domestic efforts more specifically, it is highly uncertain that a domestic R&D program alone could achieve a significant carbon reduction target through renewables uptake. Instead, an R&D subsidy could be viewed as a complementary instrument that can be used to achieve longer-term societal goals such as promoting innovation.

5.6 Sensitivity Analysis

To further test the robustness of the results presented in the preceding discussion, it was agreed that a sensitivity analysis would be conducted with respect to the following:

• An increase in the baseline electricity price
• An increase in the baseline price of natural gas.

For each of these sensitivity cases, we re-estimate the baseline scenario and the policy scenarios that achieve reductions equivalent to a $10/tonne CO2 emissions price. We do not report on the sensitivity of the R&D subsidy instrument, as it was concluded by the Scoping Group that the scenario is not plausible.

5.6.1 Increase in Baseline Forecast Electricity Price

Uncertainty exists over the future price of electricity. The assumed price of electricity impacts the supply of renewables in the baseline as well as the size of the incentives and costs required to comply with the fiscal instrument. Uncertainty in the electricity pricing assumption is introduced into the model in three ways:

• First, the price assumed in the modelling is adopted from the CEOU99. This estimate is somewhat dated and, indeed, recent forecasts predict higher oil and natural gas prices in the future, which translate into higher electricity prices.
• Second is the impact of reduced coal in the baseline, where a partial or full coal phase-out in Ontario will increase electricity prices as generation shifts to more expensive natural gas or nuclear power.
• Third, air emission control polices for carbon, criteria air contaminants and toxics will increase electricity costs in the future.

To test for the impact of an increase in electricity prices, a separate electricity price sensitivity run was conducted, in which the base electricity price was increased by 50%. A 50% increase in electricity price is a reasonable assumption given that a tripling of the assumed natural gas prices in the CEOU99, which is consistent with current natural gas prices in 2004, would impact electricity prices in the order of 50%. The impacts resulting from changing the electricity price assumption are summarized below and presented in Exhibit 5.3.

• Carbon emissions can be expected to be lower in the base case due to the higher penetration of renewables. This occurs since the higher electricity price makes renewables more cost competitive with fossil fuel generation, therefore reducing fossil output and carbon emissions. Not all of this percentage reduction translates into reductions under the policy scenarios; since additional increases in renewable production are more costly, the emissions price does not induce as much additional emissions reductions. Furthermore, the fossil generation that remains is more carbon intensive. 
• Renewable output expands due to the increased competitiveness of renewables when electricity prices are high. However, while renewable output is higher, the policy-induced increases are smaller in all scenarios. The reason is that higher prices drive renewable supply further up the marginal cost curve, so additional increases in price induce less expansion. Another effect is to flatten the path of emissions reductions, since the electricity price increase is far larger than any of the differentials created by the policies, and since this big boost is felt in both periods. 
• The main impact is that welfare costs drop significantly with a higher electricity price assumption. This result holds despite the fact that the electricity price, and thereby lost consumer surplus, is higher. Generally, the smaller the price differential between renewables and fossil fuel generation the lower the welfare implications of the instruments. A shift in renewables costs relative to fossil fuel generation costs can also be expected to trigger similar results. 
• The change in total electricity output is small, since initial output is assumed to be fixed. That said, the consumer surplus under the RPS changes dramatically due to the larger increase in electricity prices, relative to the reference scenarios. 
• These impacts are magnified under the other policies. The electricity price increase due to green certificates is higher by $0.001 in both stages, and the subsidy to achieve the same reductions as the emissions price is also higher. In welfare terms, the marginal cost of expanding renewables is so much higher that the welfare loss from the improper timing of reductions under the RPS now almost balances the inefficiency of not inducing additional conservation under the renewable generation subsidy. 
• Policy impacts on producer surplus are higher in all cases, since the higher baseline renewables output gains revenue from the increases in the price differential, while those price increases induce less expansion. 
• The change in R&D spending falls in all scenarios, due both to the greater learning by doing and to the smaller increase in renewable output induced by the policies. 

The sensitivity testing shows that when electricity prices are high and the price differential is small, a renewable generation subsidy is more desirable than an RPS (as in the reference case in Exhibit 5.1). Conversely, when electricity prices are low, an RPS is more desirable than a generation subsidy from a welfare perspective. To understand this, recall that welfare is measured through changes in three outcomes:

1. Consumer surplus is increased when the instrument raises electricity prices. Under an RPS, the price of the green certificates is passed on to consumers, thus triggering consumer surplus losses, or costs. This effect does not occur in the generation subsidy scenario, since fossil producers are not impacted by the subsidy. Under a high price differential between renewables and fossil fuel generation, the size of the consumer surplus increases significantly under an RPS but not under the generation subsidy. This effect on consumer surplus primarily explains why an RPS is less desirable than a generation subsidy when electricity prices are high. 
2. Producer surplus increases with the price differential, since higher electricity prices allow more renewable producers to supply output at a profit. Under both an RPS and generation subsidy, the producer surplus increases and thus does not significantly impact the relative desirability of either instrument. 
3. Transfers only occur under the generation subsidy and they increase as the differential increases between renewables and fossil generation (as measured by the electricity price). Thus, the size of the transfer grows when the price differential increases, making the generation subsidy less attractive than an RPS under a low electricity price scenario.

The sensitivity analysis shows that the price differential between renewables and the electricity price is an important determinant of the size of the welfare cost. As well, the desirability of an RPS versus a renewable generation subsidy is affected. These results can also be expected when the price of renewables changes, where a decrease in the price of renewables would produce results that are directionally similar to an increase in the electricity price. Exhibit 5.3 Sensitivity Case No. 1 Results – 50% Increase in Electricity Price % Change (%D) from Exhibit 5.1 (Reference Case) to Sensitivity Case

	Base Case		Emissions Price		Renewable Portfolio Standard		Renewable Generation Subsidy
	Sensitivity	%	Sensitivity	%	Sensitivity	%	Sensitivity	%
1. Policy level for 12% emissions reduction			10 $/tCO2	0%	42%	74%	0.0066	10%
2. Electricity price (in $/kWh)
1st stage	0.125		0.130	34%	0.129	36%	0.125	-86%
2nd stage	0.125		0.130	34%	0.127	37%	0.125	-86%
3. Carbon emissions (MT CO2)
1st stage	73	-31%	66.19	-33%	63.06	-31%	66.42	-33%
2nd stage	69	-32%	60.18	-29%	63.49	-31%	59.93	-28%
4. Renewable output (MWh 10^11)
1st stage	0.91	214%	1.04	160%	1.08	97%	1.04	145%
2nd stage	0.99	161%	1.15	75%	1.08	96%	1.16	61%
5. Fossil output (MWh 10^11)
1st stage	1.38	-31%	1.22	-34%	1.19	-31%	1.25	-33%
2nd stage	1.30	-32%	1.11	-30%	1.20	-31%	1.13	-28%
6. Total electricity output (MWh 10^11)
1st stage	2.29	0 %	2.261	0%	2.266	0%	2.29	0%
2nd stage	2.29	0 %	2.261	0%	2.281	0%	2.29	0%
7. Renewable R&D ($M)			221	-51%	187.39	-41%	249.2	-53%
8. Additional renewables cost reduction			7.52 %	-50%	7.57%	-42%	7.899%	-51%
9. Consumer surplus ($M)			-$11,724	0%	-$6,651	47%	$0 $	0%
10. Producer surplus ($M)			$5,193	134%	$6,167	77%	$6,570	131%
11. Transfers ($M)			$6,152	-31%	$0	0%	-$7,044	98%
12. Welfare – no benefits ($M) (9+10+11=12)			-$379	-34%	-$484	-54%	-$474	-167%

5.6.2 Increase in the Price of Natural Gas

It can be expected that higher long-term natural gas prices will increase the cost of carbon abatement from combined-cycle plants. The result of higher natural gas prices is a decrease in the competitiveness of combined cycle as a carbon mitigating technology relative to RETs. We can therefore expect to have less abatement and more penetration of RETs in the emissions price scenario. As with electricity, we increase the price of gas by 50% in the estimate of variable costs for combined-cycle natural gas plants. The impacts associated with the natural gas price increase are summarized below: The increase in natural gas price has no impact on the baseline, since without a binding policy instrument in place no abatement occurs.

• In the emissions price scenario, fossil emitters can either abate internally or use RETs to achieve their emission reduction constraint. With higher natural gas prices increasing the cost of internal carbon abatement, as expected, more RETs are deployed. This deployment does not, however, have a large relative impact on emissions. 
• For the RPS and generation subsidy, there is no incentive for the fossil sector to abate internally, and thus the rising natural gas price has virtually no impact on the predicted outcomes.

In conclusion, the sensitivity results indicate that increasing natural gas prices has a minimal impact on the outcomes with respect to the reference case. As discussed in the previous scenario, however, increasing gas prices could increase the price of electricity, and this effect would have a real impact on the scenario outcomes.

5.7 Conclusion

We conclude that the results are robust to changing key variable assumptions. Indeed, our core observation holds: the economic efficiency and environmental effectiveness of the EFR instruments are linked to their ability to influence the entire electricity market, and three decarbonizing drivers in particular. As a general rule, an EFR instrument will be more efficient and effective if it signals to multiple agents in the electricity market that carbon is more expensive: fossil producers will reduce their emissions intensity; renewables producers will supply more output when the price differential between renewables and fossil generation decreases; and consumers will take measures to conserve electricity, reduce demand and displace fossil output. This finding holds under multiple input variables and explains why an emissions price is preferable to an RPS or renewable generation subsidy. A good example of the increased risk in using a single instrument is highlighted by the R&D instrument scenario, where the emission reduction is entirely reliant on the ability of R&D investments to reduce renewables costs through innovation. While cost reductions can be expected to occur from R&D spending, the scope and scale of the cost reductions is questionable, thus increasing the overall uncertainty in the instrument.