Appendix A: Analysis and Assessment of Individual Measures in the Statement¹

1. Regulatory Framework for GHG Emissions

Summary of the Initiative and Emissions Projections

The Regulatory Framework for GHG Emissions (Canada, 2007b) imposes emissions reductions on Large Final Emitters (LFE), forcing affected firms to achieve an 18% reduction in GHG intensity from 2006 levels beginning in 2010, with a further 2% improvement required in each year thereafter. Affected firms may comply with the regulations either through internal abatement, through contributions to a climate change technology fund (at an initial rate of $15/tonne), by purchasing the right to claim emissions reductions made by other domestic firms through the emissions trading and offset systems, or by purchasing emissions reductions credits through the CDM mechanism defined under the Kyoto Protocol. Firms may also claim a one-time credit for GHG reductions between 1992 and 2006.²

Methodological Approach

The estimated emission reductions are based on output from Environment Canada’s E3MC model, which provides an integrated view of the effects of the proposed regulations. Actual industrial emission levels depend on the compliance options chosen by regulated firms, for which Environment Canada provided preliminary estimates. For the reported emissions reductions, the breakdown by compliance mechanism is shown in Figure 1.

Table 1: Emissions Reductions Attributed to the Regulatory Framework for GHG Emissions

Program	Projected Emissions Reductions in Mt					Key Determinants of Results	Predictive Accuracy
	2008	2009	2010	2011	2012
Regulatory Framework for GHG Emissions	0	0	49	53	58	relative contribution of internal emissions reductions and Technology Fund contributions accounting for Technology Fund contributions	Likely overestimate

Figure 1 - Allocation of emissions reductions across compliance mechanisms*

 

* This table was constructed by the NRTEE using the government's numbers presented in the Plan and the Reference Case.

Analysis

Intensity targets regulate the quantity of emissions allowable per unit of output or value-added, and therefore add a degree of uncertainty with regard to the eventual emissions profile as compared with an emissions cap. In order to properly estimate the eventual emissions reductions, reliable estimates should be available for each of the following:

• the evolution of emissions intensity and economic activity absent the policy,
• economic activity under the policy, and
• compliance behaviour and use of flexibility/compliance mechanisms under the policy.

In order to understand how these emissions trends will evolve under the Regulatory Framework for GHG Emissions, use of an integrated model is particularly important. The policy affects some of the largest sectors of the Canadian economy, and thus will have important spillover or secondary effects. Also, the response of other markets will determine the availability of domestic offsets to supplement internal emissions reductions and the purchase of international permits as a means of complying with the policy. While many of the details of the policy have yet to be finalized, the key driver of the results will be the marginal cost of emissions faced by affected firms, and this will be determined in the early years of the policy largely by the fixed rate of contribution to the Technology Fund, and to a greater degree in later years by the domestic and international offset markets.

Estimates in the Plan treat contributions to the Technology Fund as being equivalent to emissions reductions realized in the time period in which the contributions are made. However, these contributions will be used to finance emissions reductions programs that will result in an undetermined number of future emissions reductions; conceivably, these reductions could be more, or less, than 1 tonne of realized reductions per $15 invested. This represents an important inconsistency in accounting for emissions reductions.

While the Plan appears to overestimate realized reductions in the 2010-2012 period, the lack of realized reductions for the 2008-2009 period is also inconsistent with other modelling outcomes. As firms engage in early action to reduce their eventual compliance costs, some reductions relative to the Reference Case are likely to occur in the two initial years of the commitment period.

Conclusions

The above evidence suggests that significant emissions reductions and contributions to future emissions reductions will result from the Regulatory Framework for GHG Emissions, but that an immediate accounting of expected emissions reductions accruing as a result of contributions to the Technology Fund leads to a likely overestimation of the emissions reductions realized. Regardless of the eventual emissions reductions that occur as a result of the Fund, it is inconsistent to treat investment in potential emissions reductions as being equivalent to realized present-day emissions reductions.

2. Regulating Energy Efficiency

Summary of the Initiative and Emissions Projections

As part of the Regulatory Framework for Air Emissions, the government proposes to update existing standards for 12 product categories, and introduce new energy efficiency standards for 20 more between 2007 and 2010, as well as introduce a ban on incandescent light bulbs that would begin in 2012.

Methodological Approach

The emissions reductions provided in the Plan were calculated assuming that, in the case of appliance replacement, the energy savings resulting from the regulations would be proportional to the difference in energy consumption between a benchmark regulated unit and the least efficient product currently sold in Canada. According to information provided to the NRTEE by Natural Resources Canada (NRCan), no consideration was explicitly given to the rebound effect. No specific methodology was provided for the calculation of projected emissions reductions from the ban on incandescent bulbs.

Table 2: Emission Reductions Attributed to Energy Efficiency for Household Goods

Program	Projected Emissions Reductions in Mt					Key Determinants of Results	Predictive Accuracy
	2008	2009	2010	2011	2012
Regulating Energy Efficiency	.61	.96	1.3	1.4	7.1	rebound effect policies are compared to the lowest efficiency case pace of benefits from light bulb ban	Likely overestimate

Analysis

Reducing emissions by improving efficiency standards poses similar challenges to those presented by intensity targets. When the energy consumption of a particular device is regulated, and attempts are made to translate this regulation into a reduction in total energy consumed and/or emissions levels, several ancillary assumptions must be made. The only ways in which the decrease in energy use from the device will correspond directly to the percentage increase in energy efficiency, as is assumed in the calculations in the Plan, is if replacement rates remain the same, the use of old appliances is discontinued, and the intensity of use does not change. There are three principal reasons that estimates based on such assumptions will lead to an overestimate of the impact of efficiency standards. First, people may continue to use an older unit as a secondary device (the "beer fridge effect"), they may purchase a larger but more efficient replacement unit, or they may use the new appliance more intensively (the rebound effect). The occurrence of any of these effects implies that the reduction in total energy use will be less than the increase in energy efficiency. Also, efficiency gains mean that appliances are less costly to operate and so people who otherwise would not have purchased one may choose to do so, and the total number of appliances in use would therefore increase, which also negates some of the emissions reductions.

Empirical evidence suggests that the rebound effect is important to consider when regulating appliance efficiency, but that these changes in behaviour are unlikely to negate all the gains from energy efficiency. Nadel (1993) examined more than 40 studies of energy efficiency programs and found evidence of rebound effects only for compact fluorescent lamps and air conditioners (Nadel refers to the rebound effect as a "snapback effect"). A recent study by Davis (2007) provides strong evidence of the increased use of higher-efficiency household appliances. He shows that when randomly chosen homeowners are given washers that are on average 48% more energy efficient, they wash 5.6% more clothes. And so the total resulting energy (and emissions) reduction is just 42.4% rather than 48%. Additional studies by Hausman (1979); Dubin and McFadden (1984); Dubin (1985); and Dubin, Miedema, and Chandran (1986) show similar patterns of increased usage intensity after the acquisition of more efficient appliances.

The data do provide strong support for the role of regulation in driving incremental changes in efficiency, and therefore reductions in energy use. Between 1972 and 2001, average electricity consumption of central air conditioners and refrigerators decreased by 44% and 56% respectively, and this was representative of gains among other products (Davis, 2007). Strong evidence for the role of regulation in affecting this trend is also provided in Nadel (2002) who shows that, for refrigerators, the efficiency improvements are not smooth, but rather follow almost exactly the stringency of efficiency standards. Nadel states that in years in which market forces prevailed, and there were no new standards, there were also few efficiency improvements. As a result, treating the set of appliances currently available for sale as the no-policy benchmark may not be inconsistent for the short-term. However, assuming that the same replacement decision would have been made and each would have chosen the least-efficient model available represents a best-case scenario and so is too strong an assumption. In the absence of the policy, some replacements would certainly be undertaken with appliances other than the least-efficient model.

Empirical evidence for Canada does support the Plan’s contention that some emissions reductions will be induced by tighter emissions standards. Jaccard and Rivers (2007) discuss the likely consequences of a set of improved standards for energy-using products in Canada based on the historic relationship between energy demand and regulation, and find that likely emissions from their chosen standards are in the order of 1.5 Mt in 2010. It is not possible to directly compare the projections in the Plan to the results of Jaccard and Rivers (2007) since they are not based on the same standards as those used to compute the emissions projections specified in the Plan.

In addition to the concerns expressed above with respect to the estimation of the impact of the policies, a potential issue of additionality exists with the manner in which these results are treated. Since the impacts are expressed as reductions relative to the Reference Case, there should be careful consideration of which efficiency improvements have been already taken account of in the Reference Case projections. Specifically, consider that the Reference Case accounts for some new equipment regulations and standards that will require new gas furnace efficiency to be 90% in 2009 and new gas boilers to be 85% efficient by 2010. Since specific details of the standards used to calculate the stated reductions in the Plan are not given, it is not possible to say whether these are the same as some or all of those already included in the Reference Case, although information provided suggests that additionality of these emissions reductions has been assured.³

In the Plan, particular attention is given to an effective ban on incandescent light bulbs as a source of increased energy efficiency. While it is true that compact fluorescent light bulbs convert electricity to light at a much more efficient rate than incandescent lights, several caveats must be considered before accepting the claim that moving from a 60-watt incandescent bulb to a 15-watt compact fluorescent will lead to 75% reductions in energy use. Jaccard et al. (2006) discusses how previous projections of the market acceptability of energy-efficient substitutes for incandescent bulbs greatly over-estimated their uptake. Greater up-front costs, high premature failure rates, unfamiliar light quality, and incompatibility with household fixtures and switches led to greater financial risk and a product that was not as "cheap" as a watt-for-watt or lumen-for-lumen comparison would suggest. In this measure, the regulation of the technology will likely result in further innovation, improved quality, and-eventually-lowered cost and increased uptake.

This being said, the forecasted emissions reductions seem likely to be overestimated. Estimates produced by the Ontario Ministry of Energy state that replacing all incandescent light bulbs in Ontario would reduce electricity consumption by 6 TW-h, or just over 4% of total electricity consumption.⁴ If we extrapolate this to Canadian electricity sales, this would imply a savings of 20TW-h per year. At current emissions intensities, 20TW-h of generation produces 4.1 Mt of carbon emissions. This corresponds exactly with the Plan’s estimate of the savings in 2012.⁵ But since the Canada-wide ban is only slated to take effect in 2012, realized emissions reductions should be lower in the initial years as some incandescent bulbs remain in use. For example, in Australia, a ban imposed in 2008 is not expected to reach peak emissions reductions until 2015.⁶

Conclusions

Given the fact that the estimates provided do not explicitly account for the rebound effect of increased intensity of use or increased total appliance stock, and given that the estimates have assumed the best-case scenario that all replacements would have chosen the least energy efficient model available absent the policy, it may be concluded that the gains from improved standards are likely overestimated. However, since the exact standards are not defined, it is not possible to compare the Plan’s projections to estimates that correct the above assumptions in order to asses the magnitude of any overestimation. This lack of specific detail also makes it difficult to evaluate the degree to which stated emissions reductions should be understood to be additional to those attributed to more stringent efficiency standards already accounted for in the Reference Case. While this policy to ban incandescent light bulbs will undoubtedly result in emissions reductions, the 2012 number is a reasonable estimate of the eventual reductions once all incandescent bulbs are removed from use, but is not an accurate measurement of the reductions accruing in the first year of the ban. It is expected to take several years for this measure to reach its full potential.

3. Motor Vehicle Fuel Efficiency
Memorandum of Understanding (MoU)

Summary of the Initiative and Emissions Projections

A Memorandum of Understanding (MoU) between the Government of Canada and automakers aims to reduce GHG emissions from motor vehicles on the road by 5.3 Mt/year by 2010. The 5.3 Mt/year target is measured from a benchmark level of emissions from the vehicle fleet in absence of any action.

Methodological Approach

The MoU was signed before the development of the Reference Case, and is explicitly included in the emissions projections contained therein (see Canada 2006c, Annex II). As a result, the estimation of the reductions is not relevant for the present study.

Analysis

Measures enacted under the Motor Vehicle Fuel Consumption Standards Act will impose more stringent fuel-efficiency ratings for Canadian vehicles, however the Plan clearly states that the estimates of emissions reductions accruing from these fuel-efficiency changes are preliminary, and as such are not included since they are based on unknown standards. The stated emissions reductions thus include only those accruing as a result of the MoU. This poses a significant problem of additionality, since the MoU is already included in the Reference Case. As a result, counting these emissions as part of reductions relative to the Reference Case implies counting the same reductions twice.

Table 3: Emission Reductions Attributed to Vehicle Fuel Efficiency Standards

Program	Projected Emissions Reductions in Mt					Key Determinants of Results	Predictive Accuracy
	2008	2009	2010	2011	2012
Vehicle Fuel Efficiency Standards	3.0	3.9	5.3	5.3	5.3	additionality of reductions	Reliable, but should not be included

Conclusions

Given the additionality concerns expressed above, emissions accruing as a result of the MoU should not be counted as reductions against the Reference Case. Any emissions reductions attributed to the MoU in the context of the Plan would constitute an overestimate.

4. Regulating Renewable Fuels Content Standards

Summary of Initiative and Emissions Projections

Regulations detailed in the Plan will require 5% renewable fuel content by volume for gasoline from 2010 and 2% by volume for diesel fuel and heating oil by no later than 2012.

Methodological Approach

The estimates are derived by estimating aggregate volumes of biodiesel and ethanol produced, and calculating emission reductions using conversion factors that specify the percentage by which total GHG emissions are reduced when gasoline and diesel are produced from biomass rather than from petroleum. The emission reduction factors for Regulating Renewable Fuels Content are 1.25 Mt of GHG emissions reduced per 1 billion litres of ethanol and 2.2 Mt per billion litres for biodiesel. The former corresponds to GHG reductions of 33.1% relative to gasoline produced from petroleum sources, while the latter represents a 66.5% reduction relative to the conventional production of diesel.⁷

Analysis

Renewable fuels content standards do not directly regulate GHG emissions, but rather seek to indirectly regulate the inputs to gasoline and other fuel production. In order for a renewable fuels content standard to reduce GHG emissions, two conditions must be verifiable:

1. The renewable fuels content requirement does not lead to increases in production of gasoline and diesel fuels in general.
   
2. Based on a full-cost accounting, the production of the renewable fuel has lower GHG emissions than comparable petroleum-based production.

Table 4: Emission Reductions Attributed to Regulating Renewable Fuels Content Standards

Program	Projected Emissions Reductions in Mt					Key Determinants of Results	Predictive Accuracy
	2008	2009	2010	2011	2012
Regulating Renewable Fuels Content Standards	1.3	1.3	2.9	2.9	4.1	full cost accounting of renewable fuels' induced emissions additionality uncertainty not considered	Likely overestimate

Optimal business management suggests that the first of these conditions is likely to be satisfied. It would be implausible for a regulation to reduce the costs of providing fuel by requiring a change in input mix, at least in the short term. Simply put, if it were cheaper to provide a renewable content at or above 5%, companies would be doing so now.

Empirical estimates show that the second of these conditions is likely true, but the magnitude of the emissions reduction factor may be lower than that chosen in the Plan. Farrell et al. (2006) show that production of gasoline using ethanol reduces petroleum use on average by about 95% relative to conventional refining, but that GHG reductions are only reduced by about 13% relative to conventional gasoline production. In fact, Farrell et al. argue that "the impact of a switch from gasoline to ethanol has an ambiguous effect on GHG emissions, with the reported values ranging from a 20% increase to a decrease of 32%." Updated estimates published as a correction to the Ferrell et al. article suggest a point estimate of net GHG for corn ethanol of 18% below conventional gasoline, but with a possible range of 36% fewer emissions to 29% more emissions.⁸ Hill et al. (2006) find similar results for gasoline, and find that GHG emissions are reduced by 41% for biodiesel relative to the fossil fuels they displace. Further, they note that "these estimates assume these biofuels are derived from crops (corn) harvested from land already in production; converting intact ecosystems to production would result in reduced GHG savings or even net GHG release from biofuel production."

A further issue with regard to additionality arises when considering the quantities used to calculate emissions displacement. Figures presented are based on the expansion of ethanol production to 2.2 billion litres after 2010, and the expansion of biodiesel production to 600 million litres in 2012. For ethanol production, this represents an increase of 900 million litres over the Reference Case for 2010. Using the conversions from the Plan, this would only account for an additional 1.125 Mt of reductions in 2010, rather than the figure of 2.9 Mt stated in the Plan, which corresponds to more than the predicted emissions displaced by all ethanol produced, including units already counted in the Reference Case.⁹

Conclusion

The evidence above suggests that two factors contributed to the likely overestimation of the emissions reductions resulting from the renewable fuel content standard. First, the emissions reduction factors used are higher than those cited in the current scientific literature. Second emissions reductions already included in the Reference Case are counted a second time.

5. ecoENERGY for Renewable Power

Summary of Initiative and Emissions Projections

According to the Plan, the ecoENERGY for Renewable Power program will invest $1.5 billion dollars to provide incentives to increase Canada’s supply of clean electricity. The goal of the program is to encourage the production of 14.3 terawatt hours (TW-h) of new electricity from renewable energy sources, which is enough electricity to power about one million homes.¹⁰ The program provides an incentive of one cent per kilowatt hour for up to ten years, which will reduce the cost gap between new technologies and traditional sources of electricity.

Table 5: Emission Reductions Attributed to ecoENERGY for Renewable Power

Program	Projected Emissions Reductions in Mt					Key Determinants of Results	Predictive Accuracy
	2008	2009	2010	2011	2012
ecoENERGY for Renewable Power	2.2	3.7	5.5	6.7	6.7	actual displacement of existing capacity nature of displaced alternative additionality free ridership	Likely Overestimate

Methodological Approach

The estimates above are calculated on the basis of renewable energy supplies of 4.7 TW-h in 2008, 8.0 TW-h in 2009, 11.7 TW-h in 2010, and 14.3 TW-h for 2011 and 2012. The estimates of emissions reductions are derived using a conversion factor of 0.4564 Mt/TW-h.

Analysis

As with all subsidy programs, it is difficult to establish the incremental impact of the ecoENERGY for Renewable Power program. Program designers have recognized this problem, and according to NRCan, "where a renewable electricity generation project is developed at a site where no previous electrical generation existed, it would clearly be considered 'incremental’."¹¹ This does not, however, constitute incremental generation for the purposes of evaluating policy-induced emissions reduction. In order for emissions reductions to be clearly attributed to increased renewable generation under the RPPI, one of two factors must be demonstrable:

• The production facility would not have been built without the subsidy, and the new facility replaces an existing one with a higher rate of emissions; or
  
• The production facility would have been added without the subsidy, but the facility would have been more emissions intensive.

In either of these cases, the new capacity would represent emissions reductions as a result of displacing a more emissions-intensive alternative. Statements in the Plan suggest that all new renewable energy production eligible for financing under the RPPI would be considered as contributing to emissions reductions, thereby ignoring the potential for policy free riders, who benefit from receiving the subsidy for projects which would have been built irrespective of it.

A second important source of additionality concern arises because the original Wind Power Production Initiative (WPPI) subsidy is included in the Reference Case. Any emissions savings resulting from projects financed under this initiative (as opposed to those financed under the expanded WPPI from the 2005 Budget) would be double-counted if included in the Plan. As of 2007 Canada has seen significant growth in wind power generation, and a continuation of this pattern is part of the Reference Case: wind generation is expected to increase from 1 TW-h in 2004 to 9 TW-h by 2010 and to 24 TW-h by 2020. This represents a Reference Case for growth in wind power of 17% per year.

Table 6: Use of Information Provided for the RPPI *

Year	2005	2008	2009	2010	2011	2012
Reference Case Wind Production 12 (*=interpolated)	1.8	4.76*	6.58*	9.1	10.6*	12.5*
Provided Generation Total Numbers		4.7	8.0	11.7	14.3	14.3
Total Emissions Reductions (multiply total generation by a factor of 0.4564 Mt/TW-h)		2.2	3.7	5.3	6.5	6.5
Stated Emissions Reductions from the Plan		2.2	3.7	5.5	6.7	6.7
Incremental Emissions Reductions (multiply difference between Reference Case and Provided Generation by a factor of 0.4564 Mt/TW-h)		0	0.6	1.2	1.7	1.7

* This table was constructed by the NRTEE using numbers provided by Natural Resources Canada.

Figures provided by NRCan suggest that emissions reductions are in fact stated for all renewable generation, or at least all generation that would be eligible for RPPI or WPPI financing. Table 6 above shows the figures provided, the Reference Case for wind power (other renewables are taken to be negligible and biomass is not included), and the calculations that led to the stated reductions.

To interpret the table, consider that in 2010, Reference Case generation is to be 9.1TW-h from wind, while figures the NRTEE was provided implied 11.7 TW-h of total production from renewables, for which stated emissions reductions are 5.5 Mt, with the ratio between the two being exactly equal to the provided conversion factor (up to rounding errors). Consider also that the implied incremental production from renewables in 2010 is 2.7 TW-h (the difference between the Reference Case of 9 TW-h and the production forecast of 11.7 TW-h provided by NRCan). Based on the conversion factors provided, rather than achieving the 5.45 Mt of stated reductions, this incremental production would be equivalent to emissions reductions of 1.2 Mt. The key assumptions used above are that the provided figures do not correspond to an estimate of incremental generation. This would imply that total generation from wind and other renewables would increase two-fold under the policy, to a total of 21.8 TW-h per year. While this would negate the discussion of additionality above, such an assumption would seem unrealistic and so would be subject to a different set of critiques.

Conclusions

The evidence reported above suggests that the figures in the Statement do not represent incremental reductions in GHG emissions that will occur relative to those already accounted for in the Reference Case. The projections are a reflection of the aggregate displacement of GHG emissions associated with projects financed under the existing WPPI and the enhanced RPPI, and so would represent a double-counting of some reductions. The forecasts suggest some incremental production as a result of the RPPI that is not accounted for in the Reference Case, and emissions reduction numbers would be more accurately based on this incremental production only.

6. ecoAUTO Incentives

Summary of Initiative and Emissions Projections

The Plan describes the ecoAUTO new-vehicle-purchase incentive program, which offers rebates or charges additional fees to new vehicles based on their relative fuel economy. Under the program, purchasers may be eligible for rebates on fuel-efficient vehicles of up to $2000, or be charged fees of up to $4000 on new fuel-inefficient vehicles.

Methodological Approach

The estimates provided in the Plan are derived from a multi-step estimation procedure. First, an estimate of the fuel consumption per kilometre for vehicles being replaced by the rebate vehicles is based on lifetime on-road fuel consumption estimates for either the conventional engine equivalent of a hybrid vehicle (when available), or an average vehicle in one of two classes (cars or light trucks). This rate is then used to compute total consumption of fuel with and without the rebate program, accounting for a 15% rebound effect on kilometres travelled. These total fuel savings are converted to emissions reductions attributable to replacement vehicles. The estimate also accounts for the free rider problem by assuming that 60% of forecast increases in efficient vehicle sales cannot be directly attributed to the rebates.

Analysis

The analysis of the ecoAUTO incentives is very thorough. As specified above, the analysis clearly accounts for estimated rebound and free rider effects.

The data supports the magnitude of the corrections used for the rebound and free rider effects, as estimates both above and below the value of 15% used in the Plan can be found. Kleit (2002) examines the role of fuel- (as opposed to emissions-) efficiency standards in promoting gasoline conservation in the United States. He finds that people use efficient cars more intensively, which negates a significant proportion of the aggregate fuel savings. Specifically, Kleit finds that under a proposal to tighten Corporate Average Fuel Economy (CAFE) standards by 50%, fuel consumption would decrease by only 22% as a result of increased driving, increased congestion, and other factors. Fischer, Parry, and Harrington (2007) examine the welfare basis for tightening vehicle fuel-efficiency standards. They find that a 13% reduction in gasoline consumption can be achieved by a 15% tightening of the CAFE standards. This suggests that the rebound effect may be stronger for higher fuel economy. The rebound effect for vehicles ensuing from tightened fuel-economy standards was also studied by Greene et al. (1999), who found that the rebound effect leads to a long-run take back of about 20% of potential energy savings. The correction for the free rider problem is also consistent with estimates compiled for the NRTEE (NRTEE, 2005).

One issue with these estimates with respect to the current mandate is the way in which emissions reductions are attributed to the subsidy program. The estimates in the Plan represent a reliable estimate of the difference in total lifetime (15 year) emissions attributable to cars purchased under the subsidy program. In other words, for a car purchased in 2010, all emissions reductions-from 2010 through 2025-that will be realized as a result of that purchase are accounted for in 2010. Since these emissions reductions occur incrementally over those 15 years, accounting for them in this way likely overestimates the realized emissions reductions in the early years of the subsidy, which is our period of interest.

Table 7: Emission Reductions Attributed to EcoAUTO Rebate Program

Program	Projected Emissions Reductions in Mt					Key Determinants of Results	Predictive Accuracy
	2008	2009	2010	2011	2012
EcoAUTO Rebate Program	0.1	0.2	0.2	0.2	0.3	accounting for cumulative rather than year-over-year reductions	Likely overestimate

Conclusions

The estimates are reliable in terms of the lifetime impact of sales of vehicles in the years 2008-2012; however, they do not accurately reflect realized emissions reductions over that period. Given the way in which reductions are generally defined in the Plan, results for this policy should state the year-by-year realized emissions reductions for the subsidy program.

7. ecoENERGY Retrofit Initiative

Summary of Initiative and Emissions Projections

The ecoENERGY Retrofit Initiative offers subsidies to owners of homes and small- to medium-sized businesses upon completion of retrofits that verifiably improve the energy-efficiency rating of the building. The Plan projects reductions of 440 kt in 2008 up to 1 Mt in 2012, or roughly 250kt per cumulative-program-year of emissions savings.

Table 8: Emission Reductions Attributed to ecoENERGY Retrofit Program

Program	Projected Emissions Reductions in Mt					Key Determinants of Results	Predictive Accuracy
	2008	2009	2010	2011	2012
ecoENERGY Retrofit	0.4	0.7	0.9	1	1	rebound effect conversion of predicted energy savings to realized emissions reductions treatment of free ridership	Likely overestimate

Methodological Approach

Reductions are calculated based on differences between the forecasted energy consumption with and without the increased energy efficiency associated with NRCan program activities. Forecasted energy savings are then converted to emissions savings using emissions factors. Details on grants provided and realized emissions savings per grant were not provided.

Analysis

As with other programs in the Plan, the ecoENERGY Retrofit Initiative targets energy efficiency rather than energy consumption, and the results of the program are presented in terms of emissions reductions. It is highly likely that in most cases, utility and government retrofit programs have overestimated the impact of their investments on realized energy demand, largely as a result of directly translating potential gains in efficiency to estimated reductions in energy use, ignoring rebound effects, and/or as a result of treating all realized energy use reductions as incremental results of the subsidy programs, ignoring the free rider effect. Each of these is expanded upon below.

The projected reductions in the Plan summarize potential gains in energy efficiency realized through grants. As with any subsidy program, the free rider effect drives a wedge between the number of subsidies paid out and the incremental benefits of the policies relative to the Reference Case. For homeowners obtaining subsidies, there is no way to reward only those who are acting as a result of the program (for homeowners who would have retrofitted their homes regardless, the subsidy represents a windfall) and so only some of the dollars spent under the program actually alter retrofit behaviour. The data suggest that the level of free ridership in the case of retrofit subsidies may be substantial. In Carpenter and Chester (1984), results show that over 90% of homeowners receiving a conservation tax credit for home retrofits would have made the changes without the tax credit. In NRTEE (2005), estimates are reported for free ridership of between 40% and 80% of subsidy recipients. Overall, the academic literature suggests that there is a weak, positive relationship between fiscal incentives and retrofits, but that each dollar remitted through these programs equates to far less than one dollar of new investment. It appears that evaluations of benefits accruing from the EcoENERGY programs include the reductions resulting from all investments for which grants were received, which would overestimate the incremental effects of these programs.

Retrofit subsidies reward retrofits, they do not reward diminished total energy consumption. They provide no disincentive to invest the savings in other energy consuming goods after the grant has been received, and they may provide an incentive to increase the intensity of use or the total number of certain energy durables (the rebound effect). The role of consumer behaviour is very important in determining energy savings from energy efficiency improvements. A more energy efficient home costs the homeowner less per year to keep the furnace at a higher temperature, just as a high-efficiency washer makes it cheaper to wash clothes as discussed above. As with results from Davis (2007) cited above for appliances, results reported by Dubin, Miedema, and Chandran (1986) for a pilot project undertaken by the Florida Power and Light Company in 1981 constitute a relevant example here. The goal of the study was to evaluate how usage patterns changed upon the installation of one of three technology combinations: (1) upgraded attic insulation, (2) upgraded insulation and a high-efficiency air conditioner with a traditional furnace, and (3) upgraded insulation and a high-efficiency heat pump. The key distinguishing characteristic of this study is the ability to monitor changes after random assignment of technologies, and to compare these changes to a control group. The key results of the study show that usage increases substantially after the installation of new technologies. In particular, they find that actual energy gains from new cooling technologies would be as much as 13% below engineering estimates on average, but only 1%-2% below for peak summer cooling (where the air conditioner is used all the time). For heating, energy savings 8%-12% below engineering estimates were found. Studies undertaken by Hausman (1979), Dubin and McFadden (1984), and Bernard, Bolduc, and Belanger (1996) provide similar results.

In contrast to the evidence above, the estimates in the Plan directly translate forecast energy efficiency gains into emissions reductions, without explicitly accounting for rebound effects. The resulting emissions reductions will therefore likely be overestimated.

Evidence from similar programs in Canada supports these conclusions. Over the period 1998-2006 under the similar EnerGuide for Houses program, 270,000 audits were performed, with approximately 180,000 of these performed after retrofit grants were introduced in 2003. In 2005, 37% of homeowners receiving initial audits received a grant, and among those, the predicted emissions savings resulting from renovations was 4 tonnes, similar to projected energy savings for the EcoENERGY Retrofit. However, these estimates were based on the homeowners undertaking all recommendations.¹³ In fact homeowners undertook fewer and smaller changes than recommended by the audits, and thus saved less energy than the amount predicted by the second audit. For homeowners receiving the first sets of energy audits under the EnerGuide for Houses program, the average realized emissions savings was found to be 1.4 tons per household, or less than half of the predicted savings at the time.¹⁴

Conclusions

Given the historical evidence of overestimation of these types of programs, the evident lack of accounting for free ridership and rebound effects and the historical rates of grants-per-budget-dollar and emissions-saving-per-grant, this would likely result in an overestimate of stated emissions reductions from the eventual realized reductions.

8. Encouraging Canadians to Use Transit (EcoMOBILITY Program)

Summary of Initiative and Emissions Projections

Under this policy, persons purchasing monthly transit passes may claim a 15.5% income tax credit for the amount of the pass.

Methodological Approach

The estimates in the Plan are calculated on the basis of assumptions that using transit brings about an average emission savings of 2.8 kilograms per 10-kilometre trip, and that the tax credit is expected to increase urban transit ridership by an average of 5%. This 5% increase translated to 80 million additional public transit rides, which, with the expected average savings, is equated to an emissions reduction of 0.224 Mt.

Analysis

The success of this policy in terms of incremental transit ridership and the success in terms of displacing vehicle trips must be separated. Emissions reductions will be determined by three factors:

• The cross-price elasticity of automobile use to transit price, or the number of additional riders that will be attracted away from cars by a decrease in the effective price of a pass;
  
• The relative difference between the emissions per person kilometre on transit; and
  
• The inter-relationship between the results of this program and the effects of other programs that implicitly reduce the cost and impact of driving.

Table 9: Emissions Reductions Attributed to EcoMOBILITY Program

Program	Projected Emissions Reductions in Mt					Key Determinants of Results	Predictive Accuracy
	2008	2009	2010	2011	2012
EcoMOBILITY Program	0.2	0.2	0.2	0.2	0.2	displacement of vehicle trips is equal to increase in transit use energy savings per trip	Likely overestimate

It is important to consider the cross-price elasticity between personal vehicles and transit as a measure of potential policy effectiveness. We are not as interested in increases in aggregate ridership since each transit ride does not necessarily offset a vehicle trip. Estimates of vehicle-for-transit substitution rates in Elgar and Kennedy (2005) suggest that the 15.5% decrease in the price of transit should be expected to yield a 0.4% decrease in the use of automobiles. However, we must also consider the inverse problem. Voith (1991) estimates a 10% decrease in the fixed cost of auto ownership will decrease transit ridership by 11.3%, while a 10% decrease in the variable cost of an auto trip will decrease transit ridership by 26.9%. These interrelationships are important since the transit pass subsidy, the ecoAUTO rebate program, and the renewable fuels content standards are likely to induce changes in both the costs of transit and the fixed and variable costs of auto trips.

Estimates of the incremental impact of the public transit tax credit in Jaccard and Rivers (2007) find decreases of emissions of 0.15 Mt per year. The fact that this estimate is slightly smaller than that calculated in the Plan is likely a result of the assumption that all new transit trips result in emissions reductions. It may also reflect the fact that, given the other measures imposed in the Plan, vehicles also become slightly cheaper to drive. In both cases the predicted impact of the policy in terms of emissions reduction is very small.

Conclusions

Because of the Plan’s assumption that all new transit trips result in emissions reductions, the Plan’s projected emissions reduction will be slightly overestimated.

9. Information Programs

Information-based programs described in the Plan account for 3.4 Mt per year of emissions reductions. Gillingham et al. (2006) analyze some past empirical evidence on information programs designed to induce investment in energy efficiency, and show that these can appear to be powerful tools. Likely the best-known labelling program is EnergyStar, and the US Environmental Protection Agency credits the use of EnergyStar appliances with savings of up to 80 TW-h of electricity in 2001. Loughran and Kulick (2004) point out that, in some cases, knowledge may present a significant barrier to technology adoption and so ".programs that collect and disseminate information on the costs and benefits of particular energy efficiency investment, demonstrate to firms how they can benefit from energy efficiency investments may be more successful (than financial incentives)" (Loughran and Kulick, 2004, page 39).

However, with few exceptions, little evidence exists through which one can evaluate the incremental effect of information-provision programs for emissions control or energy conservation. With the example of the Energy Star appliances given above, it is difficult to identify the incremental role of the label in driving these decisions. While we can observe peoples’ actions after receiving information, we do not know what information they had before, what they would have acquired through other channels, and what their decisions would have been absent the programs.

In the case of the information-dissemination programs discussed in the Plan, the majority of the individual programs provide few details and the stated emissions reductions are small relative to margins of error that would exist for estimates from past or similar programs. As such, the section below discusses in turn the various measures proposed and suggests means by which their emissions reductions might be validated after the fact.

9.1 ecoENERGY for personal vehicles and ecoENERGY for fleets

In addition to the incentive and regulatory programs discussed above, the Plan attributes small emissions reductions to these education programs for vehicle owners and fleets. It will be difficult to assess the incremental impact of this approach, particularly given that other relatively large programs also raise awareness of fuel-economy concerns. Given the small emissions savings reported, the major concern with this program will be the ability to tie results to spending, since it will be difficult to observe the small projected changes in behaviour in aggregate data.

9.2 ecoENERGY for buildings and houses

The ecoENERGY for buildings and houses program is investing $60 million over four years to encourage the construction of more energy-efficient homes and buildings using ratings, labelling, and training. Programs similar to this have proven very successful in the United States, with a prime example being the US Department of Energy’s Rebuild America program. Estimates show that for each dollar investment in Rebuild America, there were energy savings valued at $18.43 (Gillingham et al., 2006). However, these estimates do not identify the incremental impact of the program. Details on the ecoENERGY for buildings and houses are limited, and so it is not possible to make a direct comparison between these programs.

9.3 ecoENERGY for Industries

The ecoENERGY for Industries program is designed to facilitate information sharing and best practices for energy use. In order to provide a reliable estimate, the analysis must specify the best practices that it hopes to disseminate and the timelines over which these best practices alter behaviour and lead to energy savings. Realizations can then eventually be benchmarked against adoption of particular practices. In the absence of this, it will be difficult to directly and confidently separate program impacts from market forces.

9.4 ecoTECHNOLOGY for Vehicles

The ecoTECHNOLOGY for Vehicles program is investing $15 million over four years to test the safety and performance of more energy-efficient, light-duty vehicles in the Canadian context. What is not clear in this program is the degree to which these measures lie outside traditional government mandates, and to what extent they can be considered incremental. As regulatory requirements for more efficient vehicles become reality, smaller and more efficient vehicles will be introduced into the marketplace. It is not clear from the program exactly which barriers currently exist and in what ways investment will lower these barriers.

Ideally, for this program to be properly evaluated a set of benchmarks against which to measure success should be specified. Market-share projections for particular technologies should be adopted before the program is launched, and these should be compared at regular intervals to realized adoptions. Secondly, a comparison should be made between adoption rates in other jurisdictions with similar conditions and no government promotion of the same technologies.

9.5 ecoMOBILITY

Another program aimed at changing the decision-making process of Canadians is the ecoMOBILITY initiative: a $10 million program to help develop products and services that make it easier for Canadians to change their behaviour. Again, there are few specifics provided on the nature of this program, or how it integrates with other initiatives discussed above. In order to be able to properly assess this program, it is necessary to provide clear results that include information on the following:

• Regions affected by the programs,
  
• Previous predictions of ridership, kilometres driven, or other benchmarks against which to evaluate the program, and
  
• Realized changes in ridership in municipalities/regions covering regions that benefited from investment under the program and regions that did not.

These recommendations for future reporting echo the recommendations of Transport Canada (2005).¹⁵

9.6 Marine Shore Power Program

The Marine Shore Power Program will invest up to $6 million to demonstrate the use of shore-based power for vessels in Canadian Ports. The goal is to demonstrate how vessels can reduce emissions by adopting shore-based power. However, the eventual adoption outside the demonstration program depends on key assumptions being satisfied:

• It must be cheaper for vessels to power using on-shore power than by idling engines.
  
• Infrastructure must be present for sufficient power to be provided for on-shore powering of vessels.
  
• Current lack of use of on-shore power must be due to a lack of information about its applicability.

In the absence of these conditions, it must be the case that vessel owners are currently ignoring a means of reducing fuel costs, which seems unlikely. Since the policy provides no additional incentive in terms of fines, penalties, or rewards, it must be shown that a lack of information or infrastructure exists that the program will provide.

9.7 Synopsis and Reasons for Lack of Evaluation

In general, information programs are very hard to evaluate. It is very difficult to collect data on what people or companies do not know. Further, it is difficult to identify the source of information used to make a purchase or investment decision, since information has significant spillover across markets. As such, we are less likely to be able to draw conclusions only based on differences between areas benefiting and not benefiting from the programs. Further, since information is changing, it is difficult to predict which sources of information will be used in future programs and how these will benefit Canadians and lead to emissions reductions. Even if such information were to be available, it is difficult to assess how the adopted technologies would have evolved in the absence of the information programs, or how non-adopted technologies would fare in the presence of more aggressive information programs.

Table 10: Summary of information-based emissions reduction programs

Program	Projected Emissions Reductions in Mt					Key Determinants of Results	Predictive Accuracy
	2008	2009	2010	2011	2012
ecoENERGY for Personal Vehicles	0	0.1	0.1	0.1	0.1	consumers are not fully aware of the consequences of their driving behaviour	Insufficient information to reach a conclusion
ecoENERGY for Buildings and Houses	0.6	0.9	1.2	1.3	1.3	lack of regulatory backstop in building codes unclear whether existing practices would be adopted	Insufficient information to reach a conclusion
ecoENERGY for Commercial and Industrial Buildings	0.2	0.3	0.4	0.4	0.4	percentage of free ridership	Insufficient information to reach a conclusion
EcoTECHNOLOGY for Vehicles	0.2	0.4	0.5	0.7	0.9	lack of information versus lack of fiscal incentives explains some of current driving behaviour	Insufficient information to reach a conclusion
ecoENERGY for Fleets	0.2	0.3	0.5	0.5	0.5	managers are now aware of the cost of energy and the ways to incite cost-reducing behaviour among employees	Insufficient information to reach a conclusion
EcoFREIGHT	0.4	0.7	1.2	1.2	1.3	separation of program-induced benefits from natural improvements in energy efficiency	Insufficient information to reach a conclusion
EcoMOBILITY	0.9	1.2	1.6	1.7	1.7	current barriers exist and can be reduced through information rather than fiscal or regulatory incentives	Insufficient information to reach a conclusion
Marine Shore Power Program	Less than 0.1 Mt total					infrastructure costs will be paid back in savings at market energy prices potential cost savings exist but there is a lack of information about these savings	Insufficient information to reach a conclusion
Eco-AGRICULTURE	No Specific Commitment						Insufficient information to reach a conclusion
Total	1.7	2.7	3.9	4.2	4.5

10. Clean Air and Climate Change Trust Fund

Summary of Initiative and Emissions Projections

Under the $1.5 billion Clean Air and Climate Change Trust Fund, a series of third-party trusts have been established to directly support provincial and territorial efforts to reduce emissions.

Methodological Approach

Annual emissions reductions of 16 Mt were attributed to the $1.519 billion Clean Air and Climate Change Trust Fund. Although information provided to the NRTEE by Environment Canada suggests that these were estimated on the basis of stated emissions reductions from the Province of Quebec, specific details were not available to us. However, the NRTEE notes that details of all provincial activities to be undertaken as a result of the Fund have yet to be determined. The Government of Quebec’s June 2006 climate change plan indicated that federal funding of $328 million would generate additional emissions reduction of 3.8 Mt per year, on average, for the years 2008-2012. An extrapolation of these numbers to the total budgeted amount would yield an estimated annual total reduction of 17.5 Mt, so a direct extrapolation was not used. Specific details on the method of accounting for provincial emissions reductions realized as a result of federal programs were not provided.

Analysis

Details are not provided on the relative contribution of specific measures to the 16 Mt total, as a result only limited analysis is possible.

The investment of $1.519 billion is credited in the Plan with generating 80 Mt of total emissions reductions, assuming that none of the programs have results beyond 2012. If this assumption holds, the rate at which emissions reductions are achieved averages $19 per tonne. If some policies have longer-term results, the average dollars per tonne will be less than $19. Modelling completed by the NRTEE in the context of the present study suggests that the total emissions reductions from Canadian industry, households, and the transportation sector would be 16-20 Mt if an emissions price of $19 per tonne were imposed. This figure represents, arguably, the most cost-effective way of achieving a comparable number of emissions reductions. However, it represents all reductions available.

While this is not a perfect proxy for the role of the Clean Air and Climate Change Trust Fund, it suggests only two eventualities. First, the incremental emissions reductions stated for provincial programs are likely overestimated, or second, the reductions counted for the Clean Air and Climate Change Trust Fund include reductions that occur as a result of all programs, federal and provincial. In this case, the stated 16 Mt would not be additional to other reductions estimated in the Plan.

Table 11: Emission Reductions Attributed to the Clean Air and Climate Change Trust Fund

Program	Projected Emissions Reductions in Mt					Key Determinants of Results	Predictive Accuracy
	2008	2009	2010	2011	2012
Clean Air and Climate Change Trust Fund	16	16	16	16	16	additionality of provincial reductions	Insufficient information to reach a conclusion

Conclusions

The nature of some of the provincial programs suggests that issues of additionality exist. For example, the Quebec plan (as communicated to the NRTEE by Environment Canada) includes reductions in GHG emissions due to projects funded under the WPPI and due to the 5% ethanol content standard. It also sets targets for Quebec industries that will already be affected under the Regulatory Framework for GHG Emissions. Without a fully integrated model that includes these transfers to the provinces, the federal policies, and the provincial policies, it is difficult if not impossible for the NRTEE to attribute incremental emissions reductions to each separately and therefore insufficient information is available to conclude the likelihood of emissions reductions at this time.

« Previous — Contents