Greenhouse Gas Emissions Forecasting: Learning from International Best Practices

Description of Additionality, Free-ridership, Rebound Effect, and Policy Interaction Effects

There are four key reasons the emissions reductions were overestimated in the Government of Canada’s 2007 KPIA Plan. First, the estimates of the reductions generated by the various initiatives suffered from biases related to additionality (including additionality concerns resulting from a lack of accounting for free-ridership). Second, the emissions-reduction factors used in the calculations were, in some cases, not consistent with recent scientific evidence. Third, rebound effects were not always taken into account in the estimates. Finally, policies were treated independently, so policy-interaction effects are ignored.41

Problems of additionality arise when the stated emissions reductions do not reflect the difference in emissions between equivalent scenarios with and without the initiative in question. If emissions reductions from an initiative have already been included in the reference case, these emissions reductions will be double-counted.

A key source of additionality issues that arises frequently and so was treated separately in the NRTEE’s 2007 KPIA analysis is the failure to account for free-ridership. Free-ridership is not properly accounted for when stated reductions include the results of behaviour that is rewarded but not influenced by the policies. This can occur when subsidies are paid to all purchasers of an item, regardless of whether they purchased the item because of the subsidy. Those who would have purchased the product regardless are termed free-riders, and their behaviour (since it would have happened regardless of the policies) has already been accounted for in the reference case. Not correcting for this implies that induced emissions reductions will be overestimated by the proportion of free-riders, which has been estimated to be between 40% and 80% (NRTEE, 2005).

The rebound effect describes the increased use of a more efficient product resulting from the implied decrease in the price of use: for example, a more efficient car is cheaper to drive and so people may drive more. While estimates vary, emissions reductions will generally be overestimated by between 5% and 20% if estimates do not account for increased consumption due to the rebound effect.

The relative successes of emissions-control policies will be interdependent, and an evaluation framework that takes this into account is important for proper interpretation of stated results. The 2007 Government KPIA Plan provides results from separate evaluations of individual policies, while these are slated to be imposed simultaneously. This approach omits any policy interaction effects and will only be accurate when the sum of all individual policy effects is equal to the total effect of all policies, which is not likely to be the case. A general finding of the NRTEE’s 2007 KPIA Response, which is consistent with the statement above, is that in order to deliver a statement of total expected emissions reductions, all policies should be imposed simultaneously in a modelled economy.